Blog

How to determine population and survey sample size?

652 comments

A survey can only be truly valuable when it’s reliable and representative for your business. However, determining the ideal survey sample size and population can prove tricky. In other words, who will you be surveying and how many people? No idea? No worries. We’re here to help!

Say you’re a market research manager at a furniture company and you are planning to launch a new furniture line by the end of 2016. However, before you launch the new line you wish to conduct an online survey on whether your line ‘Fall – 2016’ is more or less likely to be a hit or miss on the European Union (EU) market. So far, so good. Yet, the following question will almost instantly arise: “What is the population that I would like to survey?”. Or, who do you need to survey to gain valuable insights in the success of your new furniture line? In this case the answer is rather straightforward. Assuming that you are launching the new line on the European market, that minors do not buy furniture and that your furniture is reasonably priced, your population consists of all adults in the EU.

What is the survey sample size?

For obvious reasons it is impossible to survey those (roughly) 400 million adults in the EU. A sample of adults living in the EU offers the solution for this issue. A sample is a selection of respondents chosen in such a way that they represent the total population as good as possible. However, instantly a new question comes to the forefront: “How many people should my sample consist of?”. Using a correct survey sample size is crucial for your research. After all, a sample that is too big will lead to the waste of precious resources such as time and money, while a sample that is too small will not allow you to gain reliable insights.

So, how large should your sample be? Should you survey 1%, 5%, 10%, … of the adult citizens in the EU? Well, this depends largely on how accurate you want your survey data to be. In other words, how closely you want your results to match those of the entire population. There are two measures that affect the accurateness of the data.

  • First of all there is the margin of error (or confidence intervals). In short, this is the positive and negative deviation you allow on your survey results for the sample. Or, in other words, the deviation between the opinions of your respondents and the opinion of the entire population. An example will shed some light on this statistical explanation. Suppose you set your margin of error on 5%. If – let’s hope so! – 90% of your survey respondents like the ‘Fall 2016’ line, a 5% margin of error means that you can be ‘sure’ that between 85% (90%-5) and 95% (90%+5) of the entire population actually likes the ‘Fall 2016’ line.
  • Second there is the confidence level. This tells you how often the percentage of the population that likes the ‘Fall 2016’ line actually lies within the boundaries of the margin of error. Or, following on our previous example, it tells you how sure you can be that between 85% and 95% of the population likes the ‘Fall 2016’ campaign. Suppose you chose the 95% confidence level – which is pretty much the standard in quantitative research1 – then in 95% of the time between 85% and 95% of the population likes the ‘Fall 2016’ line2.

How many respondents does your survey require?

Once you have decided how accurate you want your sample data to be, you can start calculating how many respondents (people who have completely filled in the survey or completes as we call them at CheckMarket) you actually need.

Below you find an indicative table on how to calculate your number of completes. Remember that your population consist of approximately 400 million adults in the EU. As a consequence, the appropriate number of completes will be found on the last row of the table below. Depending on the confidence level and the margin of error, the number of completes will vary. As we chose a margin of error of 5% and a confidence level of 95% for our ‘Fall 2016’ campaign, you need approximately 400 completes (it is advisable to round to the nearest hundred) for your sample.

Alternatively, on the CheckMarket website, you find an easy sample size calculator to calculate the number of completes…

estimate_population_survey_sample

 

What about response rate?

Before you start sending out your survey to 400 respondents, remember there is such a thing as response rate. Response rate is the ratio of respondents that fill in the questionnaire they received compared to the total number of surveys you send out. For instance, if you send out your survey to 400 people and you receive 200 filled in surveys, your response rate is 50%.

For an online survey, conventionally, a response rate of 20% is considered as a good response rate, while a 30% response rate is considered to be really really good. As we calculated that we need 400 completes, this means that you will definitely have to send the survey to more than 400 people in order to reach those 400 completes. Obviously, you cannot predict beforehand what response rate you will achieve. However, assuming that your survey will achieve a response rate of 20%, we divide the objective of 400 completes by a response rate of 20%. As a consequence, you will have to send your survey to approximately 2.000 adults in the EU.


1 In some quantitative research, stricter confidence levels are used (e.g. the 99% confidence level)
2 To put it more precisely: 95% of the samples you pull from the population.

 

Build your survey now

Calculate your own sample size using our online calculator

 

Related Articles

652 comments

Join the conversation
  • Abdul - November, 2017 reply

    If we have around 60000 population. Divided in 2 clusters. School ( Max pop 53000) college (5000) and university(pop 2000). I distributed 55% sample to schools 30% to colleges and 15 % to universities.. Any problem in it?

    Maarten Marijnissen - November, 2017 reply

    Hi Abdul,

    If you want your sample to be fully representative, you’ll need to recalculate these percentages.
    You might want to make statements about all clusters separately and about the total population.
    If that’s the case, the sample sizes of the clusters need to be representative and accurately represent these groups in the total population.

    Good luck with your research!

  • mac - September, 2017 reply

    is there any need to sample if your total population is 100,if yes , what would be the sample size

    Maarten Marijnissen - September, 2017 reply

    Hi,

    For this population the ideal sample size is 80 (using a 5% margin of error and a 95% confidence level), but it’s best to invite all 100 people to fill in your survey.
    You can calculate the margin of error afterwards using the second part of our online calculator.

    Good luck!

  • marc - August, 2017 reply

    Hi, i’ve been trying to use the formula you presented in this site and my question is, does the sample size get pegged at 385? i tried using your SSadjusted = (SS) / 1 + [(SS – 1) / population] formula to adjust the SS based on the population but it still registers about 385… may i request some help on this?
    the study i am conducting is a survey on prevalence of drug use in junior high school students, the population size is 6,857,110 what is the sample size i need for this? (95% level of confidence)
    advance thanks for the help

    Maarten Marijnissen - August, 2017 reply

    Hi Marc,

    If your population is this large, the exact size doesn’t matter that much. Your sample size will always be around 385.
    E.g. if your population is 25,000 sample size will be 379 (for confidence level 95% and margin of error 5%). If your population increases to 2,500,000 is will still be no more than 385.

    Good luck with your research!

  • Marie - August, 2017 reply

    Hi. I am initiating a study on cognitive development for a certain country in children between 6 and 14 years old who are given a test set. In addition, it is desired to calculate the average score obtained by age. The country has 19 states. Each age has approximately 50,000 children giving a total of 450000. So what would be the best way to select the sample? I mean? how many states would be enough? What would be the right size per age? If I did it by age I would give 384 * 9 = 3456 which seems a lot in terms of cost. But if I considered the population as a whole I would give 384, which would give me about 40 per age group. This seems little to make population inferences. So what would be the right reasoning? Thank you very much for your help

    Gert Van Dessel - August, 2017 reply

    Hi Marie,

    It all depends on which subgroups you want to take to analyse your data.
    You do not necessarily have to use 9 different age groups. It is perfectly possible to make a division of 3 age groups (6-8yrs, 9-11yrs, 12-14 yrs).
    In that way you would only need 3 * 384 = 1152 respondents.
    Ideally you would use a representative split over all 19 states to pick your sample. But you could also regroup the states in wider regions (e.g. North-East, North-West, South-East, South-West) and look for a representative split across the regions.
    It all depends on the ressources you have (time, budget, people, …) and the margin of error that is acceptable for your research.

  • Pit - July, 2017 reply

    Hello, i think this will really help me determining my sample size. my problem is to determine the student satisfaction in a school’s offices. my population is 1200. how much sample do i need?

    Maarten Marijnissen - July, 2017 reply

    Hi,

    You can calculate the ideal sample size using our online sample size calculator.
    For a population of 1200 with a 95% confidence level and 5% margin of error the sample size will be 292.

    Good luck with your research!

  • Zaynah - July, 2017 reply

    Hello, I am doing my dissertation right now. i have to conduct my survey but i am so stuck. It is based on a population of 303527. i will be using stratified random sampling. i am actually using total number of students studying in my country which is 303527 based on recent stats. then i have studying population for tertiary level which is 55706 and population of students from my university is 4750. but i dont know how to calculate my sample size.

    Maarten Marijnissen - July, 2017 reply

    Hi Zayna,

    When you want to draw significant conclusions about all populations, you need to calculate one sample size for the entire population with sub sample sizes for the smaller populations.
    If you do that, you can compare your different populations afterwards.

    You can easily calculate sample sizes using our online calculator.

    Good luck!

  • onyi - May, 2017 reply

    If my population size is 2000 and there are 70 items on my research instrument. what will my sample size be?

    Maarten Marijnissen - May, 2017 reply

    Hi,

    You can calculate your sample size using our online calculator.
    For a population of 2,000 with a 95% confidence level and a 5% margin of error your sample size will be 323.

    Good luck with your research!

  • Richard Ayandele - April, 2017 reply

    Hi,
    I have a population of 116,681. What will be my sample size.
    Kindly avail me the formula and the step by step of how to arrive at the sample size.
    Thanks

    Maarten Marijnissen - April, 2017 reply

    Hi Richard,

    Your required sample size will be 383 for your population with a 5% margin of error and a 95% confidence level.
    Here’s the formula:

    SS = (Z-score)² * p*(1-p) / (margin of error)²
    SS = (1.96)² * 0.5*(1-0.5) / (0.05)²
    SS = 3.8416 * 0,25 / 0.0025
    SS = 384.16

    (Z-score is 1.96 for a 95% confidence level)

    Then you need to adjust it to your specific population.
    SSadjusted = (SS) / 1 + [(SS – 1) / population]

    Good luck with your research!

    Richard Ayandele - April, 2017

    Thank you very much for the response.

    Robert - April, 2017

    Maarten,

    Does this calculation change for attribute samples? I have a population of 2800 and I want to know if anything in the population “passes” or “fails” without sampling the entire population. What would my sample size need to be for a one-tailed test at a 99% confidence level and 2% margin of error?

    Thank you

    Maarten Marijnissen - April, 2017

    Hi Robert,

    We looked it up for you. The calculation doesn’t change for those kind of samples.
    Your ideal sample size for a population of 2800 with a 2% margin of error and a 99% confidence level would be 1672.

    You can calculate it using our online sample size calculator.

    watson - July, 2017

    i have a household population of 3755, what can be my sample size?

    Maarten Marijnissen - July, 2017

    Hi Watson,

    You can calculate this using our sample size calculator.
    Just fill in the population size, margin of error (usually 5%) and confidence level (usually 95%) and our calculator will show you your required sample size.

    Good luck!

  • chacha matoka - April, 2017 reply

    Hi
    I have population 2,872 . category A= 42, category B=99, category C=222, Category D = 516, category E = 1,549 and category F = 444.
    1. What is the sample size
    2. should I calculate sample for each category or just one single sample to the population?

    Maarten Marijnissen - April, 2017 reply

    Hi,

    If you want to compare these categories, you should calculate representative sample sizes for each category.
    You can do this using our online calculator.

    Good luck!

  • Ritesh - March, 2017 reply

    Hy
    I have being doing research work for commercial banks
    I have a population size of ;28
    and i need to calculate sample size with such small population. How should i conduct?

    Maarten Marijnissen - March, 2017 reply

    Hi Ritesh,

    Sample sizes for small populations are conducted the same way large populations are conducted.

    If your population is that small, it’s best to invite all of them to complete your survey.
    Your ideal sample size, calculated with our sample size calculator, is 27.

    Good luck with your research!

    Ritesh - March, 2017

    Thank you so much Mr Marteen

    Ali Iftikhar - March, 2017

    Hi Maarten:

    I am working on my thesis proposal and survey sample size. I am evaluating three academic buildings, and determining the total population of each building right now. Population of one building would be around 1200. I was thinking 30% of the total population which is 360 and that seems really high given that its not an online survey, and respondents are students.
    360 x 3

    Please let me know what you think

    Thanks,

    Maarten Marijnissen - March, 2017

    Hi Ali,

    If you get 360 responses per building the margin of error for a 95% confidence level will be below 5%, which is great.
    For your total population the margin of error will be 2.5%.
    You can calculate this yourself when your responses are in using our online calculator.

    This is definitely a sample size you can work with.

    Good luck with your research!

    Maarten Marijnissen - March, 2017

    Hi Ali,

    If you get 360 responses per building the margin of error for a 95% confidence level will be below 5%, which is great.
    For your total population the margin of error will be 2.5%.
    You can calculate this yourself when your responses are in using our online calculator.

    This is definitely a sample size you can work with.

    Good luck with your research!

  • Mark - March, 2017 reply

    Hi
    I’m glad that you guys are there to help for our researches.
    So, I have population of 353,534 and its says in the sample size calculator that the required sample size is 384 with a margin of error of 5% and a confidence level of 95%. I am just wondering how the calculator resulted to a 384 sample size. May I know the step by step process so I could show my professor how I got such sample size? and what is the name of the formula and the author of such. Thank you. Hope you could reply the soonest!

    Maarten Marijnissen - March, 2017 reply

    Hi Mark,

    We’re happy to help! :)
    The formula is by Cochran and this is the calculation:

    The first step is to calculate the sample size for infinite populations.
    SS = (Z-score)² * p*(1-p) / (margin of error)²
    SS = (1.96)² * 0.5*(1-0.5) / (0.05)²
    SS = 3.8416 * 0,25 / 0.0025
    SS = 384.16

    (Z-score is 1.96 for a 95% confidence level)

    Then you need to adjust it to your specific population.
    SSadjusted = (SS) / 1 + [(SS – 1) / population]
    SSadjusted = 384.16 / 1 + ((384.16 – 1) / 353,534)
    SSadjusted = 384.16 / 1.001084
    SSadjusted = 383.74 => 384

    Good luck with your research!

  • Sidra - February, 2017 reply

    Hello! I just stumbled upon this forum and it seems it might save me. I’m working on my Mphil thesis. I need to determine sampling technique and sample size. My objective is to determine satisfaction of mothers with maternity care services in public and private hospitals of Lahore, Pakistan. I have designed the sampling in such a way that I ll be choosing hospitals from both public and private sectors in equal proportions. Then I ll proceed with choosing subjects from the selected hospitals using systematic random sampling.. The sampling frame is the list of women who have delivered at selected hospitals within last six weeks, for every hospital. Now what I’m not sure about is whether I have to determine sample size for each hospital individually using the formula SS = (Z-score)² * p*(1-p) / (margin of error)² , where sampling frame would serve as N (for sample size correction) OR I have to determine just one sample size using the above formula and then distribute that number among all hospitals proportionate to the N for each hospital? A, comprehensive answer would be highly appreciated. Thanks

    Maarten Marijnissen - February, 2017 reply

    Hi Sidra,

    It depends. If you want to draw conclusions for separate groups and you want to compare results between the selected hospitals, or between public and private hospitals, you have to make sure every subpopulation has a representative sample size. If you are using one general population, you only have to calculate 1 representative sample size.

    Either way, you have to make sure the proportions within the sample match the proportions of the population.
    A representative sample size doesn’t guarantee a representative sample in general.

    Good luck with your research!

  • abeer - February, 2017 reply

    hi.
    please guide me how to calculate my sample size if i am doing a quantitative study on cosumer buying behavior and my population size is 3,117,000 in MULTAN DISTRICT of year 2015 according to (http://pcgip.urbanunit.gov.pk/docs/ADPDocumnets/ConsolidatedMultan_ADP.pdf) .what should be my sample size as my study is limited to only females of multan.how should i distribute my quesstionnaire.? there is no latest estimate of multan city population .

    Maarten Marijnissen - February, 2017 reply

    Hi,

    You can calculate your sample size using our calculator.
    Just fill in the population size (female population of Multan) and choose the margin of error and confidence level (usually 5% and 95%).
    Using these data your sample size will be 385.

    You can distribute your survey in many different ways. When you don’t have email addresses you can distribute it via social media or you can ask people in shopping streets, etc.
    These are the distribution methods of our survey tool.

    Good luck with your research.

  • Ambes Shimellis - February, 2017 reply

    Hi Gert, thank you for your support, but who is the author of the formula listed here? Just citation.

    Maarten Marijnissen - February, 2017 reply

    Hi Ambes,

    The author of the formula we use to calculate sample sizes is Cochran.
    The citation is: Cochran, W. G. (1977). Sampling techniques (3rd ed.). New York: John Wiley & Sons.

    Good luck with your research!

  • Otilia - January, 2017 reply

    Hi, nice post.
    I have a question. What is the effect size that the calculator is assuming?
    Or where can I find the formula?
    Thank you!

    Maarten Marijnissen - January, 2017 reply

    Hi Otilia,

    We’re glad you like our post!

    Our calculator assumes a population proportion of 50% (p=0.5).
    The formula that is used is the following:
    SS = (Z-score)² * p*(1-p) / (margin of error)²
    SSadjusted = (SS) / 1 + [(SS – 1) / population]

  • edge - January, 2017 reply

    good day! thanks for the calculator. but I’m having a really hard time to make it into a solution which I want to show to my professor. the given formula which I see it through comments is a standard sample size of 384.16 or 385 but I want to know if my population is 9,068. what can I do to make a feasible solution or equation format. thanks a lot for the quick response or reply.

    Maarten Marijnissen - January, 2017 reply

    Hi,

    The standard sample size of 385 is for infinite populations.
    When the population is known (9,608 in this case) you can calculate an adjusted sample size.

    Your ideal sample size with a confidence level of 95% and a margin of error of 5% is 369.
    To calculate this yourself, you can fill in the population size in our online calculator.

    Good luck with your survey!

  • Christey - December, 2016 reply

    I am doing a research on teachers with population of 1,872 teachers. My problem is, I don’t know what is the number of respondents do I need for each school. You see, each school has different number of teachers, For example, school X: 133 teachers, school Y: 68 teachers, school Z: 96 teachers (and the list goes on for 24 more schools), so I want to know how can I get number of respondents for each school so that the data can truly represent the whole population. Thanks

    Maarten Marijnissen - December, 2016 reply

    Hi Christey,

    If you want to compare all schools in the results, you need to use a representative sample size for each school.
    You can calculate these sample sizes using our online calculator.
    Fill in the number of teachers per school, select the margin of error and confidence interval (usually 5% and 95%), and our calculator will do the rest.

    If comparing different schools isn’t that important, you can just use the total population of teachers to calculate your sample size.

    Good luck with your research!

  • Peace Harriet Simfukwe - December, 2016 reply

    Am doing a qualitative research on “An assessment on the factors that affect the the reduction of flood impacts on sanitation in rural areas: A case of Sekeni village in Chikwawa, Malawi”. my sample population is 1050 and am having troubles on what formula should i use to find my sample size. what is confusing me most is that when read some other books they say that a qualitative sample should at least have 50-60 respondents but after doing calculations with different sample size formula the answers i get are more than 200. Is this possible and what formula is suitable for my research?

    Maarten Marijnissen - December, 2016 reply

    Hi,

    There are no specific sample size formulas or calculations for qualitative sample sizes.
    Your sample size should be large enough to get insights in most or all of the perceptions that are important for your research.
    It all depends on the specific project en the method you are using.

    Further information can be found in this article.

    Good luck with your research!

  • Chidi - December, 2016 reply

    Hello,i am conducting a small reseach and my problem is how to calculate or find out my sample size to use.
    My entire research population is 120.What method of sampling should i use?Is there a particular formula to use for such a small population?Thank you.

    Maarten Marijnissen - December, 2016 reply

    Hi,

    The formula is the same for all populations:
    SS = Z² x p x (1-p)/M²

    with
    SS = Sample Size for infinite population
    Z = Z value (e.g. 1.96 for 95% confidence level)
    P = population proportion (expressed as decimal) (assumed to be 0.5 (50%)
    M = Margin of Error at 5% (0.05)

    After calculation of sample size you have to correct for the total (estimated) population
    SSadjusted = (SS) / (1 + [(SS – 1) / population])

    You can calculate your sample size using our online calculator.
    With a confidence level of 95% and margin of error of 5% the ideal sample size is 92.

  • UMORU, Mohammed Lawal - November, 2016 reply

    Iam carrying out research on Effect of Relationships Education Training on Enhancement of Marital Responsibilities and Success among young couples. how can i determine my population? if im properly guided on how to get the population, i can determine the sample from your explanation so far. thank you. from umoru, mohammed lawal

    Maarten Marijnissen - November, 2016 reply

    Hi,

    The population is the set of elements you want to draw conclusions about using a sample. The population size in your research should be an estimation of the number of young couples in a certain area (country, city, …). There are no formulas or calculations to know your population size.

  • Paula - November, 2016 reply

    Am conducting a study in a small organisation with not more than 37 staff and I thought I would consider the population as my sample size. However, my supervisor thinks 37 is still too small a number to qualify my study as valid. Is there any authority I can quote to justify a sample size of 37 as valid for a quantitative study?

    Maarten Marijnissen - November, 2016 reply

    Hi Paula,

    If you only want to draw conclusions about your staff (37 people), your sample size is definitely valid.

    Good luck with your research!

  • shey - November, 2016 reply

    sir how to get the sample size of 36 using stratified random sampling.. tnx.

    Maarten Marijnissen - November, 2016 reply

    Hi,

    Is your population size 36? Then I would suggest to survey your entire population.
    The sample size with a 5% margin of error and 95% confidence interval is 33.

    You can calculate this using our sample size calculator.

  • Mir Rafi - November, 2016 reply

    I have a study population of 98,000 and according to your checker sample size would be 370. but I want use a study sample of 200, Is it be questionable? What would you suggest? Should I decrease it or go for 370?

    Maarten Marijnissen - November, 2016 reply

    Hi,

    Using the second part of our calculator you can calculate the margin of error of your sample.
    This percentage tells you how much the opinions of your sample are likely to deviate from the population.
    With your sample and population size the margin of error is 6,92%.

    You can use a sample of 200 keeping the margin of error in mind, but a larger sample will provide more statistically significant results.

    Mir Rafi - December, 2016

    Thank you!
    If I make it 6.92%, and sample size 200, is it correct? I mean I can do this, no one can make a objection why I increase the margin of error . .

    Maarten Marijnissen - December, 2016

    That’s correct! If you mention the margin of error in the reporting, there’s nothing ‘wrong’ about it.

  • Om nona - November, 2016 reply

    Good morning
    My study is hospital based case control study .The population under study is about 5949 how can I determine the sample size.
    Thanks in advance

    Maarten Marijnissen - November, 2016 reply

    Hi,

    You can determine the sample size using our sample size calculator.
    Fill in the population size, the margin of error you want to use (usually 5%) and the confidence level (usually 95%).
    Using those numbers your ideal sample size will be 361.

    Good luck!

  • Mousa - November, 2016 reply

    Hello!
    My survey is about e-bussiness Social Network & Cities for Master Degree graduation the link of survey (https://www.surveymonkey.com/r/BFYFQZ8)
    I didn’t determine the population because it’s global I received up to now only 103 responds. Since the survey website ask me to upgrade to analyse more than 100 responses.
    What should be my sample size?
    And what is your advice for continuing with survey monkey or not?
    Best regards,

    Maarten Marijnissen - November, 2016 reply

    Hi,

    For your population the sample size will be 385 with a 5% margin of error and a 95% confidence level.
    You can calculate sample sizes with different confidence levels using our sample size calculator.

    Keep in mind that you need to spread proportionally between different regions to draw conclusions on a global level.
    However, when you want to draw separate conclusions per region with the same confidence level (95%) and margin of error (5%), you will need 385 respondents for each region in your research.

    I would suggest to use a larger sample size to draw significant conclusions, or use a more specific population.

    If you’re not sure about continuing with SurveyMonkey, you should take a look at our survey tool.
    We offer all our advanced features, including fully labeled SPSS export and unlimited support, for all pricing plans.

    Good luck!

  • savi - November, 2016 reply

    Hi there,
    I am working on household consumption pattern for food and non food items . The variable which i have taken in non food are clothing and footwear. now my query is can one individual member represent the entire household’s psychology. or is there any other i can only one individual and not the entire for family as its going to be a very lengthy questionaire depending upon the the size of the family……..

    do we have any references for the same where we can say one individual of the family can be treated as entire family.
    Please revert
    regards

    Maarten Marijnissen - November, 2016 reply

    Hi,

    The number of people you select within households has consequences for the output of your survey.
    It’s best to select all people in selected households, if this is feasible. If not, one person is usually selected.

    You can find more information about different methods with their implications in this article.

    Good luck!

  • Akwasi Akomeah Agyekum - November, 2016 reply

    I want to study 150 people of a population of 1500 farmers. what justification can I give for selecting 150 people of the population of 1500 for the study?

    Maarten Marijnissen - November, 2016 reply

    You can calculate margin of error for your population and sample size using our online calculator.
    This tells you how much the opinions of your sample are likely to deviate from the population.

  • Akwasi Akomeah Agyekum - November, 2016 reply

    I am conducting a survey on about 1500 farmers and want to use a confidence level of 90%. What formular should I use and what will be the sample size?

    Maarten Marijnissen - November, 2016 reply

    Hi,

    The formula to calculate sample size (SS) is: n = Z² x p x (1-p)/M²

    with
    n = Sample Size for infinite population
    Z = Z value (e.g. 1.645 for 90% confidence level)
    P = population proportion (expressed as decimal) (assumed to be 0.5 (50%)
    M = Margin of Error at 5% (0.05)

    After calculation of sample size you have to correct for the total (estimated) population
    SSadjusted = (SS) / (1 + [(SS – 1) / population])

    The ideal sample size for your population and confidence level is 230.

  • jess - October, 2016 reply

    hi. is the maximum sample size be 384 according to Krejcie & Morgan Table? if the population in my survey will be 2 million and can i know what would be my sample size? thanks.

    Maarten Marijnissen - October, 2016 reply

    Hi Jess,

    The maximum sample size in their table is indeed 384 (using a 95% confidence level and margin of error of 5%).
    Once your population is large enough, your required sample size does not change very much anymore.
    You can calculate your ideal sample size using our online calculator.

  • Fiona - October, 2016 reply

    hi. i intend to to survey 400 participants in my research. i would conduct my study at shopping malls, which is unknown population. but the problem is i don’t know what formula to be used to specifically arrive at 400. i hope i would get a response. thank you so much.

    Fiona - October, 2016 reply

    btw, my confidence level is 95%, and margin of error 5%. thank you.

