Chapter 5.2 · Report

Tolerance of speech

In 2008 to 2018 Democrats allowed a racist speaker 6.8 points less often than Republicans did, while the party gaps on four speakers that conservatives are expected to object to were within 2.1 points of zero.

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

Every estimate of the chapter with its interval and every figure as a table, generated from the chapter's data files. Table n of this report is figure n of the chapter. The same report as Markdown: report.md.

Identification

Title

Tolerance of speech

Chapter

5.2 (Part V, Institutions and democratic attitudes; Coalitions and political differences)

Question

Whom would Americans allow to speak in public, and how do the answers differ by party and ideology?

Dek

Grain

individuals

Status

draft

Published

30 September 2026

Updated

not revised

Data sources

General Social Survey. General Social Survey cumulative file 1972-2024, allow-to-speak items (spk), college-teaching items (col) and library-book items (lib); interview mode from the raw file; years: 1972-2024 (35 study years); weights: weight_use (wtssnrps from 2004, wtssps before)

Canonical URL

https://vishalsingh.org/politics/tolerance-of-speech

Markdown

https://vishalsingh.org/politics/tolerance-of-speech/report.md

BibTeX

@misc{singh2026toleranceofspeech,
  author = {Singh, Vishal},
  title = {Tolerance of speech},
  year = {2026},
  url = {https://vishalsingh.org/politics/tolerance-of-speech},
  note = {In The Public Mind. Published 30 September 2026.}
}

Principal estimates

The facts the dek and the opening prose cite, in order of appearance. Every fact is in the table of all estimates below.

Estimand	Population	Period	Value	Interval	n	Weights	Source	Fact id
difference in weighted shares who would allow the racist speaker to speak, Republicans minus Democrats (the party gap with its sign reversed)	Democrats and Republicans (leaners included) answering the item	2008-2018 waves pooled	6.8	95% CI 4.0 to 9.7	7,054	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey	rac_rep_lead_late
largest absolute party gap among the Communist, anti-religionist, homosexual man and Muslim clergyman items	Democrats and Republicans (leaners included) answering the item	2008-2018 waves pooled	2.1			weight_use (wtssnrps from 2004, wtssps before)	General Social Survey	max_abs_pg_other
weighted share who would allow the racist speaker to speak	Democrats answering the racist item (leaners included)	2008-2018 waves pooled	56.1%		4,175	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey	rac_dem_late
weighted share who would allow the racist speaker to speak	Republicans answering the racist item (leaners included)	2008-2018 waves pooled	63.0%		2,879	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey	rac_rep_late
difference in weighted shares who would allow the speaker to speak in the community (Democrats minus Republicans)	Democrats and Republicans (leaners included) answering the item	2008-2018 waves pooled	−6.8	95% CI −9.7 to −4.0	7,054	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey	pg_rac_late
difference in weighted shares who would allow the speaker to speak in the community (Democrats minus Republicans)	Democrats and Republicans (leaners included) answering the item	2008-2018 waves pooled	−3.8	95% CI −6.4 to −1.1	7,069	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey	pg_mil_late
difference between the racist and militarist party gaps in the share who would allow the speech (same period, standard errors in quadrature)	Democrats and Republicans (leaners included) answering the item	2008-2018 waves pooled	−3.1	95% CI −6.9 to 0.8		weight_use (wtssnrps from 2004, wtssps before)	General Social Survey	pgdiff_rac_mil_late
smallest ideology gap among the Communist, anti-religionist, homosexual man and Muslim clergyman items	liberals and conservatives (self-placed) answering the item	2008-2018 waves pooled	7.2			weight_use (wtssnrps from 2004, wtssps before)	General Social Survey	ig_con_min
largest ideology gap among the Communist, anti-religionist, homosexual man and Muslim clergyman items	liberals and conservatives (self-placed) answering the item	2008-2018 waves pooled	13.7			weight_use (wtssnrps from 2004, wtssps before)	General Social Survey	ig_con_max
smallest absolute party gap among the five speakers asked	Democrats and Republicans (leaners included) answering the item	1976-1994 waves pooled	2.4			weight_use (wtssnrps from 2004, wtssps before)	General Social Survey	pg_early_abs_min
largest absolute party gap among the five speakers asked	Democrats and Republicans (leaners included) answering the item	1976-1994 waves pooled	6.3			weight_use (wtssnrps from 2004, wtssps before)	General Social Survey	pg_early_abs_max
smallest adjusted party gap among the Communist, anti-religionist, homosexual man and Muslim clergyman items (adjusted for age, education, race, sex and survey year)	Democrats and Republicans (leaners included) answering the item	2008-2018 waves pooled	3.0			weight_use (wtssnrps from 2004, wtssps before)	General Social Survey	adjpg_con_min
largest adjusted party gap among the Communist, anti-religionist, homosexual man and Muslim clergyman items (adjusted for age, education, race, sex and survey year)	Democrats and Republicans (leaners included) answering the item	2008-2018 waves pooled	5.8			weight_use (wtssnrps from 2004, wtssps before)	General Social Survey	adjpg_con_max

Findings

Whether Democrats or Republicans allowed a speaker more often depended on the speaker. In 2008 to 2018, 56.1% of Democrats and 63.0% of Republicans said that a person who believes Blacks are genetically inferior should be allowed to speak in their community. The party gap, Democrats minus Republicans, was −6.8 (95% CI −9.7 to −4.0) points (a negative gap means that Republicans allowed the speaker more often), the largest in size of the six speakers in the General Social Survey (GSS) battery, a set of questions asked in the same wording across survey years. The militarist gap was −3.8 (95% CI −6.4 to −1.1), and the interval on the difference between the racist and militarist gaps (−3.1 (95% CI −6.9 to 0.8)) includes zero. The party gaps for the Communist, the anti-religionist, the homosexual man and the Muslim clergyman, the four speakers that the chapter codes as objectionable to conservatives, had point estimates within 2.1 points of zero. Ideology separates the sides on those four speakers, and liberals allowed them 7.2 to 13.7 points more often than conservatives did. The section on measurement gives the wording of each item.

In 1976 to 1994 the party gap was negative for every speaker asked, so Republicans allowed each one more often than Democrats did, by 2.4 to 6.3 points. Among people of the same age, education, race and sex, Democrats allowed the Communist, the anti-religionist, the homosexual man and the Muslim clergyman 3.0 to 5.8 points more often than Republicans did in 2008 to 2018. These comparisons also change with the action asked about, the interview mode and the wave that opens the 2008 to 2018 window, and the sections below report each check beside its evidence.

Measurement: the six speakers and the three actions

The battery asks about six speakers, listed in Table 1. Two are people whom liberals are expected to object to, a racist and a militarist who wants the army to run the country. Four are people whom conservatives are expected to object to: a Communist, someone against all churches and religion (called the anti-religionist here), a man who admits he is a homosexual and a Muslim clergyman who preaches hatred of the United States. The chapter calls the first pair the speakers objectionable to liberals and the second group the speakers objectionable to conservatives. The grouping is a coding choice, drawing on work that finds people on each side less tolerant of targets linked with the other side [1], and the speaker-level figures let a reader regroup the speakers. Respondents are asked whether each speaker should be allowed to speak in their community. The question is whether he should be allowed to speak, not whether the respondent agrees with him. In all, 44,878 respondents gave 217,422 answers in 31 survey years.The design descends from Stouffer's national survey of 1954 [2]. A fixed list of speakers measures tolerance of groups that some respondents do not mind, so Sullivan, Piereson and Marcus asked each respondent to name the group he or she likes least and then asked about that group [3]. The GSS keeps the fixed list, which makes the answers comparable across years and across groups of respondents.

Table 1. The six speaker items began between 1972 and 2008, and the militarist and homosexual man items ended in 2021

Two more questions ask about the same six speakers: whether a book the speaker wrote should be removed from the public library, and whether he should be allowed to teach in a college. For each action the tolerant answer is the one that lets the speaker speak, teach or stay on the shelf, and each action is analyzed separately. The Communist teaching item is worded differently from the other teaching items, as a question about firing him from a college post, and the section on actions returns to it.

Groups, gaps and interview mode

Party groups Democrats and Republicans with their leaners. Ideology uses the seven-point self-placement, with liberals at the three left points and conservatives at the three right points, and Planned, The conservative label (planned) asks who takes the conservative label. Independents and moderates drop out of the group comparisons and stay in the all-adult shares. A gap is the first group minus the second, so a negative party gap means that Republicans allowed the speaker more often.Shares are weighted with the survey weights. Intervals are 95% confidence intervals built from the effective sample size [4], which treat the sample as a weighted simple random sample and ignore clustering, so they can be narrower or wider than intervals that follow the sampling design. An annual cell needs 150 respondents and an effective sample of 100.

Interviews before 2021 were conducted by an interviewer, mostly in person. In 2021, 86% of the respondents who answered the racist item were interviewed on the web, and in 2022 and 2024 51% were. The survey mode can change answers to sensitive questions [5], and answers to one speaker item differ by mode in 2022 and 2024. Estimates from 2021 on therefore carry a mode label, and the mode split appears where the sample allows one. Chapter 5.1, Confidence in thirteen institutions meets the same break in the GSS confidence items.

Tolerance rose for most speakers, but not for the racist

In 2008 to 2018 a majority of adults would allow every speaker except the Muslim clergyman, for whom the share was 42.5% (95% CI 41.3% to 43.8%). Table 2 sets the shares against 1976 to 1994. The share rose by 6.3 to 15.2 points for the militarist, the Communist, the anti-religionist and the homosexual man, and changed by −3.2 (95% CI −4.7 to −1.7) points for the racist.

Table 2. Tolerance rose for most speakers between 1976 to 1994 and 2008 to 2018, but not for the racist

Republican leads shrank for three speakers

Table 3 plots the share allowing each speaker in every survey year. For the Communist, the anti-religionist and the homosexual man, the Republican line usually lay above the Democratic line in the 1970s and 1980s, and the two lines cross each other in the years since 2008. For the militarist and the racist the Republican estimate is above the Democratic estimate in every survey year from 2010, though single-year intervals are wide.

Table 3. Party lines converged for three speakers and stayed apart for the racist and the militarist

Table 4 reduces the annual series to two periods. A negative party gap is a Republican lead, so a gap that moves toward zero is a shrinking lead. In 1976 to 1994 every party gap was negative and every interval lay below zero, from −2.4 for the homosexual man to −6.3 for the anti-religionist. By 2008 to 2018 the Communist gap was +0.8 (95% CI −1.9 to 3.5), the anti-religionist gap −1.4 (95% CI −3.8 to 1.0) and the Muslim clergyman gap +0.5 (95% CI −2.4 to 3.4), all with intervals that include zero, and the homosexual man gap was +2.1 (95% CI 0.1 to 4.0). The change from the earlier period was +5.1 (95% CI 1.8 to 8.4) points for the Communist, +4.9 (95% CI 1.9 to 7.8) for the anti-religionist and +4.5 (95% CI 1.9 to 7.0) for the homosexual man.

Table 4. Republicans allowed every speaker more often in 1976 to 1994, and by 2008 to 2018 only the racist and militarist leads remained

The two speakers whom liberals are expected to object to kept a Republican lead. The militarist gap was −4.6 in 1976 to 1994 and −3.8 (95% CI −6.4 to −1.1) in 2008 to 2018, a change of +0.8 (95% CI −2.4 to 4.0) points, an interval that includes zero. The racist gap was −4.5 and −6.8, a change of −2.4 (95% CI −5.8 to 1.0) points, an interval that includes zero. The Republican leads that shrank are the leads on speakers whom conservatives are expected to object to.

The 2008 wave, the first wave of the Muslim clergyman item, opens the 2008 to 2018 window. In that wave alone the party gaps for the racist, the Communist, the anti-religionist, the homosexual man and the Muslim clergyman ran from +1.6 to +7.6 points, and only the militarist gap was negative (−3.6). The party view without 2008 in Table 4 leaves that wave out. In 2010 to 2018 the racist change from 1976 to 1994 is −3.9 (95% CI −7.5 to −0.3) points, an interval that excludes zero, the anti-religionist gap has an interval below zero, and the homosexual man gap has an interval that includes zero. The verdicts on the racist change, the anti-religionist gap and the homosexual man gap therefore depend on one wave, while the rise for the Communist and the persistence of the racist and militarist gaps do not.

Ideology separates the sides on four speakers

The ideology gap, liberals minus conservatives, follows a different pattern. In 2008 to 2018 it was +10.3 (95% CI 7.4 to 13.2) points for the Communist, +7.2 (95% CI 4.7 to 9.8) for the anti-religionist, +7.9 (95% CI 5.9 to 9.9) for the homosexual man and +13.7 (95% CI 10.4 to 17.0) for the Muslim clergyman. These gaps are close to the ones in 1976 to 1994: +11.0 for the Communist, +8.5 for the anti-religionist and +10.5 for the homosexual man, and the changes have intervals that include zero. The Muslim clergyman item has the largest ideology gap of the six, and the interval on its difference from the Communist gap (+3.4 (95% CI −1.0 to 7.7)) includes zero. The party gap for each of these four speakers is smaller than the ideology gap. The militarist follows the same pattern at a lower level, with an ideology gap of +6.3 (95% CI 3.4 to 9.2) in 2008 to 2018.

Switch Table 3 or Table 4 to the ideology view to see the difference. Party identification with leaners includes moderates, which self-placed liberals and conservatives do not, and the groups differ in age, education, race and sex, which the next section holds fixed. The racist speaker is the one case in which the liberal lead is absent in 2008 to 2018. The ideology gap was +6.2 (95% CI 4.1 to 8.4) in 1976 to 1994 and +1.0 (95% CI −2.2 to 4.2) in 2008 to 2018, a change of −5.2 (95% CI −9.1 to −1.3) points.

Among similar adults, Democrats lead on four speakers

The raw party gaps mix party with the traits that differ between the parties. The adjusted views of Table 4 compare Democrats with Republicans of the same age, education, race and sex in the same survey year, using a weighted linear probability model with robust standard errors. In 1976 to 1994 the adjusted party gaps are within 1.7 points of zero for all five speakers asked, against raw Republican leads of 2.4 to 6.3 points. The raw Republican lead of that period is therefore smaller among people of the same age, education, race and sex. In 2008 to 2018 the adjusted gaps for the Communist, the anti-religionist, the homosexual man and the Muslim clergyman run from 3.0 to 5.8 points, each with an interval above zero, where the raw gaps were smaller. The adjusted racist gap is −3.9 (95% CI −6.9 to −0.8) points, against −1.3 (95% CI −3.1 to 0.6) in 1976 to 1994, and the adjusted militarist gap is −0.3 (95% CI −2.9 to 2.4).

The ideology gaps change little with adjustment. For the four speakers that conservatives are expected to object to they run from 6.7 to 13.4 points in 2008 to 2018. The adjustment holds four traits fixed and no others.

Web respondents allowed the racist speaker less often

Table 5 follows the party and ideology gaps on the racist speaker year by year, one view for each. The party estimate is below zero in 27 of 28 survey years since 1976, with 2008 the exception, and 7 of the annual intervals lie entirely below zero. The estimate is farthest below zero in 2021, at −22.9 (95% CI −30.3 to −15.4), which differs from the 2008 to 2018 gap by −16.0 (95% CI −24.0 to −8.0) points. The 2021 wave was mostly web. The ideology estimate is positive in most years before 2010, around zero in the 2010s and negative in 2021, 2022 and 2024. The intervals for 2021 and 2022 lie below zero, and the interval for 2024 includes zero.

