I'm a fairly inexperienced statistician fighting a huge deadline and just need some peace of mind that I'm not making a massive error here. I'd be most grateful for any pointers.

I've been working on fitting an lmer() model that explores the spatial relationships between crime and socio-economic factors. The ICC demonstrates a 40% attribution of variances to group factors, which is encouraging. The model's evolution has seen a steady improvement of AIC scores. I am now a little confused about how to demonstrate its effectiveness against other non-HLM techniques. Is this purely by standard error residuals, or am I missing something important?!

Many thanks

  • $\begingroup$ My comment about your other question also goes here. Moreover, it's not clear if you should have 2 questions, or if these are ultimately one question. $\endgroup$ Sep 5, 2012 at 2:51
  • 1
    $\begingroup$ Hi @gung, I see your point but I think they're separate :) The other is about the risk of mixed-modelling with multiple correlated variables. This is about how to compare an HLM model with other analytical techniques like OLS regression. I am exploring the relative benefits of the approach, so need an idea of how to compare different models. I'm entirely self-taught and have found it pretty tough getting into the language, so just need some guidance. $\endgroup$
    – geotheory
    Sep 5, 2012 at 9:38


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.