    Maarten Marijnissen - October, 2016

    Hi Fiona,

    Below you find the formula for calculating your sample size for an unknown population:

    First you calculate the sample size (SS). If you do not know your proportion p, it is quite common to take 0,5 as a value for p.
    SS = (Z-score)² * p*(1-p) / (margin of error)²

    Then you correct it if you know your population:
    SSadjusted = (SS) / 1 + [(SS – 1) / population]

    Once your population is large enough, your required sample size does not change very much anymore.
    Your sample size is quite large, because if you look at a population of 100,000 the required sample size is 383 (with a confidence level of 95% and margin of error of 5%).
    If you can estimate your population, you can calculate your ideal sample size yourself with our online calculator.

    I hope this helps!

  • Olanrewaju - October, 2016 reply

    Hi,
    With a population of 40 and confidence level of 95%, how can i calculate an appropiate sample size??

    Maarten Marijnissen - October, 2016 reply

    Hi,

    You can calculate your sample size using our sample size calculator.
    With a margin of error of 5% your ideal sample size for statistically significant results is 37.

  • benedict - September, 2016 reply

    How come you do not require variation of population as an input to determine sampole size?

    benedict - September, 2016 reply

    Let me rephrase my question: your formula applies only whe n estimating proportions? Correct? But what to do when you do not have variance when estimating a mean?

    Maarten Marijnissen - September, 2016

    Hi Benedict,

    That is correct.
    More explanation and further information about determining sample size for estimating the mean can be found in this article.

  • tanveer kaur - September, 2016 reply

    Hello Sir,

    Iam doing a qualitative research, having a total population of 3078 with 655 households. I wanted to know how much sample size will be appropriate for my study and how it can be calculated with a formula. Also, what all variables should be considered?

    Maarten Marijnissen - September, 2016 reply

    Hi,

    There are no specific formulas or calculations for qualitative sample sizes.
    Your samle size should be large enough to get insights in most or all of the perceptions that are important for your research.
    The size of your sample depends on the research project en the method you are using.

    More explanation and further information can be found in this article.

    tanveer - September, 2016

    Thank you so much!

  • marlene - September, 2016 reply

    hello sir, i have 140 population in my research but I only got 50 data gathered from my respondents because some are not responding on my survey. what would i do on this matter because I was not able to achieve my sample size. thank you

    Maarten Marijnissen - September, 2016 reply

    Hi Marlene,

    When you know your actual sample size, you can calculate the margin of error with our online calculator to check how much the opinions of your sample are likely to deviate from the population. You should include this margin of error in your analysis.

  • cindy - September, 2016 reply

    hi! I am conducting a descriptive research, I Have 473 total population of respondents.what would be the sample size needed in my study?

    Maarten Marijnissen - September, 2016 reply

    Hi Cindy,

    You can calculate your sample size with our sample size calculator.
    You need to fill in the population size, the margin of error (= usually 5%) and the confidence level (= usually 95%) and our calculator will do the rest.

  • mchomvu - September, 2016 reply

    am doing a qualitative research where by survey research design was used to a population of 900,000 people. from this population what will be the recommended sample size. is any formula which will guide the sample size

    Maarten Marijnissen - September, 2016 reply

    Hi,

    There are no specific rules or calculations for qualitative sample sizes.
    Your samle size should be large enough to get insights in most or all of the perceptions that are important for your research project.
    So the size of your sample depends on the project en the method you want to use.

    More explanation and further information can be found in this article.

    Kevin - September, 2016

    Hi Maarten,
    I’m working on my Thesis. My population is 423 and I’m required to come up with a sample size of between 50-60. What formula can I apply?

    Maarten Marijnissen - September, 2016

    Hi Kevin,

    The formula to calculate sample size is: SS = Z² x p x (1-p)/M²

    with
    SS = Sample Size for infinite population
    Z = Z value (e.g. 1.96 for 95% confidence level)
    P = population proportion (expressed as decimal) (assumed to be 0.5 (50%)
    M = Margin of Error at 5% (0.05)

    After calculation of sample size you have to correct for the total (estimated) population
    SSadjusted = (SS) / (1 + [(SS – 1) / population])

    You will see that the ideal sample size for this population is bigger than 50-60.

    When your sample size is given, you can calculate the margin of error to check how much the opinions of your sample are likely to deviate from the population.
    You can do this yourself with the second part of our sample size calculator.
    When you fill in your population size, the number of respondents and the confidence level, you will get the margin of error of your sample.

  • Teweldebirhan Girma - September, 2016 reply

    I want to study 125 units of a population of 500. Therefore, what justification can I give for randomly selecting 120 units of the population of 500 for the study?
    From what I taught in statistics that it is acceptable to randomly sample 5% – 30% of the population depending upon the size of the population. Am I correct? Can you suggest me reference documents about this?

    Maarten Marijnissen - September, 2016 reply

    Hi,

    You can calculate this in the second part of our sample size calculator.
    With 120 selected units of a population of 500 the margin or error is 7,81% with a confidence level of 95%.
    This is perfectly acceptable.

  • sunday - September, 2016 reply

    dear sir,
    I am doing research using qualitative methodology with a case study design .but I encounter a problem on stating sample size because there is no actual data on the total number of women food vendor from the municipal , but also very difficult to get data because of the nature of business the perform.
    so i want to know for such kind of the study, how can i determine a sample size?

    Gert Van Dessel - September, 2016 reply

    Sunday,

    If you do qualitative research as you state (e.g. face-2-face interviews or focus groups), you don’t need a minimmum sample size to get valid research.
    Qualitative research is by definition exploratory, which means you do as much interviews as you need to acquire the information you need. Commonly for qualitative research 20 to 30 interviews are sufficient.

  • Nang Aye Aye Mon - September, 2016 reply

    Hello sayar,I have collected a sample of 100 out of total population of 1533.Is it a representative?

    Maarten Marijnissen - September, 2016 reply

    Hi,

    This sample gives you a margin of error of 9,48% for 95% confidence level, which is quite high.

    This number is the plus-or-minus figure usually reported in newspaper or television opinion poll results. For example, if you use a margin of error of 10% and 47% percent of your sample picks an answer, you can be “sure” that if you had asked the question to the entire population, between 37% (47-10) and 57% (47+10) would have picked that answer.

    You can calculate your ideal sample size for yourself with our sample size calculator.
    Fill in population size, pick a smaller margin of error for 95% confidence level and you will get your required sample size.

  • Sherli - September, 2016 reply

    Hello,
    My research topic is height of upper secondary girl students. Could u plz tell me how to determine sample size if don’t give any information (standard deviation and confidence interval).
    Plz reply. Tq

    Gert Van Dessel - September, 2016 reply

    Sherli,

    You can use our sample size calculator to estimate your required sample size.
    First you need to make an estimate of your total population (note that for large populations sample size won’t change very much).
    For the confidence level often 95% is taken, for margin of error 5%.

  • ali - September, 2016 reply

    hi..i have a general question in my mind, i just want to know about it. if you have total population of 100000 students from different colleges in a district, and you have selected sample from it. is it possible to distribute your questionnaire in a single institution which can cover your sample size and collect your data.

    Gert Van Dessel - September, 2016 reply

    Ali,

    If you only distribute your questionnaire in one institution, your sample will not be representative for your total population. You must try to spread your population evenly over all colleges in the district.

  • Jon - August, 2016 reply

    I have an issue with questionnaire distribution. My sample size is 384 using sample size calculator but the population from two geographic locations are Kachia – 120,893 and Dwudu – 432285. hence I cant distribute equally so how to I get the number to distribute the questionnaire from the 384 respondents.

    Gert Van Dessel - September, 2016 reply

    Hi Jon,

    You will have to spread proportionally, which means 84 for region Kachia and 300 for region Dwudu of you want to draw conclusions on a global level with representative region spread.
    However, when you want to draw separate conclusions per region with the same confidence level (95%) and margin of error (5%), you will need 384 respondents for each of the 2 regions.

  • Erickson - August, 2016 reply

    Sir, How can you estimate the magnitude and proportion of the respondents, just for example: total number of households is 59032 magnitude is 17056 and proportion is 28.89 … what could be the possible answer?

    Gert Van Dessel - September, 2016 reply

    What do you mean exactly with magnitude?

  • Maryam - August, 2016 reply

    Hi
    My research topic is role of women in politics and i conducted survey and population is 48. Could u plz tell me which sampling is suitable.
    Plz reply

    Gert Van Dessel - August, 2016 reply

    Maryam,

    For 95% confidence level and 5% margin of error, you would need a sample of 43. But for such small populations, you might as well include the total population in your survey.
    More explanation can be found in this article (Evan Morris, University of Regina)

  • Siyam Mannan - August, 2016 reply

    I am willing to do study on Health related quality of life (QOL) of the postsurgical epilepsy patient. Study suggested about about 80% of the patients become seizure free for first one year though surgery have few expected adverse event. Most studies on surgery outcome are based on clinical improvement only few on QOL using retrospective data. Therefore not much information available on QOL. For my study, how I can calculate sample size.

    Gert Van Dessel - August, 2016 reply

    First of all you have to make an estimation on the total population, and choose the confidence level and margin of error you want to allow.
    Often a 95% confidence level and margin of error of 5% is taken.

    To calculate the sample size, you can use the following formulas:

    SS = (Z-score)² * p*(1-p) / (margin of error)²
    Z-score = 1,96 for confidence level 95%
    Proportion p will be 0,8 in your case.

    After calculation of sample size you have to correct for the total population
    SSadjusted = (SS) / (1 + [(SS – 1) / population])

  • Ravi - August, 2016 reply

    Hi,
    In manufacturing (Pharma domain), how to select suitable sample size to well represent entire lot/batch/population, and how to justify that sample is good enough to represent entire batch.
    Many thanks in advance.

    Gert Van Dessel - August, 2016 reply

    Ravi,

    First of all you have to make an estimation of your entire population (e.g. number of products in your batch).
    If you have an indication about that, you can use our sample size calculator to calculate your sample for a given confidence level and margin of error you want to accept. Often a 95% confidence level and margin of error of 5% is taken.

  • Seth@Babatunde - July, 2016 reply

    Am conducting a comparative research on early childhood education teachers in the Central Region of Ghana and the South-West Geo-political zone of Nigeria.My attempt to get the population of early childhood education teachers in both regions so as to calculate the sample size for each region has proved futile.Please how do I select the sample size from such unknown populations?What source can I cite to back any selection made as per your feedback?Thanks a bunch.

    Gert Van Dessel - July, 2016 reply

    Seth,

    Once your population is large enough, your required sample size does not change very much anymore.
    E.g. for a population of 10,000, confidence level 95% and margin of error 5%, your required sample size is 370.
    If your population rises to 100,000, the required sample size will be 383.
    You can check for yourself with our sample size calculator.

  • Dione Maluwa - July, 2016 reply

    Hello,

    Thank you so much for clarifying these statistical terms in a way much easier to understand. I have 1 question: I want to conduct research that focuses on why young engineering graduates are leaving careers in a specific engineering discipline in favor of others. The problem however is that the total number of graduates in this specific field (which originate from 4 universities) is relatively unknown. If i were to sample current young students (that are busy with their postgraduate studies) from one of the universities would this sample be representative? if not how would i use this data meaningfully?

    Gert Van Dessel - July, 2016 reply

    Dione,

    It seems to me that the 4 universities would normally be able to provide you with the number of postgraduate engineering students, which would give you an accurate indication of your population. To be representative, you will need to have respondents from all 4 universities. If you only are sampling in 1 univerity, your conclusions will be limited to the population of this university and cannot be extrapollated to the entire population.

  • bonnie bourgini - July, 2016 reply

    Helpful discussion , I was enlightened by the information ! Does anyone know if my company might be able to grab a template a form document to fill in ?

    Gert Van Dessel - July, 2016 reply

    Bonnie,

    Can you clarify your question, I do not understand what you need exactly.

  • ijeoma Mary - July, 2016 reply

    Good day Gert Van Desssel. Please what sampling technique and sample size would be suitable for my study of senior civil servants in the Ministries of education of three states in Nigeria. The total population is 1122. please kindly suggest an authority to back up the choice of the sampling techinque and sample size. thank you.

    Gert Van Dessel - July, 2016 reply

    Mary,

    For advice on the sampling technique I suggest you post your question and the details of your research on a statistical forum.
    If you would use random sampling, you would need a sample of 287 for a margin of error of 5% and confidence level 95%.

    Allen Alfred - July, 2016 reply

    I think before choosing the number of sample size you need to ask yourself that all the respondents or area of interest hs have the same characteristics?If not then you can apply your research techniques clustering technique,stratified technnique then u do random selection from there you can have the right number of sample.

  • EMMY - July, 2016 reply

    WHY SIMPLE SIZE IS EQUAL TO TOTAL POPULATION ?

    Gert Van Dessel - July, 2016 reply

    Emmy,

    Your sample size will only be equal to total population when this population is very small and margin of error also.

    Tessy - July, 2016

    Hi Gert,
    Im currently studying on UK working class. Please, I need your guide and help to know whether I should use random sampling or purposive sampling design? What will be the sample size for this population? I shall be grateful with your answers to my three questions to save me from the confusing state. Thank you in advance for kind consideration.

    Gert Van Dessel - July, 2016

    Tessy,

    For quantitative research it is most common to use random sampling, unless you are interested in a specific part of your population.
    If you know the size of your population (or if you have an estimate) you can use our sample size calculator to calculate your sample size for a given margin of error (commonly 5% is taken) and confidence level (95%)

    April J. - September, 2016

    In that case of members of the population should be included. Right?

    Maarten Marijnissen - September, 2016

    Hi April,

    Indeed, your survey sample represents the members of the population you are working with.

  • kastury - July, 2016 reply

    If the population size is unknown. How to determine the required sample size by using a formula.

    Gert Van Dessel - July, 2016 reply

    If the population is large, the exact size is not that important as sample size doesn’t change once you go above a certain treshold.
    For example, for a population of 10,000 your sample size will be 370 for confidence level 95% and margin of erro 5%. For a population of 100,000 this will be 383, for 1,000,000 it’s 384.

    If you have a smaller population, you will have to make an estimation of your population (try to define your target group the best you can).

    sandra - July, 2016

    Hi Gert,

    Sorry I could post a comment to had to click in a reply bottom from someone’s else comment :). I doing a research about the factors that influence online buying decisions (in France). I decide to administrate my questionnaires to only university students because its easier to reach them and all are internet savvy. My population will be french university students. According to statistical site there’s 427000 students enrolled in university programs. I have now a total of 115 responses. Response rate is really low and margin of error is 10% which is above the recommended (5%). But my doubt is if my study is sample is non-probabilistic or probabilistic. Because for what I have search is a non-probabilistc since not all the population studied have the chance to be select (university students will be a sub-group of the universe (French people). IS this right? Because if my sample is not probabilistic the margin error and confidence level have no importance and will be able to make a analysis based on the 115 answers collected. Do you think my sample is probabilistic or non-probabilistic?

    Gert Van Dessel - July, 2016

    Hi Sandra,

    Since you are only recruiting in the group of university students, your sample is clearly non-probabilistic and you cannot extrapolate your results to the entire French population.
    And also within the student population, your margin of error remains above the recommended 5% (9,1% for 95% confidence level). For margin of error = 5% within the student population, you will need 384 respondents

  • Mani manifest - June, 2016 reply

    When your research is based on collecting data from institutions (Research on water conservation). How do you choose the sample size, would it be based on the total number of all the ministries? or would it be based on the individual population of all the workers in the various ministries?

    E.g if the total number of ministries are 36, would you design just one questionnaire for each ministry? or would you

    Gert Van Dessel - June, 2016 reply

    Hi,

    Thanks for your question. This depends on the definition of your population, which is based on the objective of your study.
    If you would do a study about employee satisfaction in the Ministries, your populations will be the employees working in the ministries.

    If it’s for exammple more about general work processes in the ministries, you have a population of 36 ministries. For such a small sample, you will have to send out one questionnaire to all ministries.

  • Karol - June, 2016 reply

    Dear Gert,
    I understand you do market research but maybe you could help over this (same principle). I want to conduct a simple descriptive assessment on the healthcare behaviors of patients in a specific department of a public sector health facility (specifically I would like to know for eg if the patients come to this facility for 2nd opinion, if this facility is their first choice, why do they choose it, would they go to a private facility if they could afford it etc.). I am struggling a bit with: 1)sample size (as I am not sure which size of the population should I choose (is it the average number of patients admitted to this department by day, or by week ,or by month??), 2)the period over which I should be conducting the study (should I choose one day per week and question all patients coming into the department during that day say for a month, or 2; or maybe 2 days a week on a period of 1 month etc.)
    Hope you would be able to help!
    Thanks
    Karol

    Gert Van Dessel - June, 2016 reply

    Karol,

    I would go for a fixed fieldwork period, for example 1 month, and collect responses every day. Your population is the total number of patients for the department in that month.
    Let’s say that you have an average of 30 patients per day, then your monthly population will be 900. For a margin of error of 5% at 95% confidence level, you would need a sample of 270.
    You can always check with our sample size calculator

    Karol - June, 2016

    Thanks a lot Gert! That helps.. Might have more complicated sample calculations in the coming future! I have used your sample size calculator and found it very useful too.
    Thanks again
    Karol

  • Benka - June, 2016 reply

    Dear Gert,
    In my research the population was 45 people, so I have send the survey to all of them. In die end I got 30 reactions.
    Respons rate of 67%.
    Which is good according to the above – right?
    Now I need to back up my statement. Are there any sources I can use?
    How can I defend the statement that a respins rate of 67% is good and sufficient?

    Gert Van Dessel - June, 2016 reply

    Hi Benka,

    If you would go for a margin of error of 5% with 95% confidence level, you would need 41 respondents for your population of 45.
    A sample of 30 respondents corresponds with a margin of error of 10,5% for the 95% confidence level.
    More explanation can be found in following article

  • sharz - May, 2016 reply

    hye…
    my i ask you something which want better if my total population 2.2 million use Krejcie & Morgan Table or Cochran 1963? to measure my study population for cross- sectional survey?

    Gert Van Dessel - June, 2016 reply

    Sharz,

    Both tables use the same formula to calculate the sample size: n = Z² x p x (1-p)/M²

    with
    n = Sample Size for infinite population
    Z = Z value (e.g. 1.96 for 95% confidence level)
    P = population proportion (expressed as decimal) (assumed to be 0.5 (50%)
    M = Margin of Error at 5% (0.05)

  • rehema mcharo - May, 2016 reply

    Hey thank you for the most wonderful explanation my problem how can i download that calculator for my personal usage?

    Gert Van Dessel - May, 2016 reply

    Hi Rehema,

    The calculator is freely available on our website: https://www.checkmarket.com/market-research-resources/sample-size-calculator/

  • Arshad - March, 2016 reply

    Hi,
    Thank you for this valuable discussion forum. I have conducted a survey of households in three villages with household numbers (1)1254, (2)600 and (3)350. In my study I have surveyed these villages selecting random households. Sample size of households was 100, 60 and 60 for the villages 1, 2 and 3 respectively. I want to know how much is the confidence level and margin of error for these three villages independently ?

    With kind regards,

    Arshad

    Gert Van Dessel - March, 2016 reply

    Arshad,

    You can use our sample size calculator to calculate the different margins of error for your 3 samples at confidence level 95% or 99%.

  • Sarah - March, 2016 reply

    Please, How would you draw a sample of 350 students of public Administration department that will be true representative of 6300 students of the population.. Thank you.

    Gert Van Dessel - March, 2016 reply

    Sarah,

    I suggest you use random sampling. By doing this, you ensure that the sample will on average be representative of the entire population. Standard statistical measures of precision such as standard errors and confidence intervals will tell you how far off the population values your sample estimates are likely to be, so there’s no need to validate the sample is representative of the population, unless you have concerns that is was truly sampled at random

    Raghav - September, 2016

    Sir
    I want to know the method to find the no. of samples possible if the population is 10 and size of sample to be taken is 3

    Maarten Marijnissen - September, 2016

    Hi Raghav,

    In this case you have to use the ‘combination formula’.

    You can find this formula here.
    With your data it looks like this: 10! /3!*7! = (10*9*8) / (3*2*1) = 120

    There is an online calculator available that will do the math for you.
    Just fill in your data and it will calculate the number of samples.

  • Keith - March, 2016 reply

    Hi! I am conducting a research and I would like to know how am I going to calculate the number of samples I am going to hand a survey questionnaire within the final sample size? What I mean is how am I going divide the students/respondents in the given sample size?

    EXAMPLE:

    Assume that there are 4 classrooms and each classroom consists of:

    Classroom 1: 44
    Classroom 2: 56
    Classroom 3: 51
    Classroom 4: 48

    The total of the students is 199 and assuming that I have calculated 66 samples from that number of students. I would like to know how many students am I going to give a survey questionnaire in each classroom.

    I hope you get my point! Thank you very much!

    Gert Van Dessel - March, 2016 reply

    Hi Keith,

    I would make a pro rata split in this case, e.g. for classroom 1: 44*66/199 = 15 surveys
    But with this sample size you have a marging of error of almost 10% (95% confidence level).
    If you want to reduce your error margin to 5%, you need to double your total sample from 66 to 132.

    Keith - March, 2016

    What formula should I use then?

    Gert Van Dessel - March, 2016

    Keith,

    The formula that is used:

    first you calculate the sample size(SS).
    SS = (Z-score)² * p*(1-p) / (margin of error)²
    Z-score = 2,01 for confidence level 95,45%
    When the proportion p is not known, it is common to use 0,5.
    After calculation of sample size you have to correct for the total population
    SSadjusted = (SS) / (1 + [(SS – 1) / population])

    You can also check with our sample size calculator

  • umarwahab - February, 2016 reply

    Dear sir
    i m doing Mphil Env.sciences i have to total population of 3901i have survey 350 questionnaries then who i can find their sample size and sample intensity.

    Gert Van Dessel - February, 2016 reply

    Hello sir,

    With the CheckMarket sample size calculator you can calculate your error margin for various confidence levels.
    If you take a confidence level of 95%, your error margin will be exactly 5% with 350 respondents for a population of 3901.

  • Happy Mussa - February, 2016 reply

    I have managed to interview 782 informal traders from 7 border posts. i don’t know the population of informal traders from these sites. how can i statistically justify my sample size?

    Gert Van Dessel - February, 2016 reply

    Mussa,

    Your margin of erro will be maximum 3,5% with 95% confidence level, even if your population would be very large.
    If you have a rough estimation of your population, you can always use our sample size calculator to estimate the margin of error.

  • Andrew - February, 2016 reply

    OK, so sorry if I’m repeating any previous questions but, with a 95% confidence interval and 2% margin of error, what is the sample size needed to represent the average age of a 120,000 person town?

    Gert Van Dessel - February, 2016 reply

    Andrew,

    With an error margin of 2% you will need 2354 respondents in your sample. For error margin 5% this would be 383.
    You can calculate different options for yourself with our sample size calculator.

  • Nida - February, 2016 reply

    Hello! I have a project in school and in my project i am doing a survey to all the students who use Facebook. I ask them how many friends they have in facebook. Out of 10,000 populations i only choose 100 of them to answer the question. Out of 100 there are 60 students responded. I have to answer the questions my professor ask me are, ” how well you think your sample represents your populations? What proportion of your population is female? Does this compare well with your sample? Please give me an idea to answer these questions. Thank you!

    Gert Van Dessel - February, 2016 reply

    Nida,

    If your population is 10,000 and you only have 60 respondents, you will have a margin of error of 12,6% at 95% confidence level. This means if for example 30% of respondents pick a certain answer options, the actual percentage for your population will lie between 17,4% and 42,6%. Most researches take a 5% margin error as a maximum. Your sample is too small to make reliable conclusions for your population. If you know how many females you have in you sample, you can compare this with the (estimated) nr. of females in your population. The % should be more or less the same for both groups to have a representative sample.

  • David Simiyu - February, 2016 reply

    The comment to various questions raised have improved my understanding of calculating the sample size. However, I want to researching on library staff as well as students and teachers in an academic library at a University I do not have much time left. What sample size can I rely on in each strata if I am to be 95% that the sample I use in each category will be representative of the population?

    Gert Van Dessel - February, 2016 reply

    Hi David,

    If you know the population size for each of your strata, you can use our sample size calculator to calculate sample sizes for a 95% confidence level.
    Generally an error margin of 5% is considered as acceptable.

  • Emanuel - February, 2016 reply

    HI, I really appreciate your effort towards the helps your giving on research field . But one thing still confuse me when i want to calculate the size of the sample, and my problem to me is what criterion am i using to choose (2% as marginal of error),, thank you

    Gert Van Dessel - February, 2016 reply

    Emanuel,

    In general a margin of error of maximum 5% is considered as acceptable. The smaller your error margin, the better of course.

  • MT - February, 2016 reply

    Hi dear,
    I wanted to gather the perception of a community (unknown no. of population) regarding CSR as a stakeholder. How can I determine the sample size.

    Gert Van Dessel - February, 2016 reply

    Once your population gets large, your sample size doesn’t change much anymore. You can use our sample size calculator to determine your sample for different populations. Generally a confidence level of 95% and margin of error of 5% is taken.

  • Jurgita - February, 2016 reply

    Hi,
    I’m student from Lithuania. Can you help me. I am doing a research on what Lithuanian society is thing about e- demcracy and its perspectives. How can I concolate sample size, if I dont know the population (or it is too large). Thank You.

    Gert Van Dessel - February, 2016 reply

    Hi Jurgita,

    For large populations the sample size diesn’t change much anymore. For a population of 100,000 or more, you need a sample of 383 to obtain a margin of error of 5% for 95% confidence interval.
    You can calculate for yourself with our sample size calculator

  • Baljeet Singh - February, 2016 reply

    Hi, I have collected a sample of 900 out of total population of 31907. is it a representative???

    Gert Van Dessel - February, 2016 reply

    Hi Baljeet,

    This sample gives you a margin of error of 3,2% for 95% confidence level, which certainly can be considered acceptable. You can calculate forvyourself with our sample size calculator.

