PS2 exercise 2

[jgdenby]

would you like us to return the standard errors for the quasi-coefficients in addition to those for the beta parameters? Or is a plot sufficient? If so, what's the best way to compute and report those? I see in your example code how you compute standard errors for one dichotomous and one continuous variable, but i'm not 100% sure how to translate that to two continuous variables. Thanks!

[bensoltoff]

Look at the first two marginal effects plots for interaction terms with continuous variables. This is a good template for how to report the standard errors. There's not much use in a table since there would be 20 sets of standard errors (one for each value of education in the data set). The marginal effects plot for the Titanic logistic regression example also shows an appropriate graphical form for a continuous variable interacted with another continuous variable.