Skip to content
Prev 17893 / 20628 Next

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

Dear James,

Since all ID's are unique, you can rely on the implicit nesting of random
effects. (1|STATE_ID/COUNTY_ID) is equivalent to (1|STATE_ID) +
(1|COUNTY_ID) when each COUNTY_ID is unique (not reused among states).

There is a huge overlap between PLACE_ID and full_district_id. The model
will have a hard time separating both effects. Especially since you have
the same random slope.

You should also center the year variable. Now your model is estimating the
random intercept at year 0.

So in terms of basic model I would

- drop either PLACE_ID or full_district_id
- center year
- add year to the fixed effects or add it as a separate random intercept

Furthermore, I'm not sure it the Poisson distribution is the most relevant
distribution for your response. I've seen a case were the Poisson
distribution failed. The response was an area in hectares rounded to an
integer value. The model fitted well with a Gamma distribution.

Best regards,

Thierry

ir. Thierry Onkelinx
Statisticus / Statistician

Vlaamse Overheid / Government of Flanders
INSTITUUT VOOR NATUUR- EN BOSONDERZOEK / RESEARCH INSTITUTE FOR NATURE AND
FOREST
Team Biometrie & Kwaliteitszorg / Team Biometrics & Quality Assurance
thierry.onkelinx at inbo.be
Havenlaan 88 bus 73, 1000 Brussel
www.inbo.be

///////////////////////////////////////////////////////////////////////////////////////////
To call in the statistician after the experiment is done may be no more
than asking him to perform a post-mortem examination: he may be able to say
what the experiment died of. ~ Sir Ronald Aylmer Fisher
The plural of anecdote is not data. ~ Roger Brinner
The combination of some data and an aching desire for an answer does not
ensure that a reasonable answer can be extracted from a given body of data.
~ John Tukey
///////////////////////////////////////////////////////////////////////////////////////////

<https://www.inbo.be>


Op zo 8 sep. 2019 om 07:40 schreef Ades, James <jades at ucsd.edu>: