Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

Supported by

Probably dumb question about Bayesian t tests in JASP

This is probably a really dumb question, but when I run a Bayesian t test on some data using the default scaled Cauchy prior (sigma=0.707) I get one BF, yet when I define a custom prior to be a Cauchy with the SAME scale, I get a different BF. Note that my data set is huge (~20,000 points). My first thought was that the scale in the default prior is really an irrational number and 0.707 is rounded simply for display purposes. Is this true?

Comments

  • EDIT: Rather than custom, I meant to say informed.

  • Good question! The .707 from the default test is really 1/2 * sqrt(2). What you can do to check is use the informed test with "0.7071068" instead -- the result should be closer. Also, we have a better routine for the informed t-test now (much faster too); we hope to implement this soon.
    Cheers,
    E.J.

Sign In or Register to comment.