Chapter 1.3

Sort dates by attitude

By the lower bound of the 95% interval, the party gap passed 0.25 standard deviations on racial resentment in 1992, abortion in 2008 and religion in 2016; the point estimates alone put those dates at 1986 (the first reading), 1998 and 2012.

Author: Vishal Singh
Affiliation: NYU Stern
Published: 30 September 2026

Three of the four party gaps in this chapter persist once they open, and the fourth, on perceptions of the economy, changes sign whenever the presidency changes parties. A date here is the survey year in which a gap crosses a threshold, and it is not the historical beginning of a divide. By the lower bound of the 95% interval, the gap between Republican and Democratic identifiers passed 0.25 standard deviations (SD) of the answers on racial resentment in 1992, on abortion in 2008 and on the importance of religion in 2016, and stays above that mark in every later survey. The dates move with the test, the sample and the survey, as later sections show.

A sort date is the survey year in which the party gap on an attitude crosses a fixed threshold and stays above it. This chapter finds that year with one rule for every attitude. For each attitude and survey year the answers are standardized, so that the average is zero and the standard deviation is one among all respondents that year. The party gap is the average of that score among Republican identifiers minus the average among Democratic identifiers, with leaners counted with the party they lean toward. A gap of zero means the parties answer alike, and a gap of one means their averages sit one standard deviation apart. The data are the American National Election Studies (ANES) Cumulative Data File, which asked the abortion, religion and economic items from 1980 to 2024 and a racial resentment battery from 1986. The Chapter 1.1, Party and ideology chapter dates the merger of party and ideology, and this one compares four attitudes and dates three.

The lower-bound test applies the rule to the lower end of the 95% interval, so it asks whether the gap is reliably above the mark, and the sort date is the first survey year in which that end exceeds 0.25 SD and stays above it in every later survey. The point-estimate test applies the same rule to the estimate itself and shows how early the estimates alone put the date.The mark is a convention, a little above Cohen's conventional label for a small difference between two means [1]. It was set in an earlier analysis of the same surveys and kept, and the views of Figure 2 move it from 0.15 to 0.50 SD. Intervals use the effective sample size, treat the weighted sample as if respondents were drawn independently rather than in the geographic sampling clusters of the ANES design, and treat the standard deviation used for standardizing as a fixed number, so they may be narrower than intervals that follow the design, which the Cumulative Data File has no variables to test.

The three gaps opened in different decades

Figure 1 plots the gap for each attitude and survey year. On abortion the gap was −0.02 (95% CI −0.13 to 0.08) SD in 1980 and 1.06 (95% CI 0.99 to 1.14) in 2024. Between 1992 and 1996 the estimates fall on both sides of the mark and every interval includes it, and the 1996 estimate is 0.003 SD under the mark, so the point-estimate date of 1998 rests on that reading.The abortion gap was 0.06 (95% CI −0.04 to 0.16) in 1990, 0.26 (95% CI 0.17 to 0.34) in 1992, 0.19 (95% CI 0.08 to 0.30) in 1994 and 0.247 (95% CI 0.134 to 0.359) in 1996, and the change from 1992 to 1994 is −0.06 (95% CI −0.20 to 0.07) SD, so these readings do not separate a fall from sampling variation. Adams [2] tested whether abortion followed the pattern of issue evolution (a new issue slowly realigning the parties) that Carmines and Stimson [3] described for race, using congressional roll calls and General Social Survey (GSS) data through 1994. The ANES point estimate for the gap between identifiers is under the mark in 1994, the last year of his survey data, and its interval includes the mark. The Chapter 4.1, The abortion sort chapter follows the abortion item in detail.

