Mark Kelly vs Martha McSally

Poll was fielded 2015 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 11 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Poll was fielded 2016 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 11 historical scored polls the average error vs. actual margin is D+5.1pt — a measurable house effect.

Poll was fielded 2016 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Commissioned by NBC News, an independent media or academic sponsor. Independent sponsorship doesn't guarantee accuracy but rules out direct campaign funding.

Across 12 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Across 15 historical scored polls the average error vs. actual margin is D+7.7pt — a measurable house effect.

Aggregation weight is 0.49 (1.0 ≈ average). Weight combines historical accuracy, methodology, and recency.

Sample size of 333 respondents implies a margin of error around ±5.4pt — wider than typical.

Poll was fielded 2016 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Poll was fielded 2016 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

Stdev of (poll − rater consensus) across 127 historical paired polls exceeds 8pt — this pollster's results swing widely relative to where the editors place the same races.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 50 historical scored polls this pollster's average error vs. actual margin is D+1.0pt — within ±1pt.

Poll was fielded 2017 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 234 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Across 117 historical scored polls this pollster's average error vs. actual margin is D+0.7pt — within ±1pt.

Across 29 historical scored polls the average error vs. actual margin is D+5.0pt — a measurable house effect.

Poll was fielded 2017 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 62 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Poll was fielded 2018 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 10 historical scored polls the average error vs. actual margin is R+0.2pt.

Across 15 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Poll was fielded 2018 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 11 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Pollster has no historical polls with a known outcome in our database, so accuracy and bias cannot be measured.

Poll was fielded 2018 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 17 historical scored polls the average error vs. actual margin is D+4.4pt — a measurable house effect.

Poll was fielded 2018 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Commissioned by CNN, an independent media or academic sponsor. Independent sponsorship doesn't guarantee accuracy but rules out direct campaign funding.

Pollster has only 4 historical polls with a known outcome to score against — too few for stable accuracy estimates.

Poll was fielded 2019 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 33 historical scored polls the average error vs. actual margin is D+2.9pt.

Poll was fielded 2020 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

Across 49 historical paired polls, this pollster's margins don't track the editorial consensus (Pearson r < 0.5). Independence from consensus is informational — could indicate genuine methodological independence or noisy results.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Pollster has no historical polls with a known outcome in our database, so accuracy and bias cannot be measured.

Poll was fielded 2021 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 15 historical scored polls the average error vs. actual margin is D+7.7pt — a measurable house effect.

Aggregation weight is 0.49 (1.0 ≈ average). Weight combines historical accuracy, methodology, and recency.

Sample size of 286 respondents implies a margin of error around ±5.8pt — wider than typical.

Poll was fielded 2022 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Pollster has no historical polls with a known outcome in our database, so accuracy and bias cannot be measured.

Poll was fielded 2023 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Poll was fielded 2023 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 15 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Pollster has no historical polls with a known outcome in our database, so accuracy and bias cannot be measured.

Poll was fielded 2023 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Registered-voter sample — broader than likely-voter screens; typical for early-cycle polling before turnout signals stabilize.

Pollster has only 3 historical polls with a known outcome to score against — too few for stable accuracy estimates.

Poll was fielded 2024 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Pollster has no historical polls with a known outcome in our database, so accuracy and bias cannot be measured.

Poll was fielded 2024 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Commissioned by The Salt Lake Tribune, an independent media or academic sponsor. Independent sponsorship doesn't guarantee accuracy but rules out direct campaign funding.

Pollster has no historical polls with a known outcome in our database, so accuracy and bias cannot be measured.

Poll was fielded 2026 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Poll was fielded 2027 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 11 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Across 29 historical scored polls the average error vs. actual margin is D+5.0pt — a measurable house effect.

Poll was fielded 2028 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 62 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Pollster has only 4 historical polls with a known outcome to score against — too few for stable accuracy estimates.

Poll was fielded 2029 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Sample size of 232 respondents implies a margin of error around ±6.4pt — wider than typical.

