online opt-in
Ipsos
Reuters/Ipsos Poll: Issues Survey March 2023 (Large Sample) · Mar 14–20, 2023 · US national
Methodology
- Pollster
- Ipsos commercial
- Sponsor
- Thomson Reuters
- Commissioned by
- Thomson Reuters
- Fielded
- Mar 14–20, 2023
- Population
- adults
- Sample size
- 4,410
- Sample frame
- online opt-in
- Panel
- Ipsos online panel plus partner panels and river sampling
- Mode
- online - self-administered web survey; sample randomly drawn from Ipsos's online panel, partner online panel sources, and 'river' sampling; does not rely on a population frame in the traditional sense
- Margin of error
- Credibility interval of +/- 1.8 percentage points for all respondents; design-effect adjusted confidence interval +/- 3.3 points. Party CIs: Democrats +/- 2.6, Republicans +/- 3.0, independents +/- 5.7. Ipsos states statistical margins of error are not applicable to online non-probability polls.
- Weighting
- Raking-ratio adjustment to U.S. population targets from the 2022 American Community Survey; post-hoc weights on gender, age, race/ethnicity, region, and education.
- Geography
- US national
- Base notes
- 2,111 Democrats, 1,627 Republicans, 448 independents. Methodology page cites n=4,408 in the design-effect calculation while the topline tables use 4,410.
Sources
Verification note. Full 24-page Reuters/Ipsos topline & methodology PDF retrieved and text-extracted. This is the densest single Reuters/Ipsos fielding for trans content in the corpus: it re-fields the TM2062Y21 sports and youth-care trend battery, adds a drag-ban item, adds a school-pronoun item, and adds candidate-support items. Note the sharp restrictive movement on TM2062Y21_1 between April 2022 (46% agree) and March 2023 (55% agree) on the identical wording. FRAME CAUTION: online opt-in / river sample; credibility interval, not margin of error.
Figures below are transcribed from these documents. Where they disagree with the source, the source governs.
Questions asked (7)
banning doctors from providing transition-related treatment to minors
To what extent do you agree or disagree with the following statements?
To what extent do you agree or disagree with the following statements? - Doctors should be banned from providing transgender minors with treatments that help change their physical characteristics to better resemble their gender identity.
Wording notes: agree/disagree format with restrictive polarity - acquiescence bias pushes toward the ban | identical wording to the April 2022 fielding, so the 9-point increase in agreement is a real trend and not a wording artifact
tilts_restrictive — Same construction as April 2022: acquiescence runs restrictive, and the description of the care emphasizes physical alteration without distinguishing reversible from irreversible interventions. Its value is as a fixed-wording trend point - 46% agree in April 2022, 55% in March 2023 - measured on a constant instrument.
Subgroup figures (15)
transgender girls in women's sports
To what extent do you agree or disagree with the following statements?
To what extent do you agree or disagree with the following statements? - Transgender girls should be allowed to participate in women's sports.
Wording notes: identical wording to the April 2022 fielding - a clean trend point | no level of competition specified
neutral — Same neutral construction as the 2022 fielding, and again paired with the trans-boys mirror. Agreement fell from 35% to 32% and strong disagreement rose from 32% to 36% across the eleven months.
Subgroup figures (15)
transgender boys in men's sports
To what extent do you agree or disagree with the following statements?
To what extent do you agree or disagree with the following statements? - Transgender boys should be allowed to participate in men's sports.
Wording notes: mirror item to the transgender-girls version
neutral — Constant wording across fieldings. The trans-girls / trans-boys gap persists and widens slightly: in March 2023, 36% strongly disagree for trans girls versus 30% for trans boys.
Subgroup figures (9)
banning under-18s from drag performances
To what extent do you agree or disagree with the following statements?
To what extent do you agree or disagree with the following statements? - I support efforts to ban children under 18 from attending events featuring drag performers
Wording notes: double-embedded: the respondent is asked to agree with a first-person statement of support for a ban, which layers acquiescence bias on top of restrictive polarity | 'events featuring drag performers' is undifferentiated - it covers a public library story hour and an adult cabaret alike | added to the TM2062Y21 battery in 2023, tracking the wave of state drag-restriction bills
tilts_restrictive — The stem is phrased as 'I support efforts to ban...', so agreeing is doubly easy - acquiescence and the restrictive direction point the same way. More importantly, collapsing all drag events into one category means a respondent who objects only to age-inappropriate adult performance registers identically to one who objects to a library reading hour, which is precisely the elision the underlying legislation relies on.
Subgroup figures (9)
laws requiring school staff to use birth-sex pronouns
To what extent do you agree or disagree with the following statements?
To what extent do you agree or disagree with the following statements? - Laws requiring faculty and staff at K-12 schools only use the pronouns of a student's biological sex at birth, not their gender identity
Wording notes: the printed stem says 'agree or disagree' but the response options are support/oppose - an internal inconsistency in the Ipsos instrument itself | 'biological sex at birth' is contested phrasing, though it is set against 'gender identity' rather than asserted as the only real category
unclear — Substantively the item is even-handed - it names both 'biological sex at birth' and 'gender identity' and offers a symmetric four-point scale plus 'not sure'. The problem is instrument hygiene rather than slant: the stem asks respondents to agree/disagree while the scale asks them to support/oppose, so it is not certain what task respondents believed they were performing. That mismatch is recorded here rather than resolved.
Subgroup figures (12)
vote effect of a candidate supporting under-18 gender transition care
In the election for president in 2024, are you more or less likely to support a candidate who does the following?
In the election for president in 2024, are you more or less likely to support a candidate who does the following? - Supports allowing transgender children under the age of 18 to get medical treatment related to gender identity and gender transitioning
Much less likely
37%
Somewhat less likely
24%
Somewhat more likely
23%
Much more likely
16%
Wording notes: no neutral 'no difference' or 'not sure' option - every respondent is forced into a directional answer, which mechanically inflates apparent salience | measures candidate-support effect, not policy opinion
unclear — The description of the policy is neutral and accurate. The design flaw is the absent midpoint: with no 'makes no difference' option, respondents indifferent to the issue must nonetheless pick a direction, so this item cannot distinguish intensity from indifference and systematically overstates how much the issue moves votes.
Subgroup figures (12)
vote effect of a candidate supporting teachers discussing sexual orientation or gender identity
In the election for president in 2024, are you more or less likely to support a candidate who does the following?
In the election for president in 2024, are you more or less likely to support a candidate who does the following? - Supports allowing teachers to discuss sexual orientation or gender identity at school
Much less likely
31%
Somewhat more likely
28%
Somewhat less likely
22%
Much more likely
19%
Wording notes: no neutral option | bundles sexual orientation with gender identity | no grade level specified
unclear — Neutral in tone but structurally unable to register indifference, same as its neighbour, and it fuses two distinct topics. Worth recording because the split is much closer here (47% more likely / 53% less likely) than on the youth-medical item (39/61) in the same fielding.
Subgroup figures (9)