| Volume 2, Issue 2, 95-138, 2011 |
|Return to Issue List |
|Tracing the Interpersonal Web of Psychopathology: Dyadic Data Analysis Methods for Clinical Researchers|
|Pamela Sadler (a), Nicole Ethier (b) and Erik Woody (b)|
|(a) Wilfrid Laurier University|
(b) University of Waterloo
|Volume 2, Issue 2, 2011, Pages 95-138|
|Recent advances in dyadic data analysis techniques, which treat the dyad, rather than the individual, as the unit of analysis, offer great potential for clinical researchers studying psychopathology. Accordingly, the present article provides readers with a foundation for understanding how the web of interpersonal processes surrounding psychopathology can be modeled and analyzed. The authors start by describing why the analysis of dyadic behaviour may be particularly important for clinical researchers and how issues of dependence that lie at the heart of dyadic data may be productively studied. Next, they describe design issues to consider when studying the interactions of dyads, as well as different kinds of outcome and predictor variables and their data-analytic implications. They introduce the actor-partner interdependence model (APIM), and explain in detail how to estimate it using structural equation modeling (SEM) for both distinguishable and indistinguishable dyads. Extensions of the basic APIM to allow for moderation and mediation, as well as alternative dyadic models involving dyadic latent variables are also covered. Toward the end of the article, the authors describe various approaches for incorporating psychopathology into dyadic SEMs and provide a list of basic questions for clinical researchers to consider when setting up a dyadic model for data analysis. |
|Table of Contents|
What Happens to the Data when People Interact
Dependent Data: Kludges versus Opportunities
Some Basic Dyad Design Issues
Different Kinds of Dyadic Variables and their Implications for Data Analysis
Non-dyadic outcome variable.
Between-dyads outcome variable.
Within-dyads outcome variable.
Mixed outcome variable.
Between-Dyads, Within-Dyads, and Mixed Variables as Predictors
A Note on SEM versus MLM
Dyadic SEM Models
The Basic APIM
Estimation of the APIM for Dyads with Distinguishable Members
Tests of Sex Differences and the Empirical Evaluation of Distinguishability
Expanding the APIM: Between-Dyads Predictors and Moderation
Expanding the APIM: Mediation
Models Involving Dyadic Latent Variables
Estimation of the APIM for Dyads with Indistinguishable Members
Guidelines for using Dyadic Analyses in Studies of Psychopathology
Some Basic Questions in Setting up a Dyadic Model for Data Analysis
Appendix A: Downloadable Dataset and Amos Input Diagrams (.amw)
Appendix B: Using the Bias-Corrected Bootstrap in Amos to Evaluate the APIM with Mediation
Appendix C: Steps for the SEM Analysis of Indistinguishable Dyads and Examples
Example 1: APIM with mediation for indistinguishable dyads.
Example 2: Dyadic model with a mixed latent variable for indistinguishable dyads.
|Pamela Sadler, Ph.D. Department of Psychology, Wilfrid Laurier University, 75 University Avenue West, Waterloo, ON N2L 3C5, CANADA. |
|statistical dependence; dyadic dependence; distinguishable dyads; exchangeable dyads; interchangeable dyads |
|Received 3 Aug 2010; Revised 25 Oct 2010; Accepted 26 Oct 2010; In Press 5 May 2011 |