Publications:
Associations between micro- and macro level social network properties and individual productivity in virtual collaboration
Deng Dongning and Júlia Koltai
https://www.nature.com/articles/s41598-025-09309-z
Although the connection between social network properties (SNPs) and team productivity has been studied extensively, there is still room to deepen our understanding, particularly regarding individual-level dynamics, the non-linear nature of these relationships, and the interactions between individual and structural factors. To do this, we analysed 58 Open Source Software Development (OSSD) projects, using a comprehensive set of SNPs and measuring individual productivity by code editing contributions. Our findings reveal that SNPs have significant and complex dynamics in their associations with individual productivity. Highly productive individuals present SNP traits with a moderate number of connections, being indirectly connected but having influential peers, and being in a decentralised yet locally cohesive environment. Centralised team structure with direct connection with central nodes or influential clusters benefits individual productivity, especially for those who are peripheral or have powerful peers. The highly productive members in the influential clusters also form and reinforce “coordination chambers”. Low individual productivity or even the free riding phenomenon may be more prevalent in a highly closed local and global environment. This is especially true when the structure is not diverse. Taking on a brokerage role with access to diverse knowledge is generally key to active participation, especially when connections are non-redundant. However, productivity may suffer when individuals become too embedded in the bridging role. To minimise the cost of such brokerage role, how and where to be a broker matters. One can become active either with unique ties in networks with centralised bridging brokerage, decentralised accessibility, or clustered structure, or bridging disconnected groups in less clustered but locally cohesive networks with evenly-distributed influence. Our analytical framework shows how non-linear and contextual interaction dynamics can be uncovered using social network and statistical methods. The findings inform not only how open-source workspaces can be better structured according to governance goals, but also potential inequalities in OSSD teams and a possible approach for more open and inclusive team structures.
Classifying social position with social media behavioral data
Júlia Koltai, Zsófia Rakovics, Zoltán Kmetty, Kata Számel, Borbála Ungvári, Bendegúz Váradi, Ákos Huszár
https://epjdatascience.springeropen.com/articles/10.1140/epjds/s13688-025-00578-2
The main question of our study is how far social position can be predicted solely based on digital behavior. The phenomenon that offline inequalities are reflected in the digital space has been heavily researched since the digital revolution. Nevertheless, there are few data, which both measure social inequalities and digital behavior: scientists either have information on the social status of people, or on their observed digital behavior, but not on both. When analyzing digital behavioral data, however large scale it is, information on the social position of the users is hardly available. In the current paper, we analyze a special dataset collected with a data donation technique, which contains information on both the social position and the observed digital behavior of participants, and which is representative for the internet user population of Hungary. In the analysis, using diverse models, we explored how well basic indicators measuring digital behavior on Facebook can classify users’ social class measured by the 5-category version of the European Socio-economic Classification (ESeC). The results show that based on basic quantitative indicators of digital behavior and usage the models cannot classify users’ social position with a high degree neither in the classification of social class, nor in the case of socio-economic status. Nevertheless, the inclusion of socio-demographic characteristics as features increased the predictive power of the models, that could differentiate between the lowest and highest social position with a high degree. The models based on purely observed digital behavior could identify those in the lowest social position with the highest performance. Among those features, that played an important role in this classification, usage time, frequency network size and language characteristics (especially the diversity of the used language and punctuation) should be highlighted, while diverse Facebook activities and detected interest categories also played a role. These results are in line with the results of previous studies derived from smaller-scale, non-representative, or self-reported survey-based data on the same topic.
Determinants of willingness to donate data from social media platforms
Zoltán Kmetty, Ádám Stefkovics, Júlia Számely, Deng Dongning, Anikó Keller, Edit Pauló, Elisa Omodei and Júlia Koltai
https://doi.org/10.1080/1369118X.2024.2340995
Social media data donation through data download packages (DDPs) is a promising new way of collecting individual-level digital trace data with informed consent. Nevertheless, given the novelty of this approach, little is known about whether and how people would share their data with researchers, although this could seriously affect selection bias and thus, the outer validity of the results. To study the determinants of data-sharing and help future data donation studies with detecting the conditions, under which the willingness is the highest, we pre-registered two vignette experiments and embedded them in two online surveys conducted in Hungary and the US. In hypothetical requests for donating social media data via DDPs, we manipulated the amount of the monetary incentives (1), the presence or lack of non-monetary incentives (2), the number of requested platforms (3), the estimated upload/download time (4), and the type of requested data (5). The results revealed that data-sharing attitude is strongly subject to the parameters of the actual study, how the request is framed, and some respondent characteristics. Monetary incentives increased willingness to participate in both countries, while other effects were not consistent between the two countries.
