Related Experiment Videos
Using census data to investigate the causes of the ecological fallacy
Environment & Planning A
|September 26, 2002
Summary
This study combines Anonymised Records (SAR) and Small Area Statistics (SAS) data to address the ecological fallacy in census analysis. The research develops a method to adjust aggregate statistics using individual-level data, improving ecological inference.
Area of Science:
- Demography
- Social Statistics
- Geographic Information Science
Background:
- The ecological fallacy occurs when inferences about individuals are made from aggregate data.
- Existing census data analysis methods can be prone to ecological fallacies.
- Understanding within-area homogeneity is crucial for accurate ecological inference.
Purpose of the Study:
- To investigate the causes of the ecological fallacy using combined census datasets.
- To develop and demonstrate a methodology for adjusting aggregate statistics.
- To improve the accuracy of ecological analyses at the Enumeration District (ED) level.
Main Methods:
- Integration of 2% Sample of Anonymised Records (SAR) with Small Area Statistics (SAS) databases.
- Analysis of census variables across multiple SAR districts in England.
- Application of a novel methodology using individual-level data to adjust aggregate statistics.
Main Results:
- Demonstrated the feasibility of combining SAR and SAS data for ecological fallacy research.
- Identified key census variables contributing to within-area homogeneity.
- Provided comparable analysis results from the 1986 Australian census.
Conclusions:
- The proposed methodology effectively adjusts aggregate-level statistics by incorporating individual-level information.
- Combining different census data sources enhances the ability to study and correct for the ecological fallacy.
- Accurate ecological inference requires accounting for within-area individual-level variations.