Search

The psychological need to share information and impact thereof has been exacerbated with the advent of online activities and online social media. The pervasiveness of unfiltered and non-curated information in online social media induces exposure to untrustworthy information (i.e., low-factual information sources). This exposure makes it increasingly challenging for users to discern trustworthy from non-trustworthy online news sources. Moreover, the social character of these platforms raises questions around the diffusion of untrustworthy news sources. Understanding characteristics of users sharing untrustworthy online news sources could help identify and prevent the diffusion of untrustworthy content. Past work has mainly focused on the experimental investigation of specific individual differences in the interaction with untrustworthy online news sources. To comprehensively analyse the human factors related to the online dissemination of non-trustworthy news, a better understanding of user’s characteristics in the complexity of a real-world setting is called for. Here, we aim to answer the following research question: What are systematic differences in users’ demographic, psycholinguistic, and online posting behaviour regarding interaction with (un)trustworthy news sources? To address this question, we use open-access social media data from the Reddit platform (Pushshift) and independent data for political bias and trustworthiness of online media (e.g., media bias fact check) to highlight groups of users who tend to share less factual content to those platforms. We assign users into two groups: trustworthy news sharers and untrustworthy news sharers. For group assignment, we adopt communities’ embeddings to ensures that users in the two groups belong to similar communities. For predictive modelling, we extract three types of features: psycholinguistic, online posting behaviour and demographic. First, we quantify psycholinguistic characteristics by calculating the Linguistic Inquiry and Word Count scores (LIWC) using the text from the comments by users commenting a post containing a news article. Furthermore, we complement LIWC scores with general text-based characteristics such as lexical diversity and readability scores. Secondly, we extract online posting behaviour features based on the popularity of the user, posting frequency, and network-size. Lastly, we use the Bidirectional Encoder Representations from Transformers (BERT) model trained on the RedDust dataset to infer users’ demographic characteristics such as age and gender. We use the combined set of features in a machine learning approach (e.g., random forest) to predict group membership according to Redditors characteristics. Furthermore, to gain insights on which of these factors has a predominant impact on group membership we perform Shapely Values analysis to extract and interpret feature importance. Insights from this project will contribute to a more nuanced understanding of how users' characteristics can be associated to different consumption patterns of (un)trustworthy news sources, and potentially provide insights for predicting the process of misinformation spread. We anticipate that our findings can be used to identify groups who are more susceptible to consuming and spreading disinformation. This in turn could help individuals make informed decisions and avoid exposure to false information. Finally, we will discuss the implications of our findings for the development of interventions and policies to debunk disinformation and to increase media literacy.

The psychological need to share information and impact thereof has been exacerbated with the advent of online activities and online social media. The pervasiveness of unfiltered and non-curated information in online social media induces exposure to untrustworthy information (i.e., low-factual information sources). This exposure makes it increasingly challenging for users to discern trustworthy from non-trustworthy online news sources. Moreover, the social character of these platforms raises questions around the diffusion of untrustworthy news sources. Understanding characteristics of users sharing untrustworthy online news sources could help identify and prevent the diffusion of untrustworthy content. Past work has mainly focused on the experimental investigation of specific individual differences in the interaction with untrustworthy online news sources. To comprehensively analyse the human factors related to the online dissemination of non-trustworthy news, a better understanding of user’s characteristics in the complexity of a real-world setting is called for. Here, we aim to answer the following research question: What are systematic differences in users’ demographic, psycholinguistic, and online posting behaviour regarding interaction with (un)trustworthy news sources? To address this question, we use open-access social media data from the Reddit platform (Pushshift) and independent data for political bias and trustworthiness of online media (e.g., media bias fact check) to highlight groups of users who tend to share less factual content to those platforms. We assign users into two groups: trustworthy news sharers and untrustworthy news sharers. For group assignment, we adopt communities’ embeddings to ensures that users in the two groups belong to similar communities. For predictive modelling, we extract three types of features: psycholinguistic, online posting behaviour and demographic. First, we quantify psycholinguistic characteristics by calculating the Linguistic Inquiry and Word Count scores (LIWC) using the text from the comments by users commenting a post containing a news article. Furthermore, we complement LIWC scores with general text-based characteristics such as lexical diversity and readability scores. Secondly, we extract online posting behaviour features based on the popularity of the user, posting frequency, and network-size. Lastly, we use the Bidirectional Encoder Representations from Transformers (BERT) model trained on the RedDust dataset to infer users’ demographic characteristics such as age and gender. We use the combined set of features in a machine learning approach (e.g., random forest) to predict group membership according to Redditors characteristics. Furthermore, to gain insights on which of these factors has a predominant impact on group membership we perform Shapely Values analysis to extract and interpret feature importance. Insights from this project will contribute to a more nuanced understanding of how users' characteristics can be associated to different consumption patterns of (un)trustworthy news sources, and potentially provide insights for predicting the process of misinformation spread. We anticipate that our findings can be used to identify groups who are more susceptible to consuming and spreading disinformation. This in turn could help individuals make informed decisions and avoid exposure to false information. Finally, we will discuss the implications of our findings for the development of interventions and policies to debunk disinformation and to increase media literacy.

