Academic journal article Journal of Marital and Family Therapy

Spatial Statistics for Dyadic Data: Analyzing the Relationship Landscape

Academic journal article Journal of Marital and Family Therapy

Spatial Statistics for Dyadic Data: Analyzing the Relationship Landscape

Article excerpt

Systemically oriented professionals are inspired by the complexity and diversity of couples they work with. Each couple is a whole greater than the sum of its parts. However, each individual in a committed relationship is also unique; pieces that are greater than a part of the whole. Thus, a systems perspective from a quantitative research paradigm ideally retains information from both partners while also creating the ability to model their relationship as a unique whole. An additional challenge comes via communicating the results of quantitative analysis in a way that captures the mosaic of relationships while being readily applied in a clinical setting (Wood, 2014).

There are two leading strategies to address couple data in the field currently, the actor-partner interdependence model (APIM; Kenny, Kashy, & Cook, 2006) and the common fate model (CFM; Gonzalez & Griffin, 2002; Ledermann & Kenny, 2012). This article presents a primer on a third approach to working with couple data based on spatial statistics (Wood, 2014). To better place spatial statistics within current statistical models, APIM and CFM models will be briefly summarized.

Actor-partner interdependence model has been used extensively in the field allowing for a better understanding of relationship processes. APIM focuses on the association of within-partner and between partner-level variables while statistically accounting for data nonindependence. In overly simplistic terms, APIM models remove the effects of what partners in relationships have in common, the "us" within the analysis, to focus primarily on "me and you" research questions. An APIM research question could be, "how does your relationship satisfaction effect your health and my physical health, and at the same time, how does my relationship satisfaction effect your physical health as well as my own?"

Alternatively, the CFM approach allows researchers to explore couple-level variables by modeling them as latent constructs that are manifest in each partner's observed scores. To oversimplify for illustration purposes once more, the CFM attempts to model "us" variables while removing "me and you" from the interpretation of the analysis. A research question utilizing a CFM approach may be, "how does our relationship quality impact our physical health." While it is possible, most CFM models also do not typically explore questions of within-couple nor between-couple differences.

Despite the power of APIM and CFM models, obstacles to bringing readily applicable research into the practitioner's office remain. We assert that part of the barrier to bridging research and practice is the presentation of statistical analyses. Results from APIM and CFM models are typically communicated in tables and/or diagrams in the form of regression coefficients, that is, a number designating the strength and direction of the relationship between two variables. In contrast, a couple therapist is concerned about the strength and direction of the relationship between the two people sitting in front of them - people are not variables. Therefore, it is imperative that the field has analytic models that readily connect the variable-level analyses with the couple in the clinician's office without needing to extrapolate from complex tables and figures.

Wood and Crawford (2012) introduced a visual heuristic to track therapy progress at the couple level while retaining information from each partner. Couple data were placed on a 2-dimensional plane with the X- and Y-axes representing wife and husband scores on any given relational questionnaire. This enables couples to be represented by a single point on a plane while critically maintaining each partner's unique experience of their relationship (e.g., the points on Figure 1). Mapping relationships on a 2-dimensional plane simultaneously model within dyad information at the point level whereas distances between points model between dyad information. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.