The main purpose of this study was to identify a pool of discriminating variables to be used as predictors of success on the SPEAK and TEACH tests, which are tests given to prospective international teaching assistants at Iowa State University. Data were collected from four-hundred randomly selected subjects. Two sets of variables were individually investigated: background variables (sex, first language, TOEFL score, last degree completed, number of months residing in the U.S., average amount of time speaking English, and previous teaching experience) and operational variables (pronunciation, fluency, cultural ability, communication skills, interaction skills, raters' overall impression, and student-questioners' overall impression), which are diagnostic scores obtained on the SPEAK and TEACH tests. A model was designed with all variables combined and the predictive power of these variables was tested by stepwise multiple regression technique. Results revealed that the best predictors for SPEAK were pronunciation (r =.92), followed by fluency (r =.89), raters' overall impression (r =.56), and first language (r =.68). The best predictors for TEACH were raters' overall impression (r =.79), followed by pronunciation (r =.66), culture (r =.67), and fluency (r =.64). Tables, figures, and plots illustrate these findings. Implications of the findings were discussed, and recommendations for further research and for practice were also included.