Using Qualitative Methods for Ex Post Analysis of Local Population Projections

Joachim Singelmann, Louisiana State University
Michael Irwin, Duquesne University
Jack Beggs, Louisiana State University

This paper assesses population projection error associated with one time events, such as the gain or loss of a major employer, by integrating shift-share projection methods and qualitative interviews. Using back-casts from 1980 to 2000 in four Gulf of Mexico communities, we identified shifts from projected growth unique to each locality. Results of these back-cast projections were then used in qualitative interviews with local planners in each community. The interviews provided historical information on the types and timing of events that led to each specific divergence. This information was then integrated into new back-casts. Comparing the two sets of back-casts to actual population trends yields estimates of projection error associated with each type of local event and provides a quantitative measure of the increased accuracy gained by using qualitative information.