Table 5. Republicans allowed the racist speaker more often than Democrats did in 27 of 28 survey years, with the largest estimated leads in 2021, 2022 and 2024

Table 6 splits 2022 and 2024 by mode. Interviewer-administered respondents were interviewed in person, by phone or in a residual mode that the file does not label. In the web sample 40.7% (95% CI 36.5% to 44.9%) would allow the racist speaker, and in the interviewer-administered sample 54.7% (95% CI 50.0% to 59.4%), a difference of −14.0 (95% CI −20.3 to −7.7) points. The difference is similar in both parties: −16.4 (95% CI −26.2 to −6.5) points among Democrats and −12.6 (95% CI −23.9 to −1.3) among Republicans. For the Communist, the anti-religionist and the Muslim clergyman the differences in the share allowing the speech have intervals that include zero.

Table 6. Web respondents allowed the racist speaker 14.0 points less often; the other three speakers show no detectable mode difference

The party gap on the racist speaker is negative in both modes: −11.8 (95% CI −23.0 to −0.5) among interviewer-administered respondents and −15.5 (95% CI −25.4 to −5.6) on the web, and the difference between the two, −3.7 (95% CI −18.7 to 11.2), includes zero. Pooled across modes, the 2022 to 2024 gap is −14.2 (95% CI −21.9 to −6.5), which differs from the 2008 to 2018 gap by −7.3 (95% CI −15.5 to 0.9) points, an interval that includes zero. Every estimate reported here for 2021 to 2024 is more negative than the 2008 to 2018 gap, but only the 2021 wave on its own separates from it, and the data do not separate mode from period. The mode difference that the data can detect is in the share allowing the racist speaker, not in the size of the gap.

The Party gap and Ideology gap views of Table 6 show the other speakers and the ideology gap by mode. The mode difference in the party gap has an interval that excludes zero only for the Muslim clergyman, at +15.5 points, with a lower bound of 0.4 points that is close to zero. The design-based interval in the last section includes zero. That difference sits on the Republican side: the web minus interviewer difference in the share allowing the Muslim clergyman is −12.1 (95% CI −23.3 to −0.9) points among Republicans and +3.4 (95% CI −6.6 to 13.5) among Democrats. The racist ideology gap is negative in both modes, although the interviewer-administered interval includes zero, and the difference between modes, −12.2 (95% CI −28.4 to 3.9) points, includes zero. The data cannot say whether web respondents answer differently or are different people.

The average gap depends on which speakers are counted

In 2008 to 2018 the ideology gap was +8.3 (95% CI 6.3 to 10.2) points for the speakers objectionable to conservatives and +3.5 (95% CI 1.0 to 6.1) for the speakers objectionable to liberals. Table 7 averages each respondent's answers over the five speakers that were asked in every period, split into these two families. In 1976 to 1994 the two values were +9.7 and +7.7. The gap for the first family changed by −1.4 (95% CI −4.0 to 1.2) points and the gap for the second family by −4.2 (95% CI −7.3 to −1.0), and the difference between the two changes, −2.7 (95% CI −6.9 to 1.4), includes zero.

One account links conservatism to needs for order and certainty [6]. Another holds that liberals and conservatives are each intolerant of the groups that threaten their values [7], and studies that vary the target find that conservatism predicts intolerance of left-wing targets and liberalism of right-wing targets [1]. The GSS battery cannot test the dispositions behind either account, but it shows how a summary depends on the speakers it counts.

The second family has two speakers, and the racist item has the larger change in size, −5.2 (95% CI −9.1 to −1.3) points, against −2.9 (95% CI −6.6 to 0.7) for the militarist item, an interval that includes zero. The party gap for the first family went from −4.2 (95% CI −5.7 to −2.8) to +0.4 (95% CI −1.4 to 2.3), and for the second family it stayed negative, at −4.5 and −5.2.

Table 7. By 2008 to 2018 the party gap on speakers that conservatives object to was +0.4 points, and on speakers that liberals object to −5.2

A single average lies between the family values, and where it lies depends on the count of speakers on each side. Counting each of the five speakers once gives an ideology gap of +6.3 (95% CI 4.3 to 8.3) in 2008 to 2018, and counting each family once gives +5.8 (95% CI 3.8 to 7.9). Each estimate lies inside the other's interval, so the choice of weighting moves the ideology average by less than its uncertainty. The party gap is less stable. Counting each speaker once gives −1.8 (95% CI −3.7 to 0.0) in 2008 to 2018 and −4.6 (95% CI −9.3 to 0.2) in the 2021 wave, and counting each family once gives −2.4 (95% CI −4.3 to −0.5) and −6.6 (95% CI −11.5 to −1.6). The intervals for the first weighting include zero, and the 2008 to 2018 interval reaches zero at its upper end. The intervals for the second weighting lie below zero, so whether the interval on the pooled party gap excludes zero depends on the weighting.

The battery has three speakers in one family and two in the other, so the first weighting gives the larger family more say, and adding the Muslim clergyman would change the count again. Leaving the homosexual man out of the first family gives an ideology gap of +9.0 (95% CI 6.6 to 11.3) for that family in 2008 to 2018, against +8.3 with him. The 2021 wave is the only one after 2018 with all five speakers, and it is mostly web. In that wave the party gap for the speakers objectionable to liberals was −16.3 (95% CI −22.5 to −10.0) and the ideology gap was −5.0 (95% CI −11.9 to 1.8).

The racist party gap is negative on every action

Table 8 repeats the gap comparison for 2008 to 2018 with the other two actions. For the racist speaker the party gap is negative on all three: −6.8 (95% CI −9.7 to −4.0) for speech, −4.4 (95% CI −7.3 to −1.4) for teaching and −4.9 (95% CI −7.7 to −2.1) for keeping the book in the library. Fewer adults would let the racist teach (46.3%) than would let him speak (58.5%) or would keep his book (63.7%). The racist ideology gap has an interval above zero only for the library: +4.6 (95% CI 1.5 to 7.7) for the library, +1.0 (95% CI −2.2 to 4.2) for speech and +0.2 (95% CI −3.1 to 3.5) for teaching.

Table 8. The racist party gap is negative on speech, teaching and the library, and party gaps on teaching are positive for three other speakers

On teaching, the party gap is positive for three other speakers: +8.4 (95% CI 5.5 to 11.2) for the Communist, +4.2 (95% CI 1.4 to 7.0) for the anti-religionist and +3.7 (95% CI 1.7 to 5.7) for the homosexual man, each with an interval above zero. Their speech gaps were +0.8 (95% CI −1.9 to 3.5), −1.4 (95% CI −3.8 to 1.0) and +2.1 (95% CI 0.1 to 4.0). The Communist teaching item asks whether a college teacher should be fired, a different wording from the other teaching items. The anti-religionist and homosexual man items use the standard wording, so wording does not account for all of the difference. The small party gaps on speech therefore do not carry over to teaching. For the four speakers objectionable to conservatives the ideology gap is positive on every action. It runs from 9.0 to 18.4 points for teaching and from 12.1 to 14.8 for the library.

What these data cannot show

The gaps are differences between groups in one survey question, so they do not show why the groups differ, whether party or ideology shapes tolerance, or whether tolerance shapes party or ideology. The adjusted gaps hold age, education, race and sex fixed and no other trait, and they assume a linear model, so they compare similar people and do not test what party or ideology does. Periods pool many waves, so a period value averages over years, and the annual figures have wide intervals.

The intervals use the effective sample size and ignore the sampling design. Across the 64 period gaps and mode contrasts checked, standard errors that use the survey strata and primary sampling units were between 0.87 and 1.27 times the published ones, so the published intervals are sometimes narrower and sometimes wider. The annual cells, the adjusted gaps and the family gaps were not checked this way. Two verdicts sit near the edge. The homosexual man party gap in 2008 to 2018 is +2.1 (95% CI 0.1 to 4.0) with the published interval and +2.1 (95% CI 0.0 to 4.1) with a design-based one, and the Muslim clergyman mode difference in the party gap is +15.5 (95% CI 0.4 to 30.6) with the published interval and +15.5 (95% CI −3.0 to 34.0) with a design-based one, which includes zero.

The mode of interview is tied to time after 2020. The 2021 wave has too few interviewer-administered respondents to split, and the 2022 and 2024 samples are small once cut by party or ideology, so the size of the racist gap after 2020 is uncertain. Two speaker items were not asked after 2021, so the militarist and the homosexual man cannot be followed into the web era.

The battery has limits as a measure of tolerance. It asks about a fixed list of speakers, and the racist and Muslim clergyman items describe the speech itself, so the levels of the items are not comparable with one another. Grouping the speakers by which side is expected to object is a coding choice. Willingness to let a speaker speak in principle is not the same as behavior toward a speaker, and the militarist item asks about speech, not about support for army rule, which Chapter 5.3, Support for democracy takes up. Willingness to justify political violence is a separate question, taken up in Chapter 5.4, Political violence and democratic norms. The data hold no measure of dogmatism or need for closure, so they cannot say whether one side is more rigid in the sense of the psychological literature.

Tables

Table 1. The six speaker items began between 1972 and 2008, and the militarist and homosexual man items ended in 2021

Item wording, the side expected to object to each speaker in this chapter's coding, GSS variable and first and last survey year

Speaker	Item wording	Expected to object	GSS variable	First asked	Last asked
Racist	A person who believes that Blacks are genetically inferior	Liberals	spkrac	1976	2024
Militarist	A person who advocates doing away with elections and letting the military run the country	Liberals	spkmil	1976	2021
Communist	A man who admits he is a Communist	Conservatives	spkcom	1972	2024
Anti-religionist	Someone who is against all churches and religion	Conservatives	spkath	1972	2024
Homosexual man	A man who admits that he is a homosexual	Conservatives	spkhomo	1973	2021
Muslim clergyman	A Muslim clergyman who preaches hatred of the United States	Conservatives	spkmslm	2008	2024

Source: General Social Survey, 1972-2024 cumulative file; item wording and years asked, so no estimates and no weights needed

Data: vishalsingh.org/politics/report/tolerance-of-speech/figures/speaker_items.json

Table 2. Tolerance rose for most speakers between 1976 to 1994 and 2008 to 2018, but not for the racist

Share of all adults who would allow each speaker to speak in their community, independents included, by period, with 95% intervals

Speaker	Period	Share allowing (%)	Low	High	Respondents
Homosexual man	1976 to 1994	72.5	71.7	73.3	17,535
Homosexual man	2008 to 2018	87.7	86.8	88.5	8,964
Anti-religionist	1976 to 1994	69.5	68.7	70.3	17,936
Anti-religionist	2008 to 2018	78.3	77.3	79.4	9,025
Militarist	1976 to 1994	59.0	58.1	59.8	17,693
Militarist	2008 to 2018	70.0	68.8	71.2	8,952
Communist	1976 to 1994	62.4	61.6	63.3	17,660
Communist	2008 to 2018	68.7	67.5	69.9	8,806
Racist	1976 to 1994	61.7	60.9	62.6	17,711
Racist	2008 to 2018	58.5	57.3	59.8	8,955
Muslim clergyman	2008 to 2018	42.5	41.3	43.8	8,970

Source: General Social Survey, 1972-2024 cumulative file; weighted (weight_use); 95% intervals on the effective sample size

Data: vishalsingh.org/politics/report/tolerance-of-speech/figures/allow_levels.json

Table 3. Party lines converged for three speakers and stayed apart for the racist and the militarist

Share who would allow each speaker to speak in their community, by survey year, with 95% intervals

