POL 3371 Study Guide - Midterm Guide: World Values Survey, Null Hypothesis, Statistical Significance

• School: University of Ottawa
• Department: Political Science
• Course: POL 3371
• Professor: Ivan Katchanovski

Document Summary

Dependent variable: in theory: interval/ratio, in practice: ordinal variables can also be used if they assumed to be interval/ratio, need to specify explicitly/check this assumption, recode values if needed. Independent variable: nominal, two categories/values, male and female, nominal variables with more than 2 categories, can be recoded into 2 categories for t-test, or used in anova test. Independent samples t-test: assumptions, t-test uses student t-distribution/t-statistics, dependent variable is normally distributed in both independent variable levels. Each group/category of the independent variable: at least 30 o members/cases. Sample drawn from populations with equal variances on the dependent o variable. Criteria of statistical significance: statistically significant if less than one of the conventional levels of statistical significance, use equal variances is assumed p(obtained) values for t-test, levene"s test for equality of variances is statistically non-significant. Criteria of statistical significance: statistical significant is less than one of the conventional levels of statistical significance, conventional levels of statistical significance: