Online Chi-Square Test Calculator

Test Configuration

Choose your test type, configure the significance level, and enter your observed data below.

Significance Level (α)

Goodness of Fit — Data Input

Enter categories with observed and expected frequencies. Expected values can be raw counts or proportions (they will be auto-normalized to sum to the total observed count).

Category	Observed (O)	Expected (E)

Expected values must each be ≥ 5 for the Chi-Square approximation to be valid. If any expected value is less than 5, consider merging categories.

Independence — Contingency Table

Enter observed frequencies in each cell. Row and column totals are calculated automatically.

Rows

Columns

Ensure all expected cell frequencies are ≥ 5. For 2×2 tables with small samples, consider Fisher's Exact Test instead.

Result

—

—

Distribution Position

Visual comparison of your test statistic vs. the critical value on the Chi-Square distribution.

Step-by-Step Calculation

Follow every step of the Chi-Square computation to understand exactly how this result was reached.

Detailed Cell Contributions

Each cell's contribution to the overall χ² statistic: (O − E)² / E.

Assumption Checks

Verify that your data meets the requirements for a valid Chi-Square test.

How to Use This Tool

A complete walkthrough of the Chi-Square calculator — from data entry to interpreting your results.

Select Your Test

Choose Goodness of Fit when you want to compare one categorical variable's observed distribution against a hypothesized (expected) distribution.

Choose Test of Independence when you want to determine whether two categorical variables are associated using a contingency table.

Set Significance Level

The significance level (α) is the threshold for rejecting the null hypothesis. Common choices:

• α = 0.05 — Standard in most social sciences
• α = 0.01 — Stricter, used in medical studies
• α = 0.001 — Very strict, used in physics

Enter Your Data

Goodness of Fit: Enter a category name, observed frequency (actual count), and expected frequency (theoretical count) for each category.

Independence: Set the number of rows and columns, label them, then enter the observed count in each cell.

Interpret Results

The tool shows the χ² statistic, p-value, degrees of freedom, and critical value.

If p ≤ α → reject H₀ (significant result).
If p > α → fail to reject H₀ (not significant).

Formulas Used

The mathematical formulas behind every calculation in this tool.

Chi-Square Statistic Formula

χ² = Σ [ (O - E)² / E ] Where: O = Observed frequency (your actual data) E = Expected frequency (theoretical / hypothesized) Σ = Sum over all categories or cells

Each term (O−E)²/E measures how far off one category is from the expectation. Larger deviations contribute more to the final χ² value.

Degrees of Freedom

Goodness of Fit: df = k − 1 (k = number of categories) Test of Independence: df = (r − 1) × (c − 1) (r = rows, c = columns)

Expected Frequencies (Independence Test)

E(i,j) = (Row Total_i × Column Total_j) / Grand Total

For each cell (i,j), the expected frequency is calculated from the marginal totals, assuming independence between the two variables.

P-Value Calculation

The p-value is the area to the right of the χ² statistic in the Chi-Square distribution with the given degrees of freedom. It represents the probability of observing a result this extreme (or more) under the null hypothesis.

This tool uses the regularized incomplete gamma function for precise p-value computation:

p = 1 − P(df/2, χ²/2) = Q(df/2, χ²/2) (regularized upper gamma)

Cramér's V — Effect Size

V = √( χ² / (n × min(r−1, c−1)) ) Interpretation: V ≈ 0.10 → Small effect V ≈ 0.30 → Medium effect V ≈ 0.50 → Large effect

Assumptions & Conditions

These conditions must be met for the Chi-Square test to produce valid results.

Expected ≥ 5: All expected cell frequencies should be at least 5. If not, merge categories or use Fisher's Exact Test.

Independent observations: Each data point must come from a different subject — no repeated measures.

Categorical data only: Variables must be nominal or ordinal, not continuous measurements.

Adequate sample size: A total sample size of at least 20 is generally recommended for reliable results.

Worked Examples

Click "Load into Calculator" on any example to pre-fill the tool and see the results instantly.

Fair Dice Test

Goodness of Fit · 6 categories

A die is rolled 120 times. We expect each face to appear 20 times if the die is fair. Observed results:

Face	Observed	Expected
1	25	20
2	17	20
3	15	20
4	23	20
5	19	20
6	21	20

Mendel's Genetics

Goodness of Fit · 4 categories

Mendel predicted a 9:3:3:1 ratio for seed phenotypes. Observed in 560 seeds:

Phenotype	Observed	Expected
Round Yellow	315	315
Round Green	108	105
Wrinkled Yellow	101	105
Wrinkled Green	36	35

Gender × Voting Preference

Independence Test · 2×3 table

Is voting preference associated with gender? Survey of 250 people:

	Party A	Party B	Undecided
Male	40	55	30
Female	45	60	20

Drug Treatment Outcome

Independence Test · 2×2 table

Does a new drug improve recovery? Clinical trial results:

	Recovered	Not Recovered
Treatment	45	15
Placebo	20	40

Key Facts

Essential facts about the Chi-Square test you should know.

1900

Year of Origin

Karl Pearson introduced the Chi-Square test in 1900, making it one of the oldest statistical tests still in widespread use.

df+

Right-Skewed Distribution

The Chi-Square distribution is always right-skewed and non-negative. It approaches a normal distribution as degrees of freedom increase.

