Science1 distinct publisher3 min readUpdated
Annenberg finds 70% of Republicans and 69% of Democrats favor banning the herbicide, while most respondents say they have heard little about it. A one-point party gap is not a gap at all.
The Scientist · Science desk

Compiled by The ScientistSomething wrong?How this is made
Treat the headline number as 66 to 67 percent and run the weighted average backwards. MAHA supporters, 26 percent of the sample, favor a ban at 72 percent; the 40 percent who say they are not MAHA supporters favor it at 70 percent [3][6]. That leaves the third of respondents who were not sure whether they support MAHA at all, and to make the total land where it did they have to be sitting in the mid-to-high 50s [3]. The party question has the same shape: 70 percent of Republicans, 69 percent of Democrats, and fewer independents than either [4].
Support for banning a farm input that has been in American fields for roughly fifty years [8][5] is therefore highest among people who have already accepted a label and lowest among people who have not. It does not track familiarity with the chemical, because there is not much familiarity to track [2]. Ken Winneg, Annenberg's managing director of survey research, put the same observation in the release [7].
The one-point distance between the parties is not a partisan finding [1]. The two-point distance between MAHA supporters and MAHA decliners is not a MAHA finding [2]. Both are inside the rounding of Annenberg's own subgroup shares, which sum to 99 percent [6]. What the poll is measuring is the price of holding the opinion, and the price is zero: nothing in the answer requires the respondent to have looked at a literature that an overview by FactCheck, cited in the same release, calls inconsistent [14].
Cheap opinions still move policy when the route is cheap, and the routes just changed. The June ruling took the jury out of the transmission line between public suspicion of Roundup and money changing hands [13]. What remains is administrative and legislative, and that is where a number like this one is useful. Kelly Ryerson, the MAHA activist who posts as Glyphosate Girl, has already asked the EPA in writing for a permanent phase-out of glyphosate and glyphosate-based herbicides, including on agricultural land [16]. Zen Honeycutt of Moms Across America told an April rally she had spent fourteen years pursuing a ban [15].
The administration is arguing with itself in the meantime. The February order protecting glyphosate production drew an outcry from the MAHA activists whose support helped reelect Trump, which Politico summarized as MAHA unleashing on the White House [10]. Robert F. Kennedy Jr., who has said for years that glyphosate causes cancer without calling for it to be banned [11], defended that order on the Joe Rogan Experience while saying it was not something he was happy with, and that banning glyphosate overnight would destroy the American food system [12].
What is durable in the Annenberg data is the alignment, not the intensity. A sponsor who wants a state glyphosate bill does not have to trade anything to get votes from both sides, since there is no partisan seam to work. And a public that reports hearing little about the subject is a public unlikely to cross-examine anyone on the toxicology.
Follow any of these and your For You feed starts watching them — no settings page required.
Ranked by verification strength, evidence, and original report placement.
Health and Human Services Secretary Robert F. Kennedy Jr. has long said that glyphosate causes cancer but has not advocated for an outright ban.
On the Joe Rogan Experience in February 2026, Kennedy defended Trump's executive order, said of glyphosate "It's not a good thing to have in your food ... it's not something that I was particularly happy with," and added "If you banned glyphosate overnight ... it would destroy the American food system."
Two-thirds of Americans favor banning the use of the herbicide glyphosate, the active ingredient in Roundup, on crops grown in the United States, according to a nationally representative survey by the Annenberg Public Policy Center of the University of Pennsylvania.
The spring survey also finds that most Americans have read, seen or heard little about glyphosate in the few months preceding the survey, and many are uncertain about its health effects.
Seventy-two percent of people who identify as supporters of the Make America Healthy Again movement favor banning glyphosate, as do 70% of those who do not support MAHA.
Seventy percent of self-identified Republicans and 69% of self-identified Democrats favor a ban; fewer self-identified independents favor a ban.
Evidence-backed comparisons of source perspectives and observed adoption signals. Read the methodology
Which Builder, Operator, and Investor concerns the observed source mix emphasized—not a truth score.
Evidence, demonstrated adoption, hype gap, incentives, and confidence are assessed independently, each on its own current evidence. How these are measured.
One well-documented survey, no independent corroboration
The core numbers come with real methodological grounding - named wave, field window of April 14-28, 2026, a stated 1,639-respondent empaneled sample, published subgroup shares and a named research director on the record - and the release itself supplies the deflating counter-figures (83% low awareness, 46% unsure on cancer) rather than hiding them. Against that: the cluster contains a single item from a single publisher reproducing the sponsoring institution's own release, with no margin of error, no weighting detail, no question wording, and no reported ban support for the 33% unsure about MAHA, which makes the topline impossible to reconstruct from the published cross-tabs. The scientific backdrop is genuinely contested (IARC 2015 versus EPA 2017) and is characterized by an outlet owned by the same policy center.
Majority opinion, no policy uptake
Adoption here means whether the polled preference is translating into action, and the documented record points the other way. In the same year as the survey, federal policy moved to protect glyphosate-based herbicide production as a strategic input, the Supreme Court removed a labeling liability verdict 7-2, and the cabinet secretary most associated with glyphosate criticism explicitly declined to back a ban. The only uptake visible is advocacy-channel activity - EPA comments and a rally - which is input to a process rather than an outcome. The survey release itself is the fourth observation and is fresh, but publication is not adoption.
Topline outruns the depth of the opinion behind it
The 'two-thirds want a ban' framing is accurate as stated but describes an attitude held by a public that overwhelmingly reports no recent exposure to the subject (54% nothing at all, 29% a little) and is nearly evenly split between believing glyphosate causes cancer and not knowing. A near-zero party gap and a near-zero MAHA gap are more consistent with a general aversion to a chemical-sounding name than with a mobilized policy demand, and every institutional action in 2026 went the opposite direction. The gap is modest rather than large precisely because the release and the cluster headline both foreground the low-awareness caveat instead of burying it.
Institutional self-promotion, disclosed
The item is a research-center release distributed through an aggregating science outlet: the Annenberg Public Policy Center reports its own survey, quotes its own survey director, and sources the scientific framing to FactCheck.org, which the text discloses is an APPC project. That is a real self-reinforcing loop with an interest in the survey's newsworthiness, though the disclosure is explicit and the release publishes findings that undercut its own headline. No commercial, vendor or funding interest in glyphosate itself is evident in the supplied material, and the activists and officials quoted have openly stated positions.
Reported facts solid, interpretation single-sourced
Confidence is moderate: the discrete facts - poll numbers, field window, executive order, Supreme Court margin, quotations - are stated precisely and consistently within a single authoritative release, so misreporting risk on the numbers is low. What limits confidence is structural: one publisher, one institutional source, no margin of error or question wording, a missing subgroup figure that blocks internal validation, and month-only dating on two of the policy events. Adoption and hype-gap readings rest on documented events and are reasonably firm; any inference about the durability of the two-thirds majority is not.
invest
Your Landed Cost Is Being Litigated By Companies With $306,000 Problems1 distinct publisher
invest
Importers Get No Vote on a Midnight Tariff: $20bn of Canadian Goods, 50%, Decided Overnight1 distinct publisher
security
USPS finalizes federal mail-ballot eligibility rules while two injunctions stand1 distinct publisher
leadership
Trump's ballroom case is a standing case, and permitting-exposed firms should read it that way1 distinct publisher
Distinct publishers with included, body-backed reporting in this cluster.
phys.org
1 article · August 21, 2026