Poll was fielded 2029 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Poll was fielded 2029 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Poll was fielded 2030 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 10 historical scored polls the average error vs. actual margin is R+0.2pt.

Across 15 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Pollster has no historical polls with a known outcome in our database, so accuracy and bias cannot be measured.

Poll was fielded 2032 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Commissioned by CBS News, an independent media or academic sponsor. Independent sponsorship doesn't guarantee accuracy but rules out direct campaign funding.

Poll was fielded 2034 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 11 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Across 19 historical scored polls the average error vs. actual margin is D+5.0pt — a measurable house effect.

Poll was fielded 2035 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Registered-voter sample — broader than likely-voter screens; typical for early-cycle polling before turnout signals stabilize.

Across 30 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Across 29 historical scored polls the average error vs. actual margin is D+5.0pt — a measurable house effect.

Poll was fielded 2037 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 62 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Methodology not disclosed publicly; declined to participate in AAPOR Transparency Initiative. Included for transparency rather than weight-zeroing — many forecasters still include their polls — but the methodology opacity is worth surfacing. Source: https://en.wikipedia.org/wiki/Trafalgar_Group

Poll was fielded 2039 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

Across 7 historical polls where ≥2 editorial raters had rated the race (Cook / Sabato / IE / 538 / DDHQ etc.), this pollster's D-margin averaged 4.1pt MORE R-favorable than the rater consensus. Systematic R-lean relative to conventional wisdom.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Poll was fielded 2040 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 15 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Poll was fielded 2041 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 11 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Pollster has no historical polls with a known outcome in our database, so accuracy and bias cannot be measured.

Poll was fielded 2042 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Pollster has no historical polls with a known outcome in our database, so accuracy and bias cannot be measured.

Poll was fielded 2043 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Poll was fielded 2043 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 10 historical scored polls the average error vs. actual margin is R+0.2pt.

Across 15 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Pollster has no historical polls with a known outcome in our database, so accuracy and bias cannot be measured.

Poll was fielded 2043 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Sample size of 296 respondents implies a margin of error around ±5.7pt — wider than typical.

Poll was fielded 2044 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Poll was fielded 2045 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 11 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Pollster has no historical polls with a known outcome in our database, so accuracy and bias cannot be measured.

Poll was fielded 2045 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Poll was fielded 2046 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

Stdev of (poll − rater consensus) across 26 historical paired polls exceeds 8pt — this pollster's results swing widely relative to where the editors place the same races.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 11 historical scored polls the average error vs. actual margin is R+1.0pt.

Across 29 historical scored polls the average error vs. actual margin is D+5.0pt — a measurable house effect.

Poll was fielded 2047 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 62 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Poll was fielded 2048 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 15 historical scored polls the average error vs. actual margin is D+0.1pt.

Across 55 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Pollster has no historical polls with a known outcome in our database, so accuracy and bias cannot be measured.

Poll was fielded 2050 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Poll was fielded 2050 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

Stdev of (poll − rater consensus) across 127 historical paired polls exceeds 8pt — this pollster's results swing widely relative to where the editors place the same races.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 50 historical scored polls this pollster's average error vs. actual margin is D+1.0pt — within ±1pt.

Across 29 historical scored polls the average error vs. actual margin is D+5.0pt — a measurable house effect.

Poll was fielded 2050 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 62 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Poll was fielded 2056 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

Stdev of (poll − rater consensus) across 127 historical paired polls exceeds 8pt — this pollster's results swing widely relative to where the editors place the same races.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 50 historical scored polls this pollster's average error vs. actual margin is D+1.0pt — within ±1pt.

Sample size of 262 respondents implies a margin of error around ±6.1pt — wider than typical.

Poll was fielded 2058 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Pollster has no historical polls with a known outcome in our database, so accuracy and bias cannot be measured.

Poll was fielded 2058 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 29 historical scored polls the average error vs. actual margin is D+5.0pt — a measurable house effect.

Poll was fielded 2058 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 62 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Pollster has no historical polls with a known outcome in our database, so accuracy and bias cannot be measured.

