Aid on the Edge of Chaos, a book you really need to read and think about

November 8, 2013

What have we learned about women’s empowerment from a 17 country global programme?

November 8, 2013

How complexity thinking cut malnutrition in Vietnam by two thirds

November 8, 2013
empty image
empty image

To end complexity week, another of the fascinating case studies from Ben Ramalingam’s Aid on the Edge of ChaosBen Ramalingam cover

In December 1991, Jerry and Monique Sternin arrived in Vietnam so Jerry could take up the role of Save the Children US Country Director. The country was still labouring under a US-led economic embargo and had seriously high levels of child malnutrition, affecting 65 per cent of all under fives. Supplementary feeding programmes delivered by international agencies were expensive and the benefits were seldom sustained beyond their lifetime.

And the political climate was not welcoming for an American representative of a US non-governmental organization. One high-ranking official told Sternin a week into the visit, ‘There are many officials who do not want you in this country. You have six months to demonstrate impact, or I’m afraid my ministry will be unable to extend your visa.’

Just occasionally, just enough pressure is just what is needed to spark new ideas. Sternin remembered a Tufts University colleague, Marian Zeitlin, who was supported by the UN Children’s Fund and the (WHO) Organization to examine the phenomenon of positive deviance: the off-the-chart performance— in health, growth, and development—of certain children in a community compared with others.

Zeitlin trialled a version of positive deviance in Bangladesh, working with BRAC, to examine diarrhoeal infection and malnutrition in infants. The approach focused on mothers’ wisdom and knowledge, especially in relation to positive deviants: Zeitlin and her colleagues used scientific behavioural trials to identify the factors that led to positive deviance, then used this knowledge in the design and roll-out of interventions.

However, the Sternins did not have the resources or the time to do this. Instead, they used the principles of positive deviance as the basis for a different approach. Monique Sternin told me that one of their key goals was to make the positive deviance approach operational, with the community taking the lead, although Zeitlin’s work was essential in terms of providing the scientific justification.

After numerous negotiations, the Sternins finally obtained a mandate to start working in Quong Xuong district in Than Hoa province, some four hours south of Hanoi. They had just over a dozen weeks to demonstrate impacts before their visas were pulled. In the community, however, the limited amount of money they had was less than pleasing for the local officials involved, who were not happy about promises of ‘nothing more than “capacity building” and “self-reliance” ’.

Eventually, however, the Sternins were allowed to work in four communities with 2,000 under-three year olds, 63 per cent of whom were malnourished. Driven ‘more by faith than by proof ’, and armed with their belief in the evolutionary principles of positive deviance, the Sternins told the heads of the major village committees that the approach was going to be about finding solutions that were already in the community, which would have to take responsibility for their identification and application.

To their considerable surprise, the villagers were very keen on the idea. They had previously experienced only short-term aid projects after which they had watched their children’s health gains deteriorate again. This sounded different and more beneficial. After the children were weighed, they were ranked according to their family’s economic status.

Volunteer groups then identified the positive deviants. Sternin remembers asking the volunteers, ‘Is it possible for a child to be very poor and still well fed?’ and the volunteers literally leaping from their benches shouting, ‘Co co, co co!’ (‘It is, it is!’). It turned out that much of what was being done differently in the positive deviant families was tacit and unconscious: the individuals in question weren’t even consciously aware they were doing anything different or new.

Positive deviance VietnamTeams of volunteers undertook observations in their homes and found some intriguing things, some of which were common to all the positive deviant families. The two standout practices related to the content of the diet and the way food was administered. In every positive deviant family, the mother or father was collecting a number of tiny shrimps, crabs, or snails—making for a portion ‘the size of one joint of one finger’—from the rice paddies and adding these to the child’s diet. ‘Although readily available and free for the taking, the conventional wisdom held these foods to be inappropriate, or even dangerous, for young children.’

Families also varied the frequency and method of feeding. Other families fed young children only twice a day, before parents headed to the rice fields early in the morning and in the late afternoon after returning from a working day. Because these children had small stomachs, they could only eat a small amount of the available food at each sitting. The positive deviant families, however, instructed the home babysitter (an older sibling, a grandparent, or a neighbour) to feed the child regularly, four or even five times a day.

Using exactly the same amount of rice, their children were getting twice the amount of calories as their neighbours who had access to exactly the same resource. Other key factors included atypically high levels of hand hygiene in positive deviant families.

