The main resampled-GLMM approach revealed that the
population-only model was the most frequently top-ranked,
although substantially more variance was explained by including
the number of high-protection reserves (Table 1). However, the
model-averaged, standardized coefficients demonstrated that the
number of high-protection reserves was one of the two most
important predictors of increasing health of reserves (Fig. 3). The
gross national income of the country (partialled by the inclusion
of population so that it was comparable among countries) was
the only other supported predictor; as it increased, reserve health
declined (Fig. 3). The total number of protected areas by itself did
not influence average reserve health (Fig. 3). We also verified the
gross national income and corruption contributions by including
longer-term (GNI: 1990-2013; World Governance Indicators