Research on adolescents’ well-being in relation to their smartphone use
About the research
Our research aims to assess the short-term and long-term effects of smartphone use on the well-being of young people. As one of the few qualified teams in the world, we want to start a new wave of research to analyze real-time behavior.
We have developed a unique smartphone application that securely monitors the authentic online behavior of young people on their phones. This application collects accurate information, not estimates as has been common in internet research from its inception.
Apart from the application, we have also developed anonymization software that provides participants with the maximum level of privacy.
Part of our team are social scientists who oversee the research. The other part consists of experienced programmers and IT professionals who ensure that the smartphone application and the data it collects are perfectly secure. You can read more about us on the “Who We Are” page.
We also use the latest machine learning techniques in order to process the amount of data we collect from participants. These methods can better assess what young people are doing on their phones and whether the activity presents a risk or an opportunity for them. These techniques lay the groundwork for assessing the risks of specific activities on smartphones in real-time and for intervening to create a safer internet experience.
During the main data collection, which will take place in 2021, we will collect information about the authentic behavior of young people and information about their daily well-being, in addition to their online activity. This will enable us to capture the effect of technology use on the well-being of young people with much greater accuracy than is the case in previous research.
Why is it necessary to research how young people use their smartphones?
Smartphones have become more widely used than ever before and young people spend a significant part of their day with a phone in their hands. Previous research has mainly used questionnaires to assess phone and technology usage, but these are often inaccurate. A phone application can, for instance, measure the time spent on individual activities on a smartphone much more accurately and thus better assess actual behavior and online risks. Thanks to the authentic data from smartphones, we will be able to answer the question "How does the use of smartphones affect young people?" more precisely.
This type of data collection can also address some of the main research gaps present in this research field, including the accurate and objective assessment of active and passive use of smartphones and the accurate momentary assessment of changes in well-being associated with phone use in adolescents. Read more on the “Why This Research Is Needed” page.
Although this type of innovative data collection can uncover new knowledge, it proves to be challenging and it requires caution when it comes to ethical standards because it deals with sensitive data collected in real-time. We have discussed the methodology of the studies with a lawyer with expertise in the field of cybersecurity to ensure the complete privacy and security of the participants and the data. Four steps were fulfilled to uphold ethical standards:
- Ethics committee approval
The whole study procedure was approved by the Ethical Committee for Research of Masaryk University.
- Personal data protection
Personal data are protected in accordance with the General Data Protection Regulation (GDPR) and other applicable legislation.
- Informed consent
Each participant, as well as one of their parents, have to sign an informed consent form that includes an explanation of the process of the study and the handling of the data before the start of the study.
- Immediate anonymization
The data are anonymized immediately after collection with anonymization software that replaces personal data (i.e. name and surname, address, phone number) with random fictitious data of the same type (i.e. the real address with a fictional address). The original data are deleted. Only the anonymized data are analyzed.
More information about the methodology can be found on the “How It All Works” page.
- review of current research
- discussion of ethical and legal standards for study procedure creation
- initial phone app development and in-house testing
- pilot studies with adolescents
- collection of instant messaging data for training machine learning algorithms
- research questions development and questionnaire preparations
- machine learning algorithm training
- main study data collection
- data analysis
- data analysis
If you would like more information about the project, do not hesitate to contact us at firstname.lastname@example.org.
About our team
We are a team of psychologists, sociologists, computer scientists, and a lawyer who work at Masaryk University at the Faculty of Social Studies and the Faculty of Informatics (Department of Machine Learning and Data Processing). We are also part of the Interdisciplinary Research Team on the Internet and Society (IRTIS). Each of us has different expertise and knowledge. Thanks to this particular combination, we have a unique approach to exploring the authentic behavior of young people in the online world.
We work in the FUTURE project as part of a working group researching the behavior of young people on smartphones. We aim to develop innovative methods for studying young people's online behavior. You can read more about the members of this working group below by clicking on the individual names.
David is a professor at the Faculty of Social Studies and the Faculty of Informatics of Masaryk University, and he is the leader of the IRTIS research team. He has researched children in the context of the opportunities and risks of using the internet for a long time. He is the co-author of "Digital Youth: The Role of Media in Development", which has been translated into Chinese and Korean. He is the main author of the current report, which maps the use of the internet by children in 19 European countries: EU Kids Online 2020: Survey results from 19 countries. If he is not thinking up other research projects, you can find him playing with his young daughter, Rebeka. He is passionate about running, cycling, and cross-country skiing.
