Performance-contingent payment can elicit expected rather than natural responses

What it is

Performance-contingent payment ties compensation to speed, accuracy, output, or another evaluation. It changes what participants optimize: observations reflect both interface-induced behavior and strategic adaptation to the bonus rule. Meeting expectations may mean inferring the hypothesis, but it may simply mean rationally pursuing the rewarded metric.

Why it happens

When the proxy and target construct diverge, effort shifts toward what earns payment—for example, trading comprehension for speed or repeatedly checking to maximize accuracy. Bonus instructions also create demand characteristics by signaling what researchers value. Engagement may rise while behavior becomes less representative of ordinary product goals.

Studying it

Randomize fixed-pay and bonus conditions, preregister primary outcomes, and measure the rewarded metric alongside unrewarded quality, strategy changes, and perceived task goals. Debrief questions can reveal how rules were understood but cannot prove every motive. If the real product includes incentives, reproduce its rule as a distinct ecologically relevant condition.

Where it stops holding

Performance pay is not inherently bias. Incentive compatibility may be necessary when studying effort, search costs, or economic choice. If real work is paid by output or accuracy, a bonus can improve ecological validity. Fixed pay also produces expectations; the central question is whether the incentive aligns with the construct and deployment context.

Applying it

Write down the behavior a rational participant would optimize under the rule and test whether it matches the target construct. Use a fixed participation payment as a floor and add bonuses only where needed. Explain calculation and caps, and disclose how the incentive limits generalization.

Related

• Same group: Q1.12.1 Low compensation and participation bias · Q1.12.2 High compensation and profit-seeking participation · Q1.12.3 Reward format and sample composition
• Adjacent: Q1.08 Sample size · Q1.14 Researcher bias and leading questions
• Search terms: performance-contingent incentive · demand characteristics · incentive compatibility