Between December 2014 and June 2015 two MIT co-authors and Robert Blair, the Joukowsky Family Assistant Professor of Political Science and International and Public Affairs, ran a three-wave panel survey in Monrovia, Liberia. The survey was funded by the International Growth Centre and the Omidyar Foundation, and was intended first and foremost to help the response to the Ebola epidemic. The Liberian government had very little data about which Monrovia neighborhoods had been hardest hit in terms of both lives lost and livelihoods destroyed, and the goal was to help target the recovery to the communities that needed it most.
The survey taught the trio a lot about the relationship between Liberians' perceptions of government and the severity of the epidemic at the very local level. A majority of respondents expressed deep distrust of the Liberian government, and were skeptical of its attempts to control the crisis. More problematic, those who distrusted government were less likely to take precautions that might have slowed the spread of the virus. They were less likely to report washing their hands with chlorinated water; less likely to obey a nighttime curfew or ban on public gatherings; and less likely to support other government-imposed public health restrictions. The problem, moreover, wasn't that these respondents didn't understand the risks -- in fact, those who distrusted government tended to know more about the virus than those who didn't. The problem was that they didn't trust the government to implement policies that would slow the virus's spread. Nor was this distrust "irrational." While the Liberian government has improved immeasurably since the end of the country's civil war in 2003, corruption and abuse remain endemic. Skepticism was dangerous, but not unwarranted.
That was the bad news. However they got some good news when they found evidence of underreported success stories in the government's efforts to regain citizens' trust at the height of the epidemic. Aware of its own limitations, the government partnered with local civil society organizations to canvass Monrovia neighborhoods, teaching residents to recognize the symptoms of Ebola, training them in simple preventative measures and encouraging them to report known or suspected cases to the authorities. These were low-cost, day-to-day interventions, but they were extremely effective. Respondents who reported outreach of this kind were more likely to adopt presentation in the home, more likely to cooperate with the authorities in identifying cases, and more likely to support even the more contentious public health restrictions in effect at the time (like "safe burial" of Ebola-infected bodies).
Of course, as a group of social scientists they are always worried that patterns like these are merely correlations, not relationships of cause and effect. But they found a lot of evidence to support a causal interpretation of our findings and describe that evidence in a couple of working papers, which will be published soon. In the meantime, you can read an IGC memo summarizing the results. The policy punchline is that low-cost, face-to-face interactions between government and citizens can have strong positive effects on confidence and compliance, even (and perhaps especially) where the government is weak and widely distrusted.