EJ
About
- Username
- EJ
- Joined
- Visits
- 2,597
- Last Active
- Roles
- Member, Administrator, Moderator
Comments
-
My first thought is that you'd have to weigh the error percentages with the posterior model probabilities, such that error percentages have a bigger impact when they are associated with models that exert a large effect on the end result. Cheers, E.J.
-
This is a good question. I'll pass it on to our experts. My initial gut-level response was to say "why not use SEM"? But maybe we ought to expand our CFA to make this possible (or maybe the upcoing version can do this -- we'll know soon en…
-
Probably because the values are relatively large. If you express the item price in a different unit (e.g., times a thousand of what you have now) the scales will return to normal. E.J.
-
Yes, you can always transform to normality. We also offer several rank-based Bayesian methods (depending on the analysis you want to do) E.J.
-
I'll attend Johnny van Doorn to this, but if you want to stress your recommendation please post it as a feature request on our GitHub page! (https://jasp-stats.org/2018/03/29/request-feature-report-bug-jasp/) E.J.
-
Yes, definitely, this just looks like a bug -- there may even be a simple fix, let's see what they say
-
That is strange and unexpected. It would be great if you could post this on GitHub, so the programmers can fix it! Cheers, E.J.
-
Hi Laura, Thanks for noting this. I'll pass it on to the relevant team members and they will get back to you soon, I hope. Cheers, E.J.
-
Yes, but I believe that substantive hypotheses usually translate to ordinal expectations. We have developed something like this for the multinomial, see https://psyarxiv.com/bux7p/ Cheers, E.J.
-
Hi Danilinares, Yes that is a really good question, and one that is important too. In fact we currently have this under investigation. It is not something that JASP does right now (but we will in the future). Cheers, E.J.
-
That probably means the model is really big. The question is whether this is an issue in R as well (it probably is). I'll pass this on to our expert. Cheers, E.J.
-
I'll ask the expert! Cheers, E.J.
-
Hi Lea, Sometimes the evidence is so large that it cannot be represented on your computer. However, since you are interested in the interactions I would go to the "model" tab and add all of the main effects to the null model. Now your BFs …
-
Hi Sofie, The analysis of effects considers *all* models in the set; it sums the posterior probability for all models that include the effect (and then compares the posterior inclusion odds to the prior inclusion odds to obtain the inclusion BF). So…
-
Hi Firona, I am pretty sure JASP includes the covariate as you would in R. So your question would then be on a more conceptual level, I assume? Intuitively I've always thought of the covariates as follows: you include only the covariates, and this y…
-
I don't think this is possible at the moment. Would be a good feature request! I've forwarded this to our network expert. Cheers, E.J.
-
Feedback: there is one false alarm among many packages. We're currently struggling to get a new version out; maybe the new version will not have this issue, or maybe more people will experience the problem. We will reconsider the situation then. Tha…
-
That does not appear to be the case, but I'll ask our expert. E.J.
-
That is correct. The figure is meant to be descriptive, so it uses a flat prior (and therefore gives the same result as the frequentist confidence intervals). The complication with basing the credible interval on the model is that there are often mu…
-
Kruschke developed a Bayesian t-test (with the emphasis on "a"). This is not the one implemented in JASP, which is based on the work by Jeffreys. In the help file you will see a reference to Rouder et al., 2009, and Gronau et al. The assum…
-
I've checked and this does not ring a bell for anybody in the team. But we've added it to our list. Cheers, E.J.
-
Dear jmbostwick, "My understanding is that the BF01 for a given model can be roughly interpreted as the odds that the best model is a better fit than the given model. For instance, the model in Box 2 has a BF01 value of 2.064, meaning the best …
-
It uses metaBMA: https://www.dwheck.de/software/metabma-bayesian-model-averaging-for-meta-analysis/ Cheers, E.J.
-
Hi Max I'd consider group fixed in both scenarios. Cheers, E.J.
-
I'll forward this to our experts Cheers, E.J.
-
I'll forward this to our expert. Cheers, E.J.
-
I'll forward this to our expert... E.J.