# Public Use Microdata Sample **PUMS** is the census product that hands a member of the public **individual records** — one row per person and one per housing unit — rather than a published table. It is free, requires no identification and no stated purpose, and is available instantly. It is the reason the [[Census Data Access Crosswalk]] ladder inverts at its second rung: a stranger gets individual-level records of other people at no cost, while a person's own record costs $155 through the [[Age Search Service]]. What makes that possible is **coarsening rather than aggregation**. The records are real; the identifying detail is blunted. Geography is truncated to the **Public Use Microdata Area (PUMA)**, which must contain at least **100,000 people** at the time of delineation. Income and age are top-coded. Rare categories are collapsed. PUMAs are redrawn each decade following the decennial census — delineated by the **State Data Centers** in cooperation with regional, state, local and tribal bodies, not by the [[Census Bureau]] directly — which means PUMA identifiers are not comparable across decades. PUMS is drawn from the decennial census and, more often in current use, from the [[American Community Survey]]. Its resolution depends entirely on the disclosure-avoidance method in force; see [[Disclosure Avoidance for Statistical Products]] and [[Differential Privacy]]. See [[Census Data Products Tree]] · [[Census]] · [[Title 13]]