Poll was fielded 2059 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Poll was fielded 2061 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 11 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Across 29 historical scored polls the average error vs. actual margin is D+5.0pt — a measurable house effect.

Poll was fielded 2061 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 62 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Pollster has no historical polls with a known outcome in our database, so accuracy and bias cannot be measured.

Poll was fielded 2062 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

Stdev of (poll − rater consensus) across 14 historical paired polls exceeds 8pt — this pollster's results swing widely relative to where the editors place the same races.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Commissioned by Politico, an independent media or academic sponsor. Independent sponsorship doesn't guarantee accuracy but rules out direct campaign funding.

Poll was fielded 2062 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 48 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Across 19 historical scored polls the average error vs. actual margin is D+5.0pt — a measurable house effect.

Poll was fielded 2063 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Registered-voter sample — broader than likely-voter screens; typical for early-cycle polling before turnout signals stabilize.

Across 30 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Poll was fielded 2063 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 11 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Across 29 historical scored polls the average error vs. actual margin is D+5.0pt — a measurable house effect.

Poll was fielded 2064 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 62 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Pollster has no historical polls with a known outcome in our database, so accuracy and bias cannot be measured.

Poll was fielded 2065 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Pollster has no historical polls with a known outcome in our database, so accuracy and bias cannot be measured.

Poll was fielded 2065 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Registered-voter sample — broader than likely-voter screens; typical for early-cycle polling before turnout signals stabilize.

Across 33 historical scored polls the average error vs. actual margin is D+2.9pt.

Poll was fielded 2067 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

Across 49 historical paired polls, this pollster's margins don't track the editorial consensus (Pearson r < 0.5). Independence from consensus is informational — could indicate genuine methodological independence or noisy results.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Pollster has no historical polls with a known outcome in our database, so accuracy and bias cannot be measured.

Poll was fielded 2067 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Poll was fielded 2068 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 15 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Pollster has no historical polls with a known outcome in our database, so accuracy and bias cannot be measured.

Poll was fielded 2070 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Sample size of 1,600 respondents implies a margin of error around ±2.5pt — tighter than the ~600-respondent typical.

Poll was fielded 2072 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Poll was fielded 2074 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 48 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Across 29 historical scored polls the average error vs. actual margin is D+5.0pt — a measurable house effect.

Poll was fielded 2074 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 62 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Across 19 historical scored polls the average error vs. actual margin is D+2.6pt.

Poll was fielded 2077 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Pollster has no historical polls with a known outcome in our database, so accuracy and bias cannot be measured.

Poll was fielded 2078 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 29 historical scored polls the average error vs. actual margin is D+5.0pt — a measurable house effect.

Poll was fielded 2084 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 62 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Pollster has no historical polls with a known outcome in our database, so accuracy and bias cannot be measured.

Poll was fielded 2091 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 29 historical scored polls the average error vs. actual margin is D+5.0pt — a measurable house effect.

Poll was fielded 2094 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 62 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Poll was fielded 2099 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 234 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Across 117 historical scored polls this pollster's average error vs. actual margin is D+0.7pt — within ±1pt.

Poll was fielded 2100 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 29 historical scored polls the average error vs. actual margin is D+5.0pt — a measurable house effect.

Poll was fielded 2104 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 62 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Poll was fielded 2105 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 15 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Poll was fielded 2105 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

Across 10 historical polls where ≥2 editorial raters had rated the race (Cook / Sabato / IE / 538 / DDHQ etc.), this pollster's D-margin averaged 4.1pt MORE R-favorable than the rater consensus. Systematic R-lean relative to conventional wisdom.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Aggregation weight is 3.36 (1.0 ≈ average). Weight is a derived metric combining historical accuracy, methodology, and recency.

Poll was fielded 2107 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

Stdev of (poll − rater consensus) across 127 historical paired polls exceeds 8pt — this pollster's results swing widely relative to where the editors place the same races.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 50 historical scored polls this pollster's average error vs. actual margin is D+1.0pt — within ±1pt.

