Chi-Square Test of Association Fisher Exact Probability Test For two groups of subjects, each sorted according to the absence or presence of some particular characteristic or condition, this page will calculate standard measures for Rates, Risk Ratio, Odds, Odds Ratio, and Log Odds. It will also calculate the Phi coefficient of association; T perform a chi-square test of association, if the sample size is not too small; and T perform the Fisher exact probability test, if the sample size is not too large. [Although the Fisher test is traditionally used with relatively small samples, the programming for this page will handle fairly large samples, up to about n=1000, depending on how the frequencies are arrayed within the four cells.] T For intermediate values of n, the chi-square and Fisher tests will both be performed. To proceed, enter the values of X 0 Y 1 , X 1 Y 1 , etc., into the designated cells. When all four cell values have been entered, click the Calculate button. To perform a new analysis with a new set of data, click the Reset button. The logic and computational details of the Chi-Square and Fisher tests are described in Chapter 8 and Subchapter 8a, respectively, of Concepts and Applications. A briefer account of the Fisher test will be found toward the bottom of this page. Data Entry T X Expected Cell Frequencies per Null Hypothesis 0 1 Totals Y 1 146 19 165 5.97 159.03 0 35 4800 4835 175.03 4659.97 Totals 181 4819 5000 CaIcuIate CaIcuIate Peset Peset 2x2 Contingency Table with Odds Ratios, etc. http://vassarstats.net/odds2x2.html 1 of 3 11/4/2012 11:34 PM Rate Risk Ratio Odds Odds Ratio Log Odds Group 1 0.1152 0.116 0.1301 0.0009 -6.9602 Group 2 0.9928 137.1429 Rate = proportion in group with condition present Risk Ratio = Rate[1]/Rate[2] Odds[1] = present[1]/absent[1] Odds[2] = present[2]/absent[2] Odds Ratio = Odds[1]/Odds[2] Log Odds = natural logarithm of Odds Ratio .95 Confidence Intervals Observed Lower Limit Upper Limit Risk Ratio 0.116 0.076 0.1771 Odds Ratio 0.0009 0.0005 0.0017 Chi-Square Chi-square is calculated only if all expected cell frequencies are equal to or greater than 5. The Yates value is corrected for continuity; the Pearson value is not. Both probability estimates are non-directional. Phi Yates Pearson -0.84 3497.12 3522.23 P <.0001 <.0001 Fisher Exact Probability Test: T P one-tailed Sample size too large two-tailed for the Fisher test. Click this link if you did not arrive here via the VassarStats main page. Richard Lowry 2001-2012 All rights reserved. 2x2 Contingency Table with Odds Ratios, etc. http://vassarstats.net/odds2x2.html 2 of 3 11/4/2012 11:34 PM 2x2 Contingency Table with Odds Ratios, etc. http://vassarstats.net/odds2x2.html 3 of 3 11/4/2012 11:34 PM