I am conducting a quasi experimental study. Both my control and study groups consist of 30 participants. I wasn't able to compute for the effect size etc. beforehand. Please help me to justify that 30 per group is enough mo make my study valid. Thank you.
Asked
Active
Viewed 138 times
0
-
1I think you have to be a bit more specific, as it depends what you're measuring. You could theoretically justify a sample size of 5, if what you're concerned with is the median, as a sample size of 5 is sufficient to build a ~0.9 CI around the median. – chappers May 11 '15 at 03:09
-
Thank you chappers. I am actually trying to determine the significant difference on the post test scores of my control and study group. Please help me. I don't know what to do. :( – Belle May 11 '15 at 03:35
-
In that case you could perform something like a t-test, which is designed to work for things smaller than your one. It was discussed here: http://stats.stackexchange.com/questions/37993/is-there-a-minimum-sample-size-required-for-the-t-test-to-be-valid The key thing is you'll have to check whether your data is reasonable; to make sure whether or not your study is valid, that would probably assess the power of your test to determine if it is useful in a practical sense. – chappers May 11 '15 at 03:45
-
Thank you Chappers. Can I chat you in private? Please do help me more. :( – Belle May 11 '15 at 05:27
-
1On current information in the Q. I don't see any basis whatever to justify 30 per group. What was the reason you actually ended up with that number? – Glen_b May 11 '15 at 05:33