Fig. 5.

Illustration of over-interpreting null results. The scalp projections show an unthresholded z-statistic map of correlations between local scalp curvature and lifetime number of sexual partners, which has been overenthusiastically annotated with interpreted effects (i.e. the “frontal horn”). The results might be compared with those of the infamous “dead salmon” study, which highlighted the importance of correcting for multiple comparisons (Bennett, Baird, Miller, & Wolford, 2010). Please note that when thresholds for multiple comparisons were applied, none of the z-scores in this figure reached statistical significance. Also damning is the fact that the “frontal horn” area does not correspond to regions of interest predicted by 19th century phrenologists. The upper-right panel depicts a prediction for “Amativeness” on the opposite side of the skull (Fowler & Fowler, 1859).