disproportionalitet

Disproportionalitet is a concept used across disciplines to describe an imbalance where parts do not scale with the whole, or where outcomes are not distributed in proportion to a given baseline or expectation. The term is used in statistics, sociology, public policy, health, and pharmacovigilance to describe related but distinct phenomena.

In statistics, disproportionalitet refers to situations where observed frequencies differ from those expected under a null model. It is commonly assessed by comparing observed counts to expected counts, calculating risk ratios or standardized incidence ratios, and testing whether deviations are statistically significant. In practice, disproportionalitet alerts researchers to possible biases, confounding, or structural effects that merit further study.

In social sciences and public policy, disproportionalitet denotes unequal representation or impact of certain groups—such as by race, gender, age, or socioeconomic status—in areas like policing, education, housing, or health care. It is a descriptive observation that may reflect structural bias, data limitations, or policy design. It does not, on its own, establish intent or discrimination but signals areas for review.

In medicine and pharmacovigilance, disproportionalitet is used in signal detection to identify potential associations between drugs and adverse events by comparing observed reports to expected levels in large databases.

Addressing disproportionalitet involves data quality improvements, appropriate stratification, statistical adjustment, and policy or program changes aimed at fairer distributions and outcomes.

Etymology: from Latin disproportio, adapted into Scandinavian languages. See also proportionality, bias, equity.