Cover image for Counting on the census? : race, group identity, and the evasion of politics
Counting on the census? : race, group identity, and the evasion of politics
Skerry, Peter.
Personal Author:
Publication Information:
Washington, D.C. : Brookings Institution Press, [2000]

Physical Description:
x, 261 pages : illustrations ; 24 cm
"In part, this book is a brief against census adjustment. It begins by arguing that the inherent unreliability of racial and ethnic data requires a more realistic standard of accuracy than has typically been adopted by adjustment advocates. It also maintains that the implications of the undercount for both minorities and nonminorities--including the partisan interests of Democrats and Republicans--are grossly exaggerated and misunderstood. A novel intervention into a highly complex system, adjustment would produce all sorts of unpredictable results..." -- Introduction, p.1.
Census politics : boundaries and bureaucracies -- Creating racial and ethnic categories -- The undercount : causes and proposed remedies -- The politics of census adjustment -- The census in the new American political system.
Geographic Term:
Format :


Call Number
Material Type
Home Location
Central Library HT1521 .S535 2000 Adult Non-Fiction Central Closed Stacks

On Order



Since the U.S. Constitution first instructed that a slave be counted as only three-fifths of a person, the census has been caught up in America's racial dilemmas. Today it is torn by controversies over affirmative action, evolving racial identities, and minority undercounts. In Counting on the Census? Peter Skerry confirms the persistence of minority undercounts and insists that racial and ethnic data are critical to the administration of policies affecting minorities. He rejects demands thatthe census stop collecting such data. But Skerry also rejects the view that the census is a scientific exercise best left to the experts, and argues that it is necessarily and properly a political undertaking. To those advocating statistical adjustment of the census, Skerry insists that the consequences of minority undercounts have been misunderstood and exaggerated, while the risks of adjustment have been overlooked. Scrutinizing the tendency to equate census numbers with political power, Skerry places census controversies in the broader context of contemporary American politics and society. He traces our preoccupation with minority undercounts to the pervasive logic of an administrative politics that emphasizes the formal representation of minority interests over minority political mobilization and participation. Rather than confront the genuine social and political problems of the disadvantaged, political elites turn to adjustment to tweak outcomes at the margin. In such a context,where ordinary Americans already feel bewildered by and excluded from politics, the arcane techniques of adjustment would undermine public confidence in this most fundamental function of government. Finally, in a society where racial and ethnic identities are more fluid than ever, Skerry calls for greater realism about the limited accuracy of census data--and for greater tolerance of the untidy politics that accompanies the diversity we have come to value.

Author Notes

Peter Skerry is a non-resident senior fellow in the Governance Studies Program at the Brookings Institution and teaches political science at Claremont McKenna College. He is author of the award-winning Mexican Americans: The Ambivalent Minority (Harvard University Press, 1995).

Reviews 1

Choice Review

Skerry argues that there is no purely scientific way of conducting a census; the inevitably political nature of the process needs to be recognized rather than lamented, or worse yet, denied. For example, there is no "neutral" answer to the question of who should be counted--illegal aliens are included because the courts have ruled they should be. When it comes to classifying people by race and ethnicity, the ambiguities are endless. The current policy of racial and ethnic self-identification leads Skerry to conclude that "As the Census Bureau wades into increasingly subjective attitudinal phenomena, its self-conscious mission of objective 'fact-finder' makes less and less sense." Skerry's discussion of the operational, philosophical, and political challenges confronting the Census Bureau would make for an excellent case study in any research methods class. General readers will also be rewarded with a much enhanced understanding of how the census functions and of the issues surrounding its use. Skerry's analysis, along with his argument against the statistical adjustment of results to remedy the undercount of minorities, make this book an indispensable contribution to understanding the significance of the census. General readers; upper-division undergraduates; graduates; professionals. ; Alfred University



