Advances in Social Science Research Using R

Quantitative social science research has been expanding due to the ava- ability of computers and data over the past few decades. Yet the textbooks and supplements for researchers do not adequately highlight the revolution created by the R software [2] and graphics system. R is fast becoming the l- gua franca of quantitative research with some 2000 free specialized packages, where the latest versions can be downloaded in seconds. Many packages such as “car” [1] developed by social scientists are popular among all scientists. An early 2009 article [3] in the New York Times notes that statisticians, engineers and scientists without computer programming skills ?nd R “easy to use.” A common language R can readily promote deeper mutual respect and understanding of unique problems facing quantitative work in various social sciences. Often the solutions developed in one ?eld can be extended and used in many ?elds. This book promotes just such exchange of ideas across many social sciences. Since Springer has played a leadership role in promoting R, we are fortunate to have Springer publish this book. A Conference on Quantitative Social Science Research Using R was held in New York City at the Lincoln Center campus of Fordham University, June 18–19, 2009. This book contains selected papers presented at the conference, representing the “Proceedings” of the conference.

This book covers recent advances for quantitative researchers with practical examples from social sciences. The twelve chapters written by distinguished authors cover a wide range of issues--all providing practical tools using the free R software.

McCullough: R can be used for reliable statistical computing, whereas most statistical and econometric software cannot. This is illustrated by the effect of abortion on crime.

Koenker: Additive models provide a clever compromise between parametric and non-parametric components illustrated by risk factors for Indian malnutrition.

Gelman: R graphics in the context of voter participation in US elections.

Vinod: New solutions to the old problem of efficient estimation despite autocorrelation and heteroscedasticity among regression errors are proposed and illustrated by the Phillips curve tradeoff between inflation and unemployment.

Markus and Gu: New R tools for exploratory data analysis including bubble plots.

Vinod, Hsu and Tian: New R tools for portfolio selection borrowed from computer scientists and data-mining experts; relevant to anyone with an investment portfolio.

Foster and Kecojevic: Extends the usual analysis of covariance (ANCOVA) illustrated by growth charts for Saudi children.

Imai, Keele, Tingley, and Yamamoto: New R tools for solving the age-old scientific problem of assessing the direction and strength of causation. Their job search illustration is of interest during current times of high unemployment.

Haupt, Schnurbus, and Tschernig: Consider the choice of functional form for an unknown, potentially nonlinear relationship, explaining a set of new R tools for model visualization and validation.

Rindskopf: R methods to fit a multinomial based multivariate analysisof variance (ANOVA) with examples from psychology, sociology, political science, and medicine. Neath: R tools for Bayesian posterior distributions to study increased disease risk in proximity to a hazardous waste site.

Numatsi and Rengifo: Explain persistent discrete jumps in financial series subject to misspecification.