Making Sense of Factor Analysis

Marjorie A. Pett, MStat, DSW, is a Research Professor in the College of Nursing at the University of Utah, Salt Lake City, Utah, having been on the faculty since 1980. By her own admission, she is a “collector” of academic degrees: BA (Brown University), MS in sociology (University of Stockholm, Sweden), MSW (Smith College), DSW (University of Utah), and MStat (Biostatistics) (University of Utah). Dr. Pett has a strong commitment to facilitating the practical application of statistics in the social, behavioral, and biological sciences, especially among practitioners in health care settings. She has designed and taught graduate courses to students from a variety of disciplines at the beginning and advanced levels, including research design and data management, parametric and nonparametric statistics, biostatistics, multivariate statistics, instrument development, and factor analysis. She has tried to approach the teaching of statistics with humor and from a clinician’s perspective and has been the recipient of several distinguished teaching awards both at the College and University levels.Her most recent research interests include the development of client-centered assessment tools and interventions to evaluate and enhance health-related quality of life (HRQoL) for persons with intellectual disabilities. She is the author of numerous research articles and chapters, and is an author of the Sage publication, Making Sense of Factor Analysis: The Use of Factor Analysis for Instrument Development in Health Care Research. When not engaged in research, writing, or teaching, Marge is a (now retired) state soccer referee, devotee of tennis, an avid (high handicap) golfer, student of Italian and French, reader of mystery novels, grandmother to three, mother to two, and wife to (only) one. Dr John Sullivan has been a professor of management for over 26 years at San Francisco State University. His specialty is HR strategy and designing world class HR systems and tools for Fortune 200 firms. He has worked with over 200 different businesses and organizations in more than 30 countries around the world as a speaker or advisor.He has written a weekly column for ERE for over eleven years. Overall, he has written ten books, dozens of white papers and over 700 articles. He was the chief talent officer for Agilent (the 40,000+ employee HP spin off). He has appeared on the CBS and ABC national nightly news, CNN and in various publications including Fortune, the Economist, CIO, BusinessWeek, the WSJ, the Washington Post, Money, Time and every major HR magazine. Fast company called him the Michael Jordan of hiring. He was listed among the 40 most influential people in HR. Tom Peters cites and utilizes his work in his latest book Re-Imagine.

Absolutely brilliant - descriptive, easily read and motivating. Good examples from health and particularly a good description of different rotations, why we do FA, explains the print outs. All my colleagues want to borrow it too! Good for students doing questionnaires with a large number of participants.

• 1. An Overview of Factor AnalysisCharacteristics of Factor AnalysisExploratory Vs. Confirmatory Factor AnalysisAssumptions of Exploratory Factor AnalysisHistorical Developments of Factor AnalysisUses of Factor Analysis in Health Care ResearchDecision-Making Process in Exploratory Factor Analysis2. Designing and Testing the InstrumentTypes of Measurement and FrameworksThe Use of Latent Variables in Instrument DevelopmentIdentifying Empirical Indicators of Latent VariablesUsing Qualitative Research Methods to Identify Empirical IndicatorsAdditional Qualitative Approaches to Identifying Empirical IndicatorsDevelopment of the InstrumentScoring the InstrumentPilot Testing the InstrumentDetermining the Number of Subjects3. Assessing the Characteristics of MatricesCharacteristics and Types of MatricesTests of MatricesReview of the Process4. Extracting the Initial FactorsEvaluating the Correlation MatrixSources of Variance in Factor Analysis ModelsDetermining the Factor Extraction MethodSelecting the Number of Factors to RetainComparing the Two-Factor Solution Using PCA and PAF5. Rotating the FactorsAchieving a Simple StructureTypes of RotationsMapping Factors in Geometric SpaceOrthogonal RotationsOblique RotationsComparing the Orthogonal and Oblique SolutionsAdvantages and Disadvantages of the Oblique SolutionChoosing Between Orthogonal and Oblique RotationsSummary of the Process of Rotations6. Evaluating and Refining the FactorsEvaluating and Refining the FactorsAssessing the Reliability of an InstrumentEvaluating the Internal Consistency of an InstrumentEstimating the Effects on Reliability of Increasing or Decreasing ItemsCronbach′s Coefficient AlphaAssessing Reliability Using Cronbach′s Alpha: A Computer ExampleTwo Additional Reliability Estimates: Temporal Stability and Equivalence7. Interpreting Factors and Generating Factor ScoresInterpreting the FactorsNaming the FactorsInterpreting and Naming the Four Factors on the CGTS ScaleDetermining Composite Factor Scores8. Reporting and Replicating the ResultsWhen to Report the ResultsWhat to Include in the ReportAn Exemplar of a Published ReportReplicating the Factors in Other StudiesConclusionsAppendix A: Concerns About Genetic Testing ScaleAppendix B: SAS Commands and Generate OutputAppendix C: Output for 20-item CGTS ScaleAppendix D: Tables for the Chi-Square and Normal DistributionsAppendix E: Unraveling the Mystery of Principal Component ExtractionReferencesIndexAbout the Authors