undercounts

An undercount occurs when the number of people, households, or units recorded in a census, survey, or administrative count is smaller than the actual figure. Undercounts differ from overcounts, where more units are recorded than exist. They can occur in censuses, population surveys, and administrative data systems, and they produce biased estimates if not corrected.

Causes include nonresponse, sampling error, measurement error, coverage error (missing places or people), misreporting, and data processing mistakes. Certain populations are more prone to undercounting, including children, racial and ethnic minorities, immigrants, people experiencing homelessness, rural residents, and those with unstable housing or language barriers. Migration during data collection can also cause undercounts.

Consequences: undercounts distort demographic estimates, health and social indicators, and resource distribution. They can affect political representation, funding allocations, school and hospital placement, and research conclusions. The magnitude of undercounting is often reported as a rate or percentage.

Mitigation and estimation: surveys use follow-up interviews and administrative records to improve coverage. Methods such as dual-system estimation or capture-recapture help estimate the size of the undercount. Weighting adjustments, imputation, and data linkage to civil registries are used to correct estimates. Outreach, multilingual materials, trusted community partnerships, and flexible enumeration strategies are employed to reduce undercounts in future data collection cycles.