Only one question about the Principles of Equal Treatment has been replicated: in 1998 and 2000, NORC asked whether respondents thought there should be laws against interracial marriage. The trend remains steady with a persistent 12 percent of whites who would support such laws. Gallup has not repeated its intermarriage question since the 1997 edition was published, but NORC has begun to regularly ask whites if they would favor or oppose "having a close relative or family member marry a black person." As shown in the table, opposition has declined from 67 percent to 38 percent between 1990 and 2000.
Just two Implementation questions have been included in recent surveys: in 2000, ISR asked whether the government in Washington should make sure that black and white children go to the same schools and that blacks get "fair treatment" in jobs. With respect to schools, there has been some increase in those saying that the government should be involved, but this simply reflects a return to earlier levels. In other words, an apparent drop in support in 1994 seems not to have held up. The question on Federal Job Intervention shows a similar pattern. The overall message continues to be one of lukewarm support for government involvement in implementation, coupled with significant levels of "no interest" among whites.
While none of the original items
on Social Distance have been repeated, NORC has begun to ask, over what
is now a ten year period, whether respondents would favor or oppose "living
in a neighborhood where half of your neighbors were black." Opposition
to this has declined from 1990, when 48 percent opposed or strongly opposed,
to 31 percent who felt this way in 2000. Forty-six percent took the middle position
in 2000, indicating that they "neither favored nor opposed."
Questions on Explanations for Inequality are one of the types of questions
most consistently replicated by both ISR and NORC. The pattern for these items
is, again, one of little change. Just about one-third of whites agree that discrimination
is a cause of racial inequality; and about one-half agree that it is caused
by a lack of motivation and will power on the part of blacks. Gallup has continued
to ask a question related to this issue, but more directly about whether blacks
experience discrimination in the employment sector. In 1997 (the last year for
which data were reported in the earlier edition), there was some sign that whites
were less willing to acknowledge unfair treatment of blacks in the job market.
Data that have since become available now suggest that the drop was a real one-and
not just a one-time drop. In both 1998 and 1999, the percentage agreeing that
blacks "do not have as good a chance as whites to get any kind of job for
which they are qualified" dropped to just about 1 in 5.
Other items from the book on Perceptions of Discrimination have not been
replicated, but Gallup has asked two related questions that now meet our requirements
of at least three time points spanning a minimum of ten years. White respondents
show little change in the perception that "black children have as good
a chance as white children to get a good education," and the percentages
have hovered in the upper 80s since the item was first asked in 1962. In the
area of housing, fewer whites perceive that blacks have as good a chance as
whites to get housing; however, there is a trend towards denying unfair treatment
of blacks in this area: between 1989 and 2001, the percentage grew from 76 percent
to 86 percent.
Several questions about Government Expenditures and Preferential Treatment have been replicated-and the new time points suggest no change in attitudes among whites. Of the Miscellaneous questions, there is also little sign of shifting attitudes. There continues to be virtually no change in either direction with respect to whites' levels of reported warmth toward blacks (or toward whites). There is also no change in the perception that blacks "shouldn't push themselves where they're not wanted": 57-58 percent of whites disagreed with this statement in 1998 and 2000.
As with the original editions of this book, our ability to draw conclusions about black racial attitudes is severely hampered by the small sample sizes (in some cases) and possible confounding effects of race of interviewer. In our general discussion, below, we highlight this issue again. For the update, this problem is particularly acute in the case of the ISR items; thus, throughout, any apparent trends in these items must be considered cautiously.
Among African Americans, there is no change in the single Principle of Equal
Treatment question for which we have updated time points: overwhelming majorities
(96 percent) continue to oppose laws against interracial marriage. Just about
120 African Americans were asked the two Implementation questions in
the 2000 NES, and so what appears to be a drop in support for government intervention
in Schools and Jobs should be treated very cautiously because of the small sample
size, and because of mismatching of race of interviewer (see below for more
details).
As with whites, none of the Social Distance items have been repeated
since the 1997 edition. However, NORC has asked blacks how they feel about living
in a neighborhood in which "half of your neighbors are black" (recently
NORC has begun to ask about a neighborhood in which "half of your neighbors
are white" but there are no time trend data available.) The results for
this question show little movement over the ten year period from 1990 to 2000.
Among African Americans, as with whites, there has been little change in the
pattern of responses to the Explanations for Racial Inequality items
asked by NORC. In these items, about 2/3 of African Americans agree that discrimination
is a cause of racial inequality-a number unchanged since the 1996 time point.
None of the three other explanations for inequality show any shift of note.
