--- title: Race description: >- GSS trends on race and civil rights since the 1970s — open housing, segregation, interracial marriage, affirmative action, explanations of inequality — in light of Obama’s 2016 “you would choose right now” claim. --- # Race September 7, 1978 was my first day of sixth grade. I left my house in West Roxbury, which is a neighborhood of Boston, walked a few blocks to the bus stop, and got on the bus. We drove east into another neighborhood, Roslindale, and past the Washington Irving Middle School, where most kids in West Roxbury went to school. But we did not stop there. Instead we continued east into another neighborhood, Mattapan, to the Solomon Lewenberg Middle School. You might wonder why I didn't go to the closest school to my house, which was less than two miles away -- or even the second closest, the William B. Rogers Middle School, which was three miles away. Why was I bused to the Lewenberg school, more than four miles away? To understand that, you just have to understand the entire history of Boston. In 1974, a federal judge ruled that the Boston Public Schools had deliberately maintained racial segregation. Although Massachusetts had no laws requiring separate schools for Black and white students, the court found that the School Committee had intentionally drawn attendance zones, located new schools, and managed transfers in ways that kept many schools racially identifiable. Because the segregation was intentional — and not just the result of segregated housing patterns — it violated the Equal Protection Clause of the Fourteenth Amendment. To remedy the violation, Judge W. Arthur Garrity Jr. ordered a desegregation plan that reassigned thousands of students to different schools. For many children, that meant riding a bus across the city to attend a school in a different neighborhood. The first year of the plan was one of the ugliest episodes in Boston's history. In 1974, the first day of school became front-page news around the country. Hundreds of protesters gathered outside South Boston High School, where buses carrying Black students from Roxbury were met by jeering crowds. Police escorted students through lines of demonstrators who threw rocks, bottles, and eggs. Several people were injured, and many protestors were arrested. Throughout the fall, South Boston High and Hyde Park High were flashpoints of conflict. There were fights inside the schools, attacks on buses, and repeated confrontations between protesters and police. Anti-busing rallies drew thousands of participants, while supporters of desegregation organized counter-protests. Politicians, clergy, civil rights organizations, and neighborhood groups argued bitterly over the court order. Television viewers across the country watched images of buses under police escort, children running through hostile crowds, and officers in riot gear struggling to restore order. The conflict continued into the following school year. In April 1976, a newspaper photographer captured a White teenager assaulting Black attorney and civil rights activist Ted Landsmark with the staff of an American flag. The photograph appeared in newspapers nationwide and became the enduring symbol of the busing crisis. But the violence gradually subsided. When I got on that bus in 1978, the desegregation plan was starting its fifth year, and for me the academic year was uneventful. In 1978, West Roxbury's population was predominantly White, while Mattapan's was becoming predominantly Black. ## Segregation ### School segregation In 1972, the GSS started asking: > In general, do you favor or oppose the busing of Black and White school children from one school district to another?[^race-terms] [^race-terms]: In survey questions quoted in this book, I capitalize Black and White, which the codebooks usually do not. When the wording of a question changed over time, I quote the most recent version. Earlier versions of many of these questions said "Negroes" or "Blacks" where later versions say "African-Americans." The following figure shows the percentage in favor, by race of respondent. ```{figure} figs/gss/busing/composed_race_eth4_White_Black_Hispanic_time_series_period_effect.png :name: busing-by-race-time-series :align: left `busing` — Time series by race: percent in favor of busing for school integration. Hispanic is based on primary reported ancestry, for a consistent measure since 1972; White and Black exclude respondents of Hispanic ancestry. ``` Overall, support for busing rose from 19% in 1972 to 37% in 1996. Among White respondents it rose from 13% to 32%, and among Hispanic respondents from 35% to 58%. Among Black respondents, a majority supported busing throughout, rising only from 54% to 60%. ```{figure} figs/gss/busing/composed_race_eth4_White_Black_Hispanic_cohort_effect.png :name: busing-by-race-cohort :align: left `busing` — Cohort component by race: percent in favor of busing for school integration. ``` Among White respondents, the cohort component is about 19% for people born in 1900, dips slightly for people born in the 1920s, and then rises steeply, to 50% for people born in 1978. Among Hispanic respondents it rises from 36% to 60%, with more uncertainty; among Black respondents it is nearly flat, between 56% and 61%. Of the increase among White respondents, we can estimate that about 40% is due to generational replacement and about 60% to changing minds. Among Black respondents the change is too small, and among Hispanic respondents the estimates too uncertain, to divide it between the two. TODO: What should we do with the busing question from ANES? From 1972 to 1996, the GSS asked non-Black respondents: > Would you yourself have any objection to sending your children to a school where [a few / half / more than half] of the children are African-Americans? Black respondents were asked about schools where the