Speaker	Year	Group	Share allowing (%)	Low	High	Respondents
Racist	1976	Democrats	59.7	56.0	63.5	818
Racist	1976	Republicans	65.9	60.9	71.0	407
Racist	1976	Liberals	70.7	65.7	75.7	399
Racist	1976	Conservatives	63.3	58.3	68.3	431
Racist	1977	Democrats	59.0	55.3	62.7	859
Racist	1977	Republicans	62.4	57.5	67.3	451
Racist	1977	Liberals	67.6	62.5	72.6	414
Racist	1977	Conservatives	60.2	55.4	65.1	463
Racist	1980	Democrats	61.3	56.9	65.6	739
Racist	1980	Republicans	64.3	58.9	69.6	445
Racist	1980	Liberals	71.7	66.0	77.5	360
Racist	1980	Conservatives	65.1	59.8	70.3	470
Racist	1982	Democrats	57.8	53.8	61.7	1,074
Racist	1982	Republicans	64.4	59.0	69.7	487
Racist	1982	Liberals	71.7	66.5	77.0	503
Racist	1982	Conservatives	62.5	57.3	67.8	519
Racist	1984	Democrats	56.6	52.1	61.0	740
Racist	1984	Republicans	63.1	58.0	68.3	517
Racist	1984	Liberals	65.2	58.8	71.6	335
Racist	1984	Conservatives	60.7	55.4	65.9	495
Racist	1985	Democrats	56.4	51.9	61.0	737
Racist	1985	Republicans	61.5	56.7	66.3	599
Racist	1985	Liberals	66.7	60.3	73.2	361
Racist	1985	Conservatives	57.9	52.9	63.0	519
Racist	1987	Democrats	61.5	57.7	65.3	1,001
Racist	1987	Republicans	61.8	57.1	66.5	557
Racist	1987	Liberals	68.8	63.8	73.9	498
Racist	1987	Conservatives	59.2	54.1	64.2	520
Racist	1988	Democrats	58.6	53.6	63.6	474
Racist	1988	Republicans	64.3	58.8	69.7	361
Racist	1988	Liberals	66.4	59.9	72.8	267
Racist	1988	Conservatives	60.6	54.7	66.5	327
Racist	1989	Democrats	59.2	54.1	64.3	467
Racist	1989	Republicans	62.4	57.2	67.7	409
Racist	1989	Liberals	66.3	60.1	72.4	276
Racist	1989	Conservatives	67.2	61.2	73.2	297
Racist	1990	Democrats	60.1	54.5	65.8	371
Racist	1990	Republicans	68.8	63.7	73.9	406
Racist	1990	Liberals	66.0	58.8	73.2	228
Racist	1990	Conservatives	66.5	60.9	72.2	321
Racist	1991	Democrats	60.4	55.0	65.7	421
Racist	1991	Republicans	64.6	59.6	69.7	412
Racist	1991	Liberals	66.6	60.2	73.1	261
Racist	1991	Conservatives	68.5	62.6	74.5	300
Racist	1993	Democrats	62.3	57.4	67.1	497
Racist	1993	Republicans	63.9	58.6	69.2	405
Racist	1993	Liberals	71.5	65.5	77.6	271
Racist	1993	Conservatives	62.0	56.3	67.8	354
Racist	1994	Democrats	60.9	57.4	64.5	925
Racist	1994	Republicans	65.2	61.3	69.0	736
Racist	1994	Liberals	68.7	64.2	73.1	534
Racist	1994	Conservatives	62.0	57.8	66.1	676
Racist	1996	Democrats	59.1	55.4	62.8	855
Racist	1996	Republicans	63.0	59.0	66.9	699
Racist	1996	Liberals	65.2	60.2	70.1	444
Racist	1996	Conservatives	62.5	58.4	66.6	650
Racist	1998	Democrats	60.0	56.3	63.7	865
Racist	1998	Republicans	64.5	60.2	68.7	617
Racist	1998	Liberals	69.9	65.5	74.3	531
Racist	1998	Conservatives	61.7	57.3	66.2	594
Racist	2000	Democrats	58.1	54.2	62.0	813
Racist	2000	Republicans	63.1	58.9	67.3	616
Racist	2000	Liberals	66.4	61.5	71.2	459
Racist	2000	Conservatives	64.6	60.3	68.9	574
Racist	2002	Democrats	62.5	57.0	68.1	401
Racist	2002	Republicans	65.7	59.7	71.7	320
Racist	2002	Liberals	69.7	62.7	76.7	224
Racist	2002	Conservatives	60.9	54.4	67.3	298
Racist	2004	Democrats	59.8	54.1	65.4	397
Racist	2004	Republicans	65.3	59.5	71.1	330
Racist	2004	Liberals	61.6	53.9	69.4	219
Racist	2004	Conservatives	64.8	58.9	70.8	311
Racist	2006	Democrats	62.9	58.8	66.9	846
Racist	2006	Republicans	64.4	60.1	68.7	666
Racist	2006	Liberals	69.7	64.7	74.6	516
Racist	2006	Conservatives	59.8	55.3	64.2	674
Racist	2008	Democrats	60.6	55.9	65.2	630
Racist	2008	Republicans	59.0	53.6	64.3	440
Racist	2008	Liberals	66.4	60.6	72.1	354
Racist	2008	Conservatives	57.2	51.5	62.9	418
Racist	2010	Democrats	53.2	48.2	58.3	579
Racist	2010	Republicans	60.9	55.1	66.7	411
Racist	2010	Liberals	58.9	52.5	65.4	351
Racist	2010	Conservatives	60.4	54.6	66.2	397
Racist	2012	Democrats	55.2	50.3	60.1	590
Racist	2012	Republicans	62.0	55.8	68.1	393
Racist	2012	Liberals	59.1	52.9	65.4	357
Racist	2012	Conservatives	60.1	54.3	66.0	408
Racist	2014	Democrats	55.1	51.0	59.3	786
Racist	2014	Republicans	67.1	62.5	71.7	525
Racist	2014	Liberals	59.8	54.7	65.0	465
Racist	2014	Conservatives	65.5	60.9	70.0	538
Racist	2016	Democrats	56.1	52.0	60.2	897
Racist	2016	Republicans	65.2	60.6	69.8	600
Racist	2016	Liberals	62.3	57.1	67.5	512
Racist	2016	Conservatives	59.0	54.2	63.8	603
Racist	2018	Democrats	56.3	51.5	61.0	693
Racist	2018	Republicans	61.8	56.2	67.4	510
Racist	2018	Liberals	63.1	57.1	69.1	438
Racist	2018	Conservatives	60.7	54.9	66.5	461
Racist	2021	Democrats	38.2	33.4	43.0	590
Racist	2021	Republicans	61.1	55.4	66.8	400
Racist	2021	Liberals	42.0	36.2	47.7	429
Racist	2021	Conservatives	57.0	51.2	62.7	406
Racist	2022	Democrats	44.0	36.8	51.1	483
Racist	2022	Republicans	59.0	49.7	68.3	334
Racist	2022	Liberals	41.6	33.6	49.6	366
Racist	2022	Conservatives	58.7	49.0	68.4	331
Racist	2024	Democrats	39.5	32.6	46.3	429
Racist	2024	Republicans	53.0	46.1	59.9	368
Racist	2024	Liberals	48.7	41.0	56.4	322
Racist	2024	Conservatives	53.7	46.9	60.5	361
Militarist	1976	Democrats	55.1	51.3	58.9	824
Militarist	1976	Republicans	58.6	53.4	63.9	406
Militarist	1976	Liberals	68.4	63.3	73.5	398
Militarist	1976	Conservatives	53.2	48.0	58.4	426
Militarist	1977	Democrats	49.3	45.6	53.1	866
Militarist	1977	Republicans	56.9	51.8	61.9	454
Militarist	1977	Liberals	61.2	56.0	66.5	418
Militarist	1977	Conservatives	51.5	46.6	56.5	462
Militarist	1980	Democrats	57.8	53.5	62.2	740
Militarist	1980	Republicans	57.6	52.2	63.1	450
Militarist	1980	Liberals	69.3	63.4	75.1	363
Militarist	1980	Conservatives	57.5	52.2	62.9	471
Militarist	1982	Democrats	52.2	48.2	56.2	1,063
Militarist	1982	Republicans	62.1	56.7	67.4	494
Militarist	1982	Liberals	69.1	63.8	74.5	501
Militarist	1982	Conservatives	56.0	50.6	61.4	520
Militarist	1984	Democrats	57.1	52.7	61.6	743
Militarist	1984	Republicans	58.9	53.5	64.3	518
Militarist	1984	Liberals	64.8	58.0	71.5	339
Militarist	1984	Conservatives	55.6	50.3	60.9	496
Militarist	1985	Democrats	55.1	50.5	59.6	741
Militarist	1985	Republicans	59.1	54.3	64.0	602
Militarist	1985	Liberals	64.5	58.1	71.0	367
Militarist	1985	Conservatives	52.6	47.5	57.7	517
Militarist	1987	Democrats	57.7	53.9	61.6	993
Militarist	1987	Republicans	55.9	51.0	60.7	552
Militarist	1987	Liberals	62.4	57.2	67.7	494
Militarist	1987	Conservatives	58.3	53.2	63.3	516
Militarist	1988	Democrats	57.1	52.1	62.1	472
Militarist	1988	Republicans	59.7	54.2	65.3	358
Militarist	1988	Liberals	69.5	63.2	75.9	260
Militarist	1988	Conservatives	59.0	53.0	64.9	328
Militarist	1989	Democrats	59.2	54.2	64.3	474
Militarist	1989	Republicans	62.4	57.2	67.7	411
Militarist	1989	Liberals	67.5	61.3	73.6	277
Militarist	1989	Conservatives	64.4	58.3	70.6	296
Militarist	1990	Democrats	53.9	48.1	59.6	369
Militarist	1990	Republicans	63.5	58.2	68.8	404
Militarist	1990	Liberals	62.6	55.3	69.9	229
Militarist	1990	Conservatives	60.8	54.9	66.7	318
Militarist	1991	Democrats	60.2	54.9	65.5	421
Militarist	1991	Republicans	66.7	61.7	71.7	413
Militarist	1991	Liberals	73.0	67.0	79.0	267
Militarist	1991	Conservatives	62.5	56.2	68.7	297
Militarist	1993	Democrats	63.8	59.0	68.6	496
Militarist	1993	Republicans	68.1	62.9	73.2	407
Militarist	1993	Liberals	77.3	71.7	82.8	274
Militarist	1993	Conservatives	62.3	56.5	68.1	351
Militarist	1994	Democrats	63.8	60.3	67.3	924
Militarist	1994	Republicans	68.0	64.2	71.7	734
Militarist	1994	Liberals	70.6	66.3	75.0	539
Militarist	1994	Conservatives	65.7	61.7	69.8	674
Militarist	1996	Democrats	60.8	57.0	64.5	854
Militarist	1996	Republicans	65.3	61.4	69.1	704
Militarist	1996	Liberals	70.7	66.0	75.5	446
Militarist	1996	Conservatives	62.2	58.1	66.3	655
Militarist	1998	Democrats	65.7	62.1	69.3	863
Militarist	1998	Republicans	69.2	65.1	73.2	626
Militarist	1998	Liberals	75.8	71.7	80.0	528
Militarist	1998	Conservatives	65.3	61.0	69.6	601
Militarist	2000	Democrats	60.2	56.3	64.0	811
Militarist	2000	Republicans	68.1	64.0	72.2	613
Militarist	2000	Liberals	70.1	65.3	74.8	460
Militarist	2000	Conservatives	62.8	58.4	67.2	569
Militarist	2002	Democrats	68.0	62.7	73.3	406
Militarist	2002	Republicans	69.8	64.1	75.6	323
Militarist	2002	Liberals	75.9	69.5	82.4	227
Militarist	2002	Conservatives	62.5	56.1	68.8	302
Militarist	2004	Democrats	64.5	59.0	70.1	397
Militarist	2004	Republicans	68.1	62.4	73.7	331
Militarist	2004	Liberals	70.8	63.6	78.0	219
Militarist	2004	Conservatives	65.6	59.8	71.5	313
Militarist	2006	Democrats	67.2	63.3	71.2	841
Militarist	2006	Republicans	69.1	64.9	73.2	671
Militarist	2006	Liberals	73.9	69.1	78.6	514
Militarist	2006	Conservatives	63.7	59.4	68.1	676
Militarist	2008	Democrats	65.7	61.1	70.2	629
Militarist	2008	Republicans	69.3	64.3	74.3	442
Militarist	2008	Liberals	73.8	68.4	79.1	355
Militarist	2008	Conservatives	65.3	59.8	70.8	418
Militarist	2010	Democrats	65.6	60.8	70.4	578
Militarist	2010	Republicans	71.9	66.6	77.2	413
Militarist	2010	Liberals	73.9	68.2	79.7	350
Militarist	2010	Conservatives	71.5	66.2	76.9	398
Militarist	2012	Democrats	69.5	65.0	74.0	595
Militarist	2012	Republicans	71.9	66.3	77.6	399
Militarist	2012	Liberals	73.1	67.5	78.8	356
Militarist	2012	Conservatives	70.3	64.9	75.7	413
Militarist	2014	Democrats	70.0	66.2	73.9	784
Militarist	2014	Republicans	72.2	67.8	76.6	525
Militarist	2014	Liberals	77.5	73.1	81.9	460
Militarist	2014	Conservatives	69.8	65.4	74.2	539
Militarist	2016	Democrats	71.1	67.3	74.8	897
Militarist	2016	Republicans	75.8	71.6	79.9	600
Militarist	2016	Liberals	78.7	74.3	83.1	512
Militarist	2016	Conservatives	70.6	66.2	75.0	602
Militarist	2018	Democrats	71.0	66.7	75.3	699
Militarist	2018	Republicans	74.6	69.6	79.6	508
Militarist	2018	Liberals	78.1	73.0	83.2	442
Militarist	2018	Conservatives	71.7	66.3	77.1	457
Militarist	2021	Democrats	57.5	52.6	62.4	584
Militarist	2021	Republicans	67.4	61.9	72.9	400
Militarist	2021	Liberals	66.0	60.5	71.5	430
Militarist	2021	Conservatives	61.0	55.3	66.6	404
Communist	1972	Democrats	52.0	48.4	55.7	884
Communist	1972	Republicans	55.9	51.0	60.8	450
Communist	1973	Democrats	62.7	59.1	66.4	787
Communist	1973	Republicans	59.5	54.8	64.3	472
Communist	1974	Democrats	61.3	57.6	65.0	799
Communist	1974	Republicans	57.3	52.2	62.4	427
Communist	1974	Liberals	73.2	68.5	77.9	423
Communist	1974	Conservatives	57.0	51.8	62.3	404
Communist	1976	Democrats	55.0	51.2	58.8	815
Communist	1976	Republicans	60.9	55.7	66.1	410
Communist	1976	Liberals	70.7	65.7	75.6	399
Communist	1976	Conservatives	53.8	48.6	58.9	426
Communist	1977	Democrats	55.5	51.8	59.3	852
Communist	1977	Republicans	59.7	54.8	64.7	457
Communist	1977	Liberals	69.4	64.5	74.4	415
Communist	1977	Conservatives	54.4	49.5	59.4	459
Communist	1980	Democrats	54.4	50.0	58.9	738
Communist	1980	Republicans	57.6	52.2	63.1	447
Communist	1980	Liberals	66.2	60.2	72.1	363
Communist	1980	Conservatives	54.9	49.4	60.3	472
Communist	1982	Democrats	55.8	51.8	59.8	1,045
Communist	1982	Republicans	63.6	58.3	68.9	492
Communist	1982	Liberals	72.7	67.6	77.8	492
Communist	1982	Conservatives	55.7	50.3	61.1	519
Communist	1984	Democrats	60.6	56.2	64.9	737
Communist	1984	Republicans	61.8	56.5	67.1	518
Communist	1984	Liberals	67.2	60.5	73.9	333
Communist	1984	Conservatives	61.2	56.0	66.4	497
Communist	1985	Democrats	60.5	56.0	65.0	726
Communist	1985	Republicans	60.6	55.8	65.4	603
Communist	1985	Liberals	71.0	64.8	77.3	358
Communist	1985	Conservatives	57.0	52.0	62.1	519
Communist	1987	Democrats	59.1	55.2	62.9	1,004
Communist	1987	Republicans	62.5	57.8	67.2	556
Communist	1987	Liberals	70.3	65.3	75.2	500
Communist	1987	Conservatives	61.5	56.5	66.4	525
Communist	1988	Democrats	60.3	55.3	65.2	475
Communist	1988	Republicans	65.3	59.9	70.7	363
Communist	1988	Liberals	70.6	64.4	76.7	267
Communist	1988	Conservatives	61.4	55.6	67.2	330
Communist	1989	Democrats	64.5	59.6	69.5	466
Communist	1989	Republicans	65.0	59.8	70.1	409
Communist	1989	Liberals	72.8	67.0	78.7	275
Communist	1989	Conservatives	68.4	62.5	74.4	297
Communist	1990	Democrats	63.0	57.4	68.5	373
Communist	1990	Republicans	70.5	65.5	75.5	401
Communist	1990	Liberals	71.1	64.2	77.9	229
Communist	1990	Conservatives	69.2	63.7	74.8	321
