Cross-tabulation
Also called crosstab, contingency table
A table that breaks down answers to one question by answers to another.
A cross-tab puts one question down the rows and another across the columns, with counts in the cells: satisfaction by region, recommendation by tenure, choice by experiment arm. It is the workhorse of survey analysis because most survey questions are categories, and categories are compared by counting.
Read the percentages in the direction of your question. “What share of each region is satisfied” and “what share of satisfied people are in each region” are different tables from the same counts, and mixing them up is how a finding gets reversed in a summary.
A chi-square test on the table says whether the pattern is bigger than chance. It needs reasonable expected counts in every cell, so slicing a small survey by four variables at once produces tables that no test can say anything useful about.
Where it goes wrong
Percentages from small groups are unstable. Show the underlying counts and avoid treating a few responses as a precise population estimate.