SOC202H1 Lecture : SOC202 MARCH 6.docx
Document Summary
T and f test both require normaility or normal sample large enough, interval level data, research not always have assumptions meant, not interval, or two categories. When data not meet assumptions the chi square not need normality. Chance of type of 1 increase, more likely reject when retain. Looking at sig difference between table and expecting get if no sig difference. Get if null true are expected frequencies. Present in table format, raw data put in data format. Non sym and on tailed like f test. Ordinal or nominal no interval, both in and dep. Or no difference what observed or expect to get, null is no difference. One way chi square formula is sum of difference fo and fe squared divided fe, careful where symbols appear. Df = k -1 k number of groups. Observed freq for more then one variable. We compare ind and dependednt nominal or interval. Df changes is row minus 1 times column minus 1.