I’d like to share that my book, Multilevel Thinking: Discovering Variation, Universals, and Particulars in Cross-Cultural Research is available for pre-order at Oxford University Press: https://global.oup.com/academic/product/multilevel-thinking-9780197806463?cc=us&lang=en&#
* This book explains multilevel modeling building upon my 15 years of teaching this topic.
* The book contains appendices for estimating multilevel models in #Stata, #Rstats, and #Julialang.
Odd-numbered years bring new #Stata releases. But for the first time in more than a decade, my unit will probably not upgrade our licences. Which means that we will miss out on this new killer feature 
https://www.stata.com/new-in-stata/french-interface/
Quasi-Magic:Average marginal effects, marginal effects at the mean, and the margins command in #Stata
https://www.kai-arzheimer.com/me-at-the-margins-average-marginal-effects-marginal-effects-at-the-mean-and-statas-margins-command/
https://www.kai-arzheimer.com/me-at-the-margins-average-marginal-effects-marginal-effects-at-the-mean-and-statas-margins-command/
Interested in #eusilc?
#SPSS, #Stata, and #Rstats routines to transform the raw data of EU-SILC into system files are now available for the most recent EU-SILC release 2024-release1: https://www.gesis.org/en/missy/materials/EU-SILC/setups
Attain standard errors for practically anything: nlcom and the Delta Method 
#Stata
https://www.kai-arzheimer.com/delta-method-nlcom
New blogpost on offsetting points on Stata marginsplot graphs
https://www.erikreinbergs.com/offsetting-points-in-stata-marginsplot-graphs/ #stata
ARGH. I am trying to convince my study group that I am correct and they are not, in terms of our statistical analysis. We do not yet know who the prof thinks is right! But ... I legit think that one of my compatriots has gotten their answer from ChatGPT or something. And it is WRONG, in my opinion. I may die on this hill. Sigh.
The prof actively told us we could to use ChatGPT to get the right code to run in Stata. FML.
#gradschool #stata #data #stats
Useful but confusing:Average marginal effects, marginal effects at the mean, and #Stata's margins command
https://www.kai-arzheimer.com/me-at-the-margins-average-marginal-effects-marginal-effects-at-the-mean-and-statas-margins-command/
https://www.kai-arzheimer.com/me-at-the-margins-average-marginal-effects-marginal-effects-at-the-mean-and-statas-margins-command/
Get standard errors for nearly anything: nlcom and the Delta Method #Stata
https://www.kai-arzheimer.com/delta-method-nlcom
Quasi-Magic:Marginal effects at the mean, average marginal effects, and #Stata's margins command
https://www.kai-arzheimer.com/me-at-the-margins-average-marginal-effects-marginal-effects-at-the-mean-and-statas-margins-command/
https://www.kai-arzheimer.com/me-at-the-margins-average-marginal-effects-marginal-effects-at-the-mean-and-statas-margins-command/
@Carwil 1. The World Bank or UN may have some resources for translation. They obviously have tons of that stuff.
2. The proper approach should of course be multilingual category labels. #Stata supports these through their equivalents of attr(). The limited and unstructured support for labels in #rstats is available through a mix of 
haven, labelled, sjlabelled, and ipumsr. In particular, the latter has useful relabeling functions.
The article is from 2023, but I think that other ageing #Stata holdouts may find it useful, too
https://blog.stata.com/2023/09/12/from-datasets-to-framesets-and-alias-variables-data-management-advances-in-stata/
Attain standard errors for virtually anything: nlcom and the Delta Method 
#Stata
https://www.kai-arzheimer.com/delta-method-nlcom
Almost like magic:Average marginal effects, marginal effects at the mean, and #Stata's margins command
https://www.kai-arzheimer.com/me-at-the-margins-average-marginal-effects-marginal-effects-at-the-mean-and-statas-margins-command/
https://www.kai-arzheimer.com/me-at-the-margins-average-marginal-effects-marginal-effects-at-the-mean-and-statas-margins-command/
@ereinbergs #Stata: egen ultPSU = group(strata PSU)
#rstats: df %>% group_by(strata, PSU) %>% mutate(ultPSU=group_indices())
More visually appealing would be
ultPSU = stratum*1000 + PSU so that you could clearly see which stratum this comes from.
The recycled PSU numbers is a heavy legacy of SAS where you don't care as its survey procedures always treat PSUs as nested in strata. All other packages have to suffer.
I've updated the #multilevel models in https://agrogan1.github.io/multilevel-multilingual/. The results are much closer, but not yet equivalent, now that all software is estimating uncorrelated random effects, rather than some software estimating correlated random effects and other software estimating uncorrelated random effects. #rstats #Stata #julialang
#Stata question about #margins. Cross-posted at https://www.statalist.org/forums/forum/general-stata-discussion/general/1752205-margins-after-melogit-statistically-different-but-point-estimate-of-one-margin-inside-confidence-interval-of-another
As described at the above link, I am estimating `margins` (predicted probabilities) of a #multilevel #logistic regression. I am finding that in some cases the point estimate of one group is inside the confidence interval of the other group, but a test nonetheless indicates a statistically significant difference between the two groups. Any insights on this seeming paradox would be appreciated.
I’ve updated my graphing tutorial for #Stata, How To Choose A Chart - A Statistically Informed Visual Guide With #Stata Commands: https://agrogan1.github.io/Stata/how-to-choose-a-chart-a-visual-guide-with-Stata/how-to-choose-a-chart-a-visual-guide-with-Stata.pdf
