4 ms·
sample size please - how is this even publishable?
by boringg 7y ago
sample size please - how is this even publishable?
- domnomnom 7y agoCentral Limit Theorem
- robbrit 7y agoFrom the article: > In this study, we tested 77 undergraduates who were randomly assigned to play either a popular video game (Portal 2) or a popular brain training game (Lumosity) for 8 h.
- boringg 7y agoSorry my comment should have been clearer. I wasn't actually asking for the sample size - i was making a statement that the sample size was ridiculously small as to be irrelevant. It was more like an eye roll. Thank you for responding to my literal comment snd reminding me to be clearer in my messages :)
- robbrit 7y ago77 is a large enough sample size to produce statistically significant results in a lot of cases, depending on the problem. For example, using 77 men and women, you could easily see a statistically significant difference in height between genders. This paper is pretty simple, it's just a straight-forward before-and-after test. They are holding a lot of factors constant and not doing an observational study, so 77 is actually a fine sample size for something like this.
- tir 7y agoLiterally the first line of the abstract.
- mattkrause 7y agoThere's a power analysis in the paper showing that they've got plenty of statistical power. (Estimated N=34 subjects for 90% power, but they used 77 subjects instead).