Steri is a researcher at the Faculty of Informatics of Masaryk University. Currently, her research is focused on mobile technologies for the improvement of behavioral health and active aging. As part of her research, she is interested in health behavior and its psychological impacts across populations. Steri is the author of publications in many international journals and she spent 16 years in the United States, where she received her doctorate (from the University of Illinois at Champaign-Urbana) and worked as an academic (at Penn State University at University Park). Steri also works at the University of Ostrava. Her other research interests include health psychology, physical activity, subjective well-being/quality of life, adult development and aging, women's health, and mind-body therapy (e.g. yoga). Steri likes to dig in the dirt, exercise, read science fiction, and have fun outside with her children.
Jaromír is an assistant professor at the Faculty of Informatics of Masaryk University and the leader of the IT part of our team. The research part of his pragmatic self focuses mainly on modeling and simulation, machine learning, and assistive technologies to help people with disabilities. As part of his pedagogical activities, he teaches subjects from all corners of informatics, from the basics of programming through the principles of operating systems and from computer networks to the field of social informatics. In his free time, he enjoys time with his children, board games, and sports, especially if the sport involves a ball of any size.
Pavel is a lawyer who works at the Faculty of Informatics of Masaryk University. He works within the Center of Excellence and focuses on issues related to cybercrime, cybersecurity, and the protection of critical information infrastructures. He is also a researcher and a lecturer at the Department of Law and Technology at the Faculty of Law of Masaryk University. As part of this project, he oversees the legal side of the research so that the procedures protect both the research participants and the researchers themselves. He likes to go for walks with his wife and dog, and he plays in a gothic-rock band.
Aleš has been working as a researcher in the field of artificial intelligence and language technologies for more than 20 years. Since 2014, he has been the Head of the Department of Machine Learning and Data Processing of the Faculty of Informatics of Masaryk University. His research focuses on the analysis of meaning, mining information from text, the automatic recognition of text style, and the design of specialized systems to create ontologies, lexicographic databases, and e-dictionaries. He is a co-creator of systems used in research and security projects in dozens of institutions around the world.
Michał Tkaczyk, Ph.D. is a postdoc researcher at the Department of Media Studies and Journalism (Faculty of Social Studies) the Department of Machine Learning and Data Processing (Faculty of Informatics) at Masaryk University in Brno. His scholarly and research interests include associations between media use, adolescents sleep and well-being, changes of journalism in Czech press, analysis of media and communication content, computer-assisted text analysis.
Jana Blahošová, M.A. is a PhD student of Media studies and journalism at the Faculty of Social Studies at Masaryk University. In her dissertation work, she examines the excessive internet use in adolescents and how it relates to factors from family environment such as the quality of relationships with parents or parental mediation. She is also interested in behavioral addictions, especially in technological and online addictions in young people. She also assists in academic journal Cyberpsychology: Journal of Psychosocial Research on Cyberspace. In her spare time, she enjoys baking, writing and practicing yoga, and if she could, she would like to work as a professional food taster.
Michaela is a doctoral student of media studies at the Faculty of Social Studies of Masaryk University and our junior researcher. Since 2017, she has been researching sexually explicit media and their impact on society. As part of the project, she works in the social science part of the team. The rest of the time she is in love with the robot Karel, whom she wishes to adopt.
Ondřej is a doctoral student at the Faculty of Informatics of Masaryk University and plays the role of a researcher and programmer in the research team. He is responsible for the design of the technologies that are used for data collection and processing. In his Master's studies, he focused on data anonymization and the topic of his doctoral research is the application of machine learning to human-computer interaction. See his LinkedIn profile for his professional experience. Ondra enjoys cycling in his spare time.
Martin Tancoš is a Ph.D. student of psychology at the Faculty of Social Studies, Masaryk University. He performs management of intensive longitudinal data and objective data from smartphone use. He also analyses these data using linear multilevel models and structral equation modeling.
Michal worked at the Faculty of Informatics of Masaryk University and is our programmer responsible for technical solutions. In the team, together with Jaromír and Ondřej, he developed the smartphone application for data collection. Among other things, he ensured that the application can do everything that other team members request, and at the same time keep it is safe and friendly for our research participants. His current interest is app development for Android and wearables. He is also interested in progressive technologies, such as artificial intelligence. Michal is also interested in growing chili peppers and testing how spicy is too spicy.
The current research on young people and technology derives its findings mainly from questionnaires in which children, young people, parents, and teachers answer questions about the use of technology and online activities. Such inquiries are not precise because they do not make it possible to objectively capture the details of what children and young people really come into contact with and how it affects them.
With contemporary children and young people spending twice as much time on mobile devices as in 2010 (source EU Kids Online 2020), it is no longer enough to have rough estimates of times, activities, and experiences that they have online. To move research forward and provide relevant information for creating a safe online environment, we need to know exactly what children, and young people in particular, are doing on their mobile devices.
