No abstract

No abstract

No abstract

The news arguably serves to inform the quantitative reasoning (QR) of news audiences. Before one can contemplate how well the news serves this function, we first need to determine how much QR typical news stories require from readers. This paper assesses the amount of quantitative content present in a wide array of media sources, and the types of QR required for audiences to make sense of the information presented. We build a corpus of 230 US news reports across four topic areas (health, science, economy, and politics) in February 2020. After classifying reports for QR required at both the conceptual and phrase levels, we find that the news stories in our sample can largely be classified along a single dimension: The amount of quantitative information they contain. There were two main types of quantitative clauses: those reporting on magnitude and those reporting on comparisons. While economy and health reporting required significantly more QR than science or politics reporting, we could not reliably differentiate the topic area based on story-level requirements for quantitative knowledge and clause-level quantitative content. Instead, we find three reliable clusters of stories based on the amounts and types of quantitative information in the news stories.

Because people are constantly confronted with numbers and mathematical concepts in the news, we have embarked on a project to create journalism that can support news users’ number skills. But doing so requires understanding (1) journalists’ ability to reason with numbers, (2) other adults’ ability to do so, and (3) the attributes and affordances of news. In this paper, we focus on the relationship between adults’ news habits and their quantitative reasoning skills. We collected data from a sample of 1,200 US adults, testing their ability to interpret statistical results and asking them to report their news habits. The assessment we developed differentiated the skills of adults in our sample and conformed to the theoretical and statistical assumption that such skills are normally distributed in the population overall. We also found that respondents could be clustered into six distinct groups on the basis of news repertoires (overall patterns of usage, including frequency of news use overall and choice of news outlets). As often assumed in the literature on quantitative reasoning, these news repertoires predicted quantitative reasoning skills better than the amount of quantification in the outlets, but they still predicted only a small fraction of the variance. These results may suggest that news habits may play a smaller or less direct role in quantitative reasoning than has previously been assumed. We speculate that the presence (or absence) of quantification in everyday activities – namely work and hobbies – may be a better predictor of adults’ quantitative reasoning, as may additional dimensions of news habits and affective responses to numbers.