The ISR questions on explanations of inequality suffer even more from small
sample size than the Implementation items, since just 62 respondents were asked
these questions. Gallup, however, does provide some useful data. With respect
to perceptions that blacks have "as good a chance as whites" for getting
a job, there has been little change over time. But similar questions (not included
in the original book, but for which there are now trend data) about education
and housing show that there is a decline in the belief among African Americans
that black children have as good a chance as whites for a good education-a return
in 2001 to the lower 1962 level, after some increase in the intervening years.
The same pattern holds for a question about blacks' chances for good housing.
A NORC question on support for federal spending and other aid to blacks shows
a slight decrease in the proportion of African Americans who report that the
levels are "too little." The seemingly dramatic drop in the ISR questions
about Government Expenditures and Racial Policies should be interpreted
cautiously, given the small sample size and possible race of interviewer effects.
African Americans show no consistent change in either direction with respect
to their levels of warmth toward whites or blacks. And, finally, the apparent
trend among blacks in the ISR "Black Push" item towards an increase
in agreement that "blacks should not push themselves where they do not
belong" should also be interpreted cautiously, given the methodological
problems.
Of the 51 unique survey questions included in the 1997 edition of this book, just 20 have new time points by the year 2002. The items that have been replicated are, not surprisingly, not randomly distributed across the 6 categories of questions. Those most frequently repeated were those in the categories of Explanations of Inequality and Perceptions of Discrimination (8 of 12 were repeated) and Affirmative Action (6 of 7 were replicated). All of these were categories introduced in the 2nd edition of our book. Among those with scant representation in this update are Principles of Equal Treatment (1 of 10 was asked again), Implementation (2 of 6 were replicated), and Social Distance (none of the nine were asked again). In the case of the Principles questions, the disappearance of most items can be attributed to their having reached a "ceiling" (6 of the 10 items in this section already had 90 percent or more respondents giving the liberal answer the last time it had been asked). The same cannot be said for the Implementation and Social Distance questions, where just 2 of the possible 15 items have been asked again, but where only 3 had reached this same "ceiling". Although the GSS has added two items that tap, in different ways, racial intermarriage attitudes and racial residential preferences, the disappearance of items from these categories is of concern if we are interested in maintaining a rich and diverse record on racial attitudes.
In surveys that are repeated over time, there is always a tension between keeping
questions so as to preserve trends, and adding new questions that have become
more relevant to the contemporary time period. The GSS and NES have no small
task in trying to achieve this delicate balancing act. With respect to "staying
current" the NES and GSS have added some important new racial attitude
items. Just as such shifts allowed us to add important new questions between
the 1st and 2nd edition of the book (questions about the causes of inequality,
perceptions of discrimination, preferential treatment, and government expenditures
on racial issues most notably), an examination of the latest codebooks from
the NES and GSS suggest that the 3rd edition of the book will add still more
new and important topics.
Using as a rule of thumb the presence of a question on at least three years
of the study, the following "new" topics have emerged: general attitudes
toward affirmative action (NES, Gallup), assessments of how close respondents
feel to members of other racial groups (NES/GSS), modern stereotype questions
asking respondents to evaluate different racial groups along a number of characteristics,
such as intelligence, trustworthiness, violence, and laziness (NES/GSS), perceptions
of "reverse discrimination" (GSS), assessments of how fairly blacks
are treated in a variety of domains of life (Gallup), explicit questions of
African Americans about their experiences with discrimination in a number of
areas of life (Gallup) and assessments of whether respondents think they-and
others in their community-are "prejudiced" (Gallup). Many of these
questions reflect important new areas of concern in race relations and racial
attitudes. In time, they will be a welcome addition to the picture we can obtain
from surveys of how racial attitudes have and have not changed over time.
However, there are hints of a pattern in the recent surveys that give some cause
for concern. In the first and second edition of the book, it was noted how different
survey organizations tended to ask different types of questions. However, what
was powerful about the existence of three separate sources of data (ISR, NORC,
and Gallup) was that, when combined, we had a survey record that reflected a
broad and diverse set of questions on a range of different dimensions of racial
attitudes. In recent years, however, there seems to be an increase in overlap
between survey organizations: virtually identical questions on closeness, stereotypes,
preferences in hiring and promotion for blacks, and whether African Americans
should "work their way up like the Irish, Italians and Jews, without special
favors," appear with regularity in both the General Social Survey and the
National Election Studies. Again, there is a delicate balance. While a replication
across different survey organizations provides useful information about the
reliability of the results, it also means that the survey record on racial attitudes
we are constructing may be more narrow and, therefore, less able to reflect
the complexity of racial attitudes. In short, we run the risk of creating a
less complete record for the future about how race has mattered in the hearts
and minds of Americans.
