Decision Trees for Fundamental Statistics To Select a Parametric Statistic to Evaluate Group Differences Frequency Distribution Table Present entire distribution Table or graph? z-score Test Is σ known? Bar Graph Scale of measurement? Single-Sample t-Test Histogram or Polygon Skewed, open-ended, or undetermined scores? Scale of measurement? Central tendency Ordinal How many scores for each subject? Two Related-Samples t-Test Mean Median Repeated-Measures ANOVA Mode Related-Samples t-Test Summarize entire distribution Standard How many separate samples? Two Matched or independent samples? Deviation What measure of central tendency? Variability Median Independent-Measures tTest Semi-interquartile Range or SemiInterquartile range or Range Range Repeated-Measures ANOVA Nominal scale? None Matched or independent samples? z-Score Single-Factor ANOVA How many independent variables? Describe individual score Two-Factor ANOVA Percentile Ranks Chi-Square Goodness of Fit How many categories of measurement? Pearson Correlation Measure linear or monotonic relationship? Both variables measured on interval or ratio scale? How many dichotomous variables? How many scores for each subject? Spearman Correlation Sign Test Both variables measured on ordinal scales? Data consist of 2 scores for each individual Binomial Test If possible, rank order both variables and use Spearman. Otherwise, convert to frequency data and use chi-square test for independence. Can difference scores be ranked? How many separate samples? Two Yes Matched or independent samples? Point-Biserial Correlation Wilcoxon Test Mann-Whitney Can scores be ranked with few tied values? No Chi-Square Test for Independence Median Test Phi-Coefficient Chi-Square Test for Independence Data consist of frequency values for each category Chi-Square Test for Independence Independent samples only Are scores discrete or continuous? Median Test To Select a Statistic to Measure the Relationship Between Two Variables Download a copy of this poster at the department website: http://uca.edu/psychology/concept-maps/ To Select a Nonparametric Statistic to Evaluate Group Differences