underreport

Underreport is the act of presenting information at a lower level than its true value. It can apply to numerical data, events, or outcomes and may occur across many domains. Underreporting can be intentional, as in concealment or fraud, or unintentional, resulting from error, stigma, privacy concerns, or limitations in data collection.

Causes include deliberate concealment or manipulation, social desirability bias, recall bias, nonresponse, and reporting system gaps. Structural factors, such as complex reporting requirements or incentives that reward seeming lower risk, can also contribute.

Common contexts include epidemiology and public health, where underreporting of diseases or adverse events can delay detection of outbreaks; criminology, where victims may fail to report crimes; finance and taxation, where income, expenditures, or fraud may be understated; and social science research, where sensitive behaviors are underreported.

Consequences involve biased estimates of prevalence or magnitude, distorted risk assessment, misallocation of resources, and erosion of public trust. It can also create a “dark figure” of data that complicates evaluation and policy.

Mitigation strategies emphasize improving data collection and verification: standardized reporting protocols, triangulation with multiple data sources, audits and independent verification, data anonymization to reduce reporting barriers, and statistical methods to adjust for underreporting, such as capture-recapture or imputation where appropriate.