Ethnicity — variable results

Exploration digest for the survey stack behind book/ethnicity.md (Task 167). Grouped by topic; within each variable: question → time series → cohort trajectories → cohort component → period component with time-series overlay.

GSS mode (Task 175): GSS fits use interviewer responses only: 2021 (almost all web) is dropped, and 2022 and 2024 use the in-person and phone respondents, with weights raked to the full sample. Coding: the Hispanic thermometer (VCF0217) is the mean rating in degrees (Task 165); percent cold is the alternate cold. Intermarriage is shown with the alternate oppose [4, 5], as marblk in the race chapter: “favor” dipped in 2021 when “neither” jumped in the GSS’s first web year (2021 is no longer used; the coding is kept). Stereotype items are derived: _rel = 1 if the group is rated closer to the negative pole than White people on the same 7-point scale (lazy, unintelligent, violent); the rich–poor ratings are fit but left out (Allen, 2026-10-14: perceptions of wealth aren’t the point; actual wealth differences come later); part of the ANES decline in “lazier” comes from the reference group, since the share rating White people lazy doubled (8% to 16%, 1992–2024). Discrimination: a great deal or a lot (ANES), a lot (Pew). Influence in American politics is shown as too little (alternate toolittle; VCF9275 is too little by default), for Black, White, Hispanic and Asian people. 2000 asked about influence “in American life and politics”, a different question, and is not included. Sources: ANES items not in the CDF (discrimination, influence, peaceful–violent, and VCF9275 2024) come from the 2012–2024 study files; Pew pew_discrim_hisp is ATP Waves 24.5 (web group), 53, 84, 118, 143, 167, period only (no birth year), so it has no cohort figures. Items with two or three waves (the 1990–2000 GSS ratings, violent, Asian discrimination and influence) show levels across groups; their cohort components are weak. The GSS violence scale runs 1 = violence-prone, 7 = not. For these 2–3-wave series the model line is pulled toward 50% (up to 3 points), so read levels from the points. All fits pass their sampling gates (some after refits at draws 12000).

Full metadata: notebooks/utils.py. Ops: ethnicity_refresh.md.

Figure stems (Task 95 nested): figs/{gss|anes}/{var}/time_model.png, model10_cohort_trajectories.png, model10_cohort_standardized.png, model10_period_standardized[_time_series].png.

Layout template (Task 130 / 148): topic H2 → per-variable H3 with question + four figures. Regenerate with python scripts/gen_chapter_results_md.py ethnicity.

Warmth¶

VCF0217 — Feeling thermometer: Hispanics¶

Selected (y=1): [0–97] — Mean rating, degrees

Question

Feeling thermometer: Hispanics / Chicanos / Hispanic-Americans.

Time series

Cohort trajectories

Cohort component

Period component (with time series)

By respondent race (Other includes Hispanic respondents): time series

By respondent race (Other includes Hispanic respondents): period component

By respondent race (Other includes Hispanic respondents): cohort component

Social distance — intermarriage¶

marhisp — Close relative marry a Hispanic American?¶

Selected (y=1): [4, 5] — Percent opposed

Coding: oppose (alternate coding; the default is [1, 2])

Question

What about having a close relative marry a Hispanic American 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?

Time series

Cohort trajectories

Cohort component

Period component (with time series)

marasian — Close relative marry an Asian American?¶

Selected (y=1): [4, 5] — Percent opposed

Coding: oppose (alternate coding; the default is [1, 2])

Question

What about having a close relative marry an Asian American 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?

Time series

Cohort trajectories

Cohort component

Period component (with time series)

Stereotypes — relative to White people (GSS)¶

workhsp_rel — Hard-working or lazy?¶

Selected (y=1): [1] — Percent rating Hispanic people lazier than White people

Question

Derived from workhsps and workwhts: 1 if the respondent places Hispanic Americans closer to ‘lazy’ than White people on the 7-point hardworking–lazy scale.

Time series

Cohort trajectories

Cohort component

Period component (with time series)

intlhsp_rel — Intelligent or unintelligent?¶

Selected (y=1): [1] — Percent rating Hispanic people less intelligent than White people

Question

Derived from intlhsps and intlwhts: 1 if the respondent places Hispanic Americans closer to ‘unintelligent’ than White people on the 7-point scale.

