On the utility of population forecasts.

J Tayman, D A Swanson
Author Information
  1. J Tayman: San Diego Association of Governments, CA 92101, USA. jta@sandag.cog.ca.us

Abstract

Many customers demand population forecasts, particularly for small areas. Although the forecast evaluation literature is extensive, it is dominated by a focus on accuracy. We go beyond accuracy by examining the concept of forecast utility in an evaluation of a sample of 2,709 counties and census tracts. We find that forecasters provide "value-added" knowledge for areas experiencing rapid change or areas with relatively large populations. For other areas, reduced value is more common than added value. Our results suggest that new forecasting strategies and methods such as composite modeling may substantially improve forecast utility.

References

  1. J Am Stat Assoc. 1987 Dec;82(400):991-1,012 [PMID: 12155376]
  2. J Am Stat Assoc. 1995 Mar;90(429):64-71 [PMID: 12155398]
  3. J Am Stat Assoc. 1972 Jun;67(338):347-63 [PMID: 12309295]

MeSH Term

Bias
Data Interpretation, Statistical
Forecasting
Humans
Population Density
Population Growth
Reproducibility of Results
United States

Word Cloud

Created with Highcharts 10.0.0areasforecastutilitypopulationforecastsevaluationaccuracyvalueManycustomersdemandparticularlysmallAlthoughliteratureextensivedominatedfocusgobeyondexaminingconceptsample2709countiescensustractsfindforecastersprovide"value-added"knowledgeexperiencingrapidchangerelativelylargepopulationsreducedcommonaddedresultssuggestnewforecastingstrategiesmethodscompositemodelingmaysubstantiallyimprove

Similar Articles

Cited By