It is an age-old debate in statistics and measurement: when to provide a relatively large number of series of data (a dashboard) to permit the user to extract a common signal themselves, and when to create an artificial ‘single-number’ index through some process of amalgamation or aggregation such that a single index provides a simple (some say over-simplified) line of sight for users on whether matters have improved, worsened, or stayed constant.
This paper reflects on recent experience in the UK, in relation to four specific instances: the construction of Gross and Net Inclusive Income, the development of Public Service Productivity estimates, the National Wellbeing Dashboard and the creation of the Health Index to consider what these examples teach us, to consider how we can move forward in terms of when composite indices are valuable / useful and when they are not.
The paper concludes that a composite index has value when the weights used to derive it represent societal preferences or a societal reflection of value, and reviews alternative approaches which have been proposed or implemented.