Time series

Cohort trajectories

Cohort component

Period component (with time series)

Stereotypes across groups, 1990–2000 (GSS)¶

workasn_rel — Hard-working or lazy?¶

Selected (y=1): [1] — Percent rating Asian people lazier than White people

Question

Derived from workasns and workwhts: 1 if the respondent places Asian Americans closer to ‘lazy’ than White people on the 7-point scale (GSS 1990-2000).

Time series

Cohort trajectories

Cohort component

Period component (with time series)

intlasn_rel — Intelligent or unintelligent?¶

Selected (y=1): [1] — Percent rating Asian people less intelligent than White people

Question

Derived from intlasns and intlwhts: 1 if the respondent places Asian Americans closer to ‘unintelligent’ than White people on the 7-point scale (GSS 1990-2000).

Time series

Cohort trajectories

Cohort component

Period component (with time series)

violasn_rel — Peaceful or violent?¶

Selected (y=1): [1] — Percent rating Asian people more violent than White people

Question

Derived from violasns and violwhts: 1 if the respondent places Asian Americans closer to ‘violence-prone’ than White people on the 7-point scale (GSS 1990-2000).

Time series

Cohort trajectories

Cohort component

Period component (with time series)

violhsp_rel — Peaceful or violent?¶

Selected (y=1): [1] — Percent rating Hispanic people more violent than White people

Question

Derived from violhsps and violwhts: 1 if the respondent places Hispanic Americans closer to ‘violence-prone’ than White people on the 7-point scale (GSS 1990-2000).

Time series

Cohort trajectories

Cohort component

Period component (with time series)

violblk_rel — Peaceful or violent?¶

Selected (y=1): [1] — Percent rating Black people more violent than White people

Question

Derived from violblks and violwhts: 1 if the respondent places Black people closer to ‘violence-prone’ than White people on the 7-point scale (GSS 1990-2000).

Time series

Cohort trajectories

Cohort component

Period component (with time series)

workjew_rel — Hard-working or lazy?¶

Selected (y=1): [1] — Percent rating Jewish people lazier than White people

Question

Derived from workjews and workwhts: 1 if the respondent places Jews closer to ‘lazy’ than White people on the 7-point scale (GSS 1990-2000).

Time series

Cohort trajectories

Cohort component

Period component (with time series)

intljew_rel — Intelligent or unintelligent?¶

Selected (y=1): [1] — Percent rating Jewish people less intelligent than White people

Question

Derived from intljews and intlwhts: 1 if the respondent places Jews closer to ‘unintelligent’ than White people on the 7-point scale (GSS 1990-2000).

Time series

Cohort trajectories

Cohort component

Period component (with time series)

violjew_rel — Peaceful or violent?¶

Selected (y=1): [1] — Percent rating Jewish people more violent than White people

Question

Derived from violjews and violwhts: 1 if the respondent places Jews closer to ‘violence-prone’ than White people on the 7-point scale (GSS 1990-2000).

Time series

Cohort trajectories

Cohort component

Period component (with time series)

Stereotypes — relative to White people (ANES)¶

VCF9272_rel — Hard-working or lazy?¶

Selected (y=1): [1] — Percent rating Hispanic people lazier than White people

Question

Derived from VCF9272 and VCF9270: 1 if the respondent places Hispanic people closer to ‘lazy’ than White people on the 7-point hardworking–lazy scale.

Time series

Cohort trajectories

Cohort component

Period component (with time series)

VCF9273_rel — Hard-working or lazy?¶

Selected (y=1): [1] — Percent rating Asian people lazier than White people

Question

Derived from VCF9273 and VCF9270: 1 if the respondent places Asian people closer to ‘lazy’ than White people on the 7-point hardworking–lazy scale.

Time series

Cohort trajectories

Cohort component

Period component (with time series)

violent_hisp_rel — Peaceful or violent?¶

Selected (y=1): [1] — Percent rating Hispanic people more violent than White people

Question

Derived from violent_hisp and violent_white: 1 if the respondent places Hispanic people closer to ‘violent’ than White people on the 7-point peaceful–violent scale.

Time series

Cohort trajectories

Cohort component

Period component (with time series)

violent_asian_rel — Peaceful or violent?¶

Selected (y=1): [1] — Percent rating Asian people more violent than White people

Question

Derived from violent_asian and violent_white: 1 if the respondent places Asian people closer to ‘violent’ than White people on the 7-point peaceful–violent scale.

