This file is raw output from pdftotext and may not be ideal for distribution. If you are a maintainer for Hackipedia, please sit down when you have time and clean this text version up. Source PDF: /mnt/fw-js/docs/Math/Statistics/Significance of a Chi-Square Statistic.pdf Like all conversions the text below should be fully readable as UTF-8 unicode text. --------------------------------------------------------------- Appendix 3: Significance of a Chi-Square Statistic For 30 or fewer degrees of freedom, an exact series expansion is used; otherwise the Peizer-Pratt approximation is used. Notation The following notation is used in this appendix: X Value of the chi-square statistic k Degrees of freedom Q Significance level (right-tail probability) Computation • If X ≤ 0 or k < 1, Q =1 • If k = 1 , Q = 2Q N e Xj where Q N e X j is the standard normal one-tailed significance probability. • For k ≤ 30 , an exact series expansion is used (Abramowitz and Stegun, 1965, eqs. 26.4.4 and 26.4.5) 555 556 Appendix 3 R |2Q d X i + R 2 expF − X I k odd | | N π H2K Q=S | −X |expF I × a1 + Rf | H2K T k even where Ra f | k −1 2 r −1 2 | ∑ 1⋅ 3Ka2r − 1f | X k odd r =1 | R=S |a f | k −2 2 r | ∑ 2 ⋅ 4XK2r | k even T r =1 • If k > 30 , the Peizer-Pratt approximation is used (Peizer and Pratt, 1968, eq 2.24a). • If X ≥ 150 , Q=0 otherwise Q = QN Z bg where Appendix 3 557 R |FG −F 1 + 0.08 I IJ d 2k − 2 i |H H 3 k K K if X = k − 1 | | Z=S | |F d k − 1 logF k − 1I + X − k − 1 I |GH a f H X K a f JK a f X − k −1 if X ≠ k − 1 | T where d = X − k + 2 3 − 0.08 k • If Z < 0 , Q = 1 − QN Z bg References Peizer and Pratt (1968)