Hi!
In logit models, I am interested in estimating the slope of a tangent line. According to Hanushek and Jackson (1977, p.189), the interaction can be estimated directly by finding ∂2P/∂Xk∂Xj=bjbkP(1-P)(1-2P). If P is evaluated at .05, the interaction will be zero due to (1-2P)=(1-2x.5). Thus, there will be no interaction between Xk and Xj when evaluated at P=.5. Is that correct?
Can Stata estimate the interaction between the derivatives of multiple Xs? Can you provide stata codes? Thank you.
Hanushek, E. A., & Jackson, J. E. (1977). Statistical methods for social scientists. NY, NY: Academic Press, Inc.
Related Posts with RE: interaction of marginal effects in logit models
Subscribe to:
Post Comments (Atom)
0 Response to RE: interaction of marginal effects in logit models
Post a Comment