Time series

Cohort trajectories

Cohort component

Period component (with time series)

Perceived discrimination¶

discrim_hisp — How much discrimination is there against Hispanics?¶

Selected (y=1): [1, 2] — Percent saying a great deal or a lot

Question

How much discrimination is there in the United States today against Hispanics? A great deal / a lot / a moderate amount / a little / none at all.

Time series

Cohort trajectories

Cohort component

Period component (with time series)

discrim_asian — How much discrimination is there against Asians?¶

Selected (y=1): [1, 2] — Percent saying a great deal or a lot

Question

How much discrimination is there in the United States today against Asians? A great deal / a lot / a moderate amount / a little / none at all.

Time series

Cohort trajectories

Cohort component

Period component (with time series)

pew_discrim_hisp — How much discrimination is there against Hispanics?¶

Selected (y=1): [1] — Percent saying a lot

Question

Please tell us how much discrimination there is against each of these groups in our society today. Hispanics / Hispanic people. A lot / Some / Only a little / None at all

Time series

Cohort trajectories

Missing figure: figs/pew/pew_discrim_hisp/ (traj). Run the survey pipeline, then regenerate.

Cohort component

Missing figure: figs/pew/pew_discrim_hisp/ (coh). Run the survey pipeline, then regenerate.

Period component (with time series)

Missing figure: figs/pew/pew_discrim_hisp/ (per). Run the survey pipeline, then regenerate.

Influence in American politics¶

VCF9275 — Do Black people have too little influence in American politics?¶

Selected (y=1): [3] — Percent saying yes

Question

Would you say that Blacks have too much influence in American politics, just about the right amount of influence in American politics, or too little influence in American politics?

Time series

Cohort trajectories

Cohort component

Period component (with time series)

influence_white — How much influence do Whites have in politics?¶

Selected (y=1): [3] — Percent saying too little

Coding: toolittle (alternate coding; the default is [1])

Question

Would you say that whites have too much influence in American politics, just about the right amount of influence, or too little influence?

Time series

Cohort trajectories

Cohort component

Period component (with time series)

influence_hisp — How much influence do Hispanics have in politics?¶

Selected (y=1): [3] — Percent saying too little

Coding: toolittle (alternate coding; the default is [1])

Question

Would you say that Hispanics have too much influence in American politics, just about the right amount of influence, or too little influence?

Time series

Cohort trajectories

Cohort component

Period component (with time series)

influence_asian — How much influence do Asians have in politics?¶

Selected (y=1): [3] — Percent saying too little

Coding: toolittle (alternate coding; the default is [1])

Question

Would you say that Asians have too much influence in American politics, just about the right amount of influence, or too little influence?

Time series

Cohort trajectories

Cohort component

Period component (with time series)

Multiculturalism (GSS)¶

meltpot — Should groups keep distinct cultures or blend in?¶

Selected (y=1): [1, 2, 3] — Percent closer to maintaining distinct cultures (1–3)

Question

Some people say that it is better for America if different racial and ethnic groups maintain their distinct cultures. Others say that it is better if groups change so that they blend into the larger society as in the idea of a melting pot. Score 1 = maintain their distinct cultures, 7 = blend into the larger society.

Time series

Cohort trajectories

Cohort component

Period component (with time series)

meltpot1 — Should groups keep distinct customs or blend in?¶

Selected (y=1): [1] — Percent saying maintain distinct customs

Question

Some people say that it is better for a country if different racial and ethnic groups maintain their distinct customs and traditions. Others say that it is better if these groups adapt and blend into the larger society. Which of these views comes closer to your own?

Time series

Cohort trajectories

Cohort component

Period component (with time series)

mincult — Should government help minorities preserve their customs?¶

Selected (y=1): [1, 2] — Percent who agree

Question

Ethnic minorities should be given government assistance to preserve their customs and traditions. Agree strongly / agree / neither / disagree / disagree strongly.

Time series

Cohort trajectories

Cohort component

Period component (with time series)

ethclose — How close do you feel to your ethnic or racial group?¶

Selected (y=1): [1, 2] — Percent very close or close

Question

How close do you feel to your ethnic or racial group? Very close / close / not very close / not close at all.

Time series

Cohort trajectories

Cohort component

Period component (with time series)