Climbing Down from the Ivory Tower
Nava Ashraf explains why it makes sense for field researchers to co-produce knowledge with the people they study and serve.
In the late 1970s, the Chicago Police Department noticed that the city's crime rate increased when cops stopped walking the beat and started driving around in patrol cars instead. They wondered why, and asked the political scientist Elinor Olstrom to provide some insight.
She realized the problem: Cars negated the ability to interact with people on the streets. As it turned out, cops did a better job fighting crime when they sought input from the public. That had happened organically when they walked their beats. And by sharing their knowledge of the neighborhood, the citizens helped the police to protect the citizens. Olstrom dubbed the phenomenon "co-production."
Since then, it has become common for organizations to treat consumers as production partners. Many high-tech firms hire existing customers to beta-test early versions of products, in order to get critical feedback before creating the final versions. Some companies even base their business model on co-production; Build-a-Bear stores allow consumers to create personalized teddy bears.
"With the growth of customizable technology, there is an appreciation of this co-generation process in many, many domains," says Nava Ashraf, an associate professor in the Negotiations, Organizations, and Markets unit at Harvard Business School.
"Asking them to apply something that's been generated somewhere else isn't involving them in the process"
But there's a dearth of co-production in the public health sector, which often depends entirely on faraway academics and organizations to solve local problems. This frustrates Ashraf, who for more than a decade has studied behavioral economics in the context of developing countries. She believes that global health researchers and public health workers should make a point of co-producing knowledge with those who will most benefit from it: the residents of the communities they study.
"The common process of thinking is that from over here in our ivory tower at Harvard, we are supposed to find the knowledge, bestow the knowledge to people on the other side of the world, and get them to apply it and disseminate it," she says. "That seems kind of wrong."
Recognizing the importance of knowledge co-production can lead not only to successful products and services, but also to improved supply chain and delivery strategies. She justifies the importance of co-production in her elective course, Managing Global Health: Applying Behavioral Economics to Create Impact, which she teaches to second-year MBA students at HBS and select students from Harvard Medical School and the Kennedy School of Government.
The User's Perspective
The course begins with a case in which UNICEF tries to understand why there are so many diarrhea-related fatalities in developing countries. (Diarrheal diseases account for one in nine child deaths worldwide, according to the Center for Disease Control.) This, despite the existence of an effective treatment called oral rehydration therapy, which essentially involves drinking a solution of water, salt, sugar, and supplemental zinc.
The students soon realize that UNICEF and the World Health Organization have failed to consider the problem from the end-users' perspective. ORT, which provides life-saving rehydration, does not actually eliminate diarrhea, so it's hard for parents to tell whether it's working. The ORT solution tastes awful, like salty metal, so many children refuse to drink it. And pharmacists tend not to suggest ORT to mothers because they don't consider it to be "real medicine," the case explains.
Ashraf believes these oversights could have been prevented if parents and local health care workers had been involved in designing a solution. Indeed, oral rehydration therapy was successful in countries like Bangladesh, where cooperation among field workers, scientists, and mothers led to a very effective program. She cites Atul Gawande, who explained one program's progress in a 2013 New Yorker article:
"The field workers soon realized that having the mothers make the solution themselves was more effective than just showing them. The workers began looking for diarrhea cases when they arrived in a village, and treating them to show how effective and safe the remedy was. The scientists also investigated various questions that came up, such as whether clean water was required. (They found that, although boiled water was preferable, contaminated water was better than nothing.)"
"It's too common to think, we have the technology and we have the systems, and they'd work if only people would use them," Ashraf says. "But in doing that, we're not realizing that they should be part of the process. Asking them to apply something that's been generated somewhere else isn't involving them in the process."
Co-production in field research
Out of the classroom, Ashraf has been walking the talk by incorporating co-production into her own field research.
In 2010 Ashraf was part of a research team that helped the Zambian government with a nascent national program to recruit, train, and employ community health workers throughout the country, which was suffering from a severe lack of health care providers. The goal was to employ 5,000 new community health workers by 2015. The challenge was figuring out how to recruit a talented, dedicated cadre of people who wanted to serve their local communities.
Often, academic field researchers design an experiment among themselves, only going to the field when they're ready to conduct the experiment on members of the community. In this case, though, the academics designed an experiment with the head of policy development for the Zambia Ministry of Health, along with several other local officials. The academics came armed with knowledge of prior research. The local officials had practical knowledge of the community.
Co-production is a tricky endeavor for scholars. It's time-consuming to incorporate everyone's input, especially when working with people who are unfamiliar with standard research rules. It took nearly four years to design the health care worker recruitment experiment, Ashraf notes.
"It's challenging to balance the rigor of science and objectivity with the participation of the people we're studying," she says. "But sometimes it makes sense to have people participate in running the experiment, not only because they're the ones who can benefit most from the results, but because it's better for knowledge. At every step you can say, 'Look at the data. What do you think of this?' And their input is so important. Because in these cases, they're the ones who are the real experts, you know?"
In the eventual field experiment, researchers launched two recruitment campaigns. Each targeted half of 48 randomly selected rural districts across the country. In 24 districts, potential employees were wooed with posters and brochures emphasizing community service. In the other 24, the material emphasized career advancement.
The campaigns drew two distinct groups of candidates. While the career advancement campaign attracted a group that had more qualifications and was more technically knowledgeable, both campaigns attracted candidates with equally high levels of pro-social motivation. The researchers followed the health workers' progress for 18 months when they returned to their communities to provide health services.
It turned out career-focused applicants visited 29 percent more households than did community-focused candidates, and conducted over twice as many community mobilization meetings, while seeing approximately the same number of patients at the health posts. Both groups spent the same amount of time at each household, meaning the career group wasn't sacrificing quality time for quantity. And the employee retention rate was similar in both groups.
The researchers detail their findings in the paper, "Do-Gooders and Go-Getters: Career Incentives, Selection, and Performance in Public Service Delivery." The key takeaway in the paper: "Career incentives, far from selecting the 'wrong' types, attract talented workers to the public sector who deliver health services effectively."
But there's another takeaway for Ashraf, whose current research involves co-producing studies with urban planners in sub-Saharan Africa.
"We show academics that you can actually generate good science while also having impact on good policy," she says, "even though it's a little more labor intensive and might take longer."