no. par. with individual covariates

[Claudio]

To be sure (or to be wrong): when I fit models with Mark, using individual
covariates, the number of parameters in the model output is different
compared the number of parameters estimated. Actually, Mark counts the
number of link function parameters. Have I to change the number of
parameters using the number of estimated parameters, haven't I? E.g., when I fit the model phi(.)p(.), I get two estimated parameters. When I constrain phi with 2 covariates *phi(.)_cov1cov2_p(.)*, Mark reports 4 parameters and uses this number for AIC calculation. Have I adjust the no. par. of the model constrained to 2?
I think it does make sense getting 4 parameters, because actually the model with covariates is more parameterized than the one without covariates, but it is not clear to me what I have to do. E.g., in the chapter 12 of "Gentle ntroduction", no models with individual covariates seems to be adjusted for the no. par. (for say, in my example I should keep 4 instead of 2 parameters).
Thanks in advance for answers to this naive question.

[jlaake]

A model with an individual covariate will have a slope and intercept (2 parameters). If you used a covariate for Phi and for p that is 4 parameters. When you fit a constant model, it is the intercept only. Thus, 2 parameters, one for Phi and one for p. The parameter counts are correct so there is no need to adjust.

--jeff

[Claudio]

Thank you Jeff.