Abstract

Modelled estimates are an integral and essential part of compiling international statistics. But as the famous quote warns us 'all models are wrong, but some are useful'. The trick is having sufficient statistical literacy to know which models are useful and which ones are not. Does the use of model based estimates undermine confidence in statistics? Do users even know or care? How can we improve the statistical literacy of governments, businesses and citizens so that they can engage with theseand many other issues - including data governance?