Consider again the application in lecture 5.5, where we have analyzed response to a direct mailing using the following logit specification Pr[respi = 1] = exp(B0+ 1malei+ B 2 activei+ B3agei+ B4 (agei/10)) 1+exp (80+ 1malei+ B2 activei+ 83 agei+ B4 (agei/10)?) for i = 1, : : : , 925. The maximum likelihood estimates of the parameters are given by Variable Coefficient Std. Error t-value p-value Intercept -2.488 0.890 -2.796 0.005 Male 0.954 0.158 6.029 0.000 Active 0.914 0.185 4.945 0.000 Age 0.070 0.036 1.964 0.050 (Age/10)2 -0.069 0.034 -2.015 0.044 (a) The marginal effect of activity status is defined as Pr [ respi=1] = Pr[respi = 1] Pr[respi = 0]B2. d active i We could use this result to construct an activity status elasticity aPr [respi=1] active i d active i Pr respi=1] = Pr[respi = O]activeB2