More specific information
The latest research from our team found that 70% of Czech children use social networks daily, 75% of Czech children watch videos on YouTube daily, and 73% use a smartphone to communicate with family and friends. This information, unfortunately, does not include the content they view on social networks or many other specifics.
- What content are they watching on YouTube?
- How do they have fun on the internet?
- What effect does this content have on children's behavior and their well-being?
Although popular media often publicize the fact that technology has a generally negative effect on people, in reality, we are still looking for answers to these questions.
Beyond cross-sectional surveys
The mostly cross-sectional nature of internet and technology research in relation to the well-being of adolescents has so far generated inconsistent and often vague findings that do not allow for concrete implications to be made. The innovative real-time smartphone data collection process that we propose can address knowledge gaps that cross-sectional and self-report-based research cannot. It allows for more accurate measurement of technology and social media use, not only with exact and objective estimates of screen and app time, but also with an accurate assessment of active and passive use through the number of screen touches to uncover the exact content young people engage in and scroll through every day and by seeing the reasons that they turn to their phone screens (i.e. recreational, goal-oriented, procrastination).
Comprehensive approach to well-being
The research on this topic also often uses well-being measures that assess adolescents’ average well-being across time and investigates only between-person associations of technology use with these overall well-being estimates. We aim to more comprehensively look at well-being in relation to smartphones and technology.
- With pre- and post-study questionnaires, we will be able to focus on long-term well-being (physical, psychological, social).
- With short in-app EMA (ecological momentary assessment) questionnaires during four bursts of intensive two-week data collection, we will look at well-being as a momentary affective state in the situational contexts of everyday life.
By combining this information and looking at within-person effects, we will be able to understand both the long-term and short-term effects of smartphones on young people’s well-being more precisely than before.
“With contemporary children and young people spending twice as much time on mobile devices as in 2010, it is no longer enough to have rough estimates of times, activities, and experiences that they have online.”
About our method
To comprehensively investigate what young people do on their phones, our smartphone application collects several types of data:
- objective data from the phone
- data from short EMA questionnaires that the application automatically generates several times a day
We collect this data for two weeks (14 days) in both the pilot and main studies. In the pilot studies, we also ask to receive downloaded conversations from Facebook Messenger (from those participants who use this application) for the last year (i.e. 1 January 2019 until the end of the specific pilot study). All data that we obtain are stored on a secure server, where personal data are deleted and replaced with random other data of a similar type (i.e. surname is replaced with a fictitious surname). Due to the huge amount of data and in order to maintain the privacy of all of the research participants, this anonymization is done automatically using computer programs.
In the pilot studies, carefully trained social researchers come into contact with hundreds of thousands of text lines of depersonalized data, from which it is impossible to identify individual persons. In the main study, all data is processed automatically with machine learning algorithms that are trained by the data from the pilot studies. After anonymization, the original data is deleted and only the replaced personal data are kept for analysis. This procedure ensures that the data are protected and not distorted.
The whole procedure is in accordance with the valid legal regulations and it was approved by the Ethical Committee for Research of Masaryk University.
More about data collection
The application collects screenshots of the smartphone screen every 5 seconds. Screen capture can easily be paused by clicking a button in the application. The collected screenshots, as well as other data, will be sent to a secure server of Masaryk University the moment the smartphone connects to Wi-Fi (so as to avoid using the mobile data of the research participants).
We use OCR software to convert the acquired screenshots into text. The textual data will also go through anonymization software, which we have developed for this project. Non-text data, such as images and videos, will be deleted and replaced with a short descriptive text. This means that existing programs will try to automatically detect the content of the image and store only the text that describes this content in a database (e.g. a man riding a bike in the park) or tries to find the name of the video playing in the YouTube application and store only that name.
Anonymization software replaces personal data with fictitious personal data so that it is not possible to identify the participants or other persons to whom the information in the data may relate. Personal data are protected in accordance with the rules set out in the General Data Protection Regulations (GDPR) and other applicable legislation. In our study, we consider, for instance, names written together with surnames, addresses, telephone numbers, e-mail addresses, and account numbers to be personal data. In practice, this means that the first name and surname — Jana Nováková, for example — will be replaced by the software with a random name and surname — Magda Poličná, for example. This anonymization procedure is currently one of the most widely used methods of data anonymization in the technology sector and, in addition to companies such as Microsoft and Google, it was used in this year's Census of Population and Housing in the United States. Once the data from the acquired screenshots are anonymized, they will be deleted from the server, which guarantees that we will no longer have access to the original data.
