Hello, I am researching the effect of a method on children's attitudes towards the a course. Due to the nature of this method, it should be carried out in groups of 10-12 people at most. Can I create an experimental group with 10-12 students? Is there any statistical issue with this?
The experimental methodologies require at least 15 participants according to Cohen et al. (2007:102), and there should be at least 15 participants in control and experimental groups for comparison according to Gall et al. (1996). These references can be taken by researchers using small sample size.
Cohen, L., Manion, L., & Morrison, K. 2007. Research methods in education. Routledge.
Gall, M. D., Borg, W. R., & Gall, J. P. (1996). Educational research: An introduction. Longman Publishing.
The problem with small sample sizes is that your analyses with lack "power" because of large standard errors. In practical terms, this means that you will have problems generating statistically significant results.
Thank you for your answers. I can’t carry out the research with the participants more than 12….Do you have any suggestion? Can I combine the results, if I create two separate experimental groups, each consisting of 12 people, and I apply the same procedure in both groups?
I'm not sure how two separate experimental groups would help. The classic experimental design is to compare an experimental group and a control group.
The minimum sample size is 15 participants per group. You might find the tutorial by Brysbaert (2019) helpful. Here is the full citation.
Brysbaert, M. (2019). How many participants do we have to include in properly powered experiments? A tutorial of power analysis with reference tables. Journal of Cognition, 2(1), 16. https://doi.org/10.5334/joc.72
Good luck,
Just a quick PSA: the "small" and "large" notions of standardized mean differences are not informative. See table 1 here: Article Empirical Benchmarks for Interpreting Effect Sizes in Research
which shows natural growth in academic skills by grade. The effect sizes are large in early grades and much smaller in later grades. A so-called "small" effect of .2 would be 13 percent of growth for a young child but 300 percent or more of a late adolescent's growth. Thus, constant sizes for effects are unintelligible and meaningless. This is less the case for effects for changes in the rate of something, doubling the odds of death, no matter how large or small the baseline rate, has a prima facie sense to it. Sorry, no easy answers here.
I am contributing a paper that is relevant to this discussion: Article How Many Participants Do We Have to Include in Properly Powe...
- Badges
- Science topic
- Similar topics
- Psychology