Search found 16 matches

by andrewjdbell
Mon Aug 06, 2018 8:55 am
Forum: R2MLwiN user forum
Topic: saving sixway plots from R2MLwiN
Replies: 1
Views: 2176

saving sixway plots from R2MLwiN

Hi Chris (sorry, me again!) I'm trying to automatically save sixway plots from multiple R2MLwiN models - I'm running lots of models at once so don't want to do it by hand. When I try to use the sixway command, the plot seems to close down automatically after ~5 seconds, and won't let me save it. Any...
by andrewjdbell
Mon Aug 06, 2018 8:23 am
Forum: R2MLwiN user forum
Topic: Different ESS in MLwiN and R2MLwiN
Replies: 11
Views: 4459

Re: Different ESS in MLwiN and R2MLwiN

Thanks Bill, makes sense :) (and yes I checked with and without hierarchical centering at both levels - didn't make any meaningful difference). These are age-period-cohort models so perhaps unsurprising they aren't mixing well - we're interested in the deviations so in a sense it doesn't matter so m...
by andrewjdbell
Fri Aug 03, 2018 10:15 am
Forum: R2MLwiN user forum
Topic: Different ESS in MLwiN and R2MLwiN
Replies: 11
Views: 4459

Re: Different ESS in MLwiN and R2MLwiN

Aha - yep that's what makes the difference! (the code change, not the hierarchical centering) Thanks. Weird that it makes that much difference, given the correlations are so high between iterations in this model (see 6-way plot below). Any idea why it makes such a big difference, and would I be righ...
by andrewjdbell
Thu Jul 19, 2018 3:15 pm
Forum: R2MLwiN user forum
Topic: Different ESS in MLwiN and R2MLwiN
Replies: 11
Views: 4459

Re: Different ESS in MLwiN and R2MLwiN

Thanks for this Chris! That function seems to work and gives numbers that are pretty comparable to what I got in MLwiN. I guess the question then is: which one is right? I'd love to just use the MLwiN ones as it gives me the answer I want, but would be nice to have more of a justification than that ...
by andrewjdbell
Mon Jul 16, 2018 8:03 am
Forum: R2MLwiN user forum
Topic: Different ESS in MLwiN and R2MLwiN
Replies: 11
Views: 4459

Re: Different ESS in MLwiN and R2MLwiN

Possibly - not sure my coding ability is good enough to be sure - I'll see if I can work it out! Although given the parameter estimates are (slightly) different as well, possible that there's something else that's different as well? (normally wouldn't be fussed by the slight differences in MCMC but ...
by andrewjdbell
Fri Jul 13, 2018 4:21 pm
Forum: R2MLwiN user forum
Topic: Different ESS in MLwiN and R2MLwiN
Replies: 11
Views: 4459

Different ESS in MLwiN and R2MLwiN

I'm getting quite different ESS scores in my models comparing R2MLwiN and MLwiN, using the same models and data. With MLwiN, I get the following results: model S.E. ESS Response deaths Fixed Part cons -2.227 0.028 1159 (age-gm) 0.048 0.001 511 (age-gm)^2 0.001 0.000 18439 (year-gm) -0.031 0.001 512 ...
by andrewjdbell
Fri Jan 13, 2017 6:08 pm
Forum: runmlwin user forum
Topic: Cross-classified APC models
Replies: 3
Views: 2122

Re: Cross-classified APC models

Hi David (I know I emailed you too, but in case others are interested...) With your model, you should only be getting one variance for period and one for cohort. This is the variance of your cohort / period residuals (your residuals are not regression coefficients so will not be displayed as such). ...
by andrewjdbell
Wed Apr 13, 2016 1:58 pm
Forum: runmlwin user forum
Topic: runmlwin with starting values, with MLwiN 2.36
Replies: 2
Views: 1024

runmlwin with starting values, with MLwiN 2.36

Hi, I am running an MCMC model from runmlwin where I specify the parameter starting values in the stata code. However, with the new version of MLwiN, 2.36, I get a load of warnings when MLwiN opens, such as: "Column c1098 is protected, overwrite protected column ?" with a yes and no option. My quest...
by andrewjdbell
Tue Aug 18, 2015 12:30 pm
Forum: MLwiN user forum
Topic: MlWin gives different results to Stata/R
Replies: 3
Views: 1265

Re: MlWin gives different results to Stata/R

Hello Katrijn, Interesting. I don't have a simple answer I'm afraid, but a few things to check: Have you centred the variables consistently across the model - might make a difference when the interactions go in if what zero means is different. The exact zero on the education*age interaction is suspi...