Sample size of 365 respondents implies a margin of error around ±5.1pt — wider than typical.

Poll was fielded 2114 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 29 historical scored polls the average error vs. actual margin is D+5.0pt — a measurable house effect.

Poll was fielded 2114 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 62 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Across 17 historical scored polls the average error vs. actual margin is D+4.4pt — a measurable house effect.

Poll was fielded 2116 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Registered-voter sample — broader than likely-voter screens; typical for early-cycle polling before turnout signals stabilize.

Commissioned by CNN, an independent media or academic sponsor. Independent sponsorship doesn't guarantee accuracy but rules out direct campaign funding.

Poll was fielded 2117 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 48 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Across 29 historical scored polls the average error vs. actual margin is D+5.0pt — a measurable house effect.

Poll was fielded 2117 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 62 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Poll was fielded 2118 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

Across 368 historical paired polls, this pollster's margins don't track the editorial consensus (Pearson r < 0.5). Independence from consensus is informational — could indicate genuine methodological independence or noisy results.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Across 72 historical scored polls the average error vs. actual margin is D+1.1pt.

Across 11 historical scored polls the average error vs. actual margin is D+5.1pt — a measurable house effect.

Poll was fielded 2118 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Registered-voter sample — broader than likely-voter screens; typical for early-cycle polling before turnout signals stabilize.

Across 12 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Pollster has no historical polls with a known outcome in our database, so accuracy and bias cannot be measured.

Poll was fielded 2124 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Sample size of 345 respondents implies a margin of error around ±5.3pt — wider than typical.

Poll was fielded 2128 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 16 historical scored polls the average error vs. actual margin is D+3.4pt — a measurable house effect.

Poll was fielded 2130 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 23 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Poll was fielded 2133 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 15 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Poll was fielded 2141 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 10 historical scored polls the average error vs. actual margin is R+0.2pt.

Across 15 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Sample size of 311 respondents implies a margin of error around ±5.6pt — wider than typical.

Poll was fielded 2142 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 33 historical scored polls the average error vs. actual margin is D+2.9pt.

Poll was fielded 2143 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

Across 49 historical paired polls, this pollster's margins don't track the editorial consensus (Pearson r < 0.5). Independence from consensus is informational — could indicate genuine methodological independence or noisy results.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Pollster has no historical polls with a known outcome in our database, so accuracy and bias cannot be measured.

Poll was fielded 2146 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

Across 31 historical paired polls, this pollster's margins don't track the editorial consensus (Pearson r < 0.5). Independence from consensus is informational — could indicate genuine methodological independence or noisy results.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Commissioned by Politico, an independent media or academic sponsor. Independent sponsorship doesn't guarantee accuracy but rules out direct campaign funding.

Across 18 historical scored polls the average error vs. actual margin is D+3.7pt — a measurable house effect.

Poll was fielded 2147 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

Stdev of (poll − rater consensus) across 30 historical paired polls exceeds 8pt — this pollster's results swing widely relative to where the editors place the same races.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 18 historical scored polls the average error vs. actual margin is D+3.7pt — a measurable house effect.

Poll was fielded 2147 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

Stdev of (poll − rater consensus) across 30 historical paired polls exceeds 8pt — this pollster's results swing widely relative to where the editors place the same races.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Poll was fielded 2153 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 48 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Pollster has only 3 historical polls with a known outcome to score against — too few for stable accuracy estimates.

Poll was fielded 2154 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Registered-voter sample — broader than likely-voter screens; typical for early-cycle polling before turnout signals stabilize.

Poll was fielded 2155 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Registered-voter sample — broader than likely-voter screens; typical for early-cycle polling before turnout signals stabilize.

Across 19 historical scored polls the average error vs. actual margin is D+2.6pt.

Poll was fielded 2168 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Registered-voter sample — broader than likely-voter screens; typical for early-cycle polling before turnout signals stabilize.

Pollster has only 3 historical polls with a known outcome to score against — too few for stable accuracy estimates.

