ABSTRACT
All statistical significance tests yield a statistic, referred to its test value. Selecting a proper test of statistical significance includes choosing: one of difference or association; one that is appropriate for the levels of measure (LOM) of the data in question; the version that is appropriate to either independent or correlated sample groups; and either a one-tailed or two-tailed test version, depending on nature of hypothesis. The chi-square test of statistical significance generally tests independent sample differences between two or more groups on nominal LOM data. Phi coefficient tests for association when each variable has only two categories. A correlation coefficient (cc) can range from weak to strong and be either positive or negative. A degrees of freedom (df) stat is generally presented as a real number and usually, the larger the sample the smaller it is. Multivariate testing refers to tests of statistical significance that examine several groups around several variables and several variables around several groups.
