Skip to content
Contents
The Public Mind

    Type to search. Planned chapters match by title only.

    Chapter 4.7

    Sort dates by attitude

    Among ANES partisans, the party gap passed a fixed threshold by the lower bound of its interval on racial resentment in 1992, on abortion in 2008 and on the importance of religion in 2016.

    Affiliation
    NYU Stern
    Published

    In the American National Election Studies (ANES) of 1980, the gaps between Republican and Democratic identifiers on abortion and on the importance of religion were near zero, −0.02−0.02 standard deviations of the attitude (Republican minus Democratic identifiers)Estimandparty gap in the within-year standardized abortion (higher = more restrictive), all respondents, all interview modesInterval95% CI −0.13 to 0.08n1,313Weightsweight (VCF0009z)PopulationRepublican and Democratic identifiers, leaners includedPeriod1980SourceANES Cumulative Data File and 0.050.05 standard deviations of the attitude (Republican minus Democratic identifiers)Estimandparty gap in the within-year standardized the importance of religion, all respondents, all interview modesInterval95% CI −0.06 to 0.16n1,196Weightsweight (VCF0009z)PopulationRepublican and Democratic identifiers, leaners includedPeriod1980SourceANES Cumulative Data File standard deviations of the answers.Republicans and Democrats here include independents who say they lean toward a party, the 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., throughout the essay. The two gaps then grew on different schedules. The abortion gap rose in the 1990s, while the religion gap stayed small through 2000 and grew after it. 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. already separated the parties at its first reading in 1986, and the gap on whether the national economy got better changes sign whenever the presidency changes parties. To set these schedules side by side, the chapter dates each gap by the first survey year in which it passes one fixed mark and stays above it. Crossing a chosen mark is not the origin of a divide, and each date moves with the rule, the item, the sample and the survey that produce it.

    The three gaps crossed the mark in different decades

    Each gap is measured in standard deviationsStandard deviationA measure of how far values typically fall from their average. A larger standard deviation means the values are more spread out. (SD) of the answers. For every attitude and survey year the answers are standardized among all respondents who answered the item that year, partisans and independents alike, so that their average is zero and their spread is one. The party gap is the average score among Republican identifiers minus the average among Democratic identifiers. A gap of zero means the parties answer alike, and a gap of one means their averages sit one standard deviation apart. The mark is 0.250.25 standard deviations of the attitude (Republican minus Democratic identifiers)Estimandsort-date rule carried over from an earlier analysis of the same surveys and keptWeightsnone (constant)Populationthreshold applied to every party-gap seriesPeriod1980-2024SourceANES Cumulative Data File SD, a little above the conventional label for a small difference between two means [1]. A sortParty sortingThe alignment of party with other traits, such as ideology, religion, race and place, so that people who share a party also tend to share those traits. date is the first survey year in which the gap passes the mark and stays above it in every later survey.

