NonpartisanIndependent, data-first coverage of US elections — polls, forecasts, and prediction markets, refreshed daily.Updated daily.What's new·Methodology·Send feedback
Pollster · 55 of 1769

Mitchell Research

GRADE CWEIGHT TIER T3NEUTRAL LEAN</> Embed
Polls scored
8
Mean error
5.8 pt
Signed bias
R+3.77
Aggregation weight
1.09

Pollster vs prediction market

From 102 race-polls · 27 matched markets
Pre-poll P(D) (avg)
73.5%
Market price the day before each poll dropped
Post-poll market shift
+0.1 pt
Average move within 24h of release (D+ = market moved toward D)
Market alignment
+5.8 pt
Net direction markets moved toward Mitchell Research's read
Composite quality score
0–100 · weighted of 5 measured components
65
of 100 · tier · strong

Strong on the components we measure; one or two areas pull the composite below the elite threshold.

Aggregator weight55
Verification rate48
Sample size median67
Citation rate100
Outlier rate (inv.)not yet measured
Rater alignment76
Pollster vs editorial consensus
decoupled

Across 49 graded races, Mitchell Research's released topline differed from the median editorial-rater consensus by an average of +0.6pp. Decoupled from editorial consensus — directionally different from Cook/Sabato/IE on multiple races.

Mean delta
+0.6pp
Median delta
-1.0pp
Std-dev
6.6pp
N graded
49
Recent activity
Total all time
102
Last 30d
11
Last 90d
30
Last 180d
30

Last poll 6/16/2026 (20d ago)· status active

Per-cycle debrief appearances
CycleRankPollsMAESigned bias
2020#8 / 2034.99 ptD+4.99 pp
2018#4 / 2562.68 ptR+1.03 pp
2014#3 / 2142.53 ptD+2.06 pp
Recent polls