ISBN online: 978-80-244-6034-5 | DOI: 10.5507/ff.21.24460345

Linear statistical models in psychology

Daniel Dostál

Different kinds of linear regressions are useful in solving a wide range of problems encountered in psychology and other areas of empirical research. Procedures belonging to this family find their application in the processing of experimental data, in the analysis of questionnaire surveys and, for example, in the development of test norms. This text is intended for students who are familiar with the basic procedures of quantitative data analysis - descriptive statistics and statistical hypothesis testing. The aim of this textbook is to integrate and present this knowledge into the context of statistical modelling. The book describes both basic topics and specialized procedures of working with linear models. The reader is introduced to regression coefficients, indicators of model fit, handling nominal regressors and interaction terms. Several chapters are devoted to tests of statistical significance, construction of confidence bands, assumptions testing, and model diagnostics. Among more advanced techniques, the book covers the investigation of curvilinear dependencies, lognormal regression, stepwise and hierarchical regression, and mixed-effects models. The textbook also includes a set of data files for practicing the methods described.

1. edition, online: 2021, publisher: Univerzita Palackého v Olomouci, Křížkovského 8, 771 47 Olomouc