The Chi-Square Goodness-of-Fit Test

Observed vs. Expected Counts, Degrees of Freedom, and When to Reject — A TLDR Primer

The chi-square goodness-of-fit test shows up on the AP Statistics exam, in intro college stats courses, and on standardized tests — and it trips students up every time. Not because the math is hard, but because the setup, the logic, and the vocabulary all hit at once. This guide cuts straight to what you need.

Inside, you'll find a clear explanation of what the test actually asks (is this data consistent with a claimed distribution?), how to turn claimed proportions into expected counts, and how to build the chi-square statistic term by term. Degrees of freedom, the chi-square distribution, and how to read a p-value from a table or calculator are all covered plainly — no hand-waving, no skipping steps.

The guide also walks through the conditions students most often ignore (random sample, independence, expected counts of at least 5), shows exactly how to write a reject or fail-to-reject conclusion in context the way AP graders want to see it, and flags the misconceptions that cost points on exams. A full end-to-end worked example using M&M color claims ties every piece together. A final section clarifies how the goodness-of-fit test differs from the chi-square test for independence and the test for homogeneity — the confusion between these three is one of the most common mistakes in AP Statistics.

If you're prepping for the AP Statistics exam or working through an intro college statistics course and need a concise, no-filler resource on chi-square hypothesis testing for categorical data, this primer is built for you.

Grab it, work the examples, and walk into your exam ready.

What you'll learn

State null and alternative hypotheses for a goodness-of-fit problem
Compute expected counts from a claimed distribution
Calculate the chi-square statistic and degrees of freedom
Use the chi-square table or p-value to make a decision at a given significance level
Check the conditions (independence, sample size, expected counts ≥ 5) and recognize when the test does not apply
Distinguish goodness-of-fit from the chi-square tests for independence and homogeneity

What's inside

1. What the Goodness-of-Fit Test Actually Asks

Introduces the test as a way to compare observed category counts against a claimed distribution, with motivating examples like dice fairness and M&M color claims.
2. Expected Counts and the Chi-Square Statistic

Shows how to turn claimed proportions into expected counts and build the chi-square statistic term by term, with a fully worked dice example.
3. Degrees of Freedom, the Chi-Square Distribution, and P-Values

Explains why degrees of freedom equal k−1, what the chi-square distribution looks like, and how to convert a test statistic into a p-value using a table or calculator.
4. Conditions, Decisions, and Writing the Conclusion

Covers the conditions for using the test (random sample, independence, expected counts ≥ 5), how to phrase a reject/fail-to-reject conclusion in context, and common mistakes.
5. A Full Worked Example: Are M&M Color Claims True?

End-to-end walkthrough on a Mars candy color distribution problem: hypotheses, expected counts, statistic, df, p-value, and conclusion.
6. Goodness-of-Fit vs. Independence vs. Homogeneity

Clarifies how this test differs from the other two chi-square tests students often confuse it with, and previews where each one shows up on exams.

Published by Solid State Press

TLDR STUDY GUIDES

The Chi-Square Goodness-of-Fit Test

Observed vs. Expected Counts, Degrees of Freedom, and When to Reject — A TLDR Primer

Solid State Press

The Chi-Square Goodness-of-Fit Test

Observed vs. Expected Counts, Degrees of Freedom, and When to Reject — A TLDR Primer

Published by Solid State Press.

Part of the TLDR Study Guides series.

1 What the Goodness-of-Fit Test Actually Asks
2 Expected Counts and the Chi-Square Statistic
3 Degrees of Freedom, the Chi-Square Distribution, and P-Values
4 Conditions, Decisions, and Writing the Conclusion
5 A Full Worked Example: Are M&M Color Claims True?
6 Goodness-of-Fit vs. Independence vs. Homogeneity

Chapter 1

What the Goodness-of-Fit Test Actually Asks

Suppose you roll a six-sided die 60 times and record how often each face comes up. You get something like this:

Face	1	2	3	4	5	6
Count	8	12	9	14	7	10

The die should land on each face equally often — 10 times each, if the die is fair and you roll exactly 60 times. But your counts don't match that perfectly. The question is: does this mismatch reflect a genuinely unfair die, or just the ordinary randomness you'd expect even from a fair one?

That is precisely what the goodness-of-fit test answers. It takes a set of observed counts — the actual frequencies you recorded in each category — and compares them against what a specific claimed distribution predicts. The test asks whether your data are consistent with the claim, or whether the gap is large enough to cast real doubt on it.

Categorical Data and Counted Outcomes

The goodness-of-fit test applies to categorical data: data where each observation falls into one of several distinct, non-overlapping categories. Die faces are categorical. M&M colors are categorical. The distribution of blood types in a sample is categorical. What these situations share is that you count how many observations fall into each category — you don't measure a continuous quantity like height or temperature.

The categories have to be exhaustive (every observation lands in exactly one of them) and the counts have to be whole numbers. You'll also need a specific claim about the proportions each category should represent — that claim is what the test puts on trial.

The Null and Alternative Hypotheses

Every hypothesis test starts with two competing statements. The null hypothesis ($H_0$) is the "nothing unusual here" position — it states that the population distribution matches the claimed distribution exactly. The alternative hypothesis ($H_a$) says the opposite: the true distribution is different from the claim in some way.

For the die example:

About This Book

If you're staring down the AP Statistics chi-square test on an upcoming exam, working through a chapter on hypothesis testing in an intro college statistics course, or just trying to figure out how to do a chi-square test for a high school project, this book is for you. It's also useful for tutors who want a clean, no-detour reference before a session.

This guide covers everything the test actually requires: setting up null and alternative hypotheses, calculating observed vs. expected counts in statistics, computing the chi-square statistic, and interpreting chi-square p-value and degrees of freedom results against a critical-value table. It also clarifies where the goodness-of-fit test ends and the tests for independence and homogeneity begin — a distinction that trips up a lot of AP Stats hypothesis testing review sessions. Short by design, with no filler.

Read straight through, then work every example as you go. A problem set closes the guide — attempt it cold to find out what actually stuck.

Keep reading

You've read the first half of Chapter 1. The complete book covers 6 chapters in roughly fifteen pages — readable in one sitting.

Coming soon to Amazon