incompleteinformation

Incomplete information describes a situation in which a decision maker does not know all relevant aspects of the state of the world or the actions and preferences of other agents, resulting in uncertainty about outcomes. This category includes missing observations, hidden variables, and private information that is not publicly available. It is distinguished from incomplete data in a statistical sense; incomplete data is a lack of observed variables in a dataset, while incomplete information emphasizes the epistemic state of the decision maker.

In economics and game theory, incomplete information refers to environments where participants have private information about their characteristics, payoffs, or constraints. The standard analysis uses the concept of types and belief hierarchies, often modeled with priors and Bayesian updating, to reason about strategies under uncertainty about others' private information.

In statistics and data science, incomplete information often takes the form of missing data. Approaches to address it include imputation (single or multiple), data augmentation, and the Expectation-Maximization algorithm. Bayesian methods treat unknown quantities as random variables with prior distributions and update beliefs with observed data, producing posterior distributions that reflect the remaining uncertainty.

Implications of incomplete information include greater decision risk, the need for robust or conservative strategies, and sensitivity analysis to assess how conclusions depend on uncertain inputs. Transparent reporting of assumptions and uncertainty quantification are standard practices to manage incomplete information.