Skip to content
Contents
The Public Mind

    Type to search. Planned chapters match by title only.

    Chapter 5.1

    Fifty years of racial attitudes

    Among white respondents, equal-treatment answers rose from the 1970s to the 2000s, answers on discrimination and on government help rose between the 2014 and 2016 surveys, more among Democrats than among Republicans, and the tie between racial resentment and the Republican presidential vote changed by +1.1 points from 2008 to 2012, with an interval whose upper end is 4.6, and by +8.2 from 2012 to 2016.

    Affiliation
    NYU Stern
    Published

    Among white respondents to the General Social Survey (GSS), the share who oppose a law against marriages between Black people and whites rose from 62.6%62.6% percent of respondents giving the answerEstimandweighted share of white respondents who gave the answer: no (no law against black-white marriage)Interval95% CI 59.7% to 65.4%n1,306Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the itemPeriod1972SourceGeneral Social Survey in 1972 to 87.6%87.6% percent of respondents giving the answerEstimandweighted share of white respondents who gave the answer: no (no law against black-white marriage)Interval95% CI 84.9% to 90.3%n742Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the itemPeriod2002SourceGeneral Social Survey in 2002, the last year the question was asked, and the share who would not oppose a close relative marrying a Black person rose from 35.0%35.0% percent of respondents giving the answerEstimandweighted share of white respondents who gave the answer: favor, strongly favor or neither favor nor oppose (would not oppose a relative marrying a black person)Interval95% CI 31.8% to 38.1%n1,130Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the itemPeriod1990SourceGeneral Social Survey in 1990 to 92.4%92.4% percent of respondents giving the answerEstimandweighted share of white respondents who gave the answer: favor, strongly favor or neither favor nor oppose (would not oppose a relative marrying a black person)Interval95% CI 90.5% to 94.2%n1,517Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the itemPeriod2024SourceGeneral Social Survey in 2024. Items on the causes of racial inequality and on government action followed another schedule. The share who say that Black people's worse jobs, income and housing are mainly due to discrimination fell from 42.0%42.0% percent of respondents giving the answerEstimandweighted share of white respondents who gave the answer: yes (discrimination)Interval95% CI 39.0% to 45.0%n1,273Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the itemPeriod1977SourceGeneral Social Survey in 1977 to 28.8%28.8% percent of respondents giving the answerEstimandweighted share of white respondents who gave the answer: yes (discrimination)Interval95% CI 24.9% to 32.7%n694Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the itemPeriod2004SourceGeneral Social Survey in 2004 and was 30.0%30.0% percent of respondents giving the answerEstimandweighted share of white respondents who gave the answer: yes (discrimination)Interval95% CI 27.0% to 33.0%n1,215Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the itemPeriod2014SourceGeneral Social Survey in 2014, then rose to 37.1%37.1% percent of respondents giving the answerEstimandweighted share of white respondents who gave the answer: yes (discrimination)Interval95% CI 34.0% to 40.2%n1,332Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the itemPeriod2016SourceGeneral Social Survey in 2016 and 41.2%41.2% percent of respondents giving the answerEstimandweighted share of white respondents who gave the answer: yes (discrimination)Interval95% CI 37.4% to 45.1%n1,076Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the itemPeriod2018SourceGeneral Social Survey in 2018. The share who place themselves at the help end of a five-point scale on whether government has a special obligation to improve Black people's living standards (the two answers nearest to government should help) went from 11.2%11.2% percent of respondents giving the answerEstimandweighted share of white respondents who gave the answer: 1 or 2 on a five-point scale from government should help (1) to no special treatment (5) (places self at the help end of the government-help scale (1 or 2 of 5))Interval95% CI 9.2% to 13.2%n1,216Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the itemPeriod2014SourceGeneral Social Survey in 2014 to 21.4%21.4% percent of respondents giving the answerEstimandweighted share of white respondents who gave the answer: 1 or 2 on a five-point scale from government should help (1) to no special treatment (5) (places self at the help end of the government-help scale (1 or 2 of 5))Interval95% CI 18.8% to 24.0%n1,382Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the itemPeriod2016SourceGeneral Social Survey in 2016.

    A single step fitted to the in-person interviews of the 2000 to 2018 surveys falls at the 2016 survey for each of six dated explanation and policy items among white respondents, which puts the new level between the 2014 and 2016 surveys. Of the 1,0001,000 bootstrap replicatesEstimandnumber of seeded bootstrap replicates behind each step dateWeightsweight_use (wtssnrps from 2004, wtssps before)Populationeach group and itemPeriodin-person surveys 2000 to 2018SourceGeneral Social Survey bootstrap replicates for each item, the share that choose 2016 runs from 51.4%51.4% percent of bootstrap replicatesEstimandsmallest share of replicates that place the best single step at the 2016 surveyn1,000Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents, six itemsPeriodin-person surveys 2000 to 2018SourceGeneral Social Survey to 100.0%100.0% percent of bootstrap replicatesEstimandlargest share of replicates that place the best single step at the 2016 surveyn1,000Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents, six itemsPeriodin-person surveys 2000 to 2018SourceGeneral Social Survey. A step is a form assumption, and a kinked trend fits some of the items better. The 2014 and 2016 surveys were conducted mostly in person, so the date does not come from a change of survey mode. The movement is larger among white Democrats: their share attributing racial inequality to discrimination went from 44.9%44.9% percent of respondents giving the answerEstimandweighted share who gave the answer (discrimination)Interval95% CI 39.6% to 50.2%n471Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite democrats (leaners included; GSS race code, Hispanic whites included) who answered the itemPeriod2014SourceGeneral Social Survey in 2014 to 61.6%61.6% percent of respondents giving the answerEstimandweighted share who gave the answer (discrimination)Interval95% CI 55.5% to 67.6%n392Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite democrats (leaners included; GSS race code, Hispanic whites included) who answered the itemPeriod2018SourceGeneral Social Survey in 2018, and white Republicans' share went from 18.2%18.2% percent of respondents giving the answerEstimandweighted share who gave the answer (discrimination)Interval95% CI 14.2% to 22.2%n468Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite republicans (leaners included; GSS race code, Hispanic whites included) who answered the itemPeriod2014SourceGeneral Social Survey to 23.3%23.3% percent of respondents giving the answerEstimandweighted share who gave the answer (discrimination)Interval95% CI 18.0% to 28.5%n458Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite republicans (leaners included; GSS race code, Hispanic whites included) who answered the itemPeriod2018SourceGeneral Social Survey.

    The tie between racial attitudes and the presidential vote has its own timetable. In the American National Election Studies (ANES), a standard deviation more racial resentment goes with 22.122.1 percentage points of the Republican two-party presidential vote per standard deviation of the attitudeEstimandweighted linear probability slope on the standardized attitude (racial resentment index; party not held fixed)Interval95% CI 19.2 to 25.0n782Weightsweight (VCF0009z)Populationwhite non-Hispanic voters for a major-party candidatePeriod2008SourceANES Cumulative Data File points more Republican share of the two-party vote among white voters in 2008, 23.223.2 percentage points of the Republican two-party presidential vote per standard deviation of the attitudeEstimandweighted linear probability slope on the standardized attitude (racial resentment index; party not held fixed)Interval95% CI 21.3 to 25.0n2,541Weightsweight (VCF0009z)Populationwhite non-Hispanic voters for a major-party candidatePeriod2012SourceANES Cumulative Data File in 2012 and 31.331.3 percentage points of the Republican two-party presidential vote per standard deviation of the attitudeEstimandweighted linear probability slope on the standardized attitude (racial resentment index; party not held fixed)Interval95% CI 29.9 to 32.8n1,942Weightsweight (VCF0009z)Populationwhite non-Hispanic voters for a major-party candidatePeriod2016SourceANES Cumulative Data File in 2016. The change from 2008 to 2012 is +1.1 (95% CI −2.4 to 4.6)+1.1 percentage points of the Republican two-party presidential vote per standard deviation of the attitudeEstimandchange in the weighted linear probability slope on the standardized attitude (party not held fixed)Interval95% CI −2.4 to 4.6n3,323Weightsweight (VCF0009z)Populationwhite non-Hispanic voters for a major-party candidatePeriod2008 to 2012SourceANES Cumulative Data File points, an interval that allows a rise of up to 4.64.6 percentage points of the Republican two-party presidential vote per standard deviation of the attitudeEstimandupper end of the 95% interval for the change in the weighted linear probability slope (racial resentment index)n3,323Weightsweight (VCF0009z)Populationwhite non-Hispanic voters for a major-party candidatePeriod2008 to 2012SourceANES Cumulative Data File points but not one as large as the change from 2012 to 2016, +8.2 (95% CI 5.8 to 10.5)+8.2 percentage points of the Republican two-party presidential vote per standard deviation of the attitudeEstimandchange in the weighted linear probability slope on the standardized attitude (party not held fixed)Interval95% CI 5.8 to 10.5n4,483Weightsweight (VCF0009z)Populationwhite non-Hispanic voters for a major-party candidatePeriod2012 to 2016SourceANES Cumulative Data File points. This is an association across elections and not an effect of the Obama presidency.

    The GSS is a repeated cross-section of US adults, interviewed in person (with some telephone interviews from 2004) through 2018, mostly on the web in 2021, and in person and on the web in 2022 and 2024. The ANES interviewed mainly in person until 2008, added a web sample in 2012, and held no in-person interviews in 2020. The Cooperative Election Study (CES), an online opt-in panel, gives a second reading of the vote in 2008 and 2012. Results for race and party are shown separately. "White respondents" are white in the GSS race coding, which includes Hispanic whites (11.4%11.4% percent of white respondents who are HispanicEstimandweighted share who report Hispanic originInterval95% CI 10.8% to 12.1%n15,012Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite GSS respondents who answered the Hispanic-origin questionPeriod2012 to 2024SourceGeneral Social Survey of white respondents from 2012 to 2024), and white non-Hispanic in the ANES and CES. Party includes leanersPartisan leanerA respondent who first says they are independent but, when asked, says they lean toward one party. Most figures group leaners with the party they lean toward, and each chapter says which grouping it uses..Every share and mean is weighted, with 95% intervals from the Kish effective sample sizeEffective sample sizeThe number of equally weighted respondents that would give the same precision as the weighted sample. It is smaller than the raw count when the weights vary a lot, and it sets how wide the confidence interval is. [1]. The intervals treat the weighted sample as a simple random sample and ignore the surveys' clustering, and the CES intervals also ignore how respondents join an opt-in panel. The methods note in the appendix gives the sample floors, the item wording and the years each item was asked. Two neighboring chapters cover parts of this ground and are summarized here and not rebuilt. Chapter 5.2, Racial attitudes and ideology relates the ANES racial resentmentRacial resentmentA survey scale, developed by Donald Kinder and Lynn Sanders, that measures moral judgments about the work ethic and deservingness of Black Americans. Higher scores mean more resentment. index to ideology between and within parties, and Chapter 1.3, Sort dates by attitude dates the year in which the Republican minus Democratic gap on resentment passes a fixed threshold. This chapter dates the level of each item among white respondents, which is a different quantity, and uses the resentment index only in the section on the vote. Chapter 5.3, Racial attitudes after George Floyd's death (planned) covers the weeks after May 2020.

    Four kinds of items

    The surveys ask four kinds of question. Principles of equal treatment ask whether Black and white people should be treated alike, for example in marriage, in housing and in a party's choice of a presidential nominee. Explanations of racial inequality ask whether the gap in jobs, income and housing is mainly due to discrimination, to inborn ability, to a lack of chance for education or to a lack of motivation or will power. Policy implementation asks what government should do about the gap: help improve Black people's living standards, spend on it, or give preference in hiring and promotion. Affect is the ANES feeling thermometerFeeling thermometerA 0 to 100 rating of a group or person, where 0 is the coldest feeling and 100 the warmest. The American National Election Studies ask respondents to rate parties, candidates and social groups this way., a rating of a group on a scale from zero to one hundred. Schuman, Steeh, Bobo and Krysan separate principles of equality from government implementation of them and add questions about the sources of inequality [2], and Kluegel traced the GSS explanations through 1989 [3]. Each GSS series here is the weighted share of white respondents, among those who answered, who gave the answer named in its label. The government-help question is a five-point scale, and its series counts the two answers nearest to government should help. The chapter dates the answers on discrimination, will power, government help, spending, affirmative action and the special-favors statement, and splits them by party and race. It calls them the dated items.

    Equal-treatment answers rose from the 1970s to the 2000s

    Figure 1 plots the answers of white respondents by kind. The share who oppose a law against marriages between Black people and whites went from 62.6%62.6% percent of respondents giving the answerEstimandweighted share of white respondents who gave the answer: no (no law against black-white marriage)Interval95% CI 59.7% to 65.4%n1,306Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the itemPeriod1972SourceGeneral Social Survey in 1972 to 87.6%87.6% percent of respondents giving the answerEstimandweighted share of white respondents who gave the answer: no (no law against black-white marriage)Interval95% CI 84.9% to 90.3%n742Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the itemPeriod2002SourceGeneral Social Survey in 2002. The share who disagree that white people have a right to keep Black people out of their neighborhoods went from 60.8%60.8% percent of respondents giving the answerEstimandweighted share of white respondents who gave the answer: disagree slightly or strongly (whites have no right to exclude black neighbors)Interval95% CI 57.8% to 63.7%n1,251Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the itemPeriod1972SourceGeneral Social Survey in 1972 to 86.8%86.8% percent of respondents giving the answerEstimandweighted share of white respondents who gave the answer: disagree slightly or strongly (whites have no right to exclude black neighbors)Interval95% CI 84.1% to 89.5%n746Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the itemPeriod1996SourceGeneral Social Survey in 1996, and the share who would vote for a qualified Black nominee of their own party went from 74.0%74.0% percent of respondents giving the answerEstimandweighted share of white respondents who gave the answer: yes (would vote for a qualified black nominee)Interval95% CI 71.3% to 76.6%n1,262Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the itemPeriod1972SourceGeneral Social Survey in 1972 to 95.7%95.7% percent of respondents giving the answerEstimandweighted share of white respondents who gave the answer: yes (would vote for a qualified black nominee)Interval95% CI 94.2% to 97.2%n1,052Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the itemPeriod2010SourceGeneral Social Survey in 2010. The file has no readings of those questions after those years. The two equal-treatment items still asked in 2024 kept rising after 2012: the share who would vote for a law barring housing discrimination went from 35.5%35.5% percent of respondents giving the answerEstimandweighted share of white respondents who gave the answer: the law that bars refusing to sell because of race (would vote for a law barring housing discrimination)Interval95% CI 32.6% to 38.3%n1,283Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the itemPeriod1973SourceGeneral Social Survey in 1973 to 68.5%68.5% percent of respondents giving the answerEstimandweighted share of white respondents who gave the answer: the law that bars refusing to sell because of race (would vote for a law barring housing discrimination)Interval95% CI 64.9% to 72.0%n966Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the itemPeriod2012SourceGeneral Social Survey in 2012 and 82.8%82.8% percent of respondents giving the answerEstimandweighted share of white respondents who gave the answer: the law that bars refusing to sell because of race (would vote for a law barring housing discrimination)Interval95% CI 78.5% to 87.0%n790Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the itemPeriod2022SourceGeneral Social Survey in 2022, and was 77.7% (95% CI 70.7% to 84.6%)77.7% percent of respondents giving the answerEstimandweighted share of white respondents who gave the answer: the law that bars refusing to sell because of race (would vote for a law barring housing discrimination)Interval95% CI 70.7% to 84.6%n279Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the itemPeriod2024SourceGeneral Social Survey in 2024, when the item went to fewer respondents. The share who would not oppose a relative marrying a Black person went from 78.4%78.4% percent of respondents giving the answerEstimandweighted share of white respondents who gave the answer: favor, strongly favor or neither favor nor oppose (would not oppose a relative marrying a black person)Interval95% CI 75.2% to 81.6%n972Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the itemPeriod2012SourceGeneral Social Survey in 2012 to 92.4%92.4% percent of respondents giving the answerEstimandweighted share of white respondents who gave the answer: favor, strongly favor or neither favor nor oppose (would not oppose a relative marrying a black person)Interval95% CI 90.5% to 94.2%n1,517Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the itemPeriod2024SourceGeneral Social Survey in 2024.

    Figure 2 gives the change between pooled periods of surveys for every item. The equal-treatment items changed most early: from 1972-1977 to 1988-1994 the open-housing item changed by +23.7 (95% CI 21.6 to 25.9)+23.7 percentage pointsEstimandchange in the weighted share who gave the answer (would vote for a law barring housing discrimination) between two pooled periods of surveysInterval95% CI 21.6 to 25.9n10,051Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the item, pooled periodsPeriod1972-1977 to 1988-1994SourceGeneral Social Survey points, the segregation item by +20.3 (95% CI 18.2 to 22.3)+20.3 percentage pointsEstimandchange in the weighted share who gave the answer (whites have no right to exclude black neighbors) between two pooled periods of surveysInterval95% CI 18.2 to 22.3n9,565Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the item, pooled periodsPeriod1972-1977 to 1988-1994SourceGeneral Social Survey and the intermarriage-law item by +14.3 (95% CI 12.6 to 15.9)+14.3 percentage pointsEstimandchange in the weighted share who gave the answer (no law against black-white marriage) between two pooled periods of surveysInterval95% CI 12.6 to 15.9n13,447Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the item, pooled periodsPeriod1972-1977 to 1988-1994SourceGeneral Social Survey. The shares saying yes to the explanations fell. Over the same years the share attributing the gap to inborn ability changed by −9.6 (95% CI −12.4 to −6.7)−9.6 percentage pointsEstimandchange in the weighted share who gave the answer (inborn ability) between two pooled periods of surveysInterval95% CI −12.4 to −6.7n7,246Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the item, pooled periodsPeriod1972-1977 to 1988-1994SourceGeneral Social Survey points and the share attributing it to will power by −4.5 (95% CI −7.8 to −1.3)−4.5 percentage pointsEstimandchange in the weighted share who gave the answer (lack of motivation or will power) between two pooled periods of surveysInterval95% CI −7.8 to −1.3n7,103Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the item, pooled periodsPeriod1972-1977 to 1988-1994SourceGeneral Social Survey, and from 1988-1994 to 2004-2008 the share citing discrimination changed by −7.1 (95% CI −9.4 to −4.7)−7.1 percentage pointsEstimandchange in the weighted share who gave the answer (discrimination) between two pooled periods of surveysInterval95% CI −9.4 to −4.7n8,958Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the item, pooled periodsPeriod1988-1994 to 2004-2008SourceGeneral Social Survey points and the share citing will power by −9.3 (95% CI −11.8 to −6.7)−9.3 percentage pointsEstimandchange in the weighted share who gave the answer (lack of motivation or will power) between two pooled periods of surveysInterval95% CI −11.8 to −6.7n8,880Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the item, pooled periodsPeriod1988-1994 to 2004-2008SourceGeneral Social Survey.

