Differences between 'anova' and 'Anova'?

Just a general question on the suitability of anova vs Anova. I read through ? Anova and ? anova but i wasn't able to understand it to my satisfaction.

I am running some GLMs and I would like to compare between models as well as within models. When do I 开发者_开发百科use Anova (data) and when should I use anova (data)

I am also a newbie but here are my thoughts. I never had any issues with anova. I always use it for comparing 2 or more models. Write from smaller model to larger model anova(smal,large,..) and if the larger one is significant it means adding those variables to small and having large is meaningful. You can say larger one is better in a sense. But not always this works especially when you have multicollinearity. If that is the case you should eliminate some multicollienar variables for a better model even if anova says you shouldn't. Hope this helps.