I have a two-period panel. The dependent variable is number of corrective claims. There are four plan types (comp, hmo_sim, ppo, cdhp) with each but one represented by an indicator variable. After using the xtpqml command, the estimates for the independent variables of interest are:
Calculating Robust Standard Errors...
-------------------------------------------------------------------------------
nr_cor_clms_I | Coef. Std. Err. z P>|z| [95% Conf. Interval]
--------------+----------------------------------------------------------------
nr_cor_clms_I |
cdhp | -.4218268 .2474486 -1.70 0.088 -.9068171 .0631635
comp | .1411787 .0883146 1.60 0.110 -.0319149 .3142722
hmo_sim | .0856342 .1342522 0.64 0.524 -.1774954 .3487637
It seems that there are estimated 42% fewer claims for those moving to cdhp in year two, but only relative to those moving to ppos (the reference group), though weakly significant. Is my interpretation correct, or is there a better way to interpret or to express it?
Calculating Robust Standard Errors...
-------------------------------------------------------------------------------
nr_cor_clms_I | Coef. Std. Err. z P>|z| [95% Conf. Interval]
--------------+----------------------------------------------------------------
nr_cor_clms_I |
cdhp | -.4218268 .2474486 -1.70 0.088 -.9068171 .0631635
comp | .1411787 .0883146 1.60 0.110 -.0319149 .3142722
hmo_sim | .0856342 .1342522 0.64 0.524 -.1774954 .3487637
It seems that there are estimated 42% fewer claims for those moving to cdhp in year two, but only relative to those moving to ppos (the reference group), though weakly significant. Is my interpretation correct, or is there a better way to interpret or to express it?