The use of anonymization software preserves the privacy of study participants and allows researchers to interpret data without losing context. Anonymized texts from all participants will be combined into one huge corpus that will be hundreds of thousands of lines long. It will not be possible to distinguish individual persons in such a large amount of text. The text is further analyzed and processed only by trained researchers from Masaryk University (in pilot studies) and by the computer program only (in the main study). In the pilot studies, researchers will look for conversation topics (e.g. hobbies, relationships, health) and the content that the adolescents consume and produce, in text form. No one will see the photos that young people send or the content of the videos they view. These findings will then be used to train machine-learning algorithms to be able to fully automate the data processing for the main study.
In addition to screenshots, the application collects the following data:
- Information about the name of the application that is currently running in the front and the time spent in it
- Information about physical activity (which automatically collects possible states as detected by smartphone sensors, including walking, running, traveling in a vehicle, resting)
- Battery status information (percentage)
- Call and SMS history (only at the time of collection)
- Screen taps (information if someone touches the screen - yes/no)
- Connected headphones (yes/no)
- Playing music on the phone (yes/no)
This data is collected at all times (i.e. during the whole two weeks) and it is downloaded to a secure server each time the participant connects to Wi-Fi. We will never associate this data with a specific person, and it will be accessed with the maximum possible level of security. We will work with information from all participants at once, aggregate it, analyze it, and connect it to findings from screenshots and EMA questionnaires.
In addition to the collected phone data, we will also inquire about some information through questionnaires. During the pilot studies, the application briefly inquires about the operation of the application every morning and evening. The answers take only a minute and focus on the functionality of the software. These questions are for our developers to improve the application as needed. They are supplemented with EMA questions that inquire, for instance, about the quality of sleep or the current emotions of the participant. In the main study, only EMA questions related to the well-being of participants are asked several times a day. Participants also obtain pre- and post-study questionnaires that ask about demographic information, individual characteristics, and several types of well-being.
After the end of the data collection in the pilot studies, we will collect the conversations (i.e. chats) of willing participants from Messenger and WhatsApp applications (from 1 January 2019 until the end of the data collection). According to our instructions, the study participants download the conversations from the IM apps themselves to a computer and then upload them to our online application, which automatically and securely uploads only the textual part of the conversations. It does not upload pictures, photos, or other media files. It encrypts and sends the data to a secure server. Conversation texts are then anonymized, and personal data is replaced with random data, following the same procedure as the text from screenshots. Social researchers see and analyze only anonymized data. The original conversations are permanently deleted from the server after the necessary length of time. IM data is used to train machine learning algorithms for authentic data and to enhance the accuracy of OCR software for converting screenshots into text in preparation for the main study. This type of data is not collected during the main study.
About the theory
The goal of the FUTURE project, and this working group as well, is to develop a comprehensive integrated evidence-based theory depicting the complex impacts of technology usage on the physical, psychological, and social well-being of adolescents from both the short- and long-term perspectives.
We looked at existing theories that we aim to integrate and enhance. As the starting point, we use the Differential Susceptibility to Media Effects Model (DSMM) (Valkenburg & Peter, 2013), which aims to explain why some individuals are more susceptible to media effects and how and why media influences them. In essence, the DSMM distinguishes three types of differential susceptibility variables to media effects (e.g. dispositional, developmental, social) that affect the individuals’ media use. The use leads to immediate media responses (e.g. cognitive, emotional, excitable) that serve as moderators between the media use and the media effect (i.e. the main outcome). The DSMM acknowledges that susceptibility variables affect media use as predictors but also stimulate or reduce media effects as moderators. Finally, media effects can also reciprocally influence differential susceptibility variables, media use, and media responses. In our project, we adapt the DSMM model to focus on well-being as the main outcome and we enrich it with detailed consideration of both the direct and indirect impact of risky and beneficial activities, accounting for the diverse psycho-social and environmental influences in adolescence. Thus, we will also utilize other selected relevant theories that can provide more insights into adolescents’ well-being and media use and which can enhance various aspects of the DSMM model.
To conceptualize the differential susceptibility variables, we draw from Bronfenbrenner’s (1977) Ecological Systems Theory (EST) and Jessor’s (2014) Problem Behavior Theory (PBT). The EST provides a framework for understanding the influences on development by situating the adolescent within a system of relationships affected by multiple levels (i.e. systems) of the environment. Based on EST, we differentiate the susceptibility variables differently from DSMM and focus on the individual (i.e. microsystem), social (i.e. mesosystem), and country levels (i.e. macrosystem). The PBT provides a framework for adolescents’ problem behavior and specifies risk and protective factors that affect their development. It recognizes several domains within which we can identify these factors, including personality, perceived environment, and actual behavior. Other important factors in PBT include vulnerabilities and opportunities to engage in risks, controls against risks, and social support for positive behaviors. These factors thus appear among differential susceptibility variables, but also in media use, which can be conceptualized as either a risky or positive behavior. Finally, using the PBT allows us to overcome the dominant risk perspective because the PBT considers protective factors not only as protective for risks but also as promotive for positive development.