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

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
Sample	Attitude	Survey year	Party gap (SD)	Lower 95% bound	Upper 95% bound	Respondents
All respondents	Abortion	1980	−0.02	−0.13	0.08	1313
All respondents	Abortion	1982	−0.09	−0.21	0.03	1173
All respondents	Abortion	1984	0.12	0.03	0.21	1888
All respondents	Abortion	1986	−0.03	−0.12	0.06	1818
All respondents	Abortion	1988	0.04	−0.05	0.14	1751
All respondents	Abortion	1990	0.06	−0.04	0.16	1692
All respondents	Abortion	1992	0.26	0.17	0.34	2097
All respondents	Abortion	1994	0.19	0.08	0.30	1537
All respondents	Abortion	1996	0.25	0.13	0.36	1521
All respondents	Abortion	1998	0.32	0.19	0.45	1106
All respondents	Abortion	2000	0.27	0.16	0.38	1532
All respondents	Abortion	2004	0.37	0.23	0.51	931
All respondents	Abortion	2008	0.42	0.26	0.58	893
All respondents	Abortion	2012	0.65	0.58	0.72	5068
All respondents	Abortion	2016	0.70	0.63	0.78	3630
All respondents	Abortion	2020	0.96	0.91	1.02	7002
All respondents	Abortion	2024	1.06	0.99	1.14	4666
All respondents	Racial resentment	1986	0.37	0.24	0.49	927
All respondents	Racial resentment	1988	0.42	0.33	0.52	1554
All respondents	Racial resentment	1990	0.31	0.18	0.44	857
All respondents	Racial resentment	1992	0.49	0.40	0.57	1947
All respondents	Racial resentment	1994	0.52	0.41	0.62	1566
All respondents	Racial resentment	1998	0.51	0.38	0.63	1104
All respondents	Racial resentment	2000	0.57	0.45	0.68	1348
All respondents	Racial resentment	2004	0.74	0.62	0.87	948
All respondents	Racial resentment	2008	0.73	0.62	0.84	1824
All respondents	Racial resentment	2012	0.89	0.82	0.96	4741
All respondents	Racial resentment	2016	1.03	0.96	1.11	3151
All respondents	Racial resentment	2020	1.29	1.24	1.34	6495
All respondents	Racial resentment	2024	1.23	1.16	1.30	4565
All respondents	Religion	1980	0.05	−0.06	0.16	1196
All respondents	Religion	1984	−0.04	−0.13	0.06	1667
All respondents	Religion	1986	0.00	−0.13	0.13	908
All respondents	Religion	1988	−0.05	−0.15	0.05	1554
All respondents	Religion	1990	−0.06	−0.15	0.04	1713
All respondents	Religion	1992	0.00	−0.09	0.09	1989
All respondents	Religion	1994	0.04	−0.07	0.15	1569
All respondents	Religion	1996	−0.01	−0.12	0.10	1543
All respondents	Religion	1998	0.03	−0.10	0.16	1114
All respondents	Religion	2000	0.08	−0.03	0.19	1564
All respondents	Religion	2002	0.18	0.05	0.31	1366
All respondents	Religion	2004	0.11	−0.02	0.24	1068
All respondents	Religion	2008	0.21	0.10	0.32	2010
All respondents	Religion	2012	0.30	0.23	0.37	5077
All respondents	Religion	2016	0.42	0.34	0.49	3652
All respondents	Religion	2020	0.46	0.40	0.52	7265
All respondents	Religion	2024	0.52	0.44	0.59	5086
Face-to-face only	Abortion	1980	−0.02	−0.13	0.08	1313
Face-to-face only	Abortion	1982	−0.09	−0.21	0.03	1173
Face-to-face only	Abortion	1984	0.12	0.03	0.21	1888
Face-to-face only	Abortion	1986	−0.03	−0.12	0.06	1818
Face-to-face only	Abortion	1988	0.04	−0.05	0.14	1751
Face-to-face only	Abortion	1990	0.06	−0.04	0.16	1692
Face-to-face only	Abortion	1992	0.26	0.17	0.34	2097
Face-to-face only	Abortion	1994	0.19	0.08	0.30	1537
Face-to-face only	Abortion	1996	0.25	0.13	0.36	1521
Face-to-face only	Abortion	1998	0.32	0.19	0.45	1106
Face-to-face only	Abortion	2004	0.37	0.23	0.51	931
Face-to-face only	Abortion	2008	0.42	0.26	0.58	893
Face-to-face only	Abortion	2012	0.58	0.45	0.70	1791
Face-to-face only	Abortion	2016	0.77	0.63	0.91	1030
Face-to-face only	Abortion	2024	1.09	0.94	1.24	953
Face-to-face only	Racial resentment	1986	0.37	0.24	0.49	927
Face-to-face only	Racial resentment	1988	0.42	0.33	0.52	1554
Face-to-face only	Racial resentment	1990	0.31	0.18	0.44	857
Face-to-face only	Racial resentment	1992	0.49	0.40	0.57	1947
Face-to-face only	Racial resentment	1994	0.52	0.41	0.62	1566
Face-to-face only	Racial resentment	1998	0.51	0.38	0.63	1104
Face-to-face only	Racial resentment	2004	0.74	0.62	0.87	948
Face-to-face only	Racial resentment	2008	0.73	0.62	0.84	1824
Face-to-face only	Racial resentment	2012	0.78	0.66	0.90	1694
Face-to-face only	Racial resentment	2016	0.97	0.83	1.11	940
Face-to-face only	Racial resentment	2024	1.26	1.12	1.41	862
Face-to-face only	Religion	1980	0.05	−0.06	0.16	1196
Face-to-face only	Religion	1984	−0.04	−0.13	0.06	1667
Face-to-face only	Religion	1986	0.00	−0.13	0.13	908
Face-to-face only	Religion	1988	−0.05	−0.15	0.05	1554
Face-to-face only	Religion	1990	−0.06	−0.15	0.04	1713
Face-to-face only	Religion	1992	0.00	−0.09	0.09	1989
Face-to-face only	Religion	1994	0.04	−0.07	0.15	1569
Face-to-face only	Religion	1996	−0.01	−0.12	0.10	1543
Face-to-face only	Religion	1998	0.03	−0.10	0.16	1114
Face-to-face only	Religion	2004	0.11	−0.02	0.24	1068
Face-to-face only	Religion	2008	0.21	0.10	0.32	2010
Face-to-face only	Religion	2012	0.22	0.09	0.35	1797
Face-to-face only	Religion	2016	0.38	0.23	0.53	1039
Face-to-face only	Religion	2024	0.62	0.45	0.79	966
White non-Hispanic only	Abortion	1980	0.00	−0.12	0.11	1092
White non-Hispanic only	Abortion	1982	−0.06	−0.18	0.07	999
White non-Hispanic only	Abortion	1984	0.19	0.09	0.29	1531
White non-Hispanic only	Abortion	1986	0.01	−0.09	0.11	1415
White non-Hispanic only	Abortion	1988	0.12	0.01	0.23	1320
White non-Hispanic only	Abortion	1990	0.13	0.02	0.24	1283
White non-Hispanic only	Abortion	1992	0.37	0.27	0.46	1605
White non-Hispanic only	Abortion	1994	0.26	0.14	0.39	1183
White non-Hispanic only	Abortion	1996	0.33	0.20	0.45	1181
White non-Hispanic only	Abortion	1998	0.37	0.22	0.51	826
White non-Hispanic only	Abortion	2000	0.44	0.32	0.56	1144
White non-Hispanic only	Abortion	2004	0.50	0.34	0.66	672
White non-Hispanic only	Abortion	2008	0.53	0.33	0.73	444
White non-Hispanic only	Abortion	2012	0.75	0.67	0.83	3009
White non-Hispanic only	Abortion	2016	0.82	0.73	0.90	2631
White non-Hispanic only	Abortion	2020	1.12	1.06	1.18	5147
White non-Hispanic only	Abortion	2024	1.16	1.08	1.24	3389
White non-Hispanic only	Racial resentment	1986	0.19	0.05	0.34	731
White non-Hispanic only	Racial resentment	1988	0.19	0.07	0.30	1195
White non-Hispanic only	Racial resentment	1990	0.17	0.01	0.32	655
White non-Hispanic only	Racial resentment	1992	0.35	0.24	0.45	1512
White non-Hispanic only	Racial resentment	1994	0.39	0.27	0.52	1206
White non-Hispanic only	Racial resentment	1998	0.33	0.18	0.48	826
White non-Hispanic only	Racial resentment	2000	0.47	0.34	0.61	1045
White non-Hispanic only	Racial resentment	2004	0.58	0.41	0.74	685
White non-Hispanic only	Racial resentment	2008	0.71	0.57	0.85	907
White non-Hispanic only	Racial resentment	2012	0.84	0.75	0.93	2819
White non-Hispanic only	Racial resentment	2016	1.04	0.95	1.13	2307
White non-Hispanic only	Racial resentment	2020	1.33	1.27	1.39	4827
White non-Hispanic only	Racial resentment	2024	1.28	1.20	1.36	3350
White non-Hispanic only	Religion	1980	0.13	0.01	0.25	1001
White non-Hispanic only	Religion	1984	0.06	−0.04	0.17	1358
White non-Hispanic only	Religion	1986	0.09	−0.05	0.24	693
White non-Hispanic only	Religion	1988	0.07	−0.04	0.18	1194
White non-Hispanic only	Religion	1990	0.04	−0.07	0.14	1299
White non-Hispanic only	Religion	1992	0.14	0.04	0.23	1552
White non-Hispanic only	Religion	1994	0.15	0.02	0.27	1207
White non-Hispanic only	Religion	1996	0.10	−0.02	0.22	1197
White non-Hispanic only	Religion	1998	0.14	−0.01	0.28	829
White non-Hispanic only	Religion	2000	0.19	0.07	0.32	1167
White non-Hispanic only	Religion	2002	0.32	0.18	0.47	1065
White non-Hispanic only	Religion	2004	0.28	0.13	0.44	758
White non-Hispanic only	Religion	2008	0.34	0.21	0.48	990
White non-Hispanic only	Religion	2012	0.53	0.44	0.62	3012
White non-Hispanic only	Religion	2016	0.61	0.52	0.70	2642
White non-Hispanic only	Religion	2020	0.75	0.68	0.81	5348
White non-Hispanic only	Religion	2024	0.71	0.62	0.80	3698

Abortion and religion start near zero and end above the 0.25 SD rule, and the racial resentment point estimate is above the rule at its first reading in 1986 and rises from 0.37 to 1.23 SD, with the 2020 and 2024 estimates within sampling error of each other. Each panel plots the difference between the mean of Republican and Democratic identifiers on one attitude, measured in standard deviations of that attitude among all respondents in the same year and sample; a value of zero means the parties do not differ. The dashed line is the rule used to date a sort.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 gap on the importance of religion was 0.05 (95% CI −0.06 to 0.16) SD in 1980 and 0.08 (95% CI −0.03 to 0.19) in 2000, then 0.30 (95% CI 0.23 to 0.37) in 2012 and 0.52 (95% CI 0.44 to 0.59) in 2024. Layman [4] traced growing religious differences between the parties from the early 1960s to the late 1990s. The item here asks whether religion is important in the respondent's own life, and the gap on it between all Democratic and Republican identifiers stays small through 2000. Church attendance and the sample matter for the date, as a later section shows. The Chapter 9.1, The god gap in America chapter compares importance, attendance and denomination.

