Microsimulation of an educational attainment register to predict future record linkage quality

Area of Science:

• Social Sciences
• Computer Science
• Education

Background:

• Population-wide educational attainment registers are crucial for planning and research.
• Linking databases is essential for maintaining these registers.
• Unique national identification numbers are often unavailable, necessitating record linkage using quasi-identifiers.

Purpose of the Study:

• To inform legislation on the minimum set of identifiers for a national educational register.
• To empirically justify recommendations for identifier selection.
• To compare different record linkage methods and assess their impact on data quality.

Main Methods:

• Microsimulation of approximately 20 million individuals to model identifier changes and errors.
• Yearly linkage of updated records to a simulated register.
• Comparison of clear-text and privacy-preserving record linkage (PPRL) methods.

Main Results:

• Linkage bias occurs when only primary identifiers are used.
• More detailed identifiers, including place of birth, are necessary to minimize linkage bias.
• The quantity of identifying information is more critical for linkage quality than the specific linkage method used.

Conclusions:

• Microsimulation is a valuable tool for designing and evaluating record linkage procedures.
• Modeling identifier changes and errors can predict the data quality of future registers.
• Detailed identifiers are essential for high-quality educational registers.