Communist	1991	Democrats	68.6	63.6	73.7	417
Communist	1991	Republicans	69.2	64.3	74.1	414
Communist	1991	Liberals	76.9	71.2	82.6	267
Communist	1991	Conservatives	69.3	63.3	75.3	295
Communist	1993	Democrats	71.0	66.5	75.6	491
Communist	1993	Republicans	72.4	67.4	77.3	412
Communist	1993	Liberals	80.0	74.7	85.4	272
Communist	1993	Conservatives	69.4	64.0	74.9	358
Communist	1994	Democrats	66.4	63.0	69.8	924
Communist	1994	Republicans	70.7	66.9	74.4	730
Communist	1994	Liberals	78.1	74.1	82.0	541
Communist	1994	Conservatives	66.1	62.1	70.2	676
Communist	1996	Democrats	64.4	60.8	68.0	851
Communist	1996	Republicans	65.3	61.5	69.2	696
Communist	1996	Liberals	72.0	67.3	76.7	446
Communist	1996	Conservatives	62.9	58.8	67.0	652
Communist	1998	Democrats	65.7	62.1	69.3	864
Communist	1998	Republicans	69.2	65.1	73.3	613
Communist	1998	Liberals	75.3	71.1	79.4	528
Communist	1998	Conservatives	66.1	61.7	70.4	596
Communist	2000	Democrats	63.1	59.3	67.0	798
Communist	2000	Republicans	72.2	68.3	76.1	610
Communist	2000	Liberals	76.3	71.9	80.7	460
Communist	2000	Conservatives	67.0	62.7	71.3	560
Communist	2002	Democrats	69.2	63.9	74.5	402
Communist	2002	Republicans	70.6	64.8	76.3	322
Communist	2002	Liberals	76.8	70.4	83.2	225
Communist	2002	Conservatives	67.5	61.4	73.6	303
Communist	2004	Democrats	68.0	62.7	73.4	396
Communist	2004	Republicans	71.5	65.9	77.0	326
Communist	2004	Liberals	73.1	66.1	80.1	221
Communist	2004	Conservatives	67.8	61.9	73.6	308
Communist	2006	Democrats	69.0	65.2	72.9	839
Communist	2006	Republicans	73.5	69.5	77.4	670
Communist	2006	Liberals	76.9	72.4	81.5	509
Communist	2006	Conservatives	67.1	62.9	71.4	671
Communist	2008	Democrats	70.3	65.9	74.6	625
Communist	2008	Republicans	64.6	59.4	69.8	445
Communist	2008	Liberals	77.1	72.0	82.3	352
Communist	2008	Conservatives	63.8	58.3	69.3	424
Communist	2010	Democrats	66.2	61.4	71.0	565
Communist	2010	Republicans	65.7	60.1	71.3	416
Communist	2010	Liberals	71.7	65.7	77.6	348
Communist	2010	Conservatives	69.2	63.7	74.6	396
Communist	2012	Democrats	69.4	64.8	74.0	583
Communist	2012	Republicans	65.2	59.2	71.2	392
Communist	2012	Liberals	74.7	69.1	80.3	348
Communist	2012	Conservatives	69.1	63.6	74.7	402
Communist	2014	Democrats	67.6	63.7	71.5	775
Communist	2014	Republicans	71.6	67.2	76.0	522
Communist	2014	Liberals	77.5	73.0	81.9	453
Communist	2014	Conservatives	70.3	65.9	74.7	532
Communist	2016	Democrats	74.1	70.5	77.8	874
Communist	2016	Republicans	70.8	66.4	75.2	599
Communist	2016	Liberals	85.5	81.6	89.3	505
Communist	2016	Conservatives	67.3	62.7	71.8	594
Communist	2018	Democrats	69.0	64.6	73.4	685
Communist	2018	Republicans	73.4	68.3	78.6	500
Communist	2018	Liberals	78.3	73.2	83.4	437
Communist	2018	Conservatives	66.6	60.9	72.2	449
Communist	2021	Democrats	73.4	69.1	77.8	588
Communist	2021	Republicans	71.0	65.6	76.3	399
Communist	2021	Liberals	84.2	80.0	88.4	430
Communist	2021	Conservatives	66.9	61.4	72.4	404
Communist	2022	Democrats	68.3	61.6	75.0	479
Communist	2022	Republicans	65.8	56.6	74.9	328
Communist	2022	Liberals	67.7	60.1	75.3	364
Communist	2024	Democrats	66.9	60.3	73.5	425
Communist	2024	Republicans	63.3	56.6	70.0	364
Communist	2024	Liberals	83.1	77.3	88.9	318
Communist	2024	Conservatives	59.9	53.2	66.7	357
Anti-religionist	1972	Democrats	65.1	61.6	68.5	901
Anti-religionist	1972	Republicans	69.7	65.2	74.3	450
Anti-religionist	1973	Democrats	66.1	62.5	69.7	799
Anti-religionist	1973	Republicans	64.8	60.2	69.4	481
Anti-religionist	1974	Democrats	64.5	60.9	68.1	813
Anti-religionist	1974	Republicans	61.0	56.0	66.0	428
Anti-religionist	1974	Liberals	75.8	71.3	80.3	425
Anti-religionist	1974	Conservatives	62.3	57.2	67.4	409
Anti-religionist	1976	Democrats	63.2	59.6	66.9	829
Anti-religionist	1976	Republicans	68.7	63.8	73.6	410
Anti-religionist	1976	Liberals	77.7	73.1	82.3	400
Anti-religionist	1976	Conservatives	67.1	62.2	71.9	431
Anti-religionist	1977	Democrats	62.4	58.7	66.0	868
Anti-religionist	1977	Republicans	65.2	60.4	70.0	463
Anti-religionist	1977	Liberals	73.3	68.6	78.1	417
Anti-religionist	1977	Conservatives	59.7	54.8	64.5	466
Anti-religionist	1980	Democrats	63.5	59.2	67.7	753
Anti-religionist	1980	Republicans	68.5	63.4	73.7	452
Anti-religionist	1980	Liberals	75.3	69.9	80.8	365
Anti-religionist	1980	Conservatives	64.3	59.1	69.5	479
Anti-religionist	1982	Democrats	60.4	56.5	64.3	1,089
Anti-religionist	1982	Republicans	71.8	66.8	76.7	496
Anti-religionist	1982	Liberals	76.1	71.1	81.0	506
Anti-religionist	1982	Conservatives	64.3	59.1	69.5	526
Anti-religionist	1984	Democrats	66.2	62.0	70.4	749
Anti-religionist	1984	Republicans	71.2	66.3	76.2	523
Anti-religionist	1984	Liberals	74.9	68.8	81.0	339
Anti-religionist	1984	Conservatives	69.0	64.1	73.9	500
Anti-religionist	1985	Democrats	63.3	58.9	67.7	747
Anti-religionist	1985	Republicans	73.3	68.9	77.6	605
Anti-religionist	1985	Liberals	71.9	65.8	78.0	367
Anti-religionist	1985	Conservatives	68.2	63.5	73.0	523
Anti-religionist	1987	Democrats	69.5	66.0	73.1	1,016
Anti-religionist	1987	Republicans	70.9	66.5	75.3	563
Anti-religionist	1987	Liberals	76.7	72.2	81.3	502
Anti-religionist	1987	Conservatives	70.2	65.6	74.9	527
Anti-religionist	1988	Democrats	65.2	60.4	69.9	480
Anti-religionist	1988	Republicans	74.2	69.3	79.2	361
Anti-religionist	1988	Liberals	76.1	70.3	81.9	266
Anti-religionist	1988	Conservatives	70.5	65.0	75.9	331
Anti-religionist	1989	Democrats	69.1	64.3	73.8	473
Anti-religionist	1989	Republicans	74.5	69.8	79.2	414
Anti-religionist	1989	Liberals	80.0	74.8	85.3	276
Anti-religionist	1989	Conservatives	70.1	64.3	75.9	301
Anti-religionist	1990	Democrats	72.1	66.9	77.2	375
Anti-religionist	1990	Republicans	75.6	70.9	80.3	409
Anti-religionist	1990	Liberals	78.4	72.2	84.6	230
Anti-religionist	1990	Conservatives	70.6	65.2	76.0	327
Anti-religionist	1991	Democrats	69.5	64.5	74.5	423
Anti-religionist	1991	Republicans	76.4	71.9	80.9	416
Anti-religionist	1991	Liberals	80.1	74.7	85.5	269
Anti-religionist	1991	Conservatives	76.6	71.2	82.0	301
Anti-religionist	1993	Democrats	71.0	66.5	75.5	498
Anti-religionist	1993	Republicans	74.2	69.3	79.0	409
Anti-religionist	1993	Liberals	82.1	77.0	87.2	274
Anti-religionist	1993	Conservatives	69.2	63.7	74.6	356
Anti-religionist	1994	Democrats	71.2	67.9	74.4	940
Anti-religionist	1994	Republicans	75.3	71.8	78.8	736
Anti-religionist	1994	Liberals	78.6	74.7	82.5	543
Anti-religionist	1994	Conservatives	71.3	67.5	75.2	682
Anti-religionist	1996	Democrats	72.1	68.7	75.5	862
Anti-religionist	1996	Republicans	73.4	69.9	77.0	703
Anti-religionist	1996	Liberals	77.0	72.6	81.4	448
Anti-religionist	1996	Conservatives	72.8	69.1	76.6	658
Anti-religionist	1998	Democrats	72.7	69.3	76.1	870
Anti-religionist	1998	Republicans	76.9	73.2	80.6	625
Anti-religionist	1998	Liberals	81.3	77.6	85.1	536
Anti-religionist	1998	Conservatives	75.0	71.1	79.0	597
Anti-religionist	2000	Democrats	71.8	68.3	75.4	813
Anti-religionist	2000	Republicans	79.2	75.7	82.7	615
Anti-religionist	2000	Liberals	82.2	78.2	86.1	459
Anti-religionist	2000	Conservatives	74.5	70.5	78.4	576
Anti-religionist	2002	Democrats	72.5	67.4	77.6	409
Anti-religionist	2002	Republicans	77.7	72.4	82.9	323
Anti-religionist	2002	Liberals	80.5	74.5	86.5	227
Anti-religionist	2002	Conservatives	69.3	63.2	75.3	305
Anti-religionist	2004	Democrats	73.2	68.1	78.3	397
Anti-religionist	2004	Republicans	79.5	74.6	84.5	331
Anti-religionist	2004	Liberals	74.6	67.7	81.5	220
Anti-religionist	2004	Conservatives	77.2	72.0	82.4	312
Anti-religionist	2006	Democrats	76.6	73.1	80.1	848
Anti-religionist	2006	Republicans	82.5	79.0	85.9	672
Anti-religionist	2006	Liberals	84.0	80.1	88.0	515
Anti-religionist	2006	Conservatives	75.3	71.4	79.2	678
Anti-religionist	2008	Democrats	80.2	76.4	84.0	635
Anti-religionist	2008	Republicans	74.2	69.4	78.9	444
Anti-religionist	2008	Liberals	86.9	82.8	91.0	355
Anti-religionist	2008	Conservatives	73.5	68.4	78.5	424
Anti-religionist	2010	Democrats	73.2	68.7	77.7	577
Anti-religionist	2010	Republicans	77.0	72.0	81.9	416
Anti-religionist	2010	Liberals	79.7	74.4	84.9	353
Anti-religionist	2010	Conservatives	77.5	72.5	82.4	401
Anti-religionist	2012	Democrats	78.8	74.8	82.8	593
Anti-religionist	2012	Republicans	76.6	71.3	81.9	401
Anti-religionist	2012	Liberals	84.2	79.5	88.8	357
Anti-religionist	2012	Conservatives	74.5	69.3	79.6	415
Anti-religionist	2014	Democrats	77.6	74.1	81.1	784
Anti-religionist	2014	Republicans	84.2	80.6	87.7	528
Anti-religionist	2014	Liberals	84.0	80.2	87.9	463
Anti-religionist	2014	Conservatives	79.0	75.1	82.9	541
Anti-religionist	2016	Democrats	78.0	74.5	81.4	901
Anti-religionist	2016	Republicans	81.1	77.3	84.8	603
Anti-religionist	2016	Liberals	85.9	82.1	89.6	512
Anti-religionist	2016	Conservatives	77.8	73.8	81.9	604
Anti-religionist	2018	Democrats	81.7	78.0	85.3	704
Anti-religionist	2018	Republicans	82.8	78.5	87.2	512
Anti-religionist	2018	Liberals	85.5	81.1	89.8	444
Anti-religionist	2018	Conservatives	80.2	75.5	84.9	464
Anti-religionist	2021	Democrats	83.4	79.7	87.0	588
Anti-religionist	2021	Republicans	80.9	76.3	85.6	401
Anti-religionist	2021	Liberals	88.8	85.2	92.5	429
Anti-religionist	2021	Conservatives	77.3	72.4	82.1	407
Anti-religionist	2022	Democrats	80.6	74.9	86.3	487
Anti-religionist	2022	Republicans	71.9	63.4	80.5	334
Anti-religionist	2022	Liberals	83.4	77.4	89.4	367
Anti-religionist	2022	Conservatives	75.8	67.4	84.2	331
Anti-religionist	2024	Democrats	80.2	74.6	85.7	432
Anti-religionist	2024	Republicans	77.7	72.0	83.4	371
Anti-religionist	2024	Liberals	88.1	83.1	93.0	323
Anti-religionist	2024	Conservatives	76.5	70.7	82.3	361
Homosexual man	1973	Democrats	64.8	61.2	68.5	775
Homosexual man	1973	Republicans	61.2	56.4	66.0	465
Homosexual man	1974	Democrats	65.7	62.1	69.3	791
Homosexual man	1974	Republicans	62.9	57.9	68.0	413
Homosexual man	1974	Liberals	75.8	71.3	80.4	421
Homosexual man	1974	Conservatives	62.2	57.0	67.5	393
Homosexual man	1976	Democrats	63.3	59.6	67.0	812
Homosexual man	1976	Republicans	68.9	63.9	73.8	401
Homosexual man	1976	Liberals	78.6	74.1	83.1	400
Homosexual man	1976	Conservatives	62.7	57.7	67.7	425
Homosexual man	1977	Democrats	63.2	59.6	66.9	845
Homosexual man	1977	Republicans	66.7	61.8	71.5	446
Homosexual man	1977	Liberals	73.3	68.5	78.1	409
Homosexual man	1977	Conservatives	58.8	53.8	63.7	450
Homosexual man	1980	Democrats	67.1	62.9	71.3	740
Homosexual man	1980	Republicans	69.6	64.4	74.8	434
Homosexual man	1980	Liberals	75.8	70.4	81.2	361
Homosexual man	1980	Conservatives	67.0	61.8	72.2	467
Homosexual man	1982	Democrats	66.4	62.6	70.2	1,048
Homosexual man	1982	Republicans	70.5	65.4	75.5	489
Homosexual man	1982	Liberals	80.7	76.1	85.3	498
Homosexual man	1982	Conservatives	66.2	61.0	71.4	507
Homosexual man	1984	Democrats	71.6	67.5	75.6	727
Homosexual man	1984	Republicans	69.9	64.8	75.0	508
Homosexual man	1984	Liberals	74.0	67.8	80.3	334
Homosexual man	1984	Conservatives	70.8	66.0	75.7	484
Homosexual man	1985	Democrats	68.7	64.4	73.0	722
Homosexual man	1985	Republicans	73.9	69.6	78.3	599
Homosexual man	1985	Liberals	77.5	71.7	83.2	358
Homosexual man	1985	Conservatives	65.0	60.1	69.9	513
Homosexual man	1987	Democrats	72.6	69.2	76.1	1,006
Homosexual man	1987	Republicans	69.9	65.4	74.4	549
Homosexual man	1987	Liberals	74.8	70.1	79.6	500
Homosexual man	1987	Conservatives	68.8	64.1	73.6	516
Homosexual man	1988	Democrats	71.6	67.0	76.1	468
Homosexual man	1988	Republicans	73.2	68.1	78.2	357
Homosexual man	1988	Liberals	79.7	74.3	85.2	264
Homosexual man	1988	Conservatives	65.8	60.1	71.6	325
Homosexual man	1989	Democrats	77.0	72.7	81.4	469
Homosexual man	1989	Republicans	78.9	74.4	83.3	404
Homosexual man	1989	Liberals	82.9	78.0	87.8	274
Homosexual man	1989	Conservatives	76.7	71.2	82.1	293
Homosexual man	1990	Democrats	78.5	73.8	83.2	372
Homosexual man	1990	Republicans	78.7	74.1	83.2	397
Homosexual man	1990	Liberals	84.4	78.9	89.9	229

