CRIM 320 Lecture Notes - Lecture 4: Chi-Squared Distribution, Null Hypothesis, Level Of Measurement

• School: Simon Fraser University
• Department: Criminology
• Course: CRIM 320
• Professor: Steph Dawson

Document Summary

Chi square analysis - basic foundation fundamental technique for analyzing nominal data. Null hypothesis - 2 variables are independent (fail to reject) If we reject the null, we are stating that the two variables are dependent. If 2 variables is independent, knowing 1 variable doesn"t help us at all with our knowledge of the other variables. If something is dependent, if we know something about 1 variable, we will have a better idea of another variable. Need to look at observed counts and expected counts and compare those two. Reject the null if the observed counts are sufficiently (probabilities) different from the expected counts. X^2 = (observed frequency expected frequency)^2 / expected frequency. Chi-squared = sum of (observed frequency expected frequency) ^2 / expected frequency data e a frequency. Observed frequency and expected frequency in order to plug it in the formula. Calculate expected counts (observe red - expected)^2 / expected.