Dear Allstaters,
Paired data in a 2x2 table can be analysed using
McNemar's test, but has anyone some advice where
the data is bivariate (two binary responses per
pair). I suppose since there is a connection
between McNemar's and logistic regression, a bivariate
logistic regression (e.g. Gumpel's fit) might do the
trick, but I would be grateful for any and all
relevant advice and experiences.
Many Thanks,
Tim O'Brien
Loyola University Chicago Math & Statistics
www.math.luc.edu/~tobrien
ps., please respond directly to me and I shall post a
summary.
|