Hello Stephen,
Currently I am running some simulations comparing the results for marginal effects and treatment effects using the LPM when the true model is a probit and when the true model is a logit. The results will probably be the subject of the next blog.