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?
So when do we actually interact our variables? Is there a way to see if it is necessary or do we just do it and then see if the coefficient on the interaction term is statistically significant?
How does our interpretation of the B1 and B2 coefficients change now that you've included the interaction effect? In the machines example, would we say that B1(workers) and B2 (machines) each have effects independent of the interaction term B3?
It depends on whether the coefficients are significant. For example, If your model scores showed that B2 and B1*B2 were significant than you could conclude that there is a joint effect of B1 and B2 as well as an independent effect of B2 being captured.
Sir, For a logit regression, if the odds ratio is coming out to be less than 1 for the main independent variable then its favorable in my case. However, the odds ratio of the interaction term is coming out to be more than 1 which (according to me means) that people who use more ICT tools and are educated are making more mistakes. Its a mistake variable and 0 means "no mistake", and 1 means "mistake". Please correct me if I'm wrong and put some light on how to interpret the interaction term.
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?
Very nice explanation
Thanks, would appreciate that you get more to the point, but still good. regards
So when do we actually interact our variables? Is there a way to see if it is necessary or do we just do it and then see if the coefficient on the interaction term is statistically significant?
How does our interpretation of the B1 and B2 coefficients change now that you've included the interaction effect? In the machines example, would we say that B1(workers) and B2 (machines) each have effects independent of the interaction term B3?
It depends on whether the coefficients are significant. For example, If your model scores showed that B2 and B1*B2 were significant than you could conclude that there is a joint effect of B1 and B2 as well as an independent effect of B2 being captured.
Sir,
For a logit regression, if the odds ratio is coming out to be less than 1 for the main independent variable then its favorable in my case. However, the odds ratio of the interaction term is coming out to be more than 1 which (according to me means) that people who use more ICT tools and are educated are making more mistakes. Its a mistake variable and 0 means "no mistake", and 1 means "mistake".
Please correct me if I'm wrong and put some light on how to interpret the interaction term.
How i can create intrection effect when value is lying positive negative and zero values
It should be "with IE: B1+B3*workers", cuz you're deriving the impact of machines
Why don't we check for multicollinearity?
Three variables i understood but what if there are more than a dozen independent variable? which to pick?
Prasun Bhattacharjee same Problem here. Found a good Solution?
@@mehmet3979 no
Thanks