Chapter One Introduction               This book has its origins around the time of the 1990 census, when minority leaders were expressing dismay that once again hundreds of thousands of individuals would not be enumerated in the decennial count. To compensate for the undercount, some were advocating statistical adjustment of census results. Today, after a decade of litigation and controversy, the issue is hardly resolved. While the Supreme Court has virtually assured that apportionment of the Congress will not rely on adjusted data from the 2000 census, the Bureau of the Census will nevertheless produce a set of adjusted numbers that may well be used for redistricting, allocation of funds, and other purposes. Meanwhile, concern about the minority undercount continues unabated, as does the pressure for census adjustment.     In part, this book is a brief against census adjustment. It begins by arguing that the inherent unreliability of racial and ethnic data requires a more realistic standard of accuracy than has typically been adopted by adjustment advocates. It also maintains that the implications of the undercount for both minorities and nonminorities--including the partisan interests of Democrats and Republicans--are grossly exaggerated and misunderstood. A novel intervention into a highly complex system, adjustment would produce all sorts of unpredictable results. In some cases minorities could end up with relatively lower population totals with adjustment than without it. Moreover, adjusted census data would get fed into complicated fiscal formulas and redistricting calculations whose outcomes would be similarly counterintuitive. Thus I argue that adjustment would not benefit minorities as much as advocates believe and might well disadvantage them.     At the same time, the risks of adjustment have been underestimated. Adjustment could further undermine already weak incentives to cooperate with the census and thereby exacerbate the problem it was intended to remedy. Advocates have depicted the process as little more than an exercise in the science of statistical sampling. But adjustment would be much more: a huge logistical undertaking of which sampling would be merely one part. Not only would this process be difficult to explain and interpret to the American people, but its very complexities would create new opportunities for error. In my view, the possibility that adjustment could undermine public confidence in one of government's most basic functions greatly outweighs whatever benefits might be realized. But for me, and I hope for the reader, the case against adjustment is not the whole story. The controversy over census adjustment opens an invaluable window onto American politics at the end of the twentieth century. In the context of the modern administrative state, political elites have become acclimated to an increasingly demobilized electorate; rather than confront the genuine social and political problems of the disadvantaged, political elites prefer to argue about how best to tweak data for marginal gains. Indeed, the debate over census adjustment seems less about the empowerment of disadvantaged minorities than about the impoverishment of contemporary politics.     Yet here a question arises: why do the participants in this debate--Republican party leaders, Census Bureau officials, minority advocates, big city officials and politicians, editorial page writers, elected officials--take the positions they do? As in many controversies, interest calculations here are complicated and not necessarily obvious. Republicans, for example, have argued so forcefully against adjustment because of their traumatic experiences with redistricting at the hands of Democrats in 1990 and especially in 1980. City officials have sued for adjustment in part because this is one issue where local boosterism and minority concerns coincide. As for minorities, their concern about the undercount has not always translated into unhesitating support for adjustment. Actually, minorities are quite distrustful of the bureau's efforts to "modernize" the census.     Just as striking is the way affirmative action figures into this debate. Much of the energy directed against adjustment derives from hostility to affirmative action. Critics routinely express hostility to the census because they regard it as a vital cog in the machinery by which the federal government "counts by race." Yet at the individual level, there is no direct connection between affirmative action and the census: unlike someone filling out a college application, an individual identifying himself on a census form as belonging to a protected minority group does not stand to benefit directly. If members of minority groups did benefit from so identifying themselves on the census, then the undercount would be greatly diminished.     Also looming in this debate are the ongoing controversies over diversity and multiculturalism. The debate itself--and certainly the efforts by racial and ethnic groups to maximize their census counts--is routinely taken as one more indication that American society is balkanizing into contending racial and ethnic groups. Yet an important finding of this study is that much of this controversy over the census arises not because Americans are breaking up into hard-edged groups, but because we are intermixing as individuals to the point where group barriers are breaking down, making it increasingly difficult for the census to count racial and ethnic identities meaningfully. This intermixing is one reason why racial and ethnic data lack reliability. Of course, this may lend credence to the view that the census should stop asking Americans about their racial and ethnic backgrounds. But this is not my argument. My view is that racial and ethnic data--however much lacking reliability--are too important to be discontinued. At the same time, we need to be more realistic about the limitations of these data and their potential uses. Such issues underline that the census is inextricably bound up with race, and so I believed when I first embarked on this project. Fairly soon, however, I conceived doubts about focusing unduly on such controversial flashpoints as the minority undercount and adjustment. Encouraging these doubts were some of my colleagues and several of the individuals I was interviewing. They tried, usually subtly and rarely explicitly, to divert my focus. I was specifically steered away from racial issues by senior officials at the Census Bureau, some of whom directed my attention, for example, to privacy issues. It is probably no accident that during this same period the bureau in its public pronouncements was doing its best to discount the importance of race--for example, by justifying statistical adjustment in terms not of racial equity but of reducing costs.     