    Figure 2

    Equal-treatment items changed most before 2008, and the policy items changed most from 2010-2014 on

    Change in the share of white respondents (GSS race code, Hispanic whites included) giving the answer named in each row, percentage points, between pooled periods of surveys; empty cells are items not asked in both periods

    −40.00.0+40.0
    Faded: the interval includes 0.0No estimate
    A grid of items by five pairs of periods, shaded by the change in the share of white respondents giving the named answer. The equal-treatment rows carry the largest positive changes in the first two columns. The explanation rows are negative in the first two columns, and the discrimination and education rows turn positive in the fourth column. The policy rows are near zero in the third column and positive in the fourth, with the spending row the largest.
    Row1972-1977 to 1988-19941988-1994 to 2004-20082004-2008 to 2010-20142010-2014 to 2016-20182016-2018 to 2021-2024
    Principle: no law against intermarriage+14.3, interval +12.6 to +15.9, n = 13,447No estimateNo estimateNo estimateNo estimate
    Principle: would vote for a Black nominee+6.7, interval +5.1 to +8.3, n = 10,621+7.7, interval +5.7 to +9.8, n = 6,579+1.7, interval −0.7 to +4.0, n = 2,050No estimateNo estimate
    Principle: no right to exclude neighbors+20.3, interval +18.2 to +22.3, n = 9,565No estimateNo estimateNo estimateNo estimate
    Principle: open-housing law+23.7, interval +21.6 to +25.9, n = 10,051+7.4, interval +5.0 to +9.8, n = 9,327+3.4, interval +0.7 to +6.1, n = 6,393+7.2, interval +4.4 to +10.0, n = 5,696+4.3, interval +0.1 to +8.5, n = 3,570
    Principle: accepts a relative's marriageNo estimate+36.6, interval +33.0 to +40.2, n = 4,301+8.6, interval +6.2 to +11.1, n = 6,445+5.3, interval +3.0 to +7.7, n = 5,763+7.0, interval +5.0 to +9.0, n = 7,741
    Explanation: discrimination−5.0, interval −8.3 to −1.7, n = 7,178−7.1, interval −9.4 to −4.7, n = 8,958+0.6, interval −2.1 to +3.3, n = 6,234+8.4, interval +5.3 to +11.5, n = 5,589+9.2, interval +5.9 to +12.5, n = 5,964
    Explanation: inborn ability−9.6, interval −12.4 to −6.7, n = 7,246−7.9, interval −9.5 to −6.4, n = 9,096−0.3, interval −1.9 to +1.3, n = 6,352−1.5, interval −3.1 to +0.2, n = 5,695−1.7, interval −3.3 to −0.2, n = 6,046
    Explanation: lack of chance for education−1.2, interval −4.6 to +2.1, n = 7,282−6.5, interval −9.0 to −4.0, n = 9,093−1.9, interval −4.8 to +1.0, n = 6,315+9.1, interval +5.9 to +12.4, n = 5,661−1.0, interval −4.4 to +2.3, n = 5,995
    Explanation: lack of will power−4.5, interval −7.8 to −1.3, n = 7,103−9.3, interval −11.8 to −6.7, n = 8,880−2.5, interval −5.4 to +0.5, n = 6,189−9.0, interval −12.2 to −5.7, n = 5,556−10.4, interval −13.6 to −7.2, n = 5,935
    Policy: help end of the government-help scale−7.2, interval −9.8 to −4.6, n = 7,353−1.9, interval −3.6 to −0.3, n = 9,200+0.4, interval −1.5 to +2.3, n = 6,315+10.0, interval +7.6 to +12.4, n = 5,643+4.9, interval +2.3 to +7.6, n = 7,577
    Policy: too little spending+5.7, interval +3.7 to +7.7, n = 10,047−2.5, interval −5.1 to +0.1, n = 6,601−1.7, interval −4.7 to +1.3, n = 4,806+20.9, interval +17.2 to +24.5, n = 3,867−4.1, interval −7.8 to −0.4, n = 5,520
    Policy: preferential hiringNo estimate+0.6, interval −1.8 to +3.0, n = 4,215+2.1, interval +0.1 to +4.0, n = 6,216+5.7, interval +3.3 to +8.2, n = 5,500+2.1, interval −0.4 to +4.6, n = 7,373
    Policy: rejects the special-favors viewNo estimate+0.6, interval −1.7 to +3.0, n = 4,346+1.4, interval −0.5 to +3.3, n = 6,427+8.2, interval +5.8 to +10.6, n = 5,748+7.3, interval +4.7 to +9.9, n = 7,691
    Show data
    Data for Equal-treatment items changed most before 2008, and the policy items changed most from 2010-2014 on
    colconf_highconf_lowestimatenrowview
    1972-1977 to 1988-199415.9012.6014.3013,447Principle: no law against intermarriageall
    1972-1977 to 1988-19948.305.106.7010,621Principle: would vote for a Black nomineeall
    1988-1994 to 2004-20089.805.707.706,579Principle: would vote for a Black nomineeall
    2004-2008 to 2010-20144−0.7001.702,050Principle: would vote for a Black nomineeall
    1972-1977 to 1988-199422.3018.2020.309,565Principle: no right to exclude neighborsall
    1972-1977 to 1988-199425.9021.6023.7010,051Principle: open-housing lawall
    1988-1994 to 2004-20089.8057.409,327Principle: open-housing lawall
    2004-2008 to 2010-20146.100.7003.406,393Principle: open-housing lawall
    2010-2014 to 2016-2018104.407.205,696Principle: open-housing lawall
    2016-2018 to 2021-20248.500.1004.303,570Principle: open-housing lawall
    1988-1994 to 2004-200840.203336.604,301Principle: accepts a relative's marriageall
    2004-2008 to 2010-201411.106.208.606,445Principle: accepts a relative's marriageall
    2010-2014 to 2016-20187.7035.305,763Principle: accepts a relative's marriageall
    2016-2018 to 2021-20249577,741Principle: accepts a relative's marriageall
    1972-1977 to 1988-1994−1.70−8.30−57,178Explanation: discriminationall
    1988-1994 to 2004-2008−4.70−9.40−7.108,958Explanation: discriminationall
    2004-2008 to 2010-20143.30−2.100.6006,234Explanation: discriminationall
    2010-2014 to 2016-201811.505.308.405,589Explanation: discriminationall
    2016-2018 to 2021-202412.505.909.205,964Explanation: discriminationall
    1972-1977 to 1988-1994−6.70−12.40−9.607,246Explanation: inborn abilityall
    1988-1994 to 2004-2008−6.40−9.50−7.909,096Explanation: inborn abilityall
    2004-2008 to 2010-20141.30−1.90−0.3006,352Explanation: inborn abilityall
    2010-2014 to 2016-20180.200−3.10−1.505,695Explanation: inborn abilityall
    2016-2018 to 2021-2024−0.200−3.30−1.706,046Explanation: inborn abilityall
    1972-1977 to 1988-19942.10−4.60−1.207,282Explanation: lack of chance for educationall
    1988-1994 to 2004-2008−4−9−6.509,093Explanation: lack of chance for educationall
    2004-2008 to 2010-20141−4.80−1.906,315Explanation: lack of chance for educationall
    2010-2014 to 2016-201812.405.909.105,661Explanation: lack of chance for educationall
    2016-2018 to 2021-20242.30−4.40−15,995Explanation: lack of chance for educationall
    1972-1977 to 1988-1994−1.30−7.80−4.507,103Explanation: lack of will powerall
    1988-1994 to 2004-2008−6.70−11.80−9.308,880Explanation: lack of will powerall
    2004-2008 to 2010-20140.500−5.40−2.506,189Explanation: lack of will powerall
    2010-2014 to 2016-2018−5.70−12.20−95,556Explanation: lack of will powerall
    2016-2018 to 2021-2024−7.20−13.60−10.405,935Explanation: lack of will powerall
    1972-1977 to 1988-1994−4.60−9.80−7.207,353Policy: help end of the government-help scaleall
    1988-1994 to 2004-2008−0.300−3.60−1.909,200Policy: help end of the government-help scaleall
    2004-2008 to 2010-20142.30−1.500.4006,315Policy: help end of the government-help scaleall
    2010-2014 to 2016-201812.407.60105,643Policy: help end of the government-help scaleall
    2016-2018 to 2021-20247.602.304.907,577Policy: help end of the government-help scaleall
    1972-1977 to 1988-19947.703.705.7010,047Policy: too little spendingall
    1988-1994 to 2004-20080.100−5.10−2.506,601Policy: too little spendingall
    2004-2008 to 2010-20141.30−4.70−1.704,806Policy: too little spendingall
    2010-2014 to 2016-201824.5017.2020.903,867Policy: too little spendingall
    2016-2018 to 2021-2024−0.400−7.80−4.105,520Policy: too little spendingall
    1988-1994 to 2004-20083−1.800.6004,215Policy: preferential hiringall
    2004-2008 to 2010-201440.1002.106,216Policy: preferential hiringall
    2010-2014 to 2016-20188.203.305.705,500Policy: preferential hiringall
    2016-2018 to 2021-20244.60−0.4002.107,373Policy: preferential hiringall
    1988-1994 to 2004-20083−1.700.6004,346Policy: rejects the special-favors viewall
    2004-2008 to 2010-20143.30−0.5001.406,427Policy: rejects the special-favors viewall
    2010-2014 to 2016-201810.605.808.205,748Policy: rejects the special-favors viewall
    2016-2018 to 2021-20249.904.707.307,691Policy: rejects the special-favors viewall
    1972-1977 to 1988-199415.9012.6014.3013,447Principle: no law against intermarriageinperson
    1972-1977 to 1988-19948.305.106.7010,621Principle: would vote for a Black nomineeinperson
    1988-1994 to 2004-20089.805.507.706,440Principle: would vote for a Black nomineeinperson
    2004-2008 to 2010-20144−0.9001.501,800Principle: would vote for a Black nomineeinperson
    1972-1977 to 1988-199422.3018.2020.309,565Principle: no right to exclude neighborsinperson
    1972-1977 to 1988-199425.9021.6023.7010,051Principle: open-housing lawinperson
    1988-1994 to 2004-2008105.107.608,950Principle: open-housing lawinperson
    2004-2008 to 2010-20145.40−0.4002.505,597Principle: open-housing lawinperson
    2010-2014 to 2016-201810.304.307.305,064Principle: open-housing lawinperson
    2016-2018 to 2022-202410.601.606.103,175Principle: open-housing lawinperson
    1988-1994 to 2004-200839.4032.1035.803,915Principle: accepts a relative's marriageinperson
    2004-2008 to 2010-201411.706.409.105,645Principle: accepts a relative's marriageinperson
    2010-2014 to 2016-20188.1035.605,128Principle: accepts a relative's marriageinperson
    2016-2018 to 2022-20248.902.905.903,437Principle: accepts a relative's marriageinperson
    1972-1977 to 1988-1994−1.70−8.30−57,178Explanation: discriminationinperson
    1988-1994 to 2004-2008−4.50−9.40−6.908,594Explanation: discriminationinperson
    2004-2008 to 2010-20144.40−1.401.505,470Explanation: discriminationinperson
    2010-2014 to 2016-2018125.408.704,980Explanation: discriminationinperson
    2016-2018 to 2022-202418.705.30122,747Explanation: discriminationinperson
    1972-1977 to 1988-1994−6.70−12.40−9.607,246Explanation: inborn abilityinperson
    1988-1994 to 2004-2008−5.70−9−7.308,719Explanation: inborn abilityinperson
    2004-2008 to 2010-20141.20−2.30−0.6005,568Explanation: inborn abilityinperson
    2010-2014 to 2016-20180.400−3.10−1.305,069Explanation: inborn abilityinperson
    2016-2018 to 2022-20245.10−2.201.502,807Explanation: inborn abilityinperson
    1972-1977 to 1988-19942.10−4.60−1.207,282Explanation: lack of chance for educationinperson
    1988-1994 to 2004-2008−4.20−9.40−6.808,718Explanation: lack of chance for educationinperson
    2004-2008 to 2010-20142−4.20−1.105,533Explanation: lack of chance for educationinperson
    2010-2014 to 2016-201811.805.108.505,040Explanation: lack of chance for educationinperson
    2016-2018 to 2022-20243.70−9.70−32,780Explanation: lack of chance for educationinperson
    1972-1977 to 1988-1994−1.30−7.80−4.507,103Explanation: lack of will powerinperson
    1988-1994 to 2004-2008−5.70−11−8.308,515Explanation: lack of will powerinperson
    2004-2008 to 2010-2014−0.200−6.50−3.405,428Explanation: lack of will powerinperson
    2010-2014 to 2016-2018−5.30−12.10−8.704,949Explanation: lack of will powerinperson
    2016-2018 to 2022-2024−0.800−13.40−7.102,733Explanation: lack of will powerinperson
    1972-1977 to 1988-1994−4.60−9.80−7.207,353Policy: help end of the government-help scaleinperson
    1988-1994 to 2004-2008−0.400−3.70−2.108,858Policy: help end of the government-help scaleinperson
    2004-2008 to 2010-20142.80−1.300.8005,554Policy: help end of the government-help scaleinperson
    2010-2014 to 2016-201811.906.809.305,018Policy: help end of the government-help scaleinperson
    2016-2018 to 2022-20248.40−0.5003.903,338Policy: help end of the government-help scaleinperson
    1972-1977 to 1988-19947.703.705.7010,047Policy: too little spendinginperson
    1988-1994 to 2004-2008−1.20−6.60−3.906,227Policy: too little spendinginperson
    2004-2008 to 2010-20143.40−30.2004,144Policy: too little spendinginperson
    2010-2014 to 2016-201824.3016.5020.403,412Policy: too little spendinginperson
    2016-2018 to 2022-20245.10−7.70−1.302,309Policy: too little spendinginperson
    1988-1994 to 2004-20083.10−1.900.6003,847Policy: preferential hiringinperson
    2004-2008 to 2010-20144.300.2002.205,453Policy: preferential hiringinperson
    2010-2014 to 2016-20188.703.506.104,899Policy: preferential hiringinperson
    2016-2018 to 2022-202410.601.706.203,192Policy: preferential hiringinperson
    1988-1994 to 2004-20083−1.700.7003,966Policy: rejects the special-favors viewinperson
    2004-2008 to 2010-20142.90−10.9005,643Policy: rejects the special-favors viewinperson
    2010-2014 to 2016-201811.406.408.905,127Policy: rejects the special-favors viewinperson
    2016-2018 to 2022-202410.101.505.803,405Policy: rejects the special-favors viewinperson
    The equal-treatment rows have their largest cells in the first two columns and the policy rows have theirs in the last two, with the sign of each change printed in its cell. Each cell is the share in the later period minus the share in the earlier period for white respondents, and cells whose 95% interval includes zero are faded.Note: Periods pool the surveys of 1972 to 1977, 1988 to 1994, 2004 to 2008, 2010 to 2014, 2016 to 2018 and 2021 to 2024, after rescaling each year's weights to average one. The last column pools 2021, 2022 and 2024, mostly web and mixed-mode interviews, so a change into it can come from the change of mode; the in-person view uses the in-person interviews of 2022 and 2024.Note: A direction is the answer named in the row label: for example a positive change in the explanation rows is a larger share saying yes to that explanation. Intervals assume a simple random sample of the weighted respondents (the Kish effective sample size); they do not use the survey's sampling design.Source: General Social Survey, 1972-2024; weighted (weight_use); white respondents (race code white, Hispanic whites included); 95% intervals on the Kish effective sample size

    Most explanation and policy answers were level from 2004 to 2014

    From 2004-2008 to 2010-2014, the pooled periods before and during the Obama presidency, the changes in the dated items lie between −2.5−2.5 percentage pointsEstimandsmallest change among six items in the weighted share who gave the answern6,189Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the item, pooled periodsPeriod2004-2008 to 2010-2014SourceGeneral Social Survey and +2.1+2.1 percentage pointsEstimandlargest change among six items in the weighted share who gave the answern6,216Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the item, pooled periodsPeriod2004-2008 to 2010-2014SourceGeneral Social Survey points, and the interval includes zero for 55 items out of sixEstimandnumber of dated items whose change between pooled periods has an interval that includes zeroWeightsweight_use (wtssnrps from 2004, wtssps before)Populationsix dated items, white respondentsPeriod2004-2008 to 2010-2014SourceGeneral Social Survey of the six. The exception is the affirmative action item, at +2.1 (95% CI 0.1 to 4.0)+2.1 percentage pointsEstimandchange in the weighted share who favor preference in hiring and promoting Black peopleInterval95% CI 0.1 to 4.0n6,216Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the item, pooled periodsPeriod2004-2008 to 2010-2014SourceGeneral Social Survey points. The two equal-treatment items asked in both periods kept rising, by +3.4 (95% CI 0.7 to 6.1)+3.4 percentage pointsEstimandchange in the weighted share who would vote for a law barring housing discriminationInterval95% CI 0.7 to 6.1n6,393Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the item, pooled periodsPeriod2004-2008 to 2010-2014SourceGeneral Social Survey points for the open-housing law and +8.6 (95% CI 6.2 to 11.1)+8.6 percentage pointsEstimandchange in the weighted share who would not oppose a close relative marrying a Black personInterval95% CI 6.2 to 11.1n6,445Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the item, pooled periodsPeriod2004-2008 to 2010-2014SourceGeneral Social Survey for a relative's marriage.

    After 2014 the dated items moved. From 2010-2014 to 2016-2018 the share citing discrimination changed by +8.4 (95% CI 5.3 to 11.5)+8.4 percentage pointsEstimandchange in the weighted share who gave the answer (discrimination) between two pooled periods of surveysInterval95% CI 5.3 to 11.5n5,589Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the item, pooled periodsPeriod2010-2014 to 2016-2018SourceGeneral Social Survey points, the share citing will power by −9.0 (95% CI −12.2 to −5.7)−9.0 percentage pointsEstimandchange in the weighted share who gave the answer (lack of motivation or will power) between two pooled periods of surveysInterval95% CI −12.2 to −5.7n5,556Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the item, pooled periodsPeriod2010-2014 to 2016-2018SourceGeneral Social Survey, the share at the help end of the government-help scale by +10.0 (95% CI 7.6 to 12.4)+10.0 percentage pointsEstimandchange in the weighted share who gave the answer (places self at the help end of the government-help scale (1 or 2 of 5)) between two pooled periods of surveysInterval95% CI 7.6 to 12.4n5,643Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the item, pooled periodsPeriod2010-2014 to 2016-2018SourceGeneral Social Survey, the share saying too little is spent on improving the conditions of Black people by +20.9 (95% CI 17.2 to 24.5)+20.9 percentage pointsEstimandchange in the weighted share who gave the answer (too little is spent on improving conditions of black people) between two pooled periods of surveysInterval95% CI 17.2 to 24.5n3,867Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the item, pooled periodsPeriod2010-2014 to 2016-2018SourceGeneral Social Survey, the share favoring preference in hiring by +5.7 (95% CI 3.3 to 8.2)+5.7 percentage pointsEstimandchange in the weighted share who gave the answer (favors preference in hiring and promoting black people) between two pooled periods of surveysInterval95% CI 3.3 to 8.2n5,500Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the item, pooled periodsPeriod2010-2014 to 2016-2018SourceGeneral Social Survey and the share rejecting the statement that Black people should work up without special favors by +8.2 (95% CI 5.8 to 10.6)+8.2 percentage pointsEstimandchange in the weighted share who gave the answer (rejects that black people should work up without special favors) between two pooled periods of surveysInterval95% CI 5.8 to 10.6n5,748Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the item, pooled periodsPeriod2010-2014 to 2016-2018SourceGeneral Social Survey. The comparison with 2021-2024 pools one web survey and two mixed-mode surveys, so the later years are read by interview mode below. The share citing inborn ability changed by −1.5 (95% CI −3.1 to 0.2)−1.5 percentage pointsEstimandchange in the weighted share who attribute the gap to inborn ability between two pooled periodsInterval95% CI −3.1 to 0.2n5,695Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the item, pooled periodsPeriod2010-2014 to 2016-2018SourceGeneral Social Survey points from 2010-2014 to 2016-2018, an interval that includes zero, so the movement is in the explanations that place the gap in discrimination, education and motivation and not in the one that places it in ability.

    White respondents rated Black and white Americans alike by 2024

    White respondents in the ANES rated Black Americans at 60.760.7 points on the 0 to 100 feeling thermometerEstimandweighted mean rating (rating of black americans)Interval95% CI 59.5 to 61.8n1,366Weightsweight (VCF0009z)Populationall white respondents with a ratingPeriod1964SourceANES Cumulative Data File on average in 1964, 69.169.1 points on the 0 to 100 feeling thermometerEstimandweighted mean rating (rating of black americans)Interval95% CI 67.6 to 70.6n740Weightsweight (VCF0009z)Populationall white respondents with a ratingPeriod2004SourceANES Cumulative Data File in 2004, 61.861.8 points on the 0 to 100 feeling thermometerEstimandweighted mean rating (rating of black americans)Interval95% CI 60.9 to 62.7n3,254Weightsweight (VCF0009z)Populationall white respondents with a ratingPeriod2012SourceANES Cumulative Data File in 2012 and 69.569.5 points on the 0 to 100 feeling thermometerEstimandweighted mean rating (rating of black americans)Interval95% CI 68.5 to 70.6n3,374Weightsweight (VCF0009z)Populationall white respondents with a ratingPeriod2024SourceANES Cumulative Data File in 2024, and rated white Americans at 86.786.7 points on the 0 to 100 feeling thermometerEstimandweighted mean rating (rating of white americans)Interval95% CI 85.8 to 87.6n1,367Weightsweight (VCF0009z)Populationall white respondents with a ratingPeriod1964SourceANES Cumulative Data File in 1964 and 70.270.2 points on the 0 to 100 feeling thermometerEstimandweighted mean rating (rating of white americans)Interval95% CI 69.2 to 71.2n3,373Weightsweight (VCF0009z)Populationall white respondents with a ratingPeriod2024SourceANES Cumulative Data File in 2024 (Figure 3). The average of each respondent's own difference, the rating of white Americans minus the rating of Black Americans, went from 26.1 (95% CI 24.7 to 27.4)26.1 points on the 0 to 100 feeling thermometer (rating of white Americans minus rating of Black Americans)Estimandweighted mean of the respondent's own difference in ratingsInterval95% CI 24.7 to 27.4n1,366Weightsweight (VCF0009z)Populationwhite respondents who rated both groupsPeriod1964SourceANES Cumulative Data File in 1964 to 0.7 (95% CI −0.3 to 1.6)0.7 points on the 0 to 100 feeling thermometer (rating of white Americans minus rating of Black Americans)Estimandweighted mean of the respondent's own difference in ratingsInterval95% CI −0.3 to 1.6n3,269Weightsweight (VCF0009z)Populationwhite respondents who rated both groupsPeriod2024SourceANES Cumulative Data File in 2024.

