Online Ratings as Constructed Signals: A Systematic Review of Generation, Interpretation, and Consumer Outcomes

ABSTRACT

Aggregate ratings remain embedded in research on electronic word‐of‐mouth and online reviews, rather than being examined as distinct constructs. This review synthesizes how ratings are generated, interpreted, and translated into consumer outcomes across provider‐user streams. Following the SPAR‐4‐SLR, we searched Scopus, Web of Science, and EBSCOhost and retained 423 peer‐reviewed articles published between 2002 and 2025. We organized the evidence using the Theory–Context–Characteristics–Methodology (TCCM) framework and an actor‐based Antecedents–Decisions–Outcomes (ADO) framework. Ratings function as constructed signals shaped by participation, affect, social influence, platform architecture, and distortion. Rating generation and use form a recursive system, whereas effects operate through decision mechanisms whose competing theoretical explanations remain insufficiently distinguished. Behavioral influence does not establish consumer benefits; signal fidelity and judgment accuracy are distinct, whereas evidence on decision quality and welfare remains limited. These findings establish ratings as a distinct construct that integrates provider and user streams, diagnoses theoretical fragmentation, and separates fidelity, accuracy, decision quality, and benefits. Five research priorities extend these insights across recursion, theory adjudication, fidelity and accuracy, welfare, and boundary conditions. For platforms, the findings distinguish between interventions that target signal fidelity and those that support judgment accuracy.