    From those starts near zero, by 2024 the abortion gap was 1.061.06 standard deviations of the attitude (Republican minus Democratic identifiers)Estimandparty gap in the within-year standardized abortion (higher = more restrictive), all respondents, all interview modesInterval95% CI 0.99 to 1.14n4,666Weightsweight (VCF0009z)PopulationRepublican and Democratic identifiers, leaners includedPeriod2024SourceANES Cumulative Data File SD and the religion gap 0.520.52 standard deviations of the attitude (Republican minus Democratic identifiers)Estimandparty gap in the within-year standardized the importance of religion, all respondents, all interview modesInterval95% CI 0.44 to 0.59n5,086Weightsweight (VCF0009z)PopulationRepublican and Democratic identifiers, leaners includedPeriod2024SourceANES Cumulative Data File SD. Racial resentment is the mean of the standardized Kinder and Sanders items [2], scored so that a higher value means more resentment. It has no reading before 1986, and its gap was already 0.370.37 standard deviations of the attitude (Republican minus Democratic identifiers)Estimandparty gap in the within-year standardized the racial resentment index, all respondents, all interview modesInterval95% CI 0.24 to 0.49n927Weightsweight (VCF0009z)PopulationRepublican and Democratic identifiers, leaners includedPeriod1986SourceANES Cumulative Data File SD then, above the mark. It was 1.231.23 standard deviations of the attitude (Republican minus Democratic identifiers)Estimandparty gap in the within-year standardized the racial resentment index, all respondents, all interview modesInterval95% CI 1.16 to 1.30n4,565Weightsweight (VCF0009z)PopulationRepublican and Democratic identifiers, leaners includedPeriod2024SourceANES Cumulative Data File SD in 2024, within sampling error of the 2020 reading. The series begins after the realignment on race that Carmines and Stimson described [3], so for all respondents these data cannot say when that gap first passed the mark. Judged by the lower end of each 95% intervalConfidence intervalA range around an estimate that would contain the true value in 95% of repeated samples if the sampling model held. The book reports 95% intervals unless it says otherwise., the gaps passed the mark for good in 19921992 survey yearEstimandfirst survey year whose lower 95% bound exceeds the rule and stays above it (sort date), the racial resentment index, all respondents, all interview modesWeightsweight (VCF0009z)PopulationRepublican and Democratic identifiers, leaners includedPeriod1980-2024SourceANES Cumulative Data File for racial resentment, 20082008 survey yearEstimandfirst survey year whose lower 95% bound exceeds the rule and stays above it (sort date), abortion (higher = more restrictive), all respondents, all interview modesWeightsweight (VCF0009z)PopulationRepublican and Democratic identifiers, leaners includedPeriod1980-2024SourceANES Cumulative Data File for abortion and 20162016 survey yearEstimandfirst survey year whose lower 95% bound exceeds the rule and stays above it (sort date), the importance of religion, all respondents, all interview modesWeightsweight (VCF0009z)PopulationRepublican and Democratic identifiers, leaners includedPeriod1980-2024SourceANES Cumulative Data File for religion. The racial resentment date is the plainest case of a crossing that is not an origin, because that gap was already above the mark when the series began.

    Figure 1

    By the point estimates, party gaps stay above 0.25 SD from 1986 on racial resentment, 1998 on abortion and 2012 on religion

    Republican minus Democratic identifiers (leaners included) in standard deviations of each attitude in that survey year, ANES 1980-2024, 95% intervals

