Misrating

Misrating is the assignment of a rating that does not accurately reflect the quality or value of an item, service, or performance. In evaluative contexts, misrating can result from errors, bias, or deliberate manipulation, leading to distorted assessments and poorer decision making.

Misratings can be intentional or unintentional. Unintentional misratings arise from misunderstanding the rating scale, cognitive biases such as the halo effect, central tendency, leniency or severity bias, memory or recall errors, or ambiguous criteria. Intentional misratings include rating inflation, rating deflation, and the use of fake reviews or sockpuppets to influence perceptions.

Common domains include online marketplaces and streaming platforms, where user ratings affect visibility, pricing, and recommendations, as well as workplace performance reviews. Consequences include biased recommendations, unreliable quality signals, suboptimal purchases or selections, and unfair assessments of individuals or products.

Detection and mitigation strategies include data cleaning, anomaly detection, time-based weighting, user reliability scoring, and robust aggregation methods. Design choices such as clear rating scales, calibration prompts, verification of accounts, and limiting the influence of suspicious ratings can reduce misrating. Platforms may also implement transparency about scales and provide mechanisms for rating corrections or disputes.

See also: rating bias, response bias, fraud, recommender systems.