Census Study Guide

📖 Core Concepts Census: Complete, systematic count of every individual in a defined population; collects, records, and calculates data. Population & Housing Census: Most common type; records people and their dwellings. Agricultural Census: Counts agricultural holdings, livestock, land use, and production. De Facto Residence: Where a person is found on census day. De Jure Residence: Person’s usual, legal place of living (receives services). Permanent Residence: Fixed address used for official counts (e.g., family home). Register‑Based Census: Uses existing administrative registers instead of field enumeration. Post‑Enumeration Survey (Dual‑System Enumeration): Follow‑up survey that matches a sample to census records to estimate undercount. Capture‑Recapture: Statistical method (borrowed from wildlife studies) to estimate missed individuals by comparing two “captures.” --- 📌 Must Remember Universality: Census aims to count everyone in the target territory. Frequency: UN recommends a census at least every 10 years. Sampling Frame: Usually an address register; must be comprehensive to avoid missed dwellings. Short‑Form/Long‑Form: All households receive a short questionnaire; a random sample receives a detailed long form. Privacy Safeguards: Disclosure risk in small areas is mitigated by statistical disclosure control (e.g., data swapping, noise addition). Technology Trend: Handheld devices, internet self‑responses, and administrative register linkage are now standard. Key Uses: Apportioning representatives, allocating resources, planning development, monitoring poverty/inequality, informing agricultural policy. --- 🔄 Key Processes Preparation Build/refresh address register → define sampling frame. Design short‑form (all) and long‑form (sample) questionnaires. Enumeration Enumerators visit each dwelling (or respondents submit online). Special teams handle communal establishments (prisons, dorms). Post‑Enumeration Survey Select a random sample of households. Match (capture‑recapture) survey responses to census records → compute undercount rate. Data Integration Merge census data with administrative registers (population, housing, tax). Apply automated matching (propensity‑score matching) to reconcile records. Disclosure Control Apply swapping, noise injection, or model‑based synthesis before releasing micro‑data. --- 🔍 Key Comparisons Census vs. Sampling Census: Counts every individual; aims for total coverage. Sampling: Counts a subset; inference relies on statistical theory. De Facto vs. De Jure Residence De Facto: Where you happen to be on census day. De Jure: Your usual, legal residence (used for service provision). Register‑Based vs. Traditional Field Census Register‑Based: Uses existing admin data; cheaper, faster, but depends on register quality. Traditional: Requires full‑scale enumeration; higher cost, higher control over data collection. --- ⚠️ Common Misunderstandings “Census = Survey” – A census is not a sample survey; it strives for complete enumeration. “All households answer the long form” – Only a random sample receives the long‑form questionnaire. “De Facto always equals De Jure” – They differ for students, migrants, refugees, and temporary visitors. “Register‑based censuses eliminate undercount” – Registers can miss undocumented or transient populations; post‑enumeration checks may still be needed. --- 🧠 Mental Models / Intuition “Every door, one count” – Visualize the address register as a master list of doors; each door must be visited or linked to a record. “Two‑catch fish” – Capture‑recapture works like catching fish twice; the overlap tells you how many were missed the first time. “Layers of privacy” – Think of data release as a layered cake: raw data → apply swaps/noise → publish safe aggregates. --- 🚩 Exceptions & Edge Cases Multiple/Temporary Addresses – Students, holiday‑makers, refugees may be counted under de Jure, de Facto, or permanent rules depending on national guidelines. Out‑dated Address Registers – May miss new housing developments or communal establishments; requires supplemental field verification. Small‑Area Disclosure – Very small subpopulations can be re‑identified; special suppression or aggregation is required. --- 📍 When to Use Which Choose Register‑Based Census when: High‑quality, up‑to‑date administrative registers exist; Budget constraints limit field operations; Rapid data turnaround is needed. Choose Traditional Field Census when: Registers are incomplete or unreliable; Detailed socio‑economic questions (long form) are essential; Legal requirement mandates direct enumeration. Apply Short‑Form/Long‑Form Split when: Core demographic data needed from everyone, but detailed socio‑economic data only from a sample to reduce respondent burden. --- 👀 Patterns to Recognize “Address → Household → Person” hierarchy repeats across all census types. Undercount signals: high non‑response rates in transient groups, missing entries in address registers, or large discrepancies in post‑enumeration surveys. Privacy‑risk flag: any table that breaks down a variable to < 5 individuals in a geographic cell. --- 🗂️ Exam Traps Distractor: “Censuses are always conducted every 5 years.” – Wrong; UN recommends 10‑year interval. Distractor: “Long‑form questionnaires are sent to all households.” – Wrong; only a sample receives the long form. Distractor: “Register‑based censuses completely eliminate the need for field verification.” – Wrong; they still need quality checks for gaps and outdated registers. Distractor: “De Facto residence is used for allocating resources.” – Wrong; most allocation uses De Jure or permanent residence. ---