The first 500 of 600 rows; every row is in figures/allow_by_year.json.

Views: By party (grouping = party); By ideology (grouping = ideology).

Note: The GSS moved mostly online in 2021 (see the methods note at the end of the chapter). Estimates from 2021 on mix web and interviewer-administered respondents unless a figure says otherwise.

Note: Annual cells with fewer than 150 respondents or an effective sample below 100 are left out.

Source: General Social Survey, 1972-2024 cumulative file; weighted (weight_use); 95% intervals on the effective sample size

Data: vishalsingh.org/politics/report/tolerance-of-speech/figures/allow_by_year.json

Table 4. Republicans allowed every speaker more often in 1976 to 1994, and by 2008 to 2018 only the racist and militarist leads remained

Gap in the share who would allow the speaker to speak, in points, by speaker and period, with 95% intervals; the axis names the two groups

Speaker	Period	Grouping	Gap (points)	Low	High	Respondents
Racist	1976 to 1994	party	−4.5	−6.3	−2.6	15,315
Militarist	1976 to 1994	party	−4.6	−6.4	−2.7	15,329
Communist	1976 to 1994	party	−4.3	−6.2	−2.5	15,275
Anti-religionist	1976 to 1994	party	−6.3	−8.0	−4.5	15,497
Homosexual man	1976 to 1994	party	−2.4	−4.1	−0.7	15,167
Racist	2008 to 2018	party	−6.8	−9.7	−4.0	7,054
Militarist	2008 to 2018	party	−3.8	−6.4	−1.1	7,069
Communist	2008 to 2018	party	+0.8	−1.9	+3.5	6,981
Anti-religionist	2008 to 2018	party	−1.4	−3.8	+1.0	7,098
Homosexual man	2008 to 2018	party	+2.1	+0.1	+4.0	7,063
Muslim clergyman	2008 to 2018	party	+0.5	−2.4	+3.4	7,062
Racist	1976 to 1994	party_no2008	−4.5	−6.3	−2.6	15,315
Militarist	1976 to 1994	party_no2008	−4.6	−6.4	−2.7	15,329
Communist	1976 to 1994	party_no2008	−4.3	−6.2	−2.5	15,275
Anti-religionist	1976 to 1994	party_no2008	−6.3	−8.0	−4.5	15,497
Homosexual man	1976 to 1994	party_no2008	−2.4	−4.1	−0.7	15,167
Racist	2010 to 2018 (2008 left out)	party_no2008	−8.4	−11.5	−5.2	5,984
Militarist	2010 to 2018 (2008 left out)	party_no2008	−3.8	−6.6	−0.9	5,998
Communist	2010 to 2018 (2008 left out)	party_no2008	−0.1	−3.0	+2.8	5,911
Anti-religionist	2010 to 2018 (2008 left out)	party_no2008	−2.7	−5.3	−0.2	6,019
Homosexual man	2010 to 2018 (2008 left out)	party_no2008	+1.1	−0.9	+3.1	5,994
Muslim clergyman	2010 to 2018 (2008 left out)	party_no2008	−0.2	−3.4	+2.9	5,990
Racist	1976 to 1994	ideology	+6.2	+4.1	+8.4	10,399
Militarist	1976 to 1994	ideology	+9.2	+7.1	+11.4	10,398
Communist	1976 to 1994	ideology	+11.0	+8.8	+13.1	10,405
Anti-religionist	1976 to 1994	ideology	+8.5	+6.5	+10.5	10,504
Homosexual man	1976 to 1994	ideology	+10.5	+8.6	+12.5	10,322
Racist	2008 to 2018	ideology	+1.0	−2.2	+4.2	5,302
Militarist	2008 to 2018	ideology	+6.3	+3.4	+9.2	5,302
Communist	2008 to 2018	ideology	+10.3	+7.4	+13.2	5,240
Anti-religionist	2008 to 2018	ideology	+7.2	+4.7	+9.8	5,333
Homosexual man	2008 to 2018	ideology	+7.9	+5.9	+9.9	5,293
Muslim clergyman	2008 to 2018	ideology	+13.7	+10.4	+17.0	5,293
Racist	1976 to 1994	party_adjusted	−1.3	−3.1	+0.6	15,228
Militarist	1976 to 1994	party_adjusted	−0.7	−2.5	+1.2	15,243
Communist	1976 to 1994	party_adjusted	+0.2	−1.6	+2.0	15,190
Anti-religionist	1976 to 1994	party_adjusted	−1.7	−3.4	0.0	15,406
Homosexual man	1976 to 1994	party_adjusted	+1.0	−0.6	+2.7	15,080
Racist	2008 to 2018	party_adjusted	−3.9	−6.9	−0.8	7,037
Militarist	2008 to 2018	party_adjusted	−0.3	−2.9	+2.4	7,052
Communist	2008 to 2018	party_adjusted	+4.8	+2.1	+7.5	6,965
Anti-religionist	2008 to 2018	party_adjusted	+3.0	+0.5	+5.4	7,082
Homosexual man	2008 to 2018	party_adjusted	+4.7	+2.7	+6.8	7,046
Muslim clergyman	2008 to 2018	party_adjusted	+5.8	+2.9	+8.7	7,045
Racist	1976 to 1994	ideology_adjusted	+5.7	+3.5	+7.9	10,338
Militarist	1976 to 1994	ideology_adjusted	+6.8	+4.7	+8.8	10,338
Communist	1976 to 1994	ideology_adjusted	+9.3	+7.3	+11.4	10,343
Anti-religionist	1976 to 1994	ideology_adjusted	+6.9	+4.9	+8.8	10,441
Homosexual man	1976 to 1994	ideology_adjusted	+8.7	+6.8	+10.6	10,261
Racist	2008 to 2018	ideology_adjusted	+1.9	−1.4	+5.1	5,289
Militarist	2008 to 2018	ideology_adjusted	+5.0	+2.1	+7.8	5,289
Communist	2008 to 2018	ideology_adjusted	+8.6	+5.8	+11.5	5,227
Anti-religionist	2008 to 2018	ideology_adjusted	+6.7	+4.1	+9.3	5,321
Homosexual man	2008 to 2018	ideology_adjusted	+6.7	+4.6	+8.7	5,280
Muslim clergyman	2008 to 2018	ideology_adjusted	+13.4	+10.3	+16.5	5,280

Views: By party (grouping = party); By party, without 2008 (grouping = party_no2008); By ideology (grouping = ideology); By party, adjusted (grouping = party_adjusted); By ideology, adjusted (grouping = ideology_adjusted).

Note: The speakers are grouped by which side is expected to object to them: liberals for the racist and the militarist, conservatives for the other four. The grouping is a coding choice, and the rows let a reader regroup them.

Note: The Muslim clergyman was first asked about in 2008.

Note: Raw intervals use the effective sample size. Adjusted intervals come from a weighted linear regression with robust (HC3) errors on respondents with complete age, education, race and sex.

Source: General Social Survey, 1972-2024 cumulative file; weighted (weight_use); 95% intervals on the effective sample size

Data: vishalsingh.org/politics/report/tolerance-of-speech/figures/gap_by_speaker.json

Table 5. Republicans allowed the racist speaker more often than Democrats did in 27 of 28 survey years, with the largest estimated leads in 2021, 2022 and 2024

Gap in the share who would allow a racist to speak in their community, in points, by survey year, all interview modes, with 95% intervals

Year	Gap	Gap (points)	Low	High	Respondents
1976	Party gap	−6.2	−12.5	+0.1	1,225
1976	Ideology gap	+7.4	+0.3	+14.4	830
1977	Party gap	−3.4	−9.5	+2.7	1,310
1977	Ideology gap	+7.4	+0.4	+14.4	877
1980	Party gap	−3.0	−9.8	+3.9	1,184
1980	Ideology gap	+6.7	−1.1	+14.4	830
1982	Party gap	−6.6	−13.2	0.0	1,561
1982	Ideology gap	+9.2	+1.8	+16.6	1,022
1984	Party gap	−6.6	−13.4	+0.2	1,257
1984	Ideology gap	+4.6	−3.7	+12.8	830
1985	Party gap	−5.1	−11.7	+1.5	1,336
1985	Ideology gap	+8.8	+0.6	+17.0	880
1987	Party gap	−0.3	−6.3	+5.8	1,558
1987	Ideology gap	+9.7	+2.5	+16.8	1,018
1988	Party gap	−5.6	−13.0	+1.7	835
1988	Ideology gap	+5.8	−2.9	+14.5	594
1989	Party gap	−3.2	−10.5	+4.1	876
1989	Ideology gap	−1.0	−9.6	+7.6	573
1990	Party gap	−8.7	−16.3	−1.1	777
1990	Ideology gap	−0.5	−9.7	+8.6	549
1991	Party gap	−4.3	−11.7	+3.1	833
1991	Ideology gap	−1.9	−10.7	+6.9	561
1993	Party gap	−1.6	−8.8	+5.5	902
1993	Ideology gap	+9.5	+1.2	+17.9	625
1994	Party gap	−4.3	−9.5	+1.0	1,661
1994	Ideology gap	+6.7	+0.6	+12.8	1,210
1996	Party gap	−3.9	−9.3	+1.5	1,554
1996	Ideology gap	+2.6	−3.8	+9.1	1,094
1998	Party gap	−4.5	−10.1	+1.1	1,482
1998	Ideology gap	+8.2	+1.9	+14.4	1,125
2000	Party gap	−5.0	−10.8	+0.7	1,429
2000	Ideology gap	+1.8	−4.8	+8.3	1,033
2002	Party gap	−3.1	−11.3	+5.0	721
2002	Ideology gap	+8.8	−0.7	+18.4	522
2004	Party gap	−5.5	−13.7	+2.6	727
2004	Ideology gap	−3.2	−13.0	+6.6	530
2006	Party gap	−1.5	−7.4	+4.4	1,512
2006	Ideology gap	+9.9	+3.2	+16.5	1,190
2008	Party gap	+1.6	−5.5	+8.7	1,070
2008	Ideology gap	+9.2	+1.1	+17.3	772
2010	Party gap	−7.7	−15.4	0.0	990
2010	Ideology gap	−1.5	−10.1	+7.2	748
2012	Party gap	−6.8	−14.7	+1.0	983
2012	Ideology gap	−1.0	−9.5	+7.6	765
2014	Party gap	−12.0	−18.1	−5.8	1,311
2014	Ideology gap	−5.6	−12.5	+1.2	1,003
2016	Party gap	−9.1	−15.2	−2.9	1,497
2016	Ideology gap	+3.3	−3.7	+10.4	1,115
2018	Party gap	−5.6	−12.9	+1.8	1,203
2018	Ideology gap	+2.4	−6.0	+10.7	899
2021	Party gap	−22.9	−30.3	−15.4	990
2021	Ideology gap	−15.0	−23.1	−6.9	835
2022	Party gap	−15.0	−26.8	−3.3	817
2022	Ideology gap	−17.1	−29.6	−4.6	697
2024	Party gap	−13.6	−23.3	−3.8	797
2024	Ideology gap	−5.0	−15.3	+5.2	683

Views: By party (series = Party gap); By ideology (series = Ideology gap).

Note: The GSS moved mostly online in 2021 (see the methods note at the end of the chapter). Estimates from 2021 on mix web and interviewer-administered respondents unless a figure says otherwise.

Note: Years in which either group has fewer than 150 respondents or an effective sample below 100 are left out.

Source: General Social Survey, 1972-2024 cumulative file; weighted (weight_use); 95% intervals on the effective sample size

Data: vishalsingh.org/politics/report/tolerance-of-speech/figures/racist_gap_by_year.json

Table 6. Web respondents allowed the racist speaker 14.0 points less often; the other three speakers show no detectable mode difference

Share who would allow the speaker to speak (percent) or the gap in that share between groups (points), by interview mode, 2022 and 2024, with 95% intervals

Measure	Speaker	Interviewer-administered	Low	High	Web	Low	High	Respondents, interviewer	Respondents, web
all	Racist	54.7	50.0	59.4	40.7	36.5	44.9	1,118	1,153
all	Communist	65.7	61.1	70.2	67.7	63.7	71.7	1,081	1,151
all	Anti-religionist	79.4	75.6	83.2	78.0	74.4	81.6	1,130	1,150
all	Muslim clergyman	46.3	41.6	50.9	44.0	39.7	48.3	1,111	1,148
dem	Racist	49.8	42.6	57.0	33.4	26.8	40.1	443	469
dem	Communist	65.9	59.0	72.7	69.5	62.9	76.0	437	467
dem	Anti-religionist	79.1	73.2	84.9	81.9	76.4	87.3	451	468
dem	Muslim clergyman	46.5	39.3	53.6	49.9	42.8	57.0	440	468
rep	Racist	61.5	53.0	70.1	48.9	41.6	56.3	360	342
rep	Communist	65.1	56.5	73.6	63.9	56.9	70.9	350	342
rep	Anti-religionist	77.5	70.2	84.8	71.1	64.5	77.7	364	341
rep	Muslim clergyman	47.6	38.8	56.4	35.5	28.5	42.5	361	342
party_gap	Racist	−11.8	−23.0	−0.5	−15.5	−25.4	−5.6	803	811
party_gap	Communist	0.8	−10.2	11.8	5.6	−4.0	15.2	787	809
party_gap	Anti-religionist	1.6	−7.7	10.9	10.8	2.2	19.4	815	809
party_gap	Muslim clergyman	−1.1	−12.5	10.2	14.4	4.4	24.3	801	810
ideology_gap	Racist	−5.1	−17.6	7.5	−17.3	−27.4	−7.2	649	731
ideology_gap	Communist	9.3	−2.9	21.5	10.5	1.5	19.5	637	730
ideology_gap	Anti-religionist	5.8	−4.4	16.1	12.9	5.2	20.6	657	725
ideology_gap	Muslim clergyman	6.6	−6.0	19.3	16.8	6.6	27.0	644	726

Views: All adults (measure = all); Party gap (measure = party_gap); Ideology gap (measure = ideology_gap).

Note: Only the racist, Communist, anti-religionist and Muslim clergyman items were asked in 2022 and 2024. Phone and other-mode respondents are counted as interviewer-administered because in-person cells alone fall below the effective-sample floor for Republicans and conservatives.

Note: The 2021 wave is left out: it has too few interviewer-administered respondents.

Note: Shares among Democrats and among Republicans separately are in the data table only.

Source: General Social Survey, 1972-2024 cumulative file; weighted (weight_use); 95% intervals on the effective sample size

Data: vishalsingh.org/politics/report/tolerance-of-speech/figures/mode_comparison.json

Table 7. By 2008 to 2018 the party gap on speakers that conservatives object to was +0.4 points, and on speakers that liberals object to −5.2

Gap between groups in the average share of a family of speakers a respondent would allow to speak, in points, by family and period, with 95% intervals

Family	Period	Grouping	Gap (points)	Low	High	Respondents
Objectionable to liberals	1976 to 1994	party	−4.5	−6.1	−2.9	15,078
Objectionable to liberals	1976 to 1994	ideology	+7.7	+5.8	+9.6	10,246
Objectionable to liberals	2008 to 2018	party	−5.2	−7.6	−2.9	6,991
Objectionable to liberals	2008 to 2018	ideology	+3.5	+1.0	+6.1	5,250
Objectionable to liberals	2021 (mostly web)	party	−16.3	−22.5	−10.0	982
Objectionable to liberals	2021 (mostly web)	ideology	−5.0	−11.9	+1.8	833
Objectionable to conservatives	1976 to 1994	party	−4.2	−5.7	−2.8	14,809
Objectionable to conservatives	1976 to 1994	ideology	+9.7	+8.0	+11.4	10,130
Objectionable to conservatives	2008 to 2018	party	+0.4	−1.4	+2.3	6,885
Objectionable to conservatives	2008 to 2018	ideology	+8.3	+6.3	+10.2	5,169
Objectionable to conservatives	2021 (mostly web)	party	+3.2	−1.4	+7.7	976
Objectionable to conservatives	2021 (mostly web)	ideology	+13.6	+9.0	+18.1	826
All five, each speaker counts once	1976 to 1994	party	−4.3	−5.7	−2.9	14,468
All five, each speaker counts once	1976 to 1994	ideology	+8.9	+7.2	+10.5	9,907
All five, each speaker counts once	2008 to 2018	party	−1.8	−3.7	0.0	6,780
All five, each speaker counts once	2008 to 2018	ideology	+6.3	+4.3	+8.3	5,102
All five, each speaker counts once	2021 (mostly web)	party	−4.6	−9.3	+0.2	969
All five, each speaker counts once	2021 (mostly web)	ideology	+6.2	+1.3	+11.0	824
All five, each side counts once	1976 to 1994	party	−4.3	−5.7	−2.9	14,468
All five, each side counts once	1976 to 1994	ideology	+8.7	+7.0	+10.3	9,907
All five, each side counts once	2008 to 2018	party	−2.4	−4.3	−0.5	6,780
All five, each side counts once	2008 to 2018	ideology	+5.8	+3.8	+7.9	5,102
All five, each side counts once	2021 (mostly web)	party	−6.6	−11.5	−1.6	969
All five, each side counts once	2021 (mostly web)	ideology	+4.3	−0.8	+9.4	824

Views: By party (grouping = party, plotted = yes); By ideology (grouping = ideology, plotted = yes).