    AbortionRacial resentmentReligion
    Show data
    Data for By the point estimates, party gaps stay above 0.25 SD from 1986 on racial resentment, 1998 on abortion and 2012 on religion
    SampleAttitudeSurvey yearParty gap (SD)Lower 95% boundUpper 95% boundRespondents
    All respondentsAbortion1980−0.02−0.130.081313
    All respondentsAbortion1982−0.09−0.210.031173
    All respondentsAbortion19840.120.030.211888
    All respondentsAbortion1986−0.03−0.120.061818
    All respondentsAbortion19880.04−0.050.141751
    All respondentsAbortion19900.06−0.040.161692
    All respondentsAbortion19920.260.170.342097
    All respondentsAbortion19940.190.080.301537
    All respondentsAbortion19960.250.130.361521
    All respondentsAbortion19980.320.190.451106
    All respondentsAbortion20000.270.160.381532
    All respondentsAbortion20040.370.230.51931
    All respondentsAbortion20080.420.260.58893
    All respondentsAbortion20120.650.580.725068
    All respondentsAbortion20160.700.630.783630
    All respondentsAbortion20200.960.911.027002
    All respondentsAbortion20241.060.991.144666
    All respondentsRacial resentment19860.370.240.49927
    All respondentsRacial resentment19880.420.330.521554
    All respondentsRacial resentment19900.310.180.44857
    All respondentsRacial resentment19920.490.400.571947
    All respondentsRacial resentment19940.520.410.621566
    All respondentsRacial resentment19980.510.380.631104
    All respondentsRacial resentment20000.570.450.681348
    All respondentsRacial resentment20040.740.620.87948
    All respondentsRacial resentment20080.730.620.841824
    All respondentsRacial resentment20120.890.820.964741
    All respondentsRacial resentment20161.030.961.113151
    All respondentsRacial resentment20201.291.241.346495
    All respondentsRacial resentment20241.231.161.304565
    All respondentsReligion19800.05−0.060.161196
    All respondentsReligion1984−0.04−0.130.061667
    All respondentsReligion19860.00−0.130.13908
    All respondentsReligion1988−0.05−0.150.051554
    All respondentsReligion1990−0.06−0.150.041713
    All respondentsReligion19920.00−0.090.091989
    All respondentsReligion19940.04−0.070.151569
    All respondentsReligion1996−0.01−0.120.101543
    All respondentsReligion19980.03−0.100.161114
    All respondentsReligion20000.08−0.030.191564
    All respondentsReligion20020.180.050.311366
    All respondentsReligion20040.11−0.020.241068
    All respondentsReligion20080.210.100.322010
    All respondentsReligion20120.300.230.375077
    All respondentsReligion20160.420.340.493652
    All respondentsReligion20200.460.400.527265
    All respondentsReligion20240.520.440.595086
    Face-to-face onlyAbortion1980−0.02−0.130.081313
    Face-to-face onlyAbortion1982−0.09−0.210.031173
    Face-to-face onlyAbortion19840.120.030.211888
    Face-to-face onlyAbortion1986−0.03−0.120.061818
    Face-to-face onlyAbortion19880.04−0.050.141751
    Face-to-face onlyAbortion19900.06−0.040.161692
    Face-to-face onlyAbortion19920.260.170.342097
    Face-to-face onlyAbortion19940.190.080.301537
    Face-to-face onlyAbortion19960.250.130.361521
    Face-to-face onlyAbortion19980.320.190.451106
    Face-to-face onlyAbortion20040.370.230.51931
    Face-to-face onlyAbortion20080.420.260.58893
    Face-to-face onlyAbortion20120.580.450.701791
    Face-to-face onlyAbortion20160.770.630.911030
    Face-to-face onlyAbortion20241.090.941.24953
    Face-to-face onlyRacial resentment19860.370.240.49927
    Face-to-face onlyRacial resentment19880.420.330.521554
    Face-to-face onlyRacial resentment19900.310.180.44857
    Face-to-face onlyRacial resentment19920.490.400.571947
    Face-to-face onlyRacial resentment19940.520.410.621566
    Face-to-face onlyRacial resentment19980.510.380.631104
