Measuring Satisfaction in Treatment Nodes and Dyads

Welcome back. In our previous lesson, we established a formal language for asking causal questions in an interconnected system. By defining estimands like direct, spillover, and total effects, we created a precise framework for investigating how an intervention impacts each member of the vee triad. This gave us the "how" of asking questions.

Today, our focus shifts to the "what." We will operationalize the core components of our research questions: the Intervention and the Outcome. This lesson addresses the learning outcome of specifying concrete treatment variables and a suite of measurable satisfaction outcomes for every person (node) and relationship (dyad) in the system. By the end of our session, we will have a clearly defined set of variables that transform our abstract research questions into a testable experimental plan.

1. Defining the Treatment: From Structure to Variable

In our context, the "treatment" isn't a pill or a simple on/off switch; it is the very structure of the relationship system. One of the most powerful and experimentally tractable structural variables in a polyamorous vee relationship is the living arrangement. This choice fundamentally shapes the allocation of time, energy, and shared experience, making it an ideal candidate for our primary independent variable.

The academic literature on polyamory confirms that cohabitation is a key differentiator between relationship types. For instance, the distinction between "primary" and "secondary" partners is often defined by whether they share a household. This provides an empirically grounded basis for our experimental conditions.

In a 'vee' relationship, the Pivot (central person) has two distinct dyadic relationships with the Wings, who are not in a romantic relationship with each other. Our treatment will manipulate the structure of this triad.

Let's define the living arrangement as a categorical treatment variable, with four distinct levels. For our triad of Pivot (P), Wing 1 (W1), and Wing 2 (W2), we can establish the following experimental conditions:

Condition A: Triadic Cohabitation. All three individuals (P, W1, W2) live in the same household. This represents a highly integrated, non-hierarchical living structure.
Condition B: P-W1 Dyadic Cohabitation. The Pivot and Wing 1 live together, while Wing 2 lives separately. This establishes a "primary-like" status for the P-W1 dyad based on cohabitation.
Condition C: P-W2 Dyadic Cohabitation. The Pivot and Wing 2 live together, while Wing 1 lives separately. This mirrors Condition B, but with the other wing.
Condition D: Dispersed Living. All three individuals live in separate households. This represents a "solo poly" or non-cohabiting model where no dyad is privileged by a shared living space.

In our experimental design, the triad will cycle through these conditions. The treatment vector $T$ for any given period will not be a collection of individual treatments, but a single value from the set ${A, B, C, D}$ that describes the state of the entire system. This approach directly embraces the reality of interference—a change in living arrangement is an intervention on the system, not just on one person.

2. Defining the Outcomes: A Multi-Level Satisfaction Dashboard

Just as the treatment is complex, so is the outcome. "Relationship satisfaction" is a latent construct that cannot be captured by a single number. We must develop a multi-dimensional "dashboard" of metrics to get a complete picture of the system's health. Furthermore, as research into non-monogamous relationships highlights, standard instruments designed for monogamous dyads are often insufficient. We need to build a bespoke measurement model.

A robust approach is to categorize our outcomes by the level of analysis:

Individual (Node) Outcomes: Measures concerning each person's personal experience.
Dyadic Outcomes: Measures reflecting the health of the two romantic relationships (P-W1 and P-W2).
Triadic Outcomes: Measures of the stability and dynamics of the entire three-person system.

A. Individual-Level Outcomes ( $Y_{P}, Y_{W 1}, Y_{W 2}$ )

For each person $i \in {P, W 1, W 2}$ , we can measure:

Global Life Satisfaction: A general measure of well-being, independent of relationship status.
Perceived Autonomy: The degree to which the individual feels they have control over their own life and decisions.
Psychological Stress: Standardized scales can measure levels of anxiety or depression.
Role Satisfaction: A specific rating of their satisfaction with their current role (Pivot or Wing) within the given living arrangement.

B. Dyadic-Level Outcomes ( $Y_{P W 1}, Y_{P W 2}$ )

This is where we measure the health of the two core romantic bonds. Research on polyamory provides excellent, validated scales for this. The following resource is a study that empirically compares primary and secondary relationships, and its measurement section offers a blueprint for our dyadic outcomes.

Perceptions of primary and secondary relationships in polyamory

This paper by Moors, Conley, and their colleagues provides a clear and empirically tested set of variables for evaluating relationship quality in a polyamorous context. We will adapt their measurement strategy for our own experiment.

