Imagining prevention: the case of violence against women

A few years ago I was working at a research centre and a senior researcher had some money to develop a grant application.  I’d moved there from about ten years in practice in health promotion and I was still figuring out how research practice actually works. So when I was given the brief to do a literature review on what works in addressing racism and preventing violence against women, I didn’t think to ask for a narrower scope.


Alaskan flood plain (A.H. Brooks, USGS 1902)

In one way it was incredibly productive to get such a challenging brief, because it led me to develop a particular theory about how social change happens — one I was able to draw on in my next two years of work with the W3 project, and it has informed my writing on topics like why we should fund prevention technologies for sexually adventurous gay men.

(By the way, if anyone can recommend a good book on how to write up theory, I need it!  That’s where I got stuck trying to deliver that massive literature review.)

Along the way I got to spend some time thinking about Victoria’s strategy for the prevention of violence against women (PVAW), and all the published and grey literature that informed its development over a decade (see Sue Dyson’s account of it here).

Although in health promotion we talk a lot about wanting policy-making to reflect social justice values, I think the Bracks/Brumby Labor government Victoria made a tactical error in embedding their Fairer Victoria meta-strategy so deeply in health and social welfare policy.  Fairer Victoria was also, effectively, their election platform, so when there was a change of government, a whole raft of important policies – on men’s health, women’s health, and prevention of violence against women – got ‘pulled’ by the incoming government.  Some of them, like the men’s health policy, were never seen again, but Minister for Mental Health and Community Services, Mary Wooldridge, got the prevention of violence against women strategy revised and reissued within about a year, with much of the substance carried over (despite different emphases on e.g. the role of the family unit).

However, both strategies reflect a particular causal model of prevention that seems to predominate in public health — the idea of upstream intervention, and a typology of prevention that divides it up into primary (preventing incidence), secondary (mitigating harm) and tertiary (managing chronic effects) stages.  In the literature on PVAW this gets described as ‘the public health approach’ to contrast it to a crimino-legal model where it’s seen as a matter for the police to respond to when it occurs.

This post came about because I saw and responded to a tweet that called for more intervention in education to prevent violence against women (‘primary’) rather than crisis accommodation (‘secondary’).  (I was sliding into her mentions, etc, missing the point, etc.  Because what could a man know about family violence.)

There are problems with this approach.  One is a pragmatic issue of policy-making that is well-known in disease prevention — it’s that politicians want to solve the problem of family violence, so they’re enthusiastic about primary prevention and see secondary and tertiary prevention as ‘prevention failure’.  And politicians don’t want to pay for failure.

The second problem is the way both frameworks (upstream intervention and the typology) conceptualise the causation of the problem by posing a single hypothetical case and a linear logic of cause and effect.  It’s easy to see this if we visualise them:


‘Care’ as the opposite of ‘prevention’ in a linear model (source)

This framework — treating violence against women as akin to chronic disease — obscures some important patterns in the causal history of intimate partner violence: the most important being a pattern of escalation.  On this view, the most important predictor of future assault or homicide is past use of control or violence.

If we think in terms of multiple episodes of violence rather than a single progressive course of disease, primary prevention of future homicide may involve effective secondary prevention in an earlier episode.

But the typology encourages us to think of primary and secondary prevention as separate and makes it harder to think about those cycles of behaviour in the causation of VAW.  In the W3 project we explored the use of causal loop diagrams (wiki) to visualise and theorise the cyclical causal relationships that contribute to complex social problems.

A similar criticism can be made of the idea of upstream intervention.  It’s imagistic — it paints a mental picture that obscures more than it clarifies.

upstream intervention dentalcare dot com

Upstream intervention in dental health for children (source)

This conceptual image is incredibly prevalent in public health discourse.  It’s a little parable that confirms the politicians’ preference for primary prevention.  And if we look at an image of an actual river system we can see why it’s unhelpful:


What is a watershed? (source)

This is what lies upstream.  A river doesn’t start from a single point and drain to a single point: it has a branching structure at both ends — the watershed and the river delta.*

That’s not a bad metaphor for the many different social processes that contribute to the point, downstream, where we first identify a problem. We could use this as a metaphor for the different knowledge practices (e.g. epidemiology, psychology, social work, sociology, cultural studies, gender studies) that have a really deep knowledge of their particular ‘neck of the woods’ around a particular tributary that contributes downstream.

Public health needs to develop its ability to draw on all of these knowledges in a genuinely multi-disciplinary way — without relying on simplifying metaphors to handle complexity.  When the different causal tributaries combine, a systemic view of causation predicts the result will be more than the sum of its parts.  The experts on how that combination works are (a) service providers supporting survivors of violence and (b) the survivors of violence themselves.  But to actually tell that story — to ‘unbraid’ the complex causation of lived experience and explain it in a linear narrative — is incredibly challenging.  In a nutshell, that’s the challenge of writing/speaking after trauma.

* Except when we wall the river in with concrete as flood prevention, which we’re now learning drastically increases the likelihood of flooding.  This metaphor works, too.  When funding is stripped from the different services available to women facing abusive and violent relationships, they are increasingly channelled into an over-stretched, one size fits all model — the concrete channel — and overflow might represent their attrition from the system.



5 thoughts on “Imagining prevention: the case of violence against women

  1. At a guess Brooks Range was named after AH Brooks. If you want more reasons to doubt the upstream metaphor use Google Images > Brooks Range to show other limitations: sluggish bends, swamps, ox-bow lakes. Though you could do something with swamps!

  2. Another piece of succinct and well-written public health practice analysis, thank you Daniel.
    This book might be helpful: Jackson, A. Y., & Mazzei, L. A. (2012). Thinking with theory in qualitative research: Viewing data across multiple perspectives. Oxon, UK; New York, NY: Routledge. It was Megan McPherson who got me onto it.

  3. In terms of getting across the importance of seeing cause as multiplicity not unity, I think the kind of story telling you have done in other areas (e.g your recent telecom tale) works particularly well to raise awareness. And story-sharing across disciplinary boundaries potentially provokes new insights. So: how to get more space for public health researchers and practitioners in many disciplines to hear and appreciate each other ‘s stories?

Comments are closed.