Covid-19 and possible bias in statistical information

Abstract

Abstract. A widely discussed theme at this time in Italy concerns the statistical data used to evaluate the trend of Covid-19, since some indicators have had considerable influence in determining policy choices. The statistical collection of some phenomena, such as ‘’causes of death’’, is not always easy, especially if the deceased person had additional pathologies. Moreover, the way in which statistical information on Covid-19 is disseminated also contributes to create a ‘framing’ that can affect the use of the variables (or indicators) analysed, to the point of producing, in some cases, a loss of information expressed in a ‘damage function’. This becomes more relevant when an increasing amount of information contains measurement errors or interpretative framing. The aim of the considerations expressed in this article is to identify a useful tool for the serious pandemic situation that Italy and the world currently face, focusing on the correct use of statistics. The entire process of statistical collection, including its communication, should be monitored because public choices should depend on the use of statistics.