Here is a quick example that the uncorrected glm cannot fit. Suppose that freq contains the nine values from a 3x3 contingency table, fac1 and fac2 the factor variables indexing the margins, and lin1 and lin2 the corresponding linear variables. Then without my correction, glm cannot fit glm(freq~fac1+fac2+fac1:lin2+fac2:lin1,family=poisson) because there are 11 variable combinations (3 aliased) and only 9 observations. Jim =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html Send "info", "help", or "[un]subscribe" (in the "body", not the subject !) To: r-devel-request@stat.math.ethz.ch =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=