Comparing Categorical Displays

A bar chart of raw counts tells you how many people fall into each category, but if two groups have different total sizes, taller bars just mean more people, not a stronger preference. Imagine City has 200 students and Rural has 150. Even if City has more students who pick 'chips' as their favorite snack, that could simply be because City has more students overall. The fix is to convert each group's counts into percentages of that group's own total, called conditional percentages, so each group is judged against its own size.

A side-by-side bar chart puts two groups' bars next to each other for every category, letting you eyeball differences category by category. A segmented bar chart stacks each group's categories into a single bar that sums to 100%, which is better for seeing the full distribution shape and comparing proportions at a glance. To compare city and rural snack choices fairly, find each category's percentage within each group, then build one segmented bar per group. If chips take up 40% of the City bar and 25% of the Rural bar, that is a real difference in preference, independent of how many students are in each group.

This sets up the rest of Unit 1: once you can trust a comparison between groups, you can start asking whether an observed difference is large enough to matter, or whether it is just how the data happened to fall. Segmented bar charts and two-way tables are the same idea in different clothes, and both depend on this same instinct: always compare rates, never raw totals, when group sizes differ.