Community Case Study: Introduction | Social Activism/Non-Profit

Social Activism/Non-Profit

Modeling Domestic Violence as a Dynamic Process

Problem Statement: Sociological scholarship on the dynamics of domestic violence typically relies on traditional social scientific methods to identify personality traits that characterize a male batterer. A sociologist colleague of ours at Trinity College was concerned that profiling traits with correlations mistakenly overstates “static” variables. Such an approach shortchanges critical “dynamic” issues that might trigger an initial action and could define how that violence influences the relationship over time. We convened a group of active domestic violence practitioners to examine the broad question: “Might system dynamics enhance the understanding and bolster the efforts of practitioners actively combating domestic violence? “

Approach: We implicitly drew upon the structure of our Ladder of Engagement to:

• Engage the group in sharing what they know of the dynamics underlying 1) an initial or singular act of violence, as well as, 2) subsequent acts of violence;
• Organize and integrate that knowledge to develop a shared understanding of the feedbacks that influence a “typical” male abuser’s thinking and subsequent actions; and finally,
• Use the insights generated by that understanding to help both academic scholars and practitioners to design and evaluate intervention strategies with which to influence these behaviors?

A copy of a paper based on this work and presented at the 1999 Annual Meeting of the American Society of Criminology can be downloaded here.

---------------------------------------------------------------

KNOWLEDGE

What do we know about the “behavior(s)” of the system?

We focused on two broad areas of knowledge:

• The first was collecting and reviewing the lengthy list of static, “personality traits” identified by scholars and experts as predisposing factors that correlate with who “was” or “was not” likely to be(come) a male batterer.
• The second and more critical focus of the group was upon the powerful yet seemingly irregular and unpredictable dynamic cycles of repeated violent acts undertaken by batterers, many of whom exhibited few, if any, of the profile characteristics.

Initial conversations focused on identifying a vehicle through which “acts of violence” could be linked with “predisposing” and “dynamic” circumstances. Our initial efforts, following a fairly traditional path within domestic violence studies, focused on the dynamics (growth and dissipation) of anger and yielded the initial conceptual map presented below. Static predispositions, aggregated in “Probabiolity He Could Do It,” would set the stage; if “His Tension or Anger” grew to a critical level, “An Act of Domestic Violence” would occur.

However as the practitioners explored the dynamics and implications of this map, they realized that “anger” was misplaced as a central factor in this domestic violence model. Both the experience of their own practice and scholarship identified instances of violence in the absence of any sign of anger. If not anger, then what factor did belong at the center of our conceptual map?

That question, which the practitioners had never previously asked, led us in turn to identify and focus on a new element of the system, “his control.” The dynamics of that control, or eventually what we came to recognize was more properly “His Perception of His Total Accum(ulated) Control,” was initially described in the simple stock/flow map below.

Armed then with a “knowledge” that the dynamics of “perceived control” was more likely to link with violence than “his tension or anger,” the challenge then focused on filling in the “gaps” that explained how and where perceptions of control would change over time.

UNDERSTANDING

What “drives” the behavior of the system?

Once we had shifted our focus from “anger” to “control,” our efforts shifted to using the tool of Stock & Flow Concept Mapping to identify and come to further shared understanding of what in the system was controlling the interaction of “control” and “violence.” More detail, for those who are interested, can be found in the 1999 paper. The figure below offers a quick summary of the structure we built collaboratively to provide a framework for probing and testing our respective understandings and experiences.

Note the following points that represented ‘discoveries’ for the practitioners who 1) had to describe their mental models to other practitioners and to sociologically-naïve system dynamicists, and 2) were able to use the tools of system dynamics to facilitate that communication:

• “Anger” is not the central driver, having been replaced by a more generic and richer element , “his need to act,” that is affected by three dynamic influences. These are illustrated by the connectors (thin arrows) leading to “his need to act.”
• “Perceived Control” is one of the principle influences on his need to act. Not only is the total accumulation of perceived control important, but the potential batterer’s perception of how that control is changing, also plays a significant role;
• Physical violence is not the sole focus of this model, but is replaced by “intimidation” that includes a variety of physical as well as psychological abuses. Intimidation is both the result of his actions and is a major driver of her compliance to his need for control; and
• The original list of predispositions to violence has been joined by two new critical factors, a sense of male entitlement and a set of coping skills that can change in response to the man’s real or perceived situation.

Collectively, these influences constitute a rich array of feedback loops and contain delays of varying durations, all of which help to explain why domestic violence is a problem that has proven so difficult to address.

INFLUENCE

Can this understanding be used to generate better policies to manage the system?

The concept map that was developed through the collaboration of two system dynamicists, one academic sociologist, and three practitioners in domestic violence intervention provided a powerful framework with which all the participants could more clearly and explicitly present their own observations and beliefs and to collectively explore the implications of those, often novel, insights. By introducing and exploring these new relationships, the practitioners were enabled, indeed were obliged, to consider new factors and influences that clearly were enriching the approaches they could utilize in their work with their respective clients from among the battered and battering populations.

Actually converting the Stock & Flow Concept Map into a quantified and runnable Computer Simulation would be a major advance in this work. We reached a point where insights and relationships were being clarified and adjusted in novel and exciting ways for the practitioners. A Computer Simulation would allow that powerful growth to expand to the point where the efficacy or leverage of specific interventions could be tested. In any system as complex as this, we are severely limited in our ability to intuit the results of various policies; objective, quantitative modeling helps relieve those limitations. The field of human behaviors in general is an exceptionally rich domain for this modeling approach to pay great dividends.