The Tavern Take: Week of March 16, 2026
What you need to know this week
Voters will scroll past your typical spots. They'll watch ours. Matt published a piece in The Lab this week on what 150 randomized trials taught us about attention and persuasion. They're unrelated. Persuasion lives in audio. And mesmerizing content — hypnotic visuals matched to voter interests — produced a +4% treatment effect on completion and a −4.3% shift in Trump approval among completers. → Read more
Nine days into the Iran war, public support is a coin flip. 43% support the operation, 42% oppose — with no rally-around-the-flag effect in sight. The rationale is underwater by 17 points. Trump's tone at Dover drew 50% disapproval. 56% call it a major factor in his job performance. The public entered this war skeptical and nothing has changed their mind. → Read more
55% want congressional authorization. Congress said no. The public noticed. The war powers resolutions failed 47–53 in the Senate and 198–237 in the House. Voters supported authorization by a 28-point margin. 57% say they'd trust a lawmaker more for crossing party lines on this vote — net +41. Truth Social war comms poll at −29. The public has a coherent theory of wartime governance, and the administration is failing every test of it. → Read more
The Noem firing didn't close the loop. Voters are running three accountability ledgers at once. The $220M DHS ad campaign polls at net −65 — the most lopsided spending result of the week. The firing polls at +47. And 73% support independent investigations into citizen deaths involving federal agents — a 63-point margin, the widest consensus finding in the entire survey. The administration addressed one ledger. Two remain open. → Read more
The economic frame keeps winning on Iran — and exit strategy is now a weapon. Pocketbook appeals had the single largest argument-type effect measured in this series (+9.4 points). The top message hit 70.9%. "Trump started this war without asking Congress and without any plan for how it ends" hit 66.1%. Four confirmed traps: Iraq analogies (−5.4 pts), geopolitical framing (−5.5 pts), aggressive personal tone (−4.1 pts), bipartisan measured tone (−3.4 pts). Don't use any of them. → Read more
The Epstein finding isn't 53-20. It's the broken promise at −25. Trump's retreat from his 2024 pledge — telling AG Bondi to release "whatever she thinks is credible" — polls 15 points worse than his overall job approval. 54% support congressional investigation into whether the DOJ improperly withheld records. The DOJ's "coding error" explanation isn't landing. The mechanism is broken-promise politics, not partisan opposition. → Read more
THE PATTERN THIS WEEK: Every story tested last week — Iran, DHS, Epstein, message testing — returned the same result from a different angle: voters aren't uniformly anti-policy, but they are consistently and overwhelmingly pro-accountability, pro-clarity, and pro-institution. The administration keeps giving them reasons to activate that instinct.
What This Means in 90 Seconds
For campaigns: The case against this administration isn't built on any single unpopular policy — it's built on a consistent pattern of implementation that bypasses institutions and treats wartime like a PR exercise. On Iran specifically: lead with exit strategy and gas prices, not the Constitution. "Trump started this war with no plan for how it ends" is connecting economic costs, open-ended commitment, and broken promises in a single frame that tests across the political spectrum.
For lawmakers: Three oversight lanes are open with public backing right now — war powers authorization (55-27), Epstein records integrity (54-22 on the congressional investigation question), and federal enforcement accountability (73% on independent investigations). The public is explicitly rewarding members who walk through them. 57% would trust a lawmaker more for crossing party lines on war powers. The data says it's electorally safe.
For advocates: The broken-promise frame is the through-line this week. It's not just Epstein — MTG's argument that the Iran war "betrays campaign promises of no more foreign wars" resonates with 51% of voters. Voters aren't uniformly opposed to the administration's goals. They're tracking whether the president is doing what he said he would do. That's a frame that works across issues and across coalitions.
Methodology: All data from Tavern Research Public Pulse surveys, online panel, fielded March 2–10, 2026. Sample sizes range from n=335 to n=552 per survey. Margin of error ±3.5–8.0%. Message testing via MaxDiff (n=370–473).
Want the full memos behind these numbers? Email data@tavernresearch.com →