PSYC 2040 Lecture Notes - Lecture 9: Chi-Squared Test, Chi-Squared Distribution

• School: University of Guelph
• Department: Psychology
• Course: PSYC 2040
• Professor: David Stanley

Document Summary

Used for tesing for diference between observed and expected frequencies. Need a theory to establish expected frequencies (ex. rolling a die should get. 1/6 of each number: test of associaion between two nominal variables also called test of relaionship or correlaion. Don"t get degree of correlaion just get whether the variables are related or not. Therefore a poor it between the observed frequencies and the expected frequencies means the variables are related (if observed is so diferent from expected) If chi square is 0 raings are not related. If the observed frequencies are quite diferent from the expected frequencies, the variables must not be independent (reject the h0) variables must be related. If the observed frequencies are not very diferent from expected frequencies, there"s not enough evidence to conclude that the independence assumpion is wrong (fail to reject h0) A 2 test with variables having more than 2 levels. Might want to do post hocs to further understand the associaion.