Statistical methods employed to test individual differences in dual-task performance and the existence of a general time-sharing ability are reviewed and critiqued. Specifically, both the types of data being collected and the types of data analysis procedures have been inadequate to the critical evaluation of a hypothetical "time-sharing" ability. Problems resulting from unsophisticated use of correlational and factor analytic procedures are discussed. The application of such procedures to simulated data from a model with a hypothetical time-sharing factor accounting for 25 percent of the variance of scores in dual-tasks resulted in inappropriate rejection of the time-sharing factor. Recent psychometric findings and techniques, especially in the area of exploratory and confirmatory uses of factor analysis, are shown to have ramifications for research and analysis. Techniques are presented for convergent validational procedures appropriate to the time-sharing ability issue. The crucial nature of task selection, scoring methods, and considerations of practice and reliability are discussed. Based on a reanalysis of available data, the "time-sharing" or "A" factor is not rejected, but the lack of critical evidence prohibits the confirmation of such a factor. Simulation and incorporation of theory in planning models and crucial tests of the hypothetical "time-sharing" ability are discussed for use in further research. (Author/CM)