SOCI 3700 Lecture Notes - Lecture 1: Confounding, Statistical Significance, Null Hypothesis
7 Jun 2019
School
Department
Course
Professor
Document Summary
Spurious: assume that x causes y, therefore icecream sales must causes violent crime to happen, but temperature increases both. Both your x and y are caused by a confounding variable. Mediated: x causes the z, and then causes the y. Incarceration causes depression, but maybe it doesn"t have a direct effect but an indirect effect. Incarceration causes you to lose contact with family, which then causes depression. X causing y, but z moderates the relationship that x has on y. It changes the relationship between x and y. Sex causes educational attainment (women can be discriminated against). But income might moderate that, race, parental marriage. Variables may not measure construct correctly so we need to create different categories that make sense. Social researchers use statistics to describe populations and relationships between phenomena. In order to move from what questions to how or why questions, we must use multivariate methods, which examine at least 3 variables.
