bounds error

[Cody Olsen]

Hi, and thank you for the help and this google group.

I am curious about an error that I am getting, in case someone knows what it is, and if it is a big deal. The error is in the log and listing, but does not prevent the program from running and producing an imputed dataset.

I am using IVEware and SAS 9.4, a dataset with 200-ish rows and 200+ variables. 

I am getting an error when imputing a couple variables in particular. For example, one is the number of people in a household with gastrointestinal symptoms (Day5HouseholdSymp). The bounds for this variable are >=0 and <= total number of people in the household.

I get an error in the log:

Impute Day5HouseholdSymp

  Update data

  Poisson regression

  Correct data

Bounds error for PrimaryCareVisit: min=0, max=0

PrimaryCareVisit is whether the subject saw their doctor, and is a categorical (yes/no) variable. So I am assuming the bounds error is really for the Day5householdSymp variable.

I get the same error listing other related variables, like:

...Bounds error for Transmission: min=0, max=0

One other variable is giving me similar error messages. 

I think it could have something to do with co-linearity. Maybe coefficient estimates are really large/small because of co-linearity? These variables have some strong and redundant predictors. I can make the errors go away by setting a small maxpred, but i don't want to do that if the error is not going to harm anything. The final values of the imputed variables are within the bounds, so it looks OK to me.

Any ideas about what is causing the error, and if it should be fixed by modifying the maxpred, or candidate predictors?

thanks!

-Cody

[Brandon Hopkins]

Bounds error for PrimaryCareVisit: min=0, max=0

I've seen that error. For me it was the variable in the bounds statement. Like in the .set file  if you had: bounds  PrimaryCareVisit(>x, <y) ;

and you may somehow y is not being calculated correctly for at least one of the observations.