Not everyone adopted this stance. Eugene Ericksen, a sociologist who has served as an adviser to the secretary of commerce on census adjustment, observed in 1991 congressional testimony about adjustment: "Were it not for the civil rights aspects of this issue, it would not be something we would be focusing on so much." In 2000 any avoidance of race seems increasingly implausible, and it has accordingly been less evident. Nevertheless, as I will argue here, there is a persistent and pervasive tendency at the Census Bureau and at the Office of Management and Budget to downplay racial matters and even to deny their importance.     There is scant historical justification for such efforts. Starting with Article I of the Constitution--which stipulates how the enumeration is to be conducted, explicitly in the case of Indians and implicitly in the case of Negro slaves--the census has always been entwined with race. Certainly the debate over the undercount and the proposed remedies for it are all about race. This seemingly obvious point needs to be emphasized up front, before I explore the complex and somewhat technical matters that any serious discussion of the census necessarily involves.     Why has the Census Bureau sought to downplay race? One reason is that the agency has long been dominated by demographers, statisticians, and other highly trained professionals who pride themselves on their technical expertise. Emphasizing their politically neutral role as "factfinders for the nation," census professionals do not typically see themselves as political actors or policymakers. They certainly do not gravitate readily to a controversial subject such as race--unless, as we shall see, it is possible to do so in a way that transcends (or appears to transcend) politics.     A related factor is the tension between the extremely technical character of the census and the emotional, highly symbolic nature of race politics. A particular problem is the degree of misinformation and confusion that envelops public discussions of the census, whether of adjustment procedures or simply of the methods used to collect racial data. Such confusion is evident among elites and ordinary Americans alike.     Indeed, elites are not very interested in or focused on the census. Newspapers, for instance, are prone to decry the "racial injustice" of the census undercount. But editors are also leery of devoting much space to such a technical and (as far as their readers are concerned) tedious topic. The unhappy result is episodic coverage that is often ill-informed or downright misleading, punctuated by high-minded editorial posturing. Similarly, most politicians, including most members of Congress, would place the census at the very bottom of any list of compelling topics. These same politicians, however, are more than prepared to express moral outrage about the negative impact of the census undercount on disadvantaged minorities.     Elites behave this way for good reasons, of course. Although they may find the census tedious, they realize that it packs an emotional wallop for minorities, especially black Americans. Yet the technical nature of census discussions remains jarring to minorities and their allies, who tend to feel that basic human needs are getting lost in a welter of statistical minutiae. In a curious way this problem is exacerbated by the Census Bureau's openness in providing an abundance of data and a wealth of technical analyses about its operations and products. In this regard the census is an activist's nightmare and a policy wonk's dream.     This is precisely the problem: in the sea of information about this massive undertaking, it is easy to lose sight not only of the more fundamental issues at stake--for example, what this most basic of governmental functions tells us about the nature of citizenship in late-twentieth-century America--but also of the larger symbolic issues. A mantra repeated in this debate is that the census is "a national ceremony." It is taken from the statistician William Kruskal: The census is one of our relatively few national, secular ceremonies. It provides a sense of social cohesion, and a kind of non-religious communion: we enter the census apparatus as individual identities with a handful of characteristics; then later we receive from the census a group snapshot of ourselves at the ceremony date.     No one wants to be left out of a group snapshot--even if, as we shall see, the absence of many individuals is due at least as much to their own choices as to shortcomings of the Census Bureau. This desire not to be left out is all the stronger among black Americans in particular, given the Constitution's original provision that slaves would be counted as three-fifths of a person. It is impossible to overemphasize the importance of that single historical datum in this controversy.     These symbolic aspects of census politics are crucial, and I have tried to be attentive to them. But I have also resisted becoming overly preoccupied with them, out of a strong conviction that racial politics in the United States has too frequently been allowed to be overwhelmed by symbols.     But if the U.S. census has always been bound up with race, our notion of race has also been undergoing enormous change. Most obviously, race in the United States has become more diverse and more complex. The racial spectrum no longer consists merely of whites, blacks, and American Indians. In addition to Asians, that spectrum now includes Hispanics, a group whose racial status is quite ambiguous but increasingly important.     Of even greater significance is the transformation in our very conceptualization of race, and of ethnicity. Until quite recently social scientists regarded these as social constructs that reflected one's ties to other individuals and institutions. In the context of the census, these constructs were reported on by enumerators. Today, race and ethnicity are seen more as matters of individual psychology, of an individual's subjective feelings of attachment to a group, or groups. Moreover, this identification is now reported to the census not through an enumerator, but directly by each individual. Indeed, "self-identification" of a respondent's racial and ethnic background is now a principle to which the Census Bureau proudly adheres.     