#### Howdy, Stranger!

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

Supported by

# Bayesian reanalysis

Dear JASP team,

I am working towards my master’s degree and lately a Bayesian reanalysis of the literature seems to be an option to develop my dissertation – given our pandemic context. However I am struggling with the data necessary to undergo a Bayesian reanalysis on JASP – as presented on the “Webinar: Theory and Practice of Bayesian Inference Using JASP”. Specifically, I would like to know the feasibility of this method once some studies report their ηp2 while others provide me the η2 value.

As an example:

“Absolute error scores (cm) for practice were analyzed using a 2 (Choice: Self, Yoked) × 3 (Decision: Before, After, Both) × 6 (Block) mixed-model ANOVA with repeated measures on Block. All groups showed a reduction in AE across practice blocks, which was supported by a significant main effect, F(5,210) = 39.20, p < 0.001, ηp2 = 0.48”.

“Accuracy scores (AE) on the 30 practice trials were averaged across blocks of 5 trials and analyzed in a 2 (group: self-control versus yoked) x 6 (blocks of 5 trials) analysis of variance (ANOVA) with repeated measures on the last factor. […] Both groups reduced their AEs across blocks of trials (see Figure 2, left). The self-control group had generally smaller errors than the yoked group. The main effects of both block [F (5, 140) = 9.37, p<0.001, η2=.25] and group [F (1, 28) = 4.70, p<0.05, η2=.14] were significant”.

Are ηp2 and η2 equivalent so I can use either one in the reanalysis? And also, would it be appropriate to draw conclusions based on, for instance, the mean values of the obtained Bayes Factors?

Thank you for your time and I look forward to hearing from you.

Best regards,

Nathalia Venancio

• Hi Nathalia,

Hmmm, I was under the impression that ANOVAs do not lend themselves to a Bayesian reanalysis just from the reported information. I'll ask Julia.

Cheers,

E.J.

• Hi E.J,