Special Event on Big Data
Prof. Dr. Sabine Walper is Professor of Education and Youth Research at the Ludwig-Maximilians-University in Munich and Research Director at the German Youth Institute (DJI) in Munich, Germany. She received her academic training in Psychology and Educational Science at the Heinrich-Heine University Duesseldorf, the Technical University Berlin (Germany), and at the University of California at Berkeley (USA).
Her research addresses a variety of family issues and their links to child well-being, in particular developmental conditions of children and youth in separated and stepfamilies as well as families in poverty and other risk conditions. She is Co-Principle Investigator of the German Family Panel pairfam, a large interdisciplinary multi-cohort and multi-actor longitudinal study on intimate relationships and family life in Germany (http://www.pairfam.de/en/). Since 2008/2009 pairfam provides data from annual assessments to the scientific community.
As Research Director of the German Youth Institute, Sabine Walper is also involved in large replicative survey studies on families, children and youth aiming to provide insight in changing developmental conditions for young people in Germany (e.g., AID:A – Growing Up in Germany), including studies targeting early childhood and governmental programs of Early Prevention (KiD 0-3). Sabine Walper is member of the Scientific Council for Family Issues of the Federal Ministry for Family, Seniors, Women and Youth. She is Chair of the Scientific Council at the State Institute for Early Education, a member of the Commission for the Rights of Children at the Family Law Court of Germany and President of the German League for the Child. She serves on several editorial and review boards.
The Potential of Big Data in Developmental Research
Special even organized by Sabine Walper
Big Data have gained increasing relevance for research in many domains of behavioral research including developmental psychology. This session will present examples of two kinds of big data relevant for developmental research. First, we will present large-scale longitudinal studies which provide a broad, open-source data base for interested researchers. Second, we will illustrate the use of Big Data generated in the course of online-activities, be it in web-based social networks or when using mobile devices. The first three papers focus large-scale longitudinal studies and describe their potential for analyzing a broad range of developmental issues. Given the strong tradition of cohort studies in the UK, the first presentation (by Praveetha Patalay) gives an overview of four British cohort studies and describes the diverse methods used to inform about the developmental trajectories and life course of their participants. The second paper (by Sabine Walper) illustrates the research potential of the German Family Panel pairfam which comprises adolescent and adult cohorts, but also includes the target participants partner and children. Both presenters will describe how to access and use the data. The third paper (by Katariina Salmela-Aro) introduces a currently planned European cohort study. While being still at a preliminary stage this study may be a future source of research data. The last two papers focus online-generated Big Data. Paul Plener presents research on social reinforcement in web-based social networks and illustrates their role with findings on adolescents’ non-suicidal self-injuries. Finally, Clemens Stachl informs about smartphone-sensing as a novel method of data collection which gives access to large-scale behavioral and situational data. While most of these data are non-reactive and generated in the course of online-activities, they may also be complemented by self-report data assessed with the online-devices. The discussion will leave room for questions from the audience.
The British birth cohort studies
University College London, UK
The UK has the longest running series of national birth cohort studies. In this presentation I will present an overview of four UK birth cohorts: the 1946, 1958, 1970 and Millennium birth cohort studies. The studies contain data at each assessment across multiple domains of health and development, making the data valuable for researchers in a large range of disciplines. Data are collected from a range of sources including cohort members, parents, cognitive assessments, physical health assessments, data linkage to administrative records, biological samples, devices etc. In this talk I will describe the range of health, cognitive, social measures in each study at each assessment from birth. Examples of developmental research questions that the cohorts can be used to answer will be illustrated, alongside examples of how the cohorts can also be used in combination to answer questions related to cohort differences over these decades. The data can be accessed free of cost for researcher and I will present an overview of the process for using these data in your research (https://www.ukdataservice.ac.uk/
The German Family Panel pairfam and its potential for developmental research
German Youth Institute and Ludwig-Maximilians-University Munich, Germany
Family research in Germany has long suffered from a lack of large-scale longitudinal data which allow addressing issues of family formation, family development, and the role of family resources for young people’s development. To fill this gap and strengthen the data infrastructure for interdisciplinary family research, the German Family Panel pairfam has been launched in 2008/2009. Funded by the German Research Foundation, pairfam provides annually collected longitudinal data which are open to the scientific use of researchers across the world (see https://www.pairfam.de/en/). This presentation provides an overview of the sample, content, and research potential of pairfam.
