Users and inferred data in online social networks: Countering power imbalance by revealing inference mechanisms

Research output: Unpublished contribution to conferenceUnpublished abstract


In the past, much privacy research has focused on how social media use and social relationships are interrelated. Lately, more attention is given to the access and the use of personal data by Online Social Network (OSN) providers and other third parties. Here, data mining algorithms, machine learning techniques or other data extraction techniques play an essential role in creating meaningful information for understanding and predicting personal information of the user. This leads to a risk of disempowerment through the loss of user agency. Our research investigates how we could counter this data power imbalance, by confronting social groups and users with the way that their data is being collected, processed and inferred. From a theoretical perspective we build on the integration of Science and Technology Studies (STS) with Media and Communication Studies (MCS) (Gillespie et al., 2014), more in particularly taking a critical stance on the co- construction of technological systems (van Dijck, 2013; Mansell, 2012; Feenberg, 1999).
In the paper we present the results of an in-depth user study within the interdisciplinary EU project USEMP ( The study took place in Flanders (Belgium), in November and December 2014. Our findings discuss people's awareness and attitudes towards the way OSN providers and specific third parties can reason on their social media data and related inferences. Through means of 14 semi-structured qualitative interviews using a diverse and innovative set of probes, we captured insights on which personal data people generally find appropriate to share online and their attitudes towards the different ways of data gathering (volunteered, observed and inferred). Later on, we confronted our results with the data-reachability matrix (Creese et al., 2012) wherein the authors define which potential personal information can be inferred through the use of existing data extraction techniques on (a combination of) data, typically exposed on OSNs. Starting from these insights we analyze the need for and the possibility of an end-user visualization of personal data sharing behavior.
Original languageEnglish
Publication statusPublished - 23 Jun 2015
EventData Power Conference - Session 4 ‘Personal data and data literacy’ at - University of Sheffield, Sheffield, United Kingdom
Duration: 22 Jun 201523 Jun 2015


ConferenceData Power Conference - Session 4 ‘Personal data and data literacy’ at
CountryUnited Kingdom

Bibliographical note

Session chair: Joseph Turow


  • data
  • online social networks
  • privacy
  • user studies

Fingerprint Dive into the research topics of 'Users and inferred data in online social networks: Countering power imbalance by revealing inference mechanisms'. Together they form a unique fingerprint.

Cite this