Analyzing Residuals in Sociology: A Case Study on Welfare State Data, Assignments of Introduction to Sociology

Download Analyzing Residuals in Sociology: A Case Study on Welfare State Data and more Assignments Introduction to Sociology in PDF only on Docsity! Sociology 506 Problem Set 2 Due Wednesday April 5 at Noon A central part of any data analysis is a close examination of the residuals. The residuals are important for at least two reasons: • They are the keys to the step–by–step improvement of our analyses. • They are the keys to turning an investigator’s eye to the adequacy of the current analyses. These two observations were made by John Tukey (1977) in his classic work, Exploratory Data Analysis. Residuals from a regression function can be extracted in S+ by calling the function residuals. For example, >out <- lm(ws[,1] ws[,2]) >r <- residuals(out) >plot(r) Provide a brief analysis (less than 3 typed double-spaced pages) of the welfare state data (oecd.txt). In the write–up of the analysis report two tables: one with regression results (report coefficients, standard errors or t statistics, and R2 statistics), and another with residuals. Interpret the regression results and discuss the residuals. Some countries have much larger residuals than others. Why might this be so? Could the residuals be related to variables in the analysis, or perhaps other variables? The OECD dataset contains 3 variables: • decom: Decommodification measures welfare state development. Higher scores mean more equalitarian welfare states. • left: Left is a measure of the legislative and cabinet seat shares for left/labor parties. It represents the average from 1965-1980. • gdp: Per capita GDP is measured in 1,000s. Measures are from 1980, unless otherwise indicated. For more informa- tion you may wish to consult The Three Worlds of Welfare Capitalism by Gosta Esping-Andersen. 1