children are White. The following figure shows the percentage of non-Black respondents who said they would object. ```{figure} figs/_composed/composed_gss_school_ladder_time_series_period_effect.png :name: school-ladder-time-series :align: left `racfew`, `rachaf`, `racmost` — Time series: percent of non-Black respondents who would object, by share of Black children in the school (GSS, lines). Hollow circles: White parents with children in school who would object, Gallup (1958–1990, "a few") and Washington Post–Ipsos (2024, all three). ``` Between 1972 and 1996, the percentage who would object to a school where a few of the children are Black fell from 7% to 3%; for a school where half are Black, from 26% to 18%; and for a school where more than half are Black, from 58% to 49%. In addition to the GSS data, we also have data from similar questions in in other surveys: by Gallup, 13 times between 1958 and 1990, and by the [Washington Post and Ipsos](https://www.washingtonpost.com/education/2024/05/17/brown-vs-board-70-anniversary-poll/) in 2024, for the 70th anniversary of *Brown v. Board of Education* ([results](https://www.ipsos.com/sites/default/files/ct/news/documents/2024-06/Washington%20Post%20Ipsos%20Survey%20of%20Black%20Americans%20Trended%20Topline.pdf), including Gallup's). The dotted lines connect the last GSS estimate, in 1996, to 2024. These surveys are not directly comparable to the GSS. They asked a different group of people, and they asked in different ways — Gallup in person and later by phone, the Post and Ipsos online — and people tend to give more socially acceptable answers to an interviewer than to a screen. In 2024, "a few" and "half" were both 9%; the two circles are separated slightly so both are visible. ```{figure} figs/_composed/composed_gss_school_ladder_cohort_effect.png :name: school-ladder-cohort :align: left `racfew`, `rachaf`, `racmost` — Cohort component: percent of non-Black respondents who would object, by share of Black children in the school. ``` Later generations were less likely to object. For a school where a few of the children are Black, the cohort component falls from 10% for people born in 1900 to 4% for people born in 1978, leveling off after about 1950; for half, it falls steadily, from 33% to 14%. For more than half, it stays near 60% for people born before 1940 and then falls, to 40% for people born in 1978. We can estimate that about 36% of the decline in objections to "a few" is due to generational replacement and about 64% to changing minds. For "half" and "more than half," generational replacement accounts for more of the decline — roughly 62% and 57% — although these estimates are much less certain. ### Open housing Housing segregation can be the result of individual preferences. In *The Big Sort*, Bill Bishop argues that growing segregation by political beliefs is the large-scale result of many small-scale decisions. If people choose communities that match their lifestyle preferences—urban or rural, secular or religious, family-oriented or cosmopolitan—and those preferences are correlated with political beliefs, the results can be politically segregated communities. But in the case of racial segregation, individual preference is not the whole story, not by a long way. Even if we start from the 20th Century, it's a long story. Before 1917, many American cities had laws that prohibited Black and White families from living on the same blocks. These laws were struck down by the Supreme Court in *Buchanan v. Warley*, which held (unanimously!) that the laws violated the Fourteenth Amendment because they infringed on homeowners' property rights. But the decision did not prohibit discrimination by homeowners themselves. After 1917, house-sellers often added to the deed of sale a "restrictive covenant", which is a contractual requirement that the house could only be occupied by White residents. A [typical covenant](https://www.usccr.gov/files/historical/1973/73-004.pdf) included language like: > "... hereafter no part of said property or any portion thereof shall be ... occupied by any person not of the Caucasian race, it being intended hereby to restrict the use of said property ... against the occupancy as owners or tenants of any portion of said property for resident or other purpose by people of the Negro or Mongolian race." Most of these "restrictive covenants" were self-perpetuating -- that is, they applied not just to the original owner of a new house, but to all future owners. In 1926, these covenants were challenged before the Supreme Court, in *Corrigan v. Buckley*, on the grounds that they were prohibited by the Fifth Amendment, because they deprived Black purchasers and willing sellers of property rights without due process; by the Thirteenth Amendment, because they were a "badge or incident of slavery"; and by the Fourteenth Amendment, because they denied Black citizens the equal protection of the law. The Court rejected these arguments unanimously, holding that these constitutional protections restrict only *government* action, not private individuals entering into private contracts. After that, restrictive covenants became more common. In rapidly growing cities such as Chicago and Los Angeles, historians estimate that by the 1940s these covenants covered more than 80% of residential property. *(See [`race_plan.md`](race_plan.md) — Color of Law update.)