SOCIOL 3FF3 Lecture Notes - Lecture 9: Chi-Squared Distribution, Null Hypothesis, Standard Error

• School: McMaster University
• Department: Sociology
• Course: SOCIOL 3FF3
• Professor: Marisa Young

Document Summary

Chi-square tests hypothesis testing for nominal and ordinal data. Hypothesis testing w nominal/ordinal data: chi-square. 6 steps of hypothesis test: state alternative and null hypothesis, choose your distribution. Second type of hypothesis: differences btwn two groups in a sample, gender. Calculate standard error for the difference btwn means: anticipated level of error btwn measurements of two sample means. Sxbar = sx / square root n-1: describe the dispersion of the distribution. Standard error of difference btwn means: use this formula, determine alpha level and corresponding critical score. In 2 tail test df = (n1+n2) -2. In single tailed df = n-1: calculate t stat, observe the actual difference and calculate the test stat. Plot the critical and obtained t stat: make a decision. Should you retain or reject hypothesis: restate research question + answer it. Stat they are differences not due to change. 3rd type of hypothesi testing: assoc btwn 2 varibales, causation. Nominal and ordinal data: non parametric stat.