Skip to content
Prev 17948 / 20628 Next

Large mixed & crossed-effect model looking at educational spending on crime rates with error messages

I'm having trouble reading this file - it's 50MB to start with, my
Linux system thinks it's gzipped ("gzip compressed data"), then when I
un-gzip it I get a 100+MB file that it thinks is "data" (which just
means "binary file I don't understand").  Is there any way you can
make a **much smaller** reproducible example for us to inspect??

  I think the answer to your proximal question about per_race is that
you would need five *different* numerical variables, one for each
race-percentage. However, you can't/shouldn't actually use all five,
since they will be multicollinear (since the sum of all five will be 1
or 100 (or however you've scaled the total).

  Also: I wouldn't recommend Nelder-Mead, it's probably the slowest
and least reliable option ...
On Fri, Sep 27, 2019 at 7:05 PM Ades, James <jades at ucsd.edu> wrote: