New Emerson poll shows a tight TN-7 race, but key data points don’t match the electorate

NASHVILLE, Tenn. — A new Emerson College Polling/The Hill survey reports a two-point race in Tennessee’s 7th Congressional District, showing Matt Van Epps at 49 percent and Aftyn Behn at 47 percent after undecided respondents were pushed to choose. The result falls within the poll’s ±3.9-point margin of error and has quickly prompted national attention.

A closer look at the underlying numbers, however, reveals several gaps between the poll’s sample and the electorate that has actually been voting in the district. Those differences help explain why the survey presents a tighter contest than the district’s recent history or its early-vote patterns suggest.

The poll, conducted Nov. 22–24 among 600 likely voters, portrays a district that is significantly less Republican than past results. Respondents recalled their 2024 presidential vote as Trump 53.6 percent and Harris 40.9 percent — a margin of about 13 points. But Trump carried the redrawn TN-7 by roughly 22 points last year. That nine- to ten-point gap indicates the sample is more Democratic-leaning than the real 2024 electorate.

Party identification shows a similar pattern. Emerson’s respondents identify as 42.9 percent Republican and 32.2 percent Democratic. A Republican advantage of only 10 points does not match a seat that has consistently elected Republicans by large margins and is expected to draw an older, habitual-voting electorate in a December special election.

The early-vote findings diverge even more sharply from what TNPOLITICO has tracked daily. Emerson reports that 42 percent of respondents say they have already voted early or absentee, and among those voters, Behn leads 56 percent to 42 percent. That pattern is not reflected in the historical county-level voting tendencies or in the turnout composition observed so far.

Davidson County, where Behn must post substantial margins to offset the district’s Republican-leaning counties, continues to make up only the mid-20-percent range of the early electorate. That is well below the 45-percent benchmark Democrats would need for a competitive path. Meanwhile, rural and exurban counties — including Dickson, Cheatham, and Stewart — are producing turnout consistent with a traditional Republican-leaning special election. Nothing in the public turnout data indicates that early voters are breaking for Behn by double digits.

The poll’s methodology helps explain the gap. Emerson contacted voters through MMS text links and an online panel, modes that tend to over-represent younger and more internet-engaged respondents. Those groups are among Behn’s strongest demographics, while older rural voters — who typically dominate December specials — are less likely to respond to text-based surveys.

Likely-voter screens add another layer of uncertainty. In low-turnout elections, respondents frequently overestimate their likelihood of voting, and that bias tends to favor younger and more politically engaged individuals. A screen built on those self-assessments can produce a “likely voter” universe that does not match actual turnout.

None of this means the Emerson poll is misleading. It reflects one possible turnout scenario, and both campaigns are likely to promote its findings to motivate their supporters in the final days of voting. But it should be viewed in context. A poll that reconstructs 2024 as a Trump +13 district instead of a Trump +22 district — and that shows early voters breaking heavily for Behn despite long-standing county-level voting patterns and the turnout mix seen to this point — presents a narrower race than the fundamentals currently indicate.

TNPOLITICO’s tracking of daily early-vote reports, historical comparisons, and county-level benchmarks continues to show a district with a clear Republican lean. The Emerson poll describes an electorate that would yield a closer contest, but that electorate has not yet materialized in either turnout composition or voting patterns.

As national attention intensifies, these surveys should be taken for what they are: snapshots built on assumptions that may or may not match who actually votes on Dec. 2.