Sign in with your CogSci, Facebook, Google, or Twitter account.

Or register to create a new account.

We'll use your information only for signing in to this forum.

Supported by

Do we need to report Frequentist analyses when using Bayesian inference?

On the manual of BayesFactors, the only times you use frequentist inference is as a means of facilitating comparison. In The New Statistics: Why and How, Geoff Cumming wrote:

  1. Whenever possible, avoid using statistical significance or p values; simply omit any mention of nullhypothesis
    significance testing (NHST).

But I have found many papers saying that they obtained certain BF supporting the null or alternative hypothesis. Is it possible to obtain all the information needed for publishing a paper, by using **only ** the BayesFactors package? I haven't tried JASP so I don't know how much they differ from each other in terms of the information one can get using each one.


  • EJEJ Posts: 375

    Hi Aram,
    1. As far as Bayesian analyses are concerned, we are trying to keep BayesFactors and JASP in sync.
    2. There is a confusion in the literature about classical NHST and BFs. My perspective is that p-values are not to be recommended (but you might mention them in a footnote to pacify reviewers); however, hypothesis testing per se is often exactly what researchers want, and imo there is nothing wrong with that. You just have to test hypothesis the right way, and basically replace p-values with BFs. Zoltan Dienes has a more inclusive perspective: "a B for every p".
    3. After a few drinks, when I've mellowed a bit, I can see the value in reporting both p, BF, and Bayesian estimation. Sometimes the stats are really important for supporting a key claim, and it can't hurt for the readers to have all the information at their disposal.
    4. As far as the limitations of a "estimation only" approach, see part I here: https://osf.io/m6bi8/

    Thanked by 1Aram
  • sebastiaansebastiaan Posts: 2,774

    Just to avoid confusion: The term "null-hypothesis significance testing" (NHST) is generally used as a synonym for reporting p values. So if you test a null hypothesis with a Bayes Factor, most people would not call that "null-hypothesis significance testing", even though in a sense it is—a semantically confusing legacy from the times when NHST was (perceived to be) the only way to test a null hypothesis.

    Thanked by 2Aram EJ

    There's much bigger issues in the world, I know. But I first have to take care of the world I know.

Sign In or Register to comment.