chi-square
(kaɪ:ˈskwɛə(r))
[f. chi + square n. 8.]
Applied (usu. attrib.) to a test for, or statistic of, comparison between observed and theoretical values (see also quot. 1960). Also chi-squared a.
K. Pearson, who devised this test in 1900, called it the ‘goodness of fit test’ or ‘χ2 test’.
| 1936 Davenport & Ekas Statist. Methods Biol. (ed. 4) 47 The χ2 (‘chi square’) test. Find for each class the difference (δ1) between the theoretical value (y) and the observed frequency (f). Divide the square of this difference in each case by y. The sum of the quotients is χ2. 1939 Kurtz & Edgerton Statist. Dict. 24 Chi square test of goodness of fit, a function from which can be obtained approximately the probability that a given set of data..might arise if a certain law or cause were in operation. 1946 [see additive a. d]. 1949 British Birds XLII. 256 Application of the chi-square test to a comparison of my figures and Höhn's..shows that the chance of his part A and my column 3 differing only by chance is much less than 1 in 100. 1957 P. Lafitte Person in Psychol. 150 This kind of analysis..uses only non-parametric methods, mostly chi-square. Ibid. 151 These variables are compared with the criterion by means of chi-square (or the t-test, if the criterion is quasi-physical). 1960 Kendall & Buckland Dict. Statist. Terms 41 Chi-squared Statistic..For historical reasons..it more usually relates to a statistic of a particular kind which is, as a general rule, distributed more or less approximately in the χ2 form. Ibid. 42 Chi-squared Test..Such tests occur in many ways, the most prominent being: (i) An overall goodness-of-fit comparison of observed with hypothetical frequencies falling into specified classes; (ii) Comparison of an observed with a hypothetic variance in normal samples; (iii) Combination of probabilities from a number of tests of significance. 1961 L. F. Brosnahan Sounds of Lang. iii. 66 It was decided to test the significance of the co-incidence of the two distributions by the Chi-square technique. |