The data
on social mobility of men in the UK was assembled by Blane et al.
(1999) in JRSS-A. A sample of men aged 45-64 was drawn from the 1971 census and
1981 census and the social class of the man was recorded at each timepoint. The
classes are I=professional, II=semi-professional, IIIN=skilled non-manual, IIIM=skilled
manual, IV=semi-skilled, V=unskilled. Check for
symmetry, quasi-symmetry, marginal homgeneity and quasi-independence.
Question 2.
The dataset contains data on murder cases in Florida in 1977. The data is cross-classified by the race (black or white) of the victim, of the defendant and whether the death penalty was given.
Consider the frequency with wich the death penalty is applied to black and white defendants both marginally and conditionally with respect to the race of the victim. Is this an example of Simpson's paradox? Are the observed differences in the frequency of application of the death penalty statistically significant?
Determine the most appropriate dependence model between the variables.
Fit a binomial regression with death penalty as the response and show the relationship to your model found in b).