Jessor, R. (2014). Problem behavior theory: A half-century of research on adolescent behavior and development. In R. M. Lerner, A. C. Petersen, R. K. Silbereisen, & J. Brooks-Gunn (Eds.), The developmental science of adolescence: History through autobiography (p. 239–256). Psychology Press.
The positive view of technology usage for well-being improvement is also accentuated in the Health Belief Model (Ahadzadeh et al., 2015), which is widely used to explain health behavior change and how people use technology for change if they are aware of a threat from a specific health condition or behavior and perceive benefits from the considered actions. This model will be especially useful for the examination of the effects on physical well-being.
The conceptual model of the project
Based on the aforementioned theoretical inputs, we adapted the DSMM and developed our base conceptual model for the effects of online activities on adolescents’ well-being. The model is described below (Figure 1).
In our model, long-term well-being, which is conceptualized as physical, psychological, and social, corresponds to the media effect in DSMM. The same dimensions are also present in the short-term perspective (i.e. media response in DSMM), acknowledging that online activities can also lead to immediate and temporary changes in an adolescent’s well-being. We will thus understand both the short-term and long-term effects of technology on adolescents’ well-being as outcomes. The strength of this project is its comprehensive focus on all dimensions of well-being. However, it is not possible to focus on all of the possible indicators of well-being. We selected the following indicators to be captured in our research.
The indicators of physical well-being will be adolescents’ sleeping and physical in/activity (including sedentary behavior) habits. Both indicators can be affected in the short term (e.g. disrupted sleep after engaging in online gaming or higher physical activity after using mHealth applications) and in the long term (e.g. general sleeping and physical in/activity patterns over a longer period of time).
The indicators of psychological well-being in our project will be adolescents’ emotional state (e.g. distress, anxiety, happiness), self-esteem, and life satisfaction. The short-term aspects in this well-being dimension correspond to temporary emotional changes as a result of technology usage (e.g. being upset by comments on a social networking profile, feeling happy while communicating with a friend), and the long-term aspects correspond to more permanent and stable emotional levels (e.g. increased anxiety after facing repeated cyber aggression, enhanced self-esteem as a function of participation in online communities).
As for the indicators of adolescents’ social well-being, we will examine the short- and long-term aspects of relationship quality (with peers, family, school, community) including perceived social support and social skills. Short-term effects include increased perceived support, and long-term effects include increased connectedness among family members.
Differential susceptibility variables
Designed by pch.vector / Freepik
Based on the EST, we categorize the susceptibility variables differently from the DSMM and specifically focus on those that can constitute risk or protective factors in line with the PBT. The selection of specific variables is also based on our prior research on children’s and adolescents’ risky activities online (see e.g. Smahel & Wright, 2014; Smahel et al., 2014). In our conceptual model, we differentiate:
- individual-level variables with factors such as age, gender, developmental stage, and skills
- social-level variables with factors such as peers, family, school, and community, and we examine and account for the inequalities reflected in media use and well-being by focusing on the role of socioeconomic status (SES)
- country-level variables such as culture, media, values, technology provision, and welfare.
Smahel, D. & Wright, M. F.(eds) (2014). Meaning of online problematic situations for children. Results of qualitative cross-cultural investigation in nine European countries. London: EU Kids Online, London School of Economics and Political Science
Media use in our model is understood to be the extent and content of adolescents' online activities. Based on previous research (Castellacci & Tveito, 2018, Livingstone et al., 2011), we will focus predominantly on a set of activities that have shown to have a positive or negative impact on well-being or that raise substantial concern about their effect on well-being. As a basic differentiation, we distinguish between interaction activities (e.g. communication through technology with friends, family, and unknown people, participating in online groups) and consumption activities (e.g. viewing online content). In the project, we will also focus on the general pattern of technology usage (e.g. time and frequency, excessive use, usage of diverse devices). It is crucial to note that, although we will measure specific activities, our investigation will lead to the detection of more general patterns and the recognition of media affordances, such as interactivity, and textual or visual character. That is, we will focus on the meaning of the activity regardless of the specific device, platform, or application used. The focus on affordances rather than on how exactly they are achieved will allow us to provide insights and generate models that will be applicable in the future with technologies not yet developed.