Racial resentment is the mean of the standardized answers to the Kinder and Sanders items [5], scored so that a higher value means more resentment. Its gap was 0.37 (95% CI 0.24 to 0.49) SD at the first reading in 1986, 0.73 (95% CI 0.62 to 0.84) in 2008, 1.29 (95% CI 1.24 to 1.34) in 2020 and 1.23 (95% CI 1.16 to 1.30) in 2024, and the 2020 and 2024 estimates are within sampling error of each other. The series begins after the realignment on race that Carmines and Stimson described, so for all respondents these data cannot say when the gap first passed the mark (White respondents are dated below). Tesler [6] reported that racial attitudes were more closely tied to political evaluations in the Obama years than before. Among all respondents the gap rises at each survey from 2008 to 2020, including from 0.73 in 2008 to 0.89 in 2012, and part of the 2012 step comes from the web respondents added that year, whose gap is higher than the face-to-face gap (Figure 3). The Chapter 5.2, Racial attitudes and ideology chapter takes up the tie between resentment and ideology.

Two tests date religion after abortion

Figure 2 shows both tests for each attitude. For abortion the point estimate passes the mark between 1996 and 1998, and the lower bound stays above it from 2008, which is 10 years later. For religion the point estimate passes it between 2008 and 2012 and the lower bound from 2016, 4 years later. The upper bound gives the other side of the range: it stays above the mark from 1992 for abortion and from 2008 for religion, so the interval allows dates from 1992 to 2008 for abortion and from 2008 to 2016 for religion. For racial resentment the point estimate is above the mark at the first reading, 1986, and the lower bound stays above it from 1992 after falling below it in 1990. Taken as ranges, the dates for racial resentment and abortion meet in 1992 and those for abortion and religion in 2008, so the order holds test by test, and the ranges leave open that adjacent attitudes share a survey year.The abortion lower-bound date rests on a single survey year: in 2008 the item went to part of the sample, 893 respondents were in the two parties, and the lower bound clears the mark by 0.009 SD.

Figure 2

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
Attitude	Sample	Rule	First reading	Last reading at or below the rule	Point estimate above from	Lower bound above from (sort date)	Upper bound above from	Above the rule at the first reading	Drawn in the chart
Racial resentment	All respondents	0.15 SD	1986	n/a	1986	1986	1986	yes	yes
Racial resentment	White non-Hispanic only	0.15 SD	1986	n/a	1986	1992	1986	yes	yes
Racial resentment	Face-to-face only	0.15 SD	1986	n/a	1986	1986	1986	yes	no
Abortion	All respondents	0.15 SD	1980	1990	1992	1998	1990	no	yes
Abortion	White non-Hispanic only	0.15 SD	1980	1990	1992	1996	1988	no	yes
Abortion	Face-to-face only	0.15 SD	1980	1990	1992	1998	1990	no	no
Religion	All respondents	0.15 SD	1980	2004	2008	2012	1998	no	yes
Religion	White non-Hispanic only	0.15 SD	1980	1998	2000	2008	1992	no	yes
Religion	Face-to-face only	0.15 SD	1980	2004	2008	2016	1998	no	yes
Racial resentment	All respondents	0.25 SD	1986	n/a	1986	1992	1986	yes	yes
Racial resentment	White non-Hispanic only	0.25 SD	1986	1990	1992	2000	1986	no	yes
Racial resentment	Face-to-face only	0.25 SD	1986	n/a	1986	1992	1986	yes	no
Abortion	All respondents	0.25 SD	1980	1996	1998	2008	1992	no	yes
Abortion	White non-Hispanic only	0.25 SD	1980	1990	1992	2000	1992	no	yes
Abortion	Face-to-face only	0.25 SD	1980	1996	1998	2008	1992	no	no
Religion	All respondents	0.25 SD	1980	2008	2012	2016	2008	no	yes
Religion	White non-Hispanic only	0.25 SD	1980	2000	2002	2012	1998	no	yes
Religion	Face-to-face only	0.25 SD	1980	2012	2016	2024	2008	no	yes
Racial resentment	All respondents	0.35 SD	1986	1990	1992	1992	1986	no	yes
Racial resentment	White non-Hispanic only	0.35 SD	1986	1998	2000	2004	1992	no	yes
Racial resentment	Face-to-face only	0.35 SD	1986	1990	1992	1992	1986	no	no
Abortion	All respondents	0.35 SD	1980	2000	2004	2012	1996	no	yes
Abortion	White non-Hispanic only	0.35 SD	1980	1996	1998	2012	1992	no	yes
Abortion	Face-to-face only	0.35 SD	1980	1998	2004	2012	1996	no	no
Religion	All respondents	0.35 SD	1980	2012	2016	2020	2012	no	yes
Religion	White non-Hispanic only	0.35 SD	1980	2008	2012	2012	2002	no	yes
Religion	Face-to-face only	0.35 SD	1980	2012	2016	2024	2016	no	yes
Racial resentment	All respondents	0.50 SD	1986	1992	1994	2004	1992	no	yes
Racial resentment	White non-Hispanic only	0.50 SD	1986	2000	2004	2008	2000	no	yes
Racial resentment	Face-to-face only	0.50 SD	1986	1992	1994	2004	1992	no	no
Abortion	All respondents	0.50 SD	1980	2008	2012	2012	2004	no	yes
Abortion	White non-Hispanic only	0.50 SD	1980	2000	2004	2012	1998	no	yes
Abortion	Face-to-face only	0.50 SD	1980	2008	2012	2016	2004	no	yes
Religion	All respondents	0.50 SD	1980	2020	2024	n/a	2020	no	yes
Religion	White non-Hispanic only	0.50 SD	1980	2008	2012	2016	2012	no	yes
Religion	Face-to-face only	0.50 SD	1980	2016	2024	n/a	2016	no	yes

At the 0.25 SD rule, religion is dated 14 years after abortion by the point estimates and 8 years after by the lower bounds. Each row is one series. The open dot is the first survey whose point estimate stays above the rule; the solid dot, the sort date, is the first survey whose lower bound of the 95% interval stays above it. The thin bar runs from the first survey whose upper bound stays above the rule, the earliest date the interval allows (1992 for abortion and 2008 for religion), to the sort date. Racial resentment is already above the rule at its first reading. The views change the rule from 0.15 to 0.50 SD.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)

Survey spacing widens the ranges. ANES surveys are 2 to 4 years apart, so a date can only fall on a survey year. The order is the same under both tests: racial resentment first, then abortion, then religion. It holds at every rule from 0.15 to 0.50 SD. At the highest rule the point estimate passes it in 1994 for racial resentment, 2012 for abortion and 2024 for religion, and the lower bound for religion has not stayed above it by 2024. At the lowest rule the point-estimate dates are 1992 for abortion and 2008 for religion.

White respondents date the gaps differently

The all-respondent gap compares identifiers of every race. Limiting the sample to White non-Hispanic respondents moves the point-estimate date for abortion from 1998 to 1992 and for religion from 2012 to 2002, and the lower-bound dates to 2000 and 2012. One difference lies in who is in the Democratic group. In 2000, 85.1% (95% CI 80.6% to 89.7%) of Democratic identifiers who were Black, Hispanic or of another group said religion was important, against 70.0% (95% CI 65.8% to 74.1%) of White Democrats. On abortion, 36.9% (95% CI 30.7% to 43.1%) of the first group said a woman should always be able to obtain an abortion, against 52.9% (95% CI 48.3% to 57.5%) of White Democrats. Both differences raise the Democratic average on the religious and restrictive sides, so they shrink the gap among all respondents.

For racial resentment the direction reverses. In 1986, 93.7% (95% CI 89.7% to 97.7%) of Black partisans identified with the Democratic Party, and their average score was −0.84 (95% CI −0.99 to −0.69) SD, against 0.14 (95% CI 0.07 to 0.21) for White partisans. Part of the all-respondent gap at the first reading therefore separates racial groups that sit in different parties. Among White respondents the gap was 0.19 (95% CI 0.05 to 0.34) SD in 1986, its point estimate passes the mark between 1990 and 1992 and its lower bound stays above it from 2000, and by 2024 it was 1.28 (95% CI 1.20 to 1.36) SD. The answer to when racial resentment became a party divide depends on whether the population is all respondents or White respondents.

