Stata - dependent binary variable

Original post by AJStonee

I need to do a multiple linear regression in Stata. My dependent variable is a dummy variable. We are new to Stata, and need graphs to do explanatory statistics. When we create graphs, it obviously only appears as two solid lines of data points. How do we interpret, or is there an additional command that will draw a graph which we can more clearly analyse?

Not used stata, but youd really need to be a lot more specific about what youre trying to do (can you post the original question), what the data is like and what youve done. So generally you assume a linear relationship so why is it binary, what do you mean by a dummy dependent variable, what sort of exploratory stats are you trying to do, what lines are you talking about ....

Reply 2

OP

Original post by mqb2766

Not used stata, but youd really need to be a lot more specific about what youre trying to do (can you post the original question), what the data is like and what youve done. So generally you assume a linear relationship so why is it binary, what do you mean by a dummy dependent variable, what sort of exploratory stats are you trying to do, what lines are you talking about ....

The dependent variable (y) is a dummy variable and hence binary (as it takes the form 1 or 0).
Thank you for your response, however I am looking for advice specific to stata

Reply 3

Original post by AJStonee

The dependent variable (y) is a dummy variable and hence binary (as it takes the form 1 or 0).
Thank you for your response, however I am looking for advice specific to stata

Trying to model a binary variable using multiple linear regression is problematic at best. Are you really trying to do this, or just looking at summary stats of the independent variables?

Reply 4

OP

Original post by mqb2766

Trying to model a binary variable using multiple linear regression is problematic at best. Are you really trying to do this, or just looking at summary stats of the independent variables?

We have to run a regression. I would share the question, however we just have a data set that we have to use to form our own question. We think we may have to use assumptions to model expected y as rhe probability that y=1, but are unsure of how to really progress after that

Reply 5

Id agree that youre really trying to model the probability if the dependent variable is binary. Basic multiple linear regression does not do this. It would help to see the question as well as you describing what youve covered on your (uni?) courses if you want some appropriate advice.

Reply 6

OP

Original post by mqb2766

Id agree that youre really trying to model the probability if the dependent variable is binary. Basic multiple linear regression does not do this. It would help to see the question as well as you describing what youve covered on your (uni?) courses if you want some appropriate advice.

We have no question, as we have to form our own analysis, however i am studying econometrics. We have covered slightly beyond basic MLR alongside the Woolridge textbook.
Thanks so much for taking the time to respond by the way! :smile:

Reply 7

Original post by AJStonee

We have no question, as we have to form our own analysis, however i am studying econometrics. We have covered slightly beyond basic MLR alongside the Woolridge textbook.
Thanks so much for taking the time to respond by the way! :smile:

Im about to turn in for the night, but Im really not sure what your question is, what textbook youre referring to, what youve covered, what the data set is, ....

There are generally a number of ways of approaching the analysis and without some context about the previous questions its hard to really give useful advice. If you want some advice, pls provide a fair bit more info about the previous questions or have a chat with your tutor/lecturer who will know all this.

Reply 8

Original post by AJStonee

We have no question, as we have to form our own analysis, however i am studying econometrics. We have covered slightly beyond basic MLR alongside the Woolridge textbook.
Thanks so much for taking the time to respond by the way! :smile:

I did a quick google before turning in and I presume youre talking about one of the data sets in chapter 7 in
http://fmwww.bc.edu/gstat/examples/wooldridge/wooldridge.html
There are some summary stats for each of the examples/data sets.

As above though, you really need to be more specific to get some help or if youre having trouble getting started, ask your lecturer/tutor.

Reply 9

OP