Do diversity and context collapse kill an online social network?
Júlia Koltai, László Lőrincz, Johannes Wachs and Károly Takács
https://link.springer.com/article/10.1007/s41109-025-00719-6
Our social lives consist of various circles, such as family, friends, and colleagues. Differences in norms and expectations among these circles can create tension in large online social networks (OSNs) due to blurred boundaries. It is unclear whether this phenomenon, known as context collapse, outweighs the convenience of having diverse communities in one place for OSN users. To explore this trade-off, we examined if ego network characteristics indicating context collapse could explain users’ decisions to leave iWiW, a defunct Hungarian OSN with over 3.5 million active users at its peak. We assessed context collapse based on two conditions: the absence of overlapping communities measured by network modularity and social differences between those communities. We find that users with fragmented social networks indeed leave the platform earlier if these distinct communities differ significantly in their age profiles and urban-rural composition. However, the highest probability of leaving was among those with non-fragmented networks and similar communities. These seemingly contradicting results are caused by the process that network fragmentation itself decreases the probability of leaving. Thus, our results demonstrate simultaneously how brokerage can be valuable and context collapse stressful for users of OSNs.
Social Class and Digital Divide: Analyzing Digital Inequalities on Census and Digital Behavioral Data
Kata Számel, Ákos Huszár, Michelle Horváth, Szilvia Rudas, Zsófia Rakovics and Júlia Koltai
https://doi.org/10.1177/23780231251357646
The authors analyze the relationship between social class and the various levels of digital inequality, such as Internet access, inequality in digital skills, and online behavior. Two unique databases from Hungary are used. First, digital activity and digital skills are studied on a 10 percent random sample from the 2022 Hungarian census. Second, to analyze online behavior, a donation-based digital behavioral dataset is used, which is linked with a representative survey of the Hungarian Internet user population. There has been a notable decline in research activity regarding the first level of the digital divide, but these results indicate that the access level of the digital divide persists in Hungary, particularly among unskilled workers. In addition to the anticipated finding that individuals belonging to higher social classes are more likely to engage in digital activities and possess advanced digital skills, the analyses revealed that social class also influences social media use. Although a considerable proportion of the unskilled workforce lacks Internet access, those who are connected demonstrate the highest levels of digital engagement, posting the most often and spending the most time on Facebook. However, on average, the upper classes write longer posts and comments.
Ákos Máté, Zsófia Rakovics, Szilvia Rudas, Levente Wallis, Bence Ságvári, Ákos Huszár, Júlia Koltai
https://doi.org/10.3390/s23094571
The main question of this article is what factors influence willingness to participate in a smartphone application-based data collection where participants both fill out a questionnaire and let the app collect data about their smartphone use. Passive digital data collection is becoming more common, but it is still a new form of data collection. Due to the novelty factor, it is important to investigate how willingness to participate in such studies is influenced by both socio-economic variables and smartphone usage patterns. We estimate multilevel models based on a survey experiment with vignettes for different characteristics of data collection (e.g. different incentives, duration of the study). Our results show that of the socio-demographic variables, age has the largest effect, with younger age groups having a higher willingness to participate than older ones. Similarly, smartphone use also has an effect on participation, with advanced users more likely to participate than the social media user group. We also found that users who only use the basic functions of their device are less likely to participate. Finally, the interaction terms between levels showed that the circumstances of data collection matter differently for different social groups. These results provide important guidance on how to fine-tune circumstances to improve participation rates in this novel passive digital data collection.
In progress:
Cross-over social processes from real life to digital realms - Exploring gender differences and inequalities in digital trust
Júlia Koltai, Éva Fodor, Zsófia Rakovics, Szilvia Rudas
We explore how gender mediates the link between being trusting offline, in the “real world” and online, in the “digital world” to initiate a conversation on two areas: 1) the cross-over social processes from real to digital; and 2) gender differences and gender inequalities in digital trust and its conditionalities.
Digital Behavioral Patterns of Social Classes First Results of an Innovative Donation Based Digital Data Collection
Ákos Huszár, Zsófia Rakovics, Kata Számel, Szilvia Rudas, Bendegúz Váradi, Borbála Ungvári, Gergely Agócs, Júlia Koltai
Our paper aims to investigate the influence of social class on individuals' digital behavior in Hungary. Specifically, we will analyze the differences in observed behavior on Facebook among different social classes using innovative data collection methods.
