
26 Ordered Regression Models
Second, we will present AMEs in order to compare the effects of each independent vari-
able on the predicted probabilities. For binary variables, the marginal effect is the effect of
a discrete change in x
k
from 0 to 1 on the predicted probability. For continuous variables,
the marginal effect is the partial change in the predicted probability, which is the instanta-
neous rate of change or slope of the tangent line to the probability curve with respect to x
k
while holding the remaining variables constant (Powers and Xie 2008, p. 60). It essentially
represents a hypothetical, linear effect of a one-unit increase i