datadekking

Datadekking is the extent to which data are available and usable for the population, time period, and variables of interest in a study or dataset. It is a component of data quality that affects representativeness and the reliability of analyses. Coverage can refer to population coverage (who is included), variable coverage (which measures are present), and geographic or temporal coverage (where and when data are collected).

Measuring datadekking involves indicators such as response rates, coverage rates, and the completeness of variables. Analysts compare observed data to a reference population or census, and assess patterns of missingness. Coverage error is distinguished from sampling error, helping to understand potential biases in estimates.

Sources and methods to improve coverage include administrative data, surveys, sensor networks, and linked records. Techniques to enhance datadekking comprise data integration, record linkage, weighting, post-stratification, and imputation. Where coverage is incomplete, researchers must consider biases and uncertainty in their conclusions and document limitations.

Applications of datadekking span health, social sciences, environmental monitoring, and policy evaluation. High data coverage supports generalizability and robust decision-making, while low coverage can lead to biased results and misinformed policies. Addressing datadekking often involves trade-offs between granularity, accuracy, and privacy, and requires careful governance, transparent reporting, and, when possible, strategies to mitigate coverage gaps.