This conceptual shift helps to explain what I have already emphasized: the ambiguity of racial and ethnic data. Moreover, while race and ethnicity have become increasingly subjective categories, they have grown more critical to public policy. This then points to a conflict between the vagaries of personal identity and the requirements of bureaucratic rationality, a conflict most evident when the Census Bureau, seeking to impose order on what would otherwise be chaos, violates its commitment to racial and ethnic self-identification and forces the myriad of self-identified responses into established categories. This exercise of governmental authority in such a personal and controversial realm is a delicate balancing act: understandably, the bureau is not eager to acknowledge its inevitable role in the rationalization of racial and ethnic identity.     This balancing act is hardly an isolated example. On the contrary, it is typical of how authority in general is wielded in the contemporary administrative state. As various students of what has come to be called "the new American political system" point out, in recent decades substantive policy goals have come to be pursued less and less through conventional political and legislative means and increasingly through arcane legal and administrative channels that leave many ordinary Americans feeling bewildered and excluded.     In a regime that emphasizes administration over politics, it is perhaps not surprising that the numbers necessary to administer outcomes get an enormous amount of attention. The most revealing aspect of the census controversy is the tendency for all of the disputants to equate census numbers with political power. Adjustment advocates are particularly prone to this equation, deemphasizing the importance of actual voting or other forms of political effort. Some of these advocates go so far as to charge that to be uncounted in the census is to be "disenfranchised." While such perspectives reflect the administrative realities of the new American political system, this is a curious line of argument in an era when the quality of citizenship and of civic engagement is being criticized from all sides as thin.     Nor in such a regime should it be surprising that there would be an effort to cleanse the census of politics. This is no mean feat, since the census was originally designed by the Framers to deal with the fundamental political tasks of taxation and representation. Nevertheless, adjustment advocates argue that the census does--or ought to--transcend politics. Specifically, they insist that the census is a scientific undertaking into which politicians have no business intruding; alternatively, they maintain that there is a right to be counted that is denied by the census undercount. Science and rights are both characteristically American ways of seeking the advantages of political power without having to acknowledge the pursuit of politics or of power. These now enjoy renewed currency in our new American political system. And in this new regime, it is increasingly important--and difficult--to discern that the census is not only a critical tool in the administration of governmental benefits, but also an instrument of state power and authority. Readers primarily concerned with the policy debate over the undercount and adjustment might want to turn directly to chapter 5, "The Politics of Census Adjustment." There the various points sketched above are developed in detail. Chapter 6, "The Census in the New American Political System," broadens the argument by relating the census to the sweeping changes in American politics over the past thirty years. While those changes have often been criticized for contributing to the shift from the representation of individuals to that of groups, equally important has been the shift from a focus on the political participation of the disadvantaged to a formalistic notion of their representation. Chapter 6 also explores how the dynamics of public interest organizations have been fueling this controversy and how, in the context of this new regime, adjustment would benefit various elites more than the disadvantaged minorities who are its presumed beneficiaries.     But the reader is encouraged to take the scenic route by beginning at the beginning. Chapter 2 examines the politics of the census process and the bureaucratic politics of the Census Bureau. Relying on Aristotle as well as the Framers, I argue that the census is inherently--and properly--political in nature. But contrary to popular perception, politics does not typically impinge on the census through the gross manipulation of data. Where politics does play an essential role is the creation of the racial and ethnic categories that appear on the census questionnaire. At the same time, the Census Bureau, like any government agency, strives to control its environment and to protect its prerogatives and mission. It does this by presenting itself as a neutral fact finder and by maintaining a rigorous if delimited notion of professionalism. What the bureau has not done (at least until recently) is define its mission as primarily scientific in nature--a stance that, as we shall see, distinguishes it from many advocates of adjustment.     Chapter 3 scrutinizes the bureau's racial and ethnic categories, clearly the most politicized and contentious aspect of the census, and explores the necessarily limited precision of these data, which may be a bigger problem than the actual undercount of minorities. Further, chapter 3 examines the implications of the tension between our societal commitment to self-identification and our equally strong need to generate authoritative racial and ethnic data on which public policy can be based.     Chapter 4 confirms that the undercount is persistent and indeed has grown worse over the past fifty years. But after a review of the extensive body of ethnographic research sponsored by the Census Bureau into the causes of the undercount, two things become clear. First, responsibility for the undercount cannot be wholly laid at the door of the bureau. Second, noncooperation with the census occurs for a fascinating and surprising variety of reasons, many of them all too understandable from the viewpoint of those who do not cooperate. Finally, chapter 4 looks closely at the specific sampling and adjustment programs that have been proposed to remedy the undercount.     The book concludes with some admittedly imperfect crystal ball gazing in chapter 7 about possible (and impossible) alternatives to census adjustment. For a variety of reasons, the controversy is not likely to be resolved in the near future. This suggests that we may simply have to learn to live with the ambiguity that suffuses racial and ethnic data, and with the politics that inescapably pervades the census. As we begin the new millennium, one of the main challenges we face may well be learning to tolerate the untidiness that accompanies the diversity we have come to value.

Google Preview