Note: The family of speakers objectionable to conservatives is the Communist, the anti-religionist and the homosexual man, and the family objectionable to liberals is the racist and the militarist. A respondent's family value is the share of that family's speakers he or she would allow, for respondents who answered all of the family's items. The Muslim clergyman is left out because the item began in 2008.

Note: The data table also lists two combined averages of all five speakers that are not plotted: one counts each speaker once and one counts each family once.

Note: The GSS moved mostly online in 2021 (see the methods note at the end of the chapter). Estimates from 2021 on mix web and interviewer-administered respondents unless a figure says otherwise.

Source: General Social Survey, 1972-2024 cumulative file; weighted (weight_use); 95% intervals on the effective sample size

Data: vishalsingh.org/politics/report/tolerance-of-speech/figures/family_gaps.json

Table 8. The racist party gap is negative on speech, teaching and the library, and party gaps on teaching are positive for three other speakers

Gap between groups in the share giving the tolerant answer, in points, by speaker and action, 2008 to 2018, with 95% intervals; the axis names the groups

Speaker	Action	Grouping	Gap (points)	Low	High	Respondents
Racist	Allow to speak	party	−6.8	−9.7	−4.0	7,054
Racist	Allow to speak	ideology	+1.0	−2.2	+4.2	5,302
Militarist	Allow to speak	party	−3.8	−6.4	−1.1	7,069
Militarist	Allow to speak	ideology	+6.3	+3.4	+9.2	5,302
Communist	Allow to speak	party	+0.8	−1.9	+3.5	6,981
Communist	Allow to speak	ideology	+10.3	+7.4	+13.2	5,240
Anti-religionist	Allow to speak	party	−1.4	−3.8	+1.0	7,098
Anti-religionist	Allow to speak	ideology	+7.2	+4.7	+9.8	5,333
Homosexual man	Allow to speak	party	+2.1	+0.1	+4.0	7,063
Homosexual man	Allow to speak	ideology	+7.9	+5.9	+9.9	5,293
Muslim clergyman	Allow to speak	party	+0.5	−2.4	+3.4	7,062
Muslim clergyman	Allow to speak	ideology	+13.7	+10.4	+17.0	5,293
Racist	Allow to teach in a college	party	−4.4	−7.3	−1.4	6,975
Racist	Allow to teach in a college	ideology	+0.2	−3.1	+3.5	5,251
Militarist	Allow to teach in a college	party	−0.3	−3.2	+2.6	6,979
Militarist	Allow to teach in a college	ideology	+10.6	+7.3	+13.8	5,238
Communist	Allow to teach in a college	party	+8.4	+5.5	+11.2	6,810
Communist	Allow to teach in a college	ideology	+18.4	+15.4	+21.5	5,128
Anti-religionist	Allow to teach in a college	party	+4.2	+1.4	+7.0	6,983
Anti-religionist	Allow to teach in a college	ideology	+13.8	+10.7	+16.8	5,247
Homosexual man	Allow to teach in a college	party	+3.7	+1.7	+5.7	7,031
Homosexual man	Allow to teach in a college	ideology	+9.0	+6.8	+11.1	5,274
Muslim clergyman	Allow to teach in a college	party	+1.7	−1.0	+4.5	7,021
Muslim clergyman	Allow to teach in a college	ideology	+12.4	+9.3	+15.6	5,268
Racist	Keep the book in the public library	party	−4.9	−7.7	−2.1	7,017
Racist	Keep the book in the public library	ideology	+4.6	+1.5	+7.7	5,277
Militarist	Keep the book in the public library	party	−2.1	−4.7	+0.4	7,034
Militarist	Keep the book in the public library	ideology	+9.0	+6.2	+11.8	5,279
Communist	Keep the book in the public library	party	+0.1	−2.4	+2.7	6,977
Communist	Keep the book in the public library	ideology	+12.1	+9.3	+14.9	5,247
Anti-religionist	Keep the book in the public library	party	+0.8	−1.7	+3.2	7,042
Anti-religionist	Keep the book in the public library	ideology	+12.3	+9.7	+15.0	5,287
Homosexual man	Keep the book in the public library	party	+2.7	+0.4	+5.0	7,055
Homosexual man	Keep the book in the public library	ideology	+12.7	+10.2	+15.1	5,282
Muslim clergyman	Keep the book in the public library	party	−1.1	−4.0	+1.8	7,077
Muslim clergyman	Keep the book in the public library	ideology	+14.8	+11.5	+18.0	5,303

Views: By party (grouping = party); By ideology (grouping = ideology).

Note: Independents and moderates are left out of the group comparisons. The 2008 to 2018 waves pool six survey years.

Note: The Communist teaching item asks about firing a teacher; the coding treats not firing as the tolerant answer.

Source: General Social Survey, 1972-2024 cumulative file; weighted (weight_use); 95% intervals on the effective sample size

Data: vishalsingh.org/politics/report/tolerance-of-speech/figures/action_gaps.json

All estimates

Every fact in facts.json (121), sorted by id. The same values as JSON: facts.json.

