One commonly used outcome measure in multiple sclerosis (MS) clinical trials is the Multiple Sclerosis Functional Composite, which includes the Paced Auditory Serial Addition Test (PASAT) as a measure of cognitive function. Concerns have been raised about the standard PASAT scoring method, whereby the number of correct responses is summed. This method does not take into account whether the test is performed as intended, which may affect interpretation of the results. Accordingly, another scoring method has been proposed, which examines the number of times a correct response is immediately preceded by another correct response (termed a dyad). We compared the two scoring methods for the PASAT, and found that the mean percentage of correct responses not accounted for by dyads ranged from 27.5% to 49.5%, indicating that much of the time the test is not performed as instructed. We also examined disease course and the PASAT score, as studies have produced conflicting results as to whether disease course is associated with cognitive impairment. Although disease course was significantly associated with the PASAT score, it accounted for little of the variation in scores, even when adjusting for other predictors. Finally, as 14.2% of participants refused to do the PASAT or failed to complete it, we also examined whether the Perceived Deficits Questionnaire (PDQ), a self-reported measure of cognitive function, is a potential proxy measure for the PASAT. The correlation between the two tools was low (-0.14), suggesting that the PDQ is not a useful substitute for the PASAT.