Unequal Tweets: Black Disadvantage is (Re)tweeted More but Discussed Less Than White Privilege

ABSTRACT

Disadvantage and privilege work together to uphold systems of inequality. Nevertheless, racial inequality is often described as Black disadvantage, while White privilege remains less visible. This one-sided framing in public discourse may result in equally one-sided understandings of and policies aimed at reducing inequality. In the present research, we examined the use of and the reactions to Black disadvantage and White privilege frames in tweets. Twitter stands out as a public sphere inspiring both online and offline political discussions and protests around racial inequality (e.g. #BlackLivesMatter). We analyzed the framing of tweets using a combination of a rule-based and a machine-learning approach, resulting in two corpora of 11,292 (Study 1) and 31,984 tweets (Study 2, a direct replication of Study 1) using comparative frames of racial inequality. Users overall more often framed inequality as Black disadvantage than as White privilege. Moreover, tweets with a disadvantage frame were more often retweeted, but less often quoted and replied to than tweets with a privilege frame. These results show that racial inequality is often one-sidedly framed in real online conversations and that this pattern may be reinforced by other users because they preferably pass on disadvantage frames. However, focusing on White privilege may provoke more discussion about racial inequality. Although effect sizes were small, these effects can impact content and perspectives in mainstream media, public opinion, and political agendas by guiding attention to certain aspects of racial inequality, but not others.

KEYWORDS:

• Black disadvantage
• White privilege
• social media
• syntax analysis
• machine learning

Acknowledgments

We are grateful to Konstanze Arnoldussen, Katharina Ruff, and Sophia Wittenborn for their help in manually coding tweets.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Datasets used for analysis and validation, containing only tweet IDs due to restrictions set by Twitter, are available from the authors at request. Other researchers with access to the Twitter API (Application Programming Interface) can use these IDs to re-create the raw dataset, unless a tweet has been deleted in the time since we collected the data for this study. Note that, since data collection, Twitter has restricted research access to the API to paid, premium accounts.

Notes

1. https://spacy.io/models/en#en_core_web_sm

2. Note that although syntax analysis alone did not show sufficient accuracy, this analysis step was necessary to create training data for the machine-learning model. A manually annotated sample from the unfiltered population of tweets would not contain enough tweets with relevant frames.

Additional information

Funding

The work was supported by the Friedrich-Ebert-Stiftung and the German Research Foundation (research grant BR 5222/5-1).

Notes on contributors

Annette Malapally

Annette Malapally is a doctoral researcher at the Chair of Social Psychology at Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU). Her research interests comprise communication and social cognition in contexts of inequality between social groups.

Andreas Blombach

Andreas Blombach is a research assistant at the Chair of Computational Corpus Linguistics at FAU. His research interests include readability of scientific writing, the spread of misinformation and conspiracy theories in social networks, and, more broadly, statistical and computational methods in sociolinguistics and the digital humanities.

Philipp Heinrich

Philipp Heinrich is a research assistant at the Chair of Computational Corpus Linguistics at FAU. Having studied mathematics, linguistics, and philosophy, his research interests revolve around statistical and computational methods for digital hermeneutics.

Julia Schnepf

Julia Schnepf is a post-doctoral researcher at FernUniversität Hagen. She has studied at Heidelberg University and completed her PhD at the University of Koblenz-Landau on linguistic framing effects across different policy domains. Her current research focuses on the framing of inequalities and violence against women in the media.

Susanne Bruckmüller

Susanne Bruckmüller is professor and chair of Social Psychology, Gender, and Diversity at FAU. Her research centers on social cognition and communication, with a special focus on their intersection with stereotypes, intergroup relations, and social and economic inequality.