Statistical Quality Control System Overview and Underlying Principles
Official statistics are used as basic data to support establishment, monitoring and evaluation of public policies. A comprehensive quality assessment is conducted on the entire statistical business process from production to dissemination of official statistics to enhance reliability of official statistics through identifying areas for improvements and making necessary improvements.
Under the decentralized statistical system, 6 dimensions of statistical quality are defined to determine whether statistics provided to users are produced and disseminated adequately
Relevance: How well statistical information satisfies user needs in terms of coverage, concepts and contents from the user’s perspective.
Accuracy: How closely data measures the attributes or size of target population. Most statistics estimate the unknown true value. Accuracy represents the proximity of the unknown true value and estimated value.
Timeliness/Punctuality: The time difference between data reference point and data publication date. Timeliness represents how well statistics reflect the reality while punctuality represents how well data publication practice follows the pre-announced data release schedule.
Comparability: How well statistics of different time/space can be compared to each other by using the same standard (i.e. concept, classification, measurement tool, measurement process and primary data).
Coherence: How close are the results of statistics on the same socioeconomic phenomena but using different primary data or method.
Accessibility/Clarity: Accessibility represents how easily users can access statistics. Clarity represents the availability of information about how the statistics are produced.
Types of Statistical Quality Control System
Types: Regular, Ad-Hoc and Self-Assessment and quality improvement consulting
Regular Assessment: Major official statistics are assessed by outside experts every five years to determine areas for improvement and review whether necessary improvements have been made. (Statistics Act Article 9).
Ad-Hoc Assessment: Ad-hoc assessment is conducted when problems are detected based on monitoring and quality assessment is deemed necessary. (Statistics Act Article 10).
Self-Assessment: Agencies producing official statistics conduct self-assessment on statistical quality each year using a web-based program. (Statistics Act Article 11).
Monitoring on Quality Improvement: If an agency producing statistics faces difficulty in making necessary quality improvements due to lack of expertise, Statistics Korea can provide consulting service on statistics whose quality cannot be improved due to weak statistical infrastructure. (Statistics Act Article 13).