persuasion
Framing effect
What is the framing effect?
The framing effect is how the same fact, described positively or negatively, lands differently - "90% fat-free" feels healthier than "10% fat" though they are the identical tub. We react to the spin as much as to the substance.
Also known as: framing, framing bias
The demo
The same tub of yogurt, the same 90/10 split - just described two ways. Flip the framing and watch how healthy it suddenly feels, or doesn't, while the bar below holds perfectly still.
Greek-style yogurt
90% fat-free
The exact same tub either way: 90% fat-free, which is to say 10% fat.
Flip the framing and notice the same figure change character.
What this demo shows (text version)
A tub of yogurt with a headline claim and a bar showing its 90% fat-free, 10% fat split. In positive framing the claim reads "90% fat-free", which sounds healthy; in negative framing it reads "Contains 10% fat", which sounds indulgent.
The split bar underneath is identical in both modes, because the product hasn't changed - only the half of the fact chosen for the headline. That is the framing effect: the same information, framed positively or negatively, produces a different impression and can steer a decision before any real thinking happens.
"90% fat-free" sounded like a health food; "contains 10% fat" sounded like a guilty pleasure - and the split bar underneath never moved. Same number, opposite feeling. Whoever chooses which half of the fact to lead with has already nudged your judgement before you've done any thinking.
Framing is honest when both versions are true and you pick the relevant one, manipulative when you pick the misleading one. "95% of users had no issues" and "1 in 20 hit a bug" are both real; which you lead with should depend on what the reader needs to know, not what flatters you. The classic study: patients choose surgery far more often when told "90% survive" than "10% die".
It pairs with its cousins. A gain-or-loss frame leans on [loss aversion](/entries/loss-aversion/); a first number that colours the rest is [anchoring](/entries/anchoring/). My habit when I catch a persuasive stat is to flip it - state the other half out loud - and see whether the decision still feels the same. Often it doesn't, and that tells you the frame was doing the work.