Announcing our 2024 Psychology of Technology Dissertation Award Winners
Please join us in congratulating our 6th Annual Dissertation Award winners and honorable mentions: Lauren Eales, Mayaan Malter, Sumer Vaid, Huili Chen, Shahyar Mohsin, and Ertugrul Uysal.
We are thrilled to announce the winners of the 2024 Psychology of Technology Dissertation Award. As part of the Psychology of Technology Institute’s (PTI) continued efforts to build capacity for understanding and improving the human-tech relationship, we have awarded promising graduating scholars awards for outstanding dissertations for the past 5 years.
Please join us in congratulating our three 6th annual Psychology of Technology Dissertation Award winners: Lauren Eales, Mayaan Malter, and Sumer Vaid, as well as our three honorable mentions: Huili Chen, Shahyar Mohsin, and Ertugrul Uysal. Check out their fascinating dissertation abstracts below!
Award Winners:
Lauren Eales, University of Minnesota - LinkedIn
Title: Children’s Media Use, Family Psychological Functioning, and Parental Media-Based Racial Socialization during the Dual COVID-19 and Whiteness Pandemics
Abstract: The dual pandemics of COVID-19 and Whiteness have been jointly affecting numerous families’ lives since 2020 around the world, particularly in the United States. The Whiteness pandemic (Ferguson et al., 2021) has existed long before 2020, but its existence was highlighted in May 2020 following the murder of Mr. George Floyd. This dissertation includes two mixed methods studies that addressed various consequences of these dual pandemics, including child screen media use behaviors, family resilience, and media-based White racial socialization. The data from these studies come from online survey data collected with parents (mostly mothers) living in, primarily, the Twin Cities, MN metro region from 2019 to 2021, and individual interview data collected in 2022. Study 1 assessed screen media (i.e., screen time) and problematic media use (addictive-like screen behaviors) from 2019 to 2021, and found an inverted U-shape for both, peaking in 2020. Some parents reported returning to “normal” in 2021, while others reported continued struggles with media use. Study 2 assessed how White mothers were using media to talk to their children about race using a mixed methods collective case study. Active race-related media mediation was related to lower ethnic protection. Mothers across multiple White racial identity development groups were not using media to talk to their children about race, often citing their child’s age and emotions as reasons to not engage in that conversation. Findings from these studies can be used by researchers and interventionists to improve child and parent mental health, family resilience, and parental racial socialization during the dual pandemics.
Summary: Lauren's research examines the combined impact of the COVID-19 pandemic and heightened racial awareness on family dynamics in the US. The study is pivotal as it explores how families use media to navigate race conversations and manage resilience during overlapping crises. Her findings reveal trends in media use, showcasing a peak during 2020, and provide insights into the challenges and strategies of White mothers in addressing race with their children through media, shedding light on the intricacies of media-based racial socialization and its effects on child and parent mental health.
Mayaan Malter, Columbia University - LinkedIn
Title: Perceptions of Disability: Implications for New Product Design and Marketing
Abstract: My dissertation research focuses on consumers with disabilities and their unique relationship with technology. Essay 1 proposes a conceptual model for decision making with a disability. Consumers with disabilities face an inherent tradeoff due to their disability: choosing to prioritize one of two fundamental and competing needs, autonomy versus efficiency. Compared to other coping mechanisms (going at it alone or asking another person for help, respectively) which only fulfill one of the two competing needs, using technology has the potential to enable disabled consumers to increase both their autonomy and efficiency.
In contrast, able-bodied consumers often experience a technology paradox (Mick and Fournier, 1998), where new technologies (recommendation systems, AI, autonomous products) allow people to accomplish more, while simultaneously taking away their ability to do things autonomously (Leung et al., 2018; de Bellis and Johar, 2020; Puntoni et al., 2021). Although the technology may take away some degree of autonomy for an able-bodied person, it can level the playing field for a disabled consumer by allowing the disabled person to act and make choices in the same way as their able-bodied peers.
Essay 2 explores how the prioritization of fundamental needs differs between lay theories of physical disability and experienced needs of people with physical disabilities. In thirteen experiments and one survey (N = 4,541, including 254 participants with physical disabilities) I test how these misperceptions impact new product design (new technology) and marketing. I find that observers elevate high-level needs (need for acceptance, respect, autonomy, meaning, and fulfillment) of people with disabilities, and neglect lower-level needs, as compared to what disabled participants want for themselves. This focus on higher-level needs impacts consequential decisions. It affects choices of features built into new products (the choice of a choose-your-own-ingredients feature over a speed-up feature in an autonomous cooking machine), and messaging used in marketing campaigns (the choice of a tagline that reads “Increase your personal independence with this transportation service!” over a tagline that reads “Complete tasks with ease and speed with this transportation service!”). These choices are suboptimal when other needs, such as the need for an efficient method of accomplishing of tasks, are overlooked.
