Principal investigator: Károly Takács

The Computational Social Science – Research Center for Educational and Network Studies Research Group (CSS RECENS) was established on October 1, 2018 – by uniting the Group of Network Analysis of Institute for Sociology and Lendület RECENS Research Group which was finalized after a successful operation –, as an independent research group which does not belong to any institutes of CSS. The CSS RECENS Research Group is the home of an ERC Consolidator Grant research (EVILTONGUE, PI: Károly Takács), a finalized Lendület research (PI: Károly Takács) and a HAS Premium Postdoctoral Research grant (Zoltán Kmetty). In addition, the members of the Research Group conduct five NKFIH researches (Szabolcs Számadó, Tamás Keller, Bence Ságvári, Zoltán Kmetty and Dorottya Kisfalusi) and several small researches and they participate in national (Hungarian Telekom, smaller IT and data analysis companies) and international collaboration (e.g. European Social Survey and receiving guest researchers: Michael T. Gastner, Tanja Slisković).

By 2018, a network of interdisciplinary exchanges and collaborations on research in the field of computational social science was established with the involvement of the leaders of CSS. The first results of the co-operation, based on a diverse scientific disciplinary background, are published in the special edition of the journal Hungarian Science. The Research Group also played an active role in engaging with relevant stakeholders in business life, participating in business conferences, and initiating meetups. CSS-RECENS is also well-positioned in the international leader board of computational social science, which promises significant new research results; for example, through the creation of joint European tenders and an edited volume based on contributions from world-leading authors.

The ERC funded EVILTONGUE research project (eviltongue.tk.mta.hu) examines the role of rumour and reputation in maintaining co-operation and community standards in different contexts. The research can be divided into several subprojects. Agent-based simulations help to understand how third-party evaluations and third-party information help promote collaboration. We have shown that reliance on information from closed-link triads is a more successful strategy than indirect reciprocity based on open-ended network. Our simulation studies also show that the existence of information does not necessarily point to communally beneficial results, if the sincerity of communication is not ensured. In a number of contexts, we collected data in 2018 and analysed the data collected in previous years. Under laboratory experimental conditions, we have shown that people really build on evaluation information from others and that the reputation created from this is a prerequisite of cooperative behaviour. Both the experimental and the longitudinal scholastic data based on relation networks, which recorded earlier, show that many times the deviants are the target of the negative gossip. From our network-based questionnaire survey in organizations, it has emerged that jealousy and frustration by the relative payment conditions which are deemed unfair in this context may be the driving force behind the gossip. In 2018, for the first time in the world, we were gathering data with smart watches in a dormitory corridor, then among festival volunteers. Field work with this innovative method was supplemented by data collection using traditional methods. In order to do this, the Research Group has developed smart watch applications and developed a new research methodology that was audited by the Hungarian National Authority for Data Protection and Freedom of Information (NAIH) for ethical and data protection purposes. In addition, we have created, published and annotated a spontaneous and quantitatively unique Hungarian speech corpus from a large quantity of sound recordings from a closed community that will not only be used for analysing gossip and reputation, but may also be a valuable source of analysis of the spoken Hungarian language. In the framework of the project in May 2018, we organized a prestigious international conference called Gossip, Reputation, and Honesty.

The research supported by the Lendület Program examines the mechanisms and dynamics of negative social relationships and networks, especially in relation to competitive situations. This year, we continued to organize and analyse the data that we collected before from the primary school network panel research. We have also evaluated our field experiments based on our previous network results, in which we wanted to stimulate motivation and decision-making in a secondary school with the help of a session in which the opinion-forming students participated. According to our results, the intervention was successful, especially for those who had further study aspirations but were uncertain about their plans. Further findings from primary school research highlighted that desk-mates also have a significant impact on school performance. As a result, another field experiment was carried out to investigate the effect of the desk-mates. In accordance with the expectations of a randomized international field experiment, following the strictest scientific transparency guidelines, we registered our research plans in the American Economic Association's RCT record system. Within the framework of a Lendület funded project, a political elite research is also taking place, in which we try to reconstruct the network of relations between the senior officials of the period between 1948 and 1956 and its relations with the power struggle. The research expanded into new directions in 2018 using the computational social science toolkit e.g. the analytical analysis of the records of the leading bodies. During the processing stage, the personalized names were fixed in the digitized archive material and with an independently developed program identified than tagged, followed by the use of tagged files, and based on it, a primary text-based interpersonal network was created. In order to identify the members and the careers of the elite sample, the prosopographic data collection continued, while an analysis of the individual biographical attributes was started as to whether they could be interpreted as the basis of a historical relationship network. Based on experiences, the focus of the project has been expanded, identifying the links to possible national and international research to ensure a broader interpretation of the results.

In the research group, the coordination of small groups specialized in research tasks takes place in weekly meetings. We also discussed our own research results on the weekly reading circles, in addition to the processing of topical literature closely related to the themes of the research group.