ERIC Number: ED422364
Record Type: Non-Journal
Publication Date: 1998-Apr
Reference Count: N/A
Regression Residuals as Performance Measures: An Assessment of Robustness in the Context of Country-Level Data. Draft.
Wang, Jia; Jamison, Dean T.
Modeling outcomes of school, hospital, or country performance as a consequence of initial conditions and assessing the entity's actual performance relative to what the model predicts (i.e., constructing a regression residual) is a natural way to measure performance. The usefulness of residual-based measures of performance depends on their robustness with respect to alternative statistical estimation procedures and model specifications. This paper assesses robustness for an analysis of country-level performance with respect to male and female adult mortality rates controlling for the country's income and education levels or for income alone. The data, which come from several different sources, including records from the World Bank, consist of 505 observations from 77 countries. Performance assessments of the more desirable models for each indicator are highly correlated, giving confidence in the robustness of the results. Country performance with respect to female mortality often differs substantially from that with respect to male mortality, however, pointing to the importance of separate rankings. While the paper concludes that residual-based performance measures work well in the context examined, the analysis also suggests that robustness may be context specific. Methods used in this paper to assess robustness can be extended to other contexts. (Contains 9 tables, 4 figures, and 13 references.) (Author/SLD)
Publication Type: Reports - Evaluative; Speeches/Meeting Papers
Education Level: N/A
Sponsor: California Univ., Los Angeles.; Norwegian Ministry of Foreign Affairs, Oslo.
Authoring Institution: N/A