Logit and Probit Table

[Sarah Rossi]

I want to estimate a regression (shown below on the table) and fill in the data points of the table but I have been stuck on it and have no idea how. The table refers to predicted probabilities for Situation1 and Situation2. These are predicted probabilities of a child having a college education or higher in two hypothetical families, namely:

• Situation1: male = 1, kidsnumber = 3, motherseducation = 12, and fatherseducation = 12
• Situation2: male = 1, kidsnumber = 3, motherseducation = 16, and fatherseducation = 12

I made a binary variable that says child_degree, which equals 1 if the child received a college degree or higher, and equals zero otherwise like so:
My other variables are:
tall = 1 if the child is tall and = 0 if child is short
motherseducation : number of years of mothers education
fatherseducation: number of years of mothers education
kidsnumber: number of kids in the familu
male: =1 if child is male

How can i make this table and get the probit, logit, and ols models? im so very confused

[Leonardo Guizzetti]

This is a pretty complex table for a beginner. Is this a school or work assignment?

You may be better off computing each component and manually inputting them into a formatted table in Word.

[Sarah Rossi]

Yes it is that's why I'm totally lost

so you mean doing probit, logit, and ols regression separately right? how can i do the logit regression given male = 1, kidsnumber = 3, motherseducation = 12, and fatherseducation = 12? i typed it just as it, but I keep getting error

i put:
logit child_degree male = 1 kidsnumber = 3 motherseducation = 12 & fatherseducation = 12

[Leonardo Guizzetti]

so you mean doing probit, logit, and ols regression separately right?

That's right. Those coefficients would be plugged into the top half the table.

logit child_degree male = 1 kidsnumber = 3 motherseducation = 12 & fatherseducation = 12

Conditions can be added using the -if- qualifier. Note that operators are described in -help operator-. The following will run a logistic regression model for the outcome of -child_degree- for the specific subset described in the if condition.

logit child_degree if male == 1 & kidsnumber == 3 & motherseducation == 12 & fatherseducation == 12

However your instructor/supervisor probably doesn't want you to do this. They probably want you to use the -margins- command to get the respective probabilities for each situation. They can provide further guidance, and Richard Williams has a very nice guide to using -margins- here.

I'm totally lost

I feel for a situation when you are asked to do something that is both complicated to do and also to use unfamiliar tools. I strongly recommend that you spend some time getting more familiar with how to use Stata in general. The manuals have a very useful getting started guide. Start at Chapter 1 and choose the guide that matches your operating system. The time spent will be well worth it.

help getting started

Later, when you are more comfortable with Stata and reading documentation, you can start to learn user-contributed commands the Ben Jann's -estout- package or the built in suite to make Word files (putdocx), PDF files (putpdf) and Excel files (putexcel), etc.

[Sarah Rossi]

Thank you so much! It's difficult because in class they don't teach us the STATA codes and it's not in our book. So it's always a challenge to answer the STATA questions. This is very much appreciated!

[Leonardo Guizzetti]

You're welcome. Your teacher should be able to point you toward resources for Stata, even if they don't cover it in class.