Latent Dirichlet allocation (LDA) is a popular method for analyzing large text corpora, but it suffers from instability due to its reliance on random initialization. This results in different outcomes for replicated runs, hindering reproducibility. To address this, we introduce LDAPrototype, a new approach for selecting the most representative LDA run from multiple replications on the same dataset. LDAPrototype enhances the reliability of LDA conclusions by ensuring greater similarity between replications compared to traditional LDA runs or models chosen based on perplexity or NPMI. A key feature of LDAPrototype is its use of a novel model similarity measure called S-CLOP (Similarity of multiple sets by Clustering with LOcal Pruning). It is based on topic similarities, for which we compare the usage of measures like the thresholded Jaccard coefficient, cosine similarity, Jensen-Shannon divergence, and rank-biased overlap. The effectiveness of LDAPrototype is demonstrated through its application to six real datasets, including newspaper articles and tweets. The results show improved reproducibility and reliability in topic modeling outcomes. LDAPrototype’s approach is noteworthy for its practical applicability, comprehensibility, ease of implementation, and computational efficiency. Furthermore, the algorithm’s concept can be generalized to other topic modeling procedures that characterize topics through word distributions, making it a versatile tool in text data analysis.
The insular usage of media sources is a common tool for populist actors to position themselves in relation to the social mainstream and to hold up their claim of creating a counter public. In this article, we traced these dynamics by analysing differences in the insularity of right-wing populist parties’ source repertoires compared to other parties, i.e., the exclusivity of the sources used by populist politicians. By analysing the linking practices of parliamentarians in seven EU countries and Switzerland on Twitter we were able to test the impact of political and media context on these practices. In a second step, we asked to what extent the insularity of sources allows us to draw conclusions on the insularity of topics, i.e., whether right-wing populists used a broad or narrow topic agenda. Our key findings: In countries with high levels of media trust and low levels of polarization, the populist source repertoire features a comparatively low insularity, except for politically marginalized populist parties that use more insular sources. Regardless of the insularity of source use, right-wing populist parties in almost all countries feature a high degree of topic insularity, making a narrow topic agenda the cross-national characteristic of right-wing populist communication.
Zusammenfassung: Die vorliegende Studie betrachtet die Praxis des Teilens von Texten durch politische Akteure auf Twitter und analysiert den thematischen Geltungsanspruch, der sich darin spiegelt: zugespitzt populistisch oder vielfältig und inklusiv im Stil einer Volkspartei. Konkret wurde das ‚Verlinkungsverhalten‘ deutscher Bundestagsabgeordneter auf Twitter untersucht – mit besonderem Augenmerk auf der rechtspopulistischen AfD. Hierzu wurden mit Hilfe eines adaptiven Scrapers 56.107 Texte gesammelt, die Bundestagsabgeordnete in der laufenden Legislaturperiode in Tweets geteilt hatten. Der Analyse ihrer Vielfalt diente eine Maßzahl, die thematische Heterogenität auf Basis von Topic-Modellen ermittelt. Die Resultate zeigen, dass populistische Akteure zwar häufiger als andere auf Medien verlinken und dabei heterogene Quellen nutzen, dass sich dies jedoch nicht in thematische Vielfalt übersetzt, im Gegenteil: Die Themenagenda der AfD-Parlamentarier ist weniger vielfältig als die der Politiker anderer Parteien. Die klar populistische Themenselektion korreliert zudem stärker mit metrischem Erfolg (in Form von Likes, Retweets), als dies bei den Themen anderer Parteien der Fall ist.