The Voice of the Public: Why Citizen Surveys Work
Miller, Thomas I., Kobayashi, Michelle, Public Management
Just how in touch with the citizens of your jurisdiction do you or your elected officials have to be? After all, there are the town meetings, the budget hearings, and the neighborhood talk-back sessions, as well as conversations with the chamber, school board, planning board, liquor board, board of zoning adjustment, Rotarians, Shriners, Odd Fellows, Grange, Elks, Lions, and like clubs representing all manner of fauna and flora. Open-mike time at council meetings can get to look like bug-house square at a carnival.
The Usual Suspects
The truth is--and most staff suspect this--that the varied community activists who show up for every input opportunity are members of a single cadre of irate, enthusiastic, agitated curmudgeons who care deeply about the community in general, or their blocks in particular. When the dust settles and the budgets and policies have been written, the suspicion lingers that the typical resident still has not been heard from. Decision making by "wheel decibel" (a squeaky wheel gets the oil), after all, could simply be dismissed as the American way, by which those people with enough interest, energy, or money get to call the tunes. Although this genre of cynicism has been raised almost to a religion for the politically savvy, giving in to it won't work when you come to the apolitical questions that managers need answered if they intend to run their communities well and to run them for all.
The Citizen Survey Defined
The way to capture that much-vaunted voice of the typical resident is by a citizen survey, a scientifically conducted survey whose purpose is to gather the opinions of a sample of adults who represent the entire adult population of a jurisdiction. A citizen survey finds and gives voice to all types of citizens, the poorer as well as the better-off residents, those whose health may keep them from attending meetings and those in better health, shy people and outgoing people, newcomers and old-timers, and those who have a dispassionate point of view as well as those who are emotionally involved. The representative sample tapped in a citizen survey provides the point of view that can be found only in the community at large. We have found that about 15 percent of respondents to citizen surveys have attended any public meeting in the past 12 months. This means that 85 percent of the voices heard in a citizen survey are new.
This article, then, addresses citizen surveys that include an evaluation of local government services, that provide a kind of consumer scorecard. Common practice in local government service evaluation is to count citizen complaints. Typically, these "evaluations" of services come when there is a crisis--for example, right after a snowstorm, when streets are impassable and motorists are irate. But snowstorms of criticism are no way to judge the quality of services. Because evaluative surveys collect so much information so much more efficiently than any other kind of citizen participation, they are among the local government administrator's most useful management tools.
What's the Point of Surveying?
It's not that surveying is the only way or the most accurate way to know what citizens in a community think or do. Citizen surveying is a compromise made in the face of a scarcity of resources. If time was boundless and money ran like water from a spigot, no one would bother with surveys. In the world of wishes, everyone in a community would be asked and everyone would respond to the questions of interest. There would be no guessing about what the people wanted or what they liked or what they did.
In the real world--where time flies and money talks--surveys are the quickest, cheapest, and most accurate way to reckon the state of public opinion. Survey researchers are stuck with the unenviable job of figuring out what everyone would say after hearing from only a few. In the typical Gallup poll, pretty good guesses are made about the likely behavior of 100 million Americans based on reports from only about 2,000 of them--a sample of only about 0. …