NonpartisanIndependent, data-first coverage of US elections — polls, forecasts, and prediction markets, refreshed daily.Updated daily.What's new·Methodology·Send feedback
Debrief · 2010 cycle
The 2010 Debrief · post-mortem · permanent archive

How did the polling industry do this cycle?

The 2010 industry MAE landed at 5.2pt — graded across 73 historical races with a known outcome — and 289 polls were analyzed across 108 pollsters. Signed bias +1.67pt: pollsters again systematically over-predicted Republicans.

INDUSTRY MAE
5.2pt
vs midterm avg
SIGNED BIAS
+1.67pt
R-leaning
POLLS · POLLSTERS
289
108 firms
RACES SCORED
73
w/ known outcome
CYCLE200220042006200820102012201420162018202020222024
Industry mean absolute error · 2002–2010
20024.0pt20044.5pt20065.5pt20084.3pt20105.2pt20124.7pt20144.8pt20166.8pt20184.5pt20206.0pt20224.6pt20244.1ptMAE · 2-WAY MARGIN
Per-pollster scorecard · 2010 · sorted by MAE
Signed bias · ±4.7pt
MAE · 0–6.7pt
surveyusa-report
+1.52pt
3.2pt
quinnipiac
−2.11pt
3.3pt
marist-college
+3.85pt
3.9pt
cnntime-magazine
+1.47pt
4.1pt
surveyusa
+3.75pt
4.2pt
survey-usa
+0.46pt
5.2pt
mason-dixon
−2.55pt
5.4pt
public-policy-polling
−1.21pt
6.0pt
rasmussen-reports
+2.24pt
6.2pt
Biggest-miss races · 2010
RaceExpectedActualMissDirection
us-governor-2010-coD+39.00D+14.67-24.3ptD-margin overestimate
us-governor-2010-meR+15.00R+1.7013.3ptR-margin overestimate
us-governor-2010-sdR+11.00R+23.02-12.0ptR-margin underestimate
us-governor-2010-tnR+22.30R+31.95-9.7ptR-margin underestimate
us-senate-2010-nvR+3.50D+5.749.2ptR-overestimate
us-senate-2010-mdD+18.00D+26.448.4ptD-margin underestimate
us-senate-special-2010-wvD+3.00D+10.077.1ptD-margin underestimate
us-senate-2010-laR+12.00R+18.89-6.9ptR-margin underestimate