The Composite Scale of Morningness (CSM) is a widely used measure of behavioral temporal preference, and it is highly reliable across cultures. There are several competing models concerning its factor structure. We used confirmatory factor analysis (CFA) to replicate the two-and three-factor models, and, in addition, we examined the utility of the four-item morning affect scale as a proxy for the full CSM. We tested these models in a large student sample (N ¼ 1396). The chi-square result for the two- and three-factor models indicated they did not adequately fit the data, and the modification indices suggested some items could be correlated. Allowing these items to correlate failed to produce a non-significant chi-square result, but some improvements to the incremental fit indices and root mean square error of approximation (RMSEA) were found. In particular, the three-factor model seemed satisfactory. CFA of the morning-affect scale also failed to fit the data, but the incremental fit indices and RMSEA were strong. For each of these three models, there was no significant difference between the invariant (no sex difference) and variant (sex difference) models. However, in each model, a critical ratio difference was found on the standardized regression weights between males and females on item 12. The morning-affect scale had high reliability (0.83) and demonstrated satisfactory construct validity against the preferences scale and time to sleep and wake. In addition, high and low morning-affect groups reported significant sleepiness differences by time of day. These results suggest the morning-affect scale may be a useful proxy for the full scale. Furthermore, the items make no reference to time-of-day and may better facilitate cross-cultural research. Studies using population-based samples are recommended to further test the efficacy of the morning-affect scale.