    Face-to-face onlyRacial resentment20040.740.620.87948
    Face-to-face onlyRacial resentment20080.730.620.841824
    Face-to-face onlyRacial resentment20120.780.660.901694
    Face-to-face onlyRacial resentment20160.970.831.11940
    Face-to-face onlyRacial resentment20241.261.121.41862
    Face-to-face onlyReligion19800.05−0.060.161196
    Face-to-face onlyReligion1984−0.04−0.130.061667
    Face-to-face onlyReligion19860.00−0.130.13908
    Face-to-face onlyReligion1988−0.05−0.150.051554
    Face-to-face onlyReligion1990−0.06−0.150.041713
    Face-to-face onlyReligion19920.00−0.090.091989
    Face-to-face onlyReligion19940.04−0.070.151569
    Face-to-face onlyReligion1996−0.01−0.120.101543
    Face-to-face onlyReligion19980.03−0.100.161114
    Face-to-face onlyReligion20040.11−0.020.241068
    Face-to-face onlyReligion20080.210.100.322010
    Face-to-face onlyReligion20120.220.090.351797
    Face-to-face onlyReligion20160.380.230.531039
    Face-to-face onlyReligion20240.620.450.79966
    White non-Hispanic onlyAbortion19800.00−0.120.111092
    White non-Hispanic onlyAbortion1982−0.06−0.180.07999
    White non-Hispanic onlyAbortion19840.190.090.291531
    White non-Hispanic onlyAbortion19860.01−0.090.111415
    White non-Hispanic onlyAbortion19880.120.010.231320
    White non-Hispanic onlyAbortion19900.130.020.241283
    White non-Hispanic onlyAbortion19920.370.270.461605
    White non-Hispanic onlyAbortion19940.260.140.391183
    White non-Hispanic onlyAbortion19960.330.200.451181
    White non-Hispanic onlyAbortion19980.370.220.51826
    White non-Hispanic onlyAbortion20000.440.320.561144
    White non-Hispanic onlyAbortion20040.500.340.66672
    White non-Hispanic onlyAbortion20080.530.330.73444
    White non-Hispanic onlyAbortion20120.750.670.833009
    White non-Hispanic onlyAbortion20160.820.730.902631
    White non-Hispanic onlyAbortion20201.121.061.185147
    White non-Hispanic onlyAbortion20241.161.081.243389
    White non-Hispanic onlyRacial resentment19860.190.050.34731
    White non-Hispanic onlyRacial resentment19880.190.070.301195
    White non-Hispanic onlyRacial resentment19900.170.010.32655
    White non-Hispanic onlyRacial resentment19920.350.240.451512
    White non-Hispanic onlyRacial resentment19940.390.270.521206
    White non-Hispanic onlyRacial resentment19980.330.180.48826
    White non-Hispanic onlyRacial resentment20000.470.340.611045
    White non-Hispanic onlyRacial resentment20040.580.410.74685
    White non-Hispanic onlyRacial resentment20080.710.570.85907
    White non-Hispanic onlyRacial resentment20120.840.750.932819
    White non-Hispanic onlyRacial resentment20161.040.951.132307
    White non-Hispanic onlyRacial resentment20201.331.271.394827
    White non-Hispanic onlyRacial resentment20241.281.201.363350
    White non-Hispanic onlyReligion19800.130.010.251001
    White non-Hispanic onlyReligion19840.06−0.040.171358
    White non-Hispanic onlyReligion19860.09−0.050.24693
    White non-Hispanic onlyReligion19880.07−0.040.181194
    White non-Hispanic onlyReligion19900.04−0.070.141299
    White non-Hispanic onlyReligion19920.140.040.231552
    White non-Hispanic onlyReligion19940.150.020.271207
    White non-Hispanic onlyReligion19960.10−0.020.221197
    White non-Hispanic onlyReligion19980.14−0.010.28829
    White non-Hispanic onlyReligion20000.190.070.321167
    White non-Hispanic onlyReligion20020.320.180.471065
    White non-Hispanic onlyReligion20040.280.130.44758
    White non-Hispanic onlyReligion20080.340.210.48990
    White non-Hispanic onlyReligion20120.530.440.623012
    White non-Hispanic onlyReligion20160.610.520.702642
    White non-Hispanic onlyReligion20200.750.680.815348
    White non-Hispanic onlyReligion20240.710.620.803698

