EJ
About
 Username
 EJ
 Joined
 Visits
 430
 Last Active
 Roles
 Member, Administrator, Moderator
 Thanked
 69
Comments

Hi Anna, We have this under development at the moment! Will take a while but we are really happy with the analytical result. E.J.

Hi B, If your job is data management, I would stick to Excel (or OpenOffice Calc). JASP has some data editing functionality (compute columns, filtering, etc) but it is not a fullfledge editor; JASP's main strength is statistical inference. Cheers, …

Hi KenC, We try to avoid MCMC as much as we can. For many of our tests, we have analytic solutions, or only require a onedimensional integral. When we do use numerical methods, the error percentage gives an indication of the accuracy of the approxi…

Dear Tali, Hmm, the complication here is that "N" is involved both in NC and in ICN. So it seems to me that you would like some kind of contrast that tests the symmetric staircase pattern against an asymmetric staircase pattern. Julia Ha…

Dear Philip, I guess the priors are implied. BIC is an approximation to the Bayes factor under a specific prior assumption for the model parameters (I believe a multivariate normal mean centered at the MLE but with unit information to set the prior…

Update: if you present the R code, Richard can look with the BayesFactor package at what you did! Cheers, E.J.

I'll ask our EFA expert. Sorry for the tardy response, this slipped through the cracks...more to follow Cheers, E.J.

I'm not aware of a Gates' delta. It's the population mean divided by the population sd, so the population version of Cohen's d (i.e., Cohen's delta). E.J.

Hi Niklas, I have the strong sense that something went wrong with the model specification, but I'll gladly pass this on to our ANOVA expert...more to follow. Cheers, E.J.

I've attended the experts to your question...

Dear BrittJane, In general, I think commenting is the more responsible option. This is part of statistics that is still in flux. At the same time, it is a nice robustness check. Cheers, E.J.

Hi Lior, Hmm I'm not so sure. You are right about the 28% corresponding to a BF change of 4, but I would have guessed it indicates that the BF10 can fluctuate from approximately 12 to 16. I'll doublecheck... Cheers, E.J.

Correct. The sample effect is slightly positive, so changing to a onesided positive hypothesis is helpful for H1, but changing to a onesided negative hypothesis is harmful for H1 (because the sample effect is in the opposite direction of what H p…

Hi Niklas, Well, I have not Googled this, but it seems to me that when you add a covariate, you are changing the main effects. Suppose you test IQ scores of dogs to those of cats, but you have reason to believe that "number of pets per day"…

Dear Guillaume, The "a" means it is an alphabetical (text) variable. So at least one of the entries has text, and this is also why you can't change the type  JASP cannot guess what number the string should have. Cheers, E.J.

Hi scd21, A useful thesis on the topic (by one of the members of the JASP team) is here: https://psyarxiv.com/s56mk I would report the entire table. The uncorrected BFs are just the regular ttests; the penalty for conducting multiple tests comes in…

Dear Niklas, This is strange behavior! Apparently lVisC is in almost all models with appreciable posterior probability, but its modelaveraged posterior includes zero. Under usual circumstances, this would not happen, so I'm going to guess that the …

Hi Deirdre, Yes this is tricky. The ANOVA is specified through differences from a grand mean. Resources & ideas: This paper: http://www.ejwagenmakers.com/2017/RouderEtAl2017ANOVAPM.pdf (Rouder, J. N., Morey, R. D., Verhagen, A. J., Swagman, A. R…

I'm going to pass this on to people who have a higher probability of giving you a useful answer...more to follow E.J.

Hello Gabriel, In general, BFDA is the way to go. We have not implemented this for ANOVAs, however. For fixed sample sizes, Bayesian analyses are more conservative (less triggerhappy is the right way to put it), so what you could do is select an al…

Yes, you can change the number of decimals that JASP reports in the Preferences > Results section. When p=1, sometimes it is exactly 1. This happens, for instance, when an effect size in a ttest is 0 exactly. More commonly, it will happen in a …

Hmm I don't quite get this then. The discrepancy between the analyses is really large. I will ask some other people to look at this as well. Also, you could ttest just the probe vs irrelevant difference for the index case  I assume your pvalue w…

Interesting case. Could you also provide the regular table with all the models separately? Maybe also a plot of the results? Sometimes such discrepancies are due to model misspecification, for instance heteroscedasticity etc. Cheers, E.J.

Hi Cristony, I suspect that our network experts may require some more information in order to help you out. The best place to get their help is our GitHub page (for details see https://jaspstats.org/2018/03/29/requestfeaturereportbugjasp/). Che…

Hi Gabriel We are working to code up all our analyses in an R package, but it's not ready for prime time just yet. Let me attend some of the team members to your question and see what we can do. Cheers, E.J.

There is some relevant thesis work by Rivka de Vries (Richard's former PhD student), see https://www.rug.nl/research/portal/files/15849947/Title_and_contents_.pdf Cheers, E.J.

Hi Niklas, The results are based on a numerical procedure, so some variability is inevitable; To understand the impact of your covariate you could do some exploratory analyses; The modelaveraged posterior is a weighted combination of the posterior …


Strange indeed. The results should be identical. Perhaps check to see that the descriptives match up just to make sure that the two software programs are looking at the same data? E.J.

This is really something for the GitHub page, but it's important so I'll post it there. Thanks! (my guess: for the ANOVA, it may save the samples, at least if it's a Bayesian ANOVA...) E.J.