I am conducting a research study and for sample size calculation, i chose a paper with a similar study design to enter measurements in the sample size calculator. Is it correct to choose a paper with a similar design (pre-post design) and significant results to do the sample size calculation? And if it is correct, are there any conditions or guidelines to choose a paper for entering its measurements into the calculator?
Thank you
I applaude the approach to do a sample size calculation / power analysis before conducting the experiment and similar studies can be a starting point. But the question you have to answer for yourself is: how small is the smallest effect size of interest for me? So, look at your dependent variable(s) and think about, what amount of change will really matter in that variable. THIS is the smallest effect size of interest and this is exactly what you should choose for your sample size/power considerations. But read this excellent blog by Daniel Lakens for a further explanation of this topic.
http://daniellakens.blogspot.de/2017/05/how-power-analysis-implicitly-reveals.html
Dear Tanseem
1. It is possible to utilize any published scale of measurement and questionnaire in new research problem.
2. The sample size should be taken from the population of the new study and has no relation with population of other problems.
3.Research studies are usually carried out on sample of subjects rather than whole populations. The most challenging aspect of fieldwork is drawing a random sample from the target population to which the results of the study would be generalized. The key to a good sample is that it has to be typical of the population from which it is drawn. When the information from a sample is not typical of that in the population in a systematic way, we say that error has occurred.
Regards,
Zuhair
Using published results is the standard approach. The alternative is to run a preliminary experiment.
General problem: The sample sizes in either case may not be sufficient to accurately estimate the true (population) values for means and variance.
The problem in the former case: The calculations are only relevant to the degree that the population studied previously matches the population that you will be studying. The best you can hope for is that they are fairly similar.
The problem in the latter case: You expend resources (time, funds) to get the estimate, and you eliminate some cases from the population that you are trying to study. If the population has 600 individuals and you run a preliminary trial with 80 individuals, then you only have 520 individuals for the main trial.
Timothy raises some useful points in considering pilot samples. Err on the side of choosing a larger n than you calculate using Rainer's advice if cost is not a major consideration. Best, David
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size