unreliability

Unreliability is the quality of being inconsistent or prone to failure to perform as expected. It is the opposite of reliability and can refer to devices, systems, information, or human performance. In engineering and computing, unreliability is quantified as the probability that a component or system fails within a given period. Common metrics include failure rate and mean time between failures. Causes include design defects, manufacturing variation, wear, environmental stressors, software bugs, and operator error. Unreliability can manifest as intermittent faults, degraded performance, or complete outage, with consequences ranging from nuisance to safety hazards or financial loss.

In information and research, unreliability describes data, sources, methods, or measurements that are inconsistent, biased, or not verifiable, reducing trust and usefulness. In psychology and forensics, unreliability pertains to memory or testimony that is not reproducible or is subject to distortion.

Mitigation strategies include improving design with redundancy, fault tolerance, and conservative failure modes; thorough testing and validation; regular maintenance; calibration; and clear documentation. For information quality, strategies include data provenance, versioning, replication, and peer review. Assessing reliability involves statistical analysis, reliability testing, and, in social sciences, measures of internal consistency and test-retest reliability.

In everyday use, unreliability can erode confidence and influence decisions, discussions, and collaborations.