Chi-square tests are hypothesis tests for categorical data that compare observed counts to expected counts. On the AP Stats exam, the test for homogeneity compares a variable's distribution across populations or treatments, while the test for independence checks whether two variables in one population are associated.
A chi-square test is what you reach for when your data come as counts in categories instead of means or single proportions. The big idea is simple. You figure out what counts you'd expect to see if the null hypothesis were true, then measure how far your observed counts stray from those expectations. Big gaps produce a big chi-square statistic and a small p-value, which is evidence against the null.
For two-way tables, the AP exam cares about two versions. The test for homogeneity asks whether the distribution of one categorical variable is the same across several populations or treatment groups (H₀: no difference in distributions). The test for independence asks whether two categorical variables measured on a single population are associated (H₀: the variables are independent). The math is identical for both. What changes is how the data were collected and how you word your hypotheses and conclusion, which is exactly what learning objectives 8.5.A and 8.5.B test.
Chi-square tests live in Unit 8 (Inference for Categorical Data: Chi-Square), and Topic 8.5 covers setting up the homogeneity and independence tests. Three learning objectives drive it. AP Stats 8.5.A asks you to write the correct null and alternative hypotheses in words (there's no parameter symbol here, which trips people up). AP Stats 8.5.B asks you to pick the right test based on study design, where one sample with two variables means independence and multiple samples or treatment groups means homogeneity. AP Stats 8.5.C asks you to verify conditions, including the right kind of random sampling, the 10% condition when sampling without replacement, and the large counts condition on expected frequencies. Unit 8 typically carries meaningful exam weight, and chi-square is one of the few procedures where naming the correct test is itself a graded skill.
Keep studying AP Statistics Unit 8
Two-Proportion z-Test (Unit 6)
A chi-square test for homogeneity on a 2x2 table is basically a two-sided two-proportion z-test in disguise. The chi-square statistic equals the z-statistic squared. Chi-square is the upgrade you need once a variable has more than two categories or you're comparing more than two groups.
Two-Way Tables and Conditional Distributions (Unit 2)
Back in Unit 2 you described association in two-way tables informally by comparing conditional distributions. Chi-square is the formal version of that same instinct. It puts a p-value on the question 'do these conditional distributions actually differ, or is this just sampling noise?'
Chi-Square Goodness of Fit (Unit 8)
Goodness of fit, covered earlier in Unit 8, uses the same chi-square machinery but on a one-way table, testing whether one variable's distribution matches a claimed model. The 8.5 tests extend the idea to two-way tables with two variables or multiple groups.
Sampling Methods and Experimental Design (Unit 3)
How the data were collected decides which test you run. One simple random sample with two variables recorded means independence. Stratified samples from separate populations or a randomized experiment with treatment groups means homogeneity. Unit 3 design vocabulary is the key that unlocks 8.5.B.
Multiple-choice questions love three angles. First, naming the right test from a study description, like deciding that randomly assigning students to traditional, flipped, or online teaching methods calls for homogeneity, while one random sample of 200 employees recording both education level and retirement plan preference calls for independence. Second, spotting condition violations, such as expected counts that fall below 5 or a sampling method that doesn't match the test. Third, writing or recognizing correct hypotheses, which must be stated in words about distributions or association, not with p's and parameter symbols. On the FRQ side, chi-square shows up in the inference free-response slot, where you earn points by naming the test, stating hypotheses, checking conditions (random, 10%, large expected counts), computing the statistic and p-value, and writing a conclusion in context that links the p-value to a decision about H₀.
Same formula, same degrees of freedom, same table. The difference is study design. Homogeneity applies when you take separate samples from multiple populations (or randomly assign treatments) and compare the distribution of one categorical variable across them. Independence applies when you take one sample from one population and record two categorical variables on each individual. Quick check, count the samples. Several samples means homogeneity. One sample, two variables means independence.
Chi-square tests compare observed counts to expected counts for categorical data, and large differences produce a big test statistic and a small p-value.
Use the test for homogeneity when comparing the distribution of one categorical variable across multiple populations or treatment groups, and the test for independence when checking whether two variables from a single sample are associated.
Chi-square hypotheses are written in words, not symbols, such as H₀: there is no association between the two variables, versus Hₐ: the variables are associated.
Conditions to verify are appropriate random sampling or random assignment, the 10% condition when sampling without replacement, and large counts, meaning all expected counts should be at least 5.
The number of samples tells you which test to run, since several independent samples or treatment groups means homogeneity while one sample with two variables means independence.
A chi-square test on a 2x2 table gives the same conclusion as a two-sided two-proportion z-test, because the chi-square statistic is the z-statistic squared.
It's a hypothesis test for categorical data that compares observed counts to the counts you'd expect under the null hypothesis. Topic 8.5 covers the two-way table versions, the test for homogeneity and the test for independence.
Count the samples. If you have separate random samples from multiple populations or randomly assigned treatment groups, use homogeneity. If you have one random sample and recorded two categorical variables on each individual, use independence.
No. Per learning objective 8.5.A, chi-square hypotheses are stated in words, like 'H₀: there is no association between exercise frequency and sleep quality.' There's no single parameter to write a symbol for.
Expected counts. Each expected count in the table should be at least 5. Observed counts can be small, and a question describing an expected count below 5 is signaling a condition violation.
Not by itself. A significant result only tells you the distributions differ or the variables are associated somewhere in the table. To say where, you'd compare observed and expected counts or look at which cells contribute most to the chi-square statistic.
Connect this key term to the AP exam workflow: review the course, practice questions, and check related study tools.
Review units, study guides, and course resources.
Check this vocabulary in multiple-choice context.
Apply key concepts in written AP responses.
Estimate the exam score you are working toward.
Review the highest-yield facts before practice.
Put the full course together before test day.