Econometrics - Binary Variables and Categorical Variables

This video will introduce the concept of binary variables (also known as dummy variables or indicator variables) and categorical variables. What does it mean when we include them in a regression? How can we interpret them, and what can we do with them?

[luizabpr]

Your videos are very clear, thank you! I have a question: i have a big dataset of 5, 000 companies indebtedness information. The data is in percentage (total liabilities/total assets), NOT right-bounded (in many cases i have indebtedness=0% or >100%). I would like to regress it against 2 sectors variables, but the distribution of indebtedness varies a lot: while industries look more or less like an uniform distribution, the services sector provides an-almost uniform distribution.

my_regression --> lm(indebteness ~ agriculture + industry + size, data=mydata)

Is it safe to use OLS? What can i read about that?

[papitasdelaperra]

Hello, really appreciate your videos. I have a question related to categorical models. If in an ologit model the proportional odds assumptions are violated, is there an alternative to removing the offending variables or shifting to an mlogit? Thanks!

[emptyxnes]

Great explanation. Thank you. I was wondering, how would it work, if one was to attach another variable, like sex? Male Married, Male single, etc. Would one combine, e.g Beta i *1(male) +Beta j *1(married) or create a new variable BetaK* Married Male?