The Jubilee Centre has been working with Demos and the Centre for the Analysis of Social Media (CASM) on a year-long study to analyse the ways in which moral virtues are discussed on Twitter, and the connections between this discourse and online virtuous action. The full report will be available at the end of August, with some initial findings presented below. The analysis considers over one million tweets sent from the UK, each containing one or more of the virtue terms ‘courage’, ‘empathy’, ‘honesty’, and ‘humility’. An initial 'cluster' analysis of the words these terms commonly appear alongside has produced the following key findings:
- There is a large and distinct cluster of discussion involving Brexit and UK politics with all four virtue terms, suggesting that this subject is very often seen through a moral lens in the UK.
- Discussions around humility and courage (clusters 1 and 4) are also tied to religion; this is notably absent from the empathy virtue cluster.
- Gratitude, and the expression of thanks, often appears in the centre of the clusters, closely connected to other forms of expression, and especially to other virtue-related language.
This study represents a continuation of the Jubilee Centre’s interest in the field of social media with the Centre having previously undertaken a research project aimed at examining how social media use is related to young people’s experience and enactment of empathy and honesty, and their identification with moral virtues. An excerpt from the final report 'Over the Character Limit' is available, here, which focuses on the clustering of virtue terms on Twitter.
Notes on interpretation
- Colours indicate different word classes, which are collections of words that frequently occur closely together but rarely with words from other classes. These classes develop purely from these interactions within the data - their contents are not defined through human intervention.
- The position of the word classes on the graph shows how similar the classes are to one another; two classes positioned next to one another contain words which are relatively likely to appear close together, though not likely enough to be placed in the same class.
- The size of the word indicates how ‘characteristic’ it is of that class; large words are very likely to occur alongside other words from that class and very unlikely to occur alongside words from other classes.