Fact Check Report

Summary

The draft script is built on solid source material and the core thesis is well-supported. However, there is one significant factual error -- the conflation of two different voter metrics to calculate the "25-point swing" -- that must be corrected before recording. Several other claims need tightening or additional context. The grocery bill figure is unsourced and potentially misleading.

Red flags: 1
Yellow flags: 5
Blue flags: 2

Findings

Red Flags

"From plus-20 on election day to minus-13 disapproval in January. That is... a 25-point swing"

Location in script: Middle of the thesis section (paragraph beginning "Trump didn't build a political movement") and repeated in the walkthrough section.
Issue: The script conflates two different metrics and two different voter groups to calculate the "25-point swing." The 20-point figure (60% to 40%) comes from a pre-election poll measuring voters who consumed no news at all. The minus-13 disapproval figure comes from Morris's January 2026 Verasight poll measuring low-knowledge voters (those who cannot identify which party controls Congress). These are overlapping but distinct populations measured by different instruments. Morris himself calculates the 25-point swing from a different baseline: low-knowledge voters backed Trump by a net margin of 11 points in 2024, and now disapprove by 13 points -- a 24-point swing that Morris rounds to 25. The script's framing of "plus-20 to minus-13" is arithmetically a 33-point swing, not 25, and it splices together incompatible data points.
Evidence: Morris's own article states: "Low-knowledge voters backed Trump by a net margin of 11 points in 2024. Now, the same low-knowledge voters say they disapprove of the president by 13 points -- a 25-point shift." The 60-to-40 / plus-20 figure is from a separate pre-election poll about news consumption, not the same Verasight knowledge-based metric. Web search confirms this distinction across multiple summaries of the article.
Recommended fix: Replace "From plus-20 on election day to minus-13 disapproval in January" with language that accurately reflects the Morris data: low-knowledge voters backed Trump by roughly 11 points in 2024 and now disapprove by 13 -- a 25-point shift. The 60-to-40 figure can still be used earlier in the script as vivid context for how disengaged voters broke in 2024 (as it currently is in the opening), but it must not be presented as the starting point of the 25-point swing calculation. The "25-point swing" number is real -- it just comes from 11-point margin to minus-13, not from 20-point margin to minus-13.

Yellow Flags

"the erosion is coming entirely from within the Republican base itself. Not from Democrats who already opposed him. Not from independents who were already skeptical. From his own people."

Location in script: Pew Research discussion, middle section.
Issue: The script applies "entirely from the Republican base" to the overall approval decline (37% down from 40%). Pew's actual finding is more specific: the decline in policy support (27% down from 35%) came entirely from Republicans. The topline approval drop from 40% to 37% is a separate metric. Pew also notes that the decline in confidence in Trump's ethics came "nearly all" from Republicans. The script's sweeping framing -- "the erosion is coming entirely from within the Republican base itself" -- overstates the precision of the Pew finding by applying it to the broader approval number rather than the specific policy-support metric.
Context: Pew's actual language: "Only 27% of Americans say they support all or most of Trump's policies -- down since last year, with the change coming entirely among Republicans." The approval drop (40% to 37%) is presented separately and Pew does not attribute it exclusively to Republican movement.
Recommended fix: Either (a) specify that it is policy support that has eroded entirely among Republicans: "Only 27% of Americans now say they support all or most of his policies, down from 35% -- and that decline came entirely from within the Republican base," or (b) add a qualifier like "the erosion in policy support is coming entirely from within the Republican base" rather than applying it to the overall approval figure.