Prestige-based occupational homophily and societal openness
Sandeep Chowdhary, Dongning Deng, Federico Battiston, and Júlia Koltai
Homophily is the tendency of individuals to associate with others who are similar to them in some way, such as sharing similar backgrounds, interests, values, or demographic characteristics. Occupational homophily in families, in particular, refers to the inclination to choose careers that are similar to those of their parents or other family members. Research has shown that homophily of occupations within families is a common phenomenon. For example, studies have found that children of doctors are more likely to become doctors themselves, and that the children of lawyers are more likely to enter the legal profession. This phenomenon can have a significant impact on societal openness, particularly within families. In this work, we quantify occupational prestige homophily and its impact on societal openness, particularly within families. We utilized the Position Generator dataset (International Social Survey Programme 2017) to construct occupational networks in 30 countries . While the structure of occupational networks varied among countries, low prestige occupations were consistently segregated. Furthermore, highly developed countries were found to exhibit greater prestige-based segregation within families, suggesting a less open social structure. The results suggest that developing countries are more open to diverse career paths within families, while developed countries are less open. Taken together, our work sheds light on the complex relationship between occupations, social networks, societal development, and societal openness.
Keywords: homophily, occupations, networks, prestige, position generator
Conferences:
2022
Social Differences in Willingness of Participation in Donation Based Data Collection.
Presentation at the Conference on Complex Systems 2022, Palma de Mallorca, Spain. (co-authors: Zoltán Kmetty)
2023
Social Structure and Inequalities Through the Lens of Digital Data.
Presentation at the Budapest Winter Workshop organized by the Center for Collective Learning & ANET & NETI Lab. Corvinus University and HUN-REN, Budapest, Hungary.
Deadly serious: The effect of the belief in conspiracy theories on anti-vaccination in Hungary and the moderating role of experience.
Presentation as an invited speaker at the workshop 'From Stigmatized to Canonized Knowledge? Exploring Top-Down Conspiracy Theories in Central Eastern Europe and Beyond', Central European University, Budapest, Hungary. (co-authors: Ádám Stefkovics and Péter Krekó)
A társadalomtudományok helyzete és lehetőségei – módszertani kihívások és esélyek a digitális változások kontextusában. esélyek a digitális változások kontextusában [The state and opportunities of the social sciences - methodological challenges and opportunities in the context of digital change].
Invited participant in a roundtable discussion at the 13th Sociology Days of the Babes-Bolyai University - between the challenges of continuous change and adaptation in our societies. Babes-Bolyai University, Cluj-Napoca, Romania. Bolyai University.
Az összeesküvéses hiedelmek hatása az oltási hajlandóságra a tapasztalatok hatásának tükrében [The impact of conspiracy beliefs on vaccination uptake in the light of the impact of experience].
Presentation at the XXXth National Scientific Assembly of the Hungarian Psychological Society, Pécs, Hungary. (co-authors: Ádám Stefkovics and Péter Krekó)
2024
Társadalmi struktúra és egyenlőtlenség a digitális adatok tükrében.
Júlia Koltai (keynote presentation)
VII. Változó világ, változó társadalom konferencia. ELTE Pedagógiai és Pszichológiai Kar, Budapest, Magyarország.
Digital Behavioral Patterns of Social Classes First Results of an Innovative Donation Based Digital Data Collection
Ákos Huszár, Zsófia Rakovics, , Kata Számel, Szilvia Rudas, Borbála Zoé Ungvári, Bendegúz Váradi, Julia Koltai
International Sociological Association, Research Committee 28 (Structure and Inequality) Annual Summer Meeting 2024. Brown University, Providence, Rhode Island, USA.
Digital Behavioral Patterns of Social Classes.
Julia Koltai, Zsófia Rakovics, Ákos Huszár, Kata Számel, Szilvia Rudas, Borbála Zoé Ungvári, Bendegúz Váradi
International Conference on Computational Social Science 2024. University of Pennsylvania, Philadelphia, USA.
2025
Understanding social groups through multiplatform social media behavioural data
Michelle Horváth, Kata Számel (joint presentation)
Budapest Winter Workshop, Corvinus University Budapest
Combining digital location data with traditional survey data to study spatial patterns of cultural consumption along social classes
Kata Számel (előadó), Anna Sára Ligeti, Michelle Horváth
11th Conference of the European Survey Research Association, Utrecht, The Netherlands
Reproduction of Social Inequalities in the Digital Space – POSTER
Ákos Huszár, Zsófia Rakovics, Michelle Anna Horváth, Anna Sára Ligeti, Szilvia Rudas, Kata Viola Számel, Bendegúz Váradi, Júlia Koltai
11th International Conference on Computational Social Science, Norrköping, Sweden