* Then in 1948, covenants were challenged again, in *Shelley v. Kraemer*. This time, the plaintiffs took a different approach: People are free to make private discriminatory agreements, they argued, but when a state court enforces such an agreement, that enforcement is a state action. The Court agreed, and held that state enforcement of a discriminatory contract is prohibited by the Fourteenth Amendment. The decision was unanimous, again, but three justices did not participate in the decision. They didn't explain why, but the historical consensus is that they recused themselves because their own homes were subject to racially restrictive covenants. Let that sink in. The decisions in *Corrigan* and *Shelley* were almost directly contradictory, but they were both unanimous. So we might wonder what changed between 1926 and 1948. One thing that changed was the composition of the Court: Franklin Roosevelt appointed eight justices to the Court between 1937 and 1943; seven of them were still sitting in 1948. Another thing that happened was World War II: having fought to oppose Nazi racism, people might have found it harder to justify racial discrimination at home. Black veterans, in particular, came back from war with ___ to accept racial injustice and renewed energy to oppose it. Also, *Shelley* was not an isolated case. It was part of a long-term strategy developed by the National Association for the Advancement of Colored People (NAACP) under the leadership of Charles Hamilton Houston. The attorneys who argued the case were Loren Miller and Thurgood Marshall (in 1967 Marshall became the first African American appointed to the Supreme Court). Together, *Buchanan* and *Shelley* prohibited segregation by law and state enforcement of discriminatory contracts. But they did nothing to prevent discrimination by homeowners. It was still legal for a homeowner to refuse to sell to a Black family -- or to discriminate on any other basis. After *Shelley*, some states enacted laws that prohibited this kind of discrimination. Massachusetts was one of the first; in 1957, it prohibited racial discrimination in publicly assisted housing. Other states followed, but most were in the Northeast, Midwest, and West -- not the South. Finally, the federal government stepped in. On April 11, 1968, President Lyndon Johnson signed the Fair Housing Act, making it illegal to refuse to sell, rent, or finance housing because of race, color, religion, or national origin. Congress had debated the law for years, but the assassination of Martin Luther King Jr. on April 4, 1968 -- and the violent unrest that followed -- spurred them to act. In theory, the Fair Housing Act prohibits housing discrimination, but without knowing the inner thoughts of a homeowner, it was hard to enforce. The law is one thing, and public opinion is another. ### Fair Housing and Public Opinion Five years after the enactment of Fair Housing Act, the General Social Survey (GSS) asked this question: *Suppose there is a community-wide vote on the general housing issue. There are two possible laws to vote on: One law says that a homeowner can decide whom to sell his house to, even if that homeowner prefers not to sell to someone because of their race or color. The second law says that a homeowner cannot refuse to sell to someone because of their race or color. Which law would you vote for?* The following figure shows the percentage of respondents who chose the second law, which prohibits a homeowner from discriminating. ```{figure} figs/gss/racopen/time_model.png :name: racopen-time-model :align: left Time series: percent voting for the open housing law. ``` In 1973, only about 33% chose the second law. Among the majority that chose the first, presumably some were unaware that it would conflict with federal law, and some did not care. Either way, it is apparent that the Fair Housing Act was ahead of public opinion. But not for long -- support for open housing passed 50% in 1984, and by 2024 it was 82%. TODO: Decide how much of this text to repeat, and whether to include cohort trajectories When public opinion changes as fast as that, there are generally two kinds of explanation: * Generational replacement: as older people with the declining viewpoint age out of the survey, they are replaced by young adults with the ascending viewpoint. * Changing minds: people adopt different views over the course of their lives, possibly in response to events. With a repeated survey like the GSS, we can follow each generation to see how it changes over time. The following figure shows one line for each year of birth, estimating support for open housing over time. ```{figure} figs/gss/racopen/model10_cohort_trajectories.png :name: racopen-cohort-trajectories :align: left Cohort trajectories: percent voting for the open housing law, one line per birth year. ``` In the bottom left we can see that when people born in the 1900s and 1910s were interviewed in the 1970s, only about 20% of them chose the open housing law -- almost 80% said homeowners should be allowed to discriminate based on race. In the top right we can see that when people born in the 2000s were interviewed in 2024, more than 80% of them chose open housing. So there are big differences between generations. Also, following the cohorts over time, we can see that the trend is upward — that is, toward support for open housing laws. From this analysis we can decompose the period effect (holding the mixture of cohorts constant) and the cohort effect (holding the mixture of survey years constant). Here's the estimated cohort effect: ```{figure} figs/gss/racopen/model10_cohort_standardized.png :name: racopen-cohort-standardized :align: left Standardized cohort component, percent voting for the open housing law. ``` Looking at differences between cohorts, the biggest changes happened between people born in 1900 and 1960: the cohort component rises from 37% to about 70%. Since then, the cohort effect is essentially flat. And here's the period effect after factoring out the cohort effect. ```{figure} figs/gss/racopen/model10_period_standardized.png :name: racopen-period-standardized :align: left Standardized period component, percent voting for the open housing law. ``` The period component rises from 42% in 1974 to 75% in 2024, with a pause between the mid-1990s and the mid-2000s. Putting these