© 2015 – Chapman and Hall/CRC
328 pages | 42 B/W Illus.
Statistical Analysis of Questionnaires: A Unified Approach Based on R and Stata presents special statistical methods for analyzing data collected by questionnaires. The book takes an applied approach to testing and measurement tasks, mirroring the growing use of statistical methods and software in education, psychology, sociology, and other fields. It is suitable for graduate students in applied statistics and psychometrics and practitioners in education, health, and marketing.
The book covers the foundations of classical test theory (CTT), test reliability, validity, and scaling as well as item response theory (IRT) fundamentals and IRT for dichotomous and polytomous items. The authors explore the latest IRT extensions, such as IRT models with covariates, multidimensional IRT models, IRT models for hierarchical and longitudinal data, and latent class IRT models. They also describe estimation methods and diagnostics, including graphical diagnostic tools, parametric and nonparametric tests, and differential item functioning.
Stata and R software codes are included for each method. To enhance comprehension, the book employs real datasets in the examples and illustrates the software outputs in detail. The datasets are available on the authors’ web page.
Psychological Attributes as Latent Variables
Challenges in the Measurement of Latent Constructs
What Is a Questionnaire?
Main Steps in Questionnaire Construction
What Is Psychometric Theory?
Datasets Used for Examples
Classical Test Theory
Conceptual Approaches of Reliability
Reliability of Parallel and Nonparallel Tests
Procedures for Estimating Reliability
True Score Estimation
Item Response Theory Models for Dichotomous Items
Summary about Model Estimation
Item Response Theory Models for Polytomous Items
Taxonomy of Models for Polytomous Responses
Models for Ordinal Responses
Models for Nominal Responses
Estimation Methods and Diagnostics
Joint Maximum Likelihood Method
Conditional Maximum Likelihood Method
Marginal Maximum Likelihood Method
Estimation of Models for Polytomous Items
Graphical Diagnostic Tools
Infit and Outfit Statistics
Differential Item Functioning
Some Extensions of Traditional Item Response Theory Models
Models with Covariates
Models for Clustered and Longitudinal Data
Structural Equation Modeling Setting
Exercises appear at the end of each chapter.