I finally got round to reading The Power of Positive Deviance, in which management guru Richard Pascale teams up with the two key practitioners – the Sternins (Jerry and Monique) – to analyse over 20 years of experience in developing the Positive Deviance (PD) approach, building on their pioneering experience working on malnutrition in Vietnam in the early 1990s. Since then, PD approaches have spread and morphed in 50 countries North and South, and been used on everything from reducing gang violence in inner city New Jersey to trying to reducing sex trafficking of girls in rural Indonesia.
The starting point of PD is to ‘look for outliers who succeed against the odds’ – the families that don’t cut their daughters in Egypt, or the kids that are not malnourished in Vietnam’s poorest villages. On any issue, there is always a distribution of results, and PD involves identifying and investigating the outliers.
But it also matters who is doing the looking. The community must make the discovery itself – it’s no use external ‘experts’ coming in, spotting PD and turning it into a toolkit. Whether it is US hospital staff taking responsibility for tackling MRSA, or poor villagers selling their daughters in Indonesia, this gets round the defensive ‘immune defence response’ that solutions peddled by experts aren’t relevant.
The approach itself – recognizing for any given problem, someone in the community will have already identified a solution – is hugely energizing. Drop out rates in Misiones Province, Argentina were awful, but teachers and principals were initially hostile to being criticism and tried to blame the parents. All that started to change when the facilitators moved from the problem to asking the ‘somersault question’ why were drop out rates much lower in some schools. Teachers then agreed to ask parents to join in, who rapidly identified that the key was teacher attitudes to parents – the positive deviant teachers were negotiating informal annual ‘learning contracts’ with parents. That approach spread and drop out rates in test schools dropped by 50%.
PD means learning to ‘spot the novel in the familiar’ – eg noticing that a mother of an unusually well nourished child dipped the ladle deeper into the family pot and so scooped up solid veg for the kids’ bowls led to a big breakthrough on child nutrition in Bolivia. Families with lower infant mortality in a remote corner of Pakistan put their newborns on a blanket, rather than the bare earth. ‘Footnotes turned out to define the entire plot’.
PD involved huge attention to process –looking at the how, not the what, as in the Bolivian example. What happens next depends on what the authors call ‘social proof’. Rather than experts moving in, recording the PDs and turning them into a powerpoint, communities need to see (and preferably do) for themselves in order to internalize behaviour change.
In all this, experts are often part of the problem: ‘those eking out existence on the margins of society grasp the simple elegance of the PD approach – in contrast to the sceptical consideration of the more educated and/or privileged. Uptake seems in inverse proportion to prosperity, formal authority, years of schooling and degrees hanging on walls.’ Experts have to give up their role as ‘answers looking for problems to solve’ and become facilitators. Not easy. Holding back from providing the answers when you ask a group a question is ‘more difficult than trying to stifle an oncoming sneeze’ (I know that feeling). Some forced themselves to count to 20, by which time someone else has usually answered the question (must try that).
Given this upending of hierarchies, it is not surprising that PD, like other participatory approaches, often leads to cultural and social change, by empowering the lowers to find their own solutions. In corporate and other tight hierarchies PD is often about identifying the parallel processes used by practitioners – the ones they keep hidden from their managers (shades of Ros Eyben’s work on parallel processes in aid agencies).
That may also explain why PD seems to be the last resort, rather than the first one – experts and funders call the Sternins when they’ve tried everything else and it has failed. Otherwise, the ‘Standard Model ‘of identifying gaps, devising initiatives to fill them, and disseminating the guidance is incredibly hard to budge.
Despite its successes, PD, ironically, remains an outlier in the aid business. This is partly because it is slow in scaling up beyond the community – each community has to make the journey of discovery – no shortcuts. Also it typically involves a long slow start – a big consultation to socialize recognition of the problem and ownership, then an inquiry phase in which the community identifies the PDs, then a ‘learning by doing’ phase. That can be difficult for funders, who have neither quick results, or a clear project plan of activities and outputs to comfort them.
What about applying PD in Oxfam or other INGOs? One of messages of book is that tight hierarchies and command and control systems are instinctively hostile to PD. More loosely coupled systems are more likely both to throw up more deviants, and to allow them to be recognized and acted upon.
So within INGOs the best place to start is probably to identify the most independent actors – in the corporate sector, the authors identify sales reps as ideal, since they are typically out there on their own, with enough independence to improvise. The INGO equivalents would probably be country teams, with a combo of a champion in a country director, and staff willing to surrender their expert status and do PD with partners and communities. Any volunteers? Teams already doing it? Could we have made more use of a PD approach, perhaps in the inception phase, in some of the projects I have written about over the years – Chukua Hatua in Tanzania or TAJWSS in Tajikistan?
This is a conversational blog written and maintained by Duncan Green, strategic adviser for Oxfam GB and author of ‘From Poverty to Power’. This personal reflection is not intended as a comprehensive statement of Oxfam's agreed policies.