  • Monisha - January, 2016 reply

    Dear Gert,

    I am doing a research on farmers who I have classified into small, medium and large farmers. The total population for each strata is 1,005, 565 and 77 respectively. How do I calculate the sample for the population? Do I have calculate the sample for each strata?
    In any case, even if I calculate the sample size using the total number of population of farmers with an error margin of 5% and 95% confidence level, the sample size come out to be 312. I am constrained by time and resources available to survey a sample size of 312. can you suggest another way of sampling? Since stratification of farmers is a logical way to approach my research i was wondering if i could use purposive sampling in each strata. Can you please help me calculate the sample size for purposive sampling with formula?

    many thanks in advance.
    Monisha

    Gert Van Dessel - February, 2016 reply

    Hi Monisha,

    I would be very careful with using purposive sampling because of the high probability of researcher bias based on the selection you make.
    if you use random probability sampling, and want to draw conclusions for the 3 groups of farmers separately (small, medium, large), you will also have to calculate the samples separately.
    This gives following samples for 95% confidence interval and 5% error margin:

    Small = 279
    Medium = 229
    Large = 65

    For information on purposive sampling I can refer you to http://dissertation.laerd.com/purposive-sampling.php

  • jerry - January, 2016 reply

    how scientifically you can choose sample size based on non-probability technique?

  • sadaam04@hotmail.com - January, 2016 reply

    ii. Explain what the results mean when the level of confidence and the margin of error are taken into account.

  • sadaam04@hotmail.com - January, 2016 reply

    I have a table but this question has no population how I can answer?
    a. Determine the correct sample size to be used in the study at;
    95% level of confidence and 2.5% margin of error.

  • Remi - January, 2016 reply

    Hello Sir, for a population of 500 how many sample do I need to take?

  • mary - January, 2016 reply

    hello sir. if the population size is 900. how much will sample size will i need? thank you

    Gert Van Dessel - January, 2016 reply

    Mary,

    For a popuation of 900 you will need a sample of 270 for a confidence level of 95% and margin of error 5%.
    You can check it out yourself for other parameters with our sample size calculator

  • k. lagit - January, 2016 reply

    dear sir., how about if the population is less than 15 …how many sample size i need.

    thank you very much

    Gert Van Dessel - January, 2016 reply

    Dear,

    If your population is that small, you might as well include the total population in your study.

  • poonam - January, 2016 reply

    Dear sir
    i m doing ph.d(Economics). Iwant to ask ideal sample size. Total population is above one lakh and i have taken 1200 sample size it it good or not plz tell me.

    Gert Van Dessel - January, 2016 reply

    Dear,

    With a sample size of 1200 for a population of 100,000 you will get a margin of error of 2,8% at 95% confidence level. Often a maximum margin of error of 5% is used in research, so the 2,8% is certainly acceptable.
    You can calculate for yourself with our sample size calculator

  • Mohamed hassan - January, 2016 reply

    Thanks
    again
    Gert Van
    Dessel
    actually I
    don’t know
    the
    population
    in each
    hospital.
    because my
    research in
    emergency
    department.
    and no body
    know how
    many
    patients will
    present
    with
    asthma in the next few months.
    so, how I
    can choose
    between
    hospitals?
    am I have
    to collect
    the data
    from each
    hospital? or
    I can select
    some of
    them?
    and how
    many
    patients per
    hospital?

    Gert Van Dessel - January, 2016 reply

    In the ideal scenario I would select a sample from each hospital to be truly representative, and this proportionally based on the size of each hospital.
    I would go for a total sample of about 350. But this all depends on the budget you have of course, and also the timeframe. Do you only have a couple of weeks to collect your data, or a year or…
    Maybe you only have 300 asthma patients in a month, then it is practically impossible to collect 350 responses of course.

  • Mohamed hassan - January, 2016 reply

    Thanks again
    Gert Van Dessel
    actually I don’t know the population in each hospital.
    because my research in emergency department. and you know how many patients will present with asthma.
    so, how I can choose between hospitals?
    am I have to collect the data from each hospital? or I can select some of them?
    and how many patients per hospital?

    Gert Van Dessel - January, 2016 reply

    Mohamed,

    If you want to draw conclusions separately per hospital, I would go for a mninmum of 50 per hospital. If you only are interested in overall results, I would aim for a total of 350 respondents for the 10 hospitals (35 per hospital, or when it is difficult to obtain data for certain hospitals you could for example go for 50 respondents from 7 hospitals).

    Mohamed Hassan - January, 2016

    If I obtained the data from 7 hospitals. the question will be whay I chose 7 hospitals?
    Whay not 6 or 8 for example? Can I give clear explanation ?

    Gert Van Dessel - January, 2016

    The more hospitals you include, the more representative your results will be. The only justification to only include 6,7 or 8 hospitals, are practical reasons (limited time, budget, …).

  • Mohamed Hassan - January, 2016 reply

    Hi, thank you for your valuable information.
    I am doing research in asthma.
    And my population are the patients in Khartoum hospitals. We have 10 hospitals. How I can determine my sample size?
    And how I can choose between hospitals. And how many?
    With 95% confidence interval.
    The sample size for my unknown population was 348 patients

    Gert Van Dessel - January, 2016 reply

    Mohamed,

    What do you mean by ‘unknown population’? Do you not have an idea about the total nr. of patients in the 10 hospitals?
    If you have a sample size of 348 and a total population of for example 5000, your margin of error will be about 5% (for the total population of all 10 hospitals).
    You can calculate the margin of error per separate hospital with our sample size calculator (if you would know the population and sample size per hospital).

  • asim khalid - December, 2015 reply

    i have total 1010 retail outlets, I am keen to conduct survey for them. How many retail outlets should i visit to represent true sample and statistically cant be challenged.

  • Alizay - December, 2015 reply

    Hi
    m doing research in which i have more than 1 objectives n i want to know on the basis of which objective i have to calculate my sample size

    Gert Van Dessel - December, 2015 reply

    Alizay,

    Your population is the group of people or objects for which you want to draw the conclusions of your research.

  • Patience T Donnie - December, 2015 reply

    hi,
    i am doing a research. my total population is hundred.and i am asked to use the entire population. so how can i go about getting my sample size and which sampling am i going to use?

    Gert Van Dessel - December, 2015 reply

    Hi,

    For such small populations, you almost need to include everyone to get reliable results.
    For a margin of error of 5% at 95% confidence level, you will need a sample of 80 respondents.
    You can check for other error margins/confidence levels with our sample size calculator.

  • Nicholas - December, 2015 reply

    Let’s say I am conducting a research for a bank (questionnaires) for 432 branches within 2 months time. And the location of sampling have a total of 2.6 million customers.

    And my proposed sample size is 100,000 within 2 months taking account of the number of branch and the peak hours during lunch time. Will it be possible to achieve the target amount of respondents within the period of time.

    Best regards and thanks in advance :)

    Gert Van Dessel - December, 2015 reply

    Nicholas,

    This all depends on the methodology you use to collect the responses. When you dispose of email addresses of the customers, you can perfectly send out an online survey to a large group and collect responses within a couple of weeks.
    When you want to do face-2-face interviews in the bank offices, you will need a lot of interviewers to get to 100,000 within 2 months.

  • martin - December, 2015 reply

    hi didier,
    i am doing a research in a relatvely new field of ICT integration in kenya secondary schools. the population of respondents in my study location is 556 and i have decided on 60 respondents. could this lead to invalid findings?

    Gert Van Dessel - December, 2015 reply

    Martin,

    With a sample of 60 respondents for a population of 556, your margin of error is 12% for a 95% confidence level.
    Generally, 5% is taken as a maximum margin of error, this would correspond with a sample of 228.

  • irsa - December, 2015 reply

    helo.i m doing my MS my population of two groups were 7183 and 2500 and i chose equal number of sample size as 300 from both groups.is it right ,and what should be its justification because i am asked to justify it. kindly help me

    Gert Van Dessel - December, 2015 reply

    Irsa,

    With sample sizes of 300, your margin of error will be 5,5% for the population of 7183 and 5,3% for the population of 2500 (level of confidence = 95%)
    Generally, we accept a margin of error of maximum 5%, which can be achieved with sample sizes of 365 and 334 respectively.
    You can also check with our sample size calculator

    The formula that is used:

    first you calculate the sample size(SS).
    SS = (Z-score)² * p*(1-p) / (margin of error)²
    Z-score = 2,01 for confidence level 95,45%
    When the proportion p is not known, it is common to use 0,5.
    After calculation of sample size you have to correct for the total population
    SSadjusted = (SS) / (1 + [(SS – 1) / population])

  • solomon - December, 2015 reply

    how to calculate sample size to study cusomer satsfaction in 50 banks

    Gert Van Dessel - December, 2015 reply

    Solomon,

    Is your population 50 banks, or the customers of these 50 banks? In the 2nd case, how many cutomers are there approximately per bank/in total?
    If you know the total population, you can use our sample size calculator to determine your sample to obtain a margin of error of 5% (or lower) for confidence level 95%

  • Ben - November, 2015 reply

    Hi all wondering if you could help. I am planning to do a quota sample for teens. I have worked out that my population size for the location I want to target is 3,600.

    Based on a margin of error of 5% and a confidence level of 95% I have worked out that my required sample is 348 respondents.

    However, I plan to use a self-completion survey which usually has a response of 10% . Therefore I would then need to invite 3,480 to complete the survey to hopefully gain the sufficient sample.

    Is this right? I understand quota sampling is more difficult to work out sampling errors etc but I have just used the recommendations thanks!

    Gert Van Dessel - November, 2015 reply

    Hi Ben,

    Your reasoning is correct. Make sure you send out to a representative part of your population, and also check representativess of the sample you have collected afterwards.
    For example, when your population consists of 50% males and 50% females, and for your sample this distribution is 60/40, you might want do do some reweighting (this can be done in a statistical programme, e.g. SPSS)

  • Reddy - November, 2015 reply

    Hi Charlie I need to select a sample size from the population of 45 health facilities of which they consist of 2 are hospital, 2 health centre and 41 dispensaries/

    Gert Van Dessel - November, 2015 reply

    For small populations you almost need to include the entire population in your sample to get a reasonable margin of error.
    In this case, for 45 facilities, you will need a sample of 41 for a margin of error of 5% at the 95% confidence level.

  • Adeel - November, 2015 reply

    Hello sir…
    I have selected two populations for my study, one is the cotton growers and second is textile industry. I want to select 300 cotton growers and 100 textile workers, how i can i justify that?

    Gert Van Dessel - November, 2015 reply

    Adeel,

    If you have an idea of the total population of cotton growers and within the textile industry, you can calculate your margin of error for these 2 samples with our sample size calculator. In general a confidence level of 95% and margin of error of maximum 5% is taken.

  • savi - November, 2015 reply

    Sir,

    I have to calculate the sample size for a population of 25 million people.

    what should be my sample size . Plz suggest

    Gert Van Dessel - November, 2015 reply

    Savi,

    For large populations a sample size of 384 corresponds with margin of error 5% for a confidence level of 95%.
    You can check with our sample size calculator

  • Ofere Jmaes - October, 2015 reply

    I have population known that is 150 with the sample of 30 how can i calculate the population size and sample size

    Gert Van Dessel - November, 2015 reply

    Hello,

    A sample of 30 for a population of 150 corresponds with a margin of error of 16% (for 95% confidence level). This means if for instance 60% of your sample selects a certain answer option, the actual percentage for your population will be between 44% (60-16) and 76% (60+16). As you can see, this is a very wide interval. Generally a margin of error of 5% is taken as maximum. For a population of 150 you would need a sample of 109 to reduce the margin of error to 5%. You can also check with our sample size calculator.

  • Salar - October, 2015 reply

    Hi,

    I have a population unknown for which I am going to draw a sample. First I would want to clear that what is the method of sampling for an unknown population. Secondly if I have a known population, is there any suitable size defined. like 30% would be a best representative than 10%.
    Regards,

    Gert Van Dessel - October, 2015 reply

    Salar,

    If your population is quite large, the exact size doesn’t matter that much. Your sample size will always be around 385.
    E.g. if your population is 30,000 sample size will be 380 (for confidence level 95% and margin of error 5%). If your population increases to 3,000,000 is will still be no more than 385.
    For smaller (unknown) populations you can estimate the sample size with our sample size calculator

  • Alpha - October, 2015 reply

    I want to study the prevalence of a disease in 7 regions with different numbers of individuals by region. See table below
    How many individuals to choose by region for his study?
    Region effectif
    KDG 52524
    MTM 135200
    SL 250000
    SD 126137
    KLD 461870
    TMKD 667984
    TOTAL 1558515

    Gert Van Dessel - October, 2015 reply

    You can use our sample size calculator to calculate sample sizes for your different regions.
    Generally a confidence interval of 95% and margin of error of 5% is taken. e.g. for region KDG this gives a sample size of 382. For such large populations your sample size will always be around 380-385

  • Bob - October, 2015 reply

    I am conducting a survey of a potential pool of 250 people. I have no real data other than defaults to go by as far as required numbers. I have currently a response of 72 people out of the ones I sent. I used 95% confidence and a margin of error of 5%- the calculation shows about 152. If I cannot get any better response rate, can I use the data? I am trying to get a good sample size but my access is limited to this particular group of 250 people. For a social science project, would a CI of 8 % and a CL of 90% seem appropriate? In other words, if all I get is 72 responses, can’t I still use that and show the actual numbers instead of my computed ones?

    Gert Van Dessel - October, 2015 reply

    Bob,

    If you take the generally used confidence level of 95%, your margin of error will be around 10% which is rather high if you want to draw any relevant conclusions for your total population of 250 based on this sample unless you have very clear results. E.g. when 80% selects option A and 20% option B for a specific question, you can still say there’s a preference for A in your population (80% in your sample means the range for your population will be 70%-90% with the 10% MoE)

  • Patra - October, 2015 reply

    How can I can reference this article? cannot find on web and do not want to use as retrieved from webpage. Thank you.
    Patra

    Gert Van Dessel - October, 2015 reply

    Hi Patra,

    The article is only available on web: (www.checkmarket.com/2013/02/how-to-estimate-your-population-and-survey-sample-size/), published by Gert Van Dessel of research company CheckMarket on 13th February 2013

  • Buchura - October, 2015 reply

    How determined margin of error (whether 5% or other %age)?

    Gert Van Dessel - October, 2015 reply

    Buchara,

    In many researches it is standard to permit a maximum margin of error of 5% but of course you can make this stricter if you want a more accurate result and when you have the time and budget to collect more responses.

  • bewket - September, 2015 reply

    the total population in the district is 222,700. however, I am interested to include children aged 1 year to 16 years. but there is no total population for this age group nationally. how can I address this sample size issue?
    thanks!!!!!!!

    Gert Van Dessel - September, 2015 reply

    Although the exact size of your population is not known, you can take 384 respondents as a minimum for a margin of error of 5% with 95% confidence level.
    Once your population reaches a certain level, your required sample doesn’t increase that much anymore.
    E.g. for a population of 10,000 you need 370 respondents, for 100,000 383 and for 1,000,000 384 (to have a margin of error of 5% and 95% confidence).
    You can check with our sample size calculator

  • bewket - September, 2015 reply

    in the district, if I want to draw sample in the district for children only, how can I can consider the population size for this age group.

  • diane - September, 2015 reply

    how about for the sample size of a 120 population only?

    Gert Van Dessel - September, 2015 reply

    You will need 92 respondents for a margin of error 5% and confidence level 95%. See also our sample size calculator

  • Samron - September, 2015 reply

    I need to conduct a KAP survey on students in 550 schools in a certain city. How can I estimate the sample size of the students? The total number of students is not known.

    Many thanks in advance!

    Gert Van Dessel - September, 2015 reply

    Samron,

    If you are only interested in your total group of respondents, you would just need 384 to have a maximum margin of error of 5% with 95% confidence level.
    However, if you als want to draw conclusions on certain subgroups (e.g. cluster of schools) I would go for 384 per subgroup.

  • adugnaw - September, 2015 reply

    Dear Didier,

    1)My target population= 634
    sample size 240 (based on sample size calculation)
    Is this 240 the minimum required respondents? If so, what is the size of samples to be requested(questionnaires to be distributed)in order to collect 240? And how can I calculate the response rate?

    2)If I just distribute 240 questionnaires, is that ok? if I get 200 from my exact sample size(240), is it representative? This is because I was expected to collect minimum of 240 in order to conclude the research.
    I am sure you will adivse me

    Thank you in advance,
    Adugnaw Mesfin

    Gert Van Dessel - September, 2015 reply

    Adugnaw,

    It is correct 240 is the minimum sample for a population of 634 if you want a margin of error of 5% with 95% confidence.
    The response rate is the % of respondents you collect vs. the nr. of questionnaires you have distributed.
    You cannot know that before you distribute the surveys, this will always be an estimation.

    If you collect 200 responses, your margin of error for 95% confidence level will be 5,74%.

  • Mohamed - September, 2015 reply

    I am PhD student. Could you help me to calculate my study sample?
    I n my study I will use the questionnaire about PM examination methods in. The group age from 18 to 100 years old randomly. Libya population is 6 million people. How I can calculate my study sample, please?

    Many Thanks
    Mohamed

    Gert Van Dessel - September, 2015 reply

    Mohamed,

    You can use our sample size calculator to calculate sample sizes for various populations, confidence levels and margins of error.
    In your example, if you go for margin of error 5% and confidence level 95%, your sample will need to be 385.

  • Ndapu - September, 2015 reply

    hi.. I have a population of 678. how do I determine the study sample? kindly provide me with its justification.

    regards…

    Gert Van Dessel - September, 2015 reply

    Ndapu,

    You need a sample size of 245 for this population, confidence level 95% and margin of error 5%
    Formula: first you calculate the sample size(SS).
    SS = (Z-score)² * p*(1-p) / (margin of error)²
    Z-score = 1,96 for confidence level 95%
    When the proportion p is not known, it is common to use 0,5.
    After calculation of sample size you have to correct for the total population
    SSadjusted = (SS) / (1 + [(SS – 1) / population])
    In your case:
    SS = (1,96²) * 0,5*0,5 / 0,05² = 384
    SSadj. = 384 / (1 + [(384,2-1) / 1100000] = 384

    You can calculate yourself with our sample size calculator

  • sham - September, 2015 reply

    hi,
    in my study the accessible population is (small firms) 183. since i intend to apply SEM i need responses 100+. one response from each small firm. should i continue to select a sample or survey the accessible population.

    Gert Van Dessel - September, 2015 reply

    Sham,

    For a margin of error of 5% with confidence level 95%, you would need 125 respondents for a population of 183. So if possible (depending on your timing and budget) I would continue to collect responses until you get to this number.

  • NEERAJ SINGHAL - September, 2015 reply

    “IF WE WANT TO DRAW A SIMPLE RANDOM SAMPLE FROM A POPULATION OF 4000 ITEMS HOW LARGE A SAMPLE DO WE NEED IF WE DESIRE TO ESTIMATE THE PERCENT DEFECTIVE WITHIN 2% OF THE TRUE VALUE WITH 95.45% PROBABILITY.”
    PL GIVE SOLUTION WITH FORMULA AS WELL.
    REGARDS

    Gert Van Dessel - September, 2015 reply

    Neeraj,

    You need a sample size of 245 for this population, confidence level 95,45% and margin of error 2%
    Formula: first you calculate the sample size(SS).
    SS = (Z-score)² * p*(1-p) / (margin of error)²
    Z-score = 2,01 for confidence level 95,45%
    When the proportion p is not known, it is common to use 0,5.
    After calculation of sample size you have to correct for the total population
    SSadjusted = (SS) / (1 + [(SS – 1) / population])
    In your case:
    SS = (2,01²) * 0,5*0,5 / 0,02² = 2525
    SSadj. = 2525 / (1 + [(2525-1) / 4000] = 1548

  • siva - September, 2015 reply

    Respected Sir,
    I have study population of 3674 how ever i want to study about 1% of the population i.e 37 sample. will i get accuracy and good result?

    Thank you

    Gert Van Dessel - September, 2015 reply

    Siva,

    With only 37 respondents, your margin of error will be 16% for confidence level 95%. This means if let’s say 60% of your sample select a certain answer option, in reality for your total population this will be between 54% (60%-16%) and 76% (60% + 16%).
    As you see, this is not very accurate/reliable. For a margin of error of 5% you wil need 348 respondents. You can calculate yourself with our sample size calculator

  • Adebayo Kazeem - September, 2015 reply

    I am so happy to discover your website I believe God is good to me, sir, i am doing a comparative study on the prevalence of Hiv in pregnant women attending ante natal clinic in both rural and urban area of my local government, I have already known my population size to be 450 for urban and 350 for rural, now can we say we are having two sample size? i also choose 95% as my level of confidence and 5% as marginal error is this right or what do you think sir? thank you in anticipation.

    Gert Van Dessel - September, 2015 reply

    Dear sir,

    If you want to draw separate conclusions for these 2 groups (urban and rural) you can indeed consider this as 2 populations.
    With a confidence interval of 95% you will need samples of 208 (urban) and 184 (rural) for a margin of error of 5%

  • Mindy - August, 2015 reply

    Hi,
    I am researching birth characteristics in 300 mother/daughter pairs. In order to check reliability of mothers recall of long ago birth events – I conducted a memory study on 100 women who gave birth over 17 years ago and compared their responses by questionnaire, to birth records. So, I have a record of 100 women’s birth memories and compare agreement of facts with their birth records. Is 100 woman (33% of 300)a large enough sample number to be used as an extrapolation method to assume that 300 women’s birth memories are reliable even if i don’t have their birth records?

    Nadia De Vriendt - August, 2015 reply

    Hi Mindy,

    Actually with a population of 300 you would need a sample of about 169 women in order to get reliable results (based on a confidence level of 95% and a margin of error of 5%).
    With the current sample of 100 your margin of error is 8.01% which is really high.

    If at all possible you should expand your study and include more women.

  • charlie navarro - August, 2015 reply

    how to get population without given total of population?
    ex. a student researcher uses 5% margin error in computing his sample size. if the sample size he obtains is 250, what is the population?

    Gert Van Dessel - August, 2015 reply

    Charlie,

    For these numbers your population will be around 715. You can use our sample size calculator to check for other options”>sample size calculator to check.

  • Chris - August, 2015 reply

    Sir I am to survey a population of about 10000 students. And I have to consider the ratio of boys to girls. Using 95% confidence level and 5% margin error, I am able to determine the sample size to be 370 (using the table above). So my question is, in this 370 students, shall I consider the ratio of boys to girls? Thanks

    Gert Van Dessel - August, 2015 reply

    Chris,

    It is correct that your sample has to be 370 for a population of 10,000 and confidence level 95% + margin of error 5%.
    When you want to have a true representation of your population, it is indeed important to have the correct boys/girls ratio, but also age split, regional division, etc.
    You can work with quota for the different subgroups (based on your population split). If that is not possible, it is also an option to reweigh your results afterwards in a statistical software pack (e.g. SPSS).

  • Fatima - August, 2015 reply

    Hi,
    Sir i have population size of 210 ? What should be my sample size ? And what margin of error and confidence interval shall i choose ? (considering i want the survey to depict a good picture of the situation )

    Will be grateful if you’d help !

    Regards ,
    Fatima

    Gert Van Dessel - August, 2015 reply

    Fatima,

    It is common practice to take a 95% confidence interval and margin of error 5%. This gives a smaple size of 135.
    You can use our sample size calculator to calculate sample sizes for alternative confidence levels and margins of error.

  • Alice - July, 2015 reply

    I have a study population of 390,000 and I want use a study sample of 300, how can you justify this sample number statistically. Will I get good results.

    Gert Van Dessel - July, 2015 reply

    Alice,

    With a sample of 300 for a population of 390,000 your margin of error will be 5,66% for a 95% confidence level. It is common to take a maximum margin of error of 5%, to attain this you will need a sample of 384.
    You can calculate this yourself with our sample size calculator

  • Virendra - July, 2015 reply

    Hi

    Need help in Sample size calculations

    Population size=1100000
    Margin of error=4.38
    confidence level=95%

    Need with formula and proper explation

    Gert Van Dessel - July, 2015 reply

    Hi Virendra,

    You need a sample size of 500 for these numbers

    Formula: first you calculate the sample size(SS).
    SS = (Z-score)² * p*(1-p) / (margin of error)²
    Z-score = 1,96 for confidence level 95%
    When the proportion p is not known, it is common to use 0,5.
    After calculation of sample size you have to correct for the total population
    SSadjusted = (SS) / (1 + [(SS – 1) / population])
    In your case:
    SS = (1,96²) * 0,5*0,5 / 0,0438² = 500,6
    SSadj. = 500,6 / (1 + [(500,6-1) / 1100000] = 500,4

  • Kieran - July, 2015 reply

    If i have a population size of around roughly 2 billion, margin of error c.5% and confidence level of around 95%, what would my sample size be? I’ve calculated before and it seemed too small to be representative of that size of a population. I think it was around 384? Surely 384 respondents is not enough to know what the opinion of 2 billion people would be?

    Gert Van Dessel - July, 2015 reply

    Kieran,

    Although it seems counter-intuitive, a sample size of 384 is enough for a population of 2 billion to attain a margin of error of 5% with 95% confidence. For large populations the sample size does not change a lot (e.g. 370 for 10,000, 383 for 100,000 and 384 for 1 million).
    But note that this is based on sampling from a homogenous population. If you have a heterogeneous population (e.g. different proportions or properties for different sub-groups), then this variance estimate is not so reliable. You need to make sure the division of your sample matches this of your population as closely as possible.

  • Vlad - July, 2015 reply

    Kudos for the extremely informative post and replies.

    My question is less about sample size and more about its structure. More specifically, target population is half a million, meaning the 95% confidence level and 5% margin of error would require approximately 400 respondents. My question is, does it matter if I conduct all 400 surveys within a single area or should the surveys be distributed in a specific way throughout the target region (i.e. 20 surveys in 20 geographically distinct sub-communities)? Note: I’m not looking to analyse sub-communities, but the district as a whole.

    Many thanks!

    Gert Van Dessel - July, 2015 reply

    Hi Vlad,

    Thanks for your question. Sample size and sample representativeness are 2 related, but different issues.
    A sample of 400 is large enough for a maximum margin of error of 5% with a 95% confidence level, but that doen not necessarily mean your sample is representative for your population.
    For example your can include only people younger than 25 in your sample of 400. This doesn’t change your margin of error, but you cannot extrapolate the results to the entire population.
    You have to try to make the composition of your sample as similar as possible to your population, also when you don’t want to analyze subgroups. so in this particular case I would try to spread your sample over your different geographical regions. This will surely make your results more valid.

