15.7. A multination study of whether a country transitioned from autocracy to democracy during the study period7 reported the prediction equation logit[P(y = 1)] =

15.7. A multination study of whether a country transitioned from autocracy to democracy during the study period7 reported the prediction equation logit[ˆP(y = 1)] = −3.30 + 0.55t + 1.12(OECD)

−1.16m − 0.01 f − 0.07g, where y = 1 if the nation made that transition, t = number of past transitions, OECD is membership in the Organization for Economic Cooperation and Development (1 = yes, 0 = no), m = Muslim share of population, f =

fuel exports (percentage of merchandise exports), and g =

GDP growth (annual percentage).

(a) Interpret the coefficient of OECD.

(b) Interpret the coefficient of t.

(c) The estimated Muslim effect had se = 0.61 and Pvalue of 0.06. When the model was refit excluding countries regarded as the major oil-exporters, the estimated Muslim effect of −1.00 had se = 0.64 and P-value = 0.12.

The authors concluded that the statistical significance of

the Muslim effect might be explained by the major oilexporting countries tending to be high on m. Give an alternative explanation, involving the sample size.

TABLE 15.20 (Intercept) d V Estimate Std. Error z value Pr(>|z|) -2.0232 0.6137 -3.297 0.000978 1.1886 0.7236 1.643 0.100461 -1.5713 0.5028 -3.125 0.001778 Residual deviance: 0.16676 on 1 degrees of freedom