My results are counterintuitive, given previous research on vulnerable consumers and on disability, which shows that observers devalue the high-level (psychological) needs of marginalized populations (Schroeder and Epley, 2020), and the countless accounts of disabled people being neglected in the marketplace and in society. To understand these robust and surprising results, I tested several psychological mechanisms. I find that observers’ belief in an egalitarian world leads them to feel inspired by people with disabilities, which in turn elevates their perception of high-level need importance. When I break the connection between disability and feeling inspired, I mitigate the high-level needs elevation effect. I rule out an alternative mechanism that observers elevate the high-level needs of physically disabled people to compensate for lack of fulfillment of low-level needs. Apart from the mediating role of inspiration, the elevation effect is multiply determined. Further research is needed to fully understand the underlying process.
Summary: Mayaan focuses on the unique technological interactions of consumers with disabilities, highlighting the trade-offs they navigate between autonomy and efficiency. The research stands out for its depth in exploring how these consumers use technology to balance their needs, contrasting with non-disabled consumers who may experience a paradox where technology both enables and limits autonomy. This work has significant implications for new product design and marketing, aiming to better align with the actual needs of disabled individuals.
Sumer Vaid, Harvard Business School - LinkedIn
Title: Social Media Sensitivity: Probing Heterogeneity Across People, Places, Platforms, Types of Use and Time
Abstract: My dissertation examines five potential sources of variation in the relationship between social media and wellbeing: people’s psychological dispositions (e.g., tendency for psychological vulnerability), the physical and social context of social media use (e.g., where and around whom use is occurring), the platform that is being used (e.g., Instagram, YouTube), the type of use that is occurring (e.g., passive active) and time domains over which social media effects manifest (e.g., hours, weeks and months). I deploy a combination of random- slope multilevel models (frequentist and Bayesian) and multilevel structural equation modelling to analyze two large corpuses of experience sampling data collected from American young adults (n=3,233; k= 219,027). Results reveal small but persistent associations between social media use and wellbeing outcomes, moderated by a range of psychologically salient variables. First, psychologically vulnerable people tend to feel worse after using social media as compared to psychologically healthy people, regardless of how social media is used (e.g., actively vs passively) and what platform is being used. Second, using social media in social locations and around others is associated with negative wellbeing outcomes regardless of what platforms are used and how they are used. In contrast, both passive and active social media use are associated with positive wellbeing outcomes at the within-person level but decreased wellbeing outcomes at the between-person level. Lastly, preliminary evidence suggests that social media’s effects on wellbeing accumulate steadily over time instead of “drenching” users in the moments after use has occurred.
Summary: Sumer’s dissertation looks into the nuances of how social media impacts well-being across various dimensions, including individual psychology, social environments, and types of social media engagement. The research is compelling as it uses advanced statistical models to parse out these effects in a large dataset, revealing that while social media can sometimes enhance well-being at a personal level, it often leads to decreased well-being across broader comparisons. This work is particularly relevant for understanding the complex interplay between social media usage and mental health.
Honorable Mentions:
Huili Chen, MIT - LinkedIn
Title: Robots as Social Catalysts: A Multidisciplinary Framework for Designing Embodied Social Agents that Foster Long-term Human Collaboration and Connection
Abstract: As artificial intelligence (AI) devices become more common in our homes, concerns about their potential harm to human-human connections arise accordingly. This dissertation aspires to study the responsible design of embodied agents as social catalysts to purposefully enhance human-human interactions. It aims to shed light on the following three overarching research questions. Can we become more socially connected and collaborative with one another through the facilitation of a socially embodied agent? What social capabilities do these embodied agents need to acquire as social catalysts? What approaches should we take to design, develop and evaluate computing systems that enable positive social interactions between a human group and an embodied agent responsibly?
To investigate the three questions, this work proposes a multidisciplinary framework for the holistic design and evaluation of embodied social agents intended to foster human-human connection and collaboration. It argues that robots need to possess three social capabilities: social-affective perception, context awareness, and social adaptation. These capabilities are elaborated in detail within the framework, together with a comprehensive, iterative process for their design, evaluation, and enhancement. This process needs to be grounded in theories and findings in psychology, and employ a mixed-methods integrative approach that involve computing, social sciences, and interaction design.
A case study centered on parent-child reciprocal interaction is conducted to demonstrate and evaluate this proposed framework, highlighting the unique complexities and possibilities of multi-person human-robot interaction. The case study aims to facilitate enriching adult-child exchanges essential for children's development while overcoming various technological and methodological challenges posed by young children as a user group. A series of studies and experiments were conducted in this dissertation to examine all key aspects of the long-term multi-person human-agent interaction (M-HAI). These aspects include understanding the dynamics of human-human interaction, modeling social-affective dynamics in human-human interaction, introducing design guidelines for long-term M-HAI, and designing and evaluating adaptive M-HAI.