    Figure 3

    By 2024 white respondents rated Black and white Americans almost alike, after a fall in the rating of Black Americans from 2008 to 2012

    Mean feeling thermometer rating of Black Americans and of white Americans among white non-Hispanic respondents, points on a scale from 0 to 100, 1964 to 2024, all respondents and by party including leaners

    Rating of Black AmericansRating of white Americans
    Show data
    Data for By 2024 white respondents rated Black and white Americans almost alike, after a fall in the rating of Black Americans from 2008 to 2012
    conf_highconf_lowestimategroupmeasurenviewyear
    61.8059.5060.70All white respondentsRating of Black Americans1,366all1964
    61.8059.1060.40All white respondentsRating of Black Americans1,126all1966
    62.9060.6061.70All white respondentsRating of Black Americans1,337all1968
    59.9057.5058.70All white respondentsRating of Black Americans1,285all1970
    62.1060.3061.20All white respondentsRating of Black Americans1,850all1972
    64.506263.30All white respondentsRating of Black Americans1,324all1974
    58.805757.90All white respondentsRating of Black Americans1,602all1976
    61.6059.5060.60All white respondentsRating of Black Americans1,136all1980
    63.4060.8062.10All white respondentsRating of Black Americans1,108all1982
    63.5061.5062.50All white respondentsRating of Black Americans1,516all1984
    64.2062.2063.20All white respondentsRating of Black Americans1,598all1986
    58.5056.4057.40All white respondentsRating of Black Americans1,315all1988
    67.8065.5066.70All white respondentsRating of Black Americans1,443all1990
    62.2060.4061.30All white respondentsRating of Black Americans1,686all1992
    60.8058.3059.50All white respondentsRating of Black Americans1,317all1994
    64.1061.9063All white respondentsRating of Black Americans1,173all1996
    67.806566.40All white respondentsRating of Black Americans883all1998
    65.606364.30All white respondentsRating of Black Americans1,123all2000
    67.1064.2065.60All white respondentsRating of Black Americans987all2002
    70.6067.6069.10All white respondentsRating of Black Americans740all2004
    67.4064.8066.10All white respondentsRating of Black Americans1,020all2008
    62.7060.9061.80All white respondentsRating of Black Americans3,254all2012
    67.8065.9066.90All white respondentsRating of Black Americans2,591all2016
    70.6069.2069.90All white respondentsRating of Black Americans5,173all2020
    70.6068.5069.50All white respondentsRating of Black Americans3,374all2024
    61.9058.9060.40White DemocratsRating of Black Americans792all1964
    62.7058.9060.80White DemocratsRating of Black Americans586all1966
    63.306061.60White DemocratsRating of Black Americans680all1968
    61.4058.1059.80White DemocratsRating of Black Americans659all1970
    63.4060.9062.20White DemocratsRating of Black Americans877all1972
    65.7062.2064White DemocratsRating of Black Americans635all1974
    59.7057.1058.40White DemocratsRating of Black Americans728all1976
    63.8060.8062.30White DemocratsRating of Black Americans532all1980
    63.9060.4062.10White DemocratsRating of Black Americans559all1982
    64.8061.7063.20White DemocratsRating of Black Americans641all1984
    6561.9063.40White DemocratsRating of Black Americans688all1986
    60.9057.6059.20White DemocratsRating of Black Americans526all1988
    69.3065.9067.60White DemocratsRating of Black Americans673all1990
    63.706162.40White DemocratsRating of Black Americans747all1992
    62.6058.5060.60White DemocratsRating of Black Americans535all1994
    66.6063.5065.10White DemocratsRating of Black Americans552all1996
    68.9064.6066.70White DemocratsRating of Black Americans394all1998
    68.9065.1067White DemocratsRating of Black Americans484all2000
    68.8064.2066.50White DemocratsRating of Black Americans404all2002
    72.4067.7070.10White DemocratsRating of Black Americans295all2004
    70.8066.8068.80White DemocratsRating of Black Americans436all2008
    66.2063.4064.80White DemocratsRating of Black Americans1,283all2012
    737071.50White DemocratsRating of Black Americans1,024all2016
    76.4074.2075.30White DemocratsRating of Black Americans2,146all2020
    76.1073.2074.70White DemocratsRating of Black Americans1,485all2024
    64.7060.7062.70White RepublicansRating of Black Americans449all1964
    62.5058.4060.50White RepublicansRating of Black Americans385all1966
    6460.5062.30White RepublicansRating of Black Americans491all1968
    605658White RepublicansRating of Black Americans453all1970
    61.3058.4059.80White RepublicansRating of Black Americans712all1972
    64.7060.5062.60White RepublicansRating of Black Americans464all1974
    59.1056.5057.80White RepublicansRating of Black Americans614all1976
    60.7057.1058.90White RepublicansRating of Black Americans437all1980
    64.7060.4062.60White RepublicansRating of Black Americans403all1982
    63.5060.5062White RepublicansRating of Black Americans698all1984
    65.4062.3063.90White RepublicansRating of Black Americans685all1986
    5855.1056.60White RepublicansRating of Black Americans638all1988
    68.3064.8066.50White RepublicansRating of Black Americans590all1990
    61.8059.3060.60White RepublicansRating of Black Americans734all1992
    61.1057.6059.40White RepublicansRating of Black Americans636all1994
    63.1059.8061.50White RepublicansRating of Black Americans523all1996
    68.5064.1066.30White RepublicansRating of Black Americans384all1998
    65.3061.5063.40White RepublicansRating of Black Americans495all2000
    68.106466White RepublicansRating of Black Americans496all2002
    7268.1070White RepublicansRating of Black Americans373all2004
    66.4062.6064.50White RepublicansRating of Black Americans457all2008
    61.5058.9060.20White RepublicansRating of Black Americans1,518all2012
    65.2062.5063.90White RepublicansRating of Black Americans1,255all2016
    6865.9066.90White RepublicansRating of Black Americans2,512all2020
    686566.50White RepublicansRating of Black Americans1,698all2024
    87.6085.8086.70All white respondentsRating of white Americans1,367all1964
    86.8084.7085.70All white respondentsRating of white Americans1,124all1966
    83.4081.4082.40All white respondentsRating of white Americans1,343all1968
    79.9077.8078.80All white respondentsRating of white Americans1,293all1970
    80.5078.9079.70All white respondentsRating of white Americans1,867all1972
    80.9078.6079.80All white respondentsRating of white Americans1,336all1974
    75.5073.8074.70All white respondentsRating of white Americans1,635all1976
    78.4076.4077.40All white respondentsRating of white Americans1,148all1980
    74.8072.4073.60All white respondentsRating of white Americans1,109all1982
    75.807474.90All white respondentsRating of white Americans1,522all1984
    74.8072.7073.80All white respondentsRating of white Americans1,324all1988
    72.3070.5071.40All white respondentsRating of white Americans1,694all1992
    72.3070.1071.20All white respondentsRating of white Americans1,333all1994
    71.9069.6070.80All white respondentsRating of white Americans1,164all1996
    71.2068.4069.80All white respondentsRating of white Americans893all1998
    74.2071.7073All white respondentsRating of white Americans1,128all2000
    7067.1068.50All white respondentsRating of white Americans989all2002
    75.2072.2073.70All white respondentsRating of white Americans739all2004
    74.5071.9073.20All white respondentsRating of white Americans1,021all2008
    7472.4073.20All white respondentsRating of white Americans3,258all2012
    74.2072.4073.30All white respondentsRating of white Americans2,592all2016
    7270.5071.20All white respondentsRating of white Americans5,152all2020
    71.2069.2070.20All white respondentsRating of white Americans3,373all2024
    89.1086.9088White DemocratsRating of white Americans793all1964
    88.9086.1087.50White DemocratsRating of white Americans584all1966
    84.6081.9083.30White DemocratsRating of white Americans682all1968
    82.1079.4080.80White DemocratsRating of white Americans660all1970
    82.2079.8081White DemocratsRating of white Americans885all1972
    82.6079.4081White DemocratsRating of white Americans641all1974
    76.8074.3075.50White DemocratsRating of white Americans744all1976
    80.1077.2078.60White DemocratsRating of white Americans537all1980
    76.307374.70White DemocratsRating of white Americans556all1982
    77.5074.9076.20White DemocratsRating of white Americans641all1984
    76.5073.1074.80White DemocratsRating of white Americans525all1988
    73.6070.9072.30White DemocratsRating of white Americans751all1992
    74.4070.9072.70White DemocratsRating of white Americans542all1994
    73.8070.6072.20White DemocratsRating of white Americans546all1996
    71.9067.7069.80White DemocratsRating of white Americans392all1998
    75.4071.6073.50White DemocratsRating of white Americans486all2000
    70.9066.3068.60White DemocratsRating of white Americans404all2002
    75.8070.5073.20White DemocratsRating of white Americans298all2004
    75.6071.5073.50White DemocratsRating of white Americans435all2008
    75.3072.7074White DemocratsRating of white Americans1,284all2012
    74.3071.4072.80White DemocratsRating of white Americans1,024all2016
    70.5068.2069.40White DemocratsRating of white Americans2,144all2020
    70.6067.6069.10White DemocratsRating of white Americans1,489all2024
    87.208485.60White RepublicansRating of white Americans449all1964
    86.508384.70White RepublicansRating of white Americans385all1966
    83.2079.9081.50White RepublicansRating of white Americans492all1968
    79.3075.7077.50White RepublicansRating of white Americans458all1970
    80.507879.20White RepublicansRating of white Americans722all1972
    81.3077.6079.40White RepublicansRating of white Americans470all1974
    75.807374.40White RepublicansRating of white Americans623all1976
    78.8075.5077.20White RepublicansRating of white Americans443all1980
    75.7071.9073.80White RepublicansRating of white Americans404all1982
    75.8073.2074.50White RepublicansRating of white Americans703all1984
    74.9071.9073.40White RepublicansRating of white Americans646all1988
    72.6069.9071.30White RepublicansRating of white Americans738all1992
    72.4069.4070.90White RepublicansRating of white Americans640all1994
    71.306869.70White RepublicansRating of white Americans521all1996
    72.5068.2070.40White RepublicansRating of white Americans387all1998
    75.1071.6073.30White RepublicansRating of white Americans497all2000
    71.2067.2069.20White RepublicansRating of white Americans499all2002
    77.1073.1075.10White RepublicansRating of white Americans370all2004
    75.7071.9073.80White RepublicansRating of white Americans457all2008
    74.9072.6073.70White RepublicansRating of white Americans1,520all2012
    75.407374.20White RepublicansRating of white Americans1,256all2016
    74.4072.5073.50White RepublicansRating of white Americans2,495all2020
    73.1070.2071.60White RepublicansRating of white Americans1,692all2024
    67.4064.8066.10All white respondentsRating of Black Americans1,020ftf2008
    64.4061.1062.70All white respondentsRating of Black Americans852ftf2012
    68.9064.8066.90All white respondentsRating of Black Americans699ftf2016
    73.3068.8071All white respondentsRating of Black Americans605ftf2024
    70.8066.8068.80White DemocratsRating of Black Americans436ftf2008
    6862.9065.40White DemocratsRating of Black Americans356ftf2012
    76.4069.7073.10White DemocratsRating of Black Americans265ftf2016
    81.307578.10White DemocratsRating of Black Americans276ftf2024
    66.4062.6064.50White RepublicansRating of Black Americans457ftf2008
    63.7059.1061.40White RepublicansRating of Black Americans395ftf2012
    65.9060.6063.30White RepublicansRating of Black Americans368ftf2016
    69.206366.10White RepublicansRating of Black Americans306ftf2024
    74.5071.9073.20All white respondentsRating of white Americans1,021ftf2008
    74.6071.5073.10All white respondentsRating of white Americans855ftf2012
    75.3071.6073.50All white respondentsRating of white Americans699ftf2016
    71.7067.3069.50All white respondentsRating of white Americans605ftf2024
    75.6071.5073.50White DemocratsRating of white Americans435ftf2008
    75.7070.9073.30White DemocratsRating of white Americans356ftf2012
    77.2070.8074White DemocratsRating of white Americans265ftf2016
    73.4066.7070White DemocratsRating of white Americans277ftf2024
    75.7071.9073.80White RepublicansRating of white Americans457ftf2008
    74.8070.4072.60White RepublicansRating of white Americans397ftf2012
    7671.3073.70White RepublicansRating of white Americans368ftf2016
    72.5066.3069.40White RepublicansRating of white Americans305ftf2024
    62.4060.3061.40All white respondentsRating of Black Americans2,402web2012
    6865.8066.90All white respondentsRating of Black Americans1,892web2016
    70.3067.2068.80All white respondentsRating of Black Americans1,163web2024
    66.1062.9064.50White DemocratsRating of Black Americans927web2012
    72.6069.3071White DemocratsRating of Black Americans759web2016
    74.507072.20White DemocratsRating of Black Americans514web2024
    615859.50White RepublicansRating of Black Americans1,123web2012
    65.7062.6064.10White RepublicansRating of Black Americans887web2016
    7065.4067.70White RepublicansRating of Black Americans554web2024
    74.2072.4073.30All white respondentsRating of white Americans2,403web2012
    74.2072.3073.20All white respondentsRating of white Americans1,893web2016
    71.3068.2069.80All white respondentsRating of white Americans1,157web2024
    75.8072.9074.40White DemocratsRating of white Americans928web2012
    7470.9072.40White DemocratsRating of white Americans759web2016
    69.1064.6066.90White DemocratsRating of white Americans516web2024
    75.7073.1074.40White RepublicansRating of white Americans1,123web2012
    75.707374.40White RepublicansRating of white Americans888web2016
    75.7071.2073.40White RepublicansRating of white Americans548web2024
    The rating of white Americans declines across the series and the rating of Black Americans rises, so the two lines meet near the end, with a dip in the rating of Black Americans in 2012. Lines are weighted mean ratings among white non-Hispanic respondents, and the band is the 95% interval for all white respondents. The face-to-face and web views use the mode weights, which exist for 2008, 2012, 2016 and 2024.Note: Face-to-face and web estimates exist for 2008 (face to face only), 2012, 2016 and 2024; 2020 has no mode weights. The thermometer toward white Americans was not asked in 1986 or 1990.Note: Ratings of Black Americans by white respondents in 1964 to 1970 come from the same item as later years. Intervals assume a simple random sample of the weighted respondents (the Kish effective sample size); they do not use the survey's sampling design.Source: ANES Cumulative Data File, 1964-2024; weighted (VCF0009z; mode weights VCF0009x and VCF0009y for 2008, 2012, 2016 and 2024)

    Affect has its own Obama-era change. In face-to-face interviews, the only mode in 2008, the rating of Black Americans among white respondents changed by −3.4 (95% CI −5.5 to −1.3)−3.4 points on the 0 to 100 feeling thermometerEstimandchange in the weighted mean rating of Black AmericansInterval95% CI −5.5 to −1.3n1,872Weightsweight_ftf (VCF0009x)Populationall white respondents who rated Black AmericansPeriod2008 to 2012, face-to-face interviewsSourceANES Cumulative Data File points from 2008 to 2012, by −3.4 (95% CI −6.7 to −0.1)−3.4 points on the 0 to 100 feeling thermometerEstimandchange in the weighted mean rating of Black AmericansInterval95% CI −6.7 to −0.1n792Weightsweight_ftf (VCF0009x)Populationwhite democrats who rated Black AmericansPeriod2008 to 2012, face-to-face interviewsSourceANES Cumulative Data File among white Democrats and by −3.1 (95% CI −6.1 to −0.1)−3.1 points on the 0 to 100 feeling thermometerEstimandchange in the weighted mean rating of Black AmericansInterval95% CI −6.1 to −0.1n852Weightsweight_ftf (VCF0009x)Populationwhite republicans who rated Black AmericansPeriod2008 to 2012, face-to-face interviewsSourceANES Cumulative Data File among white Republicans. The fall was similar in the two parties. From 2012 to 2016 the face-to-face rating changed by +7.6 (95% CI 3.4 to 11.9)+7.6 points on the 0 to 100 feeling thermometerEstimandchange in the weighted mean rating of Black AmericansInterval95% CI 3.4 to 11.9n621Weightsweight_ftf (VCF0009x)Populationwhite democrats who rated Black AmericansPeriod2012 to 2016, face-to-face interviewsSourceANES Cumulative Data File points among Democrats and +1.9 (95% CI −1.6 to 5.4)+1.9 points on the 0 to 100 feeling thermometerEstimandchange in the weighted mean rating of Black AmericansInterval95% CI −1.6 to 5.4n763Weightsweight_ftf (VCF0009x)Populationwhite republicans who rated Black AmericansPeriod2012 to 2016, face-to-face interviewsSourceANES Cumulative Data File among Republicans, an interval that includes zero for Republicans, and the web interviews give +6.4 (95% CI 4.1 to 8.8)+6.4 points on the 0 to 100 feeling thermometerEstimandchange in the weighted mean rating of Black AmericansInterval95% CI 4.1 to 8.8n1,686Weightsweight_web (VCF0009y)Populationwhite democrats who rated Black AmericansPeriod2012 to 2016, web interviewsSourceANES Cumulative Data File and +4.6 (95% CI 2.5 to 6.8)+4.6 points on the 0 to 100 feeling thermometerEstimandchange in the weighted mean rating of Black AmericansInterval95% CI 2.5 to 6.8n2,010Weightsweight_web (VCF0009y)Populationwhite republicans who rated Black AmericansPeriod2012 to 2016, web interviewsSourceANES Cumulative Data File. The ANES has no study in 2014, so it dates this rise only to the interval from 2012 to 2016, in face-to-face interviews and on the web.

    Most dated items change between the 2014 and 2016 surveys

    Figure 4 dates the movement. Each row fits two levels to the yearly shares, one before the first new survey and one from it on, and reports the survey that fits best and how often resampling the respondents agrees. The best-fitting first new survey is 2016 for each dated item among white respondents and among white Democrats. The least certain date is for will power, where 51.4%51.4% percent of bootstrap replicatesEstimandshare of replicates that place the best single step at the 2016 surveyn1,000Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents, will-power itemPeriodin-person surveys 2000 to 2018SourceGeneral Social Survey of the replicates choose 2016 and 38.0%38.0% percent of bootstrap replicatesEstimandshare of replicates that place the best single step at a survey before 2016n1,000Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents, will-power itemPeriodin-person surveys 2000 to 2018SourceGeneral Social Survey choose an earlier survey. In single-survey comparisons the change from 2012 to 2014 has an interval that excludes zero for 00 items out of sixEstimandnumber of items whose change in the share among white respondents has an interval that excludes zeroWeightsweight_use (wtssnrps from 2004, wtssps before)Populationsix dated itemsPeriod2012 to 2014SourceGeneral Social Survey of the six dated items, and the change from 2014 to 2016 does so for 55 items out of sixEstimandnumber of items whose change in the share among white respondents has an interval that excludes zeroWeightsweight_use (wtssnrps from 2004, wtssps before)Populationsix dated itemsPeriod2014 to 2016SourceGeneral Social Survey: it is +18.2 (95% CI 12.7 to 23.8)+18.2 percentage pointsEstimandchange in the weighted share who gave the answer (too little is spent on improving conditions of black people)Interval95% CI 12.7 to 23.8n1,586Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the item, in-person interviewsPeriod2014 to 2016SourceGeneral Social Survey points for spending, +9.1 (95% CI 5.6 to 12.6)+9.1 percentage pointsEstimandchange in the weighted share who gave the answer (places self at the help end of the government-help scale (1 or 2 of 5))Interval95% CI 5.6 to 12.6n2,326Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the item, in-person interviewsPeriod2014 to 2016SourceGeneral Social Survey for the help end of the government-help scale and +6.3 (95% CI 1.7 to 10.9)+6.3 percentage pointsEstimandchange in the weighted share who gave the answer (discrimination)Interval95% CI 1.7 to 10.9n2,284Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the item, in-person interviewsPeriod2014 to 2016SourceGeneral Social Survey for discrimination, and −3.8 (95% CI −8.6 to 1.1)−3.8 percentage pointsEstimandchange in the weighted share who gave the answer (lack of motivation or will power)Interval95% CI −8.6 to 1.1n2,275Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the item, in-person interviewsPeriod2014 to 2016SourceGeneral Social Survey for will power, an interval that includes zero.

