Population estimations Flashcards
Estimation methods: Symptomatic techniques: Censal-Ratio Method using Housing Units (1)
Data requirement:
(1) Measure of persons per household at estimate date
(2) Count of occupied housing units at estimate date
Major assumptions:
(1) Assumes housing change is symptomatic of population change
Pt = (OHUt * PPHt) + GQt, where:
Pt = Total population at the time of estimate OHUt = Occupied housing units at the time of estimate GQt = Group Quarters population at the time of estimate PPHt = Household size or population per household at the time of estimate (Pt / Ht) Pt = Non Group quarters population at the time of estimate Ht = Non Group quarters Occupied Households at the time of estimate
Estimation methods: Symptomatic techniques: Censal-Ratio Method using Housing Units (2)
Pt = (H0 + U) * (P0 / H0), where:
Pt = Population for estimates date P0 = Total household populations at the time of last census H0 = Occupied Households at the time of last census U = net change in occupied households between Pt and P0 PPH0 = (P0 / H0) = average household size
Estimation methods: Symptomatic techniques: Censal-Ratio Method using Housing Units (3)
Pt = (OHU0 + BPt - DPt - VUt) * PPHt + GQt, where:
Pt = Total population at the time of estimate OHU0 = Occupied housing units on the last census date GQt = Group Quarters population at the time of estimate PPH0 = Household size or population per household at the last census date (P0 / H0) P0 = Non Group quarters population at the last census H0 = Non Group quarters Occupied Households at the last census BPt = Building permits issued between Pt and P0 DPt = Demolition permits issued between Pt and P0 VUt = Vacant units between Pt and P0
Estimation methods: Regression-based techniques: Ratio-Correlation Method
Data requirement:
(1) Symptomatic indicators of population size (e.g., school enrollment, passenger vehicle registrations, etc. for estimate date).
(2) Total population at two periods prior to desired date.
Major assumptions:
(1) Population of areal units can be added to equal parent population.
(2) Change in indicators related to population change.
Multiple regression-based technique which compares change in one aerial unit to change in a parent area.
y = B0 + BiXi + … + BnXn + e, where:
y = dependent variable to be estimated (population) B0 = intercept to be estimated Bi = coefficient to be estimated Xi = independent variables (e.g., births, deaths, etc.) e = error term
Estimation methods: Regression-based techniques: Ratio-Correlation Method (basic steps)
1) Obtain coefficients (betas) for change occurring at 2 previous points in time. (example: 1980-90 to estimate 2000).
B0 = Pop. (city, 1990) / Pop. (County, 1990) divided by Pop. (city, 1980) / Pop. (County, 1980)
B1 = Births (city, 1990) / Births (county, 1990) divided by Births (city, 1980) / Births (county, 1980)
B2 = Deaths, B3 = School Enrollment, etc.
2) Apply coefficients to change during recent time period.
Estimation methods: Component techniques: Component Method II
Data requirements: (*complex)
(1) Total population of estimate area for base date
(2) Population by single years of age for base date
(3) * Elementary school enrollment for grades 1-8 for base date and estimate date
(4) Births and deaths by area of residence between base date and estimate year
(5) * Survival rates for the school-age population from the base date to the estimated date for base year (usually assumed to remain constant from base date to estimate date)
(6) Migration adjustment factor (derived from data for base date for use at estimate date); (Hoque used .1)
(7) * Resident, nongroup quarters population
Estimation methods: Component techniques: Component Method II (basic steps)
1) Total population of area for base date.
2) Add natural increase between the base date and the estimate date.
3) Add net migration between the base date and the estimate date.
4) The sum is the estimated total population of area for estimate date.
* Note: Complex procedure employed when separates
Estimation errors
Mean percent error (MPE)
Mean Absolute Percent Error (MAPE)
Mean Percent Absolute Difference (MPAD)
Population estimate
Size of past or current population of a specific geographic area for which Census counts are not available. Based on existing data (vs projections, based on times in which no data exist).
Two types: Intercensal & postcensal
Principles of estimation
1) Accuracy based on assumptions used
2) No specific methodology guarantees accuracy
3) Estimates (& projections) are usually more accurate for: (a) areas with large population, (b) total population, (c) shorter time periods (ex: 2011 vs. 2019), (d) areas with slow or stable growth patterns.
Limits of estimation (& projection) procedures
1) Data availability and quality
2) Changes in real boundaries
3) Changes in definitions
4) Coverage errors (undercount or overcount)
Common data adjustments required
1) Adjusting period of estimate (or projection) from Census date (April 1) to estimate or projection date (such as July 1)
2) Deriving values for parts of years from annual data (usually simply assume linear rate of occurrence)
Estimation methods: Extrapolitive techniques: Arithmetic
Assumption: Historical pattern of change applies to current estimation period.
Pn = P0 + bn, where:
P0 = population at base period Pn = population at end period b = annual amount of change; b = (Pn - P0)/n n = # years between periods
Estimation methods: Symptomatic techniques
1) Housing unit method
2) Electric meter method
3) School enrollment method
Estimation methods: Regression-based techniques
Data requirement: Total population and/or symptoms for series of time periods prior to estimate date.
Major assumption: linear change in total population. Historical trends apply to current conditions.
1) OLS
2) Ratio-correlation method
