Chi Square Calculator: Formula, Steps & How to Use

A chi-square calculator computes the chi-square statistic (χ²) and its associated p-value from observed and expected frequencies you enter, letting you test whether categorical data differ from what chance would predict. You supply the observed counts for each category and either the expected counts or the proportions they imply, and the calculator returns the test statistic, degrees of freedom, and a p-value. Use it to run a chi-square goodness-of-fit test or a chi-square test of independence on a contingency table.

Formula

The chi-square statistic measures how far observed frequencies deviate from expected frequencies. The formula is:

χ² = Σ ( (Oᵢ − Eᵢ)² / Eᵢ )

• Oáµ¢ = observed frequency in category i (count of items in that category)
• Eáµ¢ = expected frequency in category i (what you would expect under the null hypothesis)
• Σ = sum over all categories (or all cells in a contingency table)

The degrees of freedom (df) depend on the test. For a goodness-of-fit test, df = (number of categories − 1). For a test of independence in a table with r rows and c columns, df = (r − 1)(c − 1). The p-value is the area under the chi-square distribution to the right of the computed statistic, using the appropriate df.

Worked Example

Using the default values from the calculator: observed counts are 50, 30, 20 and expected counts are 40, 40, 20.

1. For each category, compute (O − E)² / E:
2. Category 1: (50 − 40)² / 40 = 100 / 40 = 2.5
3. Category 2: (30 − 40)² / 40 = 100 / 40 = 2.5
4. Category 3: (20 − 20)² / 20 = 0 / 20 = 0
5. Sum these values: 2.5 + 2.5 + 0 = 5.0
6. Degrees of freedom = 3 − 1 = 2
7. Using a chi-square table or software, the p-value for χ² = 5.0 with df = 2 is approximately 0.0821.

The calculator shows χ² = 5.000, df = 2, and p = 0.0821, confirming the manual calculation.

When to Use a Chi-Square Test

Chi-square tests are used when your data are counts in categories—not measurements like height or time. Common applications include:

• Goodness-of-fit: Test whether observed counts match a theoretical distribution (e.g., a die is fair, or a sample matches population proportions).
• Test of independence: Determine whether two categorical variables are related (e.g., gender and product preference, or treatment group and recovery status).
• Test of homogeneity: Compare the distribution of a categorical variable across multiple populations (e.g., survey responses across age groups).

The null hypothesis always states there is no difference or no association; the alternative is that there is a difference or association. A small p-value (typically below 0.05) suggests the observed data are unlikely under the null, leading you to reject it.

How to Read the Result

The chi-square statistic itself is not directly interpretable without context. Always pair it with the degrees of freedom and the p-value. A larger χ² relative to the df indicates a larger discrepancy between observed and expected. The p-value tells you the probability of seeing a χ² at least this large if the null hypothesis were true. If the p-value is less than your chosen significance level (commonly 0.05), you reject the null hypothesis and conclude that the observed pattern is statistically significant.

For example, with χ² = 5.0 and df = 2, the p-value is 0.0821. Since 0.0821 > 0.05, you would not reject the null hypothesis at the 5% level—there is not strong evidence that the observed counts differ from the expected ones.

Common Mistakes and Pitfalls

• Using percentages instead of raw counts: Chi-square requires actual frequencies, not proportions. If you only have percentages, convert them to counts before entering.
• Ignoring the expected count assumption: The chi-square approximation works well when all expected counts are at least 5. If any expected count is below 5, consider combining categories or using Fisher’s exact test (for 2×2 tables).
• Mixing up the two tests: Goodness-of-fit uses one row of observed and expected values; independence uses a contingency table with rows and columns. The degrees of freedom differ, so be sure you use the correct calculation.
• Forgetting to sum correctly: The formula requires summing over all categories or cells. Missing a category or double-counting a cell will give a wrong result.
• Rounding too early: Round intermediate calculations only at the end to avoid accumulating rounding errors.

Limitations and Assumptions

The chi-square test is not valid when the data are not independent (e.g., repeated measures on the same subjects) or when the categories are not mutually exclusive. It also assumes the sample is random. For small expected counts, the test becomes unreliable; the rule of thumb is that at least 80% of expected counts should be ≥5 and none should be <1. If your data violate these assumptions, consider exact tests or other methods.

Additionally, the chi-square test is a hypothesis test, not a measure of effect size. A significant result only tells you that there is an association or difference, not how strong it is. For effect size, you can compute Cramér’s V or the phi coefficient.

Frequently Asked Questions

What is the difference between chi-square and t-test?

A chi-square test is used for categorical data (counts in categories), while a t-test compares means of continuous data. If your outcome is a measurement like weight or time, use a t-test or ANOVA; if it is a count like yes/no or color, use chi-square.

Can I use chi-square with percentages?

No, you must use raw counts. Percentages lose the sample size information, which is essential for calculating the statistic. Convert percentages to counts by multiplying the total sample size by each percentage.

What is a good chi-square value?

There is no absolute “good� value. A smaller chi-square (relative to df) indicates observed data are closer to expected, and a larger one indicates a bigger deviation. The p-value is what matters for decision-making.

How do I calculate expected counts for a test of independence?

For each cell in a contingency table, the expected count is (row total × column total) / grand total. For a goodness-of-fit test, expected counts are given by the null hypothesis (e.g., equal proportions or a specified distribution).

What does a p-value of 0.05 mean in chi-square?

It means that if the null hypothesis were true, there is a 5% chance of obtaining a chi-square statistic as large as (or larger than) the one you observed. Typically, a p-value below 0.05 is considered statistically significant, leading to rejection of the null.

Can I use chi-square for 2×2 tables?

Yes, but with a caution: if any expected cell count is below 5, the chi-square approximation may be poor. In that case, use Fisher’s exact test, which is accurate for small samples.

Related

• Binomial Cumulative Distribution Calculator
• Binomial Distribution Calculator Keys Explained
• Binomial Distribution Calculator Iowa: How to Use It

in obs | Observed counts (comma-separated) | - | 50, 30, 20 | 0
in exp | Expected counts (comma-separated) | - | 40, 40, 20 | 0
out chi2 | Chi-square statistic | - | sum((obs_i - exp_i)^2 / exp_i) | 3
out df | Degrees of freedom | - | count(obs) - 1 | 0
out p | p-value (right-tail) | - | 1 - chi2cdf(chi2, df) | 4