    In each panel, watch where the line climbs across the dashed rule and stays above it. The abortion line does so in the 1990s and the religion line after 2008, and the racial resentment line is already above the rule at its first point in 1986. Each point is the Republican minus Democratic gap in standard deviations of that year's answers, with its 95% band; the other views standardize within face-to-face interviews and within White respondents. Values in Table 1.

    Note: A positive gap means Republican identifiers score higher on the attitude. Abortion is scored so that a higher value means more restrictive, racial resentment so that a higher value means more resentment (mean of four standardized items) and religion so that a higher value means religion is more important. Racial resentment is not asked in 1996 and has only part of its battery in 1998. Abortion in 2008 was asked of part of the sample.Note: Web respondents enter the ANES in 2012, and 2020 has no weights for face-to-face or web respondents separately. The face-to-face view leaves out 2000 (a study that mixed face-to-face and telephone interviews), 2002 (telephone only) and 2020, and draws no line across 2000 or 2020.Source: ANES Cumulative Data File, 1980-2024; weighted (VCF0009z; weight_ftf and weight_web for the mode views)

    The dates come from one survey and two tests

    The data are the ANES Cumulative Data File, which asked the abortion, religion and economy items in study years from 1980 to 2024 and the racial resentment battery from 1986. The sample in each year is the respondents who identify with or lean toward a party and answered the item, weightedSurvey weightsNumbers that scale each respondent so the sample matches the population on known traits such as age, race, education and region. A weighted average counts a respondent with a large weight for more than one with a small weight. so that it matches the population on known traits. Two tests date each gap. The point-estimate test applies the rule to the estimate itself and dates abortion in 19981998 survey yearEstimandfirst survey year whose point estimate exceeds the rule and stays above it, abortion (higher = more restrictive), all respondents, all interview modesWeightsweight (VCF0009z)PopulationRepublican and Democratic identifiers, leaners includedPeriod1980-2024SourceANES Cumulative Data File and religion in 20122012 survey yearEstimandfirst survey year whose point estimate exceeds the rule and stays above it, the importance of religion, all respondents, all interview modesWeightsweight (VCF0009z)PopulationRepublican and Democratic identifiers, leaners includedPeriod1980-2024SourceANES Cumulative Data File. The lower-bound test applies it to the lower end of the 95% interval, so it asks whether the gap is reliably above the mark, and it gives the later dates. A date can fall only on a survey year, and one low draw in any later survey can push it later. The abortion lower-bound date rests on a single survey year: in 2008 the item went to part of the sample, 893893 respondentsEstimandrespondents in the two parties with an abortion answer (the item went to part of the sample)n893Weightsnone (count)PopulationRepublican and Democratic identifiers, leaners included, asked the abortion itemPeriod2008SourceANES Cumulative Data File respondents were in the two parties, and the lower bound clears the mark by 0.0090.009 standard deviations of the attitude (Republican minus Democratic identifiers)Estimandlower 95% bound of the party gap in the standardized abortion item minus the sort-date rule (the margin by which the 2008 lower bound clears the rule)n893Weightsweight (VCF0009z)PopulationRepublican and Democratic identifiers, leaners includedPeriod2008SourceANES Cumulative Data File SD.The intervals use the 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. and treat the weighted sample as if respondents were drawn independently rather than in the geographic clusters of the ANES design, so they may be narrower than intervals that follow the design. The report gives every estimate with its interval and sample size.

    The economic gap changes sign with the president

    Economic perceptions have no sort date. The item asks whether the national economy got better over the past year, scored so that a higher value means better, and its gap takes the sign of the president's party. In every survey year from 1980 to 2024 the point estimate shows the president's party as the more optimistic. The sign of the point estimate changed at every change in the party of the president. Under a Republican president Republican identifiers are the more optimistic, and under a Democratic president Democratic identifiers are. A rule that asks a gap to stay above a mark in every later survey gives this one no date. The Chapter 4.5, Perceptions of the economy chapter follows the item with a second survey.

    Figure 2

    The economic gap follows the president's party in all 18 survey years

    Republican minus Democratic identifiers (leaners included) in standard deviations of the answer to whether the national economy got better over the past year, ANES 1980-2024, 95% intervals

    Republican minus Democratic optimism about the economy (SD)

