Dear all users of Statalist,

I'm currently working on the my master thesis with the topic of farmers' use of personal protective equipment (PPE) during pesticide applications in Ghana. I want to estimate the determinants of the PPE use using standard probit model. I want to find out what kind of potential problems probit model has and solutions to each issue in this post

The models has the outcome variable is a binary variable to identify if a farmer used PPE or not with a set of independent variables associated with individual characteristics(age, education, size of land etc), village attributes(infrastructure, facilities in a village), and the binary variable to identify if a village chief used PPE or not, whose marginal effect is a main interest in my thesis.

As far as I know, there are potential problems such as
(i) Endogeneity of independent variables
(ii) Heteroskedasticity and Normality in the latent variable model

The questions are below:
(1) How can you test heteroskedasticity besides hetprobit and the suggestion by Wooldridge(2007) at Section 15.5.3?
(2) How would you justify the approach by Wooldridge(2007) to test heteroskedasticity? (This is unrelated with Stata topic, though)

Thanks in advance,

Reference:
Jeffrey M Wooldridge, 2010. "Econometric Analysis of Cross Section and Panel Data," MIT Press Books, The MIT Press, edition 2, volume 1