"Trump has spent his time focused on immigration, crime, and cultural issues -- priorities shared by just 21% of the public"

Location in script: Brookings section, middle of the script.
Issue: The script substitutes "cultural issues" for the actual Brookings list. Brookings says Trump focused on "immigration, crime, taxes and government spending, or foreign policy" -- with 21% selecting one of those. "Cultural issues" is an editorialized substitution that changes what the data actually measures. "Taxes and government spending" and "foreign policy" are substantively different from "cultural issues."
Context: The Brookings source (03-brookings-midterm-prospects.md) explicitly lists: "immigration, crime, and taxes -- issues only 21% regard as most important." The full Brookings article lists "immigration, crime, taxes and government spending, or foreign policy."
Recommended fix: Use the actual Brookings categories. Suggested: "Trump has spent his time focused on immigration, crime, and foreign policy -- priorities shared by just 21% of the public." Or simply: "the issues Trump has focused on are priorities for just 21% of the public" without specifying categories that don't match the source.

"grocery bill that's up 15%"

Location in script: Structural problem section, toward the end.
Issue: This specific figure ("grocery bill... up 15%") does not appear in any of the source materials and is not sourced in the script. Web search shows overall food inflation was approximately 3.1% in the 12 months ending December 2025 (BLS data). Specific items like ground beef are up roughly 15%, and cumulative food inflation since the pandemic is around 25-30%, but a blanket "grocery bill up 15%" is not supported by current data as a year-over-year figure. If it refers to cumulative increases over a longer period, that needs to be specified.
Context: BLS data shows food-at-home prices up approximately 1.9-3.1% year-over-year as of late 2025. Individual items vary wildly (eggs down 30% from peak, ground beef up 15.5%, coffee up 20%). The 15% figure is plausible for specific items but misleading as an overall grocery figure.
Recommended fix: Either source the 15% figure to a specific item or timeframe ("a grocery bill where ground beef alone is up 15%"), use a more defensible aggregate figure, or remove the specific percentage and say "a grocery bill that keeps going up" to keep the rhetorical force without the factual exposure.

"These are... people holding down two jobs, raising kids, just surviving -- who don't have the time or the bandwidth to track which party controls which chamber"

Location in script: Early context section defining low-knowledge voters.
Issue: This characterization, while sympathetic and well-intentioned, is an editorial gloss that goes beyond what the data supports. The Morris source and CNN/Catalist data describe low-knowledge voters as younger, lower-income, less educated, less white, and more urban. The "two jobs, raising kids" framing implies a specific demographic profile (working parents stretched thin) that is not established by the source data. Research shows political disengagement correlates with youth, lower income, and lower education -- but also with general apathy, distrust, and lifestyle factors unrelated to economic hardship.
Context: This is partly editorial/rhetorical, but the script presents it as a factual characterization of who these voters are. The demographic data from Catalist describes them as "younger, less White, more urban, less likely to have college degrees" -- none of which specifically maps to "holding down two jobs."
Recommended fix: This is a judgment call. The sympathetic framing serves the show's argument about respecting these voters. But consider softening from "These are working people. People holding down two jobs, raising kids" to something like "Many of these are working people -- lower income, younger, dealing with the daily grind" which stays sympathetic without overclaiming the demographic profile.

"Trump himself lost 40 House seats in 2018, then came back and won the presidency in 2020"

Location in script: Counterargument section.
Issue: Minor precision issue. Democrats gained a net of 40 seats on election night in November 2018, but the total net swing including a prior special election was 41 seats. Most sources cite 40 or 41 interchangeably. The claim is substantively correct but the precise number is debatable. More importantly, Trump did not personally "lose" 40 House seats -- Republicans did, during his presidency. The framing is understandable shorthand but worth noting.
Context: Wikipedia and Ballotpedia confirm Democrats gained 40 seats on election night (41 including Conor Lamb's special election earlier in 2018). Multiple major outlets use both 40 and 41.
Recommended fix: No change strictly necessary. The 40-seat figure is the more commonly cited number for the election night result and is defensible. If you want maximum precision, "Republicans lost 40 House seats in 2018" is more accurate than "Trump himself lost."