Fact id	Value	Unit	Estimand	Population	Period	Interval	n	Weights	Source
adjig_con_max	13.4	points	largest adjusted ideology gap among the Communist, anti-religionist, homosexual man and Muslim clergyman items (adjusted for age, education, race, sex and survey year)	liberals and conservatives (self-placed) answering the item	2008-2018 waves pooled			weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
adjig_con_min	6.7	points	smallest adjusted ideology gap among the Communist, anti-religionist, homosexual man and Muslim clergyman items (adjusted for age, education, race, sex and survey year)	liberals and conservatives (self-placed) answering the item	2008-2018 waves pooled			weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
adjpg_con_max	5.8	points	largest adjusted party gap among the Communist, anti-religionist, homosexual man and Muslim clergyman items (adjusted for age, education, race, sex and survey year)	Democrats and Republicans (leaners included) answering the item	2008-2018 waves pooled			weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
adjpg_con_min	3.0	points	smallest adjusted party gap among the Communist, anti-religionist, homosexual man and Muslim clergyman items (adjusted for age, education, race, sex and survey year)	Democrats and Republicans (leaners included) answering the item	2008-2018 waves pooled			weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
adjpg_early_abs_max	1.7	points	largest absolute adjusted party gap among the five speakers asked (adjusted for age, education, race, sex and survey year)	Democrats and Republicans (leaners included) answering the item	1976-1994 waves pooled			weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
adjpg_mil_late	−0.3	points (first group minus second group)	difference in the share who would allow the speaker to speak, Democrats minus Republicans, adjusted for age, education, race, sex and survey year (weighted linear probability model, HC3 errors)	Democrats and Republicans (leaners included) answering the item	2008-2018 waves pooled	95% CI −2.9 to 2.4	7,052	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
adjpg_rac_early	−1.3	points (first group minus second group)	difference in the share who would allow the speaker to speak, Democrats minus Republicans, adjusted for age, education, race, sex and survey year (weighted linear probability model, HC3 errors)	Democrats and Republicans (leaners included) answering the item	1976-1994 waves pooled	95% CI −3.1 to 0.6	15,228	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
adjpg_rac_late	−3.9	points (first group minus second group)	difference in the share who would allow the speaker to speak, Democrats minus Republicans, adjusted for age, education, race, sex and survey year (weighted linear probability model, HC3 errors)	Democrats and Republicans (leaners included) answering the item	2008-2018 waves pooled	95% CI −6.9 to −0.8	7,037	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
annual_min_n	150	respondents	minimum respondents for an annual cell	annual party and ideology cells	1972-2024			none (constant)	General Social Survey
chg_other_max	15.2	points	largest rise among the militarist, Communist, anti-religionist and homosexual man items	adults answering the item	1976-1994 to 2008-2018			weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
chg_other_min	6.3	points	smallest rise among the militarist, Communist, anti-religionist and homosexual man items	adults answering the item	1976-1994 to 2008-2018			weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
chg_rac	−3.2	points (2008-2018 minus 1976-1994)	difference between the two periods in the weighted share who would allow the speech (independent samples)	adults answering the item (Racist)	1976-1994 to 2008-2018	95% CI −4.7 to −1.7		weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
design_n	64	contrasts	count of contrasts checked against design-based standard errors	period gaps and mode contrasts in this chapter	1976-2024			none (count)	General Social Survey
design_ratio_max	1.27	ratio of standard errors	largest ratio of the design-based to the published standard error	period gaps and mode contrasts in this chapter	1976-2024			weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
design_ratio_min	0.87	ratio of standard errors	smallest ratio of the design-based to the published standard error	period gaps and mode contrasts in this chapter	1976-2024			weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
fam_con2_ig_late	+9.0	points (first group minus second group)	difference between groups in the mean share of the Communist and anti-religionist items a respondent would allow (liberals minus conservatives)	liberals and conservatives (self-placed) answering the item	2008-2018 waves pooled	95% CI 6.6 to 11.3	5,219	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
fam_con_ig_chg	−1.4	points (2008-2018 gap minus 1976-1994 gap)	change in the ideology gap in the mean share of the family's speakers allowed (independent samples)	liberals and conservatives (self-placed) answering the item	1976-1994 to 2008-2018	95% CI −4.0 to 1.2		weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
fam_con_ig_early	+9.7	points (first group minus second group)	difference between groups in the mean share of the family's speakers a respondent would allow (liberals minus conservatives)	liberals and conservatives (self-placed) answering the item	1976-1994 waves pooled	95% CI 8.0 to 11.4	10,130	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
fam_con_ig_late	+8.3	points (first group minus second group)	difference between groups in the mean share of the family's speakers a respondent would allow (liberals minus conservatives)	liberals and conservatives (self-placed) answering the item	2008-2018 waves pooled	95% CI 6.3 to 10.2	5,169	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
fam_con_pg_early	−4.2	points (first group minus second group)	difference between groups in the mean share of the family's speakers a respondent would allow (Democrats minus Republicans)	Democrats and Republicans (leaners included) answering the item	1976-1994 waves pooled	95% CI −5.7 to −2.8	14,809	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
fam_con_pg_late	+0.4	points (first group minus second group)	difference between groups in the mean share of the family's speakers a respondent would allow (Democrats minus Republicans)	Democrats and Republicans (leaners included) answering the item	2008-2018 waves pooled	95% CI −1.4 to 2.3	6,885	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
fam_eq_ig_late	+5.8	points (first group minus second group)	difference between groups in the mean share of the family's speakers a respondent would allow (liberals minus conservatives)	liberals and conservatives (self-placed) answering the item	2008-2018 waves pooled	95% CI 3.8 to 7.9	5,102	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
fam_eq_pg_2021	−6.6	points (first group minus second group)	difference between groups in the mean share of the family's speakers a respondent would allow (Democrats minus Republicans)	Democrats and Republicans (leaners included) answering the item	2021 wave	95% CI −11.5 to −1.6	969	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
fam_eq_pg_late	−2.4	points (first group minus second group)	difference between groups in the mean share of the family's speakers a respondent would allow (Democrats minus Republicans)	Democrats and Republicans (leaners included) answering the item	2008-2018 waves pooled	95% CI −4.3 to −0.5	6,780	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
fam_five_ig_late	+6.3	points (first group minus second group)	difference between groups in the mean share of the family's speakers a respondent would allow (liberals minus conservatives)	liberals and conservatives (self-placed) answering the item	2008-2018 waves pooled	95% CI 4.3 to 8.3	5,102	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
fam_five_pg_2021	−4.6	points (first group minus second group)	difference between groups in the mean share of the family's speakers a respondent would allow (Democrats minus Republicans)	Democrats and Republicans (leaners included) answering the item	2021 wave	95% CI −9.3 to 0.2	969	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
fam_five_pg_late	−1.8	points (first group minus second group)	difference between groups in the mean share of the family's speakers a respondent would allow (Democrats minus Republicans)	Democrats and Republicans (leaners included) answering the item	2008-2018 waves pooled	95% CI −3.7 to 0.0	6,780	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
fam_ig_chg_diff	−2.7	points (change for the family objectionable to liberals minus change for the other family)	difference between the two families' changes in the ideology gap (independent samples)	liberals and conservatives (self-placed) answering the item	1976-1994 to 2008-2018	95% CI −6.9 to 1.4		weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
fam_lib_ig_2021	−5.0	points (first group minus second group)	difference between groups in the mean share of the family's speakers a respondent would allow (liberals minus conservatives)	liberals and conservatives (self-placed) answering the item	2021 wave	95% CI −11.9 to 1.8	833	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
fam_lib_ig_chg	−4.2	points (2008-2018 gap minus 1976-1994 gap)	change in the ideology gap in the mean share of the family's speakers allowed (independent samples)	liberals and conservatives (self-placed) answering the item	1976-1994 to 2008-2018	95% CI −7.3 to −1.0		weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
fam_lib_ig_early	+7.7	points (first group minus second group)	difference between groups in the mean share of the family's speakers a respondent would allow (liberals minus conservatives)	liberals and conservatives (self-placed) answering the item	1976-1994 waves pooled	95% CI 5.8 to 9.6	10,246	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
fam_lib_ig_late	+3.5	points (first group minus second group)	difference between groups in the mean share of the family's speakers a respondent would allow (liberals minus conservatives)	liberals and conservatives (self-placed) answering the item	2008-2018 waves pooled	95% CI 1.0 to 6.1	5,250	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
fam_lib_pg_2021	−16.3	points (first group minus second group)	difference between groups in the mean share of the family's speakers a respondent would allow (Democrats minus Republicans)	Democrats and Republicans (leaners included) answering the item	2021 wave	95% CI −22.5 to −10.0	982	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
fam_lib_pg_early	−4.5	points (first group minus second group)	difference between groups in the mean share of the family's speakers a respondent would allow (Democrats minus Republicans)	Democrats and Republicans (leaners included) answering the item	1976-1994 waves pooled	95% CI −6.1 to −2.9	15,078	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
fam_lib_pg_late	−5.2	points (first group minus second group)	difference between groups in the mean share of the family's speakers a respondent would allow (Democrats minus Republicans)	Democrats and Republicans (leaners included) answering the item	2008-2018 waves pooled	95% CI −7.6 to −2.9	6,991	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
ig_book_con_max	14.8	points	largest gap in keeping the book among the four speakers objectionable to conservatives	liberals and conservatives (self-placed) answering the item	2008-2018 waves pooled			weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
ig_book_con_min	12.1	points	smallest gap in keeping the book among the four speakers objectionable to conservatives	liberals and conservatives (self-placed) answering the item	2008-2018 waves pooled			weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
ig_com_early	+11.0	points (first group minus second group)	difference in weighted shares who would allow the speaker to speak in the community (liberals minus conservatives)	liberals and conservatives (self-placed) answering the item	1976-1994 waves pooled	95% CI 8.8 to 13.1	10,405	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
ig_com_late	+10.3	points (first group minus second group)	difference in weighted shares who would allow the speaker to speak in the community (liberals minus conservatives)	liberals and conservatives (self-placed) answering the item	2008-2018 waves pooled	95% CI 7.4 to 13.2	5,240	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
ig_con_max	13.7	points	largest ideology gap among the Communist, anti-religionist, homosexual man and Muslim clergyman items	liberals and conservatives (self-placed) answering the item	2008-2018 waves pooled			weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
ig_con_min	7.2	points	smallest ideology gap among the Communist, anti-religionist, homosexual man and Muslim clergyman items	liberals and conservatives (self-placed) answering the item	2008-2018 waves pooled			weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
ig_homo_early	+10.5	points (first group minus second group)	difference in weighted shares who would allow the speaker to speak in the community (liberals minus conservatives)	liberals and conservatives (self-placed) answering the item	1976-1994 waves pooled	95% CI 8.6 to 12.5	10,322	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
ig_homo_late	+7.9	points (first group minus second group)	difference in weighted shares who would allow the speaker to speak in the community (liberals minus conservatives)	liberals and conservatives (self-placed) answering the item	2008-2018 waves pooled	95% CI 5.9 to 9.9	5,293	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
ig_mil_late	+6.3	points (first group minus second group)	difference in weighted shares who would allow the speaker to speak in the community (liberals minus conservatives)	liberals and conservatives (self-placed) answering the item	2008-2018 waves pooled	95% CI 3.4 to 9.2	5,302	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
ig_mslm_late	+13.7	points (first group minus second group)	difference in weighted shares who would allow the speaker to speak in the community (liberals minus conservatives)	liberals and conservatives (self-placed) answering the item	2008-2018 waves pooled	95% CI 10.4 to 17.0	5,293	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
ig_rac_book_late	+4.6	points (first group minus second group)	difference in weighted shares who would keep the speaker's book in the public library (liberals minus conservatives)	liberals and conservatives (self-placed) answering the item	2008-2018 waves pooled	95% CI 1.5 to 7.7	5,277	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
ig_rac_early	+6.2	points (first group minus second group)	difference in weighted shares who would allow the speaker to speak in the community (liberals minus conservatives)	liberals and conservatives (self-placed) answering the item	1976-1994 waves pooled	95% CI 4.1 to 8.4	10,399	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
ig_rac_late	+1.0	points (first group minus second group)	difference in weighted shares who would allow the speaker to speak in the community (liberals minus conservatives)	liberals and conservatives (self-placed) answering the item	2008-2018 waves pooled	95% CI −2.2 to 4.2	5,302	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
ig_rac_modechg	−12.2	points (web gap minus interviewer-administered gap)	difference between modes in the racist ideology gap (independent samples)	liberals and conservatives (self-placed) answering the item	2022 and 2024	95% CI −28.4 to 3.9		weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
ig_rac_teach_late	+0.2	points (first group minus second group)	difference in weighted shares who would allow the speaker to teach in a college (Communist: not fire the teacher) (liberals minus conservatives)	liberals and conservatives (self-placed) answering the item	2008-2018 waves pooled	95% CI −3.1 to 3.5	5,251	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
ig_rel_early	+8.5	points (first group minus second group)	difference in weighted shares who would allow the speaker to speak in the community (liberals minus conservatives)	liberals and conservatives (self-placed) answering the item	1976-1994 waves pooled	95% CI 6.5 to 10.5	10,504	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
ig_rel_late	+7.2	points (first group minus second group)	difference in weighted shares who would allow the speaker to speak in the community (liberals minus conservatives)	liberals and conservatives (self-placed) answering the item	2008-2018 waves pooled	95% CI 4.7 to 9.8	5,333	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
ig_teach_con_max	18.4	points	largest gap in allowing the speaker to teach among the four speakers objectionable to conservatives	liberals and conservatives (self-placed) answering the item	2008-2018 waves pooled			weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
ig_teach_con_min	9.0	points	smallest gap in allowing the speaker to teach among the four speakers objectionable to conservatives	liberals and conservatives (self-placed) answering the item	2008-2018 waves pooled			weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
igchg_mil	−2.9	points (2008-2018 gap minus 1976-1994 gap)	change between the two periods in the ideology gap (independent samples)	liberals and conservatives (self-placed) answering the item	1976-1994 to 2008-2018	95% CI −6.6 to 0.7		weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
igchg_rac	−5.2	points (2008-2018 gap minus 1976-1994 gap)	change between the two periods in the ideology gap (independent samples)	liberals and conservatives (self-placed) answering the item	1976-1994 to 2008-2018	95% CI −9.1 to −1.3		weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
igdiff_mslm_com_late	+3.4	points (Muslim clergyman ideology gap minus Communist ideology gap)	difference between the Muslim clergyman and Communist ideology gaps in the share who would allow the speech (same period, standard errors in quadrature)	liberals and conservatives (self-placed) answering the item	2008-2018 waves pooled	95% CI −1.0 to 7.7		weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
lev_mslm_late	42.5%	percent who would allow the speaker to speak	weighted share who would allow the speech	adults answering the item (Muslim clergyman)	2008-2018 waves pooled	95% CI 41.3% to 43.8%	8,970	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
lev_rac_book_late	63.7%	percent giving the tolerant answer	weighted share who would keep the speaker's book in the public library	adults answering the item (racist)	2008-2018 waves pooled	95% CI 62.4% to 64.9%	8,908	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
lev_rac_late	58.5%	percent who would allow the speaker to speak	weighted share who would allow the speech	adults answering the item (Racist)	2008-2018 waves pooled	95% CI 57.3% to 59.8%	8,955	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
lev_rac_teach_late	46.3%	percent giving the tolerant answer	weighted share who would allow the speaker to teach in a college (Communist: not fire the teacher)	adults answering the item (racist)	2008-2018 waves pooled	95% CI 45.0% to 47.5%	8,869	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
max_abs_pg_other	2.1	points	largest absolute party gap among the Communist, anti-religionist, homosexual man and Muslim clergyman items	Democrats and Republicans (leaners included) answering the item	2008-2018 waves pooled			weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
mslm_mode_diff_dem	+3.4	points (web minus interviewer-administered)	difference in the weighted share who would allow the Muslim clergyman, web against interviewer-administered respondents	Democrats (leaners included) interviewed in 2022 and 2024	2022 and 2024	95% CI −6.6 to 13.5		weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
mslm_mode_diff_rep	−12.1	points (web minus interviewer-administered)	difference in the weighted share who would allow the Muslim clergyman, web against interviewer-administered respondents	Republicans (leaners included) interviewed in 2022 and 2024	2022 and 2024	95% CI −23.3 to −0.9		weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
n_answers	217,422	answers	count of non-missing answers to the six speak items	all answers to the six speak items	1972-2024			none (count)	General Social Survey
n_pg_ci_neg	7	survey years	count of survey years in which the whole 95% interval is below zero	survey years with an annual racist party gap	1976-2024			none (count)	General Social Survey
n_pg_neg	27	survey years	count of survey years in which the estimate is below zero	survey years with an annual racist party gap	1976-2024			none (count)	General Social Survey
n_pg_years	28	survey years	count of survey years with a published annual party gap	survey years with an annual racist party gap	1976-2024			none (count)	General Social Survey
n_respondents	44,878	respondents	count of respondents	respondents who answered at least one speak item	1972-2024			none (count)	General Social Survey
n_years	31	survey years	count of survey years	years in which at least one speak item was asked	1972-2024			none (count)	General Social Survey
neff_floor	100	respondents (Kish effective sample size)	minimum effective sample for a cell to appear	party, ideology and mode cells	1972-2024			none (constant)	General Social Survey
pg_2008_max	+7.6	points (first group minus second group)	largest party gap among the racist, Communist, anti-religionist, homosexual man and Muslim clergyman items	Democrats and Republicans (leaners included) answering the item	2008 wave			weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pg_2008_min	+1.6	points (first group minus second group)	smallest party gap among the racist, Communist, anti-religionist, homosexual man and Muslim clergyman items	Democrats and Republicans (leaners included) answering the item	2008 wave			weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pg_com_late	+0.8	points (first group minus second group)	difference in weighted shares who would allow the speaker to speak in the community (Democrats minus Republicans)	Democrats and Republicans (leaners included) answering the item	2008-2018 waves pooled	95% CI −1.9 to 3.5	6,981	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pg_com_teach_late	+8.4	points (first group minus second group)	difference in weighted shares who would allow the speaker to teach in a college (Communist: not fire the teacher) (Democrats minus Republicans)	Democrats and Republicans (leaners included) answering the item	2008-2018 waves pooled	95% CI 5.5 to 11.2	6,810	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pg_early_abs_max	6.3	points	largest absolute party gap among the five speakers asked	Democrats and Republicans (leaners included) answering the item	1976-1994 waves pooled			weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pg_early_abs_min	2.4	points	smallest absolute party gap among the five speakers asked	Democrats and Republicans (leaners included) answering the item	1976-1994 waves pooled			weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pg_homo_early	−2.4	points (first group minus second group)	difference in weighted shares who would allow the speaker to speak in the community (Democrats minus Republicans)	Democrats and Republicans (leaners included) answering the item	1976-1994 waves pooled	95% CI −4.1 to −0.7	15,167	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pg_homo_late	+2.1	points (first group minus second group)	difference in weighted shares who would allow the speaker to speak in the community (Democrats minus Republicans)	Democrats and Republicans (leaners included) answering the item	2008-2018 waves pooled	95% CI 0.1 to 4.0	7,063	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pg_homo_late_design	+2.1	points (first group minus second group)	party gap for the homosexual man item with a design-based interval (Taylor linearization, strata and primary sampling units by survey year)	Democrats and Republicans (leaners included) answering the item	2008-2018 waves pooled	95% CI 0.0 to 4.1		weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pg_homo_late_lo	0.1	points	lower end of the 95% interval on the homosexual man party gap, Democrats minus Republicans	Democrats and Republicans (leaners included) answering the item	2008-2018 waves pooled			weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pg_homo_teach_late	+3.7	points (first group minus second group)	difference in weighted shares who would allow the speaker to teach in a college (Communist: not fire the teacher) (Democrats minus Republicans)	Democrats and Republicans (leaners included) answering the item	2008-2018 waves pooled	95% CI 1.7 to 5.7	7,031	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pg_mil_2008	−3.6	points (first group minus second group)	difference in weighted shares who would allow the speaker to speak in the community (Democrats minus Republicans)	Democrats and Republicans (leaners included) answering the item	2008 wave	95% CI −10.4 to 3.1	1,071	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pg_mil_early	−4.6	points (first group minus second group)	difference in weighted shares who would allow the speaker to speak in the community (Democrats minus Republicans)	Democrats and Republicans (leaners included) answering the item	1976-1994 waves pooled	95% CI −6.4 to −2.7	15,329	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pg_mil_late	−3.8	points (first group minus second group)	difference in weighted shares who would allow the speaker to speak in the community (Democrats minus Republicans)	Democrats and Republicans (leaners included) answering the item	2008-2018 waves pooled	95% CI −6.4 to −1.1	7,069	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pg_mslm_late	+0.5	points (first group minus second group)	difference in weighted shares who would allow the speaker to speak in the community (Democrats minus Republicans)	Democrats and Republicans (leaners included) answering the item	2008-2018 waves pooled	95% CI −2.4 to 3.4	7,062	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pg_mslm_modechg	+15.5	points (web gap minus interviewer-administered gap)	difference between modes in the party gap (independent samples)	Democrats and Republicans (leaners included) answering the item	2022 and 2024	95% CI 0.4 to 30.6		weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pg_mslm_modechg_design	+15.5	points (web gap minus interviewer-administered gap)	difference between modes in the Muslim clergyman party gap with a design-based interval (Taylor linearization, strata and primary sampling units by survey year)	Democrats and Republicans (leaners included) answering the item	2022 and 2024 waves pooled	95% CI −3.0 to 34.0		weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pg_mslm_modechg_lo	0.4	points	lower end of the 95% interval on the difference between modes in the Muslim clergyman party gap (web gap minus interviewer-administered gap)	Democrats and Republicans (leaners included) answering the item	2022 and 2024			weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pg_rac_2021	−22.9	points (first group minus second group)	difference in weighted shares who would allow the speaker to speak in the community (Democrats minus Republicans)	Democrats and Republicans (leaners included) answering the item	2021 wave	95% CI −30.3 to −15.4	990	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pg_rac_all	−14.2	points (first group minus second group)	difference in weighted shares who would allow the speech, Democrats minus Republicans	Democrats and Republicans (leaners included) answering the item	2022 and 2024	95% CI −21.9 to −6.5	1,614	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pg_rac_book_late	−4.9	points (first group minus second group)	difference in weighted shares who would keep the speaker's book in the public library (Democrats minus Republicans)	Democrats and Republicans (leaners included) answering the item	2008-2018 waves pooled	95% CI −7.7 to −2.1	7,017	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pg_rac_early	−4.5	points (first group minus second group)	difference in weighted shares who would allow the speaker to speak in the community (Democrats minus Republicans)	Democrats and Republicans (leaners included) answering the item	1976-1994 waves pooled	95% CI −6.3 to −2.6	15,315	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pg_rac_int	−11.8	points (first group minus second group)	difference in weighted shares who would allow the speech, Democrats minus Republicans	Democrats and Republicans (leaners included) answering the item	2022 and 2024, interviewer-administered respondents	95% CI −23.0 to −0.5	803	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pg_rac_late	−6.8	points (first group minus second group)	difference in weighted shares who would allow the speaker to speak in the community (Democrats minus Republicans)	Democrats and Republicans (leaners included) answering the item	2008-2018 waves pooled	95% CI −9.7 to −4.0	7,054	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pg_rac_modechg	−3.7	points (web gap minus interviewer-administered gap)	difference between modes in the racist party gap (independent samples)	Democrats and Republicans (leaners included) answering the item	2022 and 2024	95% CI −18.7 to 11.2		weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pg_rac_teach_late	−4.4	points (first group minus second group)	difference in weighted shares who would allow the speaker to teach in a college (Communist: not fire the teacher) (Democrats minus Republicans)	Democrats and Republicans (leaners included) answering the item	2008-2018 waves pooled	95% CI −7.3 to −1.4	6,975	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pg_rac_web	−15.5	points (first group minus second group)	difference in weighted shares who would allow the speech, Democrats minus Republicans	Democrats and Republicans (leaners included) answering the item	2022 and 2024, web respondents	95% CI −25.4 to −5.6	811	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pg_rel_early	−6.3	points (first group minus second group)	difference in weighted shares who would allow the speaker to speak in the community (Democrats minus Republicans)	Democrats and Republicans (leaners included) answering the item	1976-1994 waves pooled	95% CI −8.0 to −4.5	15,497	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pg_rel_late	−1.4	points (first group minus second group)	difference in weighted shares who would allow the speaker to speak in the community (Democrats minus Republicans)	Democrats and Republicans (leaners included) answering the item	2008-2018 waves pooled	95% CI −3.8 to 1.0	7,098	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pg_rel_teach_late	+4.2	points (first group minus second group)	difference in weighted shares who would allow the speaker to teach in a college (Communist: not fire the teacher) (Democrats minus Republicans)	Democrats and Republicans (leaners included) answering the item	2008-2018 waves pooled	95% CI 1.4 to 7.0	6,983	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pgchg_com	+5.1	points (2008-2018 gap minus 1976-1994 gap)	change between the two periods in the party gap (independent samples)	Democrats and Republicans (leaners included) answering the item	1976-1994 to 2008-2018	95% CI 1.8 to 8.4		weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pgchg_homo	+4.5	points (2008-2018 gap minus 1976-1994 gap)	change between the two periods in the party gap (independent samples)	Democrats and Republicans (leaners included) answering the item	1976-1994 to 2008-2018	95% CI 1.9 to 7.0		weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pgchg_mil	+0.8	points (2008-2018 gap minus 1976-1994 gap)	change between the two periods in the party gap (independent samples)	Democrats and Republicans (leaners included) answering the item	1976-1994 to 2008-2018	95% CI −2.4 to 4.0		weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pgchg_rac	−2.4	points (2008-2018 gap minus 1976-1994 gap)	change between the two periods in the party gap (independent samples)	Democrats and Republicans (leaners included) answering the item	1976-1994 to 2008-2018	95% CI −5.8 to 1.0		weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pgchg_rac_2010	−3.9	points (2010-2018 gap minus 1976-1994 gap)	change between the two periods in the party gap, 2008 wave left out (independent samples)	Democrats and Republicans (leaners included) answering the item	1976-1994 to 2010-2018	95% CI −7.5 to −0.3		weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pgchg_rac_2224	−7.3	points (2022-2024 party gap minus 2008-2018 party gap)	difference between the pooled 2022-2024 and the 2008-2018 racist party gaps, all interview modes (independent samples)	Democrats and Republicans (leaners included) answering the item	2022 and 2024 against 2008-2018	95% CI −15.5 to 0.9		weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pgchg_rel	+4.9	points (2008-2018 gap minus 1976-1994 gap)	change between the two periods in the party gap (independent samples)	Democrats and Republicans (leaners included) answering the item	1976-1994 to 2008-2018	95% CI 1.9 to 7.8		weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pgdiff_rac_2021_late	−16.0	points (2021 party gap minus 2008-2018 party gap)	difference between the 2021 and 2008-2018 racist party gaps (independent samples)	Democrats and Republicans (leaners included) answering the item	2021 wave against 2008-2018	95% CI −24.0 to −8.0		weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
pgdiff_rac_mil_late	−3.1	points (racist party gap minus militarist party gap)	difference between the racist and militarist party gaps in the share who would allow the speech (same period, standard errors in quadrature)	Democrats and Republicans (leaners included) answering the item	2008-2018 waves pooled	95% CI −6.9 to 0.8		weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
rac_dem_late	56.1%	percent who would allow the speaker to speak	weighted share who would allow the racist speaker to speak	Democrats answering the racist item (leaners included)	2008-2018 waves pooled		4,175	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
rac_int	54.7%	percent who would allow the speaker to speak	weighted share who would allow the speech	adults interviewed by an interviewer (Racist)	2022 and 2024	95% CI 50.0% to 59.4%	1,118	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
rac_mode_diff	−14.0	points (web minus interviewer-administered)	difference in the weighted share who would allow the speech, web against interviewer-administered respondents (independent samples)	adults interviewed in 2022 and 2024	2022 and 2024	95% CI −20.3 to −7.7		weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
rac_mode_diff_abs	14.0	points (interviewer-administered minus web)	difference in the weighted share who would allow the racist speaker, interviewer-administered minus web respondents (independent samples)	adults interviewed in 2022 and 2024	2022 and 2024	95% CI 7.7 to 20.3		weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
rac_mode_diff_dem	−16.4	points (web minus interviewer-administered)	difference in the weighted share who would allow the racist speaker, web against interviewer-administered respondents	Democrats (leaners included) interviewed in 2022 and 2024	2022 and 2024	95% CI −26.2 to −6.5		weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
rac_mode_diff_rep	−12.6	points (web minus interviewer-administered)	difference in the weighted share who would allow the racist speaker, web against interviewer-administered respondents	Republicans (leaners included) interviewed in 2022 and 2024	2022 and 2024	95% CI −23.9 to −1.3		weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
rac_rep_late	63.0%	percent who would allow the speaker to speak	weighted share who would allow the racist speaker to speak	Republicans answering the racist item (leaners included)	2008-2018 waves pooled		2,879	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
rac_rep_lead_late	6.8	points (Republicans minus Democrats)	difference in weighted shares who would allow the racist speaker to speak, Republicans minus Democrats (the party gap with its sign reversed)	Democrats and Republicans (leaners included) answering the item	2008-2018 waves pooled	95% CI 4.0 to 9.7	7,054	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
rac_web	40.7%	percent who would allow the speaker to speak	weighted share who would allow the speech	adults interviewed on the web (Racist)	2022 and 2024	95% CI 36.5% to 44.9%	1,153	weight_use (wtssnrps from 2004, wtssps before)	General Social Survey
web_share_2021	86%	percent of respondents	share interviewed on the web (unweighted count)	respondents who answered the racist speak item	2021		1,310	none (count)	General Social Survey
web_share_2224	51%	percent of respondents	share interviewed on the web (unweighted count)	respondents who answered the racist speak item	2022 and 2024		2,271	none (count)	General Social Survey

