[Modeling matching error and its effect on estimates of census coverage error]

Area of Science:

• Statistics
• Demography
• Survey Methodology

Background:

• Census data is vital for resource allocation and policy.
• Accurate census counts are essential for representative democracy.
• Undercounts in censuses can lead to significant demographic and socioeconomic misrepresentations.

Purpose of the Study:

• To propose a statistical model for evaluating the impact of matching error on census undercount estimators.
• To analyze the components of mean square error (MSE) attributable to matching discrepancies.
• To illustrate the application of the model using data from the 1990 U.S. census undercount evaluation program.

Main Methods:

• Development of a statistical model to quantify matching error effects.
• Derivation of the mean square error (MSE) for the dual system estimator under the proposed model.
• Analysis of MSE components specifically related to matching errors.

Main Results:

• The study defines and explains the components of MSE arising from matching error.
• It quantifies the effect of matching error on the estimator of census undercount under the assumed model.
• A methodology for optimizing the design of matching error evaluation studies is presented.

Conclusions:

• Matching error significantly influences the accuracy of census undercount estimates.
• The proposed model provides a framework for understanding and mitigating these errors.
• The findings support the development of more robust census evaluation methodologies.