Polling conducted this early in a US presidential cycle has historically had little to no predictive power for the election result.

True: published analyses of historical general-election or primary polling about 14 months before a US presidential election (September of the year before) show weak or negligible correlation with, or large errors relative to, the final outcome. Examples include FiveThirtyEight's analyses and academic work such as Wlezien and Erikson's. Mostly true: the research shows low but meaningful predictive power, so 'little' fits but 'none' slightly overstates it. Mixed: the evidence is split, for example early primary polls of well-known front-runners have moderate predictive power while general-election polls do not. Mostly false: the research shows early polls have substantial predictive power in most settings, though not in some. False: the research shows early polls predict outcomes well.

Set before the research began.

1. E1 The Timeline of Presidential Elections University of Chicago Press · 2012

   “Polls from the beginning of the year, they show, have virtually no predictive power.”

2. E2 More on Early Polls Columbia Journalism Review · 26 May 2011

   “The graph shows that polls 300 days [before the vote] have little predictive value at all.”

3. E3 When You Know How Polls Work, You Realize That The Early Polls Don't Matter Washington Monthly · 10 October 2023

   “The margin of error is mainly a cover for the historical evidence that there simply is no statistical correlation between early polls and who wins the election.”

4. E4 Early presidential poll favorites sometimes struggle at finish line National Constitution Center · 22 July 2015

   “Polls that test hypothetical general election matchups at this stage in the cycle are mostly wrong about who will win the White House,”

Wlezien and Erikson's book, as described by its publisher, says polls from the start of the election year have "virtually no predictive power" (E1). A CJR analysis says polls taken 300 days before the vote have "little predictive value at all" (E2). Both of those points are closer to Election Day than September of the year before. Independent commentary also says early polls show no statistical correlation with who wins (E3), and that hypothetical general-election matchups at this stage are "mostly wrong" (E4). None of the evidence shows early polls having substantial predictive power, so the claim of "little to no" predictive power holds.