χ²

Non-Negative Statistic

The Chi-Square statistic is always ≥ 0. A value of 0 means perfect agreement between observed and expected frequencies.

≥5

Minimum Expected Frequency

Each expected cell frequency must be at least 5 for the approximation to be valid — this is the most common violated assumption.

H₀

Null Hypothesis

The null hypothesis assumes no difference (GoF) or no association (Independence). You can only reject it — never "prove" it.

V

Cramér's V Effect Size

Statistical significance ≠ practical importance. Always report Cramér's V (or φ for 2×2) as the effect size alongside the p-value.

Advanced Tips

Pro tips from statisticians to help you use Chi-Square tests correctly.

Tip 1: Don't Confuse Significance with Importance

A statistically significant result (small p-value) only means your data would be unlikely if H₀ were true. With large samples, even tiny, meaningless differences become statistically significant. Always compute Cramér's V to assess practical (real-world) importance.

Tip 2: Merge Categories When E < 5

If any expected cell is below 5, try combining adjacent or conceptually similar categories. Alternatively, use Fisher's Exact Test (for 2×2 tables) or a Monte Carlo simulation-based Chi-Square test for more reliable results.

Tip 3: Chi-Square Cannot Show Causation

Even if two variables are significantly associated (Independence test), that does not mean one causes the other. Correlation and association do not imply causation. Always consider confounding variables and study design.

Tip 4: One-Tailed vs Two-Tailed

The Chi-Square test is inherently one-tailed (always right-tailed) because the statistic is squared and can only be positive. You're always testing whether the statistic is unusually large — which is why you always use the right tail of the distribution.

Tip 5: Check Cell Residuals for Insight

When you get a significant result, look at the standardized residuals for each cell: (O−E)/√E. Cells with |residual| > 2 are driving the significance — they tell you which specific categories differ most from expectation.

Tip 6: Yates' Continuity Correction

For 2×2 tables with small samples, Yates' correction adjusts the formula: χ² = Σ (|O−E|−0.5)² / E. This makes the test more conservative and is recommended when total N < 40 or any expected cell is between 5 and 10.

Real-World Use Cases

The Chi-Square test is one of the most versatile statistical tools — applied across virtually every field of science and industry.

Medical Research

Test whether a drug's effectiveness is associated with patient demographics. Compare treatment outcomes across multiple groups.

Market Research

Determine if customer preference varies by age group, region, or income level. Validate survey instrument categories.

Education Studies

Test whether pass/fail rates differ across teaching methods, schools, or demographic groups. Analyze exam score distributions.

Political Science

Examine whether voting patterns are independent of geographic region, education level, or other demographic factors.

Genetics & Biology

Test if offspring phenotype ratios match Mendelian inheritance predictions. Analyze population genetics data.

Quality Control

Determine if defect rates differ across production lines, shifts, or machine configurations. Monitor manufacturing processes.

A/B Testing

Compare click-through rates, conversions, or user behavior across different website designs or marketing campaigns.

Legal & Social Science

Analyze whether jury verdicts, sentencing, or legal outcomes are independent of defendant characteristics.

Environmental Science

Test if species distribution patterns differ across habitats. Compare pollution levels across multiple zones or time periods.

About This Tool

Everything you need to know about this Chi-Square calculator and the statistics behind it.

Purpose

This tool provides a professional, browser-based Chi-Square test calculator that requires no statistical software, no account, and no data upload. All calculations are performed locally in your browser using mathematically rigorous algorithms.

Privacy

Your data never leaves your device. All computations happen entirely in JavaScript within your browser. No server, no cookies, no tracking. Your research data remains completely private and secure.

Algorithms

P-values are computed using the regularized incomplete gamma function approximation, matching the precision of major statistical packages (R, SPSS, SAS). Critical values are derived from the Chi-Square quantile function.

Features

• Goodness of Fit test (any number of categories)
• Test of Independence (up to 10×10 tables)
• Step-by-step calculation walkthrough
• Cell contribution breakdown table
• Assumption checker with warnings
• Cramér's V effect size
• Visual distribution position indicator
• Export / print results

Chi-Square vs Other Tests

When to use Chi-Square and when to consider alternatives.

Situation	Recommended Test	Why
Categorical vs hypothesized distribution	Chi-Square GoF	Direct comparison of observed vs expected frequencies
Two categorical variables, large N	Chi-Square Independence	Standard for contingency table analysis
2×2 table, small N (<20)	Fisher's Exact Test	Chi-Square approximation breaks down for small samples
Ordinal categorical variable	Cochran-Armitage or Mann-Whitney	Captures the ordering of categories
Continuous data, normal distribution	t-test / ANOVA	Chi-Square is for counts, not measurements
Paired / repeated measures	McNemar's Test	Chi-Square assumes independent observations

Citation: 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, Series 5, 50(302), 157–175.

Copied!

Feature	Details
Price	Free
Rendering	Client-Side Rendering
Language	JavaScript
Paywall	No

Open This Tool

Checkout More Statistical Tools!

Advertise here

About This Tool

How It Works?