  • Ahmed - July, 2015 reply

    I have a question,

    I am surveying a 430 population on a 38-items questionnaire. How much is my sample size?

    Thank you,

    Gert Van Dessel - July, 2015 reply

    Ahmed,

    For a population of 430, you will need to collect 204 responses for a margin of error of 5% and confidence level 95%. You can check with our sample size calculator

    Vidath Sanura - July, 2015

    Hi Sir,

    This thread is really helpful an I also want to clarify something from you.

    My target population is 33 companies and using a preliminary study which I got 20 responses, it is calculated that average number of staff in a single company comes to around 150 people.

    How can I calculate my sample size assuming a ME of 5% and CL of 95%.

    Thanks,
    vidath

    Gert Van Dessel - July, 2015

    Vidath,

    With an average staff of 150 per company, your total population will be around 5000. For a ME 5% and CL 95% you will need 357 respondents (you can check with our sample size calculator. I would try to spread these proportionally over the different companies (around 11 for an average size company, more for larger companies and less for smaller companies).

  • jabar - July, 2015 reply

    Can I send invitation to entire population of 350 without any sampling strategy ?

    Thanks.

    Gert Van Dessel - July, 2015 reply

    Jabar,

    If your population is 350 it is best to send out to this total group. Depending on your response rate, you will get a number of responses. With our sample size calculator you are able to check the margin of error based on the confidence level you require.

  • Ujulu - July, 2015 reply

    Dear,
    the total population in the area is 410 having scattered settlement, as result difficult to contact easily these households. so, i need to have about 65 respondents because i heard that taking a sample size which is more than 10% of the total population is considered as a minimum requirement.could this be a true representative?

    Gert Van Dessel - July, 2015 reply

    Dear,

    If your total population is 410, a sample size of 65 corresponds with a margin of error of 11% for a 95% confidence level. THis means when for example 60% of your respondents selects a certain response, the actual percentage will lie between 49% en 71% for your total population

    Ujulu - July, 2015

    Dear Gert Van Dessel,

    I understand that, but my question is that can this sample size be true reprsentative of the populatio? what if i use 90% confidence level and 6% margin of error? can one use simply 10% of the total population as a minimum requirement to use the sample size?

    Gert Van Dessel - July, 2015

    Ujulu,

    No, you cannot do that. You can lower your confidence level to 90%, but then your results will be less reliable. You also cannot say 10% of your population will be sufficient as sample size

  • Amaika - June, 2015 reply

    We are profiling villages, however, people on ground only interview 18% of the population. i would like to know the quality of my data, my confidence with the data. if the result would be bad, how can i correct it using statistics?

    Kind regards

    Gert Van Dessel - June, 2015 reply

    Amaika,

    If you know the size of your total population and the nr. of answers you have collected, you can calculate your margin of error for a 95% confidence interval using our sample size calculator. If the margin of error is higher than you want to allow, you will need to collect additional data. You will also want to check whether your sample is representative for the total population (e.g. looking at gender and age split, …). If this is not the case you can use statistical software (e.g. SPSS) to weigh your data.

  • Michael - June, 2015 reply

    I have a sample size of 384,and I need to get respondents from 10 schools how do I get the number of respondents required for each of these 10 schools?

    Gert Van Dessel - June, 2015 reply

    Michael,

    I would take the samples per school proportionnaly to the population per school (if you know this).
    E.g. when school A has 15% of your total students population for 10 schools, you take 15% of your sample = 15% of 384 = 58 students for school A.

    If you don’t know the population per school, I would try to spread my sample evenly among the 10 schools (38 per school)

  • Michael - June, 2015 reply

    what if I do not know the population?

    Gert Van Dessel - June, 2015 reply

    If you take a sample size of 384, your margin of error will not exceed 5%. For large populations, the MoE does not change drastically. E.g. for a population of 5,000 your MoE will be 4,8%, for 10,000 it will be 4,9%, for 1,000,000 there is just a marginal increase to 5,0%

  • Michael - June, 2015 reply

    Supposing I am required to use a sample size of 384 and yet I am required to get them from ten sites , how do I go about this?

  • Humera - June, 2015 reply

    Dear Sir,

    I am working on airlines websites. The respondents are the general consumers and I dnt know about the population.Kindly do let me know how can I check the sample size. or what should be the sample size….I estimated 327 sample size…

    Gert Van Dessel - June, 2015 reply

    Dear,

    You can use our sample size calculator to calculate sample sizes for different populations. You will see that, once your population goes above 10,000 your sample size doesn’t change much anymore (e.g. 370 for 10,000 and 384 for 1 million with a confidence level of 95% and margin of error of 5%). This means that you’re safe when you take a sample of 384 or more.
    Your sample of 327 would be ok if your population is not higher than 2150.

  • DianaC - May, 2015 reply

    This is all really useful information. I have a question. I have a survey and the number of completes falls with the 95% confidence and 5% error. (1600 completes representing the TV owning population of Australia). How small can I break down each group. For example, if there are only 150 completes from people age 55+, are the responses good or is the segment too small?

    Gert Van Dessel - June, 2015 reply

    Hi Diana,

    The smaller the segment you are looking at, the higher your margin of error will get for a given confidence level (e.g. 95%).
    For a large population (e.g. 100,000) a segment of 150 respondents results in a margin of error of 8%.
    The minimum size of the segments you want to study depends on the margin of error you are prepared to accept.

  • gorba - May, 2015 reply

    hi dear, iam doing a survey at university level,,,, about the population of 320,, please help to calculate the sample size,, i want both formulas for calculating sample size for qualitative nature of research and quantitative nature of research,, please help me as soon as possible

    Gert Van Dessel - June, 2015 reply

    Hi Gorba,

    For a population of 320, the required sample size for a confidence level of 95% and margin of error 5% will be 175. How do I get to 175?

    First you calculate the sample size (SS).
    SS = (Z-score)² * p*(1-p) / (margin of error)²
    If you do not know your proportion p, it is quite common to take 0,5 as a value for p.
    Then you correct it if you know your population:
    SSadjusted = (SS) / (1 + [(SS – 1) / population])
    For your population of 320, confidence level of 95% and margin of error 5% you get the following calculation:
    SS = (1,96²) * 0,5*0,5 / 0,05² = 384
    SSadj. = 384 / (1 + [(384-1) / 320] = 175

    For qualitative research there are no clearly defined statistical rules. Your guiding principle should be the concept of saturation. You go on until you think you got all the information you need and there is no added value in interviewing additional respondents. You can have a look at this paper about the subject. There is a mention of about 30 respondents for a valid qualitative research, but there is a lot of variance depending on the study you are doing.

  • Saeed - May, 2015 reply

    Hi
    I have 2-3 measurements from one instrument (2-3 data per instrument)
    now, assume I have 100000 data per instrument and I want to get some samples that are representative for all 100000. how could I determine my sampling size?
    Also, the instrument was selected between 100 instrument by the experts and not in a random way
    Does your method in appropriate for this case??
    Could I use the mean and standard deviation of primary 2-3 data to calculate mean and stdev of my samples??

    Gert Van Dessel - June, 2015 reply

    Hi Saeed,

    Thanks for your question. If you have 100000 data for one instrument this is your ‘population’ for this instrument. For a confidence level of 95% you will need to have 384 data for this instrument to obtain a margin of error of 5%. For a smaller margin of error and/or higher confidence level you will need a largere sample.
    You can use our sample size calculator to calculate sample sizes for alternative confidence levels and margins of error.
    The results for this sample wil only valid for this one instrument, you cannot make any predictions for other instruments. You cannot use the mean and stdev of only 2-3 data, you need to take all sample data into account and calculate their mean and stdev.

  • Meg - May, 2015 reply

    How di I get a representative sample if my target population includes all current households

    Meg - May, 2015 reply

    Thanks a lot,but I dnt need a formular. I was asked to explain how to get a representative sample if my target population includes all current households and to specify how to carry out this research project

    Gert Van Dessel - May, 2015

    Meg,

    Whether your sample is large enough and whether it is representative are 2 different things.
    The required size of your sample based on the margin of error and confidence level you need can be calculated with our sample size calculator (e.g. for a large population, margin of error = 5% and confidence level 95% you need a sample of 384 respondents/households)

    Representativeness means that every part of the population is represented in your sample in the proportion that you see for the entire population.
    To make sure this is the case, you have to use quota (e.g. for region, age, gender, …). If 20% of the households you want to study lives in region X, you will need 77 household for the region in your sample of 384 households.

  • gbwanika - May, 2015 reply

    Am conducting a study but I want to get a basis of determining the study population. Are there authorities who for example talk of 50%, 20%

    Gert Van Dessel - May, 2015 reply

    Gbwanika,

    Your sample size depends on the margin of error (2% 5% 10%?) you want to acquire for a certain confidence level (mostly 95% or 99%).
    You can calculate your sample size based on these criteria with our sample size calculator

  • V.S.Ramalingam - May, 2015 reply

    Can you suggest a formula for calculating the sample size for mean to a known population?

  • richa gupta - May, 2015 reply

    thanks sir very useful information for any researcher regarding sample size. i want to know that can i use this formula for my survey as i am going to conduct survey regarding purchase behaviour of customers in four cities of about over 1 lakh population? what will be my sample size then?

    Gert Van Dessel - May, 2015 reply

    Richa,

    With an error margin of 5% and confidence level 95% you will need 383 respondents. You can check this with our sample size calculator

  • Bernard - April, 2015 reply

    my population is around 900, what then will be my sample size? please

    Gert Van Dessel - May, 2015 reply

    Bernard,

    You can calculate your sample size for various confidence intervals and margins of error using our sample size calculator

  • Asif - April, 2015 reply

    I am research scholar doing research on customer perception about primary market.I have to collect the data from four cities in these cities total demat account holder are 13 lakh 69 thousand 2 hundred and 79 which is my population.I want to use 95% confidence level and 5% confidence interval.What should be the sample size,i can use for my study. please suggest.

    Gert Van Dessel - April, 2015 reply

    Asif,

    Using our sample size calculator you will see that your sample size has to be 385

  • Antal aziz - April, 2015 reply

    hi,,, m doing a survey at university level,,,, about the population of 25000,, please help to calculate the sample size,, i want both formulas for calculating sample size for qualitative nature of research and quantitative nature of research,, please help me as soon as possible

    Gert Van Dessel - April, 2015 reply

    Antal,

    On our site you can find our sample size calculator where you can calculate sample sizes for different margin of errors, confidence levels and expected response rate.

    Samples for qualitative studies are generally much smaller than those used in quantitative studies. There is a point of saturation where more qualitative interviews don’t necessarily lead to more insightful information about the topic you are studying. Qualitative research is very labour intensive, therefor we suggest to limit your sample to a maximum of about 30 respondents.

    Antal aziz - April, 2015

    ok thanxxx for helping me,,,, i’ll be saerch out,,,,

  • ghadeer - April, 2015 reply

    plz can you help me to calculate my sample size on this formula
    n = (1.96)^2*P*Q/d^2
    if you know that prevalence rate is 1.2%
    how many samples should i collect
    thanks on advance

    Gert Van Dessel - April, 2015 reply

    Ghadeer,

    On our site you can find our sample size calculator where you can calculate sample sizes for different margin of errors, confidence levels and expected response rate.

  • Chikaiko chadzunda - April, 2015 reply

    I want to do quantitative relational research in a banking industry – however the data I am looking for is factual it means therefore only one responsible official will provide it – in which case with an industry of 10 banks I will have only 10 questionnaires – some literature recommends a min of 30 for quantitative relational research – how do I handle this?

    Gert Van Dessel - April, 2015 reply

    You are right, 10 questionnaires is not enough to draw any general conclusions. Your results will only make sense for the 10 banks you have investigated and you cannot extrapolate this to the total banking industry (unless these 10 banks represent almost the total banking industry in the region you are investigating) . As you point out, a minimum sample of 30 is recommended.

  • Maryiam - April, 2015 reply

    Hi Didier

    I just need to ask if i select a sample of let suppose 600 for any market study.But i don’t know the exact number of population of that particulars with in a country boundary than how to check the representatives and randomness of Sample..

    Gert Van Dessel - April, 2015 reply

    Hi Maryiam,

    The great thing about sample sizes is that once your population gets high, your error margin doesn’t change a lot anymore. If you take a sample on 600 on a total population of 10,000 the error margin is 3,88%. If your population grows to 100,000 the error margin only increases to 3,99% and for 1,000,000 it is 4,00%

  • Shaji - April, 2015 reply

    Hi Sir,
    I am doing an academic research in the area of e-governance where there are three respondents groups namely, Employees of Commercial Taxes Department (Population 3600), Tax payers (Population 200000) and Tax Practitioners (Population 2500). I am planning to take a sample of 300 tax payers, 180 employees and 120 Tax practitioners.(Thus a total of 600 respondents). will this sample is sufficient for my study? Please reply.

    Gert Van Dessel - April, 2015 reply

    Shaji,

    The margin of error associated with your sample sizes is 5,7% for tax payer, 7,1% for employees and 8,7% for practitioners.
    If you only allow an error margin of 5%, your sample sizes respectively need to be 384, 348 and 334.
    You can also check with our sample size calculator

  • phiona - April, 2015 reply

    please help me with the following question

    a manufacturer would like to survey users to determine the demand potential for its new washing machine. the new mashing machine has a capacity of 100kgs and costs about Us$55500 its used for washing, rinsing, dry cleaning and pressing.
    a) identify the population and the sampling frame that could be used
    b) describe how a simple random sample can be drawn using the identified sampling frame
    c) could a stratified sample be used? if so, how?
    d) which sampling technique would you recommend and why?
    kind regards,
    phiona

    Gert Van Dessel - April, 2015 reply

    Hello Phiona,

    For your homework assignments I will have to refer you to your tutor;-)

  • phiona - April, 2015 reply

    please help me with this question

    a manufacturer would like to survey users to determine the demand potential for its new washing machine. the new machine has a capacity of 100kgs and costs about $55500. its used for washing, dry cleaning,and pressing.
    a) identify the population and sampling frame that could be used
    b) describe how a simple random sample can be drawn using the identified sampling frame
    c) could stratified sample be used? if so, how?
    d) which sampling technique would you recommend, and why?

  • Kazuko - April, 2015 reply

    Hay alguna cosa que no entiendo, voy a volver a mirarlo y si nó te pregunto por aquí.

  • Amit Dixit - April, 2015 reply

    Sir I have 20 variables in my research report questionnaire. so how many respondents I should need yto justify my answer.

    Gert Van Dessel - April, 2015 reply

    Amit,

    Can you explain what you mean exactly with these 20 variables?
    Are these segmentation variables for your group of respondents?

  • Godfrey Mutibvu - April, 2015 reply

    i have a target population of 100. how can i calculate the sample size

    Gert Van Dessel - April, 2015 reply

    Godfrey,

    For a margin of error of 5% and confidence level 95% you will need 80 respondents.
    You can check with our sample size calculator

  • Tracy Carnelutti - April, 2015 reply

    Hi

    I would greatly appreciate your assistance and response to my question

    I am running three trials to compare a different processes used in each

    I am trialling ​8​ stores for each process being tested​

    Baseline data : 8​ weeks
    Transition: 4 weeks (to get used to the new process / equipment)
    Pilot: ​8 weeks

    This will give 64 weekly data points from the pilot to compare to 64 baseline points.

    This required sample size was based on
    – Observed Co-efficient of Variation of 10% (ie standard deviation is 10% of the average carton rate)
    – 95% Confidence Level required (chance of false positive is 5%)
    – 80% Power of test (chance of a false negative is 20%)

    My question is will this account(negate) for variables between these stores such as the capability of different staff in each store.

    Thanks

    Tracy

  • Bereket Admasu - April, 2015 reply

    Really my appreciation is very deep.10Q!!!

  • Adrian - March, 2015 reply

    Please can you help with the question below:

    Suppose you want to decrease your margin of error by a factor of 5. By what factor, do you need to increase your sample size?

    In other words, if you want to decrease your margin of error from 35% to 7% by what factor do you need to increase your sample size?

    Kim Helsen - April, 2015 reply

    Dear Adrian,

    Thank you for your question.

    The preferred margin of error is the positive or negative deviation you allow on your survey results for the sample, in other words the required precision level. Suppose in your survey 40% of the respondents pick a certain answer and your margin of error is 2%. This would mean that if you interrogate the total population, you can be sure that between 38% and 42% would pick the same answer. The smaller the allowed margin of error, the larger your sample will have to be.

    A margin of error of 35% is very undesirable, since no reliable conclusions can be made. In general, a margin of error between 1 and 5% is used.

    Your sample size depends on the size of the population you are measuring. There is no fixed factor with which you can multiply your sample size of you divide your margin of error by 5.

    An example:
    Consider a population of 10.000. When using a margin of error of 1% you need 4900 respondents. With a margin of error of 5% you need 370.

    Consider a population of 1.000. When using a margin of error of 1% you need 906 respondents. With a margin of error of 5% you need 278.

    If you would like to use a margin of error that is not standard in our calculator (https://www.checkmarket.com/market-research-resources/sample-size-calculator/), you can use the following formulas:

    Calculate the sample size (SS). If you do not know your proportion p, it is quite common to take 0,5 as a value for p. The z-score for a confidence level of 95% is 1,96.

    SS = (Z-score)² * p*(1-p) / (margin of error)²

    Then you correct it if you know your population:
    SSadjusted = (SS) / 1 + [(SS – 1) / population]

    If we use it in an example with a population of 100.000, confidence level of 95% and margin of error 2%, then you get the following calculation:
    SS = (1,96²) * 0,5*0,5 / 0,02²
    SS = 3,8416 * 0,25 / 0,0004
    SS = 0,9604 / 0,0004
    SS = 2401

    SSadj. = 2401 / 1 + [(2401-1) / 100.000]
    SSadj. = 2401 / 1 + [2400 / 100.000]
    SSadj. = 2401 / 1 + [0,024]
    SSadj. = 2401 / 1,024
    SSadj. = 2344

    kind regards,
    Kim

    Adrian - April, 2015

    Thank you for the help. Very helpful.

  • Lisa - March, 2015 reply

    Hi!

    Recently I´ve performed a survey amongst all 76 B2B customers of a company. I´ve received 38 responses, so established a 50% response rate. How do I discuss the confidence level (and how do I determine what a proper confidence level is?) and also, how do I explain that the reliability of the results?

    Thanks in advance!

    Kind regards,
    Lisa

    Kim Helsen - April, 2015 reply

    Dear Lisa,

    Thank you for your question.

    Your population size is 76 with a sample size of 38. If you calculate this in our calculator (bottom part of the page) https://www.checkmarket.com/market-research-resources/sample-size-calculator/, you will see that, with a confidence level of 95%, you obtain a margin of error of 11,32%. Confidence levels of 95% or 99% are standard. The margin of error is the positive or negative deviation you allow on your survey results for the sample, in other words the required precision level.

    Suppose in your survey 40% of the respondents pick a certain answer and your margin of error is 11,32%. This would mean that if you interrogate the total population, you can be sure that between 28,68% and 51,32% would pick the same answer.

    Kind regards,
    Kim

  • zamin ali - March, 2015 reply

    i am doing a marketing research in my district but i am not able to find out High Income, Middle Income and Low Income Households. there are 32600 house holds in the area. i want to make three groups of them in the above mentioned categories. plz guide me in the matter

    Kim Helsen - March, 2015 reply

    Dear Zamin Ali,

    Thank you for your question.

    First you need to know or estimate the distribution of the households regarding income. Maybe a governmental website exists showing such information (statistics on demographical variables or economics?).

    Then, you can enter these numbers in our calculator: https://www.checkmarket.com/market-research-resources/sample-size-calculator/

    If, for instance, all income levels are equally distributed (33,3%), you will need information from 372 households per group (confidence level of 95% and a margin of error of 5%). If you want to draw more precise conclusions, with a margin of error of 2%, you will need information on 1967 households for each income level.

    kind regards,
    Kim

  • Richard - March, 2015 reply

    i want my sample size to be 50 respondents to filling the questionnaires.So how do i calculate to get the right target population?

    Kim Helsen - March, 2015 reply

    Dear Richard,
    thank you for your question.

    A sample of 50 is sufficient when investigating a population of 57 individuals with a margin of error of 5% and a confidence level of 95%.

    However, a population size is fixed. You can make an estimation if you don’t know the exact number of individuals, but you cannot ‘choose’ or calculate your population size.

    If you conduct an employee survey for instance, your population would be the total staff.

    Kind regards,
    Kim

  • Joseph Jaeger - March, 2015 reply

    Sir – I appreciate your organization’s helping so many. I am in hopes you can answer a question. The process you outline deals with survey responses. Are the calculations the same when dealing with other “respondents.” The issue I have is trying to choose a representative sample of files I must assess. My population is 310, and when entered into your calculator (error 5%, CL 95%), the required number of respondents (172) seems high.

    Thanks,
    JJ

    Kim Helsen - March, 2015 reply

    Dear Joseph,

    as you’ve correctly calculated, you will need a sample of 172 respondents for this survey (error 5%, CL 95%).

    Kind regards,

  • FG - March, 2015 reply

    Hi, I am planning to apply for a grant application to analyse the role of psychosocial factors on wellbeing.
    I would like do questionnaires in two hospitals centrally located in London and Madrid. I would like to know first How can I have a representative sample of the population in both different regional areas and second how Can I distribute the surveys through different departments at hospitals.
    Also I would like to know if I could focus in only one specialised department (for example ophthalmology) with different work positions such as for example administratives, nurses, specialised professionals.
    Many thanks

    Kim Helsen - March, 2015 reply

    Dear FG,

    To calculate a representative sample of the population in both different regional areas, you need to fill in the population size for each hospital. If you’re investigating staff, this is the total number of employees, if you’re investigating patients, you need to know the total number of patients for each hospital.

    Use the calculator on: https://www.checkmarket.com/market-research-resources/sample-size-calculator/

    If you want to make statements for a specific department, your population number is the total number of staff/patients for this specific department.

    You can focus on one specialised department (for example ophthalmology) with different work positions such as for example administratives, nurses, specialised professionals.

    If you want to make statements about that department in general, population size is the number of staff in that department.
    If you want to write conclusions for each type of work position separately, you need to count population per working position. Often this means that you’ll need to interview all people that work in that department.

    Kind regards,

  • Ishtiaq - March, 2015 reply

    I have 20 colleges with different no. of students and a total population of 1 lakh. around 2000 to 6000 students in a college. how many students should i take from each college.
    As per your website calculator i have to select 400 students out of 1 lakh.

  • muhammad anees - February, 2015 reply

    sir…….
    i have a problem to how i find sample size of old building,,,,,,

  • asma - February, 2015 reply

    Hi, nice article. I have a question that how can we find the average no. of people in a household in given population?

  • Tarsuinn - February, 2015 reply

    Odd question. Suppose I record ten different things each year for 20 years (e.g. quarterly mean outgoings) , and two more which are calculated from other records (mean annual outgoings, for example). What is the sample size here? Thanks

  • Jules - February, 2015 reply

    Hi there!

    I have to determine the needed sample size for a population of 350 for both the 0.01 and the 0.05 significance levels. The only information I have to go on is an effect size of 0.20 and a power of 0.80. How in the world do I do this without knowing confidence levels or confidence intervals?

    Thanks for your help!!

  • malou - February, 2015 reply

    gud pm. may i want to know if the population is 114/221 what is the percentage of respondents? how many respondents should i have? Thank you.

  • Masego Jeremiah - February, 2015 reply

    Good evening. Am doing my MBA and have been asked to state the name of the formula I used to determine my sample size. What is the name of this formula? (Z-score)² – StdDev*(1-StdDev) / (margin of error)²

    Willem Schrijver - February, 2015 reply

    Dear Jeremiah,

    Here is an example that came up earlier in this post that gives detailed info on the formula:

    First you calculate the sample size (SS). If you do not know your proportion p, it is quite common to take 0,5 as a value for p.
    SS = (Z-score)² * p*(1-p) / (margin of error)²
    Then you correct it if you know your population:
    SSadjusted = (SS) / 1 + [(SS – 1) / population]
    If we use it in an example with a population of 100.000, confidence level of 95% and margin of error 2%, then you get the following calculation:
    SS = (1,96²) * 0,5*0,5 / 0,02²
    SS = 3,8416 * 0,25 / 0,0004
    SS = 0,9604 / 0,0004
    SS = 2401
    SSadj. = 2401 / 1 + [(2401-1) / 100.000]
    SSadj. = 2401 / 1 + [2400 / 100.000]
    SSadj. = 2401 / 1 + [0,024]
    SSadj. = 2401 / 1,024
    SSadj. = 2344

  • Lavanya - February, 2015 reply

    Sir, I have universe of 50,800 . will 1% sample size be accurate or 2% be good. I do not want to go for a bigger sample . Plz suggest

    Willem Schrijver - February, 2015 reply

    Please check it over here: https://www.checkmarket.com/market-research-resources/sample-size-calculator/

    Emanuel - February, 2016

    Hi, am student at Mwenge Catholic University am enjoying answers through different questions that people posed to you . But mine is that how am i going to choose marginal of error in order to use in calculating the sample size.

  • Chelsy - February, 2015 reply

    I am working on a change project (small scale) for the completion of my BSN in May. My population is MUCH smaller (roughly 550 students) than those mentioned above and once the proposal is submitted and approved I will have only 5 to 6 weeks to gather representative data. I will be encouraging parents to participate, but I can’t include them in my sample. Those directly affected by the findings will be the students. I did utilize the sample calculator, which is essentially telling me that I will need to include close to the entire population for a 5% margin of error w/ 95% confidence. Considering my time frame, is this feasible? If I only include, say, 230 students, and expect a 50% response rate, will my data be indicative of the population at large?

    Willem Schrijver - February, 2015 reply

    Hi Chelsy,

    In order for your data to be respresentative on the generally accepted 95% confidence level and 5% margin of error, you will need to sample of 227 students. With your suggestion you have only 115 respondents, so this will not be enough to be representative.

    However, if you clearly report this in your analyses and you formulate your conclusions carefully (pointing out it is not representative) I think you should be fine normally. However, I do advise you to discuss this with your teacher.