    A step is a form assumption. A trend that is constant before a kink and linear after it fits the discrimination share better than a step, with a weighted squared error of 2.42.4 weighted squared error of the yearly shares around the fitted kinked trendEstimandinverse-variance weighted squared error of the best trend with one kink (discrimination)Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents, in personPeriodsurveys 2000 to 2018SourceGeneral Social Survey against 8.08.0 weighted squared error of the yearly shares around the fitted stepEstimandinverse-variance weighted squared error of the best single step (discrimination)Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents, in personPeriodsurveys 2000 to 2018SourceGeneral Social Survey, because the share keeps rising from 2016 to 2018. It fits the will-power share better still, 3.03.0 weighted squared error of the yearly shares around the fitted kinked trendEstimandinverse-variance weighted squared error of the best trend with one kink (lack of motivation or will power)Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents, in personPeriodsurveys 2000 to 2018SourceGeneral Social Survey against 15.215.2 weighted squared error of the yearly shares around the fitted stepEstimandinverse-variance weighted squared error of the best single step (lack of motivation or will power)Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents, in personPeriodsurveys 2000 to 2018SourceGeneral Social Survey, because that share keeps falling, so will power belongs with trend items more than step items. The step fits the government-help and spending shares better than the kink. The kink falls at 2014.0 (95% CI 2012.0 to 2015.5)2014.0 survey year of the kinkEstimandyear at which the fitted trend turns from constant to linear (discrimination)Interval95% CI 2012.0 to 2015.5n9,719Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents, in personPeriodsurveys 2000 to 2018SourceGeneral Social Survey for discrimination and at 2011.4 (95% CI 2008.0 to 2014.8)2011.4 survey year of the kinkEstimandyear at which the fitted trend turns from constant to linear (lack of motivation or will power)Interval95% CI 2008.0 to 2014.8n9,620Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents, in personPeriodsurveys 2000 to 2018SourceGeneral Social Survey for will power, and the latest upper end of a kink interval across the dated items is 2015.52015.5 survey year of the kinkEstimandlatest upper end of the bootstrap interval for the kink year across the six itemsWeightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents, in person, six dated itemsPeriodsurveys 2000 to 2018SourceGeneral Social Survey. By point estimate the kink years run from about 2011 to 2014, so the kinked trend starts the rise at or before the 2014 survey, while the step puts the new level at the 2016 survey.

    Figure 4

    The best single step falls at the 2016 survey for each of the six dated items among white respondents and among white Democrats, and a kinked trend fits discrimination and will power better

    Best-fitting single step in the yearly share, by item and group, in-person surveys 2000 to 2018, among white respondents (GSS race code, Hispanic whites included); step in percentage points, share of bootstrap replicates choosing each date, and the error of a trend with one kink

    GroupItemBest-fitting first new surveyStep at that survey (points)Step at the 2016 survey (points)2016 step, lower end2016 step, upper endReplicates at that survey (%)Replicates at an earlier survey (%)Error of the best stepError of the kinked trendKink yearKink year, lower endKink year, upper end
    White respondentsDiscrimination2016+8.9+8.9+5.8+11.67418.02.42014.02012.02015.5
    White respondentsLack of motivation or will power2016−10.6−10.6−13.3−7.6513815.23.02011.42008.02014.8
    White respondentsAt the help end of the scale2016+9.2+9.2+7.0+11.610009.212.92013.92012.92014.0
    White respondentsToo little spent2016+20.3+20.3+17.3+23.7100016.124.12013.42010.92014.0
    White respondentsFavors preferential hiring2016+7.2+7.2+4.9+9.887109.69.12013.62006.02014.7
    White respondentsRejects the special-favors view2016+9.3+9.3+7.2+11.892711.24.62012.82011.42013.9
    White DemocratsDiscrimination2016+18.5+18.5+14.1+23.2782013.15.52012.42007.32013.8
    White DemocratsLack of motivation or will power2016−17.6−17.6−21.4−13.3603915.56.02008.62004.12013.4
    White DemocratsAt the help end of the scale2016+19.1+19.1+15.1+23.410009.214.52013.52012.02014.0
    White DemocratsToo little spent2016+29.1+29.1+24.0+34.3100015.320.12013.52009.62014.0
    White DemocratsFavors preferential hiring2016+13.8+13.8+9.4+18.1653112.66.42011.12006.02014.7
    White DemocratsRejects the special-favors view2016+19.6+19.6+15.4+24.6851511.27.12011.62008.02013.4
    White RepublicansDiscrimination2004−3.3+0.4−3.3+4.34106.48.32015.92004.02015.9
    White RepublicansLack of motivation or will power2018−10.4−6.8−11.2−2.539616.53.82012.62009.32015.9
    White RepublicansAt the help end of the scale2018+6.3+2.9+0.5+6.1505011.910.92015.42004.02015.9
    White RepublicansToo little spent2016+15.0+15.0+10.6+20.298210.914.82013.52010.22014.0
    White RepublicansFavors preferential hiring2016+1.4+1.4−1.0+4.228504.13.92014.22004.02015.9
    White RepublicansRejects the special-favors view2018+6.7+2.1−0.4+5.2623813.613.52015.92004.02015.9
    Show data
    Data for The best single step falls at the 2016 survey for each of the six dated items among white respondents and among white Democrats, and a kinked trend fits discrimination and will power better
    best_yearfix16fix16_highfix16_lowgroupitem_labelkinkkink_highkink_lowkink_ssep_best_pctp_earlier_pctstepstep_sse
    20168.9011.605.80White respondentsDiscrimination20142015.520122.407418.908
    2016−10.60−7.60−13.30White respondentsLack of motivation or will power2011.42014.8200835138−10.6015.20
    20169.2011.607White respondentsAt the help end of the scale2013.920142012.912.9010009.209.20
    201620.3023.7017.30White respondentsToo little spent2013.420142010.924.10100020.3016.10
    20167.209.804.90White respondentsFavors preferential hiring2013.62014.720069.1087107.209.60
    20169.3011.807.20White respondentsRejects the special-favors view2012.82013.92011.44.609279.3011.20
    201618.5023.2014.10White DemocratsDiscrimination2012.42013.82007.35.50782018.5013.10
    2016−17.60−13.30−21.40White DemocratsLack of motivation or will power2008.62013.42004.166039−17.6015.50
    201619.1023.4015.10White DemocratsAt the help end of the scale2013.52014201214.50100019.109.20
    201629.1034.3024White DemocratsToo little spent2013.520142009.620.10100029.1015.30
    201613.8018.109.40White DemocratsFavors preferential hiring2011.12014.720066.40653113.8012.60
    201619.6024.6015.40White DemocratsRejects the special-favors view2011.62013.420087.10851519.6011.20
    20040.4004.30−3.30White RepublicansDiscrimination2015.92015.920048.30410−3.306.40
    2018−6.80−2.50−11.20White RepublicansLack of motivation or will power2012.62015.92009.33.803961−10.406.50
    20182.906.100.500White RepublicansAt the help end of the scale2015.42015.9200410.9050506.3011.90
    20161520.2010.60White RepublicansToo little spent2013.520142010.214.809821510.90
    20161.404.20−1White RepublicansFavors preferential hiring2014.22015.920043.9028501.404.10
    20182.105.20−0.400White RepublicansRejects the special-favors view2015.92015.9200413.5062386.7013.60
    Every row for white respondents and for white Democrats has 2016 as the first new survey. For white Republicans the best dates run from 2004 to 2018, so their steps are also given at a fixed 2016 date, and 4 of the six dated items have best-step intervals that include zero. The last columns compare the error of the best step with the error of the best trend with one kink: among white respondents the kink fits the discrimination and will-power shares better, and the step fits the government-help and spending shares better.Note: The fit uses the surveys of 2000, 2002, 2004, 2006, 2008, 2010, 2012, 2014, 2016 and 2018, all conducted in person, so the dates cannot come from the move to the web. A step cannot begin before 2004 or leave fewer than one survey after it. The kinked trend is constant before the kink and linear after it, with the kink anywhere from 2004 to just before 2016. A step and a kink are assumptions about the form of change. The bootstrap resamples respondents within each survey year with replacement (seed 12345) and refits; the interval for a step size or a kink year is the 2.5th to 97.5th percentile. Survey spacing is two years, so a date is a survey and not a month.Source: General Social Survey, 2000-2018; weighted (weight_use); in-person interviews; white respondents (race code white, Hispanic whites included), leaners included for party; seeded bootstrap of respondents within each survey year

    The post-2018 changes depend on interview mode

    The GSS interviewed 87.3%87.3% percent of interviews by webEstimandshare of interviews conducted on the web (count)n4,032Weightsnone (interview count)Populationall 2021 GSS respondentsPeriod2021SourceGeneral Social Survey of its 2021 respondents on the web and held no in-person interviews that year, so the level after 2018 needs a mode check. Figure 5 gives, for each item asked in both modes, the in-person level in 2010-2014, 2016-2018 and 2022-2024 and the web level in 2022-2024, all for white respondents. By sign, web respondents gave the answer less often than in-person respondents on 66 items out of 9Estimandnumber of items on which the web share is below the in-person share, by the sign of the estimateWeightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the itemPeriod2022 and 2024 pooledSourceGeneral Social Survey of the 99 itemsEstimandnumber of items with an in-person and a web estimate that clear the floorWeightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the itemPeriod2022 and 2024 pooledSourceGeneral Social Survey items and more often on 33 items out of 9Estimandnumber of items on which the web share is above the in-person share, by the sign of the estimateWeightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the itemPeriod2022 and 2024 pooledSourceGeneral Social Survey. By interval, the gap excludes zero below on 55 items out of 9Estimandnumber of items on which web respondents gave the answer less often than in-person respondents, with an interval that excludes zeroWeightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the itemPeriod2022 and 2024 pooledSourceGeneral Social Survey items and above on 11 items out of 9Estimandnumber of items on which web respondents gave the answer more often than in-person respondents, with an interval that excludes zeroWeightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the itemPeriod2022 and 2024 pooledSourceGeneral Social Survey, and it is at most 9.99.9 percentage pointsEstimandlargest absolute difference across the 9 items between web and in-person respondents in the weighted share who gave the answerWeightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the itemPeriod2022 and 2024 pooledSourceGeneral Social Survey points.

    In person, the share citing discrimination changed by +12.0 (95% CI 5.3 to 18.7)+12.0 percentage pointsEstimandchange in the weighted share who gave the answer (discrimination) among in-person respondentsInterval95% CI 5.3 to 18.7n2,747Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the item, in-person interviewsPeriod2016-2018 to 2022-2024SourceGeneral Social Survey points from 2016-2018 to 2022-2024, the share citing will power by −7.1 (95% CI −13.4 to −0.8)−7.1 percentage pointsEstimandchange in the weighted share who gave the answer (lack of motivation or will power) among in-person respondentsInterval95% CI −13.4 to −0.8n2,733Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the item, in-person interviewsPeriod2016-2018 to 2022-2024SourceGeneral Social Survey, the share favoring preference in hiring by +6.2 (95% CI 1.7 to 10.6)+6.2 percentage pointsEstimandchange in the weighted share who gave the answer (favors preference in hiring and promoting black people) among in-person respondentsInterval95% CI 1.7 to 10.6n3,192Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the item, in-person interviewsPeriod2016-2018 to 2022-2024SourceGeneral Social Survey and the share saying too little is spent by −1.3 (95% CI −7.7 to 5.1)−1.3 percentage pointsEstimandchange in the weighted share who gave the answer (too little is spent on improving conditions of black people) among in-person respondentsInterval95% CI −7.7 to 5.1n2,309Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the item, in-person interviewsPeriod2016-2018 to 2022-2024SourceGeneral Social Survey, an interval that includes zero. The in-person change excludes zero for 44 items out of sixEstimandnumber of dated items whose in-person change has an interval that excludes zeroWeightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the item, in-person interviewsPeriod2016-2018 to 2022-2024SourceGeneral Social Survey of the six dated items. The web level in 2022-2024 minus the in-person level in 2016-2018 is +2.1 (95% CI −3.8 to 7.9)+2.1 percentage pointsEstimandweb share in 2022 and 2024 minus in-person share in 2016 and 2018 who said racial inequality is mainly due to discriminationInterval95% CI −3.8 to 7.9n3,003Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the itemPeriodweb interviews of 2022 and 2024 against in-person interviews of 2016 and 2018SourceGeneral Social Survey points for discrimination, so the post-2018 rise in that share holds in person and not on the web. Spending is lower on the web than in person in 2022-2024, by −7.1 (95% CI −13.9 to −0.3)−7.1 percentage pointsEstimandweb minus in-person difference in the weighted share who said too little is spent on improving conditions of Black peopleInterval95% CI −13.9 to −0.3n2,065Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race code, Hispanic whites included) who answered the itemPeriod2022 and 2024 pooledSourceGeneral Social Survey points, and the in-person change is within sampling error of zero, so the chapter does not read the lower all-interview level after 2018 as a fall.

    Figure 5

    The post-2018 level depends on interview mode: in 2022 and 2024 web respondents gave the answer less often than in-person respondents on 6 of 9 items

    Share of white respondents (GSS race code, Hispanic whites included) giving the answer named in each row, percent, in-person interviews of 2010-2014, 2016-2018 and 2022-2024 and web interviews of 2022-2024, 95% intervals

    • In person, 2010-2014
    • In person, 2016-2018
    • In person, 2022-2024
    • Web, 2022-2024
    Show data
    Data for The post-2018 level depends on interview mode: in 2022 and 2024 web respondents gave the answer less often than in-person respondents on 6 of 9 items
    conf_highconf_lowestimateitemnseries
    81.507879.80Principle: accepts a relative's marriage2,860In person, 2010-2014
    87.2083.5085.30Principle: accepts a relative's marriage2,268In person, 2016-2018
    93.6088.8091.20Principle: accepts a relative's marriage1,169In person, 2022-2024
    9692.8094.40Principle: accepts a relative's marriage1,700Web, 2022-2024
    33.6029.5031.50Explanation: discrimination2,781In person, 2010-2014
    42.8037.7040.30Explanation: discrimination2,199In person, 2016-2018
    58.5046.1052.30Explanation: discrimination548In person, 2022-2024
    47.6037.1042.30Explanation: discrimination804Web, 2022-2024
    9.507.108.30Explanation: inborn ability2,828In person, 2010-2014
    8.305.607Explanation: inborn ability2,241In person, 2016-2018
    11.8058.40Explanation: inborn ability566In person, 2022-2024
    4.100.8002.50Explanation: inborn ability806Web, 2022-2024
    44.8040.5042.70Explanation: lack of chance for education2,813In person, 2010-2014
    53.7048.5051.10Explanation: lack of chance for education2,227In person, 2016-2018
    54.3041.9048.10Explanation: lack of chance for education553In person, 2022-2024
    54.3043.7049Explanation: lack of chance for education807Web, 2022-2024
    50.404648.20Explanation: lack of will power2,760In person, 2010-2014
    42.1036.9039.50Explanation: lack of will power2,189In person, 2016-2018
    38.1026.7032.40Explanation: lack of will power544In person, 2022-2024
    28.4019.4023.90Explanation: lack of will power803Web, 2022-2024
    13.6010.7012.20Policy: help end of the government-help scale2,762In person, 2010-2014
    23.6019.4021.50Policy: help end of the government-help scale2,256In person, 2016-2018
    29.3021.5025.40Policy: help end of the government-help scale1,082In person, 2022-2024
    26.2020.5023.30Policy: help end of the government-help scale1,726Web, 2022-2024
    29.3024.5026.90Policy: too little spending1,855In person, 2010-2014
    50.5044.3047.40Policy: too little spending1,557In person, 2016-2018
    51.7040.5046.10Policy: too little spending752In person, 2022-2024
    42.8035.2039Policy: too little spending1,313Web, 2022-2024
    15.3012.2013.70Policy: preferential hiring2,755In person, 2010-2014
    2217.8019.90Policy: preferential hiring2,144In person, 2016-2018
    29.9022.1026Policy: preferential hiring1,048In person, 2022-2024
    22.5016.8019.60Policy: preferential hiring1,672Web, 2022-2024
    13.3010.4011.80Policy: rejects the special-favors view2,870In person, 2010-2014
    22.9018.7020.80Policy: rejects the special-favors view2,257In person, 2016-2018
    30.3022.8026.60Policy: rejects the special-favors view1,148In person, 2022-2024
    30.9024.6027.80Policy: rejects the special-favors view1,697Web, 2022-2024
    On the discrimination row the in-person dot rises from 2016-2018 to 2022-2024 and the web dot is within sampling error of the 2016-2018 dot. Web respondents gave the answer less often than in-person respondents on 6 rows by sign (5 with an interval that excludes zero) and more often on 3 (1). Each dot is a weighted share among white respondents; rows are the items asked in both modes with cells above the effective sample floor.Note: The open-housing item is not shown because it was not asked of web respondents. The telephone and other-mode interviews are left out of both modes. Intervals assume a simple random sample of the weighted respondents (the Kish effective sample size); they do not use the survey's sampling design.Source: General Social Survey, 1972-2024; weighted (weight_use); white respondents (race code white, Hispanic whites included); 95% intervals on the Kish effective sample size

    The movement is larger among white Democrats

    Figure 6 splits the dated items by party. Among white Democrats the share citing discrimination was 44.9% (95% CI 39.6% to 50.2%)44.9% percent of respondents giving the answerEstimandweighted share who gave the answer (discrimination)Interval95% CI 39.6% to 50.2%n471Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite democrats (leaners included; GSS race code, Hispanic whites included) who answered the itemPeriod2014SourceGeneral Social Survey in 2014, 54.6%54.6% percent of respondents giving the answerEstimandweighted share who gave the answer (discrimination)Interval95% CI 49.6% to 59.7%n541Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite democrats (leaners included; GSS race code, Hispanic whites included) who answered the itemPeriod2016SourceGeneral Social Survey in 2016, 61.6%61.6% percent of respondents giving the answerEstimandweighted share who gave the answer (discrimination)Interval95% CI 55.5% to 67.6%n392Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite democrats (leaners included; GSS race code, Hispanic whites included) who answered the itemPeriod2018SourceGeneral Social Survey in 2018, 81.5%81.5% percent of respondents giving the answerEstimandweighted share who gave the answer (discrimination)Interval95% CI 78.1% to 84.8%n858Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite democrats (leaners included; GSS race code, Hispanic whites included) who answered the itemPeriod2021SourceGeneral Social Survey in 2021 and 73.3%73.3% percent of respondents giving the answerEstimandweighted share who gave the answer (discrimination)Interval95% CI 65.6% to 81.0%n237Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite democrats (leaners included; GSS race code, Hispanic whites included) who answered the itemPeriod2024SourceGeneral Social Survey in 2024. Among white Republicans it was 18.2%18.2% percent of respondents giving the answerEstimandweighted share who gave the answer (discrimination)Interval95% CI 14.2% to 22.2%n468Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite republicans (leaners included; GSS race code, Hispanic whites included) who answered the itemPeriod2014SourceGeneral Social Survey in 2014, 23.3%23.3% percent of respondents giving the answerEstimandweighted share who gave the answer (discrimination)Interval95% CI 18.0% to 28.5%n458Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite republicans (leaners included; GSS race code, Hispanic whites included) who answered the itemPeriod2018SourceGeneral Social Survey in 2018 and 21.4%21.4% percent of respondents giving the answerEstimandweighted share who gave the answer (discrimination)Interval95% CI 15.1% to 27.8%n314Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite republicans (leaners included; GSS race code, Hispanic whites included) who answered the itemPeriod2024SourceGeneral Social Survey in 2024. From 2010-2014 to 2016-2018 the share citing discrimination changed by +14.8 (95% CI 9.7 to 19.9)+14.8 percentage pointsEstimandchange in the weighted share who gave the answer (discrimination) between two pooled periodsInterval95% CI 9.7 to 19.9n2,225Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite Democrats (leaners included; GSS race code, Hispanic whites included) who answered the itemPeriod2010-2014 to 2016-2018SourceGeneral Social Survey points among Democrats and +2.5 (95% CI −1.7 to 6.7)+2.5 percentage pointsEstimandchange in the weighted share who gave the answer (discrimination) between two pooled periodsInterval95% CI −1.7 to 6.7n2,217Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite Republicans (leaners included; GSS race code, Hispanic whites included) who answered the itemPeriod2010-2014 to 2016-2018SourceGeneral Social Survey among Republicans, and the share at the help end of the government-help scale by +18.7 (95% CI 14.2 to 23.2)+18.7 percentage pointsEstimandchange in the weighted share who gave the answer (places self at the help end of the government-help scale (1 or 2 of 5)) between two pooled periodsInterval95% CI 14.2 to 23.2n2,293Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite Democrats (leaners included; GSS race code, Hispanic whites included) who answered the itemPeriod2010-2014 to 2016-2018SourceGeneral Social Survey and +5.0 (95% CI 2.3 to 7.6)+5.0 percentage pointsEstimandchange in the weighted share who gave the answer (places self at the help end of the government-help scale (1 or 2 of 5)) between two pooled periodsInterval95% CI 2.3 to 7.6n2,179Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite Republicans (leaners included; GSS race code, Hispanic whites included) who answered the itemPeriod2010-2014 to 2016-2018SourceGeneral Social Survey. The interval for the change excludes zero for 66 items out of sixEstimandnumber of dated items whose change between pooled periods has an interval that excludes zeroWeightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite DemocratsPeriod2010-2014 to 2016-2018SourceGeneral Social Survey of the six dated items among Democrats and for 44 items out of sixEstimandnumber of dated items whose change between pooled periods has an interval that excludes zeroWeightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite RepublicansPeriod2010-2014 to 2016-2018SourceGeneral Social Survey among Republicans. The Democratic change minus the Republican change excludes zero on 44 items out of sixEstimandnumber of dated items on which the Democratic change minus the Republican change has an interval that excludes zeroWeightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite Democrats and white RepublicansPeriod2010-2014 to 2016-2018SourceGeneral Social Survey of the six, discrimination and government help among them. For will power the difference is −4.0 (95% CI −11.0 to 3.0)−4.0 percentage pointsEstimandDemocratic change minus Republican change in the weighted share who gave the answer (lack of motivation or will power)Interval95% CI −11.0 to 3.0n4,403Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite Democrats and white Republicans (leaners included)Period2010-2014 to 2016-2018SourceGeneral Social Survey points and for spending +7.3 (95% CI −0.5 to 15.0)+7.3 percentage pointsEstimandDemocratic change minus Republican change in the weighted share who gave the answer (too little is spent on improving conditions of black people)Interval95% CI −0.5 to 15.0n3,103Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite Democrats and white Republicans (leaners included)Period2010-2014 to 2016-2018SourceGeneral Social Survey, both intervals that include zero.

