The Hard Count: The Political and Social Challenges of Census Mobilization (review)

Despite its subtitle, four of this short book's five chapters do not concern the politics involved in organizing censuses as such, but rather the effects of the ambitious campaign undertaken to improve response rates for the 2000 U.S. census. The final chapter, written by the former director of the Census Bureau, Kenneth Prewitt, contains some interesting and cautiously odd reflections on the lessons to be learned from the 2000 effort for the census of 2010.

Having reminded the reader briefly of some aspects of the longer history of American census making, the authors revisit the intense partisan controversy of the 1990s surrounding proposals for a statistical remedy for the undercounting of urban minority groups in the mail-in censuses conducted from 1970. As the 2000 enumeration approached, statistical remedies were attacked in the courts and in a closely divided Congress by Republicans fearing the creation of large numbers of urban Democratic voters. Instead, an extensive and well-funded census publicity campaign was mounted and new logistical measures were introduced, aimed at raising participation rates in general and among undercounted groups in particular. However, in the midst of the campaign, Republican members of Congress and their media attack dogs began yapping at the census long form, sent to about 19 million addresses, as government intrusion into citizen privacy.

The bulk of this book examines the consequences of the 2000 mobilization campaign on participation rates and of the attack on the long form on participation rates and item responses. The campaign is shown to have raised participation rates above the 1990 level and largely to have eliminated the ethnic undercount. The attacks on the long form reduced response rates and item completion rates as well. In the course of their demonstration, the authors counter claims in several literatures that public campaigns are ineffective in influencing behavior and that publicity campaigns reinforce knowledge gaps. The census campaign is said to show that the American public is susceptible to mobilization for civic cooperation and, against claims in the social capital literature, census completion rates were found to be more closely associated with variations in household structure than with participation in civic associations.

Other readers may wish to quibble around methodological issues; what strikes me most about this book is the concluding chapter by Kenneth Prewitt. Sociologists are not often able to read the reflections of a census director on the lessons to be learned from the census making process. Prewitt had an usually challenging time of it because of the partisan political climate, which led to seventeen formal House Census Subcommittee hearings, and the creation of a new oversight body in the guise of the Census Monitoring Board, but his census came in well under budget and with an improved participation rate. Given his publisher and his own successful administrator's habitus, it is perhaps too much to hope for some biting criticism of the politicization of the 2000 census. Prewitt gives us a warmish plea for respect for the non-partisan qualities of the census as social scientific observation, and proposes that the next census "be designed as a civic ceremony," declaring indeed that "the census is in fact the only such [i.e., non-partisan] civic ceremony available to American political life." (p.118) Such sorry commentary on the quality of political society is the more striking because it follows Prewitt's discussion of the elimination of the census long form for 2010, with the introduction of the yearly American Community Survey. The average completion time for the census short form is three minutes.

Prewitt's other musings in this chapter touch on the challenges to state information-gathering posed by the interaction between citizen concerns with privacy and the massive proliferation of public and private data banks. As census supplements, commercial data banks are discounted as unscientifically produced (although I wonder about the market basket data of the world...