The project objective is to conduct a review of the current practices in measuring and communicating uncertainty of disseminated data, and to study the latest developments in the field and their applicability to official statistics. The expected results include enhanced recommendations to tackle the misinterpretation, due to errors or uncertainties, of the information the statistics reveal.
More in detail, the project will identify the main causes and sources of uncertainty, providing for each of them a clear definition, detailed description and analysis of their impact in official statistics. The project will also deal with a detailed presentation on advanced methods based on statistical and econometric technics to estimate and present uncertainty.
Theoretical, empirical and computational aspects of the alternative methods will be described, as well as their advantages and drawbacks. With particular reference to the construction of the consumer price indices based on scanner data, a case study will be carried out to show how the associated uncertainty can be estimated and presented.
At the end, the project will provide enhanced recommendations on the measurement of uncertainties in official statistics and will give practical advice on how to represent the uncertainty of datasets and data points for dissemination in an easily understandable way.