Question

What's the command for testing the significance of the treatment effect in STATA. (treatment: anykids; expansion: post 93)

Answer 1

STATA

There are some manual documents which commands on the use of data which are observational for estimating the effect which is caused by the treatment which one gets instead of another. The one who collects data does not control the assignment of the treatment under the observational data. So the treatment assignment and outcomes are affected by some common variables. The outcomes of the treatment are said to be conditionally independent when the treatment assignment and outcome variables are observable. When the treatment assignment and outcome variables are not observable then the outcome of the treatment are not conditionally independent.

The outcome of these treatments can be continuous, binary, count or nonnegative. Linear regression is used for continuous outcomes, logit is used for binary, poisson regression can be used for nonnegative and count outcomes.

Now, some sample subjects are observed among which some are there who have received the treatment and some are there who did not. Over here subjects means people and treatment means any kind of medical treatment or participation in any job training.

We would find that if there is any effect of the treatment on an outcome. The outcome can be ant person’s blood sugar level who takes his regular medicine or tries a new one. We would observe the outcome when the subject is treated and when the subject is not treated. After the observations the average of the difference between the treated and not treated will be found and this will show us the average impact of the treatment.