Percent Change Illusion

A percent change can feel much bigger or smaller than it actually is, depending entirely on a starting number that often goes unnoticed. The percentage sticks; the baseline it was computed from doesn't.

This illusion names the systematic misestimation that occurs when evaluating proportional differences between quantities. Perceptual and cognitive shortcuts distort the baseline comparison, so the relative change gets misread.

"Crime jumped 100%" sounds alarming in a headline — until you learn there were 2 crimes last year and 4 this year. The phrase "100% increase" conjures something far more dramatic than a difference of two incidents, because the tiny starting number never made it into the sentence.

A pharmaceutical company reports that a new drug reduces adverse events from 0.02% to 0.01%, marketed as a "50% reduction in risk." Clinicians reading the relative figure without renormalizing to the denominator can overestimate the therapeutic benefit substantially — the absolute risk reduction is just 0.01 percentage points, for a number needed to treat of 10,000. The tiny baseline rate simply fades from memory while the 50% figure stays vivid. Presenting the absolute risk alongside the relative one, and requiring pre-registered outcome framing, meaningfully corrects this in clinical decision-making.

One number anchors the judgment, and people adjust away from it only a little. That small adjustment is why percent differences end up looking larger or smaller than reality.

A comparator-dominant encoding, paired with an underweighted memory of the baseline, biases the system toward overrepresenting the endpoint. Working memory's limited capacity for precise ratio computation is what produces the skewed magnitude estimate.

Putting both numbers side by side, with the baseline clearly labeled, closes most of the gap. A quick percent calculator is enough to check a gut reaction.

Normalized change metrics presented alongside explicit baseline context rebalance the comparator weighting directly. An automated ratio computation offsets both the memory limitation and the salience bias.

Anchoring to endpoint; Baseline forgetting; Salience-driven exaggeration

An adversarial communicator can selectively choose an extreme baseline value to make a small absolute change appear as a dramatic percent increase or decrease, manufacturing emotional impact without substantive movement. This technique is routinely weaponized in financial journalism, political messaging, and advertising by suppressing baseline context and foregrounding the comparator endpoint — exploiting comparator-dominant encoding to exaggerate or minimize outcomes as needed. By framing the same underlying data with different reference periods or denominator choices, an actor can fabricate opposing narratives from a single dataset.

Analysts and communicators should mandate dual presentation of absolute and relative changes together, with the baseline value clearly labeled and anchored visually, to counteract comparator salience bias. Structured templates for quantitative reporting that require explicit baseline disclosure, along with automated ratio-checking tools, reduce reliance on working memory and suppress endpoint anchoring. Training in denominator awareness — specifically rehearsing the mental step of re-encoding the base rate before evaluating a percent figure — builds durable resistance at the individual level.