A goodness-of-fit test introduced in 1950. If there are n observed frequencies, f1, f2,⋯, fn with, according to a p-parameter model, corresponding expected frequencies e1, e2,⋯, en, then the Freeman–Tukey statistic, T2, is given byIf the model is correct then the distribution of T2 is approximately chi-squared with (n−p) degrees of freedom.