ABSTRACT
All statistical significance tests yield a statistic (number); a t test yields a t stat, an F test an F stat, a chi-square test a chi-square stat, etc. That stat is referred to as the test value, and with the exception of one type of test, it can range from 1 to 3, based on the three standard
standard deviation
Probability theory (always thumbing its nose!) says that our chances of finding significant difference or association are zero (nil, none, nada). So, the farther away from zero the test value, the more likely that any significance noted by the test’s math is real/true/accurate. Although a test value (except for one type of test) might range from 1 to 3, the point at which it begins to reflect significance, known as its critical value, is 1.96. So a test value must be 1.96 or greater for you to claim that a difference or association noted by that test is, in fact, significant. In this case, then, bigger is better!