pieces together, we can estimate that about 32% of the change we see from 1974 to 2024 can be attributed to generational replacement, and about 68% to changing minds. ## Support for segregation has declined The following figure shows six questions from the GSS and the American National Election Studies (ANES), each in the direction that supports segregation or exclusion: 1. GSS, 1972–2002: Do you agree or disagree: "African-Americans shouldn't push themselves where they're not wanted." (percent agreeing strongly or slightly) 2. GSS, 1973–2024: "Suppose there is a community-wide vote on the general housing issue. There are two possible laws to vote on: One law says that a homeowner can decide whom to sell his house to, even if that homeowner prefers not to sell to someone because of their race or color. The second law says that a homeowner cannot refuse to sell to someone because of their race or color. Which law would you vote for?" (percent choosing the first law -- the other side of the open housing question above) 3. GSS, 1990–2024: Would you be in favor or opposed to "living in a neighborhood where half of your neighbors were Blacks?" (percent somewhat or very much opposed). 4. GSS, 1972–1996: Do you agree or disagree: "White people have a right to keep African-Americans out of their neighborhoods if they want to, and African-Americans should respect that right." (percent agreeing strongly or slightly) 5. ANES, 1964–1976: "Which of these statements would you agree with: White people have a right to keep Black people out of their neighborhoods if they want to, or Black people have a right to live wherever they can afford to?" (percent choosing the first). 6. ANES, 1964–1978: "Are you in favor of desegregation, strict segregation, or something in between?" (percent choosing strict segregation). The circles show the same question in the [Southern Focus Poll](https://www.thearda.com/data-archive/browse-alphabetically), 1992–1999, which asked about "integration of the races" rather than "desegregation." The following figure shows levels of support for segregation. ```{figure} figs/_composed/composed_race_segregation_time_series_period_effect.png :name: race-segregation-time-series :align: left `racpush`, `racopen`, `liveblks`, `racseg`, `VCF0819`, `VCF0815` — Time series: percent supporting segregation or exclusion, numbered as in the text. Circles: Southern Focus Poll, 1992–1999. ``` All six questions show decreasing support for segregation. Agreement that Black people shouldn't push where they're not wanted (1) fell from 75% in 1972 to 36% in 2002, when the question was dropped. Support for letting homeowners discriminate (2) fell from 66% in 1973 to 16% in 2024. Opposition to living in a neighborhood that is half Black (3) fell from 41% in 1990 to 12% in 2024. Agreement that White people have a right to keep Black people out of their neighborhoods fell from 41% in 1972 to 12% in 1996 in the GSS (4), and from 35% in 1964 to 10% in 1976 in the ANES (5). Support for strict segregation (6) fell from 23% in 1964 to 7% in 1978, and in the Southern Focus Poll it stayed between 2% and 4% through the 1990s. The following figure shows the cohort component of these changes. ```{figure} figs/_composed/composed_race_segregation_cohort_effect.png :name: race-segregation-cohort :align: left `racpush`, `racopen`, `liveblks`, `racseg`, `VCF0819`, `VCF0815` — Cohort component: percent supporting segregation or exclusion, numbered as in the text. ``` For all six, support for segregation is lower in later generations. The cohort component falls from 80% for people born in 1900 to 35% for people born in 1984 (1); from 58% to about 28% for people born around 1960, and then stays about flat (2); from 43% for people born in 1901 to 13% for people born in 2006, declining steadily (3); from 46% to 11% for people born in 1978 (4); from 33% to 12% for people born in 1958 (5); and from 19% to 6% for people born in 1960 (6). For each question, we can estimate that between 14% and 38% of the decline is due to generational replacement, and the rest to changing minds. ## The experience of segregation lingers ### Segregated lives The questions so far ask what people think should be. The GSS and ANES also asked people about their own lives, in two groups of questions. The first group measures geographical segregation -- whether Black and White Americans live, study, and work in the same places. The second measures social distance -- whether they mix in churches, clubs, friendships, and families. The following figure shows six questions from the first group, each in the direction that indicates segregation: 1. ANES, 1964–1976: "Is this neighborhood you live in all White, mostly White, about half and half, mostly Black, or all Black?" (percent all White) 2. GSS, 1972–1993: "Are there any Blacks living in this neighborhood now? … Are there any African-American families living close to you?" (percent with no Black family close by, including people with no Black neighbors at all) 3. ANES, 1964–1976: "Is the grade school nearest you all White, mostly White, about half and half, mostly Black, or all Black?" (percent all White) 4. ANES, 1964–1972: "Is the high school nearest you all White, mostly White, about half and half, mostly Black, or all Black?" (percent all White) 5. GSS, 1972–2024: "Are there any Blacks living in this neighborhood now?" (percent saying no) 6. GSS, 1990–2022: "Are the people who work where you work all White, mostly White, about half and half, mostly Black, or all Black?" (percent all White, among people who work) The first figure shows non-Black respondents. The GSS asked Black respondents about White neighbors, so their answers are shown in a second figure, with the same numbers. ```{figure} figs/_composed/composed_race_lives_segregated_time_series_period_effect.png :name: race-lives-segregated-time-series :align: left `VCF0860`, `racclos_all`, `VCF0861`, `VCF0863`, `raclive`, `racwork` — Time series: percent of non-Black respondents in segregated settings, numbered as in the text. ``` All six declined. The percentage living in an all-White neighborhood (1) fell from 80% in 1964 to 61% in 1976, and the percentage with no Black family living close by (2) fell from 75% in 1972 to 58% in 1993. The percentage whose nearest grade school was all White (3) fell from 56% in 1964 to 35% in 1976, and for the nearest high school (4), from 43% in 1964 to 27% in 1972. In the longest series, the percentage with no Black people in their neighborhood (5) fell from 64% in 1972 to 22% in 2024. And the percentage of workers whose workplace was all White (6) fell from 37% in 1990 to 14% in 2022. For Black respondents, the same questions indicate segregation in the other direction: an all-Black neighborhood, school or workplace (1, 3, 4 and 6), or no White family close by and no White people in the neighborhood (2 and 5). The GSS neighborhood questions for Black respondents start in 1978. ```{figure} figs/_composed/composed_race_lives_segregated_black_time_series_period_effect.png :name: race-lives-segregated-black-time-series :align: left `VCF0860`, `racclos_all`, `VCF0861`, `VCF0863`, `raclive`, `racwork` — Time series: percent of Black respondents in segregated settings, numbered as in the text. ``` The percentage of Black respondents living in an all-Black neighborhood (1) was about the same in 1964 and 1976, 28% and 29%, and the percentage with no White family close by (2) was 35% in both 1978 and 1993. But the percentage whose nearest grade school was all Black (3) fell from 36% in 1964 to 12% in 1976, and for the nearest high school (4), from 33% in 1964 to 13% in 1972. The percentage with no White people in their neighborhood (5) fell from 25% in 1978 to 7% in 2024, and the percentage of workers whose workplace was all Black (6) fell from 12% in 1990 to 5% in 2022. ### Personal relationships The second group asks about social segregation: some are about personal and policy preferences; others are about experiences. 1. GSS, 1973–2006: "During the last few years, has anyone in your family brought a friend who was an African-American home for dinner?" (percent saying no) 2. GSS, 1978–1994: "Do African-Americans attend the church that you, yourself, attend most often, or not?" (percent saying no, among people who attend a church) 3. GSS, 1977–1994: "If you and your friends belonged to a social club that would not let Blacks join, would you try to change the rules so that Blacks could join?" (percent saying no) 4. GSS, 1972–1985: "How strongly would you object if a member of your family wanted to bring a Black friend home to dinner? Would you object strongly, mildly, or not at all?" (percent objecting strongly or mildly) 5. GSS, 1972–2002: "Do you think there should be laws against marriages between African-Americans and Whites?" (percent saying yes) 6. GSS, 1990–2024: "What about having a close relative marry a Black person? Would you be very in favor of it happening, somewhat in favor, neither in favor nor opposed to it happening, somewhat opposed, or very opposed to it happening?" (percent somewhat or very opposed) The first figure shows non-Black respondents. The GSS asked Black respondents about White friends, churchgoers and club members, so their answers are shown in a second figure, with the same numbers. ```{figure} figs/_composed/composed_race_social_distance_time_series_period_effect.png :name: race-social-distance-time-series :align: left `rachome`, `racchurh`, `racchng`, `racdin`, `racmar`, `marblk` — Time series: percent giving the answer that indicates distance, numbered as in the text. ``` All six declined. The percentage saying no one in their family had brought a Black friend home to dinner (1) fell from 79% in 1973 to 60% in 2006, and the percentage whose church had no Black members attending (2) fell from 64% in 1978 to 53% in 1994. The percentage who would not try to change the rules of a club that excluded Black people (3) fell from 57% in 1977 to 31% in 1994, and the percentage who would object to a family member bringing a Black friend to dinner (4) fell from 29% in 1972 to 19% in 1985. Support for laws against interracial marriage (5) fell from 37% in 1972 to 10% in 2002. Opposition to a close relative marrying a Black person (6) fell from 63% in 1990 to 6% in 2024, the largest decline of the six. For Black respondents, questions 1–4 ask about White friends, churchgoers and club members, question 5 is the same, and question 6 asks about a relative marrying a White person: > What about having a close relative marry a White person? Would you be very in favor of it happening, somewhat in favor, neither in favor nor opposed to it happening, somewhat opposed, or very opposed to it happening? Most of these questions were asked of Black respondents only from 1978 or later. ```{figure} figs/_composed/composed_race_social_distance_black_time_series_period_effect.png :name: race-social-distance-black-time-series :align: left `rachome`, `racchurh`, `racchng`, `racdin`, `racmar`, `marwht` — Time series: percent of Black respondents giving the answer that indicates distance, numbered as in the text. ``` The percentage of Black respondents saying no one in their family had brought a White friend home to dinner (1) fell from 51% in 1980 to 43% in 2006, and the percentage whose church had no White members attending (2) fell from 61% in 1978 to 47% in 1994. The percentage who would not try to change the rules of a club that excluded White people (3) was 17% in 1985 and 21% in 1994. Few Black respondents would object to a family member bringing a White friend to dinner (4): 7% in 1980 and 5% in 1985. Support for laws against interracial marriage (5) fell from 11% in 1980 to 4% in 2002, and opposition to a close relative marrying a White person (6) fell from 11% in 2000 to 5% in 2024. ## Hypothetical workplace ### Affirmative action The GSS and ANES ask nearly the same question about preferential hiring: 1. ANES, 1986–2024: "Some people say that because of past discrimination Blacks should be given preference in hiring and promotion. Others say that such preference in hiring and promotion of Blacks is wrong because it gives Blacks advantages they haven't earned. What about your opinion -- are you for or against preferential hiring and promotion of Blacks?" 