Team: 2 Members Duration: 10 weeks Project Type: Research Proposal
My Role: I oversaw brainstorming and frame-working. Also collaborated with my team member as the Lead Researcher throughout the project cycle from initial development to method creation and paper formatting. Cultivated a plan of action utilizing recognized research methodologies.
In my Research Methods class, I was tasked with developing a research proposal about an area of interest. Talking with my teammate, as people who had both used TikTok, we sought to explore and develop academic expertise in an area that has gained little attention despite its global hold on media and culture. Moreover, as UI/UX designers and researchers with a large interest in the intersection of tech and ethics, we were intent on navigating one of the largest social networking sites to gain insight into the future of UI/UX design and how we can be proactive and intentional about the projects and products we develop.
Over the course of 10 weeks, this project was developed collaboratively with help from peers and UI/UX Researchers. With exploratory research, we learned about dark patterns which led us to compound our research proposal around a central question: How does TikTok's algorithmic dark pattern affect consumer well-being over time.
How does TikTok’s algorithmic dark pattern affect consumer well-being over time?
Throughout the timeline of this project, I learned the importance of intentionality–particularly when curating scope. Early on in the process, we faced difficulties from failing to fully grasp what our research proposal attempted to explore. Intending to investigate TikTok, we struggled to proceed forward because the scope was far too vague and large. However, after talking and many brainstorming sessions later we were able to specifically emphasize TikTok's algorithm as a dark pattern. Ultimately this project taught me the necessity of making decisions to creating a stellar output. Oftentimes, I believe I can be uncommitted to certain decisions for the sake of avoiding conflict or pushback, but doing that here slowed down our process greatly.
Thus, moving forward, if I were to do this project again, I would focus even more on making decisions from the very beginning. I would spend the time fine-tuning a balance between exploration and focus–focused exploration.
As students who recognize the material impacts TikTok has had on cultural discourse and digital habit development, creating scholarship on the app in an area of minimal academic literature will inform UX/UI designers and the future of social networking technology. Consequently, this study aims to examine the influence of TikTok’s algorithmic dark pattern on its consumers. Exploring this, we propose 15 semi-structured interviews with a diverse group of individuals, aged 16-23 years old, that use the platform three to five times or more per week. Using grounded theory analysis, we will begin preliminary theory construction by recognizing overarching themes that arise. Expected themes include users’ experiences with relative comparison, identity construction, and the influence of aesthetics. Analyzing these concepts, we can gain a greater understanding of algorithmic categorization and the effects this can create for young consumers in digital spaces, particularly as social networking sites grow in popularity.
"As students who recognize the material impacts TikTok has had on cultural discourse and digital habit development, creating scholarship on the app in an area of minimal academic literature will inform UX/UI designers and the future of social networking technology."
With over 800 million daily active users, twice that of its social media competitors, TikTok has become a digital behemoth, dictating many of the most popular songs, trends, and personalities since its conception (Mou, 2020). Its beginnings would originate through the merge of two separate platforms: Musical.ly, a 2014 Shanghai-based app that gained traction in the United States, and Douyin, a 2016 app that drew millions of users in China. In 2018, given Douyin and Musical.ly’s success, the apps would combine thus introducing TikTok into the social media market (Anderson, 2020).
Since then, the app has become central to internet culture, made true by its young consumer base–40% of whom are between 10-19 years old (Haenlein et al., 2020). As university students who have navigated the app to communicate with friends, create content, understand trends, and interact with virality, we have seen our experiences on TikTok guide our generational culture and toxic digital habits. Having said this, the app remains the topic of very few academic studies despite its continuing rise in popularity. Specifically, many in the academic communities glaze over the role dark patterns and algorithms have played in harmfully cultivating TikTok’s user experiences, viewing them as separate areas of interest.
In a 2018 Association for Computing Machinery (ACM) Conference on Human Factors in Computing Systems (CHI) paper, dark patterns, in a general sense, are discussed as “instances where designers use their knowledge of human behavior (e.g., psychology) and the desires of end users to implement deceptive functionality that is not in the user’s best interest” (Gray et al., 2018). Within the context of TikTok, typical dark patterns manifest through features such as its scrolling interface, video lengths, and button size. However, our study aims to define TikTok’s algorithm as a central dark pattern, worthy of analysis. Given that the algorithm generates specifically curated content in rapid succession, it is designed through the psychological understanding that humans gravitate towards content that pertains to them (Grison & Gazzaniga, 2019). When one is oversaturated with media and categorized by an inanimate algorithm, it creates the potential for harm. For that reason, the context of dark patterns in this study is framed through TikTok’s use of its algorithm to foster an unhealthy digital environment that exploits the wellbeing of its users. With the utilization of semi-structured interviews that explore individual cognizance of algorithmic effects on personal health, we aim to implement grounded theory practice to increase acknowledgment of this pattern of concern and begin to generate findings. Using these methods we can start to understand the impact of algorithmic dark patterns and the necessity to improve digital literacy, and create more scholarship on these topics, specifically amongst user experience and user interface (UX/UI) designers working in social networking technologies.
