Censored Regression

[Olive Bat]

Could anyone please specify the command for censored regression models in STATA.

[Olive Bat]

Carlo Lazzaro Jeff Wooldridge .. could you please provide me with details on censored regression and the command on STATA for it for both cross sectional and panel data.

[Carlo Lazzaro]

Olive:
you may want to take a look at -help tobit- and -help cnsreg-.

[Jeff Wooldridge]

Based on your other posts I’m guessing you don’t really have censored data. You have a corner at zero. Censoring means you don’t observe the variable you want. But expenditures can be zero. They aren’t censored at zero, even though this is common language.

Do you really have data censoring?

[Olive Bat]

thanks to Carlo Lazzaro

[Olive Bat]

Jeff Wooldridge .. my data is not count data .. so i cant use Poisson..my dependent variable is expenditure on different items such as health,education and other consumer durable in the last 30 days.... so not applicable for Poisson.. and as its panel.. so for cross sectional i can use tobit bit for panel data what do ?

[Jeff Wooldridge]

-tobit- is suited more to having a corner solution outcome. -cnreg- is for the case where you have true data censoring.

[Jeff Wooldridge]

Olive: It was explained to you previously that Poisson regression applies to any nonnegative outcome, including a variable such as expenditures. Forget the name Poisson! You only require the mean to be correct. You can try tobit, too, as I showed in a previous post. But the Poisson FE estimator is a very good way to go.

Has someone now told you that you "have censored data, and need to use censored regression"? If so, they're wrong. Please don't keep asking for the same advice in new threads when you have been given detailed advice, even with Stata commands, about what to do!

[Olive Bat]

@ Jeff Wooldridge.. thank you and okay