NewsWatch NewsWatch
← Real & factually accurate

Real & factually accurate

Confidence 60% — the evidence leans this way, but alternatives remain open

Beware of bogus polling

washingtonexaminer.com ·Washington Examiner ·Oct 22, 05:00 EDT · Read the original →

What it says

A Washington Examiner house editorial warning readers about fabricated polling, asserting that for the second time in two months a polling organization used by media outlets admitted to making up results. It cites an August case in which 'Median Strategies' allegedly admitted fabricating polls showing Wisconsin Democratic gubernatorial candidate Francesca Hong up more than 20 points and Los Angeles Mayor Karen Bass leading challenger Nithya Raman by nearly 12, noting Hong actually lost to David Crowley by 0.5 points. The visible text cuts off at a subscription wall before the main, more recent example is described.

Why this verdict

The piece reads as a standard unsigned editorial from a known partisan-but-reporting outlet, with a measured tone, no hype language, and specific, checkable particulars (named pollster, named candidates, numeric margins). The named figures are real and the political context is internally coherent for a 2026 cycle: Francesca Hong and David Crowley are Wisconsin Democrats plausibly contesting a 2026 gubernatorial primary, and Karen Bass and Nithya Raman are real Los Angeles officials. None of this contradicts anything I can establish, and the events postdate my knowledge, so unfamiliarity with 'Median Strategies' or with the specific primary outcomes is not evidence of fabrication. The automated AI-text score is very low and the writing shows normal human burstiness, so the generated-filler path is unlikely. The main limitations are that the body is truncated at a paywall before the central claim (the second, larger fabrication case) is identified, and the single corroborating domain is another conservative outlet rather than independent confirmation. I therefore land on 'real' with only moderate confidence rather than high.

No meaningful indicators of machine generation: low LLM-phrase hits, high burstiness and lexical diversity, idiomatic editorial voice, and concrete named entities and figures rather than template slots. Consistent with a human-written newspaper editorial.

Claims examined

  • unverifiable

    A polling organization embraced by media outlets has admitted to fabricating results, the second such case in two months

    Post-dates my knowledge; the second, more recent case is behind the paywall cutoff and cannot be examined.

  • unverifiable

    In August, Median Strategies admitted to making up polls

    I have no record of this firm, but that is expected for a 2026-cycle outfit; no attribution to a report or statement is given in the visible text.

  • unverifiable

    Fabricated poll showed Francesca Hong leading Wisconsin's Democratic gubernatorial primary by more than 20 points

    Hong is a real Wisconsin state representative; a 2026 gubernatorial run is plausible but beyond my verification.

  • unverifiable

    Hong lost the primary to David Crowley by 0.5 points

    Crowley is a real Milwaukee County Executive; the specific result is a 2026 outcome I cannot check.

  • unverifiable

    Fabricated poll showed Karen Bass leading Nithya Raman by nearly 12 points

    Both are real Los Angeles officials; a Raman mayoral challenge in 2026 is plausible but unconfirmed here.

Signals

SignalStrengthScore
Source trust
Reputation prior for the publishing domain. A starting point, not a verdict.
0.52
Corroboration
How many independent outlets carried a matching story, discounted when they are all low-tier.
0.40
Specificity
Density of named people, direct quotes, attributed statements, figures and dates.
0.36
AI text markers
Stylometric markers of generated filler. Higher means more machine-like. A weak prior, never proof.
0.03
Emotional loading
Clickbait, outrage and hedging language. Higher means more loaded.
0.00
Byline
Whether a named human author is credited.
1.00

Also carried by

1 other domains in the current window carried what appears to be the same story: washingtontimes.com.

Caveats

  • The article body is cut off by a subscription prompt, so the central, most recent fabrication claim referenced in the lede is not visible and could not be assessed.
  • Every substantive factual claim concerns 2026 events after my knowledge cutoff; the verdict rests largely on internal coherence, real named entities, absence of fabrication markers, and the outlet's record of original reporting.
  • The single corroborating domain is an ideologically similar outlet, which is weak independent confirmation.

Provenance

Assessed byclaude-opus-5
Text available Full article body
Found viaWashington Examiner (outlet feed)
First seenOct 22, 06:04 EDT
Composite signal prior0.56