Hi,

As far as sample size is concerned, there is a very popular formula developed by Krejcie and Morgan in 1970. Article Determining Sample Size for Research Activities

Based on this formula, a table has been generated that can be used to identify the ideal sample size that one must choose in case they know the size of the universe.

You can find the table here. https://www.kenpro.org/sample-size-determination-using-krejcie-and-morgan-table/

Krejcie, R. V., & Morgan, D. W. (1970). Determining sample size for research activities. Educational and psychological measurement, 30(3), 607-610.

Hello Erhan,

You're asking, by how much should you over-sample (go beyond what the value given by what one hopes is an appropriate formula, given your sampling method) in a study?

Yes, non-completion, inattention, reluctance or refusal to answer sensitive questions, drop-outs and the like can lead to incomplete/missing data sets. This presumes that the chosen cases actually show up and participate in the first place!

If your study is such that only complete data sets may be used (no imputation will be tried for missing data points), then you're probably wise to over-sample by 10% at a minimum. If you review the literature of similar studies, you'll likely get a better idea as to the rates of incomplete or non-usable data cases that others have noted; this might be the best guidance for your intended study.

Good luck with your work.

Thank you, David. it was very useful information. Do you know of a source for the 10% rate?

Thanks Rama,

How much larger can the sample size be than calculated? Does the source you suggested include this? I will read it, thank you.

Erhan Biyik , Although the source contains a table with a maximum universe size of 100,0000, one can use the formula to calculate the sample size for any given population, irrespective of its size. There are no limitations as such.

s. Rama Gokula Krishnan,

I think I didn't explain it clearly enough.

I calculated the sample as 784 using the formula, but I collected data from 820 people.

1. Can't I go above 784?

2. If I can get. what is the limit?

I hope I could explain.