Bayesian independent sample t-test and quasi sequential Bayes factor

sloughon

Hello,

I apologize in advance for a slightly trivial question.

I am conducting a study in which I compare 2 independent groups. The first recruitment ended at n = 15 in each group (quite naively :D), I made a directional hypothesis and using Bayesian independent t-test I got BF 0.677. This value, of course, is not very convincing. For this reason, I figured I could recruit another n = 15 to each group for a total of n = 30 in each group. My question is, does double recruiting in this format (and seeing the data in the meantime) have any negative consequences on the interpretation of the results? In classical NHST it would of course have, also in a Sequential Bayes factor analysis it would produce a greater proportion of false positive and false negative results.

Thank you very much in advance

EJ

This is not a problem at all. In fact it is one of the strengths of the Bayesian paradigm. See for instance

https://psyarxiv.com/5ntkd

sloughon

WOW, that was an incredibly quick response - thanks so much for your help!