2. GSS, 1994–2024: "Some people say that because of past discrimination, Black people should be given preference in hiring and promotion. Others say that such preference in hiring and promotion of Black people is wrong because it discriminates against Whites. What about your opinion -- are you for or against preferential hiring and promotion of Black people?" The questions differ in the argument against: the ANES says preference gives Black people "advantages they haven't earned"; the GSS says it "discriminates against Whites." The following figure shows the percentage in favor, strongly or not strongly. ```{figure} figs/_composed/composed_race_affirmative_action_time_series_period_effect.png :name: race-affirmative-action-time-series :align: left `VCF0867a`, `affrmact` — Time series: percent in favor of preferential hiring and promotion of Black people, numbered as in the text. ``` Support is low in both surveys, but it has increased. In the ANES (1), it fell from 24% in 1986 to about 18% in the late 1990s and early 2000s, then rose to 35% in 2024. In the GSS (2), it was about 17% from 1994 to 2006, then rose to 27% in 2024. Where they overlap, the ANES numbers are a few points higher. ```{figure} figs/_composed/composed_race_affirmative_action_cohort_effect.png :name: race-affirmative-action-cohort :align: left `VCF0867a`, `affrmact` — Cohort component: percent in favor, numbered as in the text. ``` The cohort component is nearly flat for people born between 1900 and the early 1960s, around 16–20%, and then increases. In the ANES (1), it rises from 18% for people born in 1900 to 37% for people born in 2006; in the GSS (2), from 16% for people born in 1905 to 33% for people born in 2006. For the GSS question, we can estimate that about 35% of the increase from 1994 to 2024 is due to generational replacement, and about 65% to changing minds. ### Discrimination in the workplace The GSS also asks whether people think White workers lose out to Black workers: > What do you think the chances are these days that a White person won't get a job or promotion while an equally or less qualified Black person gets one instead? Is this very likely, somewhat likely, or not very likely to happen these days? The following figure shows the percentage who say very or somewhat likely, by race and ethnicity. ```{figure} figs/gss/discaff/composed_race_eth4_White_Black_Hispanic_time_series_period_effect.png :name: discaff-by-race-time-series :align: left `discaff` — Time series by race and ethnicity: percent saying very or somewhat likely. Hispanic is based on primary reported ancestry, for a consistent measure since 1972; the GSS's direct Hispanic-origin question begins in 2000. ``` Among White respondents, this percentage fell from 74% in 1990 to 58% in 2024. Among Black respondents, it rose from 33% in 1990 to about 50% in the mid-2000s, then fell to 36% in 2024; among Hispanic respondents, it rose from 49% to about 63%, then fell to 53%. ```{figure} figs/gss/discaff/composed_race_eth4_White_Black_Hispanic_cohort_effect.png :name: discaff-by-race-cohort :align: left `discaff` — Cohort component by race and ethnicity: percent saying very or somewhat likely. ``` Among White respondents, the cohort component rises from 55% for people born in 1901 to about 71% for people born between 1930 and 1945, then declines to 55% for people born in 2006. Among Black respondents, it is nearly flat, between 40% and 43%. For White respondents, we can estimate that about 27% of the decline from 1990 to 2024 is due to generational replacement, and about 73% to changing minds. A second question asks about people's own experience at work: > Do you feel in any way discriminated against on your job because of your race or ethnic origin? ```{figure} figs/gss/wkracism/composed_race_eth4_White_Black_Hispanic_time_series_period_effect.png :name: wkracism-by-race-time-series :align: left `wkracism` — Time series by race and ethnicity: percent of workers saying yes. Hispanic is based on primary reported ancestry. ``` Among White workers, the percentage saying yes is small and nearly constant, 2–3%. Among Black workers, it fell from 19% in 2002 to 10% in 2022, and among Hispanic workers, from 13% to 8%, though the intervals for both groups are wide. ```{figure} figs/gss/wkracism/composed_race_eth4_White_Black_Hispanic_cohort_effect.png :name: wkracism-by-race-cohort :align: left `wkracism` — Cohort component by race and ethnicity: percent of workers saying yes. ``` Within each group, the cohort component is nearly flat: about 11–14% for Black workers, about 9% for Hispanic workers, and 1–3% for White workers, rising to about 6% for the youngest. Among Black workers, almost all of the decline appears in the period component, not the cohort component: about 10 of the 11 percentage points. ## Actual workplace *(Placeholder — variables and data in [`race_plan.md`](race_plan.md) § Actual workplace.)* ## General condition *(Placeholder — variables and data in [`race_plan.md`](race_plan.md) § General condition.)