A Kardashian effect on research?

Few names are more synonymous with the cult of celebrity than the Kardashians. In less than two decades, the family have leveraged their personal lives to build a global media empire with impact across the entertainment, beauty and wellness industries, technology and social media and even politics and science. But do their names come up in research articles and academic books? If they do, what is the meaning and impact? This is what Digital Science’s Mike Taylor and Carlos Areia looked into when they authored a chapter in Kardashians, a Critical Anthology published by Routledge. 

Using Altmetric and Dimensions data to investigate trends around Kardashian related research. Mike and Carlos’ chapter, In the lab with the Kardashians: how Kardashian-linked research finds its audience presents a map of ‘Kardashian studies’, shedding light on how this research is received and discussed in social media, focusing on X (formerly known as Twitter). 

Mike, Head of Data Insights at Digital Science, has been fascinated by how popular culture and celebrity can draw people into science. “A post mortem of an infamous Earl, Alexander Lindsay, First Earl of Balcarres in the 17th Century provoked public conversations about anatomy, and more recently, Buffy the Vampire Slayer has been a lens to attract attention to areas of study such as sexuality or representations of fables in the media,” he says. So, when a friend suggested he look into the Kardashian effect on research trends he took up the challenge. Together with Carlos, Senior Data Scientist at Digital Science, they used several tools to create the analysis discussed in this chapter, that included a combination of tailored Dimensions searches to identify Kardashian related studies, and linked Altmetric attention data to evaluate the discourse and engagement with this research in the Altmetric Share of Voice dashboard.

 “We generated the map using VOS Viewer, inside Dimensions) of terms that appear in the title and abstract of the selected research, with the relationships between those terms clustered according to the frequency that they appear,” Mike explains (Figure 1). 

This map was used to identify five clusters of Kardashian-related research: 

• Cluster 1 – Wealth, Influence and Social Networks (red) 
• Cluster 2 – Celebrity, Fame and Public Identity (green) 
• Cluster 3 – Feminism, Politics and Social Change (blue)
• Cluster 4 – Sex, Gender and Identity (yellow) 
• Cluster 5 – Mental Wellbeing, Youth and Digital Life (purple)

Using a novel technique suggested by Dr Fabio Gouveia of Fiocruz/BIIST, we extracted citation data from Dimensions, X/Twitter data from Altmetric, and replaced citation counts with tweet counts to enable a direct comparison (Figure 2). “We then generated maps that weight the studies by volume of mean citations and by posts/tweets,” he adds.

There are similarities in the maps, particularly for the most cited/mentioned terms, though the weighting differs in some of the smaller nodes. Yet strikingly the colour scale almost reverses for tweets compared to citations, with academic engagement (as expressed via citations) being much more focussed on Cluster 2 – Celebrity, Fame and Public Identity; whereas public engagement on X was stronger for Cluster 4 – Sex Gender and Identify. 

These clusters are also mapped in a graph to show their growth over time, from 2011-2021. Cluster 3 – feminism, policy and social change – gets the most attention, followed by Clusters 4 and 5. Although cited more by other academics, Cluster 2 — celebrity, fame and public identity. There is a noticeable leap in interest for Clusters 3 and 4 in 2021, suggesting growing public engagement with research in these two areas.

Using Altmetric Share of Voice, which uses a machine learning model, “we analysed X/tweet sentiment across the five clusters that were outlined in the first map,” says Carlos “We found that the spread of sentiment across the clusters is relatively even, though there are some points of note,” he adds. 

Clusters 1 and 2 — wealth, influence, social networks and celebrity, fame, public identity —  attract the highest positive and least negative sentiment, whereas cluster 4 — sex, gender, identity — has the least positive and Cluster 5 — mental well-being, youth, digital life has the biggest percentage of negative sentiment. (See Figure 3 for details)

“The aim of the chapter was to show how platforms such as Altmetric and Dimensions can enable digging into a field or fields of research to follow either funding, academic or public interest,” says Mike. “Considering the Kardashian name is associated with public interest and with the hyperconnected world of social media and instant information, it is particularly pertinent that such a data analysis was applied here, yet the same could be done for any area. Our research confirms observations from 400 years ago that the use of celebrity, or eye-catching subjects continues to engage the public and draw them into engaging with scholarly research. At a time when openness and public support for the scientific method is under new scrutiny, this effect should never be ignored.” 
For more information about how to use Altmetric Share of Voice, contact the Altmetric team.