Poll was fielded 2179 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Poll was fielded 2190 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 15 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Poll was fielded 2223 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 15 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Across 11 historical scored polls the average error vs. actual margin is D+5.1pt — a measurable house effect.

Poll was fielded 2247 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Registered-voter sample — broader than likely-voter screens; typical for early-cycle polling before turnout signals stabilize.

Sample size of 2,523 respondents implies a margin of error around ±2.0pt — tighter than the ~600-respondent typical.

Across 12 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Across 19 historical scored polls the average error vs. actual margin is D+5.0pt — a measurable house effect.

Poll was fielded 2248 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Registered-voter sample — broader than likely-voter screens; typical for early-cycle polling before turnout signals stabilize.

Across 30 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Pollster has no historical polls with a known outcome in our database, so accuracy and bias cannot be measured.

Poll was fielded 2251 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Registered-voter sample — broader than likely-voter screens; typical for early-cycle polling before turnout signals stabilize.

Poll was fielded 2258 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 15 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Poll was fielded 2259 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

Across 368 historical paired polls, this pollster's margins don't track the editorial consensus (Pearson r < 0.5). Independence from consensus is informational — could indicate genuine methodological independence or noisy results.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Across 72 historical scored polls the average error vs. actual margin is D+1.1pt.

Pollster has only 3 historical polls with a known outcome to score against — too few for stable accuracy estimates.

Poll was fielded 2282 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Pollster has no historical polls with a known outcome in our database, so accuracy and bias cannot be measured.

Poll was fielded 2298 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

Stdev of (poll − rater consensus) across 14 historical paired polls exceeds 8pt — this pollster's results swing widely relative to where the editors place the same races.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Commissioned by Politico, an independent media or academic sponsor. Independent sponsorship doesn't guarantee accuracy but rules out direct campaign funding.

Poll was fielded 2318 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

Across 368 historical paired polls, this pollster's margins don't track the editorial consensus (Pearson r < 0.5). Independence from consensus is informational — could indicate genuine methodological independence or noisy results.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Across 72 historical scored polls the average error vs. actual margin is D+1.1pt.

Poll was fielded 2349 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 15 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Poll was fielded 2386 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Registered-voter sample — broader than likely-voter screens; typical for early-cycle polling before turnout signals stabilize.

Across 234 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Across 117 historical scored polls this pollster's average error vs. actual margin is D+0.7pt — within ±1pt.

Across 18 historical scored polls the average error vs. actual margin is D+3.7pt — a measurable house effect.

Poll was fielded 2416 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

Stdev of (poll − rater consensus) across 30 historical paired polls exceeds 8pt — this pollster's results swing widely relative to where the editors place the same races.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Registered-voter sample — broader than likely-voter screens; typical for early-cycle polling before turnout signals stabilize.

Pollster has no historical polls with a known outcome in our database, so accuracy and bias cannot be measured.

Poll was fielded 2432 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Registered-voter sample — broader than likely-voter screens; typical for early-cycle polling before turnout signals stabilize.

Commissioned by Politico, an independent media or academic sponsor. Independent sponsorship doesn't guarantee accuracy but rules out direct campaign funding.

Poll was fielded 2461 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 15 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Pollster has no historical polls with a known outcome in our database, so accuracy and bias cannot be measured.

Poll was fielded 2475 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Poll was fielded 2565 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 15 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Poll was fielded 2643 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 15 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.

Poll was fielded 2643 days ago. Older polls reflect an earlier state of the race and are weighted down in the rolling average.

auto-checked on 5/4/2026. Deterministic checks confirmed bounds + structure (percentages in [0,100], totals plausibly close to 100%, dates ordered). Source page itself was not re-fetched.

Likely-voter screen — pollster filters respondents by reported turnout intent. Generally tighter for predicting elections than registered- or adult-population polls.

Across 15 historical paired polls, this pollster's D-margins track the rater consensus (Cook / Sabato / IE / 538 / DDHQ etc.) with high correlation (r ≥ 0.85) and small mean delta (|Δ| ≤ 2pt). Their results align with conventional wisdom.