Dear friends
I collected data in a questionnaire form with a total number of 500. I submitted my manuscript and now the reviewer asked me to verify sample size calculation. What will be the values that I must determine to approve the size of 500?
Yeah, but then THIS IS your justification, you just followed other studies. You can't come up with a good statistical explanation and arguement, why you chose 500, if you haven't had one in the beginning, this would be dishonest.
What you can do is a sensitivity analysis, which shows the smalles effect size you were able to detect with your sample.
Dear Ahmed Atia ,
Sample size calculations depend upon the size of the effect you are looking for, the Type II error threshold level that you set (normally 0.2), and the design of your study. The effect size should not be calculated from your data - it should be established from your literature review. There are free tools available for calculating sample sizes, such as G*Power.
Asking the other way around: what did YOU do a priori, to justify 500? Where did this come from? Why 500 and not 400 oder 600?
I believe that you must have used a formula to arrived at the sample size of 500. If not kindly go through the link below to study and chose the correct formula that suit your research work.
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3775042/
Thanks all for your cooperation, but I still can't find solution how I could replay reviewer comment and clarify about the 500 questionnaire
Ahmed, please tell us, how did you came up with 500? What was your rational to collect 500 datasets and not any other number?
We must first know the size of the community and the characteristics of the community to know the size of the sample and what is the appropriate method (sampling) to choose the required sample
I wish you success
Dr. Rainer, I adapted this number from previous study. But in that study the authors didn't give sample suze justification
Yeah, but then THIS IS your justification, you just followed other studies. You can't come up with a good statistical explanation and arguement, why you chose 500, if you haven't had one in the beginning, this would be dishonest.
What you can do is a sensitivity analysis, which shows the smalles effect size you were able to detect with your sample.
Hello Ahmed,
I agree with Rainer Duesing 's advice, that your rationale should be that you based your choice on the size(s) used in previous (and, I presume, related) studies.
You still haven't indicated what your goal was in collecting the survey data. Sample size determination is slightly different in the case of wanting to estimate population parameters (e.g., what proportion of adult females work full-time), vs. testing hypotheses (e.g., do men and women see medical professionals equally often in a year). So, for some situations, it's quite possible that 500 cases would be more than enough to satisfy your research aims, whereas for others, it might be far too few.
Good luck with your work.
Informative answers, appreciate all replies
I will go with Dr. Rainer's advice
Thanks alot
Ahmed Atia, you have been given some good advice above. I wonder whether, given the nature of your research, you need to justify your sample size in the way that is often the case. In fact, I wonder whether the reviewer might be using a stock-standard "requirement" that is not really appropriate under some, or many, circumstances. (Unfortunately, reviewers are not always sensible / informed in their requests, suggestions, and criticisms.)
Perhaps it is more important for you to describe the sample you obtained in a way that provides satisfactory information related to your research objectives. For example, is your sample likely to be representative of a relevant population? Might it be deficient in some way that you can acknowledge in the limitations section of your article?
Good suggestion from Dr Rainer Duesing. But adopting someone's sample size will definitely not satisfy good reviewer. Even the study he adopted from must have came up with the sample size based on their research objectives using acceptable statistical methods.
In future studies, you need to know and justify why you select a particular sample size, in order to go along with the methodology and answer your research objectives.
Hii Ahmed
A simple explanation to verify the sample size depends upon the number of variables and type of analysis you are using in your study. For example, for regression analysis, a minimum of 10 participants per variable is considered to be appropriate and for factor analysis, it could be 300 cases. Read the work by VanVoorhis and Morgan (2007) for further insight.
Reference:
VanVoorhis, C. W., & Morgan, B. L. (2007). Understanding power and rules of thumb for determining sample sizes. Tutorials in Quantitative Methods for Psychology, 3(2), 43-50.
It makes sense to say that you were replicating the methods used in a previous study, but more important is to discuss what the sampling error is; that's what people need in order to evaluate your data.
However, you should also document how the sample was chosen. It sounds as if you might have performed a convenience sample, in which case the formulas for computing sampling errors won't apply. You should discuss whether any biases are associated with the sampling approach you used.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size