NonpartisanIndependent, data-first coverage of US elections — polls, forecasts, and prediction markets, refreshed daily.Updated daily.What's new·Methodology·Send feedback
Debrief · 2012 cycle
The 2012 Debrief · post-mortem · permanent archive

How did the polling industry do this cycle?

The 2012 industry MAE landed at 4.7pt — graded across 514 historical races with a known outcome — and 200 polls were analyzed across 88 pollsters. Signed bias +3.54pt: pollsters again systematically over-predicted Republicans.

INDUSTRY MAE
4.7pt
vs midterm avg
SIGNED BIAS
+3.54pt
R-leaning
POLLS · POLLSTERS
200
88 firms
RACES SCORED
514
w/ known outcome
CYCLE200220042006200820102012201420162018202020222024
Industry mean absolute error · 2002–2012
20024.0pt20044.5pt20065.5pt20084.3pt20105.2pt20124.7pt20144.8pt20166.8pt20184.5pt20206.0pt20224.6pt20244.1ptMAE · 2-WAY MARGIN
Per-pollster scorecard · 2012 · sorted by MAE
Signed bias · ±8.7pt
MAE · 0–8.7pt
kcts-9washington-poll
−2.31pt
2.4pt
surveyusa
+2.74pt
2.9pt
reutersipsos
+1.64pt
3.0pt
nbcwsjmarist
+3.56pt
3.6pt
pharos-research
−0.58pt
4.1pt
quinnipiac
+4.69pt
4.7pt
rasmussen-reports
+5.04pt
6.1pt
weaskamerica
+6.72pt
6.7pt
gravis-marketing
+7.64pt
7.6pt
mason-dixon
+8.22pt
8.2pt
Biggest-miss races · 2012
RaceExpectedActualMissDirection
us-senate-2012-neR+3.00R+15.54-12.5ptR-margin underestimate
us-senate-2012-moD+4.00D+15.7011.7ptD-margin underestimate
us-governor-2012-nhD+3.00D+12.099.1ptD-margin underestimate
us-senate-2012-ctD+4.00D+12.468.5ptD-margin underestimate
us-senate-2012-miD+13.00D+20.827.8ptD-margin underestimate
us-senate-2012-hiD+18.50D+25.206.7ptD-margin underestimate
us-governor-2012-utR+46.00R+40.835.2ptR-margin overestimate
us-senate-2012-flD+8.00D+13.005.0ptD-margin underestimate