    Republicans moved on will power (−8.5 (95% CI −13.7 to −3.3)−8.5 percentage pointsEstimandchange in the weighted share who gave the answer (lack of motivation or will power) between two pooled periodsInterval95% CI −13.7 to −3.3n2,184Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite Republicans (leaners included; GSS race code, Hispanic whites included) who answered the itemPeriod2010-2014 to 2016-2018SourceGeneral Social Survey points) and on spending, by +19.0 (95% CI 13.8 to 24.3)+19.0 percentage pointsEstimandchange in the weighted share who gave the answer (too little is spent on improving conditions of black people) between two pooled periodsInterval95% CI 13.8 to 24.3n1,534Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite Republicans (leaners included; GSS race code, Hispanic whites included) who answered the itemPeriod2010-2014 to 2016-2018SourceGeneral Social Survey against +26.3 (95% CI 20.5 to 32.1)+26.3 percentage pointsEstimandchange in the weighted share who gave the answer (too little is spent on improving conditions of black people) between two pooled periodsInterval95% CI 20.5 to 32.1n1,569Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite Democrats (leaners included; GSS race code, Hispanic whites included) who answered the itemPeriod2010-2014 to 2016-2018SourceGeneral Social Survey among Democrats. Republican best step dates differ by item (2004 for discrimination, 2016 for spending and affirmative action, 2018 for the others), so for discrimination the step at a fixed 2016 date is +18.5 (95% CI 14.1 to 23.2)+18.5 percentage pointsEstimandsize of the step at the 2016 survey in the weighted share (discrimination)Interval95% CI 14.1 to 23.2n3,818Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite democrats (GSS race code, Hispanic whites included) who answered the itemPeriodin-person surveys 2000 to 2018, step at the 2016 surveySourceGeneral Social Survey points among Democrats and +0.4 (95% CI −3.3 to 4.3)+0.4 percentage pointsEstimandsize of the step at the 2016 survey in the weighted share (discrimination)Interval95% CI −3.3 to 4.3n3,933Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite republicans (GSS race code, Hispanic whites included) who answered the itemPeriodin-person surveys 2000 to 2018, step at the 2016 surveySourceGeneral Social Survey among Republicans. The Republican spending share was 16.9%16.9% percent of respondents giving the answerEstimandweighted share who said too little is spent on improving conditions of Black peopleInterval95% CI 12.8% to 20.9%n462Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite Republicans (leaners included; GSS race code, Hispanic whites included) who answered the itemPeriod2021, web interviewsSourceGeneral Social Survey among web respondents in 2021 and 27.9%27.9% percent of respondents giving the answerEstimandweighted share who said too little is spent on improving conditions of Black peopleInterval95% CI 19.7% to 36.0%n309Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite Republicans (leaners included; GSS race code, Hispanic whites included) who answered the itemPeriod2022 and 2024 pooled, in-person interviewsSourceGeneral Social Survey in person in 2022 and 2024 pooled.

    Figure 6

    From 2010-2014 to 2016-2018 white Democrats moved on each of the six dated items and white Republicans on 4, and the Democratic change is separable from the Republican change on 4

    Share of white Democrats and white Republicans (GSS race code, Hispanic whites included; leaners included) giving the answer named above each panel, percent, 2000 to 2024, all interviews and by interview mode

    Share giving the answer (%)

    • Democrats
    • Republicans
    DiscriminationLack of motivation or will powerAt the help end of the scaleToo little spentFavors preferential hiringRejects the special-favors view
    Show data
    Data for From 2010-2014 to 2016-2018 white Democrats moved on each of the six dated items and white Republicans on 4, and the Democratic change is separable from the Republican change on 4
    conf_highconf_lowestimatenpanelpooledseriesviewyear
    44.9035.6040.30523DiscriminationnoDemocratsall2000
    43.9029.9036.90260DiscriminationnoDemocratsall2002
    49.1034.8042258DiscriminationnoDemocratsall2004
    42.1032.4037.20532DiscriminationnoDemocratsall2006
    45.1033.8039.40422DiscriminationnoDemocratsall2008
    47.2035.4041.30423DiscriminationnoDemocratsall2010
    47.603641.80398DiscriminationnoDemocratsall2012
    50.2039.6044.90471DiscriminationnoDemocratsall2014
    59.7049.6054.60541DiscriminationnoDemocratsall2016
    67.6055.5061.60392DiscriminationnoDemocratsall2018
    84.8078.1081.50858DiscriminationnoDemocratsall2021
    85.4071.7078.50306DiscriminationnoDemocratsall2022
    8165.6073.30237DiscriminationnoDemocratsall2024
    30.4022.1026.20524DiscriminationnoRepublicansall2000
    28.1017.3022.70289DiscriminationnoRepublicansall2002
    21.9012.5017.20316DiscriminationnoRepublicansall2004
    2819.9024585DiscriminationnoRepublicansall2006
    26.5016.8021.70389DiscriminationnoRepublicansall2008
    25.6016.1020.90408DiscriminationnoRepublicansall2010
    25.2014.8020345DiscriminationnoRepublicansall2012
    22.2014.2018.20468DiscriminationnoRepublicansall2014
    25.101721.10538DiscriminationnoRepublicansall2016
    28.501823.30458DiscriminationnoRepublicansall2018
    22.7015.9019.30732DiscriminationnoRepublicansall2021
    31.801724.40321DiscriminationnoRepublicansall2022
    27.8015.1021.40314DiscriminationnoRepublicansall2024
    52.9043.4048.20520Lack of motivation or will powernoDemocratsall2000
    5540.3047.70255Lack of motivation or will powernoDemocratsall2002
    52.7038.5045.60263Lack of motivation or will powernoDemocratsall2004
    47.2037.2042.20531Lack of motivation or will powernoDemocratsall2006
    5442.5048.30421Lack of motivation or will powernoDemocratsall2008
    45.4033.9039.70417Lack of motivation or will powernoDemocratsall2010
    46.8035.1040.90389Lack of motivation or will powernoDemocratsall2012
    4332.7037.90470Lack of motivation or will powernoDemocratsall2014
    33.2024.1028.60546Lack of motivation or will powernoDemocratsall2016
    29.9019.2024.50397Lack of motivation or will powernoDemocratsall2018
    18.9012.6015.80856Lack of motivation or will powernoDemocratsall2021
    18.807.5013.20305Lack of motivation or will powernoDemocratsall2022
    15.505.1010.30244Lack of motivation or will powernoDemocratsall2024
    59.8050.4055.10518Lack of motivation or will powernoRepublicansall2000
    60.6047.2053.90272Lack of motivation or will powernoRepublicansall2002
    60.5047.9054.20312Lack of motivation or will powernoRepublicansall2004
    62.3052.9057.60583Lack of motivation or will powernoRepublicansall2006
    60.5048.5054.50382Lack of motivation or will powernoRepublicansall2008
    6654.5060.30403Lack of motivation or will powernoRepublicansall2010
    63.6050.6057.10340Lack of motivation or will powernoRepublicansall2012
    59.2048.8054466Lack of motivation or will powernoRepublicansall2014
    57.1047.1052.10533Lack of motivation or will powernoRepublicansall2016
    50.3037.9044.10442Lack of motivation or will powernoRepublicansall2018
    49.2040.7045721Lack of motivation or will powernoRepublicansall2021
    52.3035.1043.70314Lack of motivation or will powernoRepublicansall2022
    48.203340.60315Lack of motivation or will powernoRepublicansall2024
    27.3019.4023.30551At the help end of the scalenoDemocratsall2000
    21.2011.2016.20284At the help end of the scalenoDemocratsall2002
    21.6011.3016.50278At the help end of the scalenoDemocratsall2004
    23.101519518At the help end of the scalenoDemocratsall2006
    21.7013.1017.40420At the help end of the scalenoDemocratsall2008
    26.3016.3021.30422At the help end of the scalenoDemocratsall2010
    23.1013.8018.50399At the help end of the scalenoDemocratsall2012
    23.1014.9019471At the help end of the scalenoDemocratsall2014
    42.3032.8037.50584At the help end of the scalenoDemocratsall2016
    45.1033.6039.30417At the help end of the scalenoDemocratsall2018
    61.3052.9057.10853At the help end of the scalenoDemocratsall2021
    57.4045.2051.30592At the help end of the scalenoDemocratsall2022
    53.1040.7046.90490At the help end of the scalenoDemocratsall2024
    9.705.107.40584At the help end of the scalenoRepublicansall2000
    10.9047.50280At the help end of the scalenoRepublicansall2002
    10.303.907.10307At the help end of the scalenoRepublicansall2004
    9.904.907.40584At the help end of the scalenoRepublicansall2006
    5.401.203.30412At the help end of the scalenoRepublicansall2008
    8.6035.80384At the help end of the scalenoRepublicansall2010
    6.101.303.70349At the help end of the scalenoRepublicansall2012
    5.401.603.50463At the help end of the scalenoRepublicansall2014
    104.807.40526At the help end of the scalenoRepublicansall2016
    15.307.4011.40457At the help end of the scalenoRepublicansall2018
    10.405.708.10744At the help end of the scalenoRepublicansall2021
    10.1047.10667At the help end of the scalenoRepublicansall2022
    9.603.906.70602At the help end of the scalenoRepublicansall2024
    4433.1038.50384Too little spentnoDemocratsall2000
    40.2028.9034.50368Too little spentnoDemocratsall2002
    39.8028.7034.30389Too little spentnoDemocratsall2004
    4734.6040.80384Too little spentnoDemocratsall2006
    46.2033.2039.70292Too little spentnoDemocratsall2008
    44.6030.4037.50260Too little spentnoDemocratsall2010
    48.4034.3041.30277Too little spentnoDemocratsall2012
    42.603136.80348Too little spentnoDemocratsall2014
    67.8056.4062.10396Too little spentnoDemocratsall2016
    74.9061.7068.30288Too little spentnoDemocratsall2018
    80.4071.8076.10629Too little spentnoDemocratsall2021
    78.3065.6071.90454Too little spentnoDemocratsall2022
    74.4060.9067.60370Too little spentnoDemocratsall2024
    27.201822.60374Too little spentnoRepublicansall2000
    20.3012.5016.40431Too little spentnoRepublicansall2002
    24.1015.5019.80456Too little spentnoRepublicansall2004
    22.1013.3017.70395Too little spentnoRepublicansall2006
    27.6015.4021.50253Too little spentnoRepublicansall2008
    177.6012.30277Too little spentnoRepublicansall2010
    23.2011.3017.30230Too little spentnoRepublicansall2012
    2010.9015.50320Too little spentnoRepublicansall2014
    38.5027.3032.90377Too little spentnoRepublicansall2016
    4228.3035.10330Too little spentnoRepublicansall2018
    19.9012.6016.20550Too little spentnoRepublicansall2021
    28.6015.4022442Too little spentnoRepublicansall2022
    23.5013.5018.50442Too little spentnoRepublicansall2024
    21.9014.5018.20519Favors preferential hiringnoDemocratsall2000
    20.6010.2015.40259Favors preferential hiringnoDemocratsall2002
    19.109.2014.10263Favors preferential hiringnoDemocratsall2004
    18.1011.1014.60538Favors preferential hiringnoDemocratsall2006
    19.301115.20414Favors preferential hiringnoDemocratsall2008
    23.4013.9018.60412Favors preferential hiringnoDemocratsall2010
    26.5016.6021.60386Favors preferential hiringnoDemocratsall2012
    24.9016.2020.50460Favors preferential hiringnoDemocratsall2014
    32.4023.3027.80527Favors preferential hiringnoDemocratsall2016
    41.5029.1035.30367Favors preferential hiringnoDemocratsall2018
    43.6035.1039.40839Favors preferential hiringnoDemocratsall2021
    48.8036.8042.80589Favors preferential hiringnoDemocratsall2022
    43.8031.8037.80471Favors preferential hiringnoDemocratsall2024
    11.406.208.80543Favors preferential hiringnoRepublicansall2000
    11.304.307.80288Favors preferential hiringnoRepublicansall2002
    10.804.207.50317Favors preferential hiringnoRepublicansall2004
    9.304.506.90591Favors preferential hiringnoRepublicansall2006
    10.604.407.50391Favors preferential hiringnoRepublicansall2008
    11.805.208.50402Favors preferential hiringnoRepublicansall2010
    9.1036.10345Favors preferential hiringnoRepublicansall2012
    10.104.607.30464Favors preferential hiringnoRepublicansall2014
    11.6068.80535Favors preferential hiringnoRepublicansall2016
    11.804.908.40450Favors preferential hiringnoRepublicansall2018
    4.401.502.90724Favors preferential hiringnoRepublicansall2021
    10.904.307.60627Favors preferential hiringnoRepublicansall2022
    11.105.208.10626Favors preferential hiringnoRepublicansall2024
    19.6012.8016.20550Rejects the special-favors viewnoDemocratsall2000
    177.5012.20264Rejects the special-favors viewnoDemocratsall2002
    22.9012.1017.50269Rejects the special-favors viewnoDemocratsall2004
    21.5013.9017.70553Rejects the special-favors viewnoDemocratsall2006
    19.6011.4015.50436Rejects the special-favors viewnoDemocratsall2008
    24.7015.2019.90432Rejects the special-favors viewnoDemocratsall2010
    25.4015.9020.70398Rejects the special-favors viewnoDemocratsall2012
    27.901923.50478Rejects the special-favors viewnoDemocratsall2014
    40.8031.2036557Rejects the special-favors viewnoDemocratsall2016
    443238400Rejects the special-favors viewnoDemocratsall2018
    59.9051.4055.70864Rejects the special-favors viewnoDemocratsall2021
    64.705358.90615Rejects the special-favors viewnoDemocratsall2022
    574551499Rejects the special-favors viewnoDemocratsall2024
    1168.50552Rejects the special-favors viewnoRepublicansall2000
    11.104.207.70290Rejects the special-favors viewnoRepublicansall2002
    125.108.60324Rejects the special-favors viewnoRepublicansall2004
    9.304.506.90599Rejects the special-favors viewnoRepublicansall2006
    9.803.806.80393Rejects the special-favors viewnoRepublicansall2008
    8.503.105.80420Rejects the special-favors viewnoRepublicansall2010
    6.201.303.80355Rejects the special-favors viewnoRepublicansall2012
    93.906.40476Rejects the special-favors viewnoRepublicansall2014
    104.807.40552Rejects the special-favors viewnoRepublicansall2016
    16.208.1012.10463Rejects the special-favors viewnoRepublicansall2018
    9.4057.20735Rejects the special-favors viewnoRepublicansall2021
    17.809.4013.60651Rejects the special-favors viewnoRepublicansall2022
    12.906.609.70647Rejects the special-favors viewnoRepublicansall2024
    44.9035.6040.30523DiscriminationnoDemocrats, in personbymode2000
    43.9029.9036.90260DiscriminationnoDemocrats, in personbymode2002
    53.403745.20197DiscriminationnoDemocrats, in personbymode2004
    42.3032.4037.30498DiscriminationnoDemocrats, in personbymode2006
    43.6031.4037.50362DiscriminationnoDemocrats, in personbymode2008
    49.403743.20379DiscriminationnoDemocrats, in personbymode2010
    48.1035.7041.90348DiscriminationnoDemocrats, in personbymode2012
    52.1040.6046.40397DiscriminationnoDemocrats, in personbymode2014
    61.3050.8056499DiscriminationnoDemocrats, in personbymode2016
    70.2057.5063.90355DiscriminationnoDemocrats, in personbymode2018
    85.8078.9082.40775DiscriminationnoDemocrats, webbymode2021
    84.9071.7078.30300DiscriminationyesDemocrats, webbymode2023
    30.4022.1026.20524DiscriminationnoRepublicans, in personbymode2000
    28.1017.3022.70289DiscriminationnoRepublicans, in personbymode2002
    23.8013.1018.40253DiscriminationnoRepublicans, in personbymode2004
    28.4020.1024.30548DiscriminationnoRepublicans, in personbymode2006
    26.101621.10339DiscriminationnoRepublicans, in personbymode2008
    26.5016.5021.50361DiscriminationnoRepublicans, in personbymode2010
    26.201520.60293DiscriminationnoRepublicans, in personbymode2012
    24.6015.8020.20417DiscriminationnoRepublicans, in personbymode2014
    25.3016.9021.10500DiscriminationnoRepublicans, in personbymode2016
    3119.8025.40409DiscriminationnoRepublicans, in personbymode2018
    38.1021.7029.90249DiscriminationyesRepublicans, in personbymode2023
    22.9015.7019.30621DiscriminationnoRepublicans, webbymode2021
    23.8010.9017.40319DiscriminationyesRepublicans, webbymode2023
    52.9043.4048.20520Lack of motivation or will powernoDemocrats, in personbymode2000
    5540.3047.70255Lack of motivation or will powernoDemocrats, in personbymode2002
    53.9037.8045.90201Lack of motivation or will powernoDemocrats, in personbymode2004
    49.8039.6044.70497Lack of motivation or will powernoDemocrats, in personbymode2006
    56.4043.9050.20360Lack of motivation or will powernoDemocrats, in personbymode2008
    46.5034.5040.50373Lack of motivation or will powernoDemocrats, in personbymode2010
    48.8036.1042.50339Lack of motivation or will powernoDemocrats, in personbymode2012
    43.6032.4038395Lack of motivation or will powernoDemocrats, in personbymode2014
    34.1024.4029.30504Lack of motivation or will powernoDemocrats, in personbymode2016
    29.6018.4024359Lack of motivation or will powernoDemocrats, in personbymode2018
    17.9011.5014.70774Lack of motivation or will powernoDemocrats, webbymode2021
    11.3037.20301Lack of motivation or will poweryesDemocrats, webbymode2023
    59.8050.4055.10518Lack of motivation or will powernoRepublicans, in personbymode2000
    60.6047.2053.90272Lack of motivation or will powernoRepublicans, in personbymode2002
    60.4046.5053.40249Lack of motivation or will powernoRepublicans, in personbymode2004
    62.605357.80545Lack of motivation or will powernoRepublicans, in personbymode2006
    61.2048.8055332Lack of motivation or will powernoRepublicans, in personbymode2008
    64.8052.8058.80360Lack of motivation or will powernoRepublicans, in personbymode2010
    65.5051.9058.70289Lack of motivation or will powernoRepublicans, in personbymode2012
    58.2047.2052.70416Lack of motivation or will powernoRepublicans, in personbymode2014
    56.9046.5051.70495Lack of motivation or will powernoRepublicans, in personbymode2016
    51.3038.3044.80395Lack of motivation or will powernoRepublicans, in personbymode2018
    5638.3047.20247Lack of motivation or will poweryesRepublicans, in personbymode2023
    49.5040.3044.90612Lack of motivation or will powernoRepublicans, webbymode2021
    47.1030.5038.80318Lack of motivation or will poweryesRepublicans, webbymode2023
    27.3019.4023.30551At the help end of the scalenoDemocrats, in personbymode2000
    21.2011.2016.20284At the help end of the scalenoDemocrats, in personbymode2002
    21.4010.2015.80228At the help end of the scalenoDemocrats, in personbymode2004
    2415.7019.80481At the help end of the scalenoDemocrats, in personbymode2006
    21.4012.6017371At the help end of the scalenoDemocrats, in personbymode2008
    26.701621.30370At the help end of the scalenoDemocrats, in personbymode2010
    23.5013.4018.50344At the help end of the scalenoDemocrats, in personbymode2012
    24.3015.3019.80400At the help end of the scalenoDemocrats, in personbymode2014
    42.2032.5037.40539At the help end of the scalenoDemocrats, in personbymode2016
    46.1033.9040367At the help end of the scalenoDemocrats, in personbymode2018
    53.4038.3045.90349At the help end of the scaleyesDemocrats, in personbymode2023
    61.4052.5056.90773At the help end of the scalenoDemocrats, webbymode2021
    57.9046.7052.30630At the help end of the scaleyesDemocrats, webbymode2023
    9.705.107.40584At the help end of the scalenoRepublicans, in personbymode2000
    10.9047.50280At the help end of the scalenoRepublicans, in personbymode2002
    10.603.507245At the help end of the scalenoRepublicans, in personbymode2004
    9.404.406.90547At the help end of the scalenoRepublicans, in personbymode2006
    6.301.603.90361At the help end of the scalenoRepublicans, in personbymode2008
    9.503.206.40339At the help end of the scalenoRepublicans, in personbymode2010
    7.501.704.60298At the help end of the scalenoRepublicans, in personbymode2012
    5.801.603.70401At the help end of the scalenoRepublicans, in personbymode2014
    9.804.407.10490At the help end of the scalenoRepublicans, in personbymode2016
    16.10812.10418At the help end of the scalenoRepublicans, in personbymode2018
    14.606.3010.40492At the help end of the scaleyesRepublicans, in personbymode2023
    10.505.608.10634At the help end of the scalenoRepublicans, webbymode2021
    5.501.503.50663At the help end of the scaleyesRepublicans, webbymode2023
    4433.1038.50384Too little spentnoDemocrats, in personbymode2000
    40.2028.9034.50368Too little spentnoDemocrats, in personbymode2002
    34.9022.9028.90293Too little spentnoDemocrats, in personbymode2004
    47.2035.3041.20361Too little spentnoDemocrats, in personbymode2006
    41.9028.5035.20259Too little spentnoDemocrats, in personbymode2008
    43.5028.6036.10230Too little spentnoDemocrats, in personbymode2010
    51.7036.5044.10238Too little spentnoDemocrats, in personbymode2012
    43.4030.8037.10295Too little spentnoDemocrats, in personbymode2014
    69.1057.2063.10360Too little spentnoDemocrats, in personbymode2016
    76.1062.4069.30261Too little spentnoDemocrats, in personbymode2018
    78.706270.40265Too little spentyesDemocrats, in personbymode2023
    81.2072.1076.70561Too little spentnoDemocrats, webbymode2021
    76.3064.7070.50491Too little spentyesDemocrats, webbymode2023
    27.201822.60374Too little spentnoRepublicans, in personbymode2000
    20.3012.5016.40431Too little spentnoRepublicans, in personbymode2002
    23.7013.9018.80351Too little spentnoRepublicans, in personbymode2004
    20.4011.8016.10373Too little spentnoRepublicans, in personbymode2006
    26.1013.8020218Too little spentnoRepublicans, in personbymode2008
    17.107.2012.20253Too little spentnoRepublicans, in personbymode2010
    25.201218.60192Too little spentnoRepublicans, in personbymode2012
    22.8012.4017.60271Too little spentnoRepublicans, in personbymode2014
    38.4026.7032.50351Too little spentnoRepublicans, in personbymode2016
    39.6025.8032.70292Too little spentnoRepublicans, in personbymode2018
    3619.7027.90309Too little spentyesRepublicans, in personbymode2023
    20.9012.8016.90462Too little spentnoRepublicans, webbymode2021
    18.609.8014.20483Too little spentyesRepublicans, webbymode2023
    21.9014.5018.20519Favors preferential hiringnoDemocrats, in personbymode2000
    20.6010.2015.40259Favors preferential hiringnoDemocrats, in personbymode2002
    19.508.4013.90205Favors preferential hiringnoDemocrats, in personbymode2004
    18.3011.2014.70504Favors preferential hiringnoDemocrats, in personbymode2006
    19.7010.7015.20357Favors preferential hiringnoDemocrats, in personbymode2008
    24.5014.4019.50370Favors preferential hiringnoDemocrats, in personbymode2010
    25.6015.3020.40335Favors preferential hiringnoDemocrats, in personbymode2012
    25.8016.3021387Favors preferential hiringnoDemocrats, in personbymode2014
    33.2023.5028.30483Favors preferential hiringnoDemocrats, in personbymode2016
    42.2029.2035.70332Favors preferential hiringnoDemocrats, in personbymode2018
    50.1035.3042.70337Favors preferential hiringyesDemocrats, in personbymode2023
    43.7034.7039.20758Favors preferential hiringnoDemocrats, webbymode2021
    46.2035.1040.60618Favors preferential hiringyesDemocrats, webbymode2023
    11.406.208.80543Favors preferential hiringnoRepublicans, in personbymode2000
    11.304.307.80288Favors preferential hiringnoRepublicans, in personbymode2002
    10.303.306.80254Favors preferential hiringnoRepublicans, in personbymode2004
    9.904.807.40552Favors preferential hiringnoRepublicans, in personbymode2006
    125.108.50338Favors preferential hiringnoRepublicans, in personbymode2008
    13.205.909.60358Favors preferential hiringnoRepublicans, in personbymode2010
    10.703.707.20297Favors preferential hiringnoRepublicans, in personbymode2012
    8.603.406413Favors preferential hiringnoRepublicans, in personbymode2014
    11.605.708.70497Favors preferential hiringnoRepublicans, in personbymode2016
    13.605.909.80403Favors preferential hiringnoRepublicans, in personbymode2018
    17.508.6013468Favors preferential hiringyesRepublicans, in personbymode2023
    4.301.302.80618Favors preferential hiringnoRepublicans, webbymode2021
    4.9012.90656Favors preferential hiringyesRepublicans, webbymode2023
    19.6012.8016.20550Rejects the special-favors viewnoDemocrats, in personbymode2000
    177.5012.20264Rejects the special-favors viewnoDemocrats, in personbymode2002
    2512.5018.70206Rejects the special-favors viewnoDemocrats, in personbymode2004
    2113.4017.20518Rejects the special-favors viewnoDemocrats, in personbymode2006
    20.9011.7016.30374Rejects the special-favors viewnoDemocrats, in personbymode2008
    23.2013.6018.40387Rejects the special-favors viewnoDemocrats, in personbymode2010
    23.4013.6018.50350Rejects the special-favors viewnoDemocrats, in personbymode2012
    27.6018.1022.90403Rejects the special-favors viewnoDemocrats, in personbymode2014
    40.9030.8035.80512Rejects the special-favors viewnoDemocrats, in personbymode2016
    44.8032.1038.50362Rejects the special-favors viewnoDemocrats, in personbymode2018
    62.8048.4055.60370Rejects the special-favors viewyesDemocrats, in personbymode2023
    60.9051.9056.40779Rejects the special-favors viewnoDemocrats, webbymode2021
    61.9050.8056.30627Rejects the special-favors viewyesDemocrats, webbymode2023
    1168.50552Rejects the special-favors viewnoRepublicans, in personbymode2000
    11.104.207.70290Rejects the special-favors viewnoRepublicans, in personbymode2002
    13.405.409.40259Rejects the special-favors viewnoRepublicans, in personbymode2004
    9.604.707.20561Rejects the special-favors viewnoRepublicans, in personbymode2006
    9.903.706.80341Rejects the special-favors viewnoRepublicans, in personbymode2008
    9.603.606.60373Rejects the special-favors viewnoRepublicans, in personbymode2010
    5.400.8003.10305Rejects the special-favors viewnoRepublicans, in personbymode2012
    9.5046.80424Rejects the special-favors viewnoRepublicans, in personbymode2014
    9.704.407514Rejects the special-favors viewnoRepublicans, in personbymode2016
    17.809.1013.40415Rejects the special-favors viewnoRepublicans, in personbymode2018
    189.2013.60501Rejects the special-favors viewyesRepublicans, in personbymode2023
    9.604.807.20625Rejects the special-favors viewnoRepublicans, webbymode2021
    12.6069.30662Rejects the special-favors viewyesRepublicans, webbymode2023
    The Democratic lines move after 2014 in every panel, and the Republican changes are smaller by point estimate in every panel. In the mode view the open circle at 2023 pools the 2022 and 2024 interviews, and the in-person Democratic cells for the discrimination and will-power items fall below the sample floor and are not drawn. Lines are weighted shares of white respondents with 95% bands.Note: The GSS was in person (with some telephone interviews from 2004) through 2018, had no 2020 survey, had 87.3% of its 2021 interviews on the web and mixed in-person and web interviews in 2022 and 2024. The mode view plots in-person interviews through 2018 and web interviews from 2021.Note: The 2022 and 2024 cells alone fall below the effective sample floor for several items, so the mode view pools them in one point at 2023; a point is dropped when the pooled cell is still below the floor, as for in-person Democrats on the discrimination and will-power items. The help-end panel counts the two answers nearest to government should help. Intervals assume a simple random sample of the weighted respondents (the Kish effective sample size); they do not use the survey's sampling design.Source: General Social Survey, 1972-2024; weighted (weight_use); white respondents by party, leaners included (race code white, Hispanic whites included); 95% intervals on the Kish effective sample size

