It is worth noting that the results reported above on attitudes toward affirmative action are extremely limited in one important respect.  With the exception of general questions about government expenditures, we rely on a single question focusing on a very specific and strong form of affirmative action: giving preferences in hiring and promotion to African Americans.  Absent from the long-term survey record are questions that ask about the full range of programs and policies that were envisioned under the rubric of “affirmative action.”  Indeed, the kind of policy described in the question likely conjures up the notion of “quotas” for many respondents; a policy that is both illegal (except in specific, court-ordered situations) and extremely unpopular.  Because we require in our trend analysis questions that have been asked at least three times, of national samples of respondents, and spanning 10 years, we have only this limited number of eligible items.  The level of specificity of the program as described in the survey question, the type of policy (preferences versus quotas versus economic aid versus job training and educational assistance), and the target (women or blacks or minorities) can all impact the level of support expressed by Americans for these race-targeted policies (Steeh and Krysan 1996; Bobo and Kluegel 1993; Bobo and Smith 1994).  On the one hand, Steeh and Krysan (1996) report, when the wording of the question starkly contrasts abilities with preferential treatment based on race, fewer than 10 percent of whites favored a race-based policy (survey data available between the late 1970s and early 1990s).  On the other hand, when survey questions ask about affirmative action but specifically clarify that it would not include rigid quotas, support is far higher.  Indeed, in 1988, the last year for which data are available, 73 percent of whites favored affirmative action when phrased in this manner.  These contrasting figures give a hint of the complexity of attitudes toward affirmative action; a complexity that is missed because the survey questions that are asked regularly in national surveys ask only about a few specific kinds of policies or about very general issues of “government spending.”   
 

Explanations for Inequality, Perceptions of Discrimination, and Stereotypes  
 

Tables: 3.4A, 3.4A, Supplement, 3.4B, 3.4B, Supplement 
 

Explaining Racial Inequality.
In the past few decades, there has been increasing interest in understanding how people explain current racial inequality.  This is in part because it is a useful predictor of levels of support (or opposition) to the kinds of race-targeted policies just described (e.g., Bobo and Kluegel 1993; Bobo, Kluegel and Smith 1997; Hughes 1997; Kinder and Sanders 1996).  In short, whites who perceive the causes of inequality to include structural factors—such as discrimination—are more likely to support race-based policies.  In addition, numerous scholars have argued that the very nature of racial prejudice has shifted from one based on perceptions of biological racial differences and strict segregation that were characteristic of the Jim Crow era, to one that is more contemporary and in which popular beliefs include a denial of the existence of race-based discrimination, persistent negative racial stereotypes that typically involve the sense that blacks violate cherished American values, and a belief in cultural—rather than biological—differences between racial groups.  This constellation of beliefs is the cornerstone of a new contemporary racial ideology, which has been variously labeled “colorblind racism,” “symbolic racism,” “modern racism,” “racial resentment,” and “laissez faire racism” (Bobo, Kluegel and Smith 1997; Kinder and Sanders 1996; Henry and Sears 2002; McConahay 1986; Bonilla-Silva and Forman 2000).  Beginning in the mid-1970s, survey researchers began to ask questions that tap into some of the key dimensions of this new ideology.   
 

The trends that were beginning to emerge in the late 1990s on questions of this type persist with the addition of data from the first few years of the new century.  For example, beginning in 1977, NORC asked respondents whether they thought each of four different reasons could explain why “blacks, on average, have worse jobs, income, and housing than whites.”  Updating the results reported in the 2^nd edition of the book, we see that the trend that began at the end of the 1990s has persisted up to 2006: each of the four explanations—discrimination, inborn ability to learn, education, and motivation are selected by fewer and fewer whites over time (see Figure 5).   
 

Figure 5. Trend in Whites’ Explanations for Inequality (NORC) 
 
 

This declining endorsement of all four explanations reflects a waning belief among whites in both biological (inborn ability to learn) and structural (education and discrimination) explanations of racial inequality.  The most popular explanation continues to be one focused on cultural characteristics:  that blacks lack the necessary motivation and willpower to succeed.  Interestingly, however, although it is the modal explanation for whites, it, too, has been declining in support over the past few decades.  This pattern—a declining significance of structural causes as well as of biological causes—is consistent with the transition from Jim Crow to more contemporary theories of racial ideology (Bobo, Kluegel and Smith 1997).  However, the declining support for the cultural explanation is unexplained by these theories.  One possibility is that an altogether different explanation has become salient in the contemporary racial climate—one that we have yet to include in survey questions.  A second possibility is that social desirability pressures—which are always of concern on questions that are sensitive—are at work.  It may be that this explanation has become viewed as increasingly inappropriate to admit to a survey interviewer in a contemporary climate of “colorblind” ideology (Forman 2004).  Unfortunately, the data do not permit us to adjudicate between these and other interpretations.