Periodic Reporting for period 1 - TheRealCompetition (When Public Health Campaigns Warn You, but Your Friends Like to Drink – Connecting Computational Social Science and Neuroscience to Understand Real-World Health Behavior)
Okres sprawozdawczy: 2019-09-01 do 2021-08-31
What is the problem/issue being addressed?
Malleable unhealthy behaviors like binge drinking excessively burden global societies. Public health media campaigns could offer cost-effective, large-scale interventions, but affect behavior only minimally. To date, health campaigns are optimized to outperform alternative health campaigns in their effects on behavior (e.g. comparing gain to loss-framed messages). Yet, in real life, campaigns face other competitors like campaign-related media and peer-produced content. These sources are amplified by popular social media and thus often presented in close proximity to health campaigns. Each source may favor healthy or unhealthy viewpoints, causing health campaign-consistent or -contradictory updates to an individual’s evaluation of a behavior. Ignoring real-world competitors of health campaigns is a grave oversight that may cause diminished or even boomerang effects. I seek to understand and optimize public health campaign effects in the context of TheRealCompetition, namely healthy and unhealthy media and peer content on social media. This is non-trivial. State-of-the-art self-report measures of media exposure and psychological processes underlying media effects overburden lay participants who struggle to recall and explain how they integrate multiple competing influences on their behavior. In a game-changing interdisciplinary approach, I connect computational social science and neuroimaging to objectively and unobtrusively quantify daily exposure to campaign-related social media content and understand psychological mechanisms that explain campaign effects in the context of other sources of influence. The resulting, novel model of real-world campaign effects offers actionable recommendations to practitioners and contributes to neuroscience and computational social science by providing ecological validation of key models.
Why is it important for society?
This action directly tests theoretical models about behavior change in the context of risky alcohol consumption, specifically binge drinking in young adults, which is a serious public health problem around the world. As such, the results of this action will deliver new, practical insights that may support the improvement of population-level health.
In addition, the basic science problem being addressed here is to explain how everyday decisions are being made under the influence of competing sources of information. This is a process that drives behavior in many different domains of daily life. As such the findings from this research may provide insights into behavioral domains outside the specific one being tested here. To illustrate, this action specifically examines examples of competing sources of influence on risky drinking including friends and marketing efforts that are positive about alcohol and alcohol consumption and efforts to discourage risky drinking such as public service announcements. However, this juxtaposition of competing information is characteristic of many decisions we make in daily life, both in the health context and in other situations. For instance, smokers may face a similar dilemma between smoking friends and scientific information about the dangers of smoking. Similarly, in the midsts of the COVID-19 crisis, parents may face contradictory information with governments promoting social distancing and other sources suggesting that it may have negative for their kids if they cannot play with other kids regularly. That is, the basic science work performed here is relevant for society because it may explain basic decision making processes that drive behaviors across many domains.
What are the overall objectives?
The four main research objectives were to:
- Quantify how likely young adults are to be exposed to multiple types of alcohol-related information on various social media platforms when they browse their feeds at different times of day on different days of the week.
- Unobtrusively assess whether causal effects of health campaigns on behavior and the extent to which they are dependent on daily exposure to other sources of influence on social media in a field experiment.
- Unobtrusively assess psychological processes that underlie causal effects of health campaigns in the context of competing sources using functional neuroimaging.
- Identify campaign characteristics that allow campaigns to outperform real-life competitors.
Other objectives of this action include training objectives, which were to develop coding skills, primarily in Python, develop skills in computational social science, primarily focusing on natural language processing and machine learning, reintegrating in Europe, and popularization of my work and grant writing. Finally, this action includes knowledge transfer objectives which focus on the transfer of neuroimaging expertise to ASCoR, my host institution, transfer of open science expertise to ASCoR, sharing my network with University of Amsterdam members, and starting new projects together with UvA members.