Pairfam started with a nation-wide representative sample of three age cohorts and a total of 12,000 anchor participants (4,000 per cohort) drawn from register data. In wave 1, anchor participants were adolescents (age 15-17 years), young adults (25-27 years), and middle adults (35-37 years). Employing a multi-actor design, assessments are not restricted to the anchor participants but also include their partners, parents, and children (starting at age 8) thus allowing dyadic and multi-informant analyses. Data are collected in personal interviews and by written questionnaires. The annual assessments in the pairfam program provide a rich data base for studying four broad domains: (a) fertility and family formation, (b) partnership formation, quality, and dissolution, (c) intergenerational relationships in adulthood and grandparenting, and (d) parenting and child development. Biographical information and indicators of SES, personality, values, attitudes and expectations, work-life balance, health and well-being complement the rich indicators of relationship quality and parenting. Developmental issues can be addressed not only for the adolescent and (young) adult cohorts, e.g. regarding romantic relationship and partnership trajectories or educational careers, but also for children of various ages, e.g. from birth into elementary school age or across the transition to secondary school. Sample analyses will illustrate various research options.
EuroCohort: Europe’s first harmonized birth cohort survey
University of Helsinki, Finland
At present there is no data source available to scientists to comparatively analyse the lives of children across Europe as they grow up and therefor to develop policies at an European level to improve their well-being. The European Cohort Development Project (ECDP) is an EU funded Design Study which is developing an Europe-wide Research Infrastructure that will provide comparative longitudinal survey data on children and young adults from birth onwards. The EuroCohort survey infrastructure developed by ECDP will facilitate a better understanding of life-span and life-course patterns and a predictive analysis of the next generation of European children. Through a common survey design, EuroCohort will collect truly comparable data without the need for post-hoc harmonization of items nor adjustment to do with different fieldwork periods. It will offer insights into topics such as wellbeing, health, poverty, education, employment and leisure as never before. Building the case for EuroCohort is achieved through the following three objectives: 1. Gaining support from key political policy makers with a brief which covers child wellbeing as well as national funding agencies tasked with infrastructural spending on science and survey data collection. 2. Developing a scientifically excellent research design. 3. Establishing a robust operational framework that will ensure logistic integrity of the survey. A central challenge for EuroCohort is to maximize the number of countries who participate in this survey from the outset given the importance of being present in the initial data collection rounds. To this end it is essential that the value of comparative longitudinal survey data to society as a whole is made clear to funders and policy makers in order to make a compelling argument for support.
Using Big Data to investigate social contagion: Nonsuicidal Self-Injury and (social) media
Paul L. Plener
University of Vienna, Austria
Social media facilitate the large-scale spreading of information and thus provide a tool to investigate processes of web-based social mechanism. This paper presents an example of how Big Data from social media networks can be used to study processes of web-based exchange about health-related problem behavior and mechanisms through which such behavior is reinforced in social networks. The focus is on Nonsuicidal Self-Injury (NSSI) which is quite common among adolescents. Social contagion has been discussed as one of the mechanisms that lead to spreading and distribution of NSSI in this age group. Among the functions of NSSI, both intrapersonal as well as interpersonal reinforcement mechanisms have been discussed. However, it is less clear whether theses mechanisms are of significance for the online exchange about NSSI.
We conducted a real-time download of the 14 most common German hashtags from the social media network Instagram for one month. Pictures as well as comments were rated. In a qualitative follow-up study, reasons for posting NSSI content were evaluated. We collected n=32,182 pictures from n=6,721 user accounts, out of which n=2,826 (from n=1,154 accounts) directly depicted wounds. Higher wound grade was associated with more comments. Adding to evidence on offline social processes, the findings suggest that mechanisms of reinforcement seem to be present in social media as well. Most users, who post NSSI pictures report positive experiences online. The discussion will take a wider perspective on the potential of Big Data for studying social mechanisms in social media networks.
