Hi,I am doing reserach on behavior analysis, and doing statistical analysis
to see the effects of factors across different neighborhoods. For that I
need to conduct multilevel tobit regression analysis on my (censored)
behavior data. Is there any R package that can support multilevel tobit
analysis? I googled a lot, but could not find any (only found that stata 15
supports this). But I prefer using R and would very much appreciate it if sb
can provide any info. Thanks!
You can do this in Stan by integrating out censored values. Alternatively, the R package censReg handles panel data.
Hi, Chuck, thanks for your info. But censReg only supports tobit, and does not support multilevel. What I am looking for is a R package that can do multilevel tobit analysis.
Here is the section of the Stan manual on censoring. I mentioned integrating them out, but see here that you can also treat them like missing data and estimate censored values. I hope this is helpful.
hi, Chuck, many thanks for your suggestions. I will check it out.
This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.