I am using the mle2 function to run a simple linear model with a normal distribution. I have one continuous variable (X) and one factor (treatment) with 6 levels. I have seen ANCOVA examples similar to this, but my goal is to compare the differences in the slope parameters (the interaction term) among groups, not overall group means or anything like that. I want the intercept to be the same for all groups and it seems that I do not even need to estimate a parameter for X, the continuous variable, but I am having trouble understanding the different ways to specify a model like this and why all of these give me different estimates. Any insight or ideas? mod1 <- mle2 (Y ~ dnorm (mean = B0 + B1* X, sd = sigma), parameters = list (B1 ~ X : treatment), start = as.list (coef(mod)), data=data) mod2 <- mle2 (Y ~ dnorm (mean = B0 + B1 * X, sd = sigma), parameters = list (B1 ~ X : treatment - 1), start = as.list (coef(mod)), data=data) mod3 <- mle2 (Y ~ dnorm (mean = B0 + B1, sd = sigma), parameters = list (B1 ~ X : treatment), start = as.list (coef(mod)), data=data) mod4 <- mle2 (Y ~ dnorm (mean = B0 + B1, sd = sigma), parameters = list (B1 ~ X : treatment - 1), start = as.list (coef(mod)), data=data) Thanks, Andrea -- View this message in context: http://r.789695.n4.nabble.com/Categorical-variables-using-mle2-and-slope-comparisions-tp4648140.html Sent from the R help mailing list archive at Nabble.com.