If I want to measure the programme effects by using a questionnaire with 20 items and having a control and intervention group, how I can calculate the sample size? to write my proposal.
any advices or suggestions.
thanks .
From my point of view, there is a general rule there:
Ideal case: 20 - 1 ratio, 20 samples/question
Acceptable case: 10-1 ratio
Weak case: 5-1
There are so many circumstances to determine a suitable sample size depending on the research context and statistical methods used. If your research is in the well-studied field or to confirm a hypothesis with normal distribution, the sample size should be large enough (suggestion of at least 1000), if you want to explore new knowledge, the sample size can be small (no assumption about normality). If we have no idea of which statistical technique to use now, 100 samples are a good start, however, I recommend 200-300 for a grant.
Hello Albandari,
this is a question about statistical power. Google for "power analysis". There is a free software (g-power) in which you can calculate the sample size, given an assumed / expected effect size, that you would have to gather to achieve a significant effect in 80% of the cases.
Best,
Holger
You may need to think about a few more questions to do with your design before you can figure out an appropriate minimum sample size (based on a power calculation). As Holger Steinmetz says above, G*Power is very useful. You might benefit from reading a bit more about power analysis first so you can be confident about what you're doing, though! I'd suggest at least reading a basic overview, e.g.: https://stats.idre.ucla.edu/other/mult-pkg/seminars/intro-power/
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size