A 2024 Iowa poll put presidential candidates Kamala Harris and Donald Trump within three points, but author Nate Silver said the result was more likely wrong than right amid conflicting polls.
The Des Moines Register/Mediacom Iowa Poll, conducted by Selzer & Co., found Harris leading Trump 47% to 44% among 808 likely voters. Writing for Nate Silver’s website on Nov. 3, 2024, Silver said he thought Selzer’s poll was more likely to be wrong than right.
Silver put the margin of error for the difference between the candidates at 6.6 percentage points. In his account, sampling error alone would place the underlying result between a 3.4-point Trump lead and a 9.6-point Harris lead in 95 of 100 cases. He described this as a theoretical range; the calculation did not cover every source of polling error.
One omitted factor was nonresponse. Silver said most people contacted by pollsters did not complete a survey. Other Iowa measures favored Trump: an Emerson College poll showed him ahead by nine points, while Silver’s average of four Iowa polls put his lead at 3.4 points. Silver said the average fell within the narrow overlap between Selzer’s and Emerson’s estimated ranges. He also characterized Emerson as a pollster that often followed prevailing expectations and rarely published results that ran against them.
Silver’s model raised Harris’ chance of winning Iowa from 9% to 17% after the survey. Polymarket showed a larger increase, with Harris’ implied probability rising 12 percentage points to 18%, Silver reported. He described those odds as roughly 5-to-1 against her.
Selzer’s earlier Iowa polls had shown Trump 18 points ahead of presidential candidate Joe Biden in June and Harris four points behind Trump in September. Iowa had backed Trump by eight points in 2020. Silver also cited Selzer’s record of going against conventional expectations and being right. In the Nov. 5, 2024, election, Trump received 927,019 votes in Iowa to Harris’ 707,278.