The face-to-face series dates religion later

Limiting the sample to face-to-face interviews leaves the abortion and racial resentment dates unchanged and moves religion to 2016 for the point estimate and 2024 for the lower bound. The face-to-face series is smaller. It leaves out 2000, a study that mixed face-to-face and telephone interviews, 2002, which was conducted by telephone, and 2020, which has no weights for face-to-face or web respondents separately, and the ANES adds web respondents from 2012. The face-to-face gap on religion in 2016 is 0.38 (95% CI 0.23 to 0.53) SD with a lower bound under the mark, and 2020 has no face-to-face estimate, so its lower-bound date falls in 2024, the last survey, and no later survey can test it.

Figure 3 compares respondents by interview mode, web against face-to-face, in the years that have both. The interval for the web minus face-to-face difference includes zero in 8 of 9 comparisons. The exception is racial resentment in 2012, when the web gap was 0.958 (95% CI 0.874 to 1.043) SD and the face-to-face gap 0.784 (95% CI 0.664 to 0.905), a difference of 0.174 (95% CI 0.027 to 0.321). Across 9 comparisons with no mode difference, the expected number of 95% intervals that exclude zero is 0.45, so one exclusion is close to what chance alone would produce. The intervals that include zero reach as far as 0.32 SD from zero, so the check cannot rule out a mode difference larger than the 0.25 SD mark. In 2012, the year that sets the all-respondent point-estimate date for religion, the face-to-face gap on religion is 0.22 (95% CI 0.09 to 0.35) and the web gap 0.35 (95% CI 0.26 to 0.44), on either side of the mark, and their difference is 0.13 (95% CI −0.03 to 0.28) SD. In 2024 the face-to-face gap on religion is 0.62 (95% CI 0.45 to 0.79) and the web gap 0.51 (95% CI 0.39 to 0.62).The 2024 question on the importance of religion was revised to a graded scale, and the cumulative file counts extremely, very and moderately important as yes. On the graded scale in the 2024 Time Series the gap is 0.57 (95% CI 0.50 to 0.65) SD, against 0.52 (95% CI 0.44 to 0.59) on the collapsed item used here.

Figure 3

Web and face-to-face gaps differ by 0.174 SD or less, and the interval for the difference excludes zero only for racial resentment in 2012

Party gap by interview mode in the years that have both mode weights, difference and each mode, in standard deviations, with 95% intervals, ANES

Each dot is one web minus face-to-face difference, and 8 of 9 intervals include zero

Web minus face-to-face difference in the party gap (Republican minus Democratic identifiers, leaners included) in standard deviations, with 95% intervals, ANES

A dot to the right of the dashed line means a larger gap among web respondents, and an interval that crosses the line cannot separate the two modes. Only the racial resentment row for 2012 lies wholly to the right.

Show data

Data for Web and face-to-face gaps differ by 0.174 SD or less, and the interval for the difference excludes zero only for racial resentment in 2012, Difference
Row	Web minus face-to-face (SD)	Lower 95% bound	Upper 95% bound	Face-to-face respondents	Web respondents
Abortion, 2012	0.110	−0.042	0.262	1791	3277
Abortion, 2016	−0.092	−0.257	0.073	1030	2600
Abortion, 2024	−0.019	−0.204	0.165	953	1761
Racial resentment, 2012	0.174	0.027	0.321	1694	3047
Racial resentment, 2016	0.092	−0.071	0.255	940	2211
Racial resentment, 2024	0.001	−0.179	0.181	862	1581
Religion, 2012	0.129	−0.026	0.284	1797	3280
Religion, 2016	0.047	−0.124	0.219	1039	2613
Religion, 2024	−0.113	−0.317	0.090	966	1850

The web and face-to-face gaps are close in every row, and the racial resentment difference in 2012 is the only one whose interval excludes zero. The first tab plots each difference with its 95% interval and the second the gap in each mode. Each mode is standardized within itself. In 2024 the samples are the fresh in-person and fresh web samples; panel and paper cases have no mode weight.Note: The ANES mode weights exist for 2008, 2012, 2016 and 2024. The 2008 sample was face-to-face only and 2020 has no mode weights.Source: ANES Cumulative Data File, 2012, 2016 and 2024; weighted (weight_ftf and weight_web)

The GSS and CES date the gaps differently

The GSS and the Cooperative Election Study (CES) ask different items, so Figure 4 compares timing and not height. The GSS abortion-for-any-reason item has a gap of −0.13 (95% CI −0.26 to −0.01) SD in 1977 and 0.96 (95% CI 0.84 to 1.09) in 2024. Its point estimate stays above the mark from 2010 and its lower bound from 2012, later than the ANES dates for abortion. The GSS estimate was 0.24 (95% CI 0.10 to 0.39) in 2008, just under the mark, so the point-estimate date falls after 2008.

Figure 4

The GSS abortion gap passes 0.25 SD 12 years after the ANES gap by the point estimate and 4 years after by the lower bound

Republican minus Democratic identifiers (leaners included) in standard deviations of each survey's item in that year, ANES, GSS and CES, 95% intervals