One feature of the historical record that can be constructed from survey data on racial attitudes is that the questions for which we have the longest record of data were developed at a point in time when it was believed that it was the "white" perspective that was most in need of study and when the racial system in this country was aptly described as "black and white". Over time, the relevance of the minority perspective became more fully appreciated and, more recently, the increasing racial/ethnic diversity of this country has made the "black-white" model outdated. There is some sign that the survey world is catching up to this. Race-related questions asked by NORC and ISR increasingly include items about the "melting pot," stereotype questions ask not just about whites and blacks, but also about Asians and Latinos, and surveys ask respondents about their attitudes toward living with whites, blacks, Asians and Latinos.
At the same time as there are signs of progress in questionnaire content, there
has not been equal attention to establishing a survey record that includes samples
with adequate numbers of African Americans and-even more of a problem-Latinos
and Asians. The Gallup Organization's new "Social Audit" series, begun
in 1997, is an important exception. Gallup now provides four years of data about
racial attitudes asked of just about 1,000 whites and 1,000 African Americans.
To our knowledge, this is one of the only datasets that will (assuming it continues)
provide the opportunity for detailed analyses of trends in black racial attitudes.
The General Social Survey, with its new "every other year" design
and bolstered sample sizes also helps in this matter. The NES, however, generally
interviews only about 150 African Americans. In addition,the introduction of
mode and question wording experiments, useful and important as they are, has
the unintended consequence of reducing by one-half the effective sample of African
Americans for some questions. Adequate samples of Latinos and Asians remain
virtually absent in the data available to us for tracking trends over time.
As our nation becomes more racially and ethnically diverse, there will be an
increasing need to diversify our survey samples in ways that enable comparisons
across-and within-these groups.
There is one important area where there are signs of retreat with respect to
the issue of more inclusive survey data: the problems posed by racial mismatching
of interviewer and respondent among African Americas, or at least consistency
and documentation in this regard. As noted in the discussion of results, the
problem of small sample sizes for African American responses is magnified because
of the issue of possible race of interviewer effects. Where we observe an apparent
decline in support among African Americans for the Implementation questions,
for example, there are at least three possible interpretations: (1) African
Americans have, in fact, become less interested in having the government involved
in these areas; (2) the racial composition of the interviewing staff has shifted,
so that higher proportions of African Americans are interviewed by white interviewers
(to whom they are reluctant to report liberal attitudes toward racial policies);
or (3) even in the face of no change in the proportion of blacks interviewed
by whites, it is possible that a particular issue that was not sensitive to
race of interviewer has become sensitive over time.
In the case of NORC, the key problem is that there is no information available
to analysts about the race of the interviewer. Thus, the existence and/or magnitude
of the problem is unknown. The NES does provide information on the race of the
interviewer. But what we learn from these data is not encouraging. Specifically,
both differences and changes in the degree of racial matching have become quite
severe. In 1998, 89 percent of the African American respondents were interviewed
by a white interviewer; in 2000, all of the African American respondents who
were interviewed in person (these are the only data reported above, in the update,
so as to maintain comparability in mode across the time series) were interviewed
by a white interviewer. We know from research that African Americans are less
likely to state support for various racial policies when interviewed by a white
interviewer, and so it is quite likely that these data understate the levels
of support.
Thus, progress is being made in the area of including survey questions that
ask about other racial/ethnic groups and issues of relevance to minority perspectives,
but there is stagnation and some hints of a retreat when it comes to the issue
of adequately representing the diversity of racial/ethnic groups in our samples
and/or overcoming the methodological challenges of what is an age-old problem
in survey research on racial attitudes: race of interviewer effects. The latter
is no small logistical problem as well: to be sure, it is difficult and expensive
to match race of interviewer and respondent. However, the effects of not doing
so on the quality of the data are consequential enough as to merit concerted
attention and resources toward overcoming it, if we are interested in a more
inclusive and accurate survey record for the future.
Maria Krysan
April 2002
*Bibliographic citations to this website and update should be made as follows:
Krysan, Maria. 2002. "Data Update to Racial Attitudes in America." An update and website to complement Racial Attitudes in America: Trends and Interpretations, Revised Edition, by Howard Schuman, Charlotte Steeh, Lawrence Bobo and Maria Krysan, 1997, Harvard University Press. http://tigger.uic.edu/~krysan/racialattitudes.htm.
The reference for the volume for which this is an update is:
Schuman, Howard, Charlotte Steeh, Lawrence Bobo, and Maria Krysan. 1997. Racial Attitudes in America: Trends and Interpretations, Revised Edition. Cambridge, MA: Harvard University Press.
Return to Racial Attitudes Homepage