Scaling Methods for Social Science

3L Scaling Methods for Social Science: Measuring Patterns and Preferences in Surveys and Behavior, Instructor: Prof. Royce Carroll, 2020 ESSEX SUMMER SCHOOL

This course focuses on methods to discover, understand and visualize latent patterns in data and is especially suited to students with projects using survey data and other forms of relational data. The course introduces students to measurement theory and methods of scaling techniques, including Multidimensional Scaling,Ideal Point Estimation, and Item Response Theory. The first part of the course will provide an overview of the foundations of these techniques and introduce students to the most common methods for “spatial” analysis of survey and behavior data. The course will demonstrate how to interpret, measure and visualize latent dimensions of data via a variety of scaling methods using the open-source programming language R. The course will also discuss a range of applications these methods to social science studies of relational and perception data derived from elite behaviour and surveys. The course concludes with discussions of the most recent advances in the field, including applications for text analysis, and practical advice for those seeking to use such methods in social science research, relevant to the students enrolled.

The course first covers how to analyse data from scales found in surveys, focusing on surveys that ask respondents to place themselves and / or stimuli on issue or attribute scales. The course begins with approaches to scaling to generate bias-adjusted and latent spatial data from survey responses, such as the Aldrich-McKelvey scaling and ‘Basic Space’ scaling with Anchoring Vignettes as methods for addressing perceptual bias in the form of “Differential Item Functioning.” The course next examines similarities and dissimilarities data and covers multidimensional scaling (MDS) with a focus on the SMACOF optimization method implemented in R as well as Bayesian applications to Metric Multidimensional Scaling. Next, the course covers unfolding analysis of rating scale data from surveys such as favorability scales for stimuli such as politicians or social groups. Finally, the course provides an overview of IRT and ideal point estimation, generally focused on binary choice data, which includes those used in ‘roll call voting’ analysis of elite behavior in parliaments and courts. Here we will cover Poole and Rosenthal’s W-NOMINATE and Poole’s Optimal Classification unfolding method, as well as a variety of Bayesian analysis techniques for binary and ordinal choice data using Item Response Theory (IRT). An extensive range of Bayesian techniques are discussed, including Bayesian Aldrich-McKelvey Scaling, and Ordinal and Dynamic Item Response Theory (IRT). The course will conclude with several advanced methods that extend the core themese.

This course will enable students to derive latent spatial preference information and/or dimensional structure from various types of survey and behavior data, which is applicable to a wide range of social science applications, academic and non-academic alike. Consumers of research based on these methods will also benefit from a deeper understanding of this type of methodology, its potential and its limitations.

Schedule

Lecture / Lab: 14:15 – 17:45 Monday 10th August to Friday 21st August 2020

Packages

You will need to lastest installation of R (preferably version 3.6 or above) and RStudio. You also need to install the asmcjr package and other relevant scaling packages along with this course.

Install `asmcjr` package in R environment

You can install using the install_github() function from the devtools package. asmcjr requires compilation, so Windows users will have to install Rtools first. For Mac users, you need to make sure you have already installed latest GNU Fortran(gfortran 8.2) and Xcode Developer Tools. Visit asmcjr for further instructions.

install.packages("devtools", dependencies=TRUE)
library(devtools)
devtools::install_github("uniofessex/asmcjr")

You should also install JAGS, which will be used in several contexts, along with the package rjags

Note: lecture slides will be added after each lecture, reading materials will be available beforehand

The following scaling packages and relevant toolkits are used by this course:

Lectures	Packages
01 - An Introduction / R	`basicspace`
02 - Issue Scale I	`asmcjr`, `smacof`
03 - Issue Scale II	`asmcjr`, `basicspace`
04 - Issue Scale III	`asmcjr`, `smacof`, `rgenoud`,
05 - Similarities Data and Agreement Score	`asmcjr`, `rgenoud`,`smacof`
06 - Binary Ideal Point Estimation (W-NOMINATE)	`asmcjr`, `wnominate`, `anominate`
07 - Non-parametric Ideal Point Estimation (OC and OOC)	`asmcjr`, `oc`,`ooc`, `MCMCpack`
08 - IRT and Bayesian MCMC	`asmcjr`, `MCMCpack`, `pscl`
09 - Dynamic and Ordinal IRT	`asmcjr`, `ooc`
10 - Scaling Text and EM IRT	`asmcjr`, `emIRT`