The two parameters of the Hick paradigm, the intercept (a) and the slo
pe (b), of reaction time (RT) as a function of the information load sc
aled in bits (i.e., the binary logarithm of the number of stimulus-res
ponse alternatives) differ in their (negative)correlation with IQ, a g
enerally having a larger correlation than b. The typically low and oft
en nonsignificant correlation between b and IQ appears to contradict t
he theory that rate of information processing is a component of genera
l mental ability (g) as approximated by IQ or other highly g-loaded te
sts. The a of the Hick function largely reflects individual difference
s in the sensory-motor lags in task performance, while the b supposedl
y reflects individual differences solely in the rate of information pr
ocessing. Hence b theoretically should be more highly correlated with
g or its proxy, IQ, than is a. But in fact, the opposite is commonly f
ound. The weakness of the b x IQ correlation, as compared with the cor
relation between IQ and a (and with other variables derived from the H
ick paradigm) is mainly attributable to statistical artifacts that sup
press the b parameter's correlation with any other variables, e.g., a
and IQ. When the b x IQ correlation is estimated under conditions that
reduce the statistical suppression of this relationship, the correlat
ion is appreciably increased and is consistent with prediction from in
formation processing theory.