15.3 Basic Definitions
15.3.1 Distinguishability
An important feature of dyadic data is whether or not the dyad members are distinguishable. Note, this is both a theoretical and empirical question to consider.
Importantly, dyad members are indistinguishable if there is no systematic or meaningful way to order the two scores.
Distinguishability is a critical distinction in dyadic data analysis because it dictates the analytic techniques one can use. Perhaps unintuitively, it is considerabely easier to analyze dyadic data when the members are distinguishable.
Note: It is generally not advised to pick some arbitrary feature to make dyad member distinguishable (e.g. randomly selecting the first individual listed in a data file). This decision introduces a component to the data which is not found in reality and may lead to innacurate conclusions. A variable or feature that distinguishes dyad members should be meaningful to the analyses.
Distinguishable Dyas
- parent and child
- patients and careegivers
- pet owner and pet
- older and younger siblings
Indistinguishable Dyas
- coworkers
- twins
- friends
- archenemies
15.3.2 Variable Types
The nature of the predictor variables included in an analysis plays a critical role in choosing the appropriate modeling approach. Here we outline three important classes of predictor variables on might wish to include in an analysis.
15.3.2.1 Between-Dyads Variables
Data from a between-dyads variable differs from dyad to dyad, but does not differ within a dyad. This means each member of the dyad has the same value for a between-dyads variable.
For example, in a study on couple’s relationship satisfaction the variable relationship length would be a between-dyads variable.
15.3.2.2 Within-Dyads Variables
In contrast to between-dyads variables, within-dyads variables can differ between two members of a dyad. Importantly, when averaged across the two dyad members, all dyads should have the same average score.
Examples of within-dyad variables are family role in a study on different family members (e.g. father and son). Or the experimental role when one person is asked to persuade another person.
Dyad members are distinguished by a within-dyads variable, however, whether or not this is a meaningful distinction is a theoretical question.
15.3.2.3 Mixed Variables
The last type of predictor variables are referred to as mixed independent variable. Here, variation exists both within and between dyads.
Age is an example of mixed predictor variable. In a given study, dyad members may have different ages, and certain dyads may be older than other dyads.
Most predictors and outcomes in dyadic research are mixed.