In summary, this dissertation provides insights into the potential of designing embodied social agents as social catalysts within human groups. It invites future exploration into the possibilities and challenges of machine-catalyzed group interactions, emphasizing both technical and ethical considerations. As sociable intelligent devices—from personal voice agents at home to autonomous vehicles—rapidly proliferate, humans increasingly interact with AI agents in an ecology composed of other humans and other intelligent machines. Accordingly, this work helps advance the social sophistication of intelligent machines that live with humans in this emergent human-agent ecology, as well as the understanding of the social and behavioral mechanisms underlying this ecology.
Summary: Huili Chen's dissertation proposes a framework for designing robots as social catalysts to enhance human-human interaction. This is fascinating as it not only addresses the technical capabilities required for robots to act as facilitators of social interaction but also integrates psychological and design perspectives to ensure these interactions are meaningful. The research includes case studies like parent-child interactions, offering insights into the potential for robots to positively influence human social dynamics and development.
Shahryar Mohsin, Bocconi University – LinkedIn
Title: Gender-Ambiguous Digital Voices and Consumers Judgment
Abstract: In 2022, Apple debuted a new voice for its Siri that did not sound obviously male or female. This new gender-ambiguous voice was introduced in the wake of criticism levied by the United Nations and others that subservient voice assistants, being overwhelmingly female, reinforced gender stereotypes (Schwarz et al., 2018). The new voice could avoid this issue and help introduce a wide audience to a non-binary voice. While these goals are laudable, it remains unclear how people will react when they cannot easily evaluate the gender of a speaker.
A substantial body of research in processing disfluency suggests that when it is harder to process information, people react negatively (Alter & Oppenheimer, 2009). For example, difficulty understanding a speaker due to a foreign accent (Lev-Ari & Keysar, 2010) or poor audio quality (Newman & Schwarz, 2018) can lead listeners to downgrade their assessment about the quality of the information that was presented. Could a similar process also apply to difficulty processing social information? That is, when the information itself is easy to understand and process, could difficulty processing social information, such as the gender of the speaker, lead to a similar feeling of disfluency? I predict that it could, and I call the difficulty of processing social information social disfluency.
There are at least two reasons why this might be the case. First, it may be difficult to categorize a non-binary voice, and this feeling of difficulty could spill-over into evaluations of the speaker or evaluations about what the speaker is talking about (Nass & Brave, 2005). Second, feelings of fluency are associated with feelings of familiarity (Alter & Oppenheimer, 2008). If people are unfamiliar with the sound of a non-binary voice, people may get an uneasy feeling of disfluency. Since non-binary voices are less commonly encountered in day-to-day life, people may react negatively toward them merely because they seem unfamiliar. Alternatively, it may be that people harbor prejudice against members of the LGBTQ community (Lick & Johnson, 2013), since non-binary voices are often associated with this community (Schwarz et al., 2018).
In five experiments, products described by gender-ambiguous voices received less favorable ratings compared to those described by clearly male or female narrators. In Study 1 (N=602), participants rated toothpaste when described by Siri’s gender-ambiguous voice against male and female versions of Siri. As predicted, a main effect of narrator (F(1,599)=58.41, p<.001) emerged, favoring clearly-gendered narrators. Social disfluency mediated the decrease in product liking (bindirect=-0.08, SE=.03, 95%CI [-0.14, -0.03], p<0.05). Study 2 (N=207) replicated the main effect by manipulating voice gender by shifting up or down the fundamental frequency (F(1,206)=8.63, p=.004), with social disfluency mediating (p<.01). Study 3 (N=472) replicated findings across different digital voices and product categories longitudinally. Study 4 (N=601) ruled out social categorization account. Study 5 (N=604) supported the familiarity account; additional exposure to a gender-ambiguous voice eliminated the negative effect (F(1, 602)=16.62, p<.001), while ruling out the prejudice account. Familiarity, not bias, thus appeared key in mitigating the impact of a gender-ambiguous voices on product judgments.
While I have explored how social disfluency affects product judgments, the implications may be broader. For instance, would a news story reported by a gender-ambiguous speaker be more likely to be considered “fake news?” Could a similar effect occur for other socially unfamiliar voices? Indeed, it seems plausible that intuition about social disfluency could have led managers to hire radio personalities that sound highly uniform, potentially creating bias against people who do not sound familiar to audiences. The findings elucidate reactions to gender-ambiguous digital voices. While initial negative responses may occur, increased exposure holds promise as a potential remedy.
