Dear All, I am trying to use the packadge polr () to analyse ordinal categorical data responses. Instead of using polr() directly, I have altered the script slightly (imposed a constraint to make all the parameters greater than or equal to zero) (see below), fit <- list(coefficients = exp(beta), zeta = zeta, deviance = deviance,........................) However my main problem is that when I use this "new script" I am unable to get the t-values and p-values (though able to get the parameter values and the deviance). I have downloaded and scrutinised the polr() script in an attempt to understand the section where these t- and p- values are estimated but without much success. Where is polr () are these p- and t-values estimated? Thanks Simo [[alternative HTML version deleted]]