Show data

Data for The GSS abortion gap passes 0.25 SD 12 years after the ANES gap by the point estimate and 4 years after by the lower bound
Attitude	Survey and measure	Survey year	Party gap (SD)	Lower 95% bound	Upper 95% bound	Respondents
Abortion	ANES	1980	−0.02	−0.13	0.08	1313
Abortion	ANES	1982	−0.09	−0.21	0.03	1173
Abortion	ANES	1984	0.12	0.03	0.21	1888
Abortion	ANES	1986	−0.03	−0.12	0.06	1818
Abortion	ANES	1988	0.04	−0.05	0.14	1751
Abortion	ANES	1990	0.06	−0.04	0.16	1692
Abortion	ANES	1992	0.26	0.17	0.34	2097
Abortion	ANES	1994	0.19	0.08	0.30	1537
Abortion	ANES	1996	0.25	0.13	0.36	1521
Abortion	ANES	1998	0.32	0.19	0.45	1106
Abortion	ANES	2000	0.27	0.16	0.38	1532
Abortion	ANES	2004	0.37	0.23	0.51	931
Abortion	ANES	2008	0.42	0.26	0.58	893
Abortion	ANES	2012	0.65	0.58	0.72	5068
Abortion	ANES	2016	0.70	0.63	0.78	3630
Abortion	ANES	2020	0.96	0.91	1.02	7002
Abortion	ANES	2024	1.06	0.99	1.14	4666
Abortion	GSS	1977	−0.13	−0.26	−0.01	1299
Abortion	GSS	1978	−0.08	−0.22	0.06	1263
Abortion	GSS	1980	0.05	−0.09	0.19	1165
Abortion	GSS	1982	−0.02	−0.16	0.12	1532
Abortion	GSS	1983	−0.16	−0.30	−0.02	1322
Abortion	GSS	1984	−0.10	−0.24	0.04	1245
Abortion	GSS	1985	0.04	−0.09	0.17	1322
Abortion	GSS	1987	−0.05	−0.18	0.08	1527
Abortion	GSS	1988	0.19	0.04	0.34	817
Abortion	GSS	1989	0.09	−0.06	0.24	861
Abortion	GSS	1990	0.13	−0.03	0.29	766
Abortion	GSS	1991	0.05	−0.10	0.21	816
Abortion	GSS	1993	0.31	0.16	0.46	872
Abortion	GSS	1994	0.17	0.06	0.28	1647
Abortion	GSS	1996	0.23	0.12	0.34	1513
Abortion	GSS	1998	0.30	0.18	0.41	1449
Abortion	GSS	2000	0.25	0.13	0.37	1378
Abortion	GSS	2002	0.39	0.22	0.55	719
Abortion	GSS	2004	0.36	0.19	0.53	700
Abortion	GSS	2006	0.41	0.28	0.53	1486
Abortion	GSS	2008	0.24	0.10	0.39	1051
Abortion	GSS	2010	0.39	0.24	0.54	968
Abortion	GSS	2012	0.48	0.33	0.64	966
Abortion	GSS	2014	0.47	0.35	0.60	1290
Abortion	GSS	2016	0.43	0.31	0.56	1465
Abortion	GSS	2018	0.62	0.48	0.76	1189
Abortion	GSS	2021	0.74	0.59	0.88	1044
Abortion	GSS	2022	1.08	0.90	1.26	831
Abortion	GSS	2024	0.96	0.84	1.09	1534
Abortion	CES	2014	0.86	0.83	0.88	44398
Abortion	CES	2016	0.92	0.90	0.94	51939
Abortion	CES	2018	1.07	1.05	1.09	49255
Abortion	CES	2020	1.11	1.09	1.13	49758
Racial resentment	ANES	1986	0.37	0.24	0.49	927
Racial resentment	ANES	1988	0.42	0.33	0.52	1554
Racial resentment	ANES	1990	0.31	0.18	0.44	857
Racial resentment	ANES	1992	0.49	0.40	0.57	1947
Racial resentment	ANES	1994	0.52	0.41	0.62	1566
Racial resentment	ANES	1998	0.51	0.38	0.63	1104
Racial resentment	ANES	2000	0.57	0.45	0.68	1348
Racial resentment	ANES	2004	0.74	0.62	0.87	948
Racial resentment	ANES	2008	0.73	0.62	0.84	1824
Racial resentment	ANES	2012	0.89	0.82	0.96	4741
Racial resentment	ANES	2016	1.03	0.96	1.11	3151
Racial resentment	ANES	2020	1.29	1.24	1.34	6495
Racial resentment	ANES	2024	1.23	1.16	1.30	4565
Racial resentment	GSS	1994	0.34	0.22	0.47	1211
Racial resentment	GSS	1996	0.46	0.36	0.57	1608
Racial resentment	GSS	1998	0.35	0.24	0.46	1478
Racial resentment	GSS	2000	0.38	0.27	0.50	1417
Racial resentment	GSS	2002	0.32	0.15	0.49	698
Racial resentment	GSS	2004	0.44	0.27	0.60	742
Racial resentment	GSS	2006	0.43	0.32	0.55	1556
Racial resentment	GSS	2008	0.34	0.20	0.49	1083
Racial resentment	GSS	2010	0.55	0.42	0.69	1134
Racial resentment	GSS	2012	0.51	0.36	0.66	997
Racial resentment	GSS	2014	0.54	0.41	0.66	1282
Racial resentment	GSS	2016	0.63	0.52	0.75	1496
Racial resentment	GSS	2018	0.72	0.58	0.86	1182
Racial resentment	GSS	2021	1.06	0.97	1.15	2044
Racial resentment	GSS	2022	0.89	0.75	1.03	1703
Racial resentment	GSS	2024	0.94	0.81	1.07	1538
Religion	ANES importance	1980	0.05	−0.06	0.16	1196
Religion	ANES importance	1984	−0.04	−0.13	0.06	1667
Religion	ANES importance	1986	0.00	−0.13	0.13	908
Religion	ANES importance	1988	−0.05	−0.15	0.05	1554
Religion	ANES importance	1990	−0.06	−0.15	0.04	1713
Religion	ANES importance	1992	0.00	−0.09	0.09	1989
Religion	ANES importance	1994	0.04	−0.07	0.15	1569
Religion	ANES importance	1996	−0.01	−0.12	0.10	1543
Religion	ANES importance	1998	0.03	−0.10	0.16	1114
Religion	ANES importance	2000	0.08	−0.03	0.19	1564
Religion	ANES importance	2002	0.18	0.05	0.31	1366
Religion	ANES importance	2004	0.11	−0.02	0.24	1068
Religion	ANES importance	2008	0.21	0.10	0.32	2010
Religion	ANES importance	2012	0.30	0.23	0.37	5077
Religion	ANES importance	2016	0.42	0.34	0.49	3652
Religion	ANES importance	2020	0.46	0.40	0.52	7265
Religion	ANES importance	2024	0.52	0.44	0.59	5086
Religion	GSS attendance	1977	0.06	−0.07	0.18	1333
Religion	GSS attendance	1978	0.15	0.01	0.29	1293
Religion	GSS attendance	1980	0.14	0.00	0.28	1204
Religion	GSS attendance	1982	0.04	−0.09	0.18	1607
Religion	GSS attendance	1983	0.03	−0.10	0.17	1386
Religion	GSS attendance	1984	0.09	−0.05	0.23	1275
Religion	GSS attendance	1985	0.01	−0.12	0.15	1364
Religion	GSS attendance	1986	0.09	−0.03	0.22	1269
Religion	GSS attendance	1987	0.05	−0.07	0.17	1586
Religion	GSS attendance	1988	0.23	0.11	0.35	1288
Religion	GSS attendance	1989	0.08	−0.04	0.20	1326
Religion	GSS attendance	1990	0.15	0.03	0.28	1170
Religion	GSS attendance	1991	0.04	−0.08	0.16	1289
Religion	GSS attendance	1993	0.20	0.08	0.32	1345
Religion	GSS attendance	1994	0.17	0.08	0.25	2490
Religion	GSS attendance	1996	0.11	0.02	0.20	2328
Religion	GSS attendance	1998	0.21	0.12	0.30	2251
Religion	GSS attendance	2000	0.27	0.17	0.36	2139
Religion	GSS attendance	2002	0.29	0.19	0.38	2147
Religion	GSS attendance	2004	0.21	0.11	0.30	2292
Religion	GSS attendance	2006	0.29	0.21	0.37	3415
Religion	GSS attendance	2008	0.32	0.21	0.44	1646
Religion	GSS attendance	2010	0.26	0.14	0.38	1617
Religion	GSS attendance	2012	0.27	0.15	0.40	1530
Religion	GSS attendance	2014	0.33	0.23	0.44	1945
Religion	GSS attendance	2016	0.30	0.20	0.40	2284
Religion	GSS attendance	2018	0.26	0.14	0.38	1816
Religion	GSS attendance	2021	0.37	0.28	0.45	3041
Religion	GSS attendance	2022	0.34	0.22	0.47	2551
Religion	GSS attendance	2024	0.44	0.33	0.55	2343
Religion	CES importance	2008	0.49	0.46	0.52	28578
Religion	CES importance	2010	0.42	0.39	0.44	47850
Religion	CES importance	2012	0.53	0.51	0.56	45816
Religion	CES importance	2014	0.48	0.45	0.51	44749
Religion	CES importance	2016	0.46	0.44	0.48	51957
Religion	CES importance	2018	0.61	0.58	0.63	49085
Religion	CES importance	2020	0.63	0.61	0.65	49754
Religion	CES importance	2022	0.67	0.65	0.70	49506
Religion	CES importance	2024	0.75	0.72	0.78	50837

The GSS abortion-for-any-reason gap stays above the rule from 2010 by the point estimate and 2012 by the lower bound, against 1998 and 2008 in the ANES. The GSS gap is −0.13 SD in 1977 and 0.96 in 2024, against −0.02 and 1.06 in the ANES, and the CES starts in 2014 at 0.86. The items differ across surveys, so compare the timing and not the height.Note: ANES: four-category abortion legality, racial resentment index, importance of religion. GSS: abortion for any reason, whether Black Americans should work their way up without special favors, church attendance. CES: always-allow abortion (2014 to 2020) and importance of religion (2008 to 2024), even years only. Each item is standardized within survey and year.Note: GSS interviews from 2021 include web respondents (87.3% of the 2021 respondents). In 2024 the face-to-face and web gaps are 0.84 and 1.06 SD on abortion, 0.73 and 1.14 on the work-your-way-up item and 0.40 and 0.51 on church attendance. The web gap is higher on all three items, beyond sampling error on the work-your-way-up item only.Source: ANES Cumulative Data File 1980-2024, General Social Survey 1977-2024 and Cooperative Election Study 2008-2024; weighted (VCF0009z, weight_use, weight_cumulative)

