I am running pooled regression for 67 group my dependent variable is gini disposable and independent variable is women empowerment (economic,social,political) and institutional quality ( government stability,democratic accountability,law & order) When I run regression without interaction term women empowerment and institutional quality signs are negative But when I add interaction women empowerment*institutional quality then sign of separate variable become positive and and interaction term got negative sign what it explains is it normal?
Asked
Active
Viewed 31 times
0
-
make a 3D scatterplot, with axis 1 = women's empowermnet, axis 2 = institutional quality, and axis 3 = gini disposal, to see what's going on. – John Madden Jun 18 '23 at 14:52
-
2See this page and this page, the other "related" pages listed on this page, and their many links. When a predictor is involved in an interaction, its coefficient represents its association with outcome when its interacting coefficients are at 0 or their reference levels. You can only interpret such a single-predictor coefficient in terms of the levels of the interacting predictor and the interaction coefficient(s). – EdM Jun 19 '23 at 14:13