Hi everyone, I have estimated different models with the betareg() command from the package 'betareg' (3.0-4). When I started to compare them using likelihood ratio tests, it occured to me that the logLik() of the models increased with increasing number of parameters. I confirmed this observations by repeating the MockJurors example from the betareg vignette. Now my question is, why does the loglikelihood becomes bigger with more explanatory variables included? (Applying logLik() to the standard glm() command yields a negative loglikelihood which moves toward zero with each additional predictor included.) Thanks for hints, Chris -- View this message in context: http://r.789695.n4.nabble.com/logLik-in-betareg-tp4690885.html Sent from the R help mailing list archive at Nabble.com.