    The mode view of Figure 6 shows that the Democratic web levels lie beyond the 2014 level for every dated item. The pooled 2022-2024 in-person cell clears the sample floor for 44 items out of sixEstimandnumber of dated items whose pooled in-person Democratic cell clears the sample floorWeightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite DemocratsPeriod2022 and 2024 pooled, in-person interviewsSourceGeneral Social Survey of the six dated items, and for those it also lies beyond the 2014 level. Among white Democrats the share at the help end of the government-help scale was 19.0%19.0% percent of respondents giving the answerEstimandweighted share who gave the answer (places self at the help end of the government-help scale (1 or 2 of 5))Interval95% CI 14.9% to 23.1%n471Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite democrats (leaners included; GSS race code, Hispanic whites included) who answered the itemPeriod2014SourceGeneral Social Survey in 2014 and 37.5%37.5% percent of respondents giving the answerEstimandweighted share who gave the answer (places self at the help end of the government-help scale (1 or 2 of 5))Interval95% CI 32.8% to 42.3%n584Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite democrats (leaners included; GSS race code, Hispanic whites included) who answered the itemPeriod2016SourceGeneral Social Survey in 2016, and in 2022 and 2024 pooled it was 45.9% (95% CI 38.3% to 53.4%)45.9% percent of respondents giving the answerEstimandweighted share who said government should help improve Black people's living standardsInterval95% CI 38.3% to 53.4%n349Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite Democrats (leaners included; GSS race code, Hispanic whites included) who answered the itemPeriod2022 and 2024 pooled, in person interviewsSourceGeneral Social Survey in person and 52.3% (95% CI 46.7% to 57.9%)52.3% percent of respondents giving the answerEstimandweighted share who said government should help improve Black people's living standardsInterval95% CI 46.7% to 57.9%n630Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite Democrats (leaners included; GSS race code, Hispanic whites included) who answered the itemPeriod2022 and 2024 pooled, web interviewsSourceGeneral Social Survey on the web.

    Black respondents moved in the same direction on some items

    Figure 7 adds race of respondent. Black respondents give the structural or pro-government answer more often than white respondents in every period on 55 items out of sixEstimandnumber of dated items on which the interval for Black respondents lies above the interval for white respondents in all three periodsWeightsweight_use (wtssnrps from 2004, wtssps before)PopulationBlack and white respondentsPeriod2010-2014, 2016-2018 and 2021-2024SourceGeneral Social Survey of the six dated items. From 2010-2014 to 2016-2018 the help-end share changed by +9.6 (95% CI 2.4 to 16.8)+9.6 percentage pointsEstimandchange in the weighted share who gave the answer (places self at the help end of the government-help scale (1 or 2 of 5))Interval95% CI 2.4 to 16.8n1,182Weightsweight_use (wtssnrps from 2004, wtssps before)Populationblack respondents (GSS race codes, Hispanic whites included in white) who answered the itemPeriod2010-2014 to 2016-2018SourceGeneral Social Survey points among Black respondents and +10.0 (95% CI 7.6 to 12.4)+10.0 percentage pointsEstimandchange in the weighted share who gave the answer (places self at the help end of the government-help scale (1 or 2 of 5))Interval95% CI 7.6 to 12.4n5,643Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race codes, Hispanic whites included in white) who answered the itemPeriod2010-2014 to 2016-2018SourceGeneral Social Survey among white respondents, and the share citing discrimination by +7.0 (95% CI 0.2 to 13.8)+7.0 percentage pointsEstimandchange in the weighted share who gave the answer (discrimination)Interval95% CI 0.2 to 13.8n1,219Weightsweight_use (wtssnrps from 2004, wtssps before)Populationblack respondents (GSS race codes, Hispanic whites included in white) who answered the itemPeriod2010-2014 to 2016-2018SourceGeneral Social Survey and +8.4 (95% CI 5.3 to 11.5)+8.4 percentage pointsEstimandchange in the weighted share who gave the answer (discrimination)Interval95% CI 5.3 to 11.5n5,589Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race codes, Hispanic whites included in white) who answered the itemPeriod2010-2014 to 2016-2018SourceGeneral Social Survey. Among Black respondents the interval excludes zero for 33 items out of sixEstimandnumber of items whose change in the share has an interval that excludes zeroWeightsweight_use (wtssnrps from 2004, wtssps before)PopulationBlack respondentsPeriod2010-2014 to 2016-2018SourceGeneral Social Survey of the six dated items, and among white respondents for 66 items out of sixEstimandnumber of items whose change in the share has an interval that excludes zeroWeightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondentsPeriod2010-2014 to 2016-2018SourceGeneral Social Survey. Black respondents' shares show no change distinguishable from zero on affirmative action (−0.8 (95% CI −7.7 to 6.2)−0.8 percentage pointsEstimandchange in the weighted share who gave the answer (favors preference in hiring and promoting black people)Interval95% CI −7.7 to 6.2n1,177Weightsweight_use (wtssnrps from 2004, wtssps before)Populationblack respondents (GSS race codes, Hispanic whites included in white) who answered the itemPeriod2010-2014 to 2016-2018SourceGeneral Social Survey points) or will power (−0.2 (95% CI −7.1 to 6.7)−0.2 percentage pointsEstimandchange in the weighted share who gave the answer (lack of motivation or will power)Interval95% CI −7.1 to 6.7n1,229Weightsweight_use (wtssnrps from 2004, wtssps before)Populationblack respondents (GSS race codes, Hispanic whites included in white) who answered the itemPeriod2010-2014 to 2016-2018SourceGeneral Social Survey), but those intervals are wide enough to overlap the white intervals of +5.7 (95% CI 3.3 to 8.2)+5.7 percentage pointsEstimandchange in the weighted share who gave the answer (favors preference in hiring and promoting black people)Interval95% CI 3.3 to 8.2n5,500Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race codes, Hispanic whites included in white) who answered the itemPeriod2010-2014 to 2016-2018SourceGeneral Social Survey and −9.0 (95% CI −12.2 to −5.7)−9.0 percentage pointsEstimandchange in the weighted share who gave the answer (lack of motivation or will power)Interval95% CI −12.2 to −5.7n5,556Weightsweight_use (wtssnrps from 2004, wtssps before)Populationwhite respondents (GSS race codes, Hispanic whites included in white) who answered the itemPeriod2010-2014 to 2016-2018SourceGeneral Social Survey.

    Figure 7

    From 2010-2014 to 2016-2018 white respondents moved on 6 of the six dated items and Black respondents on 3, with wider intervals for Black respondents

    Share of white (GSS race code, Hispanic whites included) and of Black respondents giving the answer named above each panel, percent, by pooled period of surveys, 95% intervals; all interviews and in-person interviews only

    • Black respondents
    • White respondents

    Discrimination

    Lack of motivation or will power

    At the help end of the scale

    Too little spent

    Favors preferential hiring

    Rejects the special-favors view

    Share giving the answer (%)

    Show data
    Data for From 2010-2014 to 2016-2018 white respondents moved on 6 of the six dated items and Black respondents on 3, with wider intervals for Black respondents
    conf_highconf_lowestimatenpanelperiodrace_labelview
    59.8050.6055.20669Discrimination2010-2014Black respondentsall
    32.5028.6030.603,181Discrimination2010-2014White respondentsall
    67.2057.2062.20550Discrimination2016-2018Black respondentsall
    41.4036.5038.902,408Discrimination2016-2018White respondentsall
    79.3070.4074.80684Discrimination2021-2024Black respondentsall
    50.4045.9048.203,556Discrimination2021-2024White respondentsall
    50.5041.2045.90671Lack of motivation or will power2010-2014Black respondentsall
    50.204648.103,156Lack of motivation or will power2010-2014White respondentsall
    50.8040.5045.60558Lack of motivation or will power2016-2018Black respondentsall
    41.6036.7039.202,400Lack of motivation or will power2016-2018White respondentsall
    34.5025.1029.80688Lack of motivation or will power2021-2024Black respondentsall
    30.8026.8028.803,535Lack of motivation or will power2021-2024White respondentsall
    44.1034.4039.20617At the help end of the scale2010-2014Black respondentsall
    13.2010.5011.903,181At the help end of the scale2010-2014White respondentsall
    54.1043.6048.90565At the help end of the scale2016-2018Black respondentsall
    23.9019.8021.902,462At the help end of the scale2016-2018White respondentsall
    64.6055.7060.10996At the help end of the scale2021-2024Black respondentsall
    28.5025.1026.805,115At the help end of the scale2021-2024White respondentsall
    75.2065.2070.20480Too little spent2010-2014Black respondentsall
    28.7024.3026.502,143Too little spent2010-2014White respondentsall
    8473.7078.90409Too little spent2016-2018Black respondentsall
    50.3044.4047.401,724Too little spent2016-2018White respondentsall
    9082.8086.40781Too little spent2021-2024Black respondentsall
    45.5041.1043.303,796Too little spent2021-2024White respondentsall
    47.1037.8042.50641Favors preferential hiring2010-2014Black respondentsall
    1512.1013.503,150Favors preferential hiring2010-2014White respondentsall
    46.9036.5041.70536Favors preferential hiring2016-2018Black respondentsall
    21.3017.3019.302,350Favors preferential hiring2016-2018White respondentsall
    48.7039.8044.301,008Favors preferential hiring2021-2024Black respondentsall
    22.9019.8021.405,023Favors preferential hiring2021-2024White respondentsall
    34.9026.5030.70685Rejects the special-favors view2010-2014Black respondentsall
    13.6010.9012.303,274Rejects the special-favors view2010-2014White respondentsall
    41.5031.5036.50555Rejects the special-favors view2016-2018Black respondentsall
    22.4018.5020.502,474Rejects the special-favors view2016-2018White respondentsall
    4738.4042.701,048Rejects the special-favors view2021-2024Black respondentsall
    29.5026.1027.805,217Rejects the special-favors view2021-2024White respondentsall
    60.2050.5055.30617Discrimination2010-2014Black respondentsinperson
    33.6029.5031.502,781Discrimination2010-2014White respondentsinperson
    67.3056.8062514Discrimination2016-2018Black respondentsinperson
    42.8037.7040.302,199Discrimination2016-2018White respondentsinperson
    58.5046.1052.30548Discrimination2022-2024White respondentsinperson
    50.704145.90620Lack of motivation or will power2010-2014Black respondentsinperson
    50.404648.202,760Lack of motivation or will power2010-2014White respondentsinperson
    52.3041.7047523Lack of motivation or will power2016-2018Black respondentsinperson
    42.1036.9039.502,189Lack of motivation or will power2016-2018White respondentsinperson
    38.1026.7032.40544Lack of motivation or will power2022-2024White respondentsinperson
    43.5033.2038.40567At the help end of the scale2010-2014Black respondentsinperson
    13.6010.7012.202,762At the help end of the scale2010-2014White respondentsinperson
    54.6043.8049.20540At the help end of the scale2016-2018Black respondentsinperson
    23.6019.4021.502,256At the help end of the scale2016-2018White respondentsinperson
    61.6045.7053.70363At the help end of the scale2022-2024Black respondentsinperson
    29.3021.5025.401,082At the help end of the scale2022-2024White respondentsinperson
    75.406570.20448Too little spent2010-2014Black respondentsinperson
    29.3024.5026.901,855Too little spent2010-2014White respondentsinperson
    84.6074.1079.40387Too little spent2016-2018Black respondentsinperson
    50.5044.3047.401,557Too little spent2016-2018White respondentsinperson
    89.5076.4083279Too little spent2022-2024Black respondentsinperson
    51.7040.5046.10752Too little spent2022-2024White respondentsinperson
    47.1037.4042.20592Favors preferential hiring2010-2014Black respondentsinperson
    15.3012.2013.702,755Favors preferential hiring2010-2014White respondentsinperson
    46.9036.1041.50501Favors preferential hiring2016-2018Black respondentsinperson
    2217.8019.902,144Favors preferential hiring2016-2018White respondentsinperson
    58.1042.2050.20363Favors preferential hiring2022-2024Black respondentsinperson
    29.9022.10261,048Favors preferential hiring2022-2024White respondentsinperson
    35.1026.2030.60632Rejects the special-favors view2010-2014Black respondentsinperson
    13.3010.4011.802,870Rejects the special-favors view2010-2014White respondentsinperson
    41.6031.3036.50521Rejects the special-favors view2016-2018Black respondentsinperson
    22.9018.7020.802,257Rejects the special-favors view2016-2018White respondentsinperson
    44.6029.8037.20386Rejects the special-favors view2022-2024Black respondentsinperson
    30.3022.8026.601,148Rejects the special-favors view2022-2024White respondentsinperson
    Black respondents are above white respondents in every period on 5 of the six dated items, and where both groups moved between 2010-2014 and 2016-2018 they moved the same way. Each dot is a weighted share for one race and period; the lines are 95% intervals, which are wide for Black respondents because their samples are small.Note: The in-person view leaves out the 2021 survey and the telephone interviews and uses the in-person interviews of 2022 and 2024; in that view Black respondents' cells for the discrimination and will-power items in 2022 and 2024 fall below the effective sample floor and are not drawn. 'Black' and 'white' are the GSS race codes; Hispanic origin is not separated, and Hispanic respondents appear in whichever race code they chose. The help-end panel counts the two answers nearest to government should help. Intervals assume a simple random sample of the weighted respondents (the Kish effective sample size); they do not use the survey's sampling design.Source: General Social Survey, 1972-2024; weighted (weight_use); white and Black respondents (GSS race codes, Hispanic whites included in white); 95% intervals on the Kish effective sample size