At this point, the Sternins resisted the powerful temptation to ‘teach’ the community what they had learnt, because ‘Our past development work failures . . . had all occurred exactly at the moment in which we now found ourselves; the moment at which the solution is discovered. The next, almost reflex step, was to go out and spread the word; teach people, tell them, educate them.’

This time, they decided, the key would be to give community members the opportunity to share and learn directly from each other, with a focus on fostering and facilitating the exchange of practices. At its core, this meant turning a conventional approach to behavioural change—that of Knowledge–Attitude–Practice (KAP)—on its head.

The basic premise of KAP is that, by changing knowledge, you change attitudes and practices. However, the failures in this approach are obvious: millions still smoke, overeat, and so on.

The Sternins reversed the process, to work on Practice–Attitude–Knowledge. ‘You start by enabling people to change their practice, which then changes their attitude, and ultimately they internalize new knowledge.’ This beguilingly simple idea would become the basis of the positive deviance mantra, and go on to inform thousands of applications around the world over the next two decades: ‘It’s easier to act your way into a new way of thinking than to think your way into a new way of acting.’

Over a two-week period, the Sternins provided guidance and tools on collecting data on the positive deviant practices. The progress was shared on aVietnam kidsboard in the town hall, and the charts quickly became a focus of attention and buzz.

A few short weeks later, district health staff assessed progress to date. The findings were remarkable: some 40 per cent of the children had already been fully rehabilitated, and a further 20 per cent were well on the way. Granted another six-month visa for their efforts, the Sternins continued their work. By the end of the first year, half the children had participated and 80 per cent were rehabilitated.

The model was taken on and applied by the Vietnamese National Institute of Nutrition, and after this by the government, which scaled it nationally. Over time, the positive deviance approach saw a sustained reduction in malnutrition rates of 65–80 per cent, and reached a population of 2.2 million.


  1. This is not a comment about the quality of the Sternin’s contributions in SE Asia, as it is one on the very loose use of concepts. Complexity theory, as commonly understood, is something altogether different than PD. In this context, complexity would seem to be about contingency and unpredictability. I would be interested in knowing what positive deviance has to do with complexity. Of note, the Sternin’s work also tends to be a gross simplification of PD as the product of activity of individual champions, when deviance is largely based on processes of social heterogeneity and differentiation. Beyond the notion that complexity and PD belong to the conceptual world of the social, I struggle to find the relationships between these concepts.

  2. Hi Stephen, the basic principles of PD as developed by Marian Zeitlein was that it was a social, cultural and biomedical adaptive phenomena. The Sternins wanted to see of they could develop a social process for sharing these adaptations, and built on evolutionary principles to do so (variation, selection, amplification). They themselves use the ideas and concepts of evolution and complexity to explain the theory of PD – see their 2010 book, and Richard Pascale’s book on management & complexity science.

    Nobel Laureate Dudley Herschbach has also written about the underlying principles and mathematics of the PD phenomena in different situations – including educational attainment in US – and shown that these are underpinned by power law relations (i.e. long tails, not bell curves) that are also signatures of complex systems. I go into both of these aspects in the book…

    Caroline: I think I am right in saying that Monique was involved initially as an observer, and facilitator where possible; and also helped write up the case studies. She has always been actively involved in the PD community – in specific case studies and in communicating the PD story; even more so since Jerry’s sad passing.

  3. Whatever the merits of complexity thinking and poverty deviance, it’s clear that complexity thinking did NOT cut malnutrition in Vietnam by two thirds.

    Malnutrition fell by two-thirds between 1991 and 2011. During those 20 years, Vietnam went from being one of the poorest countries in the world to one of the developing world’s great success stories.

    In 1991, the famine of the 1980s were recent memories and agriculture had only just been decollectivized.

    There were many factors–including health policy inspired by this PD work–in the decline in malnutrition but the overwhelming driver was surely economic growth.

  4. Gabriel – great comment! Thanks for reconnecting us with a much wider perspective on ‘what works’. So it seems that the conclusion by Ramalingam that ‘the positive deviance approach saw a sustained reduction in malnutrition rates of 65–80 per cent’ would not be correct. Seems we need more than complexity thinking and positive deviance to explain change.

  5. Hi Gabriel – thanks for comment. Very much agree.

    At the risk of offending Duncan – I think that perhaps the title of this blog piece puts a bit of unintentional spin on the stats I cite in the book. The PD approach did reduce malnutrition among focus communities by 65-80%, and it did eventually reach 2.3 million.

    Over a longer period, the levels of malnutrition nationally fell. The contribution of PD work to national levels of malnutrition reduction needs to be understood by taking into account broader factors, as you say.

Leave a comment