Malleable unhealthy behaviors like binge drinking excessively burden global societies. Public health media campaigns could offer cost-effective, large-scale interventions, but affect behavior only minimally. To date, health campaigns are optimized to outperform alternative health campaigns in their effects on behavior (e.g. comparing gain to loss-framed messages). Yet, in real life, campaigns face other competitors like campaign-related media and peer-produced content. These sources are amplified by popular social media and thus often presented in close proximity to health campaigns. Each source may favor healthy or unhealthy viewpoints, causing health campaign-consistent or -contradictory updates to an individual’s evaluation of a behavior. Ignoring real-world competitors of health campaigns is a grave oversight that may cause diminished or even boomerang effects. I seek to understand and optimize public health campaign effects in the context of TheRealCompetition, namely healthy and unhealthy media and peer content on social media. This is non-trivial. State-of-the-art self-report measures of media exposure and psychological processes underlying media effects overburden lay participants who struggle to recall and explain how they integrate multiple competing influences on their behavior. In a game-changing interdisciplinary approach, I connect computational social science and neuroimaging to objectively and unobtrusively quantify daily exposure to campaign-related social media content and understand psychological mechanisms that explain campaign effects in the context of other sources of influence. The resulting, novel model of real-world campaign effects offers actionable recommendations to practitioners and contributes to neuroscience and computational social science by providing ecological validation of key models.
Why is it important for society?
This action directly tests theoretical models about behavior change in the context of risky alcohol consumption, specifically binge drinking in young adults, which is a serious public health problem around the world. As such, the results of this action will deliver new, practical insights that may support the improvement of population-level health.
In addition, the basic science problem being addressed here is to explain how everyday decisions are being made under the influence of competing sources of information. This is a process that drives behavior in many different domains of daily life. As such the findings from this research may provide insights into behavioral domains outside the specific one being tested here. To illustrate, this action specifically examines examples of competing sources of influence on risky drinking including friends and marketing efforts that are positive about alcohol and alcohol consumption and efforts to discourage risky drinking such as public service announcements. However, this juxtaposition of competing information is characteristic of many decisions we make in daily life, both in the health context and in other situations. For instance, smokers may face a similar dilemma between smoking friends and scientific information about the dangers of smoking. Similarly, in the midsts of the COVID-19 crisis, parents may face contradictory information with governments promoting social distancing and other sources suggesting that it may have negative for their kids if they cannot play with other kids regularly. That is, the basic science work performed here is relevant for society because it may explain basic decision making processes that drive behaviors across many domains.
What are the overall objectives?
The four main research objectives were to:
- Quantify how likely young adults are to be exposed to multiple types of alcohol-related information on various social media platforms when they browse their feeds at different times of day on different days of the week.
- Unobtrusively assess whether causal effects of health campaigns on behavior and the extent to which they are dependent on daily exposure to other sources of influence on social media in a field experiment.
- Unobtrusively assess psychological processes that underlie causal effects of health campaigns in the context of competing sources using functional neuroimaging.
- Identify campaign characteristics that allow campaigns to outperform real-life competitors.
Other objectives of this action include training objectives, which were to develop coding skills, primarily in Python, develop skills in computational social science, primarily focusing on natural language processing and machine learning, reintegrating in Europe, and popularization of my work and grant writing. Finally, this action includes knowledge transfer objectives which focus on the transfer of neuroimaging expertise to ASCoR, my host institution, transfer of open science expertise to ASCoR, sharing my network with University of Amsterdam members, and starting new projects together with UvA members.
So far, progress has been made on multiple research objectives by obtaining ethics approval for all proposed studies and developing novel data collection and analysis protocols for social media data that carefully balance the needs for detailed data about participants' social media experience, participant privacy, and participant burden. I have also developed necessary resources for neuroimaging data storage and analysis as well as a set of stimuli for experimental work. Pre-test data has been collected for the proposed social media research.
To meet key training objectives, I have successfully completed courses developed by the University of Michigan on applied machine learning and text mining in Python.
Finally, I made significant progress on multiple knowledge transfer objectives and on my plan to reintegrate into the European science context by becoming board member and open science officer of the Amsterdam Center for Health Communication, attending workshops and becoming part of new collaborations with European researchers, and attending a personal leadership program at the University of Amsterdam. I am also developing a novel website to present the findings of my work more broadly (www.cobras-lab.com).
To meet key training objectives, I have successfully completed courses developed by the University of Michigan on applied machine learning and text mining in Python.
Finally, I made significant progress on multiple knowledge transfer objectives and on my plan to reintegrate into the European science context by becoming board member and open science officer of the Amsterdam Center for Health Communication, attending workshops and becoming part of new collaborations with European researchers, and attending a personal leadership program at the University of Amsterdam. I am also developing a novel website to present the findings of my work more broadly (www.cobras-lab.com).
Although this action was terminated early due to conflicts with other funding mechanisms, the work done so far has laid important ground work that will support similar research lines going forward. EU support for this ground work will be gratefully acknowledged in future progress reports beyond the state of the art.