Chi-Square Statistic in AP Statistics

In AP Statistics, the chi-square statistic (χ²) measures how far observed counts in a two-way table fall from the counts you'd expect if the null hypothesis were true, computed as χ² = Σ(Observed − Expected)²/Expected. Bigger values mean stronger evidence against the null.

What is Chi-Square Statistic?

The chi-square statistic is a single number that sums up how badly your data disagree with the null hypothesis. In a two-way table, the null says either the variables are independent (test for independence) or the distributions are the same across populations (test for homogeneity). If that's true, each cell should land near its expected count, which is (row total × column total) ÷ table total.

For every cell, you take (Observed − Expected)², divide by Expected, and add them all up: χ² = Σ(O − E)²/E. Squaring keeps every term positive, so χ² can never be negative. Dividing by E makes each gap relative. A miss of 5 matters a lot when you expected 10, but barely matters when you expected 500. Think of χ² as a "surprise meter." Zero means the data match the null perfectly, and the needle climbs as the table drifts further from what the null predicts.

Why Chi-Square Statistic matters in AP® Statistics

The chi-square statistic sits at the center of Topic 3.15, Carrying Out a Chi-Square Test for Homogeneity or Independence. It's the test statistic, so everything in the test runs through it. You compute χ², pair it with degrees of freedom, (rows − 1)(columns − 1), and use the chi-square distribution to get a p-value or compare to a critical value. Then you make a decision at your significance level. It also extends ideas you already know from inference. Like a z or t statistic, it measures distance from what the null predicts. The difference is that it works with counts across many categories instead of a single mean or proportion. If you can explain what a large χ² means in context, you can handle the conclusion step that FRQs grade heavily.

How Chi-Square Statistic connects across the course

Expected Counts and the Large Counts Condition (Topic 3.15)

The same expected counts you plug into the χ² formula also decide whether you're allowed to use the test at all. Every expected count needs to be at least 5. If some are tiny, dividing by a small E blows up that cell's contribution, and the chi-square distribution stops being a good model.

Degrees of Freedom and the Chi-Square Distribution (Topic 3.15)

A χ² value means nothing until you know its degrees of freedom. A χ² of 15.8 is very unusual with df = 4 but fairly ordinary with df = 15. The distribution is right-skewed, and its center shifts right as df grows, so df tells you which curve to measure your statistic against.

Significance Level and Critical Value (Topic 3.15)

There are two ways to make the same decision. You can compare the p-value to α, or compare χ² to the critical value for that α and df. If χ² beats the critical value, the p-value is automatically below α, and you reject H₀.

Independence and 10% Conditions (Topic 3.15)

The χ² formula assumes the counts come from independent observations. When you sample without replacement, the 10% condition (n ≤ 10% of the population) keeps that assumption close enough to true that the χ² value and its p-value can be trusted.

Is Chi-Square Statistic on the AP® Statistics exam?

Multiple-choice questions usually hand you χ² and ask you to finish the job. Typical stems give a table size and a χ² value, then ask for degrees of freedom, an approximate p-value, or a decision. For example, 4 blood types across 3 ethnic groups gives df = (4 − 1)(3 − 1) = 6. A 3 × 4 table with χ² = 23.7 also has df = 6, and that value lands past the 0.001 tail, so the p-value is below 0.001. Another stem gives df = 6, α = 0.01, and a critical value of 16.81. A χ² of 18 exceeds 16.81, so you reject H₀. On free-response questions, the work follows the full inference structure. State hypotheses in context, check conditions (random, 10%, expected counts ≥ 5), compute expected counts and χ², give df and the p-value, then conclude in context. No released FRQ has used the phrase "chi-square statistic" verbatim, but chi-square tests are standard FRQ material. Points often hinge on interpreting what a large χ² means, not just on calculating it.

Chi-Square Statistic vs Critical value

The chi-square statistic comes from your data. You calculate it from observed and expected counts. The critical value comes from the chi-square table or calculator. It depends only on α and degrees of freedom, and it marks the cutoff for "unusual enough." Your data never change the critical value. You compare your χ² to it, and if χ² is larger, you reject H₀.

Key things to remember about Chi-Square Statistic

• The chi-square statistic is χ² = Σ(Observed − Expected)²/Expected, summed over every cell in the table.

• Expected counts come from (row total × column total) ÷ table total, assuming the null hypothesis is true.

• χ² is always zero or positive, and a larger value means the data disagree more with the null hypothesis.

• For a two-way table, degrees of freedom equal (rows − 1)(columns − 1), and you need df to turn χ² into a p-value.

• If χ² is greater than the critical value at your significance level, the p-value is less than α, and you reject H₀.

• The chi-square distribution is only a valid model when all expected counts are at least 5 and the data come from random, independent observations.

Frequently asked questions about Chi-Square Statistic

What is the chi-square statistic in AP Stats?

It's the test statistic for chi-square tests, χ² = Σ(O − E)²/E. It measures how far the observed counts in a two-way table are from the counts expected under the null hypothesis. It's used in Topic 3.15 for tests of homogeneity and independence.

Can a chi-square statistic be negative?

No. Each term is a squared difference divided by a positive expected count, so χ² is always zero or positive. A χ² of 0 means the observed counts matched the expected counts exactly.

Does a large chi-square statistic always mean you reject the null?

Not by itself. Whether χ² is "large" depends on degrees of freedom. A χ² of 15.8 gives a p-value around 0.003 with df = 4, but it isn't significant with a much larger df. Always find the p-value or critical value for your df and α before deciding.

How is the chi-square statistic different from the critical value?

The chi-square statistic is calculated from your sample data. The critical value is a fixed cutoff based only on α and df. For example, with df = 6 and α = 0.01 the critical value is 16.81. A sample χ² of 18 exceeds it, so you'd reject H₀.

How do you find degrees of freedom for a chi-square test?

For a two-way table, df = (number of rows − 1)(number of columns − 1). A 3 × 4 table gives df = 2 × 3 = 6, and 4 blood types across 3 groups also gives df = 6.