Jack Bailey

Could you do this using non-linear syntax instead? Then you wouldn't need to soft-fix it. Instead, you could only provide factor loadings for x2 and x3 on ind60, effectively fixing...

I agree. My two cents, we need a generalisable syntax so that you can extend the type of latent variable modelling to categorical latent classes in the future without having...

Glad that you liked it! In response, I think that fixing needs to be more flexible than just loadings. Sometimes we need to fix correlations and variances too. These can...

I've given this a lot of thought, and I'd like to see this become a reality, so I'm going to keep making suggestions. Apologies in advance for what is likely...

This looks really promising! As for the interaction, it was just an example of flexibility that Bayesian CFA offers that's hard in Frequentist CFA. The example was overly-complicated just to...

I could well be wrong. And it's also worth keeping this thread limited to brms SEM syntax. Nevertheless, I'll explain my reasoning then we can move it. That way, you...

I'm so excited by this. It looks great. It would be neat to use `fix` both where fixing a loading to a value and to a label, in my opinion....

I have suggestions for 1 and 2. 1. How about keeping it simple and having a general function to specify correlations between variables: e.g. `cor(x, y)`, `cor(lv(x), lv(y))`. Where we...

Here's the R and Stata regression output. R: ```R Call: lm(formula = y ~ x * r, data = dta) Residuals: Min 1Q Median 3Q Max -2.86028 -0.65072 -0.00758 0.69345...

Just to add to this point -- it would be useful to use haven instead of foreign as foreign only supports Stata up to version 7 and we're now on...