* ## Explanations of different outcomes In 1977, the GSS added four questions starting with this preamble: > On the average, African-Americans have worse jobs, income, and housing than White people. Do you think these differences are ... Then they ask: > 1. ... mainly due to discrimination? > > 2. ... because most African-Americans have less in-born ability to learn? > > 3. ... because most African-Americans don't have the chance for education that it takes to rise out of poverty? > > 4. ... because most African-Americans just don't have the motivation or will power to pull themselves up out of poverty? Respondents can answer yes or no to each question, so they could affirm all four, or none, or any combination. The following figure shows the percentage affirming each explanation. ```{figure} figs/gss_racdif_battery_time_series.png :name: racdif-time-model :align: left Racial inequality explanations — time-series. ``` Over time, respondents are * Less likely to accept lack of motivation or in-born ability as explanations, and * More likely to accept discrimination. Belief in educational opportunity as an explanation was high in the 1980s, declined in the 1990s, and has increased since. The following figure shows how the responses differ between generations. ```{figure} figs/gss_racdif_battery_cohort_effect.png :name: racdif-cohort :align: left Racial inequality explanations — standardized cohort component. ``` The picture here is a little more complicated, but in general, more recent cohorts are more likely to endorse discrimination and educational opportunity, and less likely to endorse motivation and innate ability — although it looks like cohorts born after 1990 are increasingly likely to believe in innate differences. The following figure shows the period effects. ```{figure} figs/gss_racdif_battery_period_effect.png :name: racdif-period :align: left Racial inequality explanations — standardized period component. ``` The period effects have the same shapes as the original time series, but the magnitudes of the changes are a little smaller after we factor out the cohort effects. So again, the changes we see over time are a combination of generational replacement and changed minds. ### Circumstances *(Placeholder — variables and data in [`race_plan.md`](race_plan.md) § Explanations of different outcomes — Circumstances.)* ### Attributes *(Placeholder — variables and data in [`race_plan.md`](race_plan.md) § Explanations of different outcomes — Attributes.)* ### Tolerance of racist speaker *(Placeholder — variables and data in [`race_plan.md`](race_plan.md) § Explanations of different outcomes — Tolerance of racist speaker.)* ### Actions *(Placeholder — variables and data in [`race_plan.md`](race_plan.md) § Explanations of different outcomes — Actions.)* ## Policy *(Placeholder — variables and data in [`race_plan.md`](race_plan.md) § Policy.)* ## Hypothetical politics From 1972 to 2010, the GSS asked: > If your party nominated an African-American for President, would you vote for him if he were qualified for the job? In earlier years the question said "Negro" or "Black." Gallup has asked a similar question since 1958: "If your party nominated a generally well-qualified person for president who happened to be Black, would you vote for that person?" The following figure shows the percentage who said they would. ```{figure} figs/gss/racpres/composed_gallup_time_series.png :name: racpres-gallup-time-series :align: left `racpres` — Time series: percent who would vote for a qualified Black presidential nominee. Line: GSS, 1972–2010. Circles: Gallup, 1958–2024. ``` The line shows the GSS; the circles show Gallup, which fills in the years before and after. In the GSS, the percentage who would vote for a Black nominee rose from 79% in 1972 to 96% in 2010, when the question was dropped. In [Gallup's polls](https://news.gallup.com/poll/285563/socialism-atheism-political-liabilities.aspx), it rose from 38% in 1958 to 77% in 1978 and 94% in 2007; in [January 2024](https://news.gallup.com/poll/609344/felonies-old-age-heavily-count-against-candidates.aspx) it was 92%. From 1978 to 1987, Gallup's numbers are a few points lower than the GSS, possibly because Gallup counts people with no opinion, while the GSS figures leave them out. ```{figure} figs/gss/racpres/model10_cohort_standardized.png :name: racpres-cohort-standardized :align: left `racpres` — Cohort component: percent who would vote for a qualified Black presidential nominee. ``` The cohort component rises from 71% for people born in 1900 to about 90% for people born around 1950, and it is nearly flat after that. We can estimate that about 31% of the increase from 1972 to 2010 is due to generational replacement, and about 69% to changing minds. ## Representation in government *(Placeholder — variables and data in [`race_plan.md`](race_plan.md) § Representation in Govt.)* ## Too far, too fast? *(Placeholder — variables and data in [`race_plan.md`](race_plan.md) § Too Far, Too Fast?.)* ## Feelings *(TODO: the text and figures below predate the thermometer recode. The fits are now neutral to warm (50–97), not cold (0–49), and the by-race composites are `figs/anes/VCF0206/composed_race_White_Black_Other_*`; the section will also add Whites (`VCF0207`) and Asian Americans (`VCF0227`). See [`race_plan.md`](race_plan.md) § Feelings.)