That is the finding. It is directional, it is attributed, and it is worth taking seriously on its own terms without inflating it into something larger.
What “likely to support counter tariffs” actually signals
There is a meaningful difference between a survey saying Canadians are likely to support counter tariffs and a survey saying Canadians have concluded that counter tariffs will work. The first is a statement about posture. The second would be a statement about expected outcomes, and it is not what the CTV-reported framing claims.
Posture matters. When a government weighs a retaliatory tariff package, the political question is not only whether the measure is economically optimal but whether the public will tolerate the friction it creates. A survey showing widespread support for responding in kind tells a government that the tolerance exists, at least in the moment the question was asked. That is genuinely useful information, and it is the kind of information trade policy has historically lacked in real time.
What it does not tell you is durability. Support for a response measured while the measure is still abstract is a different quantity than support measured six months into visible price effects on specific goods. Polls capture the former far more easily than the latter, because the former is a values question and the latter is a lived one. Nothing in the reported findings claims to have measured the second thing.
It also does not tell you intensity. “Likely to support” spans everything from a firm conviction that Canada must answer to a mild preference expressed to a pollster in under a minute. Directional findings compress that range. Reading a strong mandate into a directional result is the most common way survey coverage overreaches, and it is the read worth resisting here.
Here is the CTV News package on the survey, with the Nanos framing as reported:
Retaliation versus absorption is the frame the survey sits inside
The survey did not create the retaliation-versus-absorption binary. It landed inside it.
Through late August and into early September 2026, the Canadian policy conversation reported in national outlets has centered on how Prime Minister Mark Carney’s government responds to American trade measures. The available options in public discussion have been broadly framed as answering with counter tariffs or absorbing the American action while pursuing other channels. That framing is not neutral. It positions absorption as passivity and retaliation as agency, and survey questions written inside that atmosphere will tend to inherit the same architecture.
This is not a criticism of the polling. It is a description of how public-opinion research works. Respondents answer the question in front of them, and the question in front of them reflects the debate as it is currently constructed. When the national conversation offers two doors, a survey measures which door people prefer. It cannot measure preferences for doors nobody has described.
The adjacent finding on turning down a U.S. trade deal fits the same logic. A public that says it would rather walk than sign a bad agreement is expressing something about national self-respect as much as something about trade economics. Both readings are legitimate. Both are worth separating in analysis, because they lead to different policy conclusions. A government reading the result as economic endorsement will behave differently than one reading it as a mandate to protect Canadian standing in the negotiation.
The honest characterization is that Canadians, in this survey window, prefer response to acceptance. That preference is real. Its economic content is unspecified.
What a directional poll cannot tell you about the checkout
Counter tariffs are taxes on imports. The mechanical effect, wherever they are applied, is to raise the landed cost of the goods they touch. Some of that cost is absorbed by importers, some by retailers, and some appears on shelves. The distribution varies by category, by supply chain, and by how substitutable the affected goods are.
A survey measuring support for counter tariffs measures none of that. It cannot, and it does not claim to.
This is the discipline point. The reported findings are directional. No sample size, margin of error, or fieldwork detail is being asserted here beyond what CTV reported, and no provincial or age breakdowns should be inferred. Treating a directional national finding as though it resolves questions about consumer prices, sectoral exposure, or business investment is a category error, and it is the error most likely to be made in the days after a poll like this circulates.
There is also the timing problem. Public opinion during a trade dispute is not static. It responds to visible outcomes, to whether the other side blinks, and to whether the costs of the response become legible in ordinary purchases. A survey conducted in early September 2026 describes early September 2026. It is not a forecast of where opinion sits after the measures have run their course, and it is not evidence that the measures will produce the outcome respondents want.
The most defensible use of this finding is narrow and still useful: it tells a government that the political cost of responding is currently lower than the political cost of not responding. That is a live input into a decision. It is not a substitute for the decision.
Canadian opinion, in this survey window and as reported by CTV News, sits on the side of retaliation. Whether retaliation delivers what Canadians want from it is a separate ledger, kept in different units, and settled over a longer horizon than any poll can reach.