Lackfit not showing output information for logistic regression.

[Tariq519]

Good morning, I am currently creating an ordinal logistical model for my job and I am trying to diagnose the model to ensue that it fits the data. I am using the "lackfit" statement right after the model statement, however every time I run the code, the usual output is displayed, however no Hosmer and Lemeshow analysis is never displayed. I am currently using SAS 9.3 through SAS Enterprise 7.1. Here is my code below. 

proc logistic data=upper_limit;
class Profitrank dlr_zip RUCA2 area_description/ param= ref;
model profitrank = num_booked--percentage_of_new average_vehicle_age--average_truist_rate RUCA2 state_median_income--state_pop State_pop*RUCA2/lackfit rsq;
output out=new p=ppred;
run;

Would greatly appreciate any help!

[PaigeMiller]

For lack of fit to be computed, you have to have at least two observations with identical values across all x-variables. With so many x-variables, I think it is unlikely that you would have such observations. Have you checked your data set to see if you do have at least two observations with identical values?

[Tariq519]

I highly doubt that two of my observations have identical values. Is that really the only way you can utilize lackfit? Because I know the pearson dispersion test wouldn't work for this either since I have a lot of predictor variables a nd a relatively large data set.

[PaigeMiller]

---

@Tariq519 wrote:
I highly doubt that two of my observations have identical values

---

Me too.

Is that really the only way you can utilize lackfit?

Yes, that's what it needs to work.

[Ksharp]

You will not get "lackfit" a.k.a Goodness Of Fitness Test , if there were more than 2 levels in Y variable (profitrank ) .