* ### Feeling thermometer: Blacks (ANES) The GSS questions above ask about policies and explanations for inequality. The American National Election Studies (ANES) also field a **feeling thermometer** toward Black people — respondents rate how warm or cold they feel on a 0–100 scale. We treat ratings **0–49** as cold or dislike and **50+** as neutral or warm. The following figures use the same cohort–period decomposition as the GSS items above, but on ANES data from 1964–2024. Each panel overlays **three series** by **respondent race** (white non-Hispanic, Black non-Hispanic, and other), using ANES summary variable `VCF0106`. ```{figure} figs/anes_VCF0206_by_race_time_series_period_effect.png :name: anes-blacks-thermometer-time-model :align: left Feeling thermometer toward Blacks — national time series by respondent race (ANES `VCF0206`). ``` Cold feelings toward Black people have declined for all three groups, but white respondents remain the most likely to give cold ratings. ```{figure} figs/anes_VCF0206_by_race_period_effect.png :name: anes-blacks-thermometer-period :align: left Feeling thermometer toward Blacks — standardized period effect by respondent race. ``` ```{figure} figs/anes_VCF0206_by_race_cohort_effect.png :name: anes-blacks-thermometer-cohort :align: left Feeling thermometer toward Blacks — standardized cohort component by respondent race. ``` The cohort profiles show the familiar pattern: younger birth cohorts are less likely to give cold ratings. The gap between white and Black respondents is largest among older cohorts and narrows somewhat among those born after 1980, though white respondents still rate Black people colder on average. ## What about now? So when Obama said, "You wouldn’t choose the fifties, or the sixties, or the seventies. You’d choose right now," was he right? If the prevalence of racist views is one of the factors you would consider, the choice is clear: public opinion was less racist in 2016 than it was in the 1970s. We don't have GSS data from the fifties and sixties, but we have data from people born in the 1900s and 1910s, and it is equally clear: younger cohorts are less racist than older cohorts. Of course, these are survey results, and people don't always say what they believe. But I think it's reasonable to believe that the changes we see in the responses reflect true changes in beliefs. For example, as opposition to interracial marriage has declined, rates of interracial marriage have in fact increased. And given that presidential elections tend to be tightly contested, if many people were unwilling to vote for a Black candidate, Obama would not have been elected (and might not have spoken at Howard University). But even if we conclude that Obama was right in 2016, would he still be right in 2024? In the most recent data, it looks like fewer people believe that unequal outcomes are "mainly due to discrimination". And it looks like the youngest respondents are more likely to believe in innate racial differences, compared to previous generations. *(Task **106** — interpret recent conservative movement in period/cohort components: see [`race_plan.md`](race_plan.md).)* Frustratingly, some of the questions where we see the biggest changes were dropped from the survey, so we can't rule out the possibility that those trends have reversed. But other surveys can fill in the blanks. For example, [Gallup reported](https://news.gallup.com/poll/354638/approval-interracial-marriage-new-high.aspx) that support for interracial marriage reached 94% in 2021, compared to only 4% in 1958. And when [they asked](https://news.gallup.com/poll/257495/americans-views-voting-candidates-various-characteristics.aspx) if people would vote for a Black presidential candidate in 2019, about 3% said no, slightly less than the proportion in the GSS in 2010. Even if racism is in decline, concern about racism is not. Since 1948, Gallup has asked "What do you think is the most important problem facing this country today?" The following figure shows the percentage who said racism or a related term ([source](https://news.gallup.com/poll/350165/heightened-racial-concern-clear-legacy-floyd-death.aspx)): ```{figure} downloaded_figs/0gwa4e54wk-jnwac6ieqww.png :name: gallup-racism-important-problem :align: left Percent naming racism (or a related term) as the most important problem facing the country (Gallup). ``` Several of the peaks align with notable historical events, including: | Year | Peak | Possible events | | ---------- | ------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------- | | 1958 | 29% | Aftermath of the [Little Rock Crisis](https://en.wikipedia.org/wiki/Little_Rock_Nine) in 1957 (federal troops escorting Black students into a desegregated school in Arkansas). | | 1967 | 52% | Urban riots and racial unrest, especially the [Detroit Riot](https://en.wikipedia.org/wiki/1967_Detroit_riot) and [Newark Riots](https://en.wikipedia.org/wiki/1967_Newark_riots). | | 1968 | 35% | Assassination of Martin Luther King Jr. followed by nationwide riots in more than 100 cities. | | 1992 | 15% | 1992 Los Angeles Riots after the acquittal of officers filmed beating Rodney King. | | 2015–2017 | 10–18% | Multiple events, possibly including the [Charleston church shooting](https://en.wikipedia.org/wiki/Charleston_church_shooting) in 2015, the killing of [Philando Castile](https://en.wikipedia.org/wiki/Killing_of_Philando_Castile) in 2016, and a [white supremacist rally in Charlottesville](https://en.wikipedia.org/wiki/Unite_the_Right_rally) in 2017. | | 2020–2021 | 19% | Killing of George Floyd in May 2020, nationwide protests, and intense discussion of systemic racism. | The salience of racism as a national issue seems to follow events, and media coverage, more than the prevalence of racist attitudes. That's not ideal — the limited resource of attention should go where it is most needed — but it may not be all bad. Even if racist beliefs are rarer than they used to be, they are not gone. I'm glad if only 3% of Americans refuse to vote for a Black candidate, but that's not zero, and in a close election, it could be a deciding factor. So we should be aware of what's going on — and we should keep an eye on the trends that have stalled or reversed. But we should not ignore the consistent decline of racist attitudes in the last fifty years. Based on the data, if I had to choose when to be born, not knowing who I was going to be, I would choose right now. --- Note: The scenario Obama posed is a nod to the [veil of ignorance](https://en.wikipedia.org/wiki/Original_position), a hypothetical used by moral philosopher John Rawls to support basic principles of justice.