Religion gives a different answer by measure. GSS church attendance passes the mark in the point estimates between 2004 and 2006, and its lower bound clears the mark only in 2024, the last survey. ANES attendance, available through 2020, passes the mark by both tests in 2008. The point-estimate dates for both attendance measures are earlier than the ANES importance point-estimate date of 2012, and by the lower bound GSS attendance (2024) is later than ANES importance (2016). The GSS item on whether Black Americans should work their way up without special favors (the work-your-way-up item) begins in the extract in 1994 already above the mark, at 0.34 (95% CI 0.22 to 0.47) SD, and its lower bound stays above it from 2010. The CES begins too late to date either sort: its first religion gap, in 2008, is 0.49 (95% CI 0.46 to 0.52) SD and its first abortion gap, in 2014, is 0.86 (95% CI 0.83 to 0.88). The ANES importance gap in 2008 is 0.21 (95% CI 0.10 to 0.32), so the CES gap is larger by 0.28 (95% CI 0.17 to 0.39) SD in the same year, and the ANES religion date is specific to its yes-or-no item.

GSS interviews since 2021 include web respondents (87.3% of respondents in 2021). The GSS race gap moves from 0.72 SD in 2018 to 1.06 in 2021. In 2024 the face-to-face gap on that item is 0.73 (95% CI 0.51 to 0.95) SD and the web gap 1.14 (95% CI 0.97 to 1.30), a difference of 0.41 (95% CI 0.13 to 0.68), so the GSS race level after 2020 depends on interview mode. On abortion in 2024 the face-to-face gap is 0.84 (95% CI 0.63 to 1.06) SD and the web gap 1.06 (95% CI 0.90 to 1.22). Figure 5 lists the dates for the ANES all-respondent series and the other surveys.

Figure 5

Sort dates differ across surveys and items for the same attitude

Survey years in which the party gap passes the sort-date rule and stays above it, by survey and measure, all respondents

Survey	Measure	Lower bound above the rule from	Point estimate passes the rule between	Upper bound above the rule from
ANES	Abortion, 4-category	2008	1996 to 1998	1992
GSS	Abortion for any reason, answer no	2012	2008 to 2010	1993
CES	Abortion, always allow	at or before 2014	at or before 2014	at or before 2014
ANES	Racial resentment, 4-item index	1992	at or before 1986	at or before 1986
GSS	Work your way up (one item)	2010	at or before 1994	at or before 1994
ANES	Religion, importance	2016	2008 to 2012	2008
ANES	Church attendance, attends often	2008	2004 to 2008	1992
GSS	Church attendance, 0 to 8 scale	2024	2004 to 2006	1998
CES	Religion, importance	at or before 2008	at or before 2008	at or before 2008
ANES	Economy got better over the past year	none	none	none

Show data

Data for Sort dates differ across surveys and items for the same attitude
first	measure	sort_date	survey	upper	window
1980	Abortion, 4-category	2008	ANES	1992	1996 to 1998
1977	Abortion for any reason, answer no	2012	GSS	1993	2008 to 2010
2014	Abortion, always allow	at or before 2014	CES	at or before 2014	at or before 2014
1986	Racial resentment, 4-item index	1992	ANES	at or before 1986	at or before 1986
1994	Work your way up (one item)	2010	GSS	at or before 1994	at or before 1994
1980	Religion, importance	2016	ANES	2008	2008 to 2012
1980	Church attendance, attends often	2008	ANES	1992	2004 to 2008
1977	Church attendance, 0 to 8 scale	2024	GSS	1998	2004 to 2006
2008	Religion, importance	at or before 2008	CES	at or before 2008	at or before 2008
1980	Economy got better over the past year	none	ANES	none	none

By the lower bound the ANES dates abortion at 2008 and the GSS at 2012, and the ANES dates religion at 2016 on the importance item while the GSS attendance item is dated 2024. The window is the last survey with a point estimate at or below the rule and the first survey from which the point estimate stays above it; the crossing lies somewhere in that window. The sort date is the first survey from which the lower bound of the 95% interval stays above the rule, and the upper-bound year is the first survey from which the upper bound stays above it. The rule is 0.25 SD in every row, and the economic row has no date because the gap changes sign. Dates for the White and face-to-face samples and for the other rules are in the dumbbell chart of sort dates.Note: A date given as at or before a year belongs to a series that is above the rule at its first reading, so the date is not identified. The CES gaps are above the rule at their first readings, in 2008 for religion and 2014 for abortion.Note: ANES attendance ends in 2020: in the Cumulative Data File the attendance question has an answer for 49.1% of respondents in 2024 against 99.0% in 2020, so 2024 is not comparable.Source: ANES Cumulative Data File 1980-2024, General Social Survey 1977-2024 and Cooperative Election Study 2008-2024; weighted (VCF0009z, weight_use, weight_cumulative)

Pooled trends hide the parties moving apart

Figure 6 shows, for the question chosen to stand for each attitude, the share giving one answer, split by party and pooled. On religion the share saying it is important falls among Democrats and stays level among Republicans, so the gap grows through a fall on the Democratic side. On abortion the share saying a woman should always be able to obtain an abortion rises among Democrats and falls among Republicans.

Figure 6

The share saying religion is important fell among Democrats and stayed level among Republicans

Weighted percent giving one answer to the question chosen to stand for each attitude, by party (leaners included) and for all respondents, ANES 1980-2024, 95% intervals

