7105.0.55.003 - National Agricultural Statistics Review - Preliminary findings, 2013-14  
ARCHIVED ISSUE Released at 11:30 AM (CANBERRA TIME) 31/03/2014  First Issue
   Page tools: Print Print Page Print all pages in this productPrint All

Data quality

A theme area of nearly all the NASR’s public submissions and a topic raised in consultations was data quality. This theme was reflected by stakeholders when reporting both how well existing agricultural statistical assets meet their information needs and the barriers to the NASIS functioning effectively.

The NASR will adopt standard terminology for this discussion by referencing the ABS’ Data Quality Framework (DQ Framework). The DQ Framework reflects the full range of dimensions of data quality and provides a framework for both the development of statistical collections to produce fit for purpose statistics, as well as for the assessment and reporting of the quality of statistics. The seven dimensions of data quality are:

    1. Institutional Environment: considers institutional and organisational factors that may have a significant influence on the effectiveness and credibility of the agency producing the statistics.

    2. Relevance: considers how well the statistics meet user needs (e.g. scope and coverage, geographic detail, classifications).

    3. Timeliness: refers to the time between the reference period and that date when data becomes available.

    4. Accuracy: considers the degree to which the statistics describe the concept they were designed to measure and should be assessed in terms of the major sources of errors that can cause inaccuracy (e.g. sample error, non-response error).

    5. Coherence: considers the internal consistency of the statistical collection as well as its comparability with other information sources.

    6. Interpretability: considers the availability of information to provide insight into the data including the availability of metadata and measures of accuracy.

    7. Accessibility: refers to the ease of access of data by users, including the ease of ascertaining the existence of information and the suitability of the form through which information is accessed.

Input from stakeholders reflected that prior to defining a specific information need, no one dimension of data quality is more important than another. When the detail of an information need is defined, this is where a particular dimension of data quality may be more important than others.

There is one aspect of the first dimension of data quality (Institutional Environment) which considers the adequacy of resources which stakeholders referred to as a driving consideration in being able to produce and/or use statistics for their purpose. Resourcing includes the funds, materials and staff (including their capability level) and is discussed in more detail below.

Stakeholders also described the need for standards and common classifications to be used by stakeholders across the NASIS, regardless of whether they are a user, producer or custodian of statistics. The DQ Framework is an example of one such standard that can be adopted by stakeholders of the NASIS to align the language with which statistical assets are described and provide a consistent approach for stakeholders to use in the development of a new statistical collection.