I need help to understand whether or not my interpretation of these coefficients are correct. I am investigating the effect of being exposed to more female peers on STEM enrollment in college. In this regression I also include dummies for whether the individual has chosen high-level physics, chemistry, natural science or social science (base) in high school and interact these dummies with an indicator of being female.
Y = B1*Female*prop. female peers + B2*Male*prop. female peers + C1*Female*Physics + C2*Female*Chemistry + C3*Female*Natural Science + D1*Physics + D2*Chemistry + D3*Natural Science + E1*Female + School FE + Cohort FE + Cohort controls
Is it correct that:
- E1 is the coefficient on being female and having high-level social science relative to being male and having high-level social science when the prop. of female peers is 0? (which is not informative to me)
- B1 is the coefficient of being exposed to more female peers when being female on STEM enrollment? Or is B1 the coefficient of being exposed to more female peers when being female and having high-level social science?
I am also interested in seeing the gender effects of having the different high-level subjects in high school on STEM enrollment in college.
- The coefficient on being female and having high-level physics in high school when the prop. of female peers is not 0 relative to being male and having high-level social science is equal to B1+C1+D1?
I am getting confused because of the female being interacted with both prop. of female peers as well as the different groups of subjects..
Thank you in advance!
Dependent variable: STEM enrollment | |
Female * Proportion female peers | -0.089*** |
(0.018) | |
Male * Proportion female peers | -0.006 |
(0.019) | |
Female*Physics | -0.123*** |
(0.014) | |
Female*Chemistry | -0.026* |
(0.014) | |
Female*Natural Science | 0.042*** |
(0.013) | |
Physics | 0.142*** |
(0.012) | |
Chemistry | -0.056*** |
(0.012) | |
Natural Science | -0.220*** |
(0.011) | |
Female | -0.033** |
(0.016) | |
Observations | 87,363 |
Mean dependent variable women | 0.166 |
Mean dependent variable men | 0.315 |
School fixed effects | YES |
Cohort fixed effects | YES |
Cohort controls | YES |
0 Response to Interpretation of multiple interaction terms
Post a Comment