POLS 3650 Lecture 15: POLS 3650 mar 4.docx
Document Summary
Bivariate stats: nominal: need to standardize values in crosstabs, chi square is not a measure of association, used for calculating the purpose of statistical significance, like f in anova, and z in compare means test, table 3 shows no relationship bw iv and dv bc percentages are equal across the rows, summation of [(observed freq expected freq) 2]/expected freq, df number of moving parts in the data, the greater the df the bigger the chi square value, df table how large chi square needs to be to reject h not at these alpha levels of the null hypothesis, if our value is less than 16. 92 in the example, we cannot reject it with 95% confidence (cid:224) 9df and 0. 05, cramer"s v turns chi square into a measure of association, get a value bw 0 and 1, 0 = no relationship, 1 = strong relationship, square root of (chi square/ total number of cases * #rows 1 or #columns .