Interaction effects in logistic regression / James Jaccard.

Format:

Book

Author/Creator:

Jaccard, James.

Series:

Quantitative applications in the social sciences ; no. 07-135.

Sage university papers. Quantitative applications in the social sciences ; no. 07-135

Language:

English

Subjects (All):

Logistic regression analysis.

Physical Description:

vii, 70 pages : illustrations ; 22 cm.

Place of Publication:

Thousand Oaks, Calif. ; London : Sage Publications, [2001]

Summary:

Oriented toward the applied researcher with a basic background in multiple regression and logistic regression, this book shows readers the general strategies for testing interactions in logistic regression as well as providing the tools to interpret and understand the meaning of coefficients in equations with product terms. Using completely worked-out examples, the author focuses on the interpretation of the coefficients of interactive logistic models for a wide range of scenarios encountered in the research literature. In addition, the author avoids complex formulas in favor of simple computer-based heuristics that permit the simple calculation of parameter estimates and estimated standard errors that will typically be of interest to applied researchers.

Contents:

Probabilities and Odds 2

The Logistic Regression Model 3

Categorical Predictors and Dummy Variables 5

Predicted Values in Logistic Regression 6

Interpretation of Coefficients 7

Probabilities, Odds, and Log Odds Revisited 9

Transformations of the Predictor Variables 11

Definition of Interaction 12

Hierarchically Well-Formulated Models 15

Product Term Analysis Versus Separate Logistic Regressions 17

2. Interactions Between Qualitative Predictors 18

Two-Way Interactions 18

Three-Way Interactions 24

3. Interactions Between Qualitative and Quantitative/Continuous Predictors 30

Two-Way Interactions With a Qualitative Moderator Variable 30

Two-Way Interactions With a Quantitative Moderator Variable 34

Three-Way Interactions 37

4. Interactions Between Quantitative/Continuous Predictors 42

Two-Way Interactions 42

Three-Way Interactions 44

5. Multicategory Models 46

Ordinal Regression Models 47

Multicategory Nominal Variables 50

6. Additional Considerations 53

Methods of Presenting Interaction Effects 53

Calculating Confidence Intervals 58

Calculating Coefficients of Focal Independent Variables at Different Moderator Values 59

The Bilinear Nature of Interactions for Continuous/Quantitative Variables 61

Partialling the Component Terms 63

Multiple Interaction Effects 63

Multicollinearity 65

Model Selection and Trimming 66

Transformations 67

Confounded Interactions 67

Computer Software 68.

Notes:

Includes bibliographical references (page 69).

ISBN:

0761922075

OCLC:

45059074