One of the critical issues that face the accounting profession today involves choosing among alternative accounting information modes. This dissertation provides comparative empirical evidence on the predictive power of accrual-based accounting signals versus cash-flow accounting signals versus both of these signals jointly. The empirical hypotheses compare the degrees of association between the market evaluative criteria, market beta and security price behavior, and the different accounting signals. The research methodology employed includes the following. 1. Market beta and changes in security prices are used as the evaluative criteria. 2. Two regression models are developed and used to test the predictive power of the alternative accounting signals. 3. Several specifications for each model are used. These specifications are simple regression, multiple regression, interaction effect, partial correlation, incremental correlation, and time series and cross sectional analysis.