chi squared test


Also found in: Dictionary, Encyclopedia.

chi squared test

X2 Statistics A test of association between 2 or more variables, used to identify a linkage between a factor or attribute and an outcome, and determine if the observed values of a variable are significantly different from those expected based on a null hypothesis. See Null hypothesis.

chi squared test

A series of statistical procedures used to test how closely the observed result of a trial or a statistical observation corresponds to an expectation, hypothesis or hoped-for result. If, for instance, the expected outcome of a trial is that half the participants will show a particular result, the expected frequency (E) is the total number of participants divided by 2. If Y is the number that show the result and N is the number that do not, then chi squared = (Y-E)2/E + (N-E)2/E. A low chi squared value supports the expectation or hypothesis; a high value rejects it. Tables relating the value to statistical significance are available.
References in periodicals archive ?
This was done with Chi squared tests separating the change in Tobin's Q into above and below average for both the one and two years measures to see which of the categories--method, type and number of advisors, has as an association.