Commonly Used Statistical Tests
With all of the different analyses available, it might be confusing to know where to begin. The following table gives you an outline of some commonly used statistical tests for a specific situation.
Before using any statistical test, the procedure should be researched to assure that it meets the needs of the study. Some tests make distributional and other assumptions about the data that may or may not hold. It is important to understand the analysis, hypotheses, statistical tests, and conclusions that can be drawn before performing the analysis.
| What I want to do | Type of data
(Measurement Scales) |
Statistical analysis/method |
|---|---|---|
| Describe one group or set of data | Interval | Basic descriptive statistics |
| Ordinal | Nonparametric statistics (e.g. median, mode) | |
| Compare the mean of one group to a hypothesized population mean | Interval | Single sample t-test |
| Compare 2 independent groups | Interval | Independent t-test |
| Ordinal | Wald-Wolfowitz Runs Test
Kolmogorov Smirnov Test Mann-Whitney U Test |
|
| Compare 2 dependent groups | Interval | Dependent t-test |
| Ordinal | Sign test
Wilcoxon Matched Pairs test |
|
| Compare 3 or more independent groups | Interval | One-way ANOVA |
| Ordinal | Kruskal-Wallis ANOVA | |
| Compare 3 or more dependent groups | Interval | Repeated measures ANOVA |
| Ordinal | Friedman ANOVA
Kendall Concordance |
|
| Quantify relationship between 2 variables | Interval | Pearson product moment correlation |
| Ordinal | Spearman rank correlation
Gamma Kendall-tau |
|
| Predict the value of a numeric variable from a set of predictors | Interval | Linear regression
Decision trees Neural networks |
| Predict the nominal level of a categorical variable with 2 or more levels from a set of predictors | Nominal | Logistic regression
Discriminant analysis Decision trees Neural networks |