    The tie to the vote rose from 2012 to 2016, and the 2008 to 2012 change is within sampling error

    The vote analysis uses the 16,29916,299 respondentsEstimandcount of respondentsWeightsnone (count)Populationwhite non-Hispanic ANES respondents who reported a major-party presidential vote and have a resentment index scorePeriodthe nine presidential years from 1988 to 2024 in which the index was askedSourceANES Cumulative Data File white non-Hispanic ANES respondents who reported voting for the Democratic or the Republican candidate and have a racial resentment index score, in the presidential years from 1988 to 2024 in which the index was asked (it was not asked in 1996). In each year a linear probability modelLinear probability modelA regression of a yes-or-no outcome on predictors, fitted by least squares so that each coefficient is a change in the probability of yes. The book weights it and reports robust standard errors. of a Republican vote on an attitude, standardized within the year, gives a slope: the points of Republican two-party vote that go with a standard deviation more of the racially conservative answer. Higher resentment, a colder rating of Black Americans, placing aid to Black Americans at the self-help end and opposing affirmative action are the conservative poles, and the other three attitudes have their own samples, which include 1996.

    For the racial resentment index, the slope was 10.1 (95% CI 7.1 to 13.1)10.1 percentage points of the Republican two-party presidential vote per standard deviation of the attitudeEstimandweighted linear probability slope on the standardized attitude (racial resentment index; party not held fixed)Interval95% CI 7.1 to 13.1n952Weightsweight (VCF0009z)Populationwhite non-Hispanic voters for a major-party candidatePeriod1988SourceANES Cumulative Data File in 1988, 17.8 (95% CI 14.1 to 21.5)17.8 percentage points of the Republican two-party presidential vote per standard deviation of the attitudeEstimandweighted linear probability slope on the standardized attitude (racial resentment index; party not held fixed)Interval95% CI 14.1 to 21.5n601Weightsweight (VCF0009z)Populationwhite non-Hispanic voters for a major-party candidatePeriod2004SourceANES Cumulative Data File in 2004, 22.1 (95% CI 19.2 to 25.0)22.1 percentage points of the Republican two-party presidential vote per standard deviation of the attitudeEstimandweighted linear probability slope on the standardized attitude (racial resentment index; party not held fixed)Interval95% CI 19.2 to 25.0n782Weightsweight (VCF0009z)Populationwhite non-Hispanic voters for a major-party candidatePeriod2008SourceANES Cumulative Data File in 2008, 23.2 (95% CI 21.3 to 25.0)23.2 percentage points of the Republican two-party presidential vote per standard deviation of the attitudeEstimandweighted linear probability slope on the standardized attitude (racial resentment index; party not held fixed)Interval95% CI 21.3 to 25.0n2,541Weightsweight (VCF0009z)Populationwhite non-Hispanic voters for a major-party candidatePeriod2012SourceANES Cumulative Data File in 2012, 31.3 (95% CI 29.9 to 32.8)31.3 percentage points of the Republican two-party presidential vote per standard deviation of the attitudeEstimandweighted linear probability slope on the standardized attitude (racial resentment index; party not held fixed)Interval95% CI 29.9 to 32.8n1,942Weightsweight (VCF0009z)Populationwhite non-Hispanic voters for a major-party candidatePeriod2016SourceANES Cumulative Data File in 2016, 34.934.9 percentage points of the Republican two-party presidential vote per standard deviation of the attitudeEstimandweighted linear probability slope on the standardized attitude (racial resentment index; party not held fixed)Interval95% CI 34.0 to 35.7n4,574Weightsweight (VCF0009z)Populationwhite non-Hispanic voters for a major-party candidatePeriod2020SourceANES Cumulative Data File in 2020 and 33.033.0 percentage points of the Republican two-party presidential vote per standard deviation of the attitudeEstimandweighted linear probability slope on the standardized attitude (racial resentment index; party not held fixed)Interval95% CI 31.6 to 34.4n2,968Weightsweight (VCF0009z)Populationwhite non-Hispanic voters for a major-party candidatePeriod2024SourceANES Cumulative Data File in 2024 (Figure 8). It gained +7.7 (95% CI 2.9 to 12.4)+7.7 percentage points of the Republican two-party presidential vote per standard deviation of the attitudeEstimandchange in the weighted linear probability slope on the standardized attitude (party not held fixed)Interval95% CI 2.9 to 12.4n1,553Weightsweight (VCF0009z)Populationwhite non-Hispanic voters for a major-party candidatePeriod1988 to 2004SourceANES Cumulative Data File points from 1988 to 2004, +4.3 (95% CI −0.4 to 9.0)+4.3 percentage points of the Republican two-party presidential vote per standard deviation of the attitudeEstimandchange in the weighted linear probability slope on the standardized attitude (party not held fixed)Interval95% CI −0.4 to 9.0n1,383Weightsweight (VCF0009z)Populationwhite non-Hispanic voters for a major-party candidatePeriod2004 to 2008SourceANES Cumulative Data File from 2004 to 2008, +1.1 (95% CI −2.4 to 4.6)+1.1 percentage points of the Republican two-party presidential vote per standard deviation of the attitudeEstimandchange in the weighted linear probability slope on the standardized attitude (party not held fixed)Interval95% CI −2.4 to 4.6n3,323Weightsweight (VCF0009z)Populationwhite non-Hispanic voters for a major-party candidatePeriod2008 to 2012SourceANES Cumulative Data File from 2008 to 2012 and +8.2 (95% CI 5.8 to 10.5)+8.2 percentage points of the Republican two-party presidential vote per standard deviation of the attitudeEstimandchange in the weighted linear probability slope on the standardized attitude (party not held fixed)Interval95% CI 5.8 to 10.5n4,483Weightsweight (VCF0009z)Populationwhite non-Hispanic voters for a major-party candidatePeriod2012 to 2016SourceANES Cumulative Data File from 2012 to 2016. The interval for 2004 to 2008 includes zero and reaches above the change from 2012 to 2016, so the data do not rule out a rise of that size during the 2008 campaign. The interval for 2008 to 2012 ends below the lower end of the interval for 2012 to 2016.

    Figure 8

    Among white voters the tie between racial resentment and the Republican vote rose from 1988 to 2004 and from 2012 to 2016, with no rise distinguishable from zero from 2008 to 2012

    Percentage points of the Republican share of the two-party presidential vote per standard deviation of each racial attitude, white non-Hispanic voters, ANES 1988 to 2024 and CES 2008 and 2012

    Points of Republican vote per standard deviation of the attitude

    Show data
    Data for Among white voters the tie between racial resentment and the Republican vote rose from 1988 to 2004 and from 2012 to 2016, with no rise distinguishable from zero from 2008 to 2012
    conf_highconf_lowestimatenotherpanelseriesyear
    13.206.8010867noall_freeBlack people should help themselves (aid item)1988
    15.109.2012.20984noall_freeBlack people should help themselves (aid item)1992
    18.7012.1015.40778noall_freeBlack people should help themselves (aid item)1996
    12.602.307.50443noall_freeBlack people should help themselves (aid item)2000
    22.8015.1019547noall_freeBlack people should help themselves (aid item)2004
    22.5015.5019688noall_freeBlack people should help themselves (aid item)2008
    24.2020.3022.302,345noall_freeBlack people should help themselves (aid item)2012
    30.6026.8028.701,797noall_freeBlack people should help themselves (aid item)2016
    33.9031.7032.804,291noall_freeBlack people should help themselves (aid item)2020
    32.6028.9030.702,710noall_freeBlack people should help themselves (aid item)2024
    7.601.304.50922noall_freeColder rating of Black Americans1988
    5.40−0.8002.301,032noall_freeColder rating of Black Americans1992
    8.300.8004.50808noall_freeColder rating of Black Americans1996
    7.2003.60827noall_freeColder rating of Black Americans2000
    5.50−3.301.10588noall_freeColder rating of Black Americans2004
    10.402.806.60767noall_freeColder rating of Black Americans2008
    9.905.207.602,526noall_freeColder rating of Black Americans2012
    13.808.9011.301,919noall_freeColder rating of Black Americans2016
    13.509.9011.704,412noall_freeColder rating of Black Americans2020
    10.505.407.902,827noall_freeColder rating of Black Americans2024
    10.204.307.201,025noall_freeOpposes affirmative action (ANES)1992
    12.504.408.40778noall_freeOpposes affirmative action (ANES)1996
    124.808.40822noall_freeOpposes affirmative action (ANES)2000
    17.8010.2014573noall_freeOpposes affirmative action (ANES)2004
    16.609.5013.10738noall_freeOpposes affirmative action (ANES)2008
    17.1011.9014.502,489noall_freeOpposes affirmative action (ANES)2012
    21.1016.4018.701,911noall_freeOpposes affirmative action (ANES)2016
    25.7022.30244,504noall_freeOpposes affirmative action (ANES)2020
    26.4022.4024.402,921noall_freeOpposes affirmative action (ANES)2024
    26.2024.8025.5019,797yesall_freeOpposes affirmative action (CES)2008
    25.6024.2024.9030,271yesall_freeOpposes affirmative action (CES)2012
    13.107.1010.10952noall_freeRacial resentment index1988
    16.3011.2013.801,052noall_freeRacial resentment index1992
    16.309.5012.90887noall_freeRacial resentment index2000
    21.5014.1017.80601noall_freeRacial resentment index2004
    2519.2022.10782noall_freeRacial resentment index2008
    2521.3023.202,541noall_freeRacial resentment index2012
    32.8029.9031.301,942noall_freeRacial resentment index2016
    35.703434.904,574noall_freeRacial resentment index2020
    34.4031.60332,968noall_freeRacial resentment index2024
    22.5015.5019688noftf_freeBlack people should help themselves (aid item)2008
    24.601720.80539noftf_freeBlack people should help themselves (aid item)2012
    32.2024.2028.20483noftf_freeBlack people should help themselves (aid item)2016
    37.1027.9032.50447noftf_freeBlack people should help themselves (aid item)2024
    10.402.806.60767noftf_freeColder rating of Black Americans2008
    11.3037.10602noftf_freeColder rating of Black Americans2012
    19.208.8014511noftf_freeColder rating of Black Americans2016
    16.906.6011.80488noftf_freeColder rating of Black Americans2024
    16.609.5013.10738noftf_freeOpposes affirmative action (ANES)2008
    15.405.9010.70566noftf_freeOpposes affirmative action (ANES)2012
    23.9014.2019.10506noftf_freeOpposes affirmative action (ANES)2016
    32.4024.8028.60476noftf_freeOpposes affirmative action (ANES)2024
    26.2024.8025.5019,797yesftf_freeOpposes affirmative action (CES)2008
    25.6024.2024.9030,271yesftf_freeOpposes affirmative action (CES)2012
    2519.2022.10782noftf_freeRacial resentment index2008
    25.6018.1021.80609noftf_freeRacial resentment index2012
    33.5027.5030.50523noftf_freeRacial resentment index2016
    37.5031.9034.70495noftf_freeRacial resentment index2024
    The resentment line climbs in steps and the 2008 to 2012 step is within sampling error, in person and in the CES, while 2012 to 2016 is the largest. Each point is a weighted linear probability slope on the standardized attitude with a 95% interval; a positive slope means the more racially conservative attitude goes with a more Republican vote. The shaded years are the Obama presidency; the vote is for the election that ended each year.Note: Attitudes are oriented so that a high value is the more racially conservative pole: higher resentment, a colder rating of Black Americans, placing Black people's aid on the self-help end, and opposing affirmative action. Each attitude is standardized within the estimation sample, so a slope is per standard deviation of that sample. Party identification is not held fixed in this figure.Note: The resentment items were not asked in 1996, so that line joins 1992 and 2000; the affirmative action item was not asked in 1988. ANES 2020 has no mode weights, so it is missing from the face-to-face view, and the CES is an online opt-in panel in both years, so its intervals are model based. Intervals are HC3.Source: ANES Cumulative Data File, 1964-2024; weighted (VCF0009z; mode weights VCF0009x and VCF0009y for 2008, 2012, 2016 and 2024); Cooperative Election Study, 2008 and 2012; weighted (single-wave weight)

    The 2008 to 2012 comparison is also the change in ANES mode, so the face-to-face views give the cleaner reading. Among face-to-face respondents the slope was 22.1 (95% CI 19.2 to 25.0)22.1 percentage points of the Republican two-party presidential vote per standard deviation of the attitudeEstimandweighted linear probability slope on the standardized attitude (racial resentment index; party not held fixed)Interval95% CI 19.2 to 25.0n782Weightsweight_ftf (VCF0009x)Populationwhite non-Hispanic voters for a major-party candidatePeriod2008, face-to-face interviewsSourceANES Cumulative Data File in 2008 and 21.8 (95% CI 18.1 to 25.6)21.8 percentage points of the Republican two-party presidential vote per standard deviation of the attitudeEstimandweighted linear probability slope on the standardized attitude (racial resentment index; party not held fixed)Interval95% CI 18.1 to 25.6n609Weightsweight_ftf (VCF0009x)Populationwhite non-Hispanic voters for a major-party candidatePeriod2012, face-to-face interviewsSourceANES Cumulative Data File in 2012, a change of −0.3 (95% CI −5.0 to 4.5)−0.3 percentage points of the Republican two-party presidential vote per standard deviation of the attitudeEstimandchange in the weighted linear probability slope on the standardized attitude (party not held fixed)Interval95% CI −5.0 to 4.5n1,391Weightsmode weightsPopulationwhite non-Hispanic voters for a major-party candidatePeriod2008 to 2012, face-to-face interviewsSourceANES Cumulative Data File points, and the 2012 to 2016 change was +8.7 (95% CI 3.9 to 13.4)+8.7 percentage points of the Republican two-party presidential vote per standard deviation of the attitudeEstimandchange in the weighted linear probability slope on the standardized attitude (party not held fixed)Interval95% CI 3.9 to 13.4n1,132Weightsmode weightsPopulationwhite non-Hispanic voters for a major-party candidatePeriod2012 to 2016, face-to-face interviewsSourceANES Cumulative Data File. The CES is online in both years and has one racial policy item in the presidential years, opposition to affirmative action. Its slope was 25.5 (95% CI 24.8 to 26.2)25.5 percentage points of the Republican two-party presidential vote per standard deviation of opposition to affirmative actionEstimandweighted linear probability slope on the standardized attitude (party not held fixed)Interval95% CI 24.8 to 26.2n19,797Weightsweight (single wave)Populationwhite non-Hispanic voters for a major-party candidatePeriod2008SourceCooperative Election Study in 2008 and 24.9 (95% CI 24.2 to 25.6)24.9 percentage points of the Republican two-party presidential vote per standard deviation of opposition to affirmative actionEstimandweighted linear probability slope on the standardized attitude (party not held fixed)Interval95% CI 24.2 to 25.6n30,271Weightsweight (single wave)Populationwhite non-Hispanic voters for a major-party candidatePeriod2012SourceCooperative Election Study in 2012, a change of −0.6 (95% CI −1.6 to 0.4)−0.6 percentage points of the Republican two-party presidential vote per standard deviation of opposition to affirmative actionEstimandchange in the weighted linear probability slope (party not held fixed)Interval95% CI −1.6 to 0.4n50,068Weightsweight (single wave)Populationwhite non-Hispanic voters for a major-party candidatePeriod2008 to 2012SourceCooperative Election Study points. Tesler finds that racial attitudes became more closely tied to evaluations of politicians and parties and to opinion on policies attached to President Obama [4] [5]. The slope here is for the vote and not for evaluations or policy opinion, and it shows no rise distinguishable from zero between the 2008 and 2012 elections in person or in the CES.

    Holding party identification fixed shrinks the slope in every year, to 7.17.1 percentage points of the Republican two-party presidential vote per standard deviation of the attitudeEstimandweighted linear probability slope on the standardized attitude (racial resentment index; party held fixed)Interval95% CI 5.5 to 8.8n2,535Weightsweight (VCF0009z)Populationwhite non-Hispanic voters for a major-party candidatePeriod2012SourceANES Cumulative Data File in 2012 and 13.413.4 percentage points of the Republican two-party presidential vote per standard deviation of the attitudeEstimandweighted linear probability slope on the standardized attitude (racial resentment index; party held fixed)Interval95% CI 11.2 to 15.6n1,941Weightsweight (VCF0009z)Populationwhite non-Hispanic voters for a major-party candidatePeriod2016SourceANES Cumulative Data File in 2016 against 23.223.2 percentage points of the Republican two-party presidential vote per standard deviation of the attitudeEstimandweighted linear probability slope on the standardized attitude (racial resentment index; party not held fixed)Interval95% CI 21.3 to 25.0n2,541Weightsweight (VCF0009z)Populationwhite non-Hispanic voters for a major-party candidatePeriod2012SourceANES Cumulative Data File and 31.331.3 percentage points of the Republican two-party presidential vote per standard deviation of the attitudeEstimandweighted linear probability slope on the standardized attitude (racial resentment index; party not held fixed)Interval95% CI 29.9 to 32.8n1,942Weightsweight (VCF0009z)Populationwhite non-Hispanic voters for a major-party candidatePeriod2016SourceANES Cumulative Data File. The split between and within parties belongs to Chapter 5.2, Racial attitudes and ideology, which reports the same pattern for ideology, and this chapter does not carry it further. Party identification is measured in the same interview and can follow racial attitudes as well as precede them [6].

