Hi, What do we mean when we refer to relativity value of level of factor? It is the same as parameter (i.e. B)? Also, what do we call Xi i.e. explanatory variable which can be made up of levels of factors or interaction terms? I can't call them covariate/factor right?
These are discussed on page 27/28 of the GLM chapter, and just give the relativities of the different levels (eg male and female) of a particular rating factor (eg gender). Beta is something different - beta is the thing that the GLM is going to find for you, and tells you the importance of each factor (eg gender) on the response variable (eg frequency). In this case, gender is our explanatory variable (aka covariate). Age might be another covariate, for example (ie another X). If we find, for example, that old males have a particularly 'extra high' frequency, over and above the individual effects of being old and male, then we might decide to introduce an interaction term ("age-gender") as another covariate. We might also find that in a particular country, old people tend to be male whereas young people tend to be female - this would be 'correlation', which is something entirely different to an interaction.
