Spreading information or engaging the public? The German police’s communication on Twitter

ABSTRACT

Police like other public organizations increasingly use social media for external communication. Due to their bureaucratic organization social media poses a communication challenge for them. This study analyses the content of tweets by German police using a three-category framework. Machine-learning based classification tasks are coupled with multilevel modelling to analyse all tweets distributed in 2019. The study demonstrates that police largely use Twitter to distribute information unidirectionally focusing on core tasks, while information gathering, and public relations only play a subordinate role. There is variation between accounts. However, the model only explains parts of it, thus inviting for follow-up research.

KEYWORDS:

• Germany
• police
• Twitter
• machine-learning
• communication

Supplementary material

Supplemental data for this article can be accessed at https://doi.org/10.1080/14719037.2022.2142653

Disclosure statement

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

Notes

1. It is important to keep in mind that we do not assess whether the request for information is successful, meaning whether the public also replies to it. We only focus on the intention of the police.

2. See: https://www.ard-zdf-onlinestudie.de/files/2020/0920_Beisch_Schaefer.pdf.

3. See: https://www.iwkoeln.de/fileadmin/user_upload/Studien/Kurzberichte/PDF/2019/IW-Kurzbericht_2019_Alterung_der_Gesellschaft.pdf.

4. The range of tweets per account is between 43 and 2151.

5. Of Germany’s 16 federal states three are city-states, namely Berlin, Hamburg, and Bremen. As their spatial spread differs vastly from the other states, we decided to classify these accounts as being organized at the state level, but as police entities from a metropolitan area.

6. https://github.com/koheiw/marimo.

7. In this case, precision describes the number of tweets in which a classifier and a human coder both identified the presence of a content category (true positives) divided by the number of tweets in which a classifier identified the presence of a content category (true positives + false positives). Recall is the number of tweets in which a classifier and a human coder both identified the presence of a content category (true positives) divided by the number of true positives plus the number of cases where the human coder identified the presence of content category but the algorithm failed to classify this tweet correctly (false negative). The F1 score is defined as 2x (precision*recall)/(precision+recall) (see: Pilny et al. 2019).

8. Since we did not formulate a hypothesis for the impact of organizational and environmental characteristics on the category information sharing, we do not include a multilevel model for this variable in the analysis. It is however in the appendix.

Additional information

Notes on contributors

Marc Jungblut

Dr. Marc Jungblut is a postdoctoral researcher at the Department of Media and Communication at LMU Munich. His research focuses on the role of media in conflict, computational social science, and strategic communication.

Jens Jungblut

Dr. Jens Jungblut works as an Associate Professor at the Department of Political Science at the University of Oslo. His main research interests include party politics, policy-making, and public governance in the knowledge policy domain (higher education & research), the role of (academic) expertise in policy advice, and communication of public organizations.