0

I'm running a linear mixed effects model using an lme4 function. I am then using a BIC function to look at whether adding more factors improves model fit. Now, the way I've done it so far was using the following: BIC(model1) to get the BIC for each model. But when I run an ANOVA to compare them, the output also gives me the BIC values, among other things, which are slightly different to the ones I get from the BIC function. Neither of them make a lot of sense by the way, because they increase instead of dropping as I add ew factors, but that's a separate issue. Should I be worried about those differences? I'm new to this kind of thing and still trying to understand both R and the mixed effects model.

Any help would be appreciated.

grizzthedj
  • 7,131
  • 16
  • 42
  • 62
Agata
  • 343
  • 3
  • 13
  • 1
    You're question is close to being migrated to stats.stachexchange because it is about statistics, not programming. You might want to add a little more context - are your models nested? Are you changing only the fixed effects, or also the random effects? – Gregor Thomas Apr 03 '18 at 18:37
  • In general, different AIC or BIC values from different functions isn't concerning, since they are both calculated up to an additive constant. You should expect the differences to be consistent across functions. – Gregor Thomas Apr 03 '18 at 18:38
  • 1
    When asking for help, you should include a simple [reproducible example](https://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example) with sample input and desired output that can be used to test and verify possible solutions. – MrFlick Apr 03 '18 at 19:13

0 Answers0