  • sweta - February, 2015 reply

    dear sir
    i want to take 90% confidence level , then what will be my sample size ?
    i am doing a research on consumer buying behavior of smart phones , so i even dont know my population size.
    sir plz help me , as fast as possible

  • Monica Eduboah - January, 2015 reply

    please calculate a sample size for me with a population of 1,840

  • Ayodele Ayodeji - January, 2015 reply

    hello there,
    kindly help me with any reference that allows me the choice of subdivisions in my sampling. for example , if i have 15 wards and i intend choosing 30%,or 10% etc of that wards. please i need a reference backing this percentage up.
    Thanks.
    Ayodele

    Willem Schrijver - January, 2015 reply

    Can you eleborate a bit more on the population size of the wards/subdivisions?

  • Monica Eduboah - January, 2015 reply

    l have a population of 5840 and want get a sample size for a study can you please help me?

    Willem Schrijver - January, 2015 reply

    361 is the minimum –
    https://www.checkmarket.com/market-research-resources/sample-size-calculator/

  • Kristen - January, 2015 reply

    Please can you clarify, is sample size the minimum number of respondents you need to have responded, or is it the number that you need to contact, from amongst your total population? For example:

    My total population is 14,625 and I had access to all of them. My sample size calculation came out at 375 for a confidence level of 95% and with a confidence interval of 5. As such, I sent out my survey to the entire population of 14,625, and received a total of 792 responses back.

    Was I wrong in contacting the whole population – that is, should I have instead drawn a random sample of 375 and contacted only that group? Or, was I right to survey all, and I can now say I met my sample size, because I had greater than 375 respondents, with a margin of error of only 3.39%?

    This is top urgent…..If I have misinterpreted and was only supposed to survey 375, I am in big trouble!

    Willem Schrijver - January, 2015 reply

    Kristen,

    You sample size is the amount of people that fully participated to your study/questions/exercise units. Contacting the whole population is not wrong, it can only be good as you have reached a higher sample size and thus you can make conclusions about your population with even more certainty and accuracy.

    You are all fine here!

    KR

    Willem

    Kristen - January, 2015

    Thanks for the quick reply! Would you also happen to know what sampling method this would be considered? I know it is a non-probability method, but I can’t seem to find a match: total population sampling, consecutive sampling, available sampling? Would my ability to generalize not be compromised by the low response rate?

    Thanks again!

    Willem Schrijver - January, 2015

    Kristen, your low response rate is not an issue as you chose to contact a very big population and the higher the population the relatively lower your sample size needs to be. I am afraid I can’t help you with the terminology.

    KR

    Willem

  • Rebaz - January, 2015 reply

    dear in charge …..
    i have 1500000 population, how can i decide to taking sample. can i just distribute only 100 questionnaire forms?
    best regards

    Willem Schrijver - January, 2015 reply

    Dear Rebaz,

    You will first need to decide on the confidence level and margin of error and then you can calculate the amount of questionnaires.

    User our calculator for this: https://www.checkmarket.com/market-research-resources/sample-size-calculator/

  • Milan - January, 2015 reply

    Hi.

    I am doing psychology research and I need a representative sample. The population is cca 3000 and my sample size is cca 120. I know it is very small when I used calculator but… My population is very homogeneous. The same age, the same education… They are all psychology students at the Universities… And I just want representative sample for this specific population… If the population is very homogeneous I know my sample could be smaller and still representative… Is it OK to use sample size 120 and say that my sample is representative for this specific population?

    Thank you.

    Willem Schrijver - January, 2015 reply

    Dear Milan,

    For a population of 3000 and 95% conf interval, 5% margin of error you will need 341 respondents for it to be representative, so I am afraid it is not enough.

    When puting the population in our calculator (https://www.checkmarket.com/market-research-resources/sample-size-calculator/) the assumption will always be that this population will be homogeneous.

    KR

    Willem

  • yohannes - January, 2015 reply

    hello sir
    i a going to be work my research with the sample size of 383 respondent. but i got others numbers of studies which had large number of population size than main and they took sample size between 200 to 267. how can i make as they did? please help me

  • Brian - December, 2014 reply

    Any chance you could help with this problem?

    I am pulling meta data fields from a population of 100M documents. These meta fields include things like Author Name, Author Location, Keywords, etc. Some documents have all of these meta fields, some documents have a few of these fields, and some documents don’t have any. We have no idea what the meta field coverage rates are for the population, but would like to know before we process the entire set ( eg. 45% of documents include an Author Name ).

    I’ve determined that I need to analyze 166 random documents to be 99% confident that my sample represents the population ( 99% Confidence Level ), while having a Confidence Interval of plus or minus 10% on the coverage rates we discover.

    Intuitively, however, this does not seem like the correct method. If a meta field occurs very rarely in the full population, it seems like I’d have to adjust my population or change the Confidence Interval for that Meta Field. My input data seems fundamentally different than a yes/no response to a survey.

    How should I design this test to get valid coverage rates for meta fields representative of the complete population of documents?

    Willem Schrijver - January, 2015 reply

    Dear Brian,

    Do I understand it right you have a population of 100.000 persons? First of all, I must assume that each of these documents respresents one unique person. I assume also that the order of these documents are at random, so the order has no influence on the results when picking out a sample.

    If these elements are respected you can use our online calculator to determine the sample size you need to get respresentative results for your population – in case of 100.000 population, 99% confidence and 5% margin of error you will need to check 660 documents.

    Our calculator: https://www.checkmarket.com/market-research-resources/sample-size-calculator/

  • Mesfin Eshetu - December, 2014 reply

    I want to study Determinants of Seasonal Migration in one woreda of Ethiopia with sampled kebele of three with house hold of 2473,so what type of Formula i can used in 95% accuracy and 5% errors and the sampled use hold be ranged between 120-180.

  • Passah Titus - December, 2014 reply

    Formula that is used to calculate sample size.
    Take a community of a population of 1200.
    what will be the ideal sample size of using the formula?

    Willem Schrijver - January, 2015 reply

    Dear Passa, please use this calculator: https://www.checkmarket.com/market-research-resources/sample-size-calculator/

  • Passah Titus - December, 2014 reply

    Formula that is used to calculate sample size.
    Take a community of a population of a population of 1200.
    What will be the ideal sample size of using the formula?

  • abdul - December, 2014 reply

    Hello , I am doing PHD research in e-commerce ,using survey and ,y respondents all general , I want to make my sample size 304 , plz help me to know how I can get my size 304 ?

  • frimpong - December, 2014 reply

    dear sir
    what formula can i use to calculate a sample size when only population is given

  • frimpong - December, 2014 reply

    Commentsir
    i was given only population of 286. how can i calculate the appropriate sample size?

  • frimpong - December, 2014 reply

    sir
    i was given only population of 286. how can i calculate the appropriate sample size?

    Willem Schrijver - January, 2015 reply

    Dear sir, please use this calculator tot determine your sample size: https://www.checkmarket.com/market-research-resources/sample-size-calculator/

  • Rayan - November, 2014 reply

    Hello There. I am planning to conduct a repetitive sampling in four groups of animals at two locations in every two months interval for three times. If my population is 8500, what are the best statistical tools that I can use? How do I calculate the sample size? Thanking you in advance for your assistance.

    Willem Schrijver - January, 2015 reply

    Dear sir, please use this calculator tot determine your sample size: https://www.checkmarket.com/market-research-resources/sample-size-calculator/

  • John Brown - November, 2014 reply

    We conducted a survey with a population of 300 with 147 who actually responded. Which gives 5.78% margin of error ( Confidence level: 95%)

    Would the results and conclusions in the survey ( social survey) be credible? Based on given results? What would be considered in general a satisfying margin of error in social surveys?

    John

    Willem Schrijver - December, 2014 reply

    Dear John, first you will need to report your margin of error clearly, so that readers can interpret your results with carefulness.

    You answers should be interpreted as follows:
    Suppose 70% of your respondents said they are satisfied with customer care. With a 95% confidence level and a margina of error of 5.78%, 70% of the time between 70%- 5.78% and 70%+5.78% of the population is satisfied with your customer care.

    As your margin of error is quiet close the the widely accepted 5% you will likely get satisfactory results about your population. However as it is not within the 5% range, you will need to be careful interpreting your results.

  • sanaullah - November, 2014 reply

    hi sir
    i am going to conduct a civic education baseline survey in three district but i do not know the actual population of my targeted group how can i take my survey sample size and which sample size i can flow ( target group is youth, religious leaders, minority, women, local leaders district administration) its really confusing for me please help me

    Willem Schrijver - November, 2014 reply

    Dear Sanaullah,

    Could you elaborate a bit more on what you are trying to achieve with your research? Please notice that if the three districts are considered as 3 different populations and within that you want to make subgroup in sex, age, ethnicity and religion, you will make sample design very complex, whereas this may not serve your purpose.

  • sarah - November, 2014 reply

    Sir,
    My population for my project is small–228 employees. Since I have to study the whole population what will my sample size be. and which formulation will i use. please break it that because i need to be able to write it in my research methodology

    Willem Schrijver - November, 2014 reply

    Dear Sarah, you can calculate your sample size by using our calculator:
    https://www.checkmarket.com/market-research-resources/sample-size-calculator/
    Kind regards
    Willem

  • ENOMA - November, 2014 reply

    i”m researching on the prevalence of HIV, TB and HIV/TB co-infection between 2012 and 2014 using two public hospital records. please i want to know how to get my sample size and calculate for it as well

  • mano - November, 2014 reply

    Plz tell how we can determine sample size for simple random sampling by keeping different factors in considerations ? I have heard that there are calculators that are especially use for sample size calculation but i didn’t find any, if there exsist any can u name of them?

  • AYESHA MIRZA - October, 2014 reply

    HI THERE I HAVE A QUESTION ;
    CRITICS OF PROBABILITY SAMPLING OBJECT THAT HOW CAN 300 OR 400 RESPONDANTS REPRESENT THE VIEWS /OPINION/ ATTITUDES OF A POPULATION AS BIG AS THE WHOLE CITY OR ANY OTHER BIG GEOGRAPHIC AREA POPULATED BY A HUGE NUMBER OF PEOPLE .WHAT DO U THINK ABOUT THIS CRITICISM

    Willem Schrijver - November, 2014 reply

    Dear Ayesha Mirza,

    I surely understand this criticism and In case of doing a respresentative study of a city, a researcher needs to critically reflect on which segmentationparameters matter eg education, gender and/or age. Once you have decided that a factor like age matters, you will have to base your sample size on these subgroups in your population. You will need to calculate your sample size of each of the subgroup and thus your sample size will also increase heavily and will not be 300-400

    See also a previous comment of my colleague in this blog post:

    Rather smart mathematicians have provzen that the size of the population does not matter that much. As a result, a sample of, say, 1000 respondents is equally sufficient for a population of 10.000.000 citizens as for a population of 100.000 citizens. For more mathematical information, I can recommend the book “Statistics” by Freedman, Pisani and Purves (2007).
    However, you should take the number of sub-groups into consideration when determining your sample size. Each sub-group should be large enough for valid estimation of parameters. As a result it is advisable to treat each sub-group as a population and determine your sample size accordingly.

    Kind regards,

    Willem

  • sintayehu - October, 2014 reply

    here i am planning to study an effectiveness of an exstension tool to diisiminate a techenology adoption .in the study districts there are 22 districts among them i select purposely district A since the district is adopting a techinology(precence of techinology). the selected districts consists of 6 villages with different popoulation size and i want to sample 3 villages and 150 samle size .how can i do this?

  • Shajan - October, 2014 reply

    Hello, I would like to conduct a baseline survey of food security and livelihood recovery project. There are 1300Households under the project. How will I calculated sample size? I know about calculator software but i want to calculate it manually .

  • Shajan - October, 2014 reply

    Hello, I would like to conduct a baseline survey of food security and livelihood recovery project. There are 1300 under the project. How will I calculated sample size? I know about calculator software but i want to calculate it manually .

    Willem Schrijver - November, 2014 reply

    Dear Shajan,

    Here is an example that came up earlier in this post:

    First you calculate the sample size (SS). If you do not know your proportion p, it is quite common to take 0,5 as a value for p.
    SS = (Z-score)² * p*(1-p) / (margin of error)²
    Then you correct it if you know your population:
    SSadjusted = (SS) / 1 + [(SS – 1) / population]
    If we use it in an example with a population of 100.000, confidence level of 95% and margin of error 2%, then you get the following calculation:
    SS = (1,96²) * 0,5*0,5 / 0,02²
    SS = 3,8416 * 0,25 / 0,0004
    SS = 0,9604 / 0,0004
    SS = 2401
    SSadj. = 2401 / 1 + [(2401-1) / 100.000]
    SSadj. = 2401 / 1 + [2400 / 100.000]
    SSadj. = 2401 / 1 + [0,024]
    SSadj. = 2401 / 1,024
    SSadj. = 2344

    Kind regards

    Willem

  • Erica - October, 2014 reply

    Hi,

    i did an origin-destination survey at selected airports. The most important question were “real origin” and “real destination”, response being a city. There’s the obvious bias of respondents who live in the cities where the survey was conducted, ie, were the airports are located.

    I want to know how to estimate the average number of respondents per city, considering the aforementioned bias. I also have the current population of each city mentioned as origin or destination and the air routes sampled.

    Thank you

  • Paula - October, 2014 reply

    As part of my master’s thesis I distributed a survey. I decided to use a non-probabilistic sampling . I only got 24 answers. I have estimated the population to be 90 companies, and I got answers from 24.
    The committee who is evaluating my thesis said I should make sure my analysis is representative of the population. Otherwise all my work does not make sense. Is this 27% representative?

    Willem Schrijver - November, 2014 reply

    Dear Paula, you can calculate your sample size by using our calculator:

    https://www.checkmarket.com/market-research-resources/sample-size-calculator/

    Kind regards

    Willem

  • sowmya - October, 2014 reply

    i am pursuing my ph.d in studying the satisfaction of an airline,wher the size of population is unknown and will be a large number,say 2,60,000.i have collected a sample of 517 in 8 months.will my research be reliable.or how should justify the taken sample

  • sowmya - October, 2014 reply

    i am pursuing my ph.d in studying the satisfaction of an airline,wher the size of population is unknown and will be a large number,say 2,60,000.i have collected a sample of 517 in 8 months.will my research be reliable.

  • nathan - October, 2014 reply

    The population for my study is 1127 and I intend to use 300 respondents which 27% of the population. Is it appropriate for the sample?

    Willem Schrijver - November, 2014 reply

    Dear Nathan,

    As explained in the article, your sample size not only depends on your population but also on the margin of error and the confidence level you choose. In your case based on a 95% confidence level and a margin of error you will need 287 respondents, so your sample size of 300 is appropriate. Please do mention the confidence level and margin of error in your reporting to be scientifically correct.

    Kind regards

    Willem

  • Kristyl - October, 2014 reply

    If you have a 280 population and you wanted to get a 20% sample size of this population, can that be justifiable and will there be any restrictions?

    Thank you for your response. :)

    Willem Schrijver - November, 2014 reply

    Dear Kristyl,

    With a population of 280 you will need 163 respondents in order it to be representative with 5% margin of error and 95% confidence level. A sample size of 20% (=56) cannot be enough to justify your results are representative for the population I am afraid.

    Kind regards

    Willem

  • pranali - September, 2014 reply

    hi,
    My total population size is 225 of all HR Managers and their immediate subordinates. I am going to conduct online survey.
    what should be my sample size? considering the 20% turnout ratio, i must send survey forms to 600+ people. But, here i am having only 225 as my true sample. Please clarify and suggest.

    Willem Schrijver - November, 2014 reply

    Dear Pranali,

    Please have a look at our calculator (https://www.checkmarket.com/market-research-resources/sample-size-calculator/) and specifically look at the results that are displayed in the last row of the calculator. Depending on what margin of error you choose and what confidence level I get a number of to be invited respondents ranging from 715 – 1115
    Kind regards
    Willem

  • anna victoria - September, 2014 reply

    Hello sir I just want to ask if this applies to studies on mice. I’m a med student and we have a study on clinical pharmacology regarding the significant differences and effects of spinach and metformin in lowering blood sugar on mice subjects. We have 40 to 50 mice, and our prof asked us the rationale and basis behind choosing 40 to 50 subjects is there a guideline regarding sample sizes on mice studies? Thank you so much sir.

  • shadi - September, 2014 reply

    Dear Sir, greeting
    I would like to ask about my sample size. Actually I have conducted my study among manufacturing firms the total number is 1800 factories and based on the Cochran’s (1977) formula it should be 317 . But I could receive 214 completed factories after 8 month data collection data collection which is cover about 68% of response rate. So I would like to know is it enough to generalize or not and please let me know how I can justify this number thanks

    Willem Schrijver - November, 2014 reply

    Dear Shadi,
    I am afraid that if you have 1800 factories, the minimum of factories you will need based on 5% margin of error and 95% confidence level results in 317 factories indeed. With a sample size of 214 respondents your margin of error will be higher than the widely accepted 1-5% range, thus making your results deviate too much from the opinions of the entire population. So you can either try to get these extra 100+ factories to participate in your research or at least clearly mention the margin of error is higher than accepted in your report so that readers know how to interpret your results.

    Good luck!

    Willem

  • Cassandra - September, 2014 reply

    Hi Didier
    I am looking at doing a survey with clinicians within health and what their attitudes are towards patients, I am looking at recruiting staff from the hospital I work in both inpatient and outpatient. Whilst the subjects will be taken from one hospital setting – which will be a limitation I am wondering if there is a specific number I should be working on to be able to do this research. Could I do less than 100 or would I need to find more? Additionally since I will be recruiting medical and allied health staff they will be a specific sample group of this population and not randomly chosen, is there a known terminology for this kind of sample group? Regards

  • Sara - September, 2014 reply

    Hi Didier:
    I need to calculte the sample size for a genetic disease with a prevalenc of less than 1% in 1000,000 people in the world. I dont know how to calculate my sample size. I dont need a control group. I just need to research that between of Iron deficiency anemia patients , what percentage of these patients have Iron refractory iron deficiency anemia? so for this research , I need to calculate my sample size first. Would you please kindly help me?
    Thank you so much in advance and best regards.

    Willem Schrijver - November, 2014 reply

    Dear Sara,

    I suggest you look at medical data of your country to check how many people have are diagnosed with this disease and based on this number you can calculate you sample size. In case your country has a million inhabitants it is maximum 10,000 people that have a genetic disease X and with our calculator you can further calculate the sample size (https://www.checkmarket.com/market-research-resources/sample-size-calculator/)

    Kind regards

    Willem

  • Ricardo - September, 2014 reply

    Please help me out ..l if a survey A response rate is 40% and survey B response rate is 80%, is the overall response rate = 60%?

    Thanks

    Didier Dierckx - September, 2014 reply

    Hi Ricardo,

    Thanks for your question.

    To answer your question, no, that is not automatically the case. To calculate your overall response rate, you should divide your total number of responses by the sum of the number of invitations of survey A and B.

    Good luck!
    Didier

  • Omid - September, 2014 reply

    dear Didier,
    thanks a lot for your amazing information, it helped a lot to clear a lot of issues that I have with near to submit research :) a quick question, according to CL of 95% and ME of 5% I get 384 population sample for a population of 1200000. My population is university students. does 384 makes sense for such a large population ? thanks in advance for your help

    Didier Dierckx - September, 2014 reply

    Hi Omid,

    Yes it does if you speak in your findings about the university students in general. If you start making segmentations between e.g. men/women, different majors, different universities, etc. then it will not be enough.

    In your data collection you should treat each subgroup (e.g. university) as a population. So, if you want to make a comparison between universities, you should probably collect 384 respondents per university.

    Good luck!

    Didier

  • ernest - September, 2014 reply

    please how do i calculate the sample size for an undefined population? for instance if i want to study the relationship between two health procedures done on humans, what is the formula for the sample size?

    Didier Dierckx - September, 2014 reply

    Hi Ernest,

    Below you find the formula for calculating your sample size for an known population:

    First you calculate the sample size (SS). If you do not know your proportion p, it is quite common to take 0,5 as a value for p.

    SS = (Z-score)² * p*(1-p) / (margin of error)²

    Good luck!

    Didier

  • Evans Perry-Quao - September, 2014 reply

    Some maths:
    Total population: 5,000 pupils
    KG to JSS3 = 11 levels = about 450 pupils per level
    If focusing at KG and P1-P3 = 5 levels = 2,250 pupils (this would make sense, because some pupils with SpLD never reach P4 or higher classes; SAP data shows that many out-of-school children are 13 years and in P3…)

    ‘At risk’ assessment takes about 45 minutes.
    If on 2 afternoons per week 2 teachers conduct assessments, this means 2×2 or 2×3 pupils per afternoon. This means 8 to 12 pupils can be assessed per week.
    Period of 12 weeks = 96 – 144 pupils to be assessed

    If all pupils are from KG – P3, it means 4.2 – 6.4% of 2,250 pupils are assessed. If from all levels, they represent 1.9 – 28% of the total population of 5,000.
    These pupils will be pre-selected based on observations by Resource Teacher and class teachers.
     Will this be valid and representative for the total population of 5,000?

    Didier Dierckx - September, 2014 reply

    Hi Evans,

    Thanks for your questions.

    Even though I am not fully up to speed with the American school system, the numbers speak for themselves.

    I am afraid that your sample will not be large enough, unless you are willing to accept a margin of error of approximately 8%. So I would try to increase the sample.

    Furthermore, I should stress that a large enough sample size does not guarantee a representative sample. To find out whether your survey sample is representative for the population you should compare the distribution of some demographic characteristics between your survey sample and the population.

    Good luck with your research!

    Didier

  • Utam - August, 2014 reply

    I am trying to draw up a sample from an unknown population of retail stores customers. I do not have the resources for a large sample size. Please how do I go about this?

    Didier Dierckx - September, 2014 reply

    Hi Utam,

    Your sample size is especially important if you want to generalize your findings to the population.

    A small sample size does not prevent you from doing some statements about your findings, but it does prevent you from generalizing your findings to the population. For instance you cannot say that X% of the retail stores customers are satisfied with the service. However, you could say that X% of the retail store customers you surveyed are satisfied with the service.

    I hope this helps!

    Regards,
    Didier

  • Innocent - August, 2014 reply

    How do i calculate a sample size when given population size and confidence level only?

  • Dylan - August, 2014 reply

    Hi Didier,
    I am conducting a study based on responses from middle managers at a number of organisation, while I do not have an exact figure as to how many middle managers are in the entire population, I handed out 130 questionnaires and received a total response of 98.
    I am just curious as to how do i justify the sample size used and number of organisations selected in my small island economy.

    Didier Dierckx - August, 2014 reply

    Hi Dylan,

    That’s an interesting question.

    I am afraid it will be quite hard to convince the skeptic reader of your study that your sample is large enough. It might be, but – as long as you do not have any information on the number of middle managers – it is impossible to know.

    I would advise you to look for more information on the number of middle managers.

    Good luck!
    Didier

  • evelyn - July, 2014 reply

    Hi Didier
    I am conducting a survey questionnaire study comparing 2 groups.. I am confuse on the calculation.
    Example : Total respondents from two ( x and y groups) are = 303. However for one particular question, the total respondents for the “x” group was 180 ( yes = 90, no= 40, non applicable = 30, missing value = 20).

    My question is : Which calculation should i use to calculate the total % of “yes” respondents for the “x” group :
    a) includes “non applicable” and “missing value’: 90/180 =50%, or… b) 90/303 =20.70%?

    Thank you very much for your kind help

    Didier Dierckx - August, 2014 reply

    Hi Evelyn,

    Thanks for your question.

    Your base for your calculation should be the respondents that answered the question (= 180). Whether you include the non-applicables is up to you. If it is ‘interesting’ to know for your research, then I would include them. HOwever, I would not include the ‘missings’. The base value for this question is thus 160.

    Hope this helps,
    Didier

  • LG - July, 2014 reply

    what do you do in a situation where you are not sure what is the population size? say patients who go to Cape Town for Cosmetic surgery?

  • Tony - July, 2014 reply

    Please I meant to ask that, can I administer my questionnaire at the offices of the DVLA in only one region? Who will be the population size, will it be all people eligible to drive or register vehicles, or the average number of people who visit the DVLA office in a year? Kindly help me

    Didier Dierckx - July, 2014 reply

    Hi Tony,

    Thanks for your questions.

    I would argue that your population is every Ghanaian that has been in contact with the Driver and Vehicle Licensing Authority. Presumably, this will only be Ghanaians that have a car or driver’s licence.

    To make your project manageable, it is perhaps wisely to limit yourself to those Ghanaians that have been in contact with this authority in the past year.

    If you want to make general statements about the service quality of the Driver and Vehicle Licensin Authority, you should argue that your sample of the offices you study is representative for all offices.

    Good luck!
    Didier

  • Tony - July, 2014 reply

    Sir, Kindly help me with this problem. I am to study the service quality of the DRIVER AND VEHICLE LICENSING AUTHORITY in Ghana. Will the population be all people eligible to drive i the country? Again, can I limit my survey to only out of the 10 regions in the country as every region have more than 5 offices.

  • neo - July, 2014 reply

    Hi, sir, what is my study is face to face distribution approach? is that need to follow the sample size table as above?

    Didier Dierckx - July, 2014 reply

    Hi neo,

    Thanks for your question.

    Yes, you should follow the same sample size table as above. The distribution method does not impact the sample size needed for your project.

    Regards,
    Didier

  • Joe Bryant - July, 2014 reply

    Please help me to derive the formula and explaination of selecting the mimimum sample size from the population of 5,000 people.

    Didier Dierckx - July, 2014 reply

    Hi Joe,

    You can use our calculator (https://www.checkmarket.com/market-research-resources/sample-size-calculator/) to calculate the minimum sample size for your population.

    Alternatively, you can use the following formula to calculate the sample size manually:

    SS = (Z-score)² * p*(1-p) / (margin of error)²

    Good luck,
    Didier

  • Umair - June, 2014 reply

    In case when population size is not known then how to calculate the sample size. Is there any method to calculate sample size ?

    Regards
    Umair

    Didier Dierckx - June, 2014 reply

    Hi Umair,

    Thank you for your question.

    If you do not know your population, you can use the formula below to calculate your sample size:

    If you do not know your proportion p, it is quite common to take 0,5 as a value for the stand. dev.

    SS = (Z-score)² * p*(1-p) / (margin of error)²

    Regards,
    DIdier

  • Vincent - June, 2014 reply

    What if you have a population of 20,000 very homogeneous population exposed to the very same service levels, how do you get get representative sample with least cost?

    Didier Dierckx - June, 2014 reply

    Hi Vincent,

    First of all, I must stress that the sheer size of your sample is in no way a guarantee for being representative of your population.

    To find out whether a sample is representative, you should compare the distribution of your sample on e.g. gender, education, age, etc. with the distribution in the population.

    Regards,
    Didier

  • Millicent - June, 2014 reply

    Am conducting a cross sectional descriptive survey. It will cover internal and external stakeholders. How do I determine the study population and sample size for the study?

    Didier Dierckx - June, 2014 reply

    Hi Millicent,

    Could you please provide me with more details as I am not too sure what you exactly mean.

    Thanks,
    Didier

  • Michael - June, 2014 reply

    Hi. I would wish to find out how the sample size formula makes provision for the population. The formula doesn’t make provision for the variable “P” (population). However, a space is provided to key in the population if known and the other variables (confidence level, margin of error). And so, what is the formula for the sample size determination and where you make use of the population. Thanks

    Didier Dierckx - June, 2014 reply

    Hi Michael,

    As you probably know, you do not have to know your population size to calculate your sample size. However, you can correct your formula if you know your population.

    Below you find the formula for calculating your sample size for an known population:

    First you calculate the sample size (SS). If you do not know your proportion p, it is quite common to take 0,5 as a value for p.

    SS = (Z-score)² * p*(1-p) / (margin of error)²

    Then you correct it if you know your population:

    SSadjusted = (SS) / 1 + [(SS – 1) / population]

    I hope this helps!

    Regards,
    Didier

  • FAIS - May, 2014 reply

    Am doing a study in adolescent nutrition in girls.I need to find out if i can use 10% of my population size as a subset and estimate their blood levels of ca, my calculated population for undernutrition is abt 1200 and i want to just sample 10% of them, any refrences to back it up.

    Didier Dierckx - June, 2014 reply

    Hi Fais,

    Thank you for your question.

    I would not “just sample 10% of them”. This choice for 10% is quite arbitrary.

    You can use our calculator to calculate the correct sample size needed for your research: https://www.checkmarket.com/market-research-resources/sample-size-calculator/

    Regards,
    Didier

  • Ina - May, 2014 reply

    Hi Didier,

    I just want to know, in case when I use other country’s proportion, say smoking prevalence (30%) of school children to calculate my sample size with 95% CI, do I need to adjust the population (eg the size of the school children in my country), assuming that both countries will have different size of population? If so, how is this done? Can I just use the proportion directly in the calculation using the formula without the finite population correction?
    How will i calculate my sample size?

    Thanks,

    Didier Dierckx - June, 2014 reply

    Hi Ina,

    Thank you for your interesting question.

    Yes, you should use the proportion of your own country. Furthermore, I think you will have to argue very clearly why you would expect that the smoking prevalence is the same in both countries …

    You can calculate your sample sizes using our calculator: https://www.checkmarket.com/market-research-resources/sample-size-calculator/

    Good luck!

    Didier

  • Kerry McColgan - May, 2014 reply

    Hi, I am currently carrying out a statistical survey to determine if females perform better than males in STEM subjects. However the numbers of males and females sitting each subject is different, so does the fact that my population sizes are different affect the way in which I can compare the means? Would it mean my results are flawed?

    Didier Dierckx - June, 2014 reply

    Hi Kerry,

    Thanks for your question.

    As the t-test takes account for the standard error of the estimates, you can use this test for comparing the means. Even if the sample sizes are unequal.

    Regards,
    Didier

  • Kabins - May, 2014 reply

    Dear Didier,

    Thanks for the help that you offer. It is really appreciated.

    In my case; the sample size is determined through Yamane’s formula which is (400). However, the responses that I have got for my research are 290. Is the response rate acceptable? I don’t know what should I do because I really suffered just to get this response rate! Your advise is really appreciated.

    Thanks.

  • Siti Zubaidah - May, 2014 reply

    Dear Sir,
    I have study the birder in Malaysia but my problem is do not know the population. For this study, the population is very small. I have surveyed about 144 respondents. Can you help me to estimate how many my population and are there 144 sample is representative to the population? how many percentage? Before this my estimation is about 300 birder in Malaysian but is not actual number. Thank you very much!

    Didier Dierckx - June, 2014 reply

    Dear Siti Zubaidah,

    You can calculate the appropriate sample size (without actually knowing your population size) via the formula below:

    First you calculate the sample size (SS). If you do not know your proportion p, it is quite common to take 0,5 as a value for p.

    SS = (Z-score)² * p*(1-p) / (margin of error)²

    Then you correct it if you know your population:

    SSadjusted = (SS) / 1 + [(SS – 1) / population]

    Good luck!

  • Maryam - May, 2014 reply

    Dear Sir
    which software can estimate sample size? please tell all software which can calculate.
    thank you so muh

    Didier Dierckx - May, 2014 reply

    Dear Maryam,

    You can calculate your sample size for instance via our online calculator: https://www.checkmarket.com/market-research-resources/sample-size-calculator/

    Good luck!

    Didier

  • Raja - May, 2014 reply

    Hi sir,

    we plan to conduct a political survey all over the state Tamil Nadu in India. can you please suggest me with sample size,confidence level and interval…etc required for survey.
    Population is 50000000 (50Million)

    kindly mail me..

    thank you in advance

    Didier Dierckx - May, 2014 reply

    Hi Raja,

    Thank you for your question.

    It depends on how accurate you want your survey data to be. I would suggest a confidence level of 95% and a margin of error of 2 or 3%.

    Alternatively, you could look for earlier scientific work on political surveys in India and see which criteria they used.

    Via our calculator, you can calculate different scenario’s: https://www.checkmarket.com/market-research-resources/sample-size-calculator/

    Good luck!

    Didier

  • Shoaib - May, 2014 reply

    Hi Didier,

    My query is with respect to the table used for Calculating the survey sample size (number of respondents needed), i am bit confused with the Population Size & Number to invite , when i put value of 500 in population size , the number invite gives me value of more than 1000, how is it possible to send invite to 1000 when the total population is of just 500, bit confused , didn’t get the point actually, ,i am not stats guy :( , thanks

    Didier Dierckx - May, 2014 reply

    Hi Shoaib,

    This is in fact a good question.

    If you put in a population size of 500 – with a margin of error of 5% and confidence level 95% – your sample size will have to be 218. This is the number of respondents you need.

    However, you should also take the response rate into account. This is the ratio of respondents that fill in the questionnaire they received compared to the total number of surveys you send out. For instance, if you send out your survey to 400 people and you receive 200 filled in surveys, your response rate is 50%.

    If you have an estimate response rate of e.g. 20%, the ‘number to invite’ will be larger than your population. Practically, this means that you will have to invite your entire population.

    I hope this helps!

    Regards,
    Didier

  • Hope - May, 2014 reply

    I use a five percent (5%) of the population as my criteria for selecting my sample. one of the village was a having 1880 households and the other village 578 households. From 1880 I took 94 households, and from 578 I took 56- here I doubled the percent because I thought the number was too small for comparison between the two villages. What is the impact of my selection to generalization of my conclusions?

    Didier Dierckx - May, 2014 reply

    Hi Hope,

    This means that your margin of error will be larger. For the village of 1880 households this results in a margin of error of 9,85% and for the other village 12,46%.

    You should definitely report these numbers as they will allow the reader of your study to interpret the results.

    Regards,
    Didier

  • monika - May, 2014 reply

    hi, sir i had collected the data of using electric equipment in Ajmer a big city i have 300 surveyed forms threw which i have calculated total energy consumption by individual equipment now i want to check my data is correct or not and at which level it is correct like 95% or 99%

    Didier Dierckx - May, 2014 reply

    Hi Monika,

    Thank you for your question.

    The table in the article shows that you approximately need 400 respondents (confidence level 95% and margin of error 5%). You are slightly below this target.

    As long as you do not segment your data (i.e. look for differences between different groups of respondents), I think the data will be fine. You can check the margin of error using our calculator: https://www.checkmarket.com/market-research-resources/sample-size-calculator/

    Good luck!

    Didier

    monika - May, 2014

    thank you very much for your answer

  • Kurt - April, 2014 reply

    Hi, doesn’t number of questions on survey (degrees of freedom) also help determine the requisite sample size given confidence interval and confidence level. If I have a 40 question survey vs. a 2 question survey, it seems to me if I’m doing some basic statistical tests that I’m going to need a larger sample in order to test each of the 40 questions — I’m thinking of chi sq tests, etc. If so, is it a materially different sample size? If so, how do I feed it into calculations.

    Didier Dierckx - May, 2014 reply

    Hi Kurt,

    Yes it does, however this is especially relevant with smaller samples, as they tend to underestimate the population value.

    If you sample is large enough (you can use our calculator: https://www.checkmarket.com/market-research-resources/sample-size-calculator), this will not affect your analysis.

    Hope this helps.

    Didier

  • Ashima - April, 2014 reply

    Dear Didier,

    Does the example and calculator provided by you cuts across kind of tests. I mean if I am looking forward to doing a Triangular Test methodology in research, should I take a more robust sample. Please advise.

    Thanks,
    Ashima

    Didier Dierckx - May, 2014 reply

    Dear Ashima,

    Thank you for your interesting question.

    I am unfortunately not familiar with Triangular Test methodology, so I am afraid I am not able to say whether our calculator supports this kind of test.

    Good luck with your research!

    Didier

  • Lorena - April, 2014 reply

    Sir,
    Hi! Please help me. My target population are married people in one city.The total population is 1,400,767. Research teacher advised me to have 30% of it. That’s too much?Or she said 20 respondents per variable (13×20=260). Then I saw your way of getting the right sample.I’m confused. I think yours is scientific.I’m into Psychology. Thanks.How many sample do i need in 1,400,767 population?

    Didier Dierckx - April, 2014 reply

    Hi Lorena,

    Thanks for your question. I am afraid both approaches are ‘wrong’.

    You can calculate the sample size via our calculator (https://www.checkmarket.com/market-research-resources/sample-size-calculator/) or via the formula below:

    SS = (Z-score)² * p*(1-p) / (margin of error)²

    Then you correct it if you know your population:

    SSadjusted = (SS) / 1 + [(SS – 1) / population]

    If I use your data, a sample (using a margin of error of 5% and confidence level of 95%) of 385 would be sufficient. However, if you want to segment between different socio-demographic groups (e.g. on educational level, neighbourhood, etc.) you will need a larger sample.

    I hope this helps!

    Regards,
    Didier

  • blitz ryan earl flores - April, 2014 reply

    good day sir..

    i would like you to help me on my feseability studies, it is about a commercial building and it has a office for rent and the next floor is a dormitory… i am so confused on what will i use, for the sloven formula is not applicable on it coz i have no population to determine my sample size…i need someone whom is not living on this place for i need is someone who lives far and whom is only renting.. so the problem is i dont know how do start this one… and it so urgent…

    thank you sir…

    Didier Dierckx - April, 2014 reply

    Dear sir,

    Thank you for your question.

    I am afraid I will not be able to help you. I am not familiar with this type of study.

    Good luck with your research!

    Kind regards,
    Didier

    blitz - April, 2014

    we are done now sir… we solve it by using the purposive sampling technique….. thankz

  • mamata - April, 2014 reply

    Namaste Sir, I am doing one masters’research where i am taking all the population example pregnant women attending for antenatal check up which sampling method it will called as….. thank you sir

    Didier Dierckx - April, 2014 reply

    Dear Mamata,

    Could you please explain your question a bit further?

    You can calculate your sample size via: https://www.checkmarket.com/market-research-resources/sample-size-calculator/

    Regards,
    Didier

  • Phil O - April, 2014 reply

    Hi, what about VALIDATION studies? I mean, if you want to validate a new questionnaire (Q) by a gold standard one, and let assume that you want to examine how well the new Q estimates the mean value of a variable of interest (compared to the mean as obtained from the gold standard Q). What should be the sample size if you design the validation study in order to apply both questionnaires in the same individuals (of the estimated sample size)? I think this differs a bit from your proposed calculations. Any reference will be highly appreciated. Thnx for your time and congratulations for your work!

    Didier Dierckx - April, 2014 reply

    Hi Phil,

    Thanks for that very interesting question!

    As far as I am aware, there are no fixed rules regarding the sample size for validation studies.

    However, you should make sure that you use a representative sample compared to the population that you will be surveying. Also, make sure that you have a large enough sample if you want to validate latent variables through factor analysis.

    I do not know what field of study you are active in, but if you go to Google Scholar you will definitely find some earlier work.

    I hope this helps.

    Regards,
    Didier

  • Ali - April, 2014 reply

    I work in a call-centre where there are 73 agents, who on average take 600 calls a month. A third party company performs monthly customer satisfaction surveys and each agent gets scored on 6 out of the 600 calls a month. I’m trying to argue the point with management, on behalf of all agents, that 6 out of 600 responses is not a large enough sample to get a true reflection of an agents performance. Am I correct?
    However, I take it from the table above that a response of 1% (73*6=43873) completed surveys for a population of 43, 800 (73*600) is sufficient to give a true reflection on customer satisfaction with the company itself?
    Thank you for your help!

    Didier Dierckx - April, 2014 reply

    Hi Ali,

    Thanks for your interesting question. You are right on both accounts.

    A sample of 438 completed surveys is large enough to make statements about the company in general (i.e. all 73 agents). For instance: “X% is satisfied about the general consumer friendliness of our agents”.

    However, it is certainly not large enough to make statements about individual agents. In order to be able to make statements about an agent, you actually have to treat the 600 calls of each agent as a population.

    I hope this helps?

    Good luck!
    Didier

    Ali - April, 2014

    Thank you so much for your help!

  • Darja - April, 2014 reply

    Help please. I have already conducted my primary research, however since I didn’t know the population, I decided to do as many questionnaires as possible. I only had a short amount of time. I have 67 questionnaires, which is a low statistical realiability. (To support my findings I did interviews.)
    What is my confidence level and what is my confidence interval?
    Thanks a million.

    Didier Dierckx - April, 2014 reply

    Hi Darja,

    You do not always have to know your exact population for calculating the margin of error, etc.

    However, you do need to know whether your sample was drawn from a population larger than 1.000.000 or not. If is drawn from a smaller sample, I am afraid that you will need to know your population size as you have to correct for the ‘finite population size’.

    Regards,
    Didier

  • Alzira Pedro - April, 2014 reply

    Hi! Please Sr help me I need to finish my assignment and I have one quetion is : How I shoul sample my population in three group…

  • Alzira Pedro - April, 2014 reply

    Hi! Please Sr help me I need to finish my assignment and I have one quetion o How I shoul sample my population in three group…

    Didier Dierckx - April, 2014 reply

    Dear Alzira Pedro,

    Thank you for your questions.

    If you want to sample three different groups, you should calculate a sample for each group. As such you will have the guarantee that your overall sample will be big enough.

    You can calculate your sample via: https://www.checkmarket.com/market-research-resources/sample-size-calculator/

    Good luck!

  • Sam - April, 2014 reply

    Hi Didier… I need advice please.
    I approached 150 companies to complete a online survey for my thesis. Only 51 completed.

    I then tried to do interviews by approaching the 51 who completed the survey. Only 5 agreed but I received detailed information, the remainder of the 51 outright refused.

    My supervisor said it’s not acceptable but has given no insights.
    I don’t know how to theoretically justify the survey sample size or how to justify the number of interviews?

    or how I can reduce it to make it acceptable?

    Any guidance is appreciated.
    Thank you.

    Didier Dierckx - April, 2014 reply

    Hi Sam,

    Based on your numbers, you have a margin of error of 11%. As you can imagine, this is very high. The only way to lower your margin of error is increasing your number of respondents.

    If I were you, I would …
    1) try to increase the number of companies that filled in the online survey. Rather than emailing them, my experience is that calling them is far more effective. It is easier to ignore an email than a phone call.
    2) use some relevant quotes from your interviews to highlight some of your findings.

    Good luck!

    Didier

  • frederick - April, 2014 reply

    Please , how can i determine a sample size for a research project

    Didier Dierckx - April, 2014 reply

    Hi Frederick,

    You can calculate your sample size easily via our calculator: https://www.checkmarket.com/market-research-resources/sample-size-calculator/

    I hope this helps!

    Didier

  • tilahun - April, 2014 reply

    dear sir, i am Msc student and i want to study the contribution of agroforstry to biodiversity and rural livelihood improvement in the case of one district in there villages which has a total population 1504. the survey data includes two types. one is household interview and the second is farm tree identification and counting. I have only two month to summit the paper. so, to finished i want to reduce the sample size to 120 house hold. how to calculate the sample size statically.

    Didier Dierckx - April, 2014 reply

    Hi Tilahun,

    Thanks for your question.

    You can use our sample size calculator via this link: https://www.checkmarket.com/market-research-resources/sample-size-calculator/

    Or below you find the formulas for calculating your sample size for an known population:

    First you calculate the sample size (SS). If you do not know your proportion p, it is quite common to take 0,5 as a value for p.

    SS = (Z-score)² * p*(1-p) / (margin of error)²

    Then you correct it if you know your population:

    SSadjusted = (SS) / 1 + [(SS – 1) / population]

    I hope this helps!

    Regards,
    Didier

  • Parmal - April, 2014 reply

    R/Sir, In a study stratified random sampling was used for the selection of the sample. a city has a total urban population of 1053305 with 588753 males and 513229 females. the city is divided into 60 wards of which 10 wards were selected randomly using random number number table and in each selected wards, 15 eligible participants were identified using house to house survey. this study includes all the married women who were living with their husband for last one year. the The sample size was calculated using the formula
    Sample size n = [DEFF × Np (1 − p)]/[(d2/Z2 1-α/2 (N − 1)+p (1 − p)]
    P (prevalence rate) was taken as 30% (it based on the
    fi nding of in depth analysis of NFHS-3 data by Kimuna
    et al. who showed the prevalence of physical violence
    to be 31%).(3)
    D = Absolute precision of 5%
    DEFE = design effect = 1.5
    The sample was calculated to be 121 using a 95% of
    confi dence interval. The sample was further infl ated to
    150 to round up the fi gure.
    Stratifi ed random sampling technique was used for the
    selection of the samples. Gwalior city has a total of 60
    wards which were numbered from 1 to 60 and of this 10
    wards were selected randomly using random number
    table and in each selected wards, 15 eligible participants
    were identifi ed using house-to-house visit.
    Sir , my query is that how this sample size is calculated by using the above formula.

    Didier Dierckx - April, 2014 reply

    Hi Parmal,

    I am not sure what you mean.

    Could you please elaborate a bit more?

    Regards,
    Didier

  • Janice - April, 2014 reply

    Sir, I would like to ask how to determine the sample size if the population is unknown.

    Thanks!

    Best regards,
    Janice

    Didier Dierckx - April, 2014 reply

    Hi Janice,

    The reasoning is even the other way round. You have to correct your formula if you know your population.

    Below you find the formula for calculating your sample size for an unknown population:

    First you calculate the sample size (SS). If you do not know your proportion p, it is quite common to take 0,5 as a value for p.

    SS = (Z-score)² * p*(1-p) / (margin of error)²

    Then you correct it if you know your population:

    SSadjusted = (SS) / 1 + [(SS – 1) / population]

    I hope this helps!

    Regards,
    Didier

  • harry - March, 2014 reply

    Hi, i got data of 300 students and my sample size is 80 students, how can i justify this on drug and substance use research projects

    Didier Dierckx - April, 2014 reply

    Hi Harry,

    Thanks for your question.

    What do you mean exactly by “justify this on drug and substance use research projects”?

    Regards,
    Didier

  • Tesso - March, 2014 reply

    Dear Sir/Madam,
    I am a PhD student in applied linguistics & communication. Currently, I am conducting my research on the topic: language use in inter-group communication between ethnolinguistic groups in one small multilingual town. The town has three small sub-districts & the total population of the town including these three sub-districts is 48,000. This is the permanent settlers population. I also intend to involve migrant groups from different ethnic groups. These migrants are seasonal (transitory). But, I do not have the total number of these migrant population. So, how do I select my sample? My advisor told me to take large no. of sample. Also, I got some ideas from colleagues that the good %age is to take 10% of the sample from the total population. So, my total population will be 48,000+the unknown no. of migrants(possibly small compared to permanent population). Please tell me how i will select my sample. Thank you.

    Kind Regards,
    Tesso

    Didier Dierckx - April, 2014 reply

    Hi Tesso,

    Thanks for your interesting question.

    I have given it some thought and I can understand why your supervisor advises you to survey a sample as large as possible. My only worry is that you might be surveying a group of migrants that is not representative for the total migrant population.

    I would argue that other researchers – perhaps in other fields of study – have already encountered the same issues regarding surveying seasonal workers. That is where I would start looking.

    I am sorry I am not able to help you better.

    Regards,
    Didier

  • Adetol - March, 2014 reply

    Please I need explanation on how to pick a true representative sample of market market of population over 5000 market in 6 political zone of Nigeria

    Didier Dierckx - April, 2014 reply

    Hi Adetol,

    Thanks for your question.

    As I already mentioned in an earlier comment, calculating an appropriate sample size does not guarantee representativity.

    You should find out the distribution in the population of various crucial variables (e.g. gender, age, occupation, etc.). Based on these you should start calculating your sample and make sure that the distribution in your sample roughly equals the distribution in the population.

    I would advice you to treat each political zone as a separate population and thus calculate 6 sample sizes.

    I hope this helps!

    Regards,
    Didier

  • TT - March, 2014 reply

    my question is,
    i am doing a research in hospitals who serves 230 patient per week (7 days) and i need to distribute a questionnair for the patient who is treated their in one specific week, therefore, to calculate my sample size, how much is my population?…can i take the one week total patient as a population and calcualte the sample size or i have to calculate one month data by mulitplying the (230*4) and calculate my sample size.
    Please clear me with this.
    Thanks!

    Didier Dierckx - April, 2014 reply

    Hi,

    That is a good question.

    I have given it some thought and I would argue that – if there are no clear differences on e.g. age, gender, disease, etc. between the patients in week 1 and 2 – you could calculate your sample based on 1 week.

    If there are clear differences between weeks, then you should use multiple weeks to calculate your sample.

    I hope this helps!

    Regards,
    Didier

  • Md Ashrf Al Mamun - March, 2014 reply

    Dear Sir,
    ASSALAMUALIKUM. I am undertaking a research that is going to evaluate the possibility of using indigenous/alternative transport (like animal transports) for supporting military operations when the conventional military transports fails. Now, what should be my population: the whole military, all the indigenous transports available in the land? the drivers of the indigenous transports? I am confused. pl suggest.
    Best Regards.

    Didier Dierckx - March, 2014 reply

    Dear Md Ashrf Al Mamun,

    Wa alaykumu s-salam.

    Thank you for your interesting questions.

    It actually depends on who you would like to survey. If you would like to survey the military on the feasibility of indigenous/alternative transports, then I would argue that your population is the military (or the decision-makers within the military). If you want to survey on e.g. how many drivers are willing to do military transports, I would argue that the drivers of indigenous transports is your population.

    In short, you should first and foremost decide which group you would like to survey.

    I hope this helps.

    Kind regards,
    Didier

    Md Ashrf Al Mamun - March, 2014

    Thank You.

  • Abdulaziz - March, 2014 reply

    Dear Didier,

    Thanks for your continuous support you provide to online people. A team that I belong to plans to conduct a household survey in a town with approximately 5,000 households. The town is also divided into six sections. We managed to determine the sample size. However, we are wondering how we can assign (distribute) the sample size to the six sections. I mean how systematic can we allocate the sample proportionally to the six sections. Is there a way we can do this?

    Many thanks

    Didier Dierckx - March, 2014 reply

    Dear Abdulaziz,

    Thank you for your interesting questions.

    I think you should treat each section as a separate population. As a consequence, you should calculate 6 different samples, based on the population of each section. As such, you will have by definition a correct allocation of the samples.

    I hope this helps!

    Kind regards,
    Didier

  • Adebayo Tajudeen Sanni - March, 2014 reply

    Hi Didder,
    I am carry out a survey study on population of 94,393. I want you to assist me in sample selection. My topics is on”Gender relations and household income utilization among elite working couples” What kind of study can you recommend for this kind topics? Which kind of research design can you recommend?
    Thank you
    Adebayo

    Didier Dierckx - March, 2014 reply

    Hi Adebayo,

    Thanks for your questions.

    As I am not familiar with this field of study, I would suggest you try to find some state-of-the-art scientific articles on this topic. Google scholar is a great place to start!

    Good luck!

    Didier

  • Boby - March, 2014 reply

    Hi Didier,
    I have a 387 respondents reported data from a total population of 4730. I just want to know how you would generalize your finding and say that the total population so and so.
    Thanks

    Didier Dierckx - March, 2014 reply

    Hi Boby,

    Thank you for your question.

    Well, you should first start by determining your actual margin of error. You can do this by using our calculator: https://www.checkmarket.com/market-research-resources/sample-size-calculator/

    Assuming a confidence level of 95%, your margin of error is 4,77%. If you are discussing your results you should clearly indicate (for instance in subscript to a table) that you have a margin of error of 4,77% and confidence level of 95%. Your readers will then be able to interpret your results. They will know that the actual population results might differ +/- 4,77% from your survey results.

    Hope this helps!

    Didier

  • shibulal - March, 2014 reply

    hi didian
    i am came up with a research situation to deal with a popuation size of 2,00,00,000 of households in a particular state . the research deals with a social problem..pls suggest me simple sampling procudure or size to get a 95% con. levl. and 5% margin of error………
    shibulal,

    Didier Dierckx - March, 2014 reply

    Hi Shibulal,

    Thanks for your questions.

    Regarding the sample size, you calculate it with our calculator: https://www.checkmarket.com/market-research-resources/sample-size-calculator/
    However, keep in mind that if you are using sub-populations, that you should be calculating your sample size based on these sub-populations.

    Regarding the sampling procedure, I would advice you to take a random sample from those 20.000.000 households.

