Juxtaposition drops the evaluation load from good-or-bad to which-is-better

What it is

One poster sits alone. Asked “is it good,” people stall: the scale of good has no anchor. Two side by side, the question becomes “which is fitter as the hero,” and a finger can point in thirty seconds. Comparative evaluation via juxtaposition swaps absolute good/bad for a relative take. Load drops not because there are more options, but because contrast itself supplies a scale.

Generation turning from production into selection is a change of task type. This entry governs the evaluation mode: side-by-side makes “which is better” doable.

Why it happens

Absolute evaluation has to build a standard internally, holding quality dimensions and a pass line in working memory — costly and unstable. Relative evaluation borrows the other artefact as a temporary ruler: contrast, information density, strength of commitment show up on the spot. In psychophysics, a difference threshold is far lower than absolute identification; on-screen juxtaposition is that threshold put to work.

Side-by-side also forces the dimension from “overall good” into a pointable attribute. Saying “this title is larger” is already a checkable comparison, not an undecomposable score.

Studying it

The same pair, sequential isolated ratings versus simultaneous forced choice. Dependent variables: time to decide, test–retest stability, agreement with an expert ranking. Independent variables: juxtaposition or not, an intervening unrelated task (to break maintenance of an absolute standard). Code “cannot say why but can pick” separately — relative judgment can outrun verbalisable reasons.

The expert gold should be a relative ranking, not absolute scores. Noise in absolute scores is what this claim is avoiding.

Where it stops holding

With no shared task dimension (one is an icon, one is a poster), juxtaposition mints a false comparison. Two answers to a closed right/wrong item, side by side, let relative judgment override checking. When options cannot be seen at once (one-at-a-time on a narrow screen), relative evaluation falls back to absolute and the advantage goes. This entry does not treat count crushing the person, nor position sucking the choice.

Applying it

• When a take is required, put candidates in one viewport at the same time. Do not carousel one by one and score from memory.
• Ask “fitter for ××,” not “rate each.” Scores push people back into absolute evaluation.
• Write the shared task beside the pair (hero, subject line, summary for whom). Relative judgment has no ruler otherwise.
• Check: the same pair, first “is it good” in isolation, then “which is fitter” juxtaposed. If juxtaposed time drops and the pick is stable, relative evaluation is working. If people still score one by one after juxtaposition, the interface is still demanding an absolute judgment.

Related

• Same group: L3.07.2 When options differ too little, comparison collapses into picking noise · L3.07.3 More options bring choice difficulty and decision regret · L3.07.4 People favour the first and left positions; order must be controlled or randomised · L3.07.5 Multiple options expose the system’s instability — a cost and a form of honesty
• Nearby: L3.01 Generating Multiple Options · L1.01 Mismatch Between Stochastic Output and Deterministic UI
• Search terms: comparative evaluation · juxtaposition · relative judgment