Please read the Measures subsection within the "Method" section. As you read, focus on the specific constructs they measured and how they were operationalized. Pay particular attention to the items used for: Investment, Satisfaction, and Commitment: These are the core components of relationship health, based on the well-established Investment Model Scale. Communication: Note how they measure both the frequency of communication on specific topics and the perceived overall quality. Sexual Activity: A simple but powerful behavioral measure. Acceptance and Secrecy: These capture the social context and internal dynamics of the relationship.

Based on this, for each dyad (e.g., P-W1), and from the perspective of each person in that dyad, we will collect data on:

Satisfaction: "I feel satisfied with our relationship."
Commitment: "I am oriented toward the long-term future of our relationship."
Investment: "I have put a great deal into this relationship that I would lose if it ended."
Communication: A composite score based on frequency and quality.
Sexual Satisfaction: A rating of satisfaction with the sexual aspect of the relationship.
Conflict Level: A measure of the frequency and intensity of disagreements.

C. Triadic-Level Outcomes

Finally, we need measures that capture the functioning of the vee system as a whole.

As this diagram of interference in a social network illustrates, an individual's outcome (Y) is affected not just by their own treatment but by the treatments and covariates of their partners. Our triadic-level outcomes aim to measure the health of this entire interconnected system.

Some of these can be composite scores, while others are unique emergent properties of the triad:

Triadic Stability Index: A composite measure calculated from the commitment scores across both dyads. It represents the perceived likelihood of the three-person system persisting.
Compersion/Jealousy Balance: Compersion, the experience of joy from a partner's happiness with another, is a key concept in polyamory. We can measure the prevalence of compersive versus jealous feelings for each member regarding the "other" dyad.
Perceived Equity: A rating from each member on the fairness of the distribution of time, attention, and resources within the triad.
System-Wide Secrecy: Whereas dyadic secrecy is about hiding things from outsiders, this measures any information being withheld within the triad (e.g., W1 hiding something from P and W2).

3. The Need for Custom-Tailored Variables

The process we've just completed—designing a bespoke set of treatments and outcomes—is not just an academic exercise. It is a necessary response to the limitations of existing research tools, which are often "mononormative." The following dissertation highlights this exact point, arguing that lumping different polyamorous configurations together is a methodological flaw and that new instruments are needed.

[PDF] Assessing Contributing Factors of Relationship Satisfaction in ...

This doctoral dissertation by Dr. Timea Farkas directly addresses the problem of using one-size-fits-all measures for diverse relationship structures. Its conclusion validates our approach of designing a specific measurement plan for the vee triad.

Please read the Recommendations and Implications sections of the final chapter. Focus on these key points: The call to avoid lumping different non-monogamies together, specifically mentioning the need to differentiate vee’s from other structures. The author's reflection that participants in their study felt existing instruments failed to capture their experience. The argument that new instruments are crucial and must be flexible enough to accommodate the diverse realities of these relationships.

By specifying our treatments (living arrangements) and our multi-level outcomes (the satisfaction dashboard), we are directly answering this call for more rigorous, tailored research. We can now insert these operationalized variables into the research questions from our last lesson. A vague question like "What is the spillover effect of a treatment?" becomes a precise, testable hypothesis:

"What is the effect on the dyadic commitment score in the P-W2 relationship ( $Y_{P W 2 - C o mmi t m e n t}$ ) when the system moves from a Dispersed Living arrangement (Condition D) to a P-W1 Dyadic Cohabitation arrangement (Condition B)?"

This question is now fully specified and ready for experimental design.

Conclusion

In this lesson, we have built the substantive core of our experiment. We moved from abstract concepts to concrete, measurable variables that are grounded in the lived reality of polyamorous relationships and informed by existing academic research.

The key takeaways are:

Treatments can be structural. In social systems, the "intervention" is often a change to the environment or rules of the system, such as living arrangements.
Outcomes must be multi-dimensional and multi-level. To understand a complex system like a relationship, we need a dashboard of indicators measuring individuals, the bonds between them, and the system as a whole.
Customization is essential. Applying off-the-shelf tools to novel social structures can lead to invalid conclusions. Careful specification of variables tailored to the population is a cornerstone of good experimental design.

We now have the "what" of our experiment. In our next lesson, we will turn to the "how" and "why" by mapping the causal pathways that connect our treatments to our outcomes. We will discuss potential mediating variables (like feelings of security or autonomy) and shared confounders (like household financial stress) and begin constructing the full causal graph of the system.

Recommended resources