(en route to Austin, TX) Yesterday, we heard an exemplary presentation on Community-Based Participatory Research (CBPR) by Berkeley professor Meredith Minkler, who is one of the leaders of the field. She told a perfect story of a research project that was a collaboration between university-based scholars and laypeople: in this case, workers in Chinese restaurants in San Francisco. The workers provided guidance that made the research interesting, original, and important. My colleague Sarah Shugars has a summary.

I was struck by a minor point that brings up a larger issue. The Berkeley researchers knew that non-fatal accidents are very common among restaurant workers. Their worker colleagues noted that restaurants’ first aid kits tend to contain only Band-Aids. The team therefore calculated the percentage of restaurants that do not maintain satisfactory first aid kits. In order to generate that statistic, they had to know how many Chinese restaurants there are. The City of San Francisco did not know. The restaurant workers contributed a precise count.

Now, a city could count its Chinese restaurants. It could send one of its paid employees to count, or it could hire a contractor for that purpose. But many of our political theories just assume that the state knows things. We take that for granted. On the contrary, knowing something as simple as how many Chinese restaurants exist raises layers of problems:

• The state must care about the topic in order to collect the data. (To be fair to San Francisco, its health department participated in this project. My point is a general one about the need for the state to care.)
• The state must pay for the data, which is not free. Collecting the data may be more expensive for the state than for other parties. For instance, Chinese restaurant workers can read signs in Chinese; most city government employees cannot.
• The state must define the concept, which almost always raises value questions. (What is a genuinely “Chinese” restaurant? Why separate Chinese restaurants into their own category?)
• The state must employ agents who act with integrity. For instance, a state employee who counts Chinese restaurants could take bribes to leave out some restaurants so that they would avoid scrutiny. That would ruin the data.
• The state must collect the information competently. As I noted recently, “The same US government that can apparently tap almost any telephone in the world cannot harvest information that people voluntarily provide on the government’s own website regarding their eligibility for insurance.”
• The state must pay attention to the data it collects. After all, “the state” is actually a whole bunch of human beings who do not automatically know what their colleagues know, let alone act on that knowledge.

Clearly, states should collect information as the basis of sound policy. We shouldn’t ask restaurant workers to do all the research about restaurants. But collecting good data is itself a political achievement. We can’t just presume it will happen. Nor is the best way to obtain it always for the state to buy it. For one thing, citizens can benefit from being the researchers. In this case, the same restaurant workers who collected the basic data also won significant reforms in city law. The process of data-collection took effort (for which they were paid on the grant), but it also gave them political power.

(See also “Why Engineers Should Study Elinor Ostrom“)