EJ
About
- Username
- EJ
- Joined
- Visits
- 2,557
- Last Active
- Roles
- Member, Administrator, Moderator
Comments
-
Given the data and the uniform prior (and the likelihood), the probability is 95% that mu lies between the lower and upper bound. [I would use mu in order to distinguish between the sample value and the inference] Cheers, E.J.
-
Hi Rohanp16, I guess it is a little weird to add credible intervals in a "descriptive" table, because the intervals are an inference. These credible intervals are for the group means, under a uniform prior. You can change the priors (under…
-
That analysis uses the BayesFactor package in R. Have you checked their documentation? Cheers, E.J.
-
Hi Szymon, Do you mean the Bayesian linear mixed model, or the Bayesian ANOVA? E.J.
-
I'll attend Richard to your issue (he is busy so may not have time soon -- you could email him and let us know here what he says). Cheers, E.J.
-
Hi Scribe42, In many scenarios, the inference is very robust to the type of prior that is used. The primary Bayesian movement in statistics right now is "objective Bayes" and it uses general-purpose priors that meet specific desiderata. My…
-
Hi Scribe42, The priors can be set under "Advanced options" (linear regression) or "additional options" (ANOVA). The default options are not uninformed in the sense that they are not uniform, but they are fairly spread out and ce…
-
Yes, so the model including Groups get strong support from the data (over the null model). You might want to look at R2 that is not model-averaged, but based on the model including Groups. Cheers, E.J.
-
Hi Markus, Can you paste the model table? If the BF supports the inclusion of the group factor than this is what counts. The model-averaged R2 credible interval informs you about the size of the effect, not about whether it is present or absent -- i…
-
Hi Janina, This is a tricky one! Right now I don't think this can be done in JASP, since the parameter priors for ANOVA are all centered on zero. However, perhaps it is possible to consider a t-test reformulation of the main idea? Cheers, E.J.
-
Hi Brandt, -- Reviewer 2, nice :-) This would be a great feature request! (for details see https://jasp-stats.org/2018/03/29/request-feature-report-bug-jasp/) In the meantime, I am sure that there is an R package for this...the easiest solution is t…
-
Indeed, it is as Famondir says. The differences are due to the stochastic nature of the MCMC process. By increasing the number of MCMC iterations this variability can be reduced, but it will never be zero exactly. E.J.
-
He's trying to finish up his thesis before Jan 1st, but I'm sure it will be his first order of business in the new year :-) E.J.
-
Hi ARF, I do not know why you would want to subsample, but I will take this as a given. One issue is whether the subsamples are partially overlapping, as you'd get from a standard bootstrap approach. If so, this overlap needs to be taken into accoun…
-
I am pretty sure we offer a complete analysis, but I'll forward this to the experts. Cheers, E.J.
-
Hi Noortje, Well, if you just plug in the numbers you'll automatically get a test that compares against zero (which is not what you are interested in, as it makes no sense). But you will also get a confidence interval, which would be what you are lo…
-
Hi Michif, Sorry for the tardy response. Yes, the app you found only does t-tests Correct, the power analysis functionality is still missing in JASP Reducing alpha to .005 is indeed a sensible alternative approach The classical concept of power refe…
-
Hi Tim, I'm not an expert on the LASSO either (why not use the Bayesian option?) Anyway, I think that for general background you'll have to consult books, Wikipedia, blog posts, online courses, YouTube, etc. Cheers, E.J.
-
Hi Eustis11, Sorry for the tardy responses, I have been a bit swamped. No, JASP does not do time series analysis yet, but we are taking it under development. Cheers, E.J.
-
Hi Eustis11, Not yet. There are methods in R to do this; specifically, Rivka de Vries and Richard Morey worked on this problem. Rivka's dissertation is here: https://www.rug.nl/research/portal/files/15849947/Title_and_contents_.pdf Cheers, E.J.
-
Hi Michael, We are wrapping up the paper now, and I suspect it is more efficient if we send it along when it is done. It is surprising how what complications lurk beneath the surface of a relatively straightforward model. Cheers, E.J.
-
Hi Michael, Thanks for your question. Some quick thoughts: "BF of the model including the effect in question divided by the highest-ranking model without that effect." I do not like this procedure because it cherry-picks the highest-rankin…
-
I would need to see a screenshot of the table, but I can already note that the BF measures relative predictive performance -- it is the evidence or the degree to which the data change our opinion. It is therefore *not* the probability or plausibilit…
-
I believe that "condition on" allows you to compute a partial correlation. So it seems that this is what you want. I will ask our experts. Cheers, E.J.
-
So for me, this happened once after I upgraded to a new version of JASP. The solution that seems to work is re-running the analyses by typing CTRL-R (Refresh), or CMD-R on Mac. Does this work? E.J.
-
If you have too few observations in your design, higher level interactions cannot be estimated. I suspect this is the problem here. E.J.
-
Hi KiKiSoul, Perhaps your variables are not recognized as scale? Do they have a ruler icon next to them? Maybe you can send a screenshot. If you want JASP to recognize variables as scales rather than ordinal values, you can set the threshold under &…
-
Let me check...
-
[I have now tweeted this, https://twitter.com/JASPStats/status/1332345849700114433]
-
I will tweet this (sorry about the tardy response). We are in the process of constructing a forum specifically for teachers, but that is work in progress. Cheers, E.J.