Cross-tabulation Analysis
Cross-tabulation Analysis (Contingency Table Analysis) · Also known as: crosstab, contingency table analysis, two-way frequency table, bivariate frequency analysis
Cross-tabulation analysis (contingency table analysis) is a foundational descriptive and inferential technique for examining the relationship between two or more categorical variables. It arranges observed frequencies into a table of rows and columns, enabling visual inspection of patterns and formal chi-square testing of independence between the variables.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
Use cross-tabulation whenever you want to describe or test the association between two categorical variables on the same cases — for example, comparing treatment versus control on a categorical outcome, examining whether two survey items co-vary, or checking demographic differences in categorical responses. Both variables must be categorical (nominal or ordinal); at least 80 percent of cells should have expected counts of 5 or more. Do not use cross-tabulation for continuous outcomes (use t-tests or ANOVA), for more than two variables simultaneously without log-linear or logistic-regression modeling, or as a substitute for measures of agreement when the same raters rate the same objects.
Strengths & limitations
- Straightforward to compute, interpret, and communicate to non-statistical audiences.
- Handles any number of categories in each variable without distributional assumptions on the data.
- Provides both a descriptive view (observed and percentage tables) and a formal significance test.
- Pairs naturally with effect-size measures (phi, Cramer's V, odds ratio) that go beyond the p-value.
- Works with small or large samples; Fisher's exact test is available as a fallback for sparse cells.
- Assumes independent observations; cannot handle repeated measures or matched pairs without modification.
- Chi-square approximation is unreliable when many cells have expected counts below 5.
- Detects association but does not establish direction of causation or control for confounders.
- With many categories or variables the table becomes large and hard to interpret without further analysis.
- Collapsing ordinal categories into fewer levels to meet cell-count requirements can discard information.
Frequently asked
When should I use Fisher's exact test instead of the chi-square test?
Use Fisher's exact test when the sample is small and one or more cells have an expected count below 5. Fisher's test calculates the exact probability of the observed table and all more extreme tables, without relying on the chi-square approximation. For larger tables with sparse cells, consider collapsing categories or using Monte Carlo simulation of the p-value.
How do I measure the strength of the association?
For 2 x 2 tables, the phi coefficient ranges from 0 to 1 and is directly comparable to a correlation. For larger tables, use Cramer's V, which adjusts phi for table dimensions and also ranges from 0 to 1. Values near 0 indicate weak association; values above 0.3 are often considered moderate and above 0.5 strong, though context matters.
Does a significant chi-square tell me which cells drive the association?
Not directly. Inspect the adjusted standardized residuals (the difference between observed and expected counts divided by its standard error) for each cell. Residuals with absolute value above 2 flag cells that contribute most to the significant chi-square.
Can I use cross-tabulation with ordinal variables?
Yes, but standard chi-square ignores order. When both variables are ordinal, the Mantel-Haenszel chi-square test for linear trend or gamma and Kendall's tau-b are more appropriate because they exploit the ordering and have more statistical power to detect monotone associations.
Is cross-tabulation the same as frequency analysis?
Frequency analysis typically tabulates one variable at a time (a one-way frequency table). Cross-tabulation extends this to two or more variables simultaneously, producing a two-way (or higher-order) table that reveals how categories of different variables co-occur.
Sources
- Pearson, K. (1900). On the criterion that a given system of deviations from the probable in the case of a correlated system of variables is such that it can be reasonably supposed to have arisen from random sampling. Philosophical Magazine, 50(302), 157–175. DOI: 10.1080/14786440009463897 ↗
- Agresti, A. (2002). Categorical Data Analysis (2nd ed.). Wiley-Interscience. ISBN: 978-0471360933
How to cite this page
ScholarGate. (2026, June 3). Cross-tabulation Analysis (Contingency Table Analysis). ScholarGate. https://scholargate.app/en/statistics/cross-tabulation-analysis
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- Chi-square testStatistics↔ compare
- Cramer's VStatistics↔ compare
- Fisher's exact testStatistics↔ compare
- Frequency analysisStatistics↔ compare
- Logistic RegressionResearch Statistics↔ compare