Percent saying religion is important in their life

Show data

Data for The share saying religion is important fell among Democrats and stayed level among Republicans, so the pooled trend hides the party gap
Attitude	Group	Survey year	Percent	Lower 95% bound	Upper 95% bound	Respondents
Abortion	All respondents	1980	36.8	34.4	39.2	1547
Abortion	All respondents	1982	36.4	33.9	39.0	1359
Abortion	All respondents	1984	36.4	34.4	38.5	2171
Abortion	All respondents	1986	39.3	37.2	41.4	2120
Abortion	All respondents	1988	35.6	33.5	37.7	1996
Abortion	All respondents	1990	40.2	38.0	42.4	1932
Abortion	All respondents	1992	46.8	44.8	48.8	2406
Abortion	All respondents	1994	40.5	37.9	43.1	1730
Abortion	All respondents	1996	40.3	37.6	42.9	1679
Abortion	All respondents	1998	40.6	37.7	43.6	1258
Abortion	All respondents	2000	40.3	37.8	42.9	1758
Abortion	All respondents	2004	35.9	32.7	39.1	1047
Abortion	All respondents	2008	40.1	36.4	43.8	1031
Abortion	All respondents	2012	44.4	42.7	46.2	5855
Abortion	All respondents	2016	45.0	43.2	46.8	4207
Abortion	All respondents	2020	49.7	48.2	51.2	7900
Abortion	All respondents	2024	48.9	46.9	50.9	5021
Abortion	Democrats	1980	37.3	34.0	40.7	801
Abortion	Democrats	1982	36.0	32.6	39.5	741
Abortion	Democrats	1984	39.4	36.5	42.4	1032
Abortion	Democrats	1986	40.9	37.9	43.9	1061
Abortion	Democrats	1988	37.0	33.9	40.1	938
Abortion	Democrats	1990	43.3	40.2	46.4	993
Abortion	Democrats	1992	52.8	49.9	55.6	1201
Abortion	Democrats	1994	46.4	42.5	50.2	810
Abortion	Democrats	1996	46.1	42.4	49.8	879
Abortion	Democrats	1998	47.2	43.0	51.4	647
Abortion	Democrats	2000	47.2	43.5	50.9	868
Abortion	Democrats	2004	44.6	39.7	49.4	498
Abortion	Democrats	2008	51.0	45.9	56.2	600
Abortion	Democrats	2012	59.0	56.5	61.4	3081
Abortion	Democrats	2016	60.4	57.7	63.0	1915
Abortion	Democrats	2020	72.1	70.1	74.1	3713
Abortion	Democrats	2024	73.7	71.2	76.2	2459
Abortion	Republicans	1980	33.8	29.7	37.9	512
Abortion	Republicans	1982	36.1	31.6	40.6	432
Abortion	Republicans	1984	32.8	29.7	36.0	856
Abortion	Republicans	1986	38.2	34.7	41.6	757
Abortion	Republicans	1988	33.5	30.2	36.7	813
Abortion	Republicans	1990	38.6	35.0	42.2	699
Abortion	Republicans	1992	38.7	35.5	41.9	896
Abortion	Republicans	1994	33.9	30.1	37.7	727
Abortion	Republicans	1996	33.7	29.6	37.9	642
Abortion	Republicans	1998	31.7	27.1	36.4	459
Abortion	Republicans	2000	31.0	27.2	34.9	664
Abortion	Republicans	2004	25.7	21.3	30.2	433
Abortion	Republicans	2008	29.3	23.3	35.3	293
Abortion	Republicans	2012	29.0	26.4	31.6	1987
Abortion	Republicans	2016	27.5	24.9	30.0	1715
Abortion	Republicans	2020	24.0	22.1	25.9	3289
Abortion	Republicans	2024	21.4	18.9	23.8	2207
Racial resentment	All respondents	1986	55.4	52.4	58.4	1062
Racial resentment	All respondents	1988	57.6	55.2	59.9	1746
Racial resentment	All respondents	1990	54.3	51.2	57.4	967
Racial resentment	All respondents	1992	61.1	59.1	63.1	2227
Racial resentment	All respondents	1994	63.4	60.9	65.9	1749
Racial resentment	All respondents	2000	51.8	49.0	54.6	1532
Racial resentment	All respondents	2004	56.3	53.0	59.6	1058
Racial resentment	All respondents	2008	57.7	55.1	60.4	2079
Racial resentment	All respondents	2012	50.9	49.1	52.7	5484
Racial resentment	All respondents	2016	40.3	38.4	42.2	3626
Racial resentment	All respondents	2020	30.5	29.0	31.9	7367
Racial resentment	All respondents	2024	30.8	28.9	32.7	4902
Racial resentment	Democrats	1986	50.1	45.9	54.3	549
Racial resentment	Democrats	1988	52.9	49.5	56.3	828
Racial resentment	Democrats	1990	52.2	47.7	56.7	479
Racial resentment	Democrats	1992	54.4	51.4	57.3	1100
Racial resentment	Democrats	1994	58.2	54.5	62.0	826
Racial resentment	Democrats	2000	45.2	41.2	49.3	743
Racial resentment	Democrats	2004	46.0	41.2	50.9	502
Racial resentment	Democrats	2008	49.3	45.7	52.9	1212
Racial resentment	Democrats	2012	40.3	37.8	42.8	2880
Racial resentment	Democrats	2016	25.8	23.2	28.3	1669
Racial resentment	Democrats	2020	13.7	12.1	15.3	3441
Racial resentment	Democrats	2024	14.9	12.8	17.0	2393
Racial resentment	Republicans	1986	61.0	56.0	65.9	374
Racial resentment	Republicans	1988	61.4	57.9	65.0	718
Racial resentment	Republicans	1990	56.2	51.1	61.2	372
Racial resentment	Republicans	1992	68.7	65.5	71.8	841
Racial resentment	Republicans	1994	68.7	65.0	72.3	733
Racial resentment	Republicans	2000	56.8	52.4	61.2	595
Racial resentment	Republicans	2004	66.9	62.2	71.7	438
Racial resentment	Republicans	2008	69.9	65.6	74.2	598
Racial resentment	Republicans	2012	63.8	60.9	66.6	1852
Racial resentment	Republicans	2016	57.4	54.4	60.4	1474
Racial resentment	Republicans	2020	49.2	46.9	51.5	3042
Racial resentment	Republicans	2024	47.7	44.7	50.8	2167
Religion	All respondents	1980	75.3	73.0	77.5	1398
Religion	All respondents	1984	79.3	77.5	81.1	1911
Religion	All respondents	1986	78.5	76.1	81.0	1071
Religion	All respondents	1988	77.8	75.9	79.8	1760
Religion	All respondents	1990	79.2	77.4	81.0	1958
Religion	All respondents	1992	78.5	76.8	80.2	2295
Religion	All respondents	1994	76.9	74.7	79.1	1760
Religion	All respondents	1996	78.2	76.0	80.4	1706
Religion	All respondents	1998	77.2	74.7	79.7	1270
Religion	All respondents	2000	76.2	74.0	78.4	1799
Religion	All respondents	2002	75.4	72.7	78.1	1504
Religion	All respondents	2004	77.3	74.7	79.9	1203
Religion	All respondents	2008	71.5	69.2	73.8	2309
Religion	All respondents	2012	67.1	65.5	68.7	5882
Religion	All respondents	2016	65.1	63.4	66.8	4243
Religion	All respondents	2020	66.6	65.2	68.0	8249
Religion	All respondents	2024	65.8	64.0	67.7	5494
Religion	Democrats	1980	75.2	72.0	78.4	718
Religion	Democrats	1984	79.9	77.3	82.5	896
Religion	Democrats	1986	80.2	76.7	83.6	524
Religion	Democrats	1988	80.4	77.7	83.1	831
Religion	Democrats	1990	81.3	78.9	83.7	1004
Religion	Democrats	1992	79.5	77.1	81.9	1138
Religion	Democrats	1994	77.1	73.9	80.3	829
Religion	Democrats	1996	78.9	75.9	81.9	891
Religion	Democrats	1998	76.8	73.2	80.3	652
Religion	Democrats	2000	75.4	72.2	78.6	887
Religion	Democrats	2002	72.1	68.0	76.2	698
Religion	Democrats	2004	75.0	71.1	78.9	588
Religion	Democrats	2008	68.6	65.4	71.8	1358
Religion	Democrats	2012	61.1	58.7	63.5	3086
Religion	Democrats	2016	57.1	54.4	59.8	1930
Religion	Democrats	2020	57.1	54.9	59.2	3830
Religion	Democrats	2024	54.0	51.3	56.8	2636
Religion	Republicans	1980	77.4	73.7	81.2	478
Religion	Republicans	1984	78.5	75.6	81.4	771
Religion	Republicans	1986	80.2	76.2	84.2	384
Religion	Republicans	1988	78.3	75.3	81.3	723
Religion	Republicans	1990	79.0	76.0	82.0	709
Religion	Republicans	1992	79.5	76.8	82.2	851
Religion	Republicans	1994	78.8	75.5	82.0	740
Religion	Republicans	1996	78.5	74.9	82.0	652
Religion	Republicans	1998	77.8	73.7	82.0	462
Religion	Republicans	2000	78.9	75.5	82.2	677
Religion	Republicans	2002	79.7	76.0	83.4	668
Religion	Republicans	2004	79.5	75.6	83.5	480
Religion	Republicans	2008	78.1	74.4	81.8	652
Religion	Republicans	2012	75.3	72.8	77.7	1991
Religion	Republicans	2016	77.0	74.6	79.4	1722
Religion	Republicans	2020	78.7	76.9	80.6	3435
Religion	Republicans	2024	78.6	76.2	80.9	2450

Among Democrats the share falls from 75.2% to 54.0% while among Republicans it moves from 77.4% to 78.6%; the dashed pooled line falls from 75.3% to 65.8%. Lines are weighted shares with 95% bands.Note: Questions chosen to stand for each attitude: religion important (yes or no); abortion: by law a woman should always be able to obtain an abortion; racial resentment: agree strongly or somewhat that if blacks would only try harder they could be just as well off as whites.Source: ANES Cumulative Data File, 1980-2024; weighted (VCF0009z)

On racial resentment the share agreeing that Black Americans would be as well off if they tried harder falls in both parties, and falls further among Democrats. The pooled line describes the country, and the two party lines show in which party's identifiers the level changed.