    Hope this helps!

    Didier

  • Jerome - February, 2014 reply

    Hi There

    Please can you help me. I have a target population of 2500 people, how many people it is advisable to sample.

    Regards

    Didier Dierckx - March, 2014 reply

    Hi Jerome,

    Well, this depends on how accurate you want your sample data to be.

    If you want your survey data to reflect very closely the entire population, then you should choose a high confidence level (99%) and a low margin of error (1%). In that case you would need to sample 2.173 respondents.

    However, at the other end of the spectrum, if you do not mind that your survey results do not entirely reflect the entire population, you can use a ‘low’ confidence level (95%) and a high margin of error (5%). In that case you would need ‘only’ 334 respondents.

    Regards,
    Didier

  • Mani - February, 2014 reply

    Hi Didier,

    I’ve been tasked with trying to extrapolate income bands from a survey of around 16,000 people. The survey itself went to a random selection of people and no sampling methodology was used.

    Now I’m very new to this, so I was wondering if you could tell me the best method of taking these results and applying it to a base of around 4 million people.

    I initially looked into using multinomial logistic regression, but the accuracy was only around 30%. Is this the best method of trying to achieve my goal or should I be doing something else?

    Any help would be much appreciated.

    Mani

    Didier Dierckx - March, 2014 reply

    Hi Mani,

    Thanks for this interesting question.

    I am not entirely familiar with this kind of analysis, but I would first try to create several respondent profiles out of the survey using structural equation modelling and second using multinomial logistic regression to link the 4 million people to a corresponding profile.

    Perhaps someone can chip in with a suggestion?

  • Aliyu A. - February, 2014 reply

    So many difficulties have been resolved for me.
    Thank you very much.

    Didier Dierckx - February, 2014 reply

    Dear Aliyu,

    Thanks for your feedback. Glad to hear I was able to help you!

  • Mutassim - February, 2014 reply

    Dear Didier,
    Good day, I hope you have the time to help me in below inquiry…
    I am about to conduct a survey for my dissertation. The topic is about passengers’ satisfaction on international airlines.
    The number of my population will be hundreds of millions if I consider any international passenger in my population.
    I am going to do use a particular international airline and do the research on its passengers from JFK airport to AMM. In this case, my population will be all international passengers worldwide or only the number of international passengers for that particular airline on that particular rout.
    Again, my topic will not be on this particular airline it will be about international passengers in general, but I am going to use it passenger for my study.
    I would appreciate your advice.

    Didier Dierckx - February, 2014 reply

    Dear Mutassim,

    Thank you for your interesting questions. Customer satisfaction is a very interesting topic which we encounter very often at CheckMarket.

    However, I feel it will be very difficult for you to generalize your findings to all international passengers. After all, you will be conducting a survey on one route at one airline. It will be rather impossible to argue that this is representative for all international passengers.

    I would advice you to limit your generalization scope or conduct much more surveys on much more routes at much more airlines.

    Good luck!

    Didier

    Mutassim - February, 2014

    Thank you for your reply,
    Does this mean that I have to change also the title of my topic “The factors that influence passengers’ decision in selecting their international airline”. to be for that particular airline in that particular route? Or can I keep the title as is and customize my population within the text.
    Appreciate your advice.
    Mutassim

    Didier Dierckx - February, 2014

    Dear Mutassim,

    As the title of your dissertation is a reflection of its content, I would indeed change the title of your topic. Otherwise it would be rather misleading.

    Regards,
    Didier

    mutassim - March, 2014

    Dear Didier,
    many thanks.

  • Irene - February, 2014 reply

    Hi Didier,

    I have a population of 883, according to your website calculator the required sample size with 95% confidence level and 5% error margin is 268. The population are people within a certain company intranet I read the estimated respond rate for this is 30%. Concluding the number of invites is 894, which is bigger than my population… How is this possible and what should I do?

    Didier Dierckx - February, 2014 reply

    Hi Irene,

    Thank you for your question.

    The “number to invite” is only meant to give you an indication of the number of people that you need to invite to reach the calculated sample size.

    I would advice you to invite your entire population and try to increase your response rate. There are various ways to do so: send reminders, offer an incentive, etc.

    Good luck!

  • Maf - February, 2014 reply

    Hi Didier,

    below is the table for my samples.

    Sector Frequency Sector distribution Sample allocation
    AGRICULTURE, FORESTRY AND FISHING 3053 3% 21
    COMMUNITY,PERSONAL AND HOUSEHOLD 4544 5% 32
    CONSTRUCTION 5578 6% 39
    ELECTRICITY, GAS AND WATER 403 0% 3
    FINANCING, INSURANCE, REAL ESTATE 46826 49% 327
    MANUFACTURING 12814 14% 90
    MINING AND QUARRYING 153 0% 1
    TRANSPORT, STORAGE AND COMMUNICATION 3316 3% 23
    WHOLESALE & RETAIL TRADE 18188 19% 127
    Grand Total 94875 100% 663
    ……………………………….
    ELECTRICITY, GAS AND WATER have the frequency of 402 and the MINING AND QUARRYING have 153 and the sample is 1.

    How do i treat them?

  • Maf - February, 2014 reply

    Hi Didier,

    how can one treat the smaller populations when sampling?

    regards.

    Maf.

    Didier Dierckx - February, 2014 reply

    Hi Maf,

    Well, the formula for calculating your sample size is a bit different from the formula that is usually used.

    Please find the formula below:

    SS = [N * z² * (p*q)] / [(ME² * (N – 1)) + (z² * (p*q))]

    Alternatively, our calculator keeps the size of your population into account when calculating your sample size. So, you can calculate your sample size via our calculator: https://www.checkmarket.com/market-research-resources/sample-size-calculator/

    Regards,
    Didier

  • Jackie - February, 2014 reply

    Bonjour Didier,
    I have a survey for an MPH that involves 3 referral hospitals, each with a different population of staff; say 5,000, 2,000 and 500 respectively.
    Which sampling technique should I use to get a representative sample?

    How should I calculate a representative sample size for (a) each of them; and (b) all 3 combined?

    Didier Dierckx - February, 2014 reply

    Hi Jackie,

    I would treat each hospital as a separate population and, as such, calculate a sample size for each of them. Your overall sample size will then be the sum those three sample size.

    You can use our sample size calculator via this link: https://www.checkmarket.com/market-research-resources/sample-size-calculator/

    Good luck!

  • HEBA - February, 2014 reply

    dear Didier ,
    i got it thanks a lot , wish you the best always and i know those were a lot of questions to follow so million thanks for your kind help

    kind regards
    heba

  • naeem hijazi - February, 2014 reply

    an excellent way to find the sample size of a research and number of respondant?

  • Dominic Josephmary - February, 2014 reply

    1. Should one always restrict himself to to the confidence interval of 5%, 2% or 1% as the case may be and confidence level of 95% or 99% if the population size is known to calculate the sample size?
    2. How can this be calculated mannually without using the sample calculator. please simplify it.

    Didier Dierckx - February, 2014 reply

    Dear Dominic,

    Thank you for your questions.

    There are no rules carved in stone regarding the size of the confidence interval/level. The percentages you mention are however conventional. Furthermore, if you decide to increase your confidence interval and your confidence level, the data of your sample will be less accurate.

    Regarding your second question. Please find below the standard formula for calculating the sample size. If you do not know your proportion p, it is quite common to take 0,5 as a value for p.

    SS = (Z-score)² * p*(1-p) / (margin of error)²

    In a second step – if you know the size of the population – you can adjust your sample size via the formula below:

    SSa = (SS) / 1 + [(SS – 1) / population]

    Good luck with your research!

    Kind regards,
    Didier

  • Barry - January, 2014 reply

    I am involved in analyzing and presenting the results of a survey as part of a position paper that an organization is putting together and the results will accompany the report. The survey targeted current students receiving financial assistance from a federally funded but locally administered post-secondary assistance program. Since it is locally administered, we did not have access to contact information for each student. The survey was sent to the local organizations, posted on numerous websites where the student would have access, sent to known colleges and universities where the students were attending, and to counsellors and program managers. Participants were also encouraged to forward the link to the survey to their contacts.

    The total population receiving assistance is 4,400, with 303 surveys completed (some data cleansing was necessary). Obviously there is no response rate, since no attempt was made to target students – rather it was made available but not sent directly to students. The sample size calculator determined that I would need 353 respondents – but does that even apply in this case?

    How do i present the results? I am able to discuss the demographics of the sample from the questions that were asked (age, geographic residence, program type, gender, etc.) – but with so many limitations, am I able to generalize any of the findings to the population?

    In this case, what type of study would this fall under? Or is it? Is there any vaildity or reliability? Am I able to make any generalizations?

    I’d really appreciate your thoughts and expertise.

    Didier Dierckx - February, 2014 reply

    Hi Barry,

    Thanks for your questions.

    I would say that the sample size certainly applies in your case. Calculating and collecting data from the appropriate sample size is important for generalizing your findings. After all, when you collect the right sample size, you know what your margin of error is, so you know how to interpret you results.

    In your case, i.e. 303 respondents out of a population of 4.400, you have a margin of error of 5,4% (assuming a confidence level of 95%). In other words, in 95% of the time the answer of the population differs +/- 5,4% from your survey results. You should clearly report these numbers in your report.

    However, it is important to stress – as I already pointed out in earlier comments – that the sheer size of your sample is in no way a guarantee for being representative of your population. To find out whether your sample is representative, you should compare the distribution of your sample on e.g. gender, age, geographic residence, etc. with the distribution in the population.

    I hope this helps.

    Kind regards,
    Didier

    Barry - February, 2014

    I appreciate the quick response. I will be able to make some camparisons to the total population – so that will really be helpful. Again – thank you for your reply.

  • Ana - January, 2014 reply

    Hello Didier,

    Thank you so much for all the help in this forum.
    I would like to ask if you can give me a hand with a statistical problem I am facing.
    I need to calculate my ‘n’ or study population size, however, not in a survey type of thing.
    I am doing my Masters in Toxicology and I’ll explain you my project:
    From all the patients (12.000/year) I will analyse them all for cocaine and then x will be positive and y will be negative.
    For all of them (both x and y) I will analyse them for the presence of polymorphisms and then from x population z% will be positive and from the y population w% will also be positive.
    So, then I need to calculate if this difference between z and w is statistically significant or not. Basically, I want to reach a conclusion if whether having a polymorphism and cocaine use at the same time affects at all or not. So, the optimum would be that z% would be much higher than w%.

    In any case I am just in the start of my project so for now I just need to define my ‘n’, however, I do not know how to do it.

    Can you help me at all?

    Many thanks,

    Ana.

    Didier Dierckx - January, 2014 reply

    Hi Ana,

    Thank you for your questions.

    First of all the question I am familiar with, i.e. regarding the ‘N’ of your research project. As far as I understand your project, your population is 12.000. In other words, the number of patients per year. If you are studying various years, it is obviously the number of patients of those years combined.

    You can then choose to either analyse your whole population or a sample of it. If you want to calculate an appropriate sample size, you can use our online calculator (https://www.checkmarket.com/market-research-resources/sample-size-calculator/).

    Regarding the second question, I am obviously not familiar with toxicology and polymorphisms. However, you can easily compare the proportions of each category using e.g. Custom Tables in SPSS. You can then compare the proportion of having a Polymorphism or not for using cocaine or not.

    Good luck with your research!

    Kind regards,
    Didier

  • Ahmad - January, 2014 reply

    Hi
    I am planning to work on audiences attitude when they watch TV, their motivations and their orientations. Population size is 100000 but I can not collect samples from more than 150 bcz it will require so much time and resources. Will it be suffice If I have stratified my sample size also. Thanks

    Ahmad

    Didier Dierckx - January, 2014 reply

    Hi Ahmad,

    Thank you for your questions.

    Stratification will certainly help to improve the representativeness of your sample. Nevertheless, your margin of error will still be 8%, which is quite high.

    So, I would advice to collect more data and/or clearly report the margin of error.

    Kind regards,
    Didier

    Ahmad - January, 2014

    Thanks very much,

  • Paulo Jefferson Dela Cruz - January, 2014 reply

    Hi,

    I am planning to do a survey with a total population of 132 workers. How people should i survey to get 95% confidence. Thank you very much!

    Paulo

    Didier Dierckx - January, 2014 reply

    Hi Paulo,

    Thank you for your question.

    This depends on the margin of error you are willing to accept. With a margin of error of 1%, you would have to survey 131 respondents. Alternatively, you would ‘only’ have to survey 99 respondents with a margin of error of 5%.

    So, basically, it depends on how accurate you want your data to be. Nevertheless, you should always mention the margin of error and the confidence levels. As such your readers can interpret your results correctly.

    You can use our sample size calculator for your calculations: (https://www.checkmarket.com/market-research-resources/sample-size-calculator/).

    Good luck!

    Kind regards,
    Didier

  • richard - January, 2014 reply

    with a population of 3321, what should be my sample size

    Didier Dierckx - January, 2014 reply

    Richard,

    Thanks for your question.

    That depends entirely on how accurate you want your survey results to be. Below I give you the two extremes:

    Option 1: Very loose: Confidence level 95% and Margin of error 5%: Sample size = 345
    Option 2: Very strict/accurate: Confidence level 99% and Margin of error 1%: Sample size = 2.768

    You can use our calculator (https://www.checkmarket.com/market-research-resources/sample-size-calculator/) to test various scenario’s.

    Good luck!

    Didier

  • Radwa - January, 2014 reply

    I have 179 household and I want to make a satisfaction survey. what is the percentage i should talk to make the results representable.

    Thanks
    Radwa

    Didier Dierckx - January, 2014 reply

    Dear Radwa,

    Thank you for your question.

    You can calculate your sample size via our calculator (http://www.checkmarket.com/market-research-resources/sample-size-calculator/).

    I would suggest to survey at least 120 households.

    Good luck!

    Kind regards,
    Didier

  • richard - January, 2014 reply

    I have a population of 23,245 from the tables am required to use a sample size of 370, but I want to use 200 as my sample, can you justify a sample of 200 statistically.

    Didier Dierckx - January, 2014 reply

    Dear Richard,

    The sample size you use is closely linked with how accurate you want your survey data to be.

    If you use a sample size of 200, you have a margin of error of 6,90%, assuming a 95% confidence level. As I explained in the article, the margin of error is the positive/negative deviation you allow on your survey results for the sample.

    The margin of error is quite large if you use a sample of 200 respondents, which basically makes your results less robust. So you should definitely report the margin of error so that readers can interpret your results.

    Kind regards,
    Didier

  • Alemtsehay - January, 2014 reply

    Hi, I need to do research the topic is prevalence of adult under nutrion and associated factors in Hospitalized patients . the study period is 2 months
    What is my sample size caluclation formula ?
    I use all entire population in the study period ,still I am confused please Help me
    Thank you
    Africa, Ethiopia

    Didier Dierckx - January, 2014 reply

    Dear Alemtsehay,

    Thank you for your question.

    The formula for calculating the sample size when you do not know the size of the population, is the one below:

    SS = (Z-score)² * p*(1-p) / (margin of error)²

    In a second step, you can adjust your sample size if you do know the population by using the formula below:

    SSa = (SS) / 1 + [(SS – 1) / population]

    Alternatively, you can calculate the sample size via our sample size calculator: http://www.checkmarket.com/market-research-resources/sample-size-calculator

    I hope this helps.

    Kind regards,
    Didier

  • Gia - January, 2014 reply

    Hello Didier,

    Thank you for your comment. 494138 is the approximate number of SA population in 4 cities. You have mentioned that to find out whether my sample is representative, I should compare the distribution of my sample on e.g. gender, education, age, etc. with the distribution in the population…. At this initial stage I am in no position to do that. Please help.

    Didier Dierckx - January, 2014 reply

    Hello Gia,

    I suppose there are in total 494.138 South-Asian immigrants living in the 4 cities?

    In order to calculate your sample size – you can use our sample size calculator: http://www.checkmarket.com/market-research-resources/sample-size-calculator – you should calculate 4 different sample sizes. One for each city, based on the number of SA immigrants living in that particular city. As such, you will definitely have an appropriate sample size.

    If you draw a random sample, you can check afterwards whether it is representative for your population. If it is not, you can for instance weigh your results on a couple of socio-demographic variables.

    Kind regards,
    Didier

  • Gia - January, 2014 reply

    Hello,

    I am trying to study diet behaviors of South Asian population in Canada after migration. If I select 350 participants for interview and diet survey (50 from one city and 100 each from 3 other cities) will it be reasonable?
    Will they be representative of the South Asian immigrants in Canada? Please comment. Thank you!

    Didier Dierckx - January, 2014 reply

    Hello Gia,

    Thank you for your questions.

    As I explained in an earlier comment, sample size is in no way a guarantee for having a representative sample of your population. After all, you could – for example – survey 100.000 citizens in each city but if your respondents primarily contain the ’35-54 higher-educated’ citizens, you can hardly speak of a representative sample. To find out whether your sample is representative, you should compare the distribution of your sample on e.g. gender, education, age, etc. with the distribution in the population.

    Could you please provide me with more details regarding the size of the population of South Asian immigrants in the 4 cities?

    Kind regards,
    Didier

  • Manuel - January, 2014 reply

    Hello Didier,

    I am surprised how many comments are in this post!

    Would you like to give some help with the sample of a survey? I am trying to deliver an online survey for users of a website. The amount of total users is very large but we have the mails from everybody. However, that number counts people that just visited once and people that are active.

    Which kind of sample do you recommend to use? We have the possibility to deliver the survey to the whole population. Would that be worth it? Would you recommend any specific sample size?

    Many thanks for your help.

    Kind Regards,
    Manuel

    Didier Dierckx - January, 2014 reply

    Hi Manuel,

    Thanks for your questions.

    First of all, I would certainly survey both profiles, i.e. the people that just visited once and the people that are active. Via the CheckMarket survey tool, you can easily make the distinction between different profiles in your survey. More information via this link: http://www.checkmarket.com/kb/what-is-branching/

    You could survey your whole population, but it is probably more efficient as well as more cost-effective, to survey a sample of your population. It is important to treat each profile (one-time visitors and more active visitors) as different populations and calculate a sample size for each group. As such you can be sure that you will have a total sample that is large enough for both profiles.

    You can calculate your sample size via this calculator on our website: http://www.checkmarket.com/market-research-resources/sample-size-calculator/

    If you have any further questions or if you would like a non-committal tender for your survey project, please do not hesitate to contact me.

    Kind regards,
    Didier

    Manuel - January, 2014

    Hi Didier,

    Thank you very much for your reply and advises. The calculator has been a great tool for calculating the sample size!

    Kind regards,
    Manuel

  • Kevin Tran - January, 2014 reply

    Hello Sir,

    Currently I’m conducting a final year project for my undergraduate degree. The target number of my sampling size was 350 students. But unfortunately, It turned out that I was able to get only 183 respondents. Would it be sufficient for me to proceed the data analysis with only 183 respondents Sir? I really need a proof from a previous research which has encountered the same issue like me in order to make my empirical findings valid.

    Thank you so much Sir,

    Have an awesome day ahead,

    Didier Dierckx - January, 2014 reply

    Dear Kevin,

    Thanks for your question.

    While it obviously will have an effect on the interpretation of your research results, you can proceed with the data analysis. However, there are 2 issues you should take into account.

    1) Doing the reverse calculation (assuming that your population is approx. 10.000 and a confidence level of 95%), you have in fact a margin of error of approx. 7%. This is quite large. So, you should clearly report this number so that readers can interpret your results.

    2) As you probably will, segmenting between socio-demographic groups (e.g. gender, prior education, etc.) should be done with care. After all it is possible that your subgroups will be (too) small.

    Hope this helps! Good luck with your project!

    Kind regards,
    Didier

  • Hassan - January, 2014 reply

    Just to know are the sample size calculators and the error issues/level of confidence applicable to quantitative reseraches only?Can I apply the same in a qualitative study?

    Thanks

    Didier Dierckx - January, 2014 reply

    Dear Hassan,

    Thank you for your question.

    You cannot use the same criteria in a qualitative study. A useful reference for your qualitative sample size is:

    Sarah Elsie Baker & Rosalind Edwards (2012), “How many qualitative interviews is enough? Expert voices and early career reflections on sampling and cases in qualitative research”.

    I hope you will find this helpful.

    Kind regards,
    Didier

  • Nasir - December, 2013 reply

    Hello Dear all,
    I would like to conduct a survey on menstrual hygiene among the girls students 9-12. I dont know either number of school students at this grade nor number of school in these threes districts. How I can calculate my sample size and methodology of students selection.

    Didier Dierckx - January, 2014 reply

    Dear Nasir,

    Thank you for your questions. As I explained earlier, you do not have to know the size of your population to calculate your sample.

    If you do not know your population, you can use the formula below:

    SS = (Z-score)² * p*(1-p) / (margin of error)²

    If you do not know your proportion p, it is quite common to take 0,5 as a value for p.

    However, it is advisable to calculate your sample size on the smallest “population” you will be using in your research. In other words, on the smallest district.

    Hopefully this helps!

    Kind regards,
    Didier

  • zafar - December, 2013 reply

    im a lay man in statistic and im taking a consecutive sampling now i dont know how to calculate margin of error and confidence inter to decide about the sample size.plz i need ans by today

    Didier Dierckx - January, 2014 reply

    Dear Zafar,

    Thank you for your questions.

    As far as I am aware the idea behind consecutive sampling is to include all accessible subjects as part of the sample. As a result, this makes the sample a better representation of the population.

    Furthermore – in contrast with probability sampling – there is no distinct formula to determine the appropriate sample size. After all, the calculation of the sample is not based on known probabilities and has some subjective judgments (the parameters of the sample depend on what you define as being important).

    I hope this helps.

    Kind regards,
    Didier

  • ashish kumar singh - December, 2013 reply

    Dear Sir,
    I am doing research on “An Anthropometric Investigation and body type” in the age span of 6 to 15 years children and population size of (as yet unknown) three districts. Could you tell me how many samples are required for my research work.
    Ashish Kumar Singh, research scholar
    Dept. of physical education & sports . Pondicherry University (India)

    Didier Dierckx - January, 2014 reply

    Dear Ashish Kumar Singh,

    Thank you for your questions.

    I would calculate a sample for each district you want to consider in your research. As such, you can be confident you will have enough cases for each district.

    Kind regards,
    Didier

  • Ramesh Neupane - December, 2013 reply

    I just want to calculate the sample size from 150 populations, would u plz inform me what will be my sample size and how… i just want to know the formula to calculate a sample size from known population. Thank you

    Didier Dierckx - December, 2013 reply

    Dear Ramesh Neupane,

    Below you find the standard formula for calculating the sample size:

    SS = (Z-score)² * p*(1-p) / (margin of error)²

    In a second step, you can adjust your sample size if you do know the population by using the formula below:

    SSa = (SS) / 1 + [(SS – 1) / population]

    Good luck!

    Didier

  • Aaron Ato-Davies - December, 2013 reply

    Like for an example ; according to ….(2013) for every a population of 500 , 377 respondents can be used with 2.5 confident level

    Didier Dierckx - December, 2013 reply

    Dear Aaron,

    Thanks for your comments. Glad to hear that you found the article useful.

    I do not have a reference for the table as I constructed it myself. However, the formula on which the table is based, can be found in any statistics manual.

    Good luck!

    Didier

  • Aaron Ato-Davies - December, 2013 reply

    Sir , I found the information you gave very useful and I want to use it but I need to reference it , can I get a reference for that table of sample selection?
    Thank you.

  • heba - December, 2013 reply

    Hi ,
    thanks for the great article it was very helpful
    just want to ask if i can get a reference (published paper for example)
    to add to my references beside your website
    will be thankful for your help
    best regards :)

    Didier Dierckx - December, 2013 reply

    Dear Heba,

    Nice to hear the article was helpful for you.

    I would advice you to look on Google Scholar for such an article. There are countless articles on determining the sample size in many scientific fields, so it is advisable if you look for an article that is relevant to your field of study.

    It is possible that you will have to access Google Scholar via the library of your college/university. More often than not you need a subscription to access a scientific journals.

    Good luck!

    Didier

    HEBA - February, 2014

    dear Didier.
    thanks for your nice follow up on our questions and im looking about some references now and it seems good way, but does this formula you wrote i mean this one
    “SS = (Z-score)² * StdDev*(1-StdDev) / (margin of error)²”
    has a name to search for it specifically?
    thanks again and again
    best regards :)

    Didier Dierckx - February, 2014 reply

    Dear Heba,

    I googled on ‘google books’ the search term ‘sample size social sciences’ and got several books that should have this formula in it.

    Good luck!

    Didier

    HEBA - February, 2014

    dear Didier ,
    i got it thanks a lot , wish you the best always and i know those were a lot of questions to follow so million thanks for your kind help
    kind regards
    heba

  • Bet - November, 2013 reply

    Thanks for the post. I have a few clarification and your suggestions. I have determined a sample size of 400 from City A and 200 from city B, given my time and resources. The total population size of City A is 628,780 and City B is 474,990. Each city have got districts. City A have 10 districts and City B have got 6 districts. I selected all districts. Again each districts have got local districts. I have randomly selected local districts from each each districts and distributed the sample size proportional to the size of the population in each local districts. Finally, I have randomly selected households from each local districts. I want to get your feedback if the sample is representative of the two city. Can the result be generalized?

    Didier Dierckx - December, 2013 reply

    Dear Bet,

    Thank you for your questions.

    First of all, I must stress that the sheer size of your sample is in no way a guarantee for being representative of your population. You could – for example – survey 100.000 citizens in both cities but if your respondents primarily contain the ’35-54 higher-educated’ citizens, you can hardly speak of a representative sample. To find out whether your sample is representative, you should compare the distribution of your sample on e.g. gender, education,
    age, etc. with the distribution in the population.

    Second, regarding the division of your initial sample over various districts and local-districts, it is advisable to calculate your sample size on the smallest “population” you will be using in your research. In your specific case, these are the local districts. In short, I would urge you to increase your sample size if you would like to take into account the districts and local-districts.

    Good luck with your research!