"...the context of dark patterns in this study is framed through TikTok’s use of its algorithm to foster an unhealthy digital environment that exploits the wellbeing of its users."
As social media and digital experiences become pertinent topics of discussion, research surrounding dark patterns and their effect on users has grown. In common literature, dark patterns are traditionally described as online practices that mislead the consumer through the intentional exercise of manipulative UX/UI design patterns. The level of impact that dark patterns have on consumers is relative to the degree to which the designer manipulates them. This undermining of user autonomy could range from being coerced into buying expensive products to becoming addicted to particular platforms. (Mathur et al., 2021).
In a study conducted at the University of Luxembourg, researchers attempted to understand consumer perception of traditional dark patterns. Collecting 406 survey responses, they found that participants were aware of behavior manipulation by dark design principles but “were uncertain if manipulative designs online [could] cause them harm.” Having said this, the “most prominent harm identified by participants was harm to themselves (135) both of psychological and physical nature,” (Bongard-Blanchy et al., 2021). As demonstrated by this study, participants were often able to understand the negative ways their online spaces were designed to exploit them, yet the connection between dark patterns and wellbeing remained vague; many participants failed to grasp the long-lasting repercussions.
Given TikTok’s global hold, findings about dark patterns have materialized in a way that is distinct to the platform and intrinsically tied to the algorithm. When a user first creates an account, they are asked, but not obligated, to pick from a list of popular interests (such as comedy, art, and sports.). After this initial screening, the algorithm slowly starts to settle into a personalized formula distinct from many of its competitors: “...algorithmic recommendation that structures the user's experience to a greater extent than any major social media platform to date...” (Guinaudeau et al., 2021). Data about who the user follows, interacts with, and even information about the amount of time they spend on certain posts are used to cater to their For You Page (FYP)–TikTok’s personalized recommendation feed. Consequently, the app is controlled by the user’s content and app interactions, which can open up a user to harm.
"Data about who the user follows, interacts with, and even information about the amount of time they spend on certain posts are used to cater to their For You Page (FYP)–TikTok’s personalized recommendation feed. Consequently, the app is controlled by the user’s content and app interactions, which can open up a user to harm."
Building on this idea, a December 2020 study by Ellen Simpson and Bryan Semaan (2021), sought to understand the impact of TikTok’s algorithm on intersectional, marginalized populations. Interviewing 16 self-identified members from the LGBTQ+ community, a shared feeling of passivity manifested towards the algorithm’s limited representation of themselves despite their regular use of the platform. In one salient example, a 21-year-old Black cisgender queer woman expressed:
Sadly, there’s a lot of creators of color or LGBTQ creators that are not really featured even though there’s so many – the majority of [my] feed is white people. . . which there is nothing wrong with that obviously. . . Like yeah there’s so many others using the app; you guys need some Black people up here. (2021).
Failing to understand this woman’s racial identity yet computing her queerness, this study highlighted the harmful impact of being incompletely understood through an algorithm. While this would be a common thread throughout the study–the understanding of being misunderstood by algorithms–the research would fail to connect these reductive categories to a large-scale dark pattern within the app.
Referencing our understanding of the algorithm as a dark pattern, we want to create room to be critical of the algorithm’s deceptive intentions in the same way traditional dark patterns face criticism. In the case of TikTok, traditional dark patterns related to the UI compound with a dark algorithmic pattern of categorization to create users that risk addiction, depleting self-esteem, and overall lesser well-being. In a study that investigates TikTok’s influence on the cognitive behavior of young consumers, it is written that “although more pre-adolescents produced content via TikTok, older adolescent members of the sample...were significantly more likely to agree that they did so to fulfill needs for praise and social recognition as well as self-identity creation and fame-seeking desires...” (Bossen & Kottasz, 2020). Therefore, examining Social Networking Sites (SNS) through a biopsychosocial lens, they present as an online space for essential identity development that is nurtured through peer interaction and reaction. Subsequently, TikTok becomes a site of communal dependency that could harbor self-destructive attitudes from systematized model constructs, such as TikTok’s algorithm.
