#### momo

##### New Member
Hi, all

I am working on a bivariate probit model and I nearly got the results, but I have no idea to check about collin.... and hetro... problems.

I tries to run a linear regression just to use vif command on stata and I got the
next results where its low. but can anyone help me to explain the results and what is the limit to determine the numbers.

Variable | VIF 1/VIF
-------------+----------------------
Costsofemp~l | 1.57 0.637600
logemploye~r | 1.50 0.665066
naceav | 1.40 0.713640
inddummy | 1.31 0.764253
affiliate | 1.05 0.947891
Intangible~v | 1.04 0.961732
logage | 1.04 0.965437
auto | 1.02 0.980988
-------------+----------------------
Mean VIF | 1.24

so what do you think

#### Dragan

##### Super Moderator
Hi, all

I am working on a bivariate probit model and I nearly got the results, but I have no idea to check about collin.... and hetro... problems.

I tries to run a linear regression just to use vif command on stata and I got the
next results where its low. but can anyone help me to explain the results and what is the limit to determine the numbers.

Variable | VIF 1/VIF
-------------+----------------------
Costsofemp~l | 1.57 0.637600
logemploye~r | 1.50 0.665066
naceav | 1.40 0.713640
inddummy | 1.31 0.764253
affiliate | 1.05 0.947891
Intangible~v | 1.04 0.961732
logage | 1.04 0.965437
auto | 1.02 0.980988
-------------+----------------------
Mean VIF | 1.24

so what do you think
You're fine i.e. proceed....Your Variance Inflation Factors are all okay....Way less than 5.

#### momo

##### New Member
Thank you Dragan

but what about the check of the heteroscedasticity, how it could be done on stata for the bivariate model.

another think to do the VIF should I do it after OLS regression, there isn't another way depending on the biprobit results.

Thanks again
momo