EJ
About
- Username
- EJ
- Joined
- Visits
- 2,531
- Last Active
- Roles
- Member, Administrator, Moderator
Comments
-
Hi Alexa, The correlation with the default (uniform) parameter prior under H1: I have no problems with it, and it is what Jeffreys proposed. But you might try to make H1 directional (you probably have a direction in mind? either a positive or a nega…
-
Hi Ellen, Hf = full model (with interaction); Hm = main effects only model; H0 = null model BFf0 = 1.475; BFm0 = 5.058 BFf0 * BF0m = BFfm, or BFf0 / BFm0 = BFfm, meaning 1.475 / 5.058 = BFfm = 0.29, so data are about 1/0.29 = 3.45 times more likely…
-
Strange. This is really a topic for our GitHub page, but I'll attend the team to this post. Cheers, E.J.
-
If you'd like some twitter credits you can send your complete name or twitter handle -- otherwise you'll be acknowledged be Maurizio. BTW, do you have this as a gif? (this is the format we generally use)
-
Hi Maurizio, are you OK if we add this video to our website and point to it on Twitter? E.J.
-
Hi Mila, I don't think (1) is an option, and I'm not sure about (2). There are some Bayesian network models you could check out. I believe there is some recent work by Joris Mulder (Tilburg Uni) that may be relevant here. Cheers, E.J.
-
Hi Nancysara_123, Please let us know on our GitHub page what meta-analysis features you'd like JASP to have! We aim to make students and researchers less dependent on commercial software. (for details see https://jasp-stats.org/2018/03/29/request-fe…
-
Dear Tali, A similar question on this forum was answered by Dr.Mark, as follows: "Hi Stat_b The default output for linear regression is Model summary ANOVA Coefficients - with the intercept and unstandardised B values Mark"
-
OK, that's informative, I'll forward your post. E.J.
-
Actually, before I call in outside help, did you inspect these resources? https://jasp-stats.org/2018/06/27/how-to-filter-your-data-in-jasp/ and the gifs listed under https://jasp-stats.org/how-to-use-jasp/ E.J.
-
I'll forward your post to those in the know. E.J.
-
Hi Leonie, This is a good example of why it is not universally true that an interaction is tested by comparing the main effects model to the full model. The entire pattern of results across all models is important. I would present the entire table a…
-
Hi Stat_b, Lisa, Eric-Jan will ask Erik-Jan. E.J.
-
Hi Naomi, In general, asymptotically, and on average (!), increasing sample size leads to more conclusive BFs. But for a specific data set, a BF might become less compelling as N grows (up to a point). Also, the BFs may increase (again, on average) …
-
Hi Raquel, No, they don't (you can check by eliminating some correlations -- the BF between that remains should not change). We could actually implement a correction, as we did for our post-hoc test for ANOVA. I'll take this up with the team. See al…
-
Hi Rudale, When you say that your interval measure is continuous, do you mean that it can take on any value higher than 0 seconds? If you manipulated it, it seems it might be a factor with several levels (e.g., 2s, 4s, 8s). If it is the delay before…
-
Hi Arthur, I'll pass this on to Richard. Cheers, E.J.
-
Hi Olivier, There actually exists a workaround, based on transitivity. See https://psyarxiv.com/u8m2s/ Cheers, E.J.
-
No. I'll ping them again. E.J.
-
I don't think it can, at the moment --link functions need to be added, and the default prior distributions will need to change-- but I'll ask how Richard feels about this. E.J.
-
Me too! I'll ping Johnny. E.J.
-
Hi Mila, For some reason the png files do not work for me (I'll ask Sebastiaan what's up). This makes it difficult to answer your question. Could you specify exactly which analysis option you are using? As soon as I know I will pass on your question…
-
Hi Mark, Can't see the png's, but I think it is the result of rounding. When all of the models with non-zero (up to rounding~) posterior probability include the effect, the summed posterior inclusion probability is 1.0, and the corresponding inclusi…
-
Hi Mark, Sorry for the tardy reply. I'll ask Koen Derks. Maybe it's a known issue. Cheers, E.J.
-
Dear Kiner, It does seem as if MANOVA is the way to go. I'll ask Johnny van Doorn about documention and possible JASP examples. Actually, before I do, does this gif help? https://jasp-stats.org/wp-content/uploads/2019/06/Manova.gif Cheers, E.J.
-
Hi Georges, It actually was more or less a joke, to tease the Bayesians who (correctly) felt that evidence is continuous, not discrete. However, maybe some people find it useful -- you could argue that the non-diagnostic BFs deserve less attention t…
-
Hi Flix, Thanks -- I'll pass this on! Cheers, E.J.
-
Hi OCOD, Good point. I will pass this on to Don van den Bergh, who did most of the recent work on this. On a side note, we are really close to having R syntax support, which should hopefully make everything clearer than it is now. Seems we are two v…
-
Hi Rick. Sorry for the tardy reply. I can only respond with some brief suggestions and general ideas: If you want more flexibility (long format for instance) take a look at the BayesFactor package in R. This is not a trivial stats problem, it seems …
-
Hi Simon1, These thresholds are somewhat arbitrary, and the evidential value is in the continuous measure. See for instance https://www.bayesianspectacles.org/lets-poke-a-pizza-a-new-cartoon-to-explain-the-strength-of-evidence-in-a-bayes-factor/ So …