I am working on "Nephrotic syndrome and behavioral problems". I want to know what should be the sample size or how to decide sample size in clinical data. Is there any method to decide or it is just a purposive sampling?
to do a sample size calculation you first need to know what you are trying to measure, i.e. what is your outcome measure, what difference are you trying to find. You also need to know the spread/SD of that measurement among your population - you may get this from pilot data or previously published research that accurately reflects your population.
With this info you can discuss it with a statistician or use a sample size calculator if you have some stats knowledge. I use GraphPad Statmate 2 which you can download.
Hope this is helpful,
C
Catherine has made a valid point. Before venturing to a biostatistician for help I would suggest the following.
1) Decide on the timeframe of your study- how many months/ years?
2) Any cost factors involved?, any incentives offered to the study participants?- what is the amount you are willing to bear out of your own pocket in case the study is not funded.
3) How much time can you realistically afford to spend on the study?
4) What is the purpose of your study?, do you have an assumption in mind? Do you think your study can throw significant light on the problem?
5) What is the protocol going to be?- have a flowchart ready
6) How many patients with nephrotic syndrome do you hope to encounter?
7) Realistically - how many nephrotic syndrome cases come to the hospital in which you are doing your study?
8) Is any follow up required? Are you planning any interventional behaviour therapy?
9) The sample size depends partly on the end goal, prevalence of NS in your area/ no. of cases frequenting your hospital. Usually the more common the condition- the greater the sample size required for validating your findings.
10) If it is just going to be a questionnaire based - cross sectional study- then the work might actually be easier rather than a prospective intervention based study- so you do have to be careful there. An intervention based study with follow up will always have more value than a simple cross sectional study.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size