Appendix G: Glossary

This glossary defines terms used frequently in this guide. The definitions are intended to explain how the terms are used in the 2030 Census context rather than provide formal technical or legal definitions.

Administrative and supplemental data
Information collected originally for another purpose that Census can use to help conduct the census. Administrative data generally come from government programs or records, while supplemental data may come from other sources, including commercial or third-party datasets. Examples can include information used to identify addresses, determine whether a housing unit is occupied, link people to addresses, or fill gaps when a household does not respond directly.

Coarsening
Reducing the detail in published data to protect confidentiality. Census might coarsen data by combining geographic areas, grouping population categories, publishing broader ranges, or otherwise making the information less specific.

Coverage
Whether the census includes the people and housing units it is supposed to count and places them in the correct location. A coverage error can be an omission, an erroneous inclusion, a duplicate, or a person counted in the wrong place.

Coverage Estimation
The planned 2030 operation that will estimate how many people and housing units the census missed, counted more than once, or counted incorrectly. It is the successor to the 2020 Post-Enumeration Survey and will combine independent field interviews, administrative records, matching, and statistical estimation.

Demographic Analysis
An independent way of evaluating census coverage using demographic records and estimates rather than another census-style survey. It builds population estimates primarily from births, deaths, international migration, and other administrative data and compares those estimates with census results.

Disclosure avoidance
Methods Census uses to publish useful statistics while reducing the risk that someone could use those statistics to learn confidential information about a particular person or household. Differential privacy, swapping, suppression, and coarsening are different approaches to disclosure avoidance.

Dual-system estimation
A statistical method for estimating people missed by the census by comparing two largely independent sources: the census itself and a separate coverage survey or measurement system. People found in both sources, only one source, or neither source are used to estimate the total population and the size of the undercount or overcount.

Enumeration
The process of counting people and determining where they should be counted for the census. Enumeration can occur through self-response, an interview with a Census worker, information from another knowledgeable person, administrative records, or other approved methods.

Historically undercounted population
A population believed to be missed or counted inaccurately at higher rates than other populations. Census often uses this term more narrowly for groups whose undercounts or overcounts can be measured through Coverage Estimation, such as young children or particular racial and ethnic populations. Advocates may use it more broadly to include communities that are likely to experience census undercounts or other coverage problems even when existing coverage-measurement methods cannot estimate those errors directly, such as LGBTQ people and people with disabilities. In this guide, the term is generally used in this broader sense. It describes a census-coverage risk, not an inherent characteristic of the community.

Imputation
A statistical or processing method used to fill in information that Census cannot obtain directly. Depending on the situation, imputation can be used to determine whether a housing unit is occupied, how many people live there, or missing characteristics about the people counted.

In-Office Enumeration (IOE)
A new 2030 operational area that uses administrative and supplemental data, models, and other information to help resolve nonresponding housing units from an office rather than entirely through additional field visits. In some cases, IOE may result in people being counted without Census obtaining a direct household response.

Matching
The process of determining whether records from different sources refer to the same person, address, or housing unit. Matching is important for linking administrative data, identifying duplicate responses, processing responses without Census IDs, building frames, and estimating census coverage.

Near real-time processing
Processing and evaluating census information while data collection is still underway rather than waiting until an operation is complete. This can allow Census to identify duplicates, anomalies, missing information, or quality problems early enough to change fieldwork or correct cases.

Person Characteristic Frame (PCF)
A planned 2030 Census resource that links information about people to addresses and associated characteristics, household information, and measures of data quality. It will draw on Census and outside data sources and support matching, In-Office Enumeration, quality control, and other census operations.

Post-Enumeration Survey (PES)
The independent survey used in prior censuses to estimate coverage errors by comparing a separate sample of households with census results. For 2030, Census is redesigning this work as Coverage Estimation rather than relying exclusively on independent post-census field data collection.

Proxy response
Information about a household provided by someone other than a member of that household, such as a neighbor, landlord, or building manager. Census may seek a proxy response when repeated attempts to reach household residents are unsuccessful.

Response sufficiency
The standard Census uses to decide whether it has enough information from a response to use or process it without seeking additional information. A response could be sufficient for one purpose, such as establishing the number of people in a household, while still lacking information needed for individual characteristics.

Self-response
A census response provided directly by a household or person without an enumerator obtaining the information through field follow-up. Self-response can occur through the internet, paper, telephone, or other authorized response channels.

Suppression
Withholding a value, cell, table, or other piece of published information because releasing it could create an unacceptable confidentiality risk. Additional values may sometimes need to be suppressed as well so that the original value cannot simply be calculated from surrounding totals. This is often called complementary suppression.