The Palanor approval average summarises published national polls asking whether people approve of how the president is handling the job, the economy and inflation (or prices, or the cost of living). It describes what pollsters have released. It is not a forecast, an election model or a view on any official.
The signals
Nine daily series, named Presidential Approval … (Palanor Average): approve, disapprove and net (approve minus disapprove, in points) for each of overall, the economy and inflation. Their slugs start with `politics.approval.president.`. Every reading stores a band (`lo` and `hi`), the number of pollsters behind it and the newest poll's field end date.
How it is built
- Each pollster counts once, with its most recent poll, so a daily tracker weighs no more than a monthly poll.
- Recent polls count more (a half-life of 10 days overall, 21 days for the economy and inflation, which are asked less often), and larger samples count more.
- Registered- and likely-voter results are shifted to an all-adults basis by the gap seen in polls that report both.
- Each pollster's lean is estimated from its record against the other pollsters over the past year and taken out.
- A day has a reading only when at least three pollsters are in the window, and the average restarts at each inauguration.
The method and its test were written down and committed before any result was computed. The first test failed and nothing was published; a second registration changed the test only, and passed. Both results are on the methodology page.
Reading it
- Readings for the last few days can change as releases that ended earlier are added.
- The band is the uncertainty of the average from the polls in hand. In testing it was narrower than a true 95% range, so do not read it as a range for the next poll.
- Charts of these series carry no headline, and approve and disapprove are drawn in neutral colours.
- Palanor does not project these series (no Foresight). The methodology page shows Kalshi and Polymarket approval contracts beside the average, credited to each exchange and never blended in.
Sources
The FiveThirtyEight polls archive (ABC News, CC BY 4.0) for 2017 to early 2025, then each pollster's own public release: Quinnipiac and Emerson read automatically, others entered from the release with its link. Toplines only. Not used: Gallup, YouGov and Ipsos releases, any aggregator's numbers, and polls sponsored by campaigns, parties or advocacy groups.
Also published: the Cooperative Election Study's yearly question on whether the national economy got better or worse over the past year (public domain), as plain survey data.
Related: [Charts](/help/modules/charts)