    Kind regards,
    Didier

  • Joseph - November, 2013 reply

    i m designing a survey for prevalence of brucellosis in Lesotho. we currently do nt have cattle statistics and would like to know how to determine the sample size which would significantly represent cattle population

  • Otieno Mike - November, 2013 reply

    hi, am pleased to learn a lot. am doing an academic research using using non probability method specifically convenient sampling.my target population is 20,000 people and i would like to put them into Focus Group Discussion. Kindly assist in how to calculate a sample size.
    Thanks

    Didier Dierckx - November, 2013 reply

    Dear Mike,

    Thank you for your questions.

    Unfortunately – in contrast with probability sampling – there is no distinct formula to determine the appropriate sample size. After all, the calculation of the sample is not based on known probabilities and has some subjective judgments (the parameters of the sample depend on what you define as being important).

    Assuming that you know the distribution of some parameters in your population, my advice would be to try to get as close as possible to representativeness.

    Kind regards,
    Didier

  • celestin - November, 2013 reply

    please, I have a question on how to construct a sample. I am conducting an agricultural survey on different sites, where each site is. independent of the other. the population in one site is 4024, the other one is 3960. My. question is how I can get.a representative sample

    thanks

    Didier Dierckx - November, 2013 reply

    Dear Celestin,

    Thank you for your question.

    Via the sample size calculator (http://www.checkmarket.com/market-research-resources/sample-size-calculator/) you can calculate the appropriate sample for both sites.

    I would calculate a separate sample size for each site.

    Kind regards,
    Didier

  • Kingsley Simon - November, 2013 reply

    I am a novice in regards to calculating sample size, population size etc. Please I will need any help from you guys.

    I want to do a questionnaire targeting all mobile and internet banking users.

    How do i calculate my sample size?

    Thanks
    Kingsley

    Didier Dierckx - November, 2013 reply

    Dear Simon,

    Thank you for your questions.

    You can calculate your simple size using our sample size calculator (http://www.checkmarket.com/market-research-resources/sample-size-calculator/).

    If you do not know your population, you can use the formula below:

    SS = (Z-score)² * p*(1-p) / (margin of error)²

    If you do not know your proportion p, it is quite common to take 0,5 as a value for p.

    If we use it in an example with a population of 100.000, confidence level of 95% and margin of error 2%, then you get the following calculation:

    SS = (1,96²) * 0,5*0,5 / 0,02²
    SS = 3,8416 * 0,25 / 0,0004
    SS = 0,9604 / 0,0004
    SS = 2401

    Hopefully this helps!

    Kind regards,
    Didier

    Kingsley Simon - November, 2013

    Dear Didier,

    Thanks for your response. This helps a lot but with this formula, how do i factor in my population if for example, i am looking at a population size of 5000?

    Thanks
    Kingsley Simon

    Didier Dierckx - November, 2013 reply

    Dear Simon,

    Thanks for your follow-up question. I was not aware of the fact that you know your population size.

    If you know your population size, in a second step, you can adjust your sample size via the formula below:

    SSa = (SS) / 1 + [(SS – 1) / population]

    So first, you should calculate your sample size and then, in a second step, you can adjust your calculated sample size for the size of the population.

    Kind regards,
    Didier

  • Clarisse meboka - November, 2013 reply

    Dear sir from your article while explaining the measures that affect the accurateness of the data I really did not get it . Pleas sir can you be more explicit and simplify them for me. So I can better understand.

    Thanks. Clarisse

    Didier Dierckx - November, 2013 reply

    Dear Clarisse,

    In an ideal world, you would survey the entire population. However, for obvious reasons, this is not possible.

    So, when you survey a part of the population (your sample), you want the results of your survey to reflect the opinions of the entire population as closely as possible. Two measures affect this closeness: margin of error and the confidence level.

    Margin of error: the deviation between the opinions of your respondents and the opinion of the entire population. This deviation is formulated as a percentage. A margin of error of 5% means that – if 90% of the survey respondents like the “Fall 2013” line – in the real population between 85% and 95% will actually like the line.

    But how can we be sure that this is actually the case, i.e. between 85% and 95% of the population likes the line? For this we use the second measure: the confidence level. It tells you how often the percentage of the population actually lies between the boundaries of the margin of error. In other words, how often between 85% and 95% of the population likes the line. If you chose a 95% confidence interval and you would repeat your survey a 100 times, 95 times the percentage of the population will lie within those boundaries.

    I hope this clarifies some things.

    Kind regards
    Didier

  • Smile Nkets - November, 2013 reply

    Please Didier,
    I have been following your great assistance rendered. With a population of 1000 respondents (small and medium scale enterprise owners), how many do i sample taking say 95% confidence level? How do I calculate the right sample size. Is there any authority backing your answers?
    Thanks in anticipation for your answer

    Smile

    Didier Dierckx - November, 2013 reply

    Dear Smile,

    Thank you for your questions.

    You can use either our sample size calculator (http://www.checkmarket.com/market-research-resources/sample-size-calculator/) or calculate it manually via the formula below.

    The sample size calculator uses the same formulas.

    SS = (Z-score)² * p*(1-p) / (margin of error)²

    and in a second step you can adjust your formula because you know your population:

    Correction: ss = (ss) / 1 + [(SS – 1) / population]

    Regarding, the authority backing my answers, these are standard formulas that are widely used and can be found in every statistics manual.

    Kind regards,
    Didier

  • Zainab Mohammed - November, 2013 reply

    Also how estimate margin of error?..

    Didier Dierckx - November, 2013 reply

    Dear Mohammed,

    Thank you for your questions.

    You can calculate your sample size via this calculator: http://www.checkmarket.com/market-research-resources/sample-size-calculator/

    The z-score depends on the confidence interval you chose. It has fixed values. For a confidence interval of 95%, the z-score is 1,96. For instance of a confidence interval of 90% it is 1,645.

    The formula for the margin of error is as follows:

    ME = z * [ROOT [[p * (1-p)] / n]]

    Kind regards,
    Didier

    Pam Bogdan - December, 2014

    I have tried to replicate the margin of error calculator using the above formula, but this does not seem to work.. for example if I have a population of 100 and a response of 5 the calculator gives a MOE of 42.93%, yet the formula gives a MOE of 4.27%.

  • Zainab Mohammed - November, 2013 reply

    Dear sir
    If l have 1600000 population then how I estimate sample size for survey study and I dont understand how calculated z score?..
    With thanks

  • Zulfaa - November, 2013 reply

    if my total population is 172 for the cross-sectional survey, how many people should I included for my study population?

    Thanks

    Didier Dierckx - November, 2013 reply

    Dear Zulfaa,

    Thank you for your question.

    You would need – assuming a margin of error of 5% and a confidence level of 95% – a sample size of 120 respondents.

    You can calculate it yourself via this link: http://www.checkmarket.com/market-research-resources/sample-size-calculator/

    Kind regards,
    Didier

  • Grace - November, 2013 reply

    Miss Tina owns a game player shop and 30%of her customers are unemployed,She wants to conduct a small study to identify factors contributing towards purchasing behavior of her unemployed customers.
    1. identify the population and measure number of sample for this study.
    2. identify dependent and independent variables to be used in this study.

    Can sir help me these questions??thank you

    Didier Dierckx - November, 2013 reply

    Dear Grace,

    The unemployed customers of Miss Tina’s shop is your population. In order to determine its size, you would need an indication of the total number of customers per day/month/year/…

    Having said that, as I explained in previous comments you do not necessarily need the population size to calculate the sample size. The basic formula is always the same (please find it below) and depends on the confidence level and margin of error you chose:

    SS = (Z-score)² * p*(1-p) / (margin of error)²

    Regarding your second question, I would advise you to look for similar studies in scientific journals. Those will give you an idea of which variables to use. On top of my head, I would guess as a dependent variable: a) the amount of money spent or b) whether they buy something or not. As independent variables, I would definitely include gender, age, education, etc.

    Good luck!

    Kind regards,
    Didier

  • liew cassie - November, 2013 reply

    To find out the employee contribution to SOCSO towards their productivity at workplace. Study conducted for 1 year period in 30 selected companies.
    1. identify the population and measure number of sample for this study.
    2. identify dependent and independent variables to be used in this study.

    Can you kindly help me on this? thank you.

    Didier Dierckx - November, 2013 reply

    Dear Liew Cassie,

    Could you please provide me with additional information regarding your questions? Based on this limited information, it is rather difficult for me to answer your questions.

    Kind regards,
    Didier

  • bounyadeth - October, 2013 reply

    is the calculcaiton the same between sample size for populations and for villages. if it is different, how the sample size for villages be calculated?

  • neda - October, 2013 reply

    how we decide for a sample size when we dont have the population size excatly? for example, we randomly selected 5 public university from four zone of malaysia. but we dont know about the population of undergraduate students .how we can decide for sample size and justify it?

    Didier Dierckx - October, 2013 reply

    Dear Neda,

    As I explained in an earlier comment, you do not need to know the exact population in order to be able to calculate the sample size. The basic formula is always the same (please find it below) and depends on the confidence level and margin of error you chose:

    SS = (Z-score)² * p*(1-p) / (margin of error)²

    Only if you know your population size, then you will correct the initial formula for the population by using the formula below:

    SSadjusted = (SS) / 1 + [(SS – 1) / population]

    Good luck with your research.

    Kind regards,

    Didier

  • LT - October, 2013 reply

    Dear Sir

    We are planning a Gambling Prevalence Survey of Provincial Government Employees in our Province and would like to receive your suggestions regarding an appropriate sample and an approach that you will suggest.

    Kindest regards and hope to hear from you.

    LT

    Didier Dierckx - October, 2013 reply

    Dear LT,

    Thank you for your questions.

    Could you please provide me with some additional information on the population of the Provincial Government Employees?

    Alternatively, you can use our ease sample size calculator via the following link: http://www.checkmarket.com/market-research-resources/sample-size-calculator/

    Kind regards,

    Didier

  • Glory - September, 2013 reply

    How large would your population have to be for a sample to be appropriate (i.e., rather than measuring the whole population)? I need ideas on how to do this homework.

    Didier Dierckx - October, 2013 reply

    Dear Glory,

    Thank you for your question.

    As I have explained in the article and in earlier comments, it is primarily about how accurate you want your survey data to be.

    If you want your survey data to reflect very closely the entire population, then you should choose a high confidence level (99%) and a low margin of error (1%). However, if you do not mind that your survey results do not entirely reflect the entire population, you can use a ‘low’ confidence level (95%) and a high margin of error (5%).

    As long as you report which confidence level and margin of error you have chosen, people will be able to interpret your findings.

    Kind regards,

    Didier

  • Rachel - September, 2013 reply

    Dear Sir,

    I am doing research for my PhD which includes surveying tourists who are visiting specific sites. One of these sites does not maintain visitor statistics, and since I have been here, there have been very few visitors in general. I have managed, over 2 weeks, to collect only about 100. How do I know if this is significant given that I don’t know my population size though I have observed it is quite low? Thank you very much for your help.

    Didier Dierckx - October, 2013 reply

    Dear Rachel,

    Thank you for your question.

    As I already mentioned above, you do not really have to know your population in order to calculate your sample size. As you will see, in the formula below, there is no mention of the population:

    SS = (Z-score)² * p*(1-p) / (margin of error)²

    However, I do understand your worries. Isn’t there a way you can reasonably estimate your population? For instance through extrapolation of the amount of visitors you have seen during the 2 weeks you were there?

    Kind regards,

    Didier

  • Nazish Azhar - September, 2013 reply

    Hello sir! can u kindly help me in my research. I am working on my thesis including a research questionnaire for construction industry in developing countries.I have taken Pakistan and Saudi Arabia as case studies because i have access to these two countries for data collection.My target respondents are all the stakeholders in construction industry in both countries. the total number of research population and the response rate is unknown for both countries. So how can i calculate my research population for both countries and also the required response rate. Kindly guide.

    Didier Dierckx - September, 2013 reply

    Dear Nazish Azhar,

    Thank you for your questions.

    The great thing about calculating the sample size is that you do not necessarily need to know your population size. As you see in the formula below, there is no mention of the population size:

    SS = (Z-score)² * p*(1-p) / (margin of error)²

    I would advise you to expect a rather low response rate. In market research a response rate of 20% is often considered as a good response rate, so I would recommend a response rate of 5 to 10% for your research.

    Kind regards,
    Didier

  • muhammad jabbar - September, 2013 reply

    I am working on a comparative study of universities according to different attributes of two different areas, say A and B.
    A having 10 universities.
    B having 4 universities.
    i have selected 5 universities from A and 2 universities from B, we may consider that total students of all 7 selected universities are greater than 1,00,000.

    I would like to compare different attributes according to natural sciences group, bio sciences group, social sciences group.

    1. Now my question is that how much sample size i require over all?
    2. how much from each university?
    and 3. how much from each area of study ( natural sciences group, bio sciences group, social sciences group) ??

    waiting for kind response also kindly try to give some suitable reference.

    Didier Dierckx - September, 2013 reply

    Dear Muhammad Jabbar,

    Thank you for your questions.

    Assuming that you have only 1 population (i.e. the 100.000 students of 7 universities), you need roughly 2.345 respondents (margin of error 2% and confidence level 95%). You can use the table in the article above or our sample size calculator to calculate this.

    However, in your case, you have many (sub)populations. Each university and each group should be considered as a (sub)population. From each university and each group you need a sample that is large enough to draw meaningful conclusions.

    So, if you know the population size of each university and each group within the universities, you can calculate for each university and each group the sample size you need. If you add all these subsamples, you will get the total sample you should be aiming for.

    Good luck!

    Kind regards,
    Didier

    zain - October, 2016

    sir my name is zain
    i am m.phil student my population is 2000 and i have took 151 sample size so i could i justify that

    Maarten Marijnissen - October, 2016

    Hi Zain,

    You can calculate the margin of error to justify that. This number indicates how much the opinions of your sample are likely to deviate from the population.

    You can use our online calculator to do that. At the bottom of the page you will find a margin of error calculator. Fill in your population and sample size with a confidence level of choice (often 95%) and your margin of error will be calculated automatically.

  • appgratuit.es - August, 2013 reply

    I like the helpful information you provide in your articles.

    I will bookmark your blog and take a look at once more right here regularly.
    I am moderately certain I’ll be informed lots of new stuff proper right here! Best of luck for the next!

    Didier Dierckx - September, 2013 reply

    Thank you very much!

    Kind regards,

    Didier

  • Franklin A. Manzano - July, 2013 reply

    hello sir, can you please help me on my research, I just want to ask sir if how many respondents do I need to survey and need to interview for my research with a 6,123 populations. Can I just get the 10% or 20% of the population to be my respondents? thank you sir…. I really need it bad sir for may master’s, because im starting to give up. I only have 1 month to finished it. thank very much…

    Didier Dierckx - July, 2013 reply

    Dear Franklin,

    Thank you for your question.

    As I have explained in the article and in earlier comments, it is primarily about how accurate you want your survey data to be.

    If you want your survey data to reflect very closely the entire population, then you should choose a high confidence level (99%) and a low margin of error (1%). Using the formulas, this would mean that you need at approximately 4.215 respondents.

    However, if you do not mind that your survey results do not entirely reflect the entire population, you can use a ‘low’ confidence level (95%) and a high margin of error (5%). Using the table from the article, this would mean that you need at least approximately 361 respondents.

    Both approaches are defensible, as long as you clearly report which approach you have chosen.

    Kind regards,

    Didier

  • Auwal Zaria - July, 2013 reply

    I want a lirature that I can site in a research which support 10% as sample in a research work.

    Didier Dierckx - July, 2013 reply

    Dear Auwal Zaria,

    Thank you for your questions.

    Could you please elaborate on what you are looking for exactly?

    Kind regards

    Didier

  • Ludovic - July, 2013 reply

    i forgot to mention.i wish to know the formulae.indeed in my school they will not be pleased if i just pick a value from an online statistical table.they like to see how you came to your answer.thanks

    Didier Dierckx - July, 2013 reply

    Dear Ludovic,

    Thank you for your questions.

    The formulas are quite simple and can be found in any statistical manual. These are also the formulas we use. Please find them below:

    First you calculate the sample size (SS). If you do not know your stand. dev., it is quite common to take 0,5 as a value for the stand. dev.

    SS = (Z-score)² * p*(1-p) / (margin of error)²

    Then you correct it if you know your population:

    SSadjusted = (SS) / 1 + [(SS – 1) / population]

    If we use it in an example with a population of 100.000, confidence level of 95% and margin of error 2%, then you get the following calculation:

    SS = (1,96²) * 0,5*0,5 / 0,02²
    SS = 3,8416 * 0,25 / 0,0004
    SS = 0,9604 / 0,0004
    SS = 2401

    SSadj. = 2401 / 1 + [(2401-1) / 100.000]
    SSadj. = 2401 / 1 + [2400 / 100.000]
    SSadj. = 2401 / 1 + [0,024]
    SSadj. = 2401 / 1,024
    SSadj. = 2344

    This the same number you will find via our calculator: http://www.checkmarket.com/market-research-resources/sample-size-calculator/

    Kind regards,

    Didier

    V.S.Ramalingam - May, 2015

    can you suggest a formula to sampple size for the mean of a known population

    V.S.Ramalingam - May, 2015

    can you suggest a formula to sample size for the mean of a known population

    Gert Van Dessel - May, 2015

    Thank you for your questions.

    The formula for calculating sample size is quite simple and can be found in any statistical manual.

    First you calculate the sample size (SS). If you do not know your proportion p, it is quite common to take 0,5 as a value for p.
    SS = (Z-score)² * p*(1-p) / (margin of error)²
    Then you correct it if you know your population:
    SSadjusted = (SS) / 1 + [(SS – 1) / population]

    If we use it in an example with a population of 100.000, confidence level of 95% and margin of error 2%, then you get the following calculation:
    SS = (1,96²) * 0,5*0,5 / 0,02²
    SS = 3,8416 * 0,25 / 0,0004
    SS = 0,9604 / 0,0004
    SS = 2401
    SSadj. = 2401 / 1 + [(2401-1) / 100.000]
    SSadj. = 2401 / 1 + [2400 / 100.000]
    SSadj. = 2401 / 1 + [0,024]
    SSadj. = 2401 / 1,024
    SSadj. = 2344

    This the same number you will find via our calculator

  • Ludovic - July, 2013 reply

    hello. I am working on an academic report for submision by the end of the year. i have to work in an area with a population of 750 450 people.i wish to have a sample which is very representative of the population. indeed the formulae i find over internet exploit some parameters that i dont know hw to assess it in my case. i wish to know if there exist a statistical way to have the sample size from a known population such as in my case. i wish to work with a confidential interval of 95%. thanks in advance.

  • Enoch - July, 2013 reply

    Hi! I conducted survey in three village which have total household population of 2008. Village A 1022 population, village B 603, and Village C 383. I used 15% for each population to be my sample size. In total 300 households were actually surveyed from three villages. How can i do for my sample to present 5% marginal error and 95?% of confidence interval? as it sounds like use of % is not much recommended.
    Thanks

    Didier Dierckx - July, 2013 reply

    Dear Enoch,

    Thank you for your question.

    It appears to me that – if you consider your three villages as 1 population – your sample of 300 surveyed households is large enough to have 95% certainty of your 5% margin of error. You can calculate it yourself via this link: http://www.checkmarket.com/market-research-resources/sample-size-calculator/

    However, if you consider them as three separate populations, then I would need to know the distribution of the surveyed households for each village.

    Kind regards,

    Didier

    Josh t - July, 2014

    How can I determine a power level for a survey size of 150 with an 8% margin of error and 95% confidence level.

    Josh Ty - July, 2014

    How can I determine a power level for a sample size of 150 with an 8% margin of error and 95% confidence level. Population size is 3000.

  • Carol - June, 2013 reply

    I need to do a study in a village. The population is 475. How many large should my survey be?

  • Tasi'u Rilwanu Yalwa - June, 2013 reply

    Thanks, but if total population is 596411 what percentage should I take and what is the sample?

    Didier Dierckx - June, 2013 reply

    Dear Tasi’u,

    Thank you for you question.

    Please see my comment above as your questions is highly similar to the question of Victor Gloria.

    Kind regards,

    Didier

  • victor gloria - June, 2013 reply

    i can get the explanation a bit,,,am working on population of about 429300,,how ll i calculate my margin error and the confident interval please

    Didier Dierckx - June, 2013 reply

    Dear Gloria,

    Thanks for your question.

    As I have explained in the article, it is primarily about how accurate you want your survey data to be.

    If you want your survey data to reflect very closely the entire population, then you should choose a high confidence level (99%) and a low margin of error (1%). Using the table from the article, this would mean that you need at least 16.000 respondents.

    However, at the other end of the spectrum, if you do not mind that your survey results do not entirely reflect the entire population, you can use a ‘low’ confidence level (95%) and a high margin of error (5%). Using the table from the article, this would mean that you need at least 400 respondents.

    Both approaches are defensible, as long as you clearly report which approach you have chosen.

    Kind regards,
    Didier

  • Ed - June, 2013 reply

    Do the numbers not increase after 1,000,000? Why do you not need to survey 400x the number listed to approximate a 400,000,000 population?

    Didier Dierckx - June, 2013 reply

    Dear Ed,

    Thank you for your question.

    In brief, no they do not. Rather smart mathematicians have proved that the size of the population does not matter that much. As a result, a sample of, say, 1000 respondents is equally sufficient for a population of 10.000.000 citizens as for a population of 100.000 citizens. For more mathematical information, I can recommend the book “Statistics” by Freedman, Pisani and Purves (2007).

    However, you should take the number of sub-groups into consideration when determining your sample size. Each sub-group should be large enough for valid estimation of parameters. As a result it is advisable to treat each sub-group as a population and determine your sample size accordingly.

    Regards,
    Didier

  • Rutger - May, 2013 reply

    How do you calculate for instance if you’ve got a population of 3417 people and you just asked to 80 of them questions… how do you calculate then how accurate you are? if i wanted to be accurate for 95% i needed to interview for about 250 people… now how do you calculate how accurate my enquête is when i just interviewed 80 people…

    how do you calculate that?

    kind regards

    Didier Dierckx - May, 2013 reply

    Dear Rutger,
    Thanks for your question.
    First of all, you probably should have aimed for approximately 350 respondents instead of 250 respondents (margin of error 5% and confidence level 95%).
    On our website you find a calculator that allows you to calculate your ‘actual’ margin of error after completion of the survey (http://www.checkmarket.com/market-research-resources/sample-size-calculator/).
    If I insert your values, the actual margin of error is 10,83%. This is quite significant. With a confidence level of 95%, this means that you can be 95% sure that if you asked the same questions to the entire population between -10% and +10% would pick the same answer.
    In short, I would be really cautious when formulating your conclusions.
    Kind regards,
    Didier

  • Reneg - March, 2013 reply

    Thank you for the good explanation, that clears some questions up. But what if let’s say your population contains 800 existing customers (for example 800 companies in a B2B satisfaction survey). And you send an invitation trough email to fill in the survey online. (so the whole population is asked to participate).
    When you recieve 110 filled in surveys, the respons is 14%. How can you prove that the results represent the opinion of the total population?

    You would realy help me out!

    Kind regards.

    Didier Dierckx - March, 2013 reply

    Thank you for your questions.

    If 1) the survey sample is large enough and 2) it is representative on some characteristics for the population, you have already ‘proved’ that the results represent the opinion of the population.

    If we go back to your example, regarding the 1st point, then I am afraid that the survey sample is not large enough for the population at hand. Based on a population of 800 customers, a 5% confidence interval and a 95% confidence level, you would need a 260 a 300 respondents.

    Regarding the 2nd point, it is not inconceivable that a certain segment of the population will have a disproportionately higher/lower response rate to the survey. For instance, it could be that customers with a larger turnover, customers from a specific sector, etc. tend to participate more or less. To find out whether your survey sample is representative for the population you should compare the distribution of these characteristics between your survey sample and the population. Via logistic regression you can then find out whether there is some bias.

    Kind regards,

    Didier

    richard - January, 2014 reply

    I have a study population of 23,000 and I want use a study sample of 200, how can you justify this sample number statistically. Will I get good results.

    Didier Dierckx - January, 2014

    Dear Richard,

    Thank you for your question.

    It is hard to statistically justify a sample size of 200. Assuming a confidence level of 95%, you have a margin of error of 6,90%. This is quite large so you should definitely report this number so that readers can interpret your results.

    However, it is advisable to use a larger sample size for your research. You can calculate your sample size via our calculator: http://www.checkmarket.com/market-research-resources/sample-size-calculator/

    Kind regards,
    Didier

    melaku - October, 2014

    Dear Didier
    I just want to know how to estimate my sample size. the total population of vulnerable group in ten region is 11000 but I want to study only in one region how to now the population of these group in one region and how to calculate my sample size

    Yesuf ahmed - January, 2015 reply

    Dear,
    This is to ask question. I am doing my thesis on a population of 1406 non government Organizations. Then I am going to select responsdants from the saple organizations. What can i do first to select the organizations and the respondents from each organizations.

    Thanks

    Mike Gallo - July, 2015 reply

    A voluntary survey (one where people receive a mailer, email or phone call and choose to respond) will never accurately represent a population. In these instances, rather than obtaining data from a simple random sample (such as assigning a number to each potential candidate and using a randomizer program to choose your sample), you are obtaining data only from those that wish to participate in your survey. Beware of the results attained by voluntary sampling, as it fails to accurately represent the population and instead gives insight into only those that wish participate.

    Gert Van Dessel - July, 2015

    Mike, thanks for your comment.
    To deal with this it is always advisable to set quota for different subgroups (e.g. age, gender, region, …) to make sure the composition of your sample group is comparable to the profile of your entire population.

    Mehboob Ul Hassan - August, 2017

    Respected Sir/Madam
    I am doing Doctorate in Education. I am too much confused about my sample size. Total Population of my study is 125116 secondary school teachers from 9 districts of Punjab Province. Let me help in sample selection with reference so that I can defend

    Maarten Marijnissen - August, 2017

    Hi,

    If you want to compare those 9 districts you have to use 9 sub-samples for the 9 sub-populations of those districts.
    When this isn’t necessary, you can just use the total population and calculate sample size using our calculator.

    Just type in the population size, select a margin of error (usually 5%) and confidence level (usually 95%) and you’ll get your required sample size. It’s 383 in the second case.

    Good luck!

Leave a Reply

Your email address will not be published. Required fields are marked *