    Show data
    Data for The economic gap follows the president's party in all 18 survey years
    Survey yearPresidentPresident's partyParty gap (SD)Lower 95% boundUpper 95% boundRespondents
    1980CarterDemocratic−0.25−0.35−0.151341
    1982ReaganRepublican0.570.440.691193
    1984ReaganRepublican0.860.780.941881
    1986ReaganRepublican0.420.330.511798
    1988ReaganRepublican0.560.470.651721
    1990George H. W. BushRepublican0.07−0.030.171710
    1992George H. W. BushRepublican0.480.390.572135
    1994ClintonDemocratic−0.11−0.220.001560
    1996ClintonDemocratic−0.45−0.56−0.341537
    1998ClintonDemocratic−0.35−0.47−0.221114
    2000ClintonDemocratic−0.51−0.61−0.401555
    2002George W. BushRepublican0.340.210.471368
    2004George W. BushRepublican0.930.811.051064
    2008George W. BushRepublican0.220.100.342013
    2012ObamaDemocratic−0.97−1.03−0.905083
    2016ObamaDemocratic−0.83−0.90−0.763660
    2020TrumpRepublican0.890.830.947254
    2024BidenDemocratic−1.02−1.09−0.954877

    The line sits above zero in every red band, a Republican presidency, and below zero in every blue band, a Democratic one, and it crosses zero at each change of party. The series ends at 0.890.89 standard deviations of the attitude (Republican minus Democratic identifiers)Estimandparty gap in the within-year standardized the retrospective economy item (higher = more optimistic), all respondents, all interview modesInterval95% CI 0.83 to 0.94n7,254Weightsweight (VCF0009z)PopulationRepublican and Democratic identifiers, leaners includedPeriod2020SourceANES Cumulative Data File SD in 2020 under a Republican president and −1.02−1.02 standard deviations of the attitude (Republican minus Democratic identifiers)Estimandparty gap in the within-year standardized the retrospective economy item (higher = more optimistic), all respondents, all interview modesInterval95% CI −1.09 to −0.95n4,877Weightsweight (VCF0009z)PopulationRepublican and Democratic identifiers, leaners includedPeriod2024SourceANES Cumulative Data File SD in 2024 under a Democratic one; the intervals for 1990 and 1994 include zero. A positive value means Republican identifiers are the more optimistic. Values in Table 7.

    Note: Presidential terms are from the historical record, not from the ANES. Each survey year is assigned the president recorded for that year in the ANES file.Source: ANES Cumulative Data File, 1980-2024; weighted (VCF0009z; weight_ftf and weight_web for the mode views)

    The rule, the sample and the survey move the dates

    The order is the same under both tests, racial resentment first, then abortion, then religion, and it holds at every rule from 0.150.15 standard deviations of the attitude (Republican minus Democratic identifiers)Estimandlowest or highest threshold in the sensitivity checkWeightsnone (constant)Populationthreshold applied to every party-gap seriesPeriod1980-2024SourceANES Cumulative Data File to 0.500.50 standard deviations of the attitude (Republican minus Democratic identifiers)Estimandlowest or highest threshold in the sensitivity checkWeightsnone (constant)Populationthreshold applied to every party-gap seriesPeriod1980-2024SourceANES Cumulative Data File SD. At the highest rule the lower bound for religion has not stayed above it by 2024. The dates themselves move. Among White non-Hispanic respondents the abortion and religion gaps pass the mark in earlier survey years than among all respondents. One difference lies in who is in the Democratic group. In 2000, Democrats who were Black, Hispanic or of another group said religion was important more often than White Democrats did, and said a woman should always be able to obtain an abortion less often. Both differences shrink the gap among all respondents. For racial resentment the direction reverses. Among White respondents the gap is under the mark in 1986 and crosses it later, and part of the all-respondent gap at the first reading separates racial groups that sit in different parties. Limiting the sample to face-to-face interviews leaves the abortion and racial resentment dates unchanged and moves the religion dates later; its lower-bound date falls in 2024, the last survey, which no later survey can test. The General Social Survey, with a different abortion item, dates abortion later than the ANES does. Its church attendance item dates religion earlier than the ANES importance item by the point estimate and later by the lower bound.

