Report on a multiple regression method for making population estimates

Area of Science:

• Demography
• Statistical Modeling
• Geospatial Analysis

Background:

• Accurate population estimates are crucial for resource allocation and policy-making.
• Traditional ratio-correlation methods have limitations in temporal stability and variable intercorrelations.

Purpose of the Study:

• Introduce and evaluate a new population estimation technique: the difference-correlation method.
• Compare the performance of the difference-correlation method against the standard ratio-correlation method.

Main Methods:

• Developed the difference-correlation method for population estimation.
• Applied both methods to estimate populations for Michigan counties.
• Analyzed mean percentage error and intercorrelation stability.

Main Results:

• The difference-correlation method yielded population estimates with a smaller mean percentage error than the ratio-correlation method.
• Differences of proportions in the new method have zero means, unlike varying means in ratios.
• Observed higher intercorrelations and increased temporal stability with the difference-correlation method.

Conclusions:

• The difference-correlation method is a more accurate and stable approach for population estimation.
• This method offers advantages in variable intercorrelation and temporal consistency over traditional techniques.