DataStep Statistical Selection Guide
The DataStep Statistical Selection Guide provides a basic guide for selecting data analysis models and choosing statistical tests based on the type of data you'll be using. The guide is presented in two sections: measuring association and measuring differences. Each type of measure is based on the characteristics of the primary dependent variables.
Note: This selection guide is intended as an introduction to choosing statistical tests and measures. More complex data sets will require more extended or more sophisticated measures. For our purposes, we have selected only the most common statistical procedures.
Measures of association include the Pearson product moment correlation, Spearman's rho or Kendall's tau-b, partial or multiiple correlations, multiple regressions, and non-parametric tests. The measure used here assumes a single dependent variable which is either (a) a frequency or count or (b) a measure. For measures, the next step is to determine whether you are using two dependent variables or more than two dependent variables. The next step in selecting a measure is determined by whether the dependent variables are discrete or continuous.
Measures of difference include tests such as analysis of variance (ANOVA), analysis of covariance (ANCOVA), multivariate analysis of variance (MANOVA), and multivariate analysis of co-variance (MANCOVA). As with measures of association, the selection of the tests is determined by the number of dependent variables, and the combination of characteristics of the independent variables (such as all cattegorical or discrete, some categorical and some continuous), and the number of values in the independent variables.
To download the DataStep Statistical Selection Guide, click here.
If you have suggestions as to how the statistical selection guide might be improved, please forward them to firstname.lastname@example.org.