I echo the comment section. Thank you Rohen for deciding to create this video. My professor is trash when it comes to how to communicate this subject to the students. Eternal blessings.
This is 100 times better than what our econometrics lecturer did, he was reading the slides to us the whole time while explains nothing. You are a champion!
One question about three-way interaction terms. Let's label each variable A(main variable), B(1st moderator), C (2nd moderator). I'm interested in (hypothesize) the relationships A-B and A-B-C. Should all two-way (AB, AC, BC) and three-way interaction terms (A * B * C) be included in a regression model and result or would be it fine to include some of interest (AB, ABC) only?
Thank for sharing. Quick point. Why to bother with any interaction therm ?. Simply, fit two lines. One for female (y_f=b0_f+b1_fX_f) and one for male (y_m=b0_m+b1_mX_m). Compute difference in tangents = b1_m - b0_m. The result should be the same without any quadratic term. I am just interested why to introduce an additional term. Cheers Piotr
Thank you for your great question! You're absolutely right that it is "equivalent" to doing two simple regressions. That said, in the real world of research there are usually many variables to consider simultaneously, and the number of equations can go up exponentially in that case. If you want to see for example how the impact varies simultaneously by a combination of one's race, gender, age, you would need 100+ equations depending on how many race and age groups there are. And it's much easier to have one big regression rather than going back and forth between 100+ equations, especially when it comes to reporting outputs. Hope that helps!