Verification Needed

"That invisible army -- over 40 million people who'd skipped the 2022 midterms entirely -- showed up in 2024"

Location in script: Opening paragraph.
Note: The CNN/Catalist source (04-cnn-less-engaged-voters.md) confirms "More than 40 million people who didn't vote in the 2022 midterms voted in 2024." This checks out against the source material. However, the Catalist data more precisely discusses the general midterm-to-presidential dropoff pattern (roughly 40% of presidential-year voters skip the next midterm), and the "40 million" figure refers to people who voted in 2024 but did not vote in 2022 -- not necessarily people who "skipped the midterms entirely" in a broader sense (some may have voted in 2018 or 2020 but not 2022). The script's characterization is consistent with the CNN source's language but the host should confirm they are comfortable with the "skipped the 2022 midterms entirely" framing, since these voters may have voted in other prior elections.

"polls have underestimated Trump in every single election he's run in"

Location in script: Counterargument section.
Note: This is confirmed by multiple sources. Polls underestimated Trump's support in 2016, 2020, and 2024. The claim is accurate. However, the degree of underestimation varied (3.2 points in 2016, 4.1 points in 2020, smaller in 2024), and the 2024 underestimation was the smallest of the three cycles. Some pollsters argue the gap is narrowing. The host should be comfortable with the unqualified "every single election" framing, which is factually true but could imply the problem is getting worse when it may be getting better.

Sources Consulted

G. Elliott Morris, "The less voters knew, the more they liked Trump in 2024. Not Anymore," Strength In Numbers/Substack, February 12, 2026 -- https://www.gelliottmorris.com/p/trump-lost-low-info-voters
Pew Research Center, "Confidence in Trump Dips in 2026, and Fewer Now Say They Support His Policies and Plans," January 29, 2026 -- https://www.pewresearch.org/politics/2026/01/29/confidence-in-trump-dips-and-fewer-now-say-they-support-his-policies-and-plans/
Brookings Institution, "As President Trump loses support, Republican prospects in the 2026 midterms grow darker," December 4, 2025 -- https://www.brookings.edu/articles/as-president-trump-loses-support-republican-prospects-in-the-2026-midterms-grow-darker/
CNN/Catalist, "Less engaged voters were key to Trump's 2024 victory, new analysis finds," May 20, 2025 -- https://www.cnn.com/2025/05/20/politics/2024-election-voters-analysis
KFF, "ACA Marketplace Premium Payments Would More than Double on Average Next Year if Enhanced Premium Tax Credits Expire" -- https://www.kff.org/affordable-care-act/aca-marketplace-premium-payments-would-more-than-double-on-average-next-year-if-enhanced-premium-tax-credits-expire/
ABC News, "1.4 million fewer people enrolled in ACA plans as premiums spike" -- https://abcnews.go.com/Health/14-million-fewer-people-enrolled-aca-plans-premiums/story?id=129221228
Center for American Progress, "A Year in Review: How the Trump Administration's Economic Policies Made Life Less Affordable" -- https://www.americanprogress.org/article/a-year-in-review-how-the-trump-administrations-economic-policies-made-life-less-affordable-for-americans/
Bureau of Labor Statistics, food inflation data via U.S. Inflation Calculator -- https://www.usinflationcalculator.com/inflation/food-inflation-in-the-united-states/
Wikipedia, "2018 United States House of Representatives elections" -- https://en.wikipedia.org/wiki/2018_United_States_House_of_Representatives_elections
ABC News/538, "2024 polls were accurate but still underestimated Trump" -- https://abcnews.go.com/amp/538/2024-polls-accurate-underestimated-trump/story?id=115652118
Brookings, "The polls underestimated Trump's support -- again" -- https://www.brookings.edu/articles/the-polls-underestimated-trumps-support-again/
NPR, "After Texas ruling, Trump and Republicans head to 2026 with a redistricting edge" -- https://www.npr.org/2025/12/08/nx-s1-5634585/redistricting-2026midterm-election-trump-congress
Morning Consult, "2026 Midterm Elections Generic Ballot Tracker" -- https://pro.morningconsult.com/trackers/2026-midterm-election-generic-ballot-polls
Washington Times, "Trump approval rating falls to 37%, Pew poll finds" -- https://www.washingtontimes.com/news/2026/jan/30/trump-approval-rating-falls-37-pew-poll-finds/
WSBA Insurance Marketplace, "2026 ACA Open Enrollment Guide" (confirming $638-to-$2,179 premium scenario) -- https://wsba.memberbenefits.com/2025/10/20/2026-aca-open-enrollment-guide/

