Selecting Statistical Analyses

 Variable(s) Graphical Displays Numerical Summaries Inferential Statistics one qualitative (categorical) variable bar chart pie chart raw and relative frequencies mode one-sample z test about a proportion one-sample z confidence interval for a proportion chi-square goodness-of-fit test one quantitative (numerical) variable boxplot histogram frequency polygon stem-and-leaf display dotplot ogive five-number summary mean median mode range interquartile range standard deviation one-sample t test about a mean one-sample t confidence interval for a mean repeated measures of one quantitative (numerical) variable on the same group boxplot(s) histogram(s) frequency polygon(s) stem-and-leaf display(s) dotplot(s) ogive(s) line graph (if the repeated measures are made over time) five-number summary of differences mean difference median difference range of differences interquartile range of differences standard deviation of differences paired t test about a mean difference paired t confidence interval for a mean difference repeated measures ANOVA (with more than one repeated measurement) repeated measures of one ordinal-qualitative (categorical with a natural ordering on the categories) variable on the same group bar charts raw or relative frequencies Wilcoxon signed rank test   Friedman nonparametric ANOVA (with more than one repeated measurement) comparison of one nominal-qualitative (categorical with no natural ordering on the categories) variable for two or more groups or relationship between two nominal-qualitative (categorical with no natural ordering on the categories) variables stacked bar chart multiple bar charts multiple pie charts raw or relative frequencies in a contingency table two-sample z test about a difference between proportions two-sample z confidence interval for a difference between proportions chi-square test about independence comparison of one ordinal-qualitative (categorical with a natural ordering on the categories) variable for two or more groups or relationship between one nominal-qualitative (categorical with no natural ordering on the categories) variable and one ordinal-qualitative (categorical with a natural ordering on the categories) variable stacked bar chart multiple bar charts multiple pie charts raw or relative frequencies in a contingency table Mann Whitney rank sum test (with two samples)   Kruskal-Wallis test (with more than two samples) relationship between two ordinal-qualitative (categorical with a natural ordering on the categories) variables stacked bar chart multiple bar charts multiple pie charts Spearman rank correlation Spearman rank correlation test relationship between one ordinal-qualitative (categorical with a natural ordering on the categories) variable and one quantitative (numerical) variable   or   comparison of one quantitative (numerical) variable for two or more ordered groups multiple boxplots multiple histograms multiple frequency polygons multiple stem-and-leaf displays multiple dotplots multiple ogives Spearman rank correlation   five-number summaries means medians ranges interquartile ranges standard deviations Spearman rank correlation test comparison of one quantitative (numerical) variable for two or more groups or relationship between one qualitative (categorical) variable and one quantitative (numerical) variable multiple boxplots multiple histograms multiple frequency polygons multiple stem-and-leaf displays multiple dotplots multiple ogives five-number summaries means medians ranges interquartile ranges standard deviations two-sample t test about a difference between means two-sample t confidence interval for a difference between means one-way analysis of variance (with more than two samples) relationship between two quantitative (numerical) variables scatterplot line graph (if one variable is time) least squares line Pearson correlation simple linear regression t test (or Pearson correlation t test)