Consider a model for whether a person has ever smoked.

Consider a model for whether a person has ever smoked. Let P=Pr(Y = 1) be the

probability that a person has ever smoked and assume the logit model is given by

log[P/(1-P)] = α(1-X) + βX

where 1-P is Pr(Y = 0) (never smoked) and X is a dummy variable that is 1 if the person

is female and zero if male.

(i) If your prior expectations are that males are more likely to smoke than females

and that both males and females have a less than 50% chance of ever smoking,

what would you expect about the values of α and β?