S.M.M.

Stop Manipulating Me · A Field Guide to Psychological Influence

ENTRY No. T13.1
CATEGORY Information Control
CLEARANCE Public / Essential
EDITION 01
Dossier · Manipulation Tactic

Cherry-Picking

Selection · Curated Evidence Base · Pervasive
Red Flag
How It WorksSEC 01

Cherry-picking builds a false impression out of entirely true parts. Rather than fabricating anything, the manipulator selects only the facts, studies, quotes, or figures that support a predetermined conclusion and omits everything that would complicate or reverse it. Because each individual item survives fact-checking, the technique is far more durable than an outright lie — a skeptic who verifies the pieces can still be led to the wrong whole.

The engine is confirmation bias (T8.1) and the availability heuristic: an audience judges by the vivid evidence in front of them and rarely asks for the evidence that was left out. A supplement is sold on the one small trial that showed an effect while a dozen null results go unnamed; a policy record is defended with the good quarter and the bad decade is skipped; a chart begins in the single year that makes the line rise. This is the raw material of spin, pseudoscience, and disinformation, and it scales up into one-sided stories (T13.7) and, when contrary scraps are selected instead, into manufactured doubt (T13.18).

The defense is structural, not fact-by-fact. Because the deception lives in what is absent, verifying the present items cannot catch it. You have to reach for the full evidence base — base rates, systematic reviews, the widened timeframe, the strongest counter-case — and treat any presentation where all the evidence conveniently agrees as a signal to go looking for what was left on the cutting-room floor.

Warning SignsSEC 02
  • Only supportive data appears. Every number, study, or quote points the same direction — real evidence rarely lines up that cleanly.
  • No denominator. You are shown the wins but never the total attempts, the base rate, or the sample size.
  • Contrary findings go unmentioned. Well-known counter-evidence in the field is simply absent, not addressed and rebutted.
  • Outlier as headline. A single dramatic case or a single favorable year stands in for the general trend.
  • Quotes stripped of context. A source's words are excerpted so a qualification or reversal right beside them is cut away.
  • Cherry-picked timeframe. The start and end dates of a chart are chosen precisely because they flatter the claim.
Frequently Paired WithSEC 03
  • One-Sided Stories · T13.7
    Scales selection from data points to a whole narrative
  • Uncertainty Flooding · T13.18
    Selects contrary scraps to manufacture false doubt
  • Confirmation Bias · T8.1
    The cognitive vulnerability it exploits in the audience
How the Selection WorksSEC 04
  • Stage 01 · Fix the Verdict
    The desired conclusion comes first; the search for evidence then runs backward from it rather than toward it.
  • Stage 02 · Harvest the Supportive
    Only data points that fit are gathered — genuine studies, real quotes, accurate numbers — while contrary findings are quietly dropped from the pile.
  • Stage 03 · Present as Complete
    The curated subset is displayed as if it were the entire record, so a true-sounding sample is mistaken for a true whole.
Counter-ProtocolSEC 05
Defense: Ask what is missing — judge the full evidence base, not the curated sample.
  • Demand the denominator. For any success or correlation, ask for the total attempts, the base rate, and the sample size. Wins are meaningless without the count of tries.
  • Seek disconfirming evidence. Deliberately search for the strongest case against the claim. If none is offered, that absence is itself the tell.
  • Weigh the whole literature. For scientific or statistical claims, look to systematic reviews and meta-analyses that aggregate the full body of work, not a hand-picked study.
  • Check the timeframe and framing. Extend the chart's start and end dates; a trend that reverses when you widen the window was an artifact of the cut.
  • Follow the incentive. Ask who benefits from this exact conclusion. A curated case built by an interested party warrants an independent second source.