Inferences from early voting data have a fairly bad track record in predicting election outcomes, including in 2016.

Judge as of November 2016. True: evidence from analysts or academics shows that predictions based on early-vote data have a historically poor or unreliable record, and that prominent early-vote inferences in 2016 pointed the wrong way, e.g. claims that early-vote data showed Clinton winning Florida or North Carolina. Mostly true: the general poor record is supported and 2016 inferences were mostly wrong, but some notable early-vote readings were accurate (e.g. Nevada). Mixed: the evidence on the track record is balanced, or 2016 early-vote inferences were about as often right as wrong. Mostly false: early-vote inferences have a reasonably good record and the 2016 failures were isolated. False: early-vote inferences have been generally accurate, including in 2016.

Set before the research began.

1. E1 Early Voting a Poor Predictor of Final Results | RealClearPolitics RealClearPolitics · 2 November 2016

   “Second, predictions from early voting have a decidedly mixed track record.”

2. E2 Early Voting a Poor Predictor of Final Results | RealClearPolitics RealClearPolitics · 2 November 2016

   “But if predictions from 2012 were iffy, predictions from 2010 and 2014 were awful.”

3. E3 Early voting in Nevada looks very good for Democrats Vox · 3 November 2016

   “Now, a gigantic caveat: Early voting does not have a long history in the United States. Any time you make assertions about presidential election outcomes, you are stuck dealing with small sample sizes.”

4. E4 Do the Early-Voting Patterns in Florida Mean Clinton Will Win? New York Magazine (Intelligencer) · 7 November 2016

   “All in all, Florida will remain a mystery until Tuesday.”

5. E5 So far, early voting looks like good news for Hillary Clinton Vox · 27 October 2016

   “The relationship is positive, but it’s pretty noisy. In other words, knowing how a party is doing in early voting doesn’t tell you much about how it will do once all the votes are counted.”

6. E6 So far, early voting looks like good news for Hillary Clinton Vox · 27 October 2016

   “Democrats maintained substantial leads among early voters in North Carolina, Louisiana and West Virginia, and were trailing by a relatively narrow margin in Oklahoma, but still lost those states when all the votes were counted.”

7. E7 Early voting: Record levels in 2016 may give Clinton edge Associated Press (via Fox 4 News) · 7 November 2016

   “In Florida, a record 6.4 million early ballots are already in, or more than three-fourths of the expected vote. Democrats are ahead, 39.9 percent to 38.5 percent.”

8. E8 Minorities are a huge force in the early vote — and this year Latinos are making their mark Vox · 7 November 2016

   “In Nevada, a surge in the last few days of early voting made the state, according to state politics expert Jon Ralston, all but unwinnable for Republicans.”

The evidence supports a mixed-to-poor general track record for early-vote predictions (E1, E2, E5, E6). It also shows that 2016 early-vote readings favoured Democrats in Florida and Nevada (E3, E7, E8). However, no evidence item gives the 2016 final results, so the evidence cannot show whether the 2016 early-vote inferences were right or wrong, as the criteria require.