2 ms·

>> This sounds fishy. Sounds like their primary result was nothing If you combine results from a bunch of studies that really didn't do anything, it's going to

• mlyle 4y ago
  Sure... but conversely, if you keep dividing up and analyzing in different ways, eventually you'll get a p<0.05 result. This is why you "call your shot" before beginning the analysis. Otherwise, people will be suspicious you just jiggered things around until you found something.
• oldgradstudent 4y ago
  Maybe, but with 14 studies you have ~2^14 ways to split the studies into two groups. If you test multiple hypotheses, you have too adjust the p-value accordingly. Have the authors disclosed how many alternative hypotheses were tested until the result in the article was found to be significant?
  • anonymouskimmer 4y ago
    P-values don't prove anything about whether an effect is real or not. They, at most, provide some degree of confidence with the presumption that the independent and dependent variables are indeed independent and dependent. P-values should, at most, be used to direct further mechanistic studies (if possible). Only use them on their lonesome if this hasn't yet been done, or isn't possible. And if that's the case, reverify them independently (such as by doing another meta-analysis using different data).