Research Article: Using the Dirichlet process to form clusters of people’s concerns in the context of future party identification

Date Published: March 4, 2019

Publisher: Public Library of Science

Author(s): Patrick Meyer, Fenja M. Schophaus, Thomas Glassen, Jasmin Riedl, Julia M. Rohrer, Gert G. Wagner, Timo von Oertzen, Floriana Gargiulo.


Connections between interindividual differences and people’s behavior has been widely researched in various contexts, often by using top-down group comparisons to explain interindividual differences. In contrast, in this study, we apply a bottom-up approach in which we identify meaningful clusters in people’s concerns about various areas of life (e.g., their own health, their financial situation, the environment). We apply a novel method, Dirichlet clustering, to large-scale longitudinal data from the German Socioeconomic Panel Study (SOEP) to investigate whether concerns of people living in Germany evaluated in 2010 (t0) cluster participants into robust and separable groups, and whether these groups vary regarding their party identification in 2017 (t0 + 7). Clustering results suggest a range of different groups with specific concern patterns. Some of these notably specific patterns of concerns indicate links to party identification. In particular, some patterns show an increased identification with smaller parties as the ‘Bündnis 90/Die Grünen’ (‘Greens’), the left wing party ‘Die Linke’ (‘The Left’) or the right-wing party ‘Alternative für Deutschland’ (‘Alternative for Germany’, AfD). Considering that we identify as many as 37 clusters in total, among them at least six with clearly different party identification, it can also be concluded that the complexity of political concerns may be larger than has been assumed before.

Partial Text

Fig 2 shows the distribution of concerns and party affiliations in the full sample before the application of the clustering. A total of 8170 participants answered the questions regarding their concerns in 2010 as well as the questions about their party affiliation in 2017. The grey area on the right under ‘Employment Safety’ represents the number of people without current employment. Note that this also includes students, retirees etc. The section ‘Other Concerns’ represents how many people named at least one other concern in the open response category. The full sample shows a high number of respondents with no party identification (52.73%) and—compared to the other parties—a higher identification with the grand people’s parties (CDU/CSU = 18.64% and SPD = 14.93%).

The aim of this study was to investigate whether there are robust sets of participants with specific concern patterns in the German population. Furthermore, we investigated whether these sets can be linked to specific party identification. Using a Dirichlet clustering algorithm, the results indicate that there are in fact robust and clearly separable clusters. This robustness was conclusively shown by the small partition difference between the results from two independent runs with a sufficiently high number of samples. Therefore, it can be concluded that this method yields reliable and reproducible results.

From the current study it can be concluded that (1) Dirichlet clustering seems to be a helpful tool to find clusters in social groups, (2) there are clearly discernable profiles of concerns in the German population which vary strongly between clusters but are very stable within clusters, (3) these profiles displace structures which coincide with specific party identification, even though the concern data was collected seven years prior to the party identification, and (4) that this overlap of clusters and later party identification was particularly strong for the Green party and the AfD party, even though the AfD did not even exist at the point of time where the concern data was collected. Our results suggest that it is possible to also predict future party identification using the concerns of the population in the present, and to identify regarding what concerns isolated subgroups in the population drift away from the majority. Considering that we identified as many as 37 clusters in total, among them at least six with clearly different party identification, it can also be concluded that the complexity of political concerns may be larger than has been assumed before.




Leave a Reply

Your email address will not be published.