I am using R and the selection() function to run 2 seperate Heckman models.

1. Unadjusted model: For the sampling probability, the independent and dependent variable are binary, and for the second equation my independent variable is binary and my dependent variable is continuous.

                  Estimate Std. Error t value Pr(>|t|)
invMillsRatio       NA         NA      NA       NA
sigma               NA         NA      NA       NA
rho                 NA         NA      NA       NA

2. Adjusted model: When I run the same model but add binary and continuous independent variables to adjust for some things, I get :

                Estimate Std. Error t value            Pr(>|t|)    
invMillsRatio  506.807     29.125    17.4 <0.0000000000000002 ***
sigma          504.052         NA      NA                  NA    
rho              1.005         NA      NA                  NA    

Everything works fine, except that I get a lot of NA´s for the Error terms. Can anyone explain to me why this is the case? What does this mean, and are there any solutions to obtain error terms?

Thanks for any help!