Measures of latent psychological constructs have changed over the course of millennia. While psychological assessments have incrementally changed in the last 100-years, Psychology is currently facing an epochal transition into computational psychology that will require major shifts in education and knowledge creation. In this keynote I will discuss how big data and machine learning continues to change psychometrics and psychology more broadly. I will discuss how the development of more objective and data-driven assessment tools will help to understand human behavior and how we are now entering a new area in psychological science – the extension of the discipline to the systematic investigation of non-human intelligences. This new era more than ever requires a radical open science approach to allow research to be replicable and robust. I will conclude with an outlook perspective on challenges and most importantly, opportunities.

Pre-reg Workshop - let's do it!

This study investigates the expression of the traditional concept of sensation seeking in the everyday-usage of smartphones. Data used in this study is part of the PhoneStudy Project.

This study investigates the influence of students' daily experience at the university on their retention. Data used in this study was collected via a mobile app by a corporate partner. Through the app use data we gain insights to students behavior on campus and their interactions with others.

We aim at predicting personality traits from natural music listening behavior logged from smartphones. This research project is part of the PhoneStudy project and re-uses data from those studies.

Project repository for the article of the same name accepted for the special issue on "Behavioral Personality Science in the Age of Big Data" at the European Journal of Personality

This project contains materials, data and code for the study "Predicting Personality from Patterns of Behavior Collected with Smartphones", published in PNAS in 2020. This repository contains all the necessary data, syntax and other materials to reproduce the results from the variable-extraction point onwards. The Markdown script "paper.Rmd" knits the corresponding manuscript into a pdf file. Despite our strong commitment to open science practices, the raw log data cannot be made available openly, due to unsolved privacy implications of geolocation data and issues arising from the combination of other behavioral indicators (e.g., communication, app-usage, music consumption etc.). Below you also find a link to an interactive website that allows for the exploration of additional results for and provides resources to enable a better understanding of the used machine learning and data analytical tools.

This project serves as the collection of all research, teaching, and development activities, conducted with and around the PhoneStudy Android application. The aim of the PhoneStudy project is the development of smartphone applications, a server infrastructure, and adjacent data-tools for both the efficient collection and analyses of behavioral data in psychology and other sciences.

OSF repository of an introductory one-day workshop presented by Clemens Stachl (clemens.stachl@unisg.ch), Ramona Schödel (Ramona.Schoedel@psy.lmu.de), and Florian Pargent (Florian.Pargent@psy.lmu.de) to audiences from Psychology and other Social Sciences. The materials in this repository will be periodically updated to the latest version of the workshop.

interdisciplinary project about automatic personalization in automotive context

Coping with Corona (CoCo): Understanding individual differences in well-being during the COVID-19 pandemic The COVID-19 pandemic not only poses major threats to people’s physical health but also to their well-being. To manage the pandemic, governments around the world have mandated policies that restrict people’s activities and limit social contact with others. Importantly, individuals differ in how they cope with these social challenges psychologically. The Coping with Corona (CoCo) Project aims to predict (build up theoretically derived and data-driven prediction models) and understand (unpack mediating situation selection, interpersonal perception, and emotional co-regulation processes) individual differences in well-being during the pandemic. These goals will be pursued by a comprehensive mega-analysis of existing data and large-scale international smartphone-based assessments (including experience sampling of everyday feelings and thoughts, mobile sensing of everyday social behaviors and contexts, and personalized feedback). The CoCo team integrates interdisciplinary expertise of leading researchers in the fields of personality, relationship, developmental, network, and behavioral data science. This project will generate insights of empirical, theoretical and practical importance. Specifically, results provide a rich description of how different people cope with pandemics, and novel insights on person-environment transactions that explain these differences. This will constitute a sorely needed step towards the large-scale availability of personalized tools to foster psychological resilience in individuals, groups, and societies in the wake of pandemics.