Introducing Multivariate Data Analysis

See the map before selecting a technique

The contents separate data preparation, interdependence techniques such as EFA and clustering, and dependence techniques including regression, MANOVA, discriminant analysis and logistic regression. Later sections introduce SEM, CFA, structural models, advanced topics and PLS-SEM.

This organization is more useful than asking first whether data can run in a particular program. Grouping indicators differs from grouping respondents; modeling a binary outcome differs from modeling a continuous one.

Read preparation as part of the analysis

Data examination establishes what information is available and how processing choices affect later work. Readers developing measurement models can also use the EFA–CFA–SEM progression to distinguish exploration, measurement assessment and structural questions. It does not imply that every study must run the entire sequence on the same data.

Exercise created by Lê Minh: write three hypothetical service-research questions about item structure, respondent segments and loyalty predictors. Assign each to a methodological branch and specify required inputs and outputs. Being able to run all three does not make their purposes interchangeable.

Then study one branch deeply, including assumptions and interpretation. Explain the choice without relying only on the fact that previous papers used it.

A text and a reference

Graduate students and researchers seeking connections between techniques may benefit most. The inspected eighth edition carries a 2019 copyright, while release information can appear in 2018. Check the actual edition rather than applying the date of another printing. The worthwhile outcome is a defensible map of choices before producing another results table.

Choose a companion reading

Continue with: Using Multivariate Statistics (Seventh edition · English); Principles and Practice of Structural Equation Modeling (Fifth edition · 2023 · English).

Source and edition