EViews.com

Hi,
can someone tell me briefly and simply how to estimate in practise a tobit model, please?
is it a problem if there some dummy variables among the regressors?
is it applicable to a dependent variable having values between 0 and 100?
Should I delete all the observations that have a missing value for the dependent variable?

I hope someone can help me..

Many Thanks,
Danida

Hi,
can someone tell me briefly and simply how to estimate in practise a tobit model, please?
is it a problem if there some dummy variables among the regressors?
is it applicable to a dependent variable having values between 0 and 100?
Should I delete all the observations that have a missing value for the dependent variable?

I hope someone can help me..

Many Thanks,
Danida

Use the tobit command with the dependent variable followed by the explanatory variable.
Having a dummy variable on the right is fine. In principle, observations that are not censored are conditionally normal so there should not be an upper limit. If you have many observations at the upper limit, you may want both an upper and lower censoring point--which EViews will do for you.

EViews will automatically drop observations with missing values.

Thanks for this.

When I use 3 dummy toghether it gives me a message of "near singular matrix"...what can I do to solve this problem?
Which option is it good to use to improve the prob?

Thanks,
Danida

You've probably fallen into the dummy variable trap with a complete set of dummies and a constant.

You've probably fallen into the dummy variable trap with a complete set of dummies and a constant.

there are non-categorical variables..
and there are any options i can tick to improve the significance of the coefficients?

Thanks,
Danida

You've probably fallen into the dummy variable trap with a complete set of dummies and a constant.

there are non-categorical variables..
and there are any options i can tick to improve the significance of the coefficients?

Thanks,
Danida

No. You probably need a better model or better data.