The following log-odds ratio model is calculated (Green 2003):(1)where: β is matrix of parameters in the logit modelX is matrix of explanatory variables in the logit modelIn order to apply a linear form, the function above can be written as follows: + + (2)where: P is the vector of probabilities of behavior option, is the ith individual and ε is the error termFor the case of onion farmers above, the empirical binomial logit model is presented as:According to Green (2003), the marginal effect refers to the change in probability given a unit change in the independent variable, holding other variables constant. Marginal effect is computed as follows:
đang được dịch, vui lòng đợi..