IQ, the worst performance rule, and Spearman's law: a reanalysis and extension

Abstract

This paper focused on two hypotheses: the worst performance rule, which states that worst performance on a cognitive task predicts g-loaded measures such as IQ better than best performance, and Spearman's law, which states that the g-loading of IQ decreases with aptitude and with age during childhood. Combining these two hypotheses, the present study tested a new hypothesis: the worst performance rule should hold for younger and lower-IQ children, whose IQs should be highly g-loaded, but perhaps not for older and higher-IQ children, whose IQs should be less g-loaded. The new hypothesis was tested by reanalyzing archival data and extending Coyle's [Intelligence 29 (2001) 117] worst performance analysis. Nongifted ( n=81, mean IQ=108.94) and gifted ( n=85, mean IQ=140.19) children in second through fourth grade received five study-recall trials of words. IQ was correlated with worst performance (fewest words recalled across trials), best performance (most words recalled across trials), and strategy use. As predicted, only the younger nongifted children showed evidence of the worst performance rule (i.e., worst performance was the best predictor of IQ). However, gifted–nongifted difference tests of the IQ-worst performance correlation were not significant at any grade level, although the magnitude of the differences varied in the predicted direction (i.e., gifted–nongifted differences were most pronounced at the lower grades and declined with age). The results were interpreted in the context of the differentiation hypothesis, which states that mental ability becomes more differentiated and less g-loaded with increases in age and IQ.