Quality Assessment is an important part of a Quality Assurance plan and is defined as the testing of datasets to determine their quality for documentation purposes. Since this project integrates existing datasets as well as new ones, it may not have control over original production in many cases. Quality Assessment, therefore, becomes a retrospective activity that results in new documentation of problems and artifacts (published in the Dataset Documentation).
Quality Assurance objectives are discussed here and in the related Peer Review section that follows. It is also expanded upon in Appendix A, Task 4, 5 and 6. Quality Assurance guidelines for the project are also described in the Project Development Plan.
Quality Assurance, addressing issues of lineage, standards, and uncertainty in large spatial datasets, is a primary goal in building a commonly accessible characterization database. It is complimentary to our goal of scientific support, and is the motivation for much of the cooperative activity associated with the project, and for adopting an approach that employs community-wide participation by the global data community as a whole, and methodological conventions, as emerging in GIS and related technology for handling large databases. This project addresses Quality Assurance on two levels, first in regard to specific Quality Control procedures at NGDC, and second in regard to uncertainty in the datasets and their usefulness for global change research
The first level -- Quality Control procedures at NGDC -- involves internal processing, testing, and documentation procedures, as well as the pre-release review established under the project. These operating procedures are described in various documents within the Data Center, and summarized in the Project Development Plan and Technical Work Plans (not repeated here).
The second level -- the issue of uncertainty in the datasets and their usefulness for global change research -- requires a much broader approach. This is done through efforts to improve technical and scientific verifiability of the datasets individually and as an integrated whole. Improved documentation and meta-data contribute to this effort, as do internal quality checks performed during processing. The greatest contribution, however, may be the integration of these data into a completely functional package, with common structures and useful tools for intercomparison and analysis. This step provides a common method for verification and improvement of the data. By also establishing a widely-based peer-review effort, it is possible to assess the effectiveness of database design, integration, and documentation efforts, and to incorporate review comments into the documentation itself.
Feedback from public use is considered an important part of the review process, although it is clearly less direct than selective reviews. The value of public review can be greatly enhanced, however, by adopting data publication conventions similar to those for reviewed literature, where specific datasets can be commonly referenced in a stable form and all necessary information is immediately available for their evaluation.
Following these recommendations, the practice is continued here of peer-reviewing datasets and increasing emphasis on scientific standards for data publication. Publishing these datasets in a geographically integrated GIS structure was also continued to facilitate quality control and assessment, data intercomparison, and combined use. No priority or application is implied by the specific combination of data provided.
Procedure: Assess scientific grounding, statistical nature, sampling design, and derivation method; perform appropriate analyses and verification tests; assess technical presentation of data; assess accuracy and completeness of documentation; assess potential applications and scientific value.
Stage 2: Technical review of processing and publishing procedures: This involves internal testing of all processing, integration, documentation, and production procedures.
Purpose: Quality assurance of all work performed for publication.
Procedure: Evaluate and quality control data processing, additions to documentation, editing, database structure and functionality, product development and testing.
Stage 3: Public Distribution and User Support: This provides public availability on an at-cost or exchange basis, dissemination to specific groups, and user support.
Purpose: Obtain feedback from users and improve support for scientific applications and outreach activities.
Procedure: Dissemination to the scientific community, data exchange, updates, and additions, conduct and respond to user needs assessments.
Return to top of Quality Assurance page