| 释义 |
chi-squared test A test, normally written as the χ2-test, to determine how well a set of observations fits a particular discrete distribution or some other given null hypothesis (see hypothesis testing). The observed frequencies in different groups are denoted by Oi, and the expected frequencies from the statistical model are denoted by Ei. For each i, the value (Oi-Ei)2/Ei is calculated, and these are summed. The result is compared with a chi-squared distribution with an appropriate number of degrees of freedom. The number of degrees of freedom depends on the number of groups and the number of parameters being estimated. The test requires that the observations are independent and that the sample size and expected frequencies exceed minimum numbers depending on the number of groups.
|