Is the significance of race declining in the political arena? Yes, and no

Abstract

The significance of class is increasing in the USA, in the sense that economic inequality is rising within the black and Latino populations as well as among whites. Growing inequality is associated with increasing disparities in lived experiences. Is class also increasingly significant in political life? Survey evidence shows that the answer is yes: compared with previous decades, well-off blacks and Latinos are less strongly liberal in some policy preferences and feel more politically efficacious, while poor blacks and Latinos tend to move in the opposite direction. Well-off non-whites have not, however, lost any commitment to racial justice or identity, so the USA is not becoming ‘post-racial’. Given the complex patterns of change and persistence in opinions, Wilson's arguments about when and how race is significant remain as important and controversial as when first expressed.

Keywords:

• inequality
• within-group differences
• public opinion
• post-racial society
• black Americans
• Latinos
• class conflict
• racial attitudes

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. Arguably, wealth disparities are rising at an even faster rate, and with greater consequences, both in the American population as a whole and within each racial and ethnic group. Data on wealth-holding, however, are less systematic and require more extensive explication than data for income as revealed through Gini coefficients, so we limit ourselves here to the latter.

2. The yearly starting points for Gini coefficients differ in the US census for different groups. The Gini index was 0.378 in 1968 for whites, 0.412 in 1968 for blacks, 0.454 in 2002 for Asians, and 0.373 in 1974 for Hispanics (of any race). In short, income inequality has risen within each racial or ethnic group as well as across the population as a whole (http://www.census.gov/hhes/www/income/data/historical/household – Table H-4).

Note that ‘White, not Hispanic’ begins only in 1972; however, the Hispanic population in the USA was too small in 1968 to affect the results for that year more than marginally.

3. That pattern does not hold for Hispanics; data are not available for Asian Americans. Data are from Institute of Education Sciences, National Center for Education Statistics (2013). This report was generated using the NAEP Data Explorer (http://nces.ed.gov/nationsreportcard/lttdata). Our analysis spans the earliest and latest years in which data are available.

4. Authors’ analyses of the National Crime Victimization Survey. Data for 1978 are from Bureau of Justice Statistics (1980; Table 15). Data for 2008 are from Bureau of Justice Statistics (2011; Table 15). Due to data limitations, these reports do not show breakdowns for Hispanics or Asian Americans.

5. Asian Americans show a more mixed or perhaps volatile profile, due to some combination of methodological complexities in surveys, small sample sizes, and genuinely libertarian or conservative sentiments that accord poorly with liberal social policy views.

6. The first period includes 6,473 respondents, and the second (2012) 5,916. Even with these large sample sizes, the ANES does not include a large enough sample of Asian Americans in either period to permit an analysis of class patterns.

Analyses using income and occupation instead of educational attainment yielded similar results; analyses using class self-definitions found many fewer statistically significant results.

7. Four other items (spending on schools, health insurance, welfare, job guarantees) show few class differences or little change in class differences between the two periods.

8. For the other four items, class differences among Latinos already existed in the 1980s (spending on the poor, government guarantee of jobs), or were slight in both periods (crime control, public schools).

9. On a few items, however, the ANES shows slightly less class disparity among whites in 2012 than in the 1980s, which does not accord with our expectations.

10. Figure 2 displays expected values of ordinary least squares regressions, holding other variables at their mean values. Models control for age group, gender, region, native born, and in the case of Hispanics, origin nationality (Mexican, Puerto Rican, Cuban). Hispanics may identify with any race; blacks and whites do not identify as Hispanic.

11. Blacks showed no changes in class disparities on the issue of whether public officials care about the views of ‘people like me’.

12. The results reported in the text are predicted probabilities of logit regressions, holding other variables at their mean values. Models control for age group, gender, region, native born, and in the case of Hispanics, origin group (Mexican, Puerto Rican, Cuban). Hispanics may identify with any race; blacks and whites do not identify as Hispanic.

Additional information

Notes on contributors

Jennifer Hochschild

JENNIFER HOCHSCHILD is Henry LaBarre Jayne Professor of Government in the Department of Government at Harvard University.

Vesla Weaver

VESLA WEAVER is an Assistant Professor of African American Studies and Political Science at Yale University.