One line per respondent connecting their two paired values, shared by
t_test_pair and wilcoxon_pair (both carry vars = c(x_column, y_column) on the same respondents). The standard visual for a paired
design: it shows the direction and consistency of individual change,
which a plain bar-of-means would hide.
Usage
sframe_plot_paired_comparison(result, data, palette = c("web", "print"))