Scholarlink

Data quality is of utmost importance in large-scale assessments, as it directly impacts the reliability, validity, and comparability of survey outcomes. It is the role of psychometric data forensics to identify poor-quality or invalid data, enabling the detection of anomalies and potential error sources. Such forensics include harnessing process data to extract meaningful insights into test-taker behavior, leveraging cumulative databases to compare data congruency across multiple reference points, automating procedures to expedite anomaly detection and optimize resources, and collecting information about testing conditions and environmental contexts to better illuminate observable patterns in the data. Using examples from the OECD’s Programme for the International Assessment of Adult Competencies (PIAAC) and the Programme for International Student Assessment (PISA), this paper outlines these various data quality procedures and monitoring approaches to safeguard data integrity throughout the assessment lifecycle, while recognizing the dynamic and time-sensitive nature of data validation.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Psychometric Forensics to Maintain and Improve Data Quality in Large-Scale Assessments

  • 章节

  • 发表日期：

    2025年

查看全文
  • Kentaro Yamamoto,
  • Usama S. Ali,
  • Frederic Robin,
  • Lokesh Kapur,
  • Mathew M. Kandathil

摘要

Data quality is of utmost importance in large-scale assessments, as it directly impacts the reliability, validity, and comparability of survey outcomes. It is the role of psychometric data forensics to identify poor-quality or invalid data, enabling the detection of anomalies and potential error sources. Such forensics include harnessing process data to extract meaningful insights into test-taker behavior, leveraging cumulative databases to compare data congruency across multiple reference points, automating procedures to expedite anomaly detection and optimize resources, and collecting information about testing conditions and environmental contexts to better illuminate observable patterns in the data. Using examples from the OECD’s Programme for the International Assessment of Adult Competencies (PIAAC) and the Programme for International Student Assessment (PISA), this paper outlines these various data quality procedures and monitoring approaches to safeguard data integrity throughout the assessment lifecycle, while recognizing the dynamic and time-sensitive nature of data validation.