1. in REGRESSION ANALYSIS, test of the statistical significance of a regression coefficient. It involves basically two steps: (1) compute the T-VALUE of the regression coefficient as follows: t-value = coefficient/standard error of the coefficient; (2) compare the value with the t table value. High t-values enhance confidence in the value of the coefficient as a predictor. Low values (as a rule of thumb, under 2.0) are indications of low reliability of the coefficient as a predictor.

2. general statistical test for hypotheses, based on t-distribution, known as a small sample distribution. The t-test is used to estimate and test hypotheses about population means, the difference between two means, a population variance, and a comparison of two popu-lation variances. For example, an accounting instructor wishes to test to determine if the use of a new and old textbook had anything to do with the difference in performance of the two classes.