Clean Claims

The following major factual claims in the script checked out and can be treated as solid ground:

Trump won low-knowledge/disengaged voters by large margins in 2024. Confirmed across Morris, CNN/Catalist, and NBC News sources. The 60-to-40 figure for voters consuming no news is accurate as a pre-election poll finding.
Over 40 million people who skipped the 2022 midterms voted in 2024. Confirmed by CNN/Catalist analysis.
Morris's data shows low-knowledge voters now disapprove of Trump by 13 points. Confirmed by the Verasight January 2026 poll.
The shift among low-knowledge voters is roughly double that among high-knowledge voters. Confirmed by Morris's analysis.
Pew has Trump's overall approval at 37%, down from 40%. Confirmed by Pew Research Center, January 29, 2026.
Only 27% of Americans support all or most of Trump's policies, down from 35%. Confirmed by Pew.
Republicans won the House popular vote by 2.6 points in 2024. Confirmed by Brookings and election data (49.8% to 47.2%).
Democrats lead generic ballot polling by 5.3 points. Confirmed by Brookings (December 2025 data). Note: more recent polling from Morning Consult (February 2026) shows the Democratic lead at 3 points, so this figure may be slightly dated depending on which average is cited.
21 Republican-held seats won by margins under 8 points. Confirmed by Brookings.
Hispanic support for Republicans at 29%, independents at 15%, young adults at 19%. Confirmed by Brookings.
Low-knowledge voters disapprove of Trump's handling of prices by 40 points; high-knowledge by 30 points. Confirmed by Morris.
74% of high-knowledge voters hold strong convictions; 58% of low-knowledge voters. Confirmed by Morris.
ACA enhanced premium subsidies expired end of 2025. Confirmed by multiple sources.
Net premium payments up 114%. Confirmed by KFF analysis as an average for subsidized enrollees staying in the same plan.
Premiums projected to cost 4x more for those under 250% FPL. Confirmed by economic impact source material.
$638/month to $2,179/month premium example for 55-year-old couple earning $90K. Confirmed by WSBA Insurance Marketplace guide and consistent with KFF calculator scenarios.
1.4 million fewer Americans selected marketplace plans in 2026. Confirmed by CMS data reported by ABC News and KFF.
19% of low-knowledge respondents report coverage loss vs. 11% of high-knowledge. Confirmed by Morris.
Only 22% of Trump's own 2024 voters say tariffs are helping. Confirmed by Brookings and economic impact source.
Manufacturing lost 72,000 jobs April-December 2025. Confirmed by BLS data, reported by multiple outlets including CBS News, Center for American Progress, and Media Matters.
Two-thirds of Americans concerned about tariff impact on finances. Confirmed by Brookings and economic impact source.
50% prioritize inflation, jobs, or healthcare. Confirmed by Brookings.
Republicans attempting mid-decade redistricting. Confirmed. Major efforts underway in Texas, Ohio, Missouri, North Carolina, and Florida, with Democratic counter-redistricting in California, Virginia, and Maryland.
Democrats got destroyed in 2010 midterms; Obama won reelection in 2012. Confirmed. Democrats lost 63 House seats in 2010 (larger than stated 2018 comparison). Obama won 332-206 in 2012.
Polls have underestimated Trump in every election he's run in (2016, 2020, 2024). Confirmed by AAPOR reviews, 538 analysis, and Brookings.
Morris methodology: hybrid probability and non-probability sampling, Census-weighted. Confirmed by Verasight methodology description.

The Invisible Army Goes AWOL

Fact Check