Introducing Using Multivariate Statistics
When an analysis produces unexpected coefficients, trying another method can seem attractive. Barbara G. Tabachnick and Linda S. Fidell’s Using Multivariate Statistics encourages a more careful starting point: reconsider the question, data and requirements first (Tabachnick & Fidell, 2019).
Start reading
Using Multivariate Statistics
Pearson · Seventh edition · EnglishISBN 9780134790541Original book cover; the language and edition are identified alongside it.
A broad map with a preparatory foundation
The contents include technique selection and screening before regression, covariance analysis, MANOVA, discriminant and logistic methods, survival analysis, PCA, factor analysis, SEM, multilevel models and time series. The range supports selecting a suitable branch, not trying everything and retaining a favorable result.
If missing data, unusual cases or assumptions are the immediate concern, read screening before returning to the technique. Keep a processing record so that justified changes remain distinguishable from attempts to improve the appearance of output.
Examine sensitivity rather than chase significance
An unusual observation needs investigation, not automatic deletion. Check errors, context and influence. No universal exclusion rule is proposed in this introduction.
Exercise created by Lê Minh: document a problem in illustrative data, make one justified processing change and repeat the same analysis. Compare sample size, coefficients and interpretation as well as p-values. The purpose is understanding sensitivity, not obtaining significance.
The SEM chapter is contributed by Jodie B. Ullman. She should not therefore be described as a coauthor of the entire book; cite the chapter appropriately when using its specific contribution.
A useful notebook entry links a diagnostic finding to a decision and its consequence. For example, distinguish evidence of a data-entry error from an unusual but plausible response before considering any processing change.
Edition and readers
This article uses the seventh edition carrying a 2019 copyright, ISBN 9780134790541. Later digital or regional releases can carry other dates. Readers with basic statistics may use it as a demanding but practical reference. A chapter addressing an actual difficulty can provide something more durable than a new procedure: a reasoned explanation of the analytical choice already being made.
Choose a companion reading
Continue with: Multivariate Data Analysis (Eighth edition · English).
Source and edition
Tabachnick, B. G., & Fidell, L. S. (2019). Using multivariate statistics (7th ed.). Pearson.