Hello,
I am attempting to do a multiple regression analysis, but my dataset contains multiple missing values. The values were recoded as ".", but I still cannot get the number of observations to match between the two models and, therefore, cannot run a LRT test. Most of the variables are categorical and I am doing forward stepwise regression. What am I doing wrong??? Thank so much for any help!!!
Example:
1st model: logistic prevcat i.CountryCat if MWQTempCat !=. & DayNightCode !=. & AgeRangeCat !=. & AMPrecCat !=. & AMTempCat ! =.
estimates store a
2nd model: logistic prevcat i.CountryCat i.YrSurveyCat if MWQTempCat !=. & DayNightCode !=. & AgeRangeCat !=. & AMPrecCat !=. & AMTempCat ! =.
estimates store b
lrtest a b
I am attempting to do a multiple regression analysis, but my dataset contains multiple missing values. The values were recoded as ".", but I still cannot get the number of observations to match between the two models and, therefore, cannot run a LRT test. Most of the variables are categorical and I am doing forward stepwise regression. What am I doing wrong??? Thank so much for any help!!!
Example:
1st model: logistic prevcat i.CountryCat if MWQTempCat !=. & DayNightCode !=. & AgeRangeCat !=. & AMPrecCat !=. & AMTempCat ! =.
estimates store a
2nd model: logistic prevcat i.CountryCat i.YrSurveyCat if MWQTempCat !=. & DayNightCode !=. & AgeRangeCat !=. & AMPrecCat !=. & AMTempCat ! =.
estimates store b
lrtest a b