Smartphone-sensing as a novel method for the efficient collection of large-scale behavioral and situational data
Ludwig-Maximilians- University Munich, Germany
The last decades of psychological and social science research have been highly dominated by the use of self-report questionnaires as the primary method of data collection. While being convenient, questionnaire data is known to be subject to a series of biases such as response styles, social desirability, ecological invalidity and memory. We argue that the ongoing evolution of mobile-computing- and sensor-technology enable the development of new methodologies for the measurement of human behavior in a more objective, reliable and valid way. Due to their pervasiveness and diverse functionalities, smartphones in particular, are suitable as a research instrument.
In this talk, I will present the current state of the ongoing PhoneStudy mobile sensing project at Ludwig-Maximilians-Universität München. The aim of this project is the development of a scientific tool to collect and to analyze data about behavior and situations in an unobtrusive objective and ecologically-valid way. First, I will introduce the current functionalities of the application and provide insights into the collected data. Second, I will talk about challenges of this and similar projects and provide recommendations for other researchers.
Third, I will present research that has been conducted using the PhoneStudy application. This part will particularly focus on personality and behavior and will feature our latest study about the prediction of personality traits from a variety of behaviors. In this study we collected more than 10.000 behavioral variables from 624 participants over the course of 30 days and used machine learning models to predict big-five personality traits on factor and facet level. I will present both, the accuracy of personality predictions as well as a more detailed look at the most influential variables of the respective models. Finally, I will conclude with a short outlook on current developments in the field of mobile sensing.
Open Science and developmental Psychology: a roundtable and plenary discussion
Convenor: Marcel van Aken
Participants: Ingrid Schoon, Richard Lee, Stefanos Mastrotheodoros.
The idea of Open Science represents changes in the way we do our science, in the way we develop our hypotheses, design our studies, conduct our analyses, and report our results. But also in the way we share our data, for replication and further studies. It also has consequences for the rewards and incentives we have built in our system, including decisions on grants and tenure. As such, it comprises fundamental challenges (and opportunities) to our field.
In a thematic session, the implications of Open Science for our work in developmental psychology will be discussed. After an introduction by Marcel van Aken, four scholars will give a short reflection. After that, the floor is open for a plenary discussion on the topic.
From Developmental Science to Policy: A Discussion between Researchers and Stakeholders
Professor Frosso Motti-Stefanidi, National and Kapodistrian University of Athens, Greece
Professor Ersilia Menesini, University of Florence, Italy
Professor Cintia Rodríguez, Universidad Autónoma de Madrid, Spain
Professor María Ángeles Espinosa. University Autónoma of Madrid / UNICEF-Spain
Dr. Ramya Subrahmanian, Chief of Child Rights and Protection at UNICEF-Innocenti
In recent years, growing attention has been paid to translational research in psychology.
The European Association of Developmental Psychology has been working intensively on the dissemination of research-based results and projects to inform European policy makers, professionals, psychologists, educators, and other stakeholders on the relevant findings of developmental research studies. The decision to hold a policy-relevant meeting in each biennial conference starting from 2015 has been a great appointment for every conference.
This year the special event will take the form of a debate between researchers and stakeholders trying to address three main challenges in children development: (1) How can developmental science findings help improve early childhood education? (2) How can we map and prevent violence towards children within a global approach? (3) How can we promote the integration and resilience of migrant and refugee children and youth? Panel members will give a 5-minute presentation of the implications of research findings or of applications of developmental science research in the field (from preschool to high school, families, government). These presentations will form the basis of a discussion between researchers, stakeholders and the wider audience. Have we succeeded in disseminating scientific evidence to influence public and educational policy and practice? What are some good practices that will allow us to better achieve this goal? If not, why have we not succeeded? The special event of science and policy is directed to a larger audience, involving discussion with practitioners, journalists, politicians and policy makers.