    Figure 3

    At the 0.25 SD rule, the lower bound dates abortion and religion later than the point estimates do, and both tests date religion after abortion

    First survey year in which the party gap exceeds the rule and stays above it in every later survey, by the point estimate (open dot) and by the lower bound of the 95% interval (solid dot), ANES

    • Point estimate
    • Lower 95% bound
    Show data
    Data for Racial resentment, abortion and religion are dated in that order at every rule
    AttitudeSampleRuleFirst readingLast reading at or below the rulePoint estimate above fromLower bound above from (sort date)Upper bound above fromAbove the rule at the first readingDrawn in the chart
    Racial resentmentAll respondents0.15 SD1986n/a198619861986yesyes
    Racial resentmentWhite non-Hispanic only0.15 SD1986n/a198619921986yesyes
    Racial resentmentFace-to-face only0.15 SD1986n/a198619861986yesno
    AbortionAll respondents0.15 SD19801990199219981990noyes
    AbortionWhite non-Hispanic only0.15 SD19801990199219961988noyes
    AbortionFace-to-face only0.15 SD19801990199219981990nono
    ReligionAll respondents0.15 SD19802004200820121998noyes
    ReligionWhite non-Hispanic only0.15 SD19801998200020081992noyes
    ReligionFace-to-face only0.15 SD19802004200820161998noyes
    Racial resentmentAll respondents0.25 SD1986n/a198619921986yesyes
    Racial resentmentWhite non-Hispanic only0.25 SD19861990199220001986noyes
    Racial resentmentFace-to-face only0.25 SD1986n/a198619921986yesno
    AbortionAll respondents0.25 SD19801996199820081992noyes
    AbortionWhite non-Hispanic only0.25 SD19801990199220001992noyes
    AbortionFace-to-face only0.25 SD19801996199820081992nono
    ReligionAll respondents0.25 SD19802008201220162008noyes
    ReligionWhite non-Hispanic only0.25 SD19802000200220121998noyes
    ReligionFace-to-face only0.25 SD19802012201620242008noyes
    Racial resentmentAll respondents0.35 SD19861990199219921986noyes
    Racial resentmentWhite non-Hispanic only0.35 SD19861998200020041992noyes
    Racial resentmentFace-to-face only0.35 SD19861990199219921986nono
    AbortionAll respondents0.35 SD19802000200420121996noyes
    AbortionWhite non-Hispanic only0.35 SD19801996199820121992noyes
    AbortionFace-to-face only0.35 SD19801998200420121996nono
    ReligionAll respondents0.35 SD19802012201620202012noyes
    ReligionWhite non-Hispanic only0.35 SD19802008201220122002noyes
    ReligionFace-to-face only0.35 SD19802012201620242016noyes
    Racial resentmentAll respondents0.50 SD19861992199420041992noyes
    Racial resentmentWhite non-Hispanic only0.50 SD19862000200420082000noyes
    Racial resentmentFace-to-face only0.50 SD19861992199420041992nono
    AbortionAll respondents0.50 SD19802008201220122004noyes
    AbortionWhite non-Hispanic only0.50 SD19802000200420121998noyes
    AbortionFace-to-face only0.50 SD19802008201220162004noyes
    ReligionAll respondents0.50 SD198020202024n/a2020noyes
    ReligionWhite non-Hispanic only0.50 SD19802008201220162012noyes
    ReligionFace-to-face only0.50 SD198020162024n/a2016noyes

    Read each row from its open dot, the first survey whose estimate stays above the rule, to its solid dot, the first survey whose lower bound does. The thin bar reaches back to the earliest year the interval allows. The rows for all respondents keep the order racial resentment, abortion, religion, and the White and face-to-face rows date the same attitudes in other survey years. A star marks a series already above the rule at its first reading, so its open dot marks that reading and not a crossing. The other views change the rule. Values in Table 2.

