$($a$)$ Suppose you have a continuous latent variable $Y^{**}$ that you want

to make a function of a constant, a continuous variable $x$ and a

dummy variable $D.$ However, in your data you only observe a binary

version of $Y^{**}$ where $Y=1$ when $Y^{**}>0$ and $Y=0$ otherwise. De$-$

velop a binary model $($for a non$-$specific distribution $.$