Definitions

Percentage point: The unit for the difference between two percentages. A change from one share to another is stated in points, not in percent, so that absolute and relative change are not confused.
Partisan leaner: A 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.
Survey weights: Numbers 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.
Confidence interval: A 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.
Effective sample size: The 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.
Survey mode: How 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.
Linear probability model: A regression of a yes-or-no outcome on predictors, fitted by least squares so that each coefficient is a change in the probability of yes. The book weights it and reports robust standard errors.
HC3 robust standard errors: A way of computing standard errors that stays valid when the spread of the outcome differs across respondents, with a correction for small samples. The book uses it for regressions that have no clusters.

Methods

Sample: respondents to the General Social Survey (cumulative file, 1972-2024) who answered at least one of six speak items (44,878 respondents, 217,422 answers; the racist and militarist items start in 1976, the homosexual man item in 1973, the Communist and anti-religionist items in 1972 and the Muslim clergyman item in 2008; the militarist and homosexual man items were last asked in 2021). Outcomes: 1 if the respondent would allow the speaker to speak in the community, would allow him to teach in a college (for the Communist, would not fire him from a college post), or would not remove his book from the public library; direction is taken from the item wording and checked by the agreement of the three items within each speaker. Groups: Democrats (party identification 0 to 2) and Republicans (4 to 6), leaners included; liberals (self-placement 1 to 3) and conservatives (5 to 7); independents, other-party identifiers and moderates are left out of group comparisons. Weights: weight_use (wtssnrps from 2004, wtssps before). Estimator: weighted share; a gap is the first group minus the second, with the two standard errors combined in quadrature. Adjusted gaps: a weighted linear probability model per speaker and period, on the two compared groups with complete covariates, of the 0 or 100 outcome on a group indicator, age, age squared, degree (five categories), race (three), sex and survey-year effects, with HC3 robust errors; the group coefficient is the adjusted gap. Family values average a respondent's answers over the family's speakers (five speakers asked in every period) for respondents who answered all of them. Intervals: 95% normal intervals on the Kish effective sample size, which are model based and ignore stratification and clustering. Periods: 1976-1994, 1996-2006 and 2008-2018, each pooled across waves; the 2021 wave alone for the family comparison; 2022 and 2024 pooled for the mode comparison. Sensitivity: the 2010-2018 waves (2008 left out) and the 2008 wave alone for the party and ideology gaps, the 2021 wave alone for the racist gap, and the family comparison without the homosexual man. Design check: Taylor-linearized standard errors with strata (survey year by vstrat) and primary sampling units (survey year by vstrat by vpsu), for the period gaps and the mode contrasts only; a stratum with one unit contributes its squared score. Suppression: annual party and ideology cells need 150 respondents and an effective sample of 100; period, mode and family cells need 100 and 100; adjusted models need 100 respondents. Mode: interviewer-administered means in person, by phone or the unlabelled residual mode (code 3); web is self-completion; before 2004 the file records no mode and interviews were in person. The GSS moved mostly online in 2021 (86% of the 2021 respondents on the racist item), and 2022 and 2024 mix interviewer-administered and web respondents.

What these data cannot show

The gaps are differences between groups in one survey question, so they do not show why the groups differ, whether party or ideology shapes tolerance, or whether tolerance shapes party or ideology.
The adjusted gaps hold age, education, race and sex fixed and no other trait, and they assume a linear model, so they compare similar people and do not test what party or ideology does.
Periods pool many waves, so a period value averages over years, and the annual figures have wide intervals.
The intervals use the effective sample size and ignore the sampling design.
Across the 64 period gaps and mode contrasts checked, standard errors that use the survey strata and primary sampling units were between 0.87 and 1.27 times the published ones, so the published intervals are sometimes narrower and sometimes wider.
The annual cells, the adjusted gaps and the family gaps were not checked this way.
Two verdicts sit near the edge.
The homosexual man party gap in 2008 to 2018 is +2.1 (95% CI 0.1 to 4.0) with the published interval and +2.1 (95% CI 0.0 to 4.1) with a design-based one, and the Muslim clergyman mode difference in the party gap is +15.5 (95% CI 0.4 to 30.6) with the published interval and +15.5 (95% CI −3.0 to 34.0) with a design-based one, which includes zero.
The mode of interview is tied to time after 2020.
The 2021 wave has too few interviewer-administered respondents to split, and the 2022 and 2024 samples are small once cut by party or ideology, so the size of the racist gap after 2020 is uncertain.
Two speaker items were not asked after 2021, so the militarist and the homosexual man cannot be followed into the web era.
The battery has limits as a measure of tolerance.
It asks about a fixed list of speakers, and the racist and Muslim clergyman items describe the speech itself, so the levels of the items are not comparable with one another.
Grouping the speakers by which side is expected to object is a coding choice.
Willingness to let a speaker speak in principle is not the same as behavior toward a speaker, and the militarist item asks about speech, not about support for army rule, which Chapter 5.3, Support for democracy (https://vishalsingh.org/politics/support-for-democracy) takes up.
Willingness to justify political violence is a separate question, taken up in Chapter 5.4, Political violence and democratic norms (https://vishalsingh.org/politics/political-violence-and-democratic-norms).
The data hold no measure of dogmatism or need for closure, so they cannot say whether one side is more rigid in the sense of the psychological literature.

References

Crawford, Jarret T.; Pilanski, Jane M. (2014). Political Intolerance, Right and Left. Political Psychology 35(6): 841-851. doi:10.1111/j.1467-9221.2012.00926.x
Stouffer, Samuel A. (1955). Communism, Conformity, and Civil Liberties: A Cross-Section of the Nation Speaks Its Mind. Doubleday.
Sullivan, John L.; Piereson, James; Marcus, George E. (1982). Political Tolerance and American Democracy. University of Chicago Press.
Kish, Leslie (1965). Survey Sampling. John Wiley and Sons.
Tourangeau, Roger; Yan, Ting (2007). Sensitive Questions in Surveys. Psychological Bulletin 133(5): 859-883. doi:10.1037/0033-2909.133.5.859
Jost, John T.; Glaser, Jack; Kruglanski, Arie W.; Sulloway, Frank J. (2003). Political Conservatism as Motivated Social Cognition. Psychological Bulletin 129(3): 339-375. doi:10.1037/0033-2909.129.3.339
Brandt, Mark J.; Reyna, Christine; Chambers, John R.; Crawford, Jarret T.; Wetherell, Geoffrey (2014). The Ideological-Conflict Hypothesis: Intolerance Among Both Liberals and Conservatives. Current Directions in Psychological Science 23(1): 27-34. doi:10.1177/0963721413510932

Provenance

Pipeline version: 1.0.0
Build script: book3/chapters/tolerance-of-speech/build.py (sha256 67504a9be858471faa06530db0a6718b13e536ae67372da310cb0c82654696eb)
Built on: 2026-09-30
Site commit: 16414b2c7257546a7042d0b1734657f208999f60

Inputs

File	Size (bytes)	SHA-256, first 1 MB	SHA-256, last 1 MB
data/gss/gss_all_raw2024.parquet	35,044,704	6f6560d1929e8a100f576317a296ee88f777ca061ab7dbc1471cbdc7b3563538	5c3294f206beb53d9a8387fccb1983bf3088818df090bd8ed3f378adadafcf9b
data/gss/gss_sub2024.parquet	4,925,270	09c4894730735ec7ae5a057c99d97705866b0b4d485a8b14030435944f909fba	fbbacc5d2586b9f6a774f02031f3686c14fa23a0effee92366bb90a6de8e0ad3

Outputs

File	SHA-256
facts.json	858da2ce7d8ae950f57c12e6b46c63a4db276415649d0ddc41e624aa1782bc4e
figures/action_gaps.json	119c9f79083a4a2cd2af2f64a62c97f0d90cc170fa1e822a13e52966ca4e36e9
figures/allow_by_year.json	c3c661faa63796f7625dafbf31d0c361964835049d374d8899b811a17e2bd428
figures/allow_levels.json	e909a91733c66db3c0c1885e69e2e70fd96f4111683b8ffc88709125a53c5e6e
figures/family_gaps.json	5b423b90ba5c086a9346b400be175b2911de47c756a58e72c2a2c1905a8fa8e6
figures/gap_by_speaker.json	ee76d831afbad815e22d18e2584ec2219c59c05e7485ad4fc5efe361f728ea17
figures/mode_comparison.json	fa055ac130e1fd3eaeef2bad9a0e4e390c2ace60633e6cafbed7c5c94371453a
figures/racist_gap_by_year.json	074d4bd2717cea5539dae2500da718a5d057e0f6426f13df2edbc5431f387f71
figures/speaker_items.json	c9a5c507e84fea4ec9244d9e7ffb8650edad1d0b1bd81cbab25e2721a5519259
meta.json	5c09a54f3eef451dc46ed3eee88b68b38d0e546601cc8e4e7e742249feb4acfd
refs.json	c7cbf0b29113b62e9f1a744b46fd382ff1b061ef258c328c0f38d7cb49132365

Machine formats

Report (Markdown): https://vishalsingh.org/politics/tolerance-of-speech/report.md
Every estimate (JSON): https://vishalsingh.org/politics/report/tolerance-of-speech/facts.json
Table 1, figure "speaker_items" (JSON): https://vishalsingh.org/politics/report/tolerance-of-speech/figures/speaker_items.json
Table 2, figure "allow_levels" (JSON): https://vishalsingh.org/politics/report/tolerance-of-speech/figures/allow_levels.json
Table 3, figure "allow_by_year" (JSON): https://vishalsingh.org/politics/report/tolerance-of-speech/figures/allow_by_year.json
Table 4, figure "gap_by_speaker" (JSON): https://vishalsingh.org/politics/report/tolerance-of-speech/figures/gap_by_speaker.json
Table 5, figure "racist_gap_by_year" (JSON): https://vishalsingh.org/politics/report/tolerance-of-speech/figures/racist_gap_by_year.json
Table 6, figure "mode_comparison" (JSON): https://vishalsingh.org/politics/report/tolerance-of-speech/figures/mode_comparison.json
Table 7, figure "family_gaps" (JSON): https://vishalsingh.org/politics/report/tolerance-of-speech/figures/family_gaps.json
Table 8, figure "action_gaps" (JSON): https://vishalsingh.org/politics/report/tolerance-of-speech/figures/action_gaps.json
Index of every report: https://vishalsingh.org/politics/llms.txt; all reports in one file: https://vishalsingh.org/politics/llms-full.txt