The lack of a conceptually and psychometrically strong measure of human mental workload has undermined interpretations of research based on cognitive load theory. Applying a confirmatory factor analysis approach to construct validation, a Subjective Mental Workload Survey was developed with four distinct subscales: Difficulty, Incompetence, Affect, and Effort. Participants were 52 English-speaking students in Australia and 12 Chinese-speaking students in Hong Kong. For these participants, there were 383 records of scores. Although Difficulty typically used in previous cognitive load research, reasonably well reflected cognitive load, a Workload measure incorporating Difficulty, Incompetence, and Effort proved to be a better predictor of cognitive load. This Workload factor is applicable to analysis of group differences using statistical procedures such as analysis of variance, a finding that shows that tasks more likely to overload working memory score higher in Workload estimates. An appendix contains a list of variables used in the study and alpha reliability estimates. (Contains 50 references.) (Author/SLD)