This study tested four theoretical models in terms of their fit with demands placed on our cognitive system by traditional tests of cognitive ability. We did so by administering seven tests of cognitive ability known to require varying types of processing demands to a large group of college undergraduates (N=193). We compared the models using confirmatory factor analyses, including those based upon a unitary factor, speed and capacity, crystallized and fluid intelligence, and verbal and spatial ability. The crystallized/fluid model provided the best fit with the data. This finding is consistent with previous research. Implications for education and future research are discussed.