PS296 Chapter Notes - Chapter 19: Chi-Squared Test, Chi-Squared Distribution, Contingency Table
Document Summary
The results of a chi square test when the data contain few observations can be very discontinuous. Yates" correction: for 2x2 tables, you apply a correction for continuity by reducing each numerator by one half unit before squaring, however this is usually not recommended. Take all possible 2x2 tables that could be formed from the fixed set of marginal totals. Calculate a chi square for each table and then determine the proportion of those tables whose results (chi square values) are as extreme, or more so, than the table we obtained from our data. If this proportion is less than alpha, we reject the null conclude conclude the two variables are independent and there is a statistically significant relationship between the two variables that make up our contingency table. Usually done with software and not by hand. Pearson s chi square and fisher"s exact test are two statistical tests for 2x2 contingency tables.