Chi-Square Test for Compatibility of K Counts

Tests the null hypothesis that all $s$ categories of a categorical variable, $x$ , are the same size. The test statistic is

$X^{2} = \sum_{k = 1}^{s} \frac{(o_{k j} - e_{k j})^{2}}{e_{k j}}$

where:

o_{k} = \sum_{i = 1}^{n} w_{i} I_{x = k}

,

w_{i}

i

th of

n

observations,

e_{k} = \frac{\sum_{i = 1}^{n} w_{i}}{s}

p \approx \Pr (χ_{(s - 1)}^{2} \geq X^{2})

See also