What does the chi-square test tell you in genetics?

It measures whether the difference between the offspring counts you observed and the counts a genetic ratio predicts is small enough to be due to chance. A small chi-square value means the observed and expected numbers are close, supporting your hypothesized ratio. A large chi-square value means the deviation is bigger than random sampling easily explains, which is evidence against the predicted ratio.

How do I find the degrees of freedom?

For a goodness-of-fit test the degrees of freedom equal the number of phenotype categories minus one. A monohybrid 3 to 1 cross has two categories, so one degree of freedom. A dihybrid 9 to 3 to 3 to 1 cross has four categories, so three degrees of freedom. You subtract one because the total count is fixed, so once all but one category is known the last is determined.

What does the p-value mean here?

The p-value is the probability of seeing a deviation at least as large as yours if the predicted ratio were true. By the usual 0.05 cutoff, a p-value above 0.05 means you fail to reject the hypothesis, so the data are considered a good fit to the ratio. A p-value of 0.05 or below means the deviation is statistically significant and the observed data do not fit the expected ratio.

Why is a high p-value good in a genetics chi-square test?

Because the hypothesis being tested is that the data fit the predicted ratio. A high p-value means the observed counts are close to the expected counts, so you fail to reject that hypothesis and conclude the cross behaved as predicted. This is the opposite of many experiments, where a low p-value is the desired result. In a goodness-of-fit test, a high p-value supports your genetic model.

When should I not use a chi-square test?

The chi-square goodness-of-fit test assumes you are working with actual counts, not percentages or proportions, and it becomes unreliable when expected counts in any category are very small, roughly below five. It also tests only whether the data fit one specific ratio; it cannot by itself tell you which alternative ratio is correct. For very small samples an exact test is more appropriate.

Chi-Square Genetic Ratio Calculator, Free

Expected ratio (pick a preset or edit the ratio column)

Phenotype	Observed count	Ratio part

Verdict at α = 0.05

—

Chi-square χ²

—

Deg. freedom

—

p-value

—

Phenotype	Observed	Expected	O − E	(O−E)²/E
χ² total =				—

How the chi-square test works

A genetic hypothesis predicts a ratio, and real data always wander from it a little through pure chance. The chi-square test asks whether the wandering is small enough to be chance, or too big to ignore. For every category you compute (observed − expected)² / expected and add them up: that sum is χ². A small χ² means the counts hug the prediction; a large χ² means they stray. You then compare χ² against the chi-square distribution for your degrees of freedom (categories minus one) to get a p-value. By the standard 0.05 cutoff, a p-value above 0.05 means the data fit the ratio, and 0.05 or below means the deviation is significant. Note that here a high p-value is the good result, because the hypothesis is that the cross matches the ratio.

Critical values at α = 0.05

Degrees of freedom	Critical χ²	Meaning
1 (2 categories, e.g. 3:1)	3.84	χ² above this rejects the fit
2 (3 categories, e.g. 1:2:1)	5.99
3 (4 categories, e.g. 9:3:3:1)	7.81
4 (5 categories)	9.49
5 (6 categories)	11.07

If your χ² is below the critical value (equivalently, p > 0.05), you fail to reject the hypothesis and the data fit the ratio. This calculator reports the exact p-value, so you do not have to read a table.

FAQ

What does the test tell you?
Whether the gap between observed and expected counts is small enough to be chance (data fit the ratio) or too large (significant deviation).

How do I get degrees of freedom?
Categories minus one. A 3:1 cross has df = 1; a 9:3:3:1 cross has df = 3.

Why is a high p-value good here?
The hypothesis is that the cross fits the ratio, so a high p-value (deviation is just chance) supports your genetic model.

Percentages or counts?
Always use raw counts, never percentages. Chi-square depends on sample size.

When is it unreliable?
When any expected count is very small (roughly under 5). Use larger samples or an exact test.