    Note: * The series is above the rule at its first reading, so the open dot marks that reading and not a crossing, and the date is not identified. A star marks each such row.Note: A face-to-face row with the same three dates as the all-respondent row is not drawn: that is racial resentment at every rule and abortion at every rule but the highest. Show data lists them. At the highest rule no lower bound stays above the rule through 2024 for religion among all respondents or among face-to-face respondents, so those two rows are not drawn.Note: Survey years are 2 to 4 years apart, so a date is known to the survey year, not to the month. A date in 2024, the last survey, cannot be tested for staying above the rule.Source: ANES Cumulative Data File, 1980-2024; weighted (VCF0009z; weight_ftf and weight_web for the mode views)

    What the dates describe, and what they do not

    A sort date describes an association between party and an answer in one survey year, and the three dates give an order. They do not show whether an attitude drew people to a party, whether party led people to change their answers, or whether new and departing voters changed the makeup of each party. Panel studies address direction: Margolis [4] found that for many Americans partisanship shapes religious identity, and Engelhardt [5] found White respondents aligning their racial attitudes with their party. A sort date is the year a survey-based gap passes a threshold. The ANES cannot show when party leaders took their positions, so the dates say nothing about them. The Chapter 4.1, The abortion sort, Chapter 3.8, The god gap in America and Companion 4.6a, Racial attitudes and ideology chapters follow the attitudes one at a time.

    What these dates cannot tell you

    The dates depend on choices the data cannot settle: the rule, the item, the sample and the survey. Each 95% interval is computed on its own, and the lower-bound test needs the interval of every later survey to stay above the mark, so one low draw can push a date later. The standardization is within year, so a gap in standard deviations can move when the spread of answers changes and the parties do not. Web respondents enter the ANES in 2012, the year that sets the point-estimate date for religion. In that year the face-to-face and web gaps on religion fall on either side of the mark, so a change in interview modeSurvey modeHow respondents were interviewed: in person, by phone, on paper or online. Mode can change answers, so a change in mode can look like a change in opinion. sits inside that date.

    Go deeper

    • The report: every estimate with its interval and sample size.
    • Tables: every figure of the report as a table.
    • Issues: the explorer for this part.

    Data sources

    • ANES Cumulative Data FileCumulative Data File 1948-2024, study years 1980-2024; 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: weight (VCF0009z); weight_ftf and weight_web for the mode views · License: Free download after accepting the ANES terms of use; cite the release.
    • ANES 2024 Time Series2024 Time Series, importance of religion on the five-point scaleYears: 2024 (pre-election and post-election interviews) · Weights: V240107a (pre-election) · License: Free download after accepting the ANES terms of use; cite the release.
    • General Social SurveyGeneral Social Survey extract, 1977-2024Years: 1972-2024 (35 study years) · Weights: weight_use · License: See the provider's terms of use.
    • Cooperative Election StudyCooperative Election Study, even years 2008-2024Years: 2006-2024 (every year; self-reported turnout is available in the 10 even years) · Weights: weight_cumulative · License: See the provider's terms of use.

    All data sources

    References

    1. Cohen, Jacob (1988). Statistical Power Analysis for the Behavioral Sciences, 2nd edition. Lawrence Erlbaum Associates.
    2. Kinder, Donald R.; Sanders, Lynn M. (1996). Divided by Color: Racial Politics and Democratic Ideals. University of Chicago Press.
    3. Carmines, Edward G.; Stimson, James A. (1989). Issue Evolution: Race and the Transformation of American Politics. Princeton University Press. doi:10.1515/9780691218250
    4. Margolis, Michele F. (2018). From Politics to the Pews: How Partisanship and the Political Environment Shape Religious Identity. University of Chicago Press. doi:10.7208/chicago/9780226555812.001.0001
    5. Engelhardt, Andrew M. (2021). Racial Attitudes Through a Partisan Lens. British Journal of Political Science 51(3): 1062-1079. doi:10.1017/S0007123419000437

    How to cite

    @misc{singh2026sortdatesbyattitude,
  author = {Singh, Vishal},
  title = {Sort dates by attitude},
  year = {2026},
  url = {https://vishalsingh.org/politics/sort-dates-by-attitude},
  note = {In The Public Mind. Published 30 September 2026.}
}