I calculated the dispersion parameters for each of two blocks in a hierarchical logistic regression. I am not quite sure how to interpret the dispersion parameter.
First I calculated the parameters by taking the chi-square goodness-of-fit statistic (hosmer-lemeshow-test) and dividing it by the degrees of freedom. For both blocks the values were less than 1.
In the second step (to support the prior calculation) I took the 2-log-likelihood value and divided it by the degrees of freedom. In this case the parameters for both blocks were above 10.
So is the first or the second calculation right? What can I do if there overdispersion?