Statistical inference and spatial patterns in correlates of IQ

摘要: Cross-national comparisons of IQ have become common since the release a large dataset international scores. However, these studies consistently failed to consider potential lack independence scores based on spatial proximity. To demonstrate importance this omission, we present re-evaluation several hypotheses put forward explain variation in mean among nations namely: (i) distance from central Africa, (ii) temperature, (iii) parasites, (iv) nutrition, (v) education, and (vi) GDP. We quantify strength autocorrelation (SAC) predictors, response variables residuals multiple regression models explaining national IQ. outline procedure for control SAC such analyses highlight differences results before after SAC. find that incorporating additional terms interdependence increases fit with no loss parsimony. Support is provided finding index parasite burden are strongly linked temperature also features models. tentatively recommend physiological – via impacts host–parasite interactions rather than evolutionary explanation effect temperature. study primarily danger ignoring spatially extended data, an appropriate approach should explicit analysis be considered necessary.