A conceptual model for demographic statistics
Abdulla Gozalov, Sabine Warschburger United Nations Statistics Division Demographic Data Collection at the UN Statistics Division
{ Data on Population Estimates { Data on Vital Statistics
{ Data on Population Censuses (population, economic and household characteristics) { Data on Housing Censuses
{ Data on Migration Flows { Data on Disability Current Means of Data Collection
{ Collecting data directly from countries through Excel-based questionnaires using XML
{ Joint collection of data with organizations
{ Obtaining data from other organizations
{ Obtaining data from national publications and websites Current Means of Data Dissemination
{ Printed publications
{ Website: http://unstats.un.org/unsd/demographic/pr oducts/dyb/dyb2.htm (Demographic Yearbook in pdf and excel) http://data.un.org/ (UNdata Portal)
{ CDs
{ Data files Potential Use of SDMX
{ Data delivery from other international organizations { Data delivery to other international organizations
{ Data collection from countries { Dissemination Conceptual Model for Demographic Statistics
{ Based on the United Nations Principles and Recommendations z P&R For Population and Housing Censuses Rev. 2 z P&R For Vital Statistics Rev. 2
{ Adapted for international-level macro data Diverse Characteristics of Demographic Data
{ Dozens of applicable concepts, hundreds of possible tabulations, e.g. z Population by age, sex, and marital status z Live births by birth order and age of mother { Stocks and flows { Varied data collection practices { Overlap with other subject-matter domains (education, health, labor) The Model: Basic Considerations
{ Macro data { Impractical to create a single key family for demographic data { Should be sufficiently abstract to support various underlying systems { Should be sufficiently specific to maintain interoperability Approach to Conceptualization
{ A simple model: subjects of measurement (“Subjects”) and their possible demographic breakdowns (“Qualifiers”) z Subject: Total Population; Applicable Qualifiers: Sex, Age, Marital Status,… z Subject: Live Births; Applicable Qualifiers: Sex, Birth Weight, Age of Mother, … { Meta-data { Cross-domain concepts Subjects
{ Population census, population estimates, and vital statistics subjects { Stocks subjects and flows subjects { Characteristics such as unit of measure, applicable qualifiers { Key families not to be used with more than one subject each z Do not mix data with different characteristics z E.g. sex applies to births but not divorces Qualifiers
{ Provide breakdowns with various levels of detail { Data points in different breakdowns can be cross-referenced and validated
Observation 1 (key family A) Observation 2 (key family B) Subject = “Foreign population” Subject = “Foreign population” Reference Area = “Country X” Reference Area = “Country X” Reference Date = “July 1, 2008” Reference Date = “July 1, 2008” Sex = ”Male” Sex =”Male” Value = “1500” Age = “Total” Value = “1500” Key families
{ Key family combines a subject and a subset of its applicable qualifiers z Total Population: Age, Marital Status z Total Population: Age, Sex, Occupation { More flexible and usable than always using the full set; relationships better reflected { But focus on concepts requires tools to handle multiple key families with different breakdowns The Way Forward
{ Coordinate with other agencies { Finalize the model, incl. meta-data { Clarify use of cross-domain concepts { Develop code lists { Implement For More Information and Comments
{ Sabine Warschburger [email protected]
{ Abdulla Gozalov [email protected]