    Figure 9 gives the same relation in levels, with the resentment index cut into thirds inside each year so that the thirds are equal in size. The Republican share among the most resentful third was 79.4% (95% CI 73.9% to 84.9%)79.4% percent voting RepublicanEstimandweighted share who voted for the Republican candidateInterval95% CI 73.9% to 84.9%n323Weightsweight (VCF0009z)Populationwhite non-Hispanic two-party voters in the most resentful third of the resentment indexPeriod2008SourceANES Cumulative Data File in 2008, 78.7%78.7% percent voting RepublicanEstimandweighted share who voted for the Republican candidateInterval95% CI 75.3% to 82.1%n849Weightsweight (VCF0009z)Populationwhite non-Hispanic two-party voters in the most resentful third of the resentment indexPeriod2012SourceANES Cumulative Data File in 2012 and 89.1%89.1% percent voting RepublicanEstimandweighted share who voted for the Republican candidateInterval95% CI 86.3% to 91.9%n702Weightsweight (VCF0009z)Populationwhite non-Hispanic two-party voters in the most resentful third of the resentment indexPeriod2016SourceANES Cumulative Data File in 2016, and among the least resentful third 28.8% (95% CI 22.9% to 34.7%)28.8% percent voting RepublicanEstimandweighted share who voted for the Republican candidateInterval95% CI 22.9% to 34.7%n302Weightsweight (VCF0009z)Populationwhite non-Hispanic two-party voters in the least resentful third of the resentment indexPeriod2008SourceANES Cumulative Data File, 27.0%27.0% percent voting RepublicanEstimandweighted share who voted for the Republican candidateInterval95% CI 23.3% to 30.6%n962Weightsweight (VCF0009z)Populationwhite non-Hispanic two-party voters in the least resentful third of the resentment indexPeriod2012SourceANES Cumulative Data File and 17.1%17.1% percent voting RepublicanEstimandweighted share who voted for the Republican candidateInterval95% CI 13.7% to 20.5%n672Weightsweight (VCF0009z)Populationwhite non-Hispanic two-party voters in the least resentful third of the resentment indexPeriod2016SourceANES Cumulative Data File. The gap between the most and least resentful thirds changed by +1.2 (95% CI −8.3 to 10.7)+1.2 percentage pointsEstimandchange in the gap in the Republican share between the most and least resentful thirdsInterval95% CI −8.3 to 10.7Weightsweight (VCF0009z)Populationwhite non-Hispanic two-party votersPeriod2008 to 2012SourceANES Cumulative Data File points from 2008 to 2012 and +20.3 (95% CI 13.6 to 27.0)+20.3 percentage pointsEstimandchange in the gap in the Republican share between the most and least resentful thirdsInterval95% CI 13.6 to 27.0Weightsweight (VCF0009z)Populationwhite non-Hispanic two-party votersPeriod2012 to 2016SourceANES Cumulative Data File from 2012 to 2016. Fixed index cuts keep the score and not the rank fixed, so the groups change in size. With them the gap changed by +15.7 (95% CI 2.5 to 28.8)+15.7 percentage pointsEstimandchange in the gap in the Republican share between the groups above the upper and at or below the lower fixed index cutInterval95% CI 2.5 to 28.8Weightsweight (VCF0009z)Populationwhite non-Hispanic two-party votersPeriod2004 to 2008SourceANES Cumulative Data File points from 2004 to 2008, +0.5 (95% CI −9.2 to 10.3)+0.5 percentage pointsEstimandchange in the gap in the Republican share between the groups above the upper and at or below the lower fixed index cutInterval95% CI −9.2 to 10.3Weightsweight (VCF0009z)Populationwhite non-Hispanic two-party votersPeriod2008 to 2012SourceANES Cumulative Data File from 2008 to 2012 and +9.7 (95% CI 2.8 to 16.6)+9.7 percentage pointsEstimandchange in the gap in the Republican share between the groups above the upper and at or below the lower fixed index cutInterval95% CI 2.8 to 16.6Weightsweight (VCF0009z)Populationwhite non-Hispanic two-party votersPeriod2012 to 2016SourceANES Cumulative Data File from 2012 to 2016. The size of the 2012 to 2016 rise depends on the cut, and the 2008 to 2012 change is within sampling error under both. The Republican share among all white voters was 56.4%56.4% percent voting RepublicanEstimandweighted share who voted for the Republican candidateInterval95% CI 52.5% to 60.3%n783Weightsweight (VCF0009z)Populationwhite non-Hispanic two-party votersPeriod2008SourceANES Cumulative Data File in 2008, 57.1%57.1% percent voting RepublicanEstimandweighted share who voted for the Republican candidateInterval95% CI 54.7% to 59.5%n2,555Weightsweight (VCF0009z)Populationwhite non-Hispanic two-party votersPeriod2012SourceANES Cumulative Data File in 2012 and 57.4%57.4% percent voting RepublicanEstimandweighted share who voted for the Republican candidateInterval95% CI 54.8% to 60.0%n1,957Weightsweight (VCF0009z)Populationwhite non-Hispanic two-party votersPeriod2016SourceANES Cumulative Data File in 2016, so the widening gap came with no comparable change in the white vote as a whole. Sides, Tesler and Vavreck describe changes in the party coalitions along racial and ethnic identities that preceded the 2016 election and were reinforced by the campaign [7].

    Figure 9

    The Republican share among the most resentful third of white voters rose from 79.4% in 2008 to 89.1% in 2016, and fell among the least resentful third from 28.8% to 17.1%

    Republican share of the two-party presidential vote among white non-Hispanic voters in the lowest, middle and highest group of the racial resentment index (thirds inside each election year, or fixed index cuts), percent, ANES 1988 to 2024, 95% intervals

    Republican share of the two-party vote (%)

    Show data
    Data for The Republican share among the most resentful third of white voters rose from 79.4% in 2008 to 89.1% in 2016, and fell among the least resentful third from 28.8% to 17.1%
    conf_highconf_lowestimategroupgroupingnyear
    53.2042.2047.70Least resentful thirdexact3221988
    37.1027.6032.30Least resentful thirdexact3851992
    44.6032.7038.70Least resentful thirdexact3182000
    43.202936.10Least resentful thirdexact2472004
    34.7022.9028.80Least resentful thirdexact3022008
    30.6023.3027Least resentful thirdexact9622012
    20.5013.7017.10Least resentful thirdexact6722016
    9.3067.60Least resentful thirdexact1,9272020
    11.706.609.10Least resentful thirdexact1,1072024
    69.1058.6063.80Middle thirdexact4001988
    56.4046.5051.40Middle thirdexact5051992
    62.2050.1056.20Middle thirdexact4182000
    71.6057.8064.70Middle thirdexact2272004
    67.605561.30Middle thirdexact3042008
    69.5061.9065.70Middle thirdexact1,1262012
    7061.8065.90Middle thirdexact8212016
    66.7060.3063.50Middle thirdexact1,6682020
    63.905559.50Middle thirdexact1,1892024
    76.2066.3071.20Most resentful thirdexact4261988
    64.4054.3059.40Most resentful thirdexact3651992
    73.7062.4068.10Most resentful thirdexact3312000
    80.7067.8074.30Most resentful thirdexact2452004
    84.9073.9079.40Most resentful thirdexact3232008
    82.1075.3078.70Most resentful thirdexact8492012
    91.9086.3089.10Most resentful thirdexact7022016
    928890Most resentful thirdexact1,4752020
    91.1085.4088.20Most resentful thirdexact1,0282024
    53.3042.4047.80Index of 50 or lessfixed3221988
    38.502933.70Index of 50 or lessfixed3851992
    44.7032.7038.70Index of 50 or lessfixed3182000
    43.1028.8036Index of 50 or lessfixed2062004
    31.2018.7024.90Index of 50 or lessfixed2282008
    2719.4023.20Index of 50 or lessfixed7802012
    28.6021.6025.10Index of 50 or lessfixed8382016
    25.9021.4023.70Index of 50 or lessfixed2,5302020
    28.6022.3025.50Index of 50 or lessfixed1,5612024
    69.5060.2064.90Index above 50 to 75fixed4071988
    5646.3051.20Index above 50 to 75fixed4111992
    635157Index above 50 to 75fixed3162000
    7258.4065.20Index above 50 to 75fixed2272004
    67.2054.8061Index above 50 to 75fixed3042008
    69.1061.3065.20Index above 50 to 75fixed9122012
    76.9068.1072.50Index above 50 to 75fixed5552016
    82.4076.2079.30Index above 50 to 75fixed1,1262020
    8071.1075.50Index above 50 to 75fixed7822024
    78.5066.8072.60Index above 75fixed2231988
    68.905763Index above 75fixed2561992
    7562.6068.80Index above 75fixed2532000
    82.1067.7074.90Index above 75fixed1682004
    85.107479.50Index above 75fixed2502008
    81.8074.9078.30Index above 75fixed8492012
    9386.9090Index above 75fixed5492016
    94.5090.1092.30Index above 75fixed9182020
    93.8087.2090.50Index above 75fixed6252024
    The lines fan out after 2008: the line for the most resentful third rises, the line for the least resentful third falls, and the gap between them widens most from 2012 to 2016. Each point is the weighted Republican share of two-party voters in a group defined by rank on the index inside each election year, with ties split so that the groups are equal in size; the shaded years are the Obama presidency.Note: The resentment index was not asked in 1996, so no group is drawn for that year. In the default view the groups are exact weighted thirds inside each year, with respondents who share an index score split in proportion across the cut. In the fixed-cut view the cuts are at the same index scores in every year (50 and 75, the pooled terciles of white two-party voters), so the groups change in size from year to year. All interview modes are pooled; the mode comparison is in the previous figure. Intervals assume a simple random sample of the weighted respondents (the Kish effective sample size); they do not use the survey's sampling design.Source: ANES Cumulative Data File, 1964-2024; weighted (VCF0009z; mode weights VCF0009x and VCF0009y for 2008, 2012, 2016 and 2024)

    What these data cannot show

    The GSS, ANES and CES interview new respondents in each wave, so a change in a share mixes people who changed their answers, people who changed party and the replacement of older adults by younger ones, and this chapter cannot separate them. The step dates describe when a new level first appears in surveys fielded in alternate years. The step and the kink are assumptions about the form of change and give different starts, the 2016 survey for the step and the 2014 survey or earlier for the kink. The dates do not say why the level changed, and the 2016 GSS was fielded during the presidential campaign.

    The items measure answers to questions and not behavior. The principle items reached ceilings that leave little room to move, three of them have no readings after 2010, and the data cannot tell a change in belief from a change in willingness to say what the question invites. The GSS wording of the group changed over the period from Negroes to Blacks to African Americans, and in 2022 and 2024 the effective samples for party and race cells are small enough that the chapter pools those years. In the in-person view the Black respondents' cells for the discrimination and will-power items in 2022 and 2024 are suppressed, and Hispanic respondents cannot be separated in the GSS before 2000. The post-2018 levels depend on interview mode on several items, so the later comparisons are read by mode and not as changes over time. Results for Hispanic, Asian and other voters are not shown, and Chapter 3.4, Hispanic, Black and Asian voters (planned) is where voters of those groups belong.

    The vote section describes association across elections. It cannot separate the Obama presidency from the 2012 re-election campaign, the change in candidates in 2016, or the changes in the parties' coalitions, so a rise from 2012 to 2016 is not an effect of either. The vote is self-reported after the election and leans slightly Democratic, the attitude and party identification are measured in the same interview, the resentment index has no 1996 reading, ANES 2020 has no mode weights, and the CES holds one racial policy item in 2008 and 2012. The chapter does not test whether racial attitudes changed how people voted, and the surveys cannot say whether the movement after 2014 comes from the events of those years. Policing items and the weeks after May 2020 are in Chapter 5.4, Policing (planned) and Chapter 5.3, Racial attitudes after George Floyd's death (planned).

    Data sources

    • General Social SurveyCumulative data 1972-2024 (race, party and the race items; interview mode from the raw file)Years: 1972-2024 (35 study years) · Weights: weight_use · License: See the provider's terms of use.
    • ANES Cumulative Data FileCumulative Data File 1964-2024: thermometers, resentment items, aid to Black Americans, affirmative action, party and presidential vote; mode weights for 2008, 2012, 2016 and 2024Years: 1948-2024 (33 study years; even years only after 2004; no 2006, 2010 or 2014 study) · Weights: VCF0009z; VCF0009x and VCF0009y for the mode views · License: Free download after accepting the ANES terms of use; cite the release.
    • Cooperative Election StudyCumulative Common Content 2008 and 2012: affirmative action item, party and presidential voteYears: 2006-2024 (every year; self-reported turnout is available in the 10 even years) · Weights: weight (single wave) · License: See the provider's terms of use.

    All data sources

    Methods

    Sample, GSS. White and Black respondents in the GSS cumulative file, 1972 to 2024, read from the raw file because interview mode is only there (59,937 white respondents in all). 'White' and 'Black' are the GSS race codes; Hispanic origin cannot be separated before 2000, so white includes Hispanic whites (11.4% of white respondents from 2012 to 2024 report Hispanic origin). Party is partyid 0 to 2 (Democrats) and 4 to 6 (Republicans), leaners included, and pure independents are left out of party splits. Interview mode is in person (mode in person from 2004, and every survey before 2004 by design), telephone, web or other; the in-person views use in-person interviews only and leave out telephone interviews, and 2021 had no in-person interviews. Items, GSS (answer counted, years asked, then the label). racmar (No, 1972 to 2002): No law against Black-white marriage; racpres (Yes, 1972 to 2010): Would vote for a qualified Black nominee; racseg (Disagree slightly or strongly, 1972 to 1996): Whites have no right to exclude Black neighbors; racopen (The law that bars refusing to sell because of race, 1973 to 2024): Would vote for a law barring housing discrimination; marblk (Favor, strongly favor or neither favor nor oppose, 1990 to 2024): Would not oppose a relative marrying a Black person; racdif1 (Yes, 1977 to 2024): Discrimination; racdif2 (Yes, 1977 to 2024): Inborn ability; racdif3 (Yes, 1977 to 2024): Lack of chance for education; racdif4 (Yes, 1977 to 2024): Lack of motivation or will power; helpblk (1 or 2 on a five-point scale from government should help (1) to no special treatment (5), 1975 to 2024): Places self at the help end of the government-help scale (1 or 2 of 5); natrace (Too little, 1973 to 2024): Too little is spent on improving conditions of Black people; affrmact (Strongly favors or favors, 1994 to 2024): Favors preference in hiring and promoting Black people; wrkwayup (Disagree somewhat or strongly, 1994 to 2024): Rejects that Black people should work up without special favors. The denominator of each share is the respondents with a substantive answer. The value labels of every item were read from the raw file and asserted before recoding. The government-help item (helpblk) is a five-point scale with only the end points and the midpoint labeled; the series counts codes 1 and 2, the two answers nearest to government should help (code 1 alone is a smaller share). Sample, ANES. White non-Hispanic respondents (VCF0106) in the ANES Cumulative Data File. Affect: the feeling thermometers toward Black Americans (VCF0206) and white Americans (VCF0207), 1964 to 2024. Vote: the two-party presidential vote (VCF0704, major-party candidates only; 17,265 white non-Hispanic voters in the presidential years 1988 to 2024, of whom 16,299 have a racial resentment index score in the nine presidential years that asked it). Attitudes: the racial resentment index (the mean of four 0 to 1 items, VCF9039 to VCF9042, for respondents who answered at least two, times 100; the book's one definition, not asked in 1996), a colder rating of Black Americans (100 minus the thermometer), aid to Black Americans (VCF0830, 7 is the self-help end) and opposition to affirmative action (VCF0867a: raw codes 1, 2, 4 and 5 recoded to 1 to 4, 4 is strongly oppose). Party identification is the seven-point scale (pid_7pt), leaners included in Democrats (1 to 3) and Republicans (5 to 7). Sample, CES. White non-Hispanic respondents in the 2008 and 2012 surveys of the cumulative file who reported a major-party presidential vote, with opposition to affirmative action on the four-point item (1 strongly support, 4 strongly oppose) and the seven-point party scale. The single-wave weight is used; a post-election weight exists for 2012 only and gives a slope of the same size. The party-gap sign of the item was checked in every year it exists (Republicans more opposed in all seven). The CES is an opt-in panel, so its intervals are model based. Weights. GSS weight_use (wtssnrps from 2004, wtssps before); ANES VCF0009z, with weight_ftf and weight_web (VCF0009x and VCF0009y) for the mode views of 2008, 2012, 2016 and 2024, and no mode weights for 2020; CES weight. In a pooled period each year's weights are rescaled to average one over the respondents with an answer. Estimators and intervals. A share or mean is a weighted mean of the 0/1 or rating with a 95% normal interval on the Kish effective sample size; a change between two periods or elections is the difference of two independent estimates with the standard errors combined in quadrature, and so is the Democratic minus Republican difference in a change. These intervals treat the weighted sample as a simple random sample and ignore the strata and clusters of the designs. A vote slope is the coefficient on the standardized attitude (weighted mean 0 and SD 1 within the year and sample) in a weighted linear probability model of a Republican vote against a Democratic vote, with HC3 errors; the one 'party held fixed' sentence adds a dummy for each of the seven party-identification categories. Groups of the resentment index: in the default view exact weighted thirds inside each year, with respondents who share an index score split in proportion across a cut, so the three thirds hold the same weighted number of voters; in the fixed-cut view the groups are an index of 50 or less, above 50 to 75 and above 75 in every year (50 and 75 are the pooled weighted terciles of white two-party voters, 1988 to 2024), so the group sizes change by year. A gap between groups combines the two Kish standard errors. Step dates and form check. For one group and item, the yearly weighted shares of the in-person surveys of 2000, 2002, 2004, 2006, 2008, 2010, 2012, 2014, 2016 and 2018 are fitted by two constants, one before the first new survey s and one from s on, with s no earlier than 2004 and at least one survey after it. The best s minimizes the inverse-variance weighted squared error. The step at the fixed survey of 2016 is also reported, because the best date differs by item for white Republicans. As a check on the form, a trend that is constant before a kink and linear after it is fitted to the same shares (kink between 2004.0 and 2015.95 in steps of 0.05, weighted least squares); its weighted squared error is compared with the step's. The bootstrap (1,000 replicates, seed 12345) resamples respondents with replacement within each survey year, recomputes the yearly shares and variances and refits all three; the table reports the share of replicates that choose each date and the 2.5th and 97.5th percentiles of the step size, the step at 2016 and the kink year. Suppression. A yearly series cell needs at least 150 respondents and an effective sample of 100 (timeline floor); a party, race, mode, period or pooled cell needs at least 100 respondents and an effective sample of 100 (subgroup floor); a vote slope needs 100 respondents and an effective sample of 100. The GSS 2022 and 2024 cells alone are often under the effective-sample floor, so the mode views pool the two years and plot the pooled cell at 2023. Rounding: every displayed number and every plotted or tabled value is rounded half up from the unrounded estimate by b3.fmt, and a printed change can differ by 0.1 from the difference of two printed shares. Survey mode. GSS: in person through 2018, no 2020 survey, 2021 mostly web (no in-person interviews), 2022 and 2024 in person and web; every estimate after 2020 is shown in the all-interview view and in an in-person or web view or in the mode figure. ANES: in person through 2008, web added in 2012; the mode views use the mode weights. CES: online in every year.

    References

    1. Kish, Leslie (1965). Survey Sampling. John Wiley and Sons.
    2. Schuman, Howard; Steeh, Charlotte; Bobo, Lawrence; Krysan, Maria (1997). Racial Attitudes in America: Trends and Interpretations, revised edition. Harvard University Press.
    3. Kluegel, James R. (1990). Trends in Whites' Explanations of the Black-White Gap in Socioeconomic Status, 1977-1989. American Sociological Review 55(4): 512-525. doi:10.2307/2095804
    4. Tesler, Michael (2012). The Spillover of Racialization into Health Care: How President Obama Polarized Public Opinion by Racial Attitudes and Race. American Journal of Political Science 56(3): 690-704. doi:10.1111/j.1540-5907.2011.00577.x
    5. Tesler, Michael (2016). Post-Racial or Most-Racial? Race and Politics in the Obama Era. University of Chicago Press. press.uchicago.edu/ucp/books/book/chicago/P/bo22961444.html
    6. Engelhardt, Andrew M. (2021). Racial Attitudes Through a Partisan Lens. British Journal of Political Science 51(3): 1062-1079. doi:10.1017/S0007123419000437
    7. Sides, John; Tesler, Michael; Vavreck, Lynn (2018). Identity Crisis: The 2016 Presidential Campaign and the Battle for the Meaning of America. Princeton University Press. press.princeton.edu/books/hardcover/9780691174198/identity-crisis

    How to cite

    @misc{singh2026fiftyyearsofracialattitudes,
  author = {Singh, Vishal},
  title = {Fifty years of racial attitudes},
  year = {2026},
  url = {https://vishalsingh.org/politics/fifty-years-of-racial-attitudes},
  note = {In The Public Mind. Published 3 October 2026.}
}

    Reading paths