1. Solution of Q&A4.6 says: the scaled deviance is scaled by parameter phi so that it can be used to compare different models. Can't we compare deviance residuals (non-scaled) or Pearson Residuals amongst different models? 2. Solution of Q&A4.7 says: 'The deviance residual considers the square root of each observation's contribution to the deviance, adjusted for the direction in which the raw residual acts. It corrects for the skewness of the distributions used, meaning that we would expect the deviance residuals to be more closely normally distributed than the raw residuals.' Can you please explain possibly by looking at the formula for deviance residuals, how it corrects for the skewness of the distribution used? Thanks!
Very good questions! The Core Reading is a little vague and doesn't go into much depth on this area, so don't worry too much as it's (hopefully) a little beyond what you need to know for the exam. Our gut feel is that it's possible that the Core Reading is omitting the word 'standardised' in some instances and not in others. However, we shall check with the Profession to try and get some more detail. Will post on here when we get a response. Thanks for raising it.
