Perform the LRT to determine if the relationship between age and the logarithm of cancer rate significantly depends on the location. You will need to fit two models (one with and one without the interaction term) to obtain the residual deviances for the full and restricted models. Is your conclusion identical to that found in Question 30? Do you think the LRT is valid to use for the data? Why or why not?