I am a beginner in the field of econometrics/regression analysis and would like to ask about the interpretation of the coefficients when using a dummy IV.
I want to evaluate the effect of a variable called CIVISM on the likelihood of joining the public sector (binary variable) in a bilingual region (German and French speakers).
Since I suspect CIVISM to be endogenous, I try to alleviate this issue by instrumenting the variable. As a result, I use a binary variable called French (whether the individual is a French speaker) as an instrument. I already did a 1st stage estimation and the IV looks correct so far.
My model:
One copy of the results table (in the good format hopefully):
I have diffculty in interpreting the coefficient of CIVISM with respect to the IV. More precisely, I am trying to understand the meaning of the coefficient according to the the value of French (0 or 1).
I know that probit model does not allow to interpect direclty the coefficients, but my question is more related to the IV, and the difference of interpretation with a regression with no instrument.
Thank you.
I want to evaluate the effect of a variable called CIVISM on the likelihood of joining the public sector (binary variable) in a bilingual region (German and French speakers).
Since I suspect CIVISM to be endogenous, I try to alleviate this issue by instrumenting the variable. As a result, I use a binary variable called French (whether the individual is a French speaker) as an instrument. I already did a 1st stage estimation and the IV looks correct so far.
My model:
Code:
ivprobit PUBLIC i.kanton (CIVISM=French) $CONTROLS i.year, vce(robust)
| Robust | ||||||
| Coef. | Std. Err. | z | P>z | [95% Conf. | Interval] | |
| CIVISM | -.1078692 | .0117544 | -9.18 | 0.000 | -.1309074 | -.0848311 |
| kanton | ||||||
| FR | .1004076 | .096608 | 1.04 | 0.299 | -.0889407 | .2897558 |
| VS | -.1114069 | .0639421 | -1.74 | 0.081 | -.2367312 | .0139173 |
| age | .1056905 | .0138313 | 7.64 | 0.000 | .0785816 | .1327993 |
| age_sq | -.0008263 | .0001456 | -5.68 | 0.000 | -.0011116 | -.0005411 |
| female | .3303467 | .0547671 | 6.03 | 0.000 | .2230051 | .4376884 |
| kid | .0601874 | .0221381 | 2.72 | 0.007 | .0167975 | .1035773 |
| female_kid | -.0520341 | .0281653 | -1.85 | 0.065 | -.1072372 | .003169 |
| college | .2894405 | .0517888 | 5.59 | 0.000 | .1879363 | .3909446 |
| confess | ||||||
| Catholic | .1535563 | .0500758 | 3.07 | 0.002 | .0554096 | .2517031 |
| Other | .0923529 | .0463577 | 1.99 | 0.046 | .0014934 | .1832124 |
| ideology | -.0214935 | .0106815 | -2.01 | 0.044 | -.0424288 | -.0005582 |
| union | .4001181 | .0656321 | 6.10 | 0.000 | .2714816 | .5287546 |
| charity | -.0944223 | .0456408 | -2.07 | 0.039 | -.1838765 | -.004968 |
| environment_protection | -.0457087 | .0439215 | -1.04 | 0.298 | -.1317932 | .0403759 |
I know that probit model does not allow to interpect direclty the coefficients, but my question is more related to the IV, and the difference of interpretation with a regression with no instrument.
Thank you.
