Improving Local Outcome Indicators Project
Criteria for indicators
Any indicator chosen to track performance of Single Outcome Agreements needs to be relevant and practical. The list below outlines a range of criteria that all potential indicators should be assessed against before use in SOAs.
1. Relevant and unambiguous
The indicator should be clearly and directly relevant to at least one of the high level outcomes that are being sought. It may not be a direct measure of the outcome but it should be a clear and unambiguous indicator of progress toward that outcome. The definition should allow for non-experts to understand the indicator and there should be no possibility of misinterpretation.
2. Harmonised with other frameworks and concepts
The definition of the indicator should be harmonised with any similar measures being used in other frameworks, performance management systems, legislation or national or international conventions.
3. Timely and accessible
The data should be published regularly enough to tie in with the SOA reporting arrangements, the time lag between recording and reporting of data should be minimal, and the data should be easily accessible to all (i.e. available publicly).
4. Statistically Robust
For data from surveys:
The data should be precise enough to measure change (i.e. the confidence intervals should be narrow enough that it can be reliably reported whether or not the target has been achieved at the relevant geography or for the given sub-group).
The data should be based on a sample that is representative of the relevant population and collected using recognised best practice in surveys.
For data from administrative systems:
All bias and error in the data should be recognised and the implications assessed against the usefulness of the data. There should be minimal risk to changes in systems and recording practice over time. The data should be fully quality assured and auditable.
The cost of collecting the data to a sufficient quality standard should be outweighed by the usefulness and utility of the data.