Summary: Shahryar’s research addresses the societal reactions to gender-ambiguous digital voices, like the new voice of Siri introduced by Apple in 2022. His dissertation explores the concept of "social disfluency," which occurs when people struggle to process social information such as the gender of a voice, leading to negative reactions similar to those experienced with disfluencies caused by accents or poor audio quality. Shahryar hypothesizes that this social disfluency could affect how information is perceived, possibly causing gender-ambiguous voices to be viewed unfavorably. His findings suggest that while initial reactions may be negative, increased exposure could mitigate these effects. This research is particularly relevant in discussions about gender stereotypes in technology and could influence how digital voices are designed in the future.
Ertuğrul Uysal, ETH Zürich – LinkedIn
Title: The Rise of Intelligent Technologies and Social Media: Implications for Human-Technology Relationships
Abstract: This dissertation explores the evolving dynamics of human-technology relationships, emphasizing the implications of intelligent technologies and social media. The dissertation encompasses three distinct articles. The first article critically evaluates anthropomorphism in AI, exploring its theoretical conceptualization and application across fields like computer science, robotics, psychology, and marketing. Despite the widespread use of humanlike features in technologies, a comprehensive understanding of anthropomorphism in AI is lacking. As a remedy, this piece conducts a systematic literature review of existing empirical work (N=69), offers a conceptual framework, and recommends best practices for future exploration. For managers, this conceptual framework identifies a key take-home message: there could be great promise in employing anthropomorphism in AI wisely, but its careless use may undermine the potential of cutting-edge new AI technologies for both consumers and firms.
The second article adopts a relationship-centric lens to analyze the impacts of anthropomorphic AI assistants (AIAs), such as Alexa, on consumers (Gray, Gray, and Wegner, 2007). Consumers spend large amounts of time with their AIAs, potentially developing a relationship over time that builds on an exchange of benefits and (psychological) costs (Dwyer, Schurr, and Oh, 1987). A preliminary survey (N=200) and user interviews (N=11), a field study (N=200) and a field experiment (N=600) with AIA users show that AIA anthropomorphism may threaten users’ identity, which disempowers them, creates data privacy concerns and ultimately undermines their well-being. These harmful effects particularly emerge in close, long
relationships. The field experiment uncovers three practical interventions to empower users for handling their data privacy and attenuate harmful effects of AIA anthropomorphism in relationships with consumers: (1) informing consumers about the data practices of the firm, (2) informing consumers about ways to protect their data privacy and (3) encouraging consumers to take action to protect the privacy of their data. Our findings suggest that, to minimize the harmful effects of AIA anthropomorphism and maximize its benefits, managers should empower consumers regarding the protection of their personal data.
The third article contemplates the profound societal shifts induced by social media. We suggest that social media could be an important catalyst for a shift in two basic individual values, achievement and conformity, transforming the very fabric of our societies (Festinger 1954; Gouveia, Milfont, and Guerra, 2014). We used a difference-in-differences analysis of the European Social Survey data from 2002 to 2008, involving approximately 135,000 participants from 15 countries. Our analysis revealed that countries with higher social media adoption, compared to culturally similar countries, experienced a significant rise in achievement and conformity orientation. We replicated these findings in a consumer survey study with a nationally representative sample (N=1,000) and explored the pertinent mechanism: consumers’ social media use increases achievement and conformity orientation through an increased need for approval, particularly for those who tend to self-ruminate. In a third experimental study (N=212), we manipulated social media use and established a causal link between social media use and the activation of achievement and conformity values. Overall, our findings suggest that consumption of social media contributes to the emergence of a societal culture that places a strong emphasis on achievement-seeking and conformity. While our findings do not necessarily suggest that social media intensifies harmful negative values, the fact that values evolve not as a result of a cultural and systemic process within societies but are transformed by consumers using a specific digital technology raises questions on the role of those technologies in societies. Therefore, an important responsibility falls on technology firms to design a healthy and sustainable digital life by adopting consumer-friendly attributes for their platforms.
In essence, this dissertation seeks to illuminate the complexities of consumer-technology relationships and their implications.
Summary: Ertuğrul’s dissertation delves into the evolving dynamics of human-technology relationships, with a focus on the implications of intelligent technologies and social media. His research is structured around three articles that collectively examine the impact of anthropomorphism in artificial intelligence (AI). Ertuğrul critically evaluates how anthropomorphic features are utilized in AI, proposing a systematic framework for understanding and employing anthropomorphism effectively to avoid undermining user autonomy and well-being. His findings from extensive surveys and experiments indicate that while anthropomorphic AI can enhance user relationships with technology, it also raises significant concerns about identity and privacy. This work is crucial for tech developers and managers, providing insights on how to balance the benefits of AI with ethical considerations.
Please join us in congratulating the above winners and honorable mentions. They are in good company with our previous winners, who have gone on to contribute great things to the field. Onward!