QUALITY GATES AND THE ABS DATA QUALITY FRAMEWORK
In May 2009, the ABS Data Quality Framework was released on the ABS website. The focus of that information paper was on the application of the ABS Data Quality Framework for use in activities relating to assessment of statistical outputs (quality about a product), using the seven dimensions of quality. However, the ABS Data Quality Framework can also be applied to other aspects of organisational quality including the Institutional Environment, statistical processes and statistical inputs (Eurostat 2009).
This paper discusses the aspect of managing the quality of statistical processes using quality gates. Quality gates link back to the ABS Data Quality Framework dimensions of Institutional Environment, Relevance, Timeliness, Accuracy, Coherence, Interpretability and Accessibility to ensure that all dimensions of quality have been considered and maintained in the production of the statistical outputs. Quality gates do this through their creation and use.
For example quality gates reflect:
- Institutional environment through their implementation as a statistical risk mitigation strategy, so that the reputation of the organisation is protected;
- Timeliness and Relevance by ensuring that they are placed at appropriate junctures in the process where they can influence the direction and outcome of the quality of the process. For example: preventing delays by identifying problems in time, confirming that the process is okay at a given point in time in order to continue to the next steps, or confirming that the output requirements are achievable from the design of the inputs;
- Accuracy and Coherence through the use of quality measures and their corresponding tolerance levels;
- Interpretability by encouraging documentation to ensure knowledge management and a shared understanding across all stakeholders of the quality gate and the underlying processes; and
- Accessibility through the provision of information on the quality of the process at the individual components. In some cases the desired information, for example a particular quality measure, may not be available due to current reporting limitations in the processes, but is something that should still be acknowledged as a requirement for future development opportunities.