The economic gap follows the president's party

Economic perceptions have no sort date. The item asks whether the national economy got better over the past year, and its gap takes the sign of the president's party: in 18 of 18 survey years the point estimate shows the president's party as the more optimistic, and in 16 of them the interval also excludes zero on that side. The survey years whose intervals include zero are 1990 and 1994. The sign of the point estimate changed at 6 of 6 changes in the party of the president, and at 5 of them the intervals at both readings exclude zero. The gap was 0.89 (95% CI 0.83 to 0.94) SD in 2020 under a Republican president and −1.02 (95% CI −1.09 to −0.95) SD in 2024 under a Democratic one. The Chapter 4.4, Perceptions of the economy chapter follows that item with a second survey.

Figure 7

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 year	President	President's party	Party gap (SD)	Lower 95% bound	Upper 95% bound	Respondents
1980	Carter	Democratic	−0.25	−0.35	−0.15	1341
1982	Reagan	Republican	0.57	0.44	0.69	1193
1984	Reagan	Republican	0.86	0.78	0.94	1881
1986	Reagan	Republican	0.42	0.33	0.51	1798
1988	Reagan	Republican	0.56	0.47	0.65	1721
1990	George H. W. Bush	Republican	0.07	−0.03	0.17	1710
1992	George H. W. Bush	Republican	0.48	0.39	0.57	2135
1994	Clinton	Democratic	−0.11	−0.22	0.00	1560
1996	Clinton	Democratic	−0.45	−0.56	−0.34	1537
1998	Clinton	Democratic	−0.35	−0.47	−0.22	1114
2000	Clinton	Democratic	−0.51	−0.61	−0.40	1555
2002	George W. Bush	Republican	0.34	0.21	0.47	1368
2004	George W. Bush	Republican	0.93	0.81	1.05	1064
2008	George W. Bush	Republican	0.22	0.10	0.34	2013
2012	Obama	Democratic	−0.97	−1.03	−0.90	5083
2016	Obama	Democratic	−0.83	−0.90	−0.76	3660
2020	Trump	Republican	0.89	0.83	0.94	7254
2024	Biden	Democratic	−1.02	−1.09	−0.95	4877

The point estimate is above zero under a Republican president and below zero under a Democratic one in 18 of 18 survey years, and the 95% interval excludes zero on that side in 16; the interval includes zero in 1990 and 1994. A positive gap means Republican identifiers are more optimistic about the economy. Shaded bands mark presidencies, blue for a Democratic president and red for a Republican one: Carter to 1981, Reagan and Bush to 1993, Clinton to 2001, Bush to 2009, Obama to 2017, Trump to 2021 and Biden to 2025. Show data lists the president and party for each survey year.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)

What these data cannot show

The gaps are associations between party and an answer in one survey year. They do not show whether an attitude drew people to a party, whether party led people to change their answers, or whether new voters and departing voters changed the makeup of each party. Panel studies address direction. Margolis [7] found that for many Americans partisanship shapes religious identity, and Engelhardt [8] found White respondents aligning their racial attitudes with their party. The Chapter 1.4, Cohort change in partisan sorting chapter describes the change in party and ideology by birth cohort and cannot show whether any individual switched.

A sort date is the year a survey-based gap passes a threshold. It is not the year party elites took positions, and the ANES cannot show it. The dates also 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.

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

Methods

Sample: respondents of the ANES Cumulative Data File in the study years 1980 to 2024 who identify with or lean toward a party (pid_7pt 1 to 3 Democratic, 5 to 7 Republican) and answered the item. Respondents with a positive weight in those study years, before any item is required: 49,597 in the all view, 26,570 in the face-to-face view and 35,623 in the White view. Items: abortion (4-category legality, 17 study years), racial resentment (mean of four standardized Kinder-Sanders items, at least two answered, 13 years; not asked in 1996, two items in 1998), importance of religion (yes or no, 17 years) and the retrospective national economy (worse, same, better, 18 years). Each attitude is scored before estimation so that a higher value means more restrictive on abortion, more resentment on the race items, religion more important and the economy better; the direction comes from the stored value labels, not from a correlation with party. The 2024 importance-of-religion item was revised to a five-point scale; the cumulative file codes extremely, very and moderately important as yes, which matches the 2024 Time Series on the same collapse. Estimand: the party gap, the weighted mean of the within-year z-score among Republican identifiers minus the weighted mean among Democratic identifiers, where the z-score uses the weighted mean and standard deviation of all respondents with the item in that year and sample. Weights and samples: the ANES analysis weight (VCF0009z). The face-to-face view uses weight_ftf from 2008 (interviews before 2008 are in person by design, apart from a few abbreviated telephone post-election interviews in 1984). It leaves out 2000, a study that mixed face-to-face and telephone interviews and has no interview-form variable in the file, 2002, which was conducted entirely by telephone, and 2020, which has no mode weights. The mode check uses weight_ftf and weight_web for 2012, 2016 and 2024, standardizing each mode within itself. The White view keeps White non-Hispanic respondents and standardizes within them. Intervals: 95% normal intervals from the Kish effective sample size, combining the standard errors of the two party means in quadrature. They treat the weighted sample as if it were drawn without clustering and treat the standard deviation used for standardizing as a fixed number, so they ignore the ANES strata and clusters and may be narrower or wider than design-based intervals. Suppression: a study year enters a series with 150 respondents in the two parties and an effective sample of 100 in the smaller party; a party-share cell needs 100 respondents and an effective sample of 100. Sort dates: the rule is a party gap of 0.25 SD, carried over from an earlier analysis of the same surveys and kept; the sensitivity rules are 0.15, 0.35 and 0.50 SD. The point-estimate date is the first study year whose estimate exceeds the rule and stays above it in every later study year. The sort date is the first study year whose lower 95% bound exceeds the rule and stays above it in every later study year. The upper-bound year, the earliest date the interval allows, is the first study year whose upper 95% bound exceeds the rule and stays above it. A date in the last study year cannot be tested for staying above the rule. GSS and CES: the GSS extract uses weight_use, party 0 to 2 against 4 to 6, abortion for any reason (no), the item on whether Black Americans should work their way up without special favors (reversed) and church attendance, 1977 to 2024; GSS interviews from 2021 include web respondents, and the 2024 face-to-face and web gaps are reported for each item. The CES uses weight_cumulative, even years only, party 1 to 3 against 5 to 7, always-allow abortion (2014 to 2020) and importance of religion (2008 to 2024).

References

Cohen, Jacob (1988). Statistical Power Analysis for the Behavioral Sciences, 2nd edition. Lawrence Erlbaum Associates.
Adams, Greg D. (1997). Abortion: Evidence of an Issue Evolution. American Journal of Political Science 41(3): 718-737. doi:10.2307/2111673
Carmines, Edward G.; Stimson, James A. (1989). Issue Evolution: Race and the Transformation of American Politics. Princeton University Press. doi:10.1515/9780691218250
Layman, Geoffrey C. (2001). The Great Divide: Religious and Cultural Conflict in American Party Politics. Columbia University Press.
Kinder, Donald R.; Sanders, Lynn M. (1996). Divided by Color: Racial Politics and Democratic Ideals. University of Chicago Press.
Tesler, Michael (2016). Post-Racial or Most-Racial? Race and Politics in the Obama Era. University of Chicago Press. doi:10.7208/chicago/9780226353159.001.0001
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
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.}
}