Over the last 20 years, so-called regression-based normative methods have become increasingly popular. In this approach, regression models for the mean and the residual variance structure are used to derive the normative data. The regression-based normative approach has some important advantages over the traditional normative approach, e.g., it allows for deriving more fine-grained norms and typically requires a substantially smaller sample size to derive accurate norms.
This book focuses on regression-based methods to derive normative data. The target audience are psychologists and other researchers in the behavioral sciences who are interested in deriving normative data for psychological tests (e.g., cognitive tests, questionnaires, rating scales, etc.). The book provides the essential theoretical background that is needed to understand the methodology, with a strong emphasis on the practical/real-life application of the methodology. To this end, the book is also accompanied by an open-source software package (the R library Norm Data) that is used to exemplify how normative data can be derived in several case studies.
Innehållsförteckning
General introduction.-The R programming language.- Normative data accounting for a binary independent variable.- Assumption of the normal error regression model.- Normative data accounting for a non-binary qualitative independent variable.- Normative data accounting for a quantitative independent variable.- Normative data accounting for multiple qualitative and/or quantitative independent variables.- Quantifying uncertainty in regression-based norms
Om författaren
Wim Van der Elst (Ph D) has a background in psychology and statistics, and is currently employed as a statistician in the pharmaceutical industry. He has published extensively on regression-based normative data, psychometrics, psychological assessment, and the statistical evaluation of biomarkers. He is also the lead programmer of several R libraries.