"...examining Social Networking Sites (SNS) through a biopsychosocial lens, they present as an online space for essential identity development that is nurtured through peer interaction and reaction. Subsequently, TikTok becomes a site of communal dependency that could harbor self-destructive attitudes from systematized model constructs, such as TikTok’s algorithm."
Hence, in examining TikTok we must be aware of how “our habits of thought and cultural expression are being increasingly shaped by the logic of these algorithms," (Collie & Wilson-Barnao, 2020). The metrics these algorithms use to curate content can be replicated in its users with devastating consequences. Moreover, research literature acknowledging TikTok’s algorithmic dark pattern as a considerable factor of its users' health is scarce, therefore, our proposed study seeks to fill this gap by addressing the following question: how does TikTok’s algorithmic dark pattern affect consumer well-being over time?
RESEARCH METHOD & METHODOLOGY
With the intention of investigating the effect that TikTok’s dark algorithmic pattern has on its consumers, our proposed research design focuses on discerning individual perceptions of the algorithm and their general sentiment regarding their consumption. By recognizing common attitudes associated with a user’s time on the app, we can become conscious of direct influences the algorithm produces and the larger implications it generates for our relationship with algorithms and technology.
Taking what we learned from our literature review and engaging with Kathy Charmaz’s (1996) work on theoretical development, we recognize a space to conduct research using a grounded theory approach. Through semi-structured in-depth interviews, the collected research would aid our theoretical evolution, particularly as key issues materialize around our larger topic.
"Taking what we learned from our literature review and engaging with Kathy Charmaz’s (1996) work on theoretical development, we recognize a space to conduct research using a grounded theory approach. Through semi-structured in-depth interviews, the collected research would aid our theoretical evolution, particularly as key issues materialize around our larger topic."
Given that algorithmic dark patterns influence the user, we reasoned that the best and most accessible way to understand the complexity of the impact would be through interviews. Interviews would provide insightful, qualitative information into a specific individual’s experience with the app (their FYP) based on their demographics and personal interests, rather than using methods such as a survey, which limits theory development as it does not fully capture the depth of their experiences. Thus, this method of choice aligns with a constructivist approach because the interpretation of the individual’s data shapes the meanings and active progress towards algorithm dark pattern influences on the TikTok community (Charmaz, 1996).
SAMPLE / POPULATION
We plan to concentrate our study on TikTok users that are 16-23 years old to facilitate a sizable portion of the consumer base and incorporate users that are digitally literate, having had prior social media experience, exposure, and knowledge. This sample population will contain 15 users and a mix of genders, sexuality, and races, all of whom use TikTok at least three to five times or more per week. In this manner, our data is gathered with a myriad of viewpoints and a recognition of how technology often replicates the same inequities in the material world. To find these participants, our main method of recruitment will be a short survey distributed online that checks if an individual meets the aforementioned criteria as well as monetary compensation to those we interview (see Appendix A).
Our interviews will be conducted over the course of three weeks, either electronically through Zoom conference calls or in a socially-distanced classroom setting that respects COVID-19 protocols. Each participant will be interviewed separately for 60-90 minutes. Given our choice of a semi-structured interview style, we will maintain the general flow of the interview questions and record field notes of open-ended responses on paper. We will also audio-record (with permission) for reference later, in order to capture the respondent’s emotions/feelings properly. Our interview flow is as follows (see Appendix B):
1) Opening: These questions will confirm the information participants entered in their survey and explore the period of time they've been on TikTok and their current amount of usage.
2) Central I: These questions explore participants’ initial motivations for using TikTok, how it compares to their motivations today and what keeps them engaged on the app. In exploring these questions we can understand how their engagement has evolved with the app.
3) Central II: These questions allow us to understand the user’s personalized experiences with the FYP and TikTok’s algorithm. Entering this portion of the interview, we seek to grasp the impact the algorithm has had on the users and more about the ways the algorithm has categorized them.
4) Closing: These questions will show us how the interviewees think in regards to their awareness of the algorithm and its influences on them within and outside of the app. This section will also explore final thoughts about the explicit technical aspects of the app without influencing any of the previous answers the participants have given us.
Utilizing the grounded theory process, we will be able to enforce focused coding and categorize similar patterns that arise in the collected participant responses. For instance, participants that have similar usage styles and emotions associated with the app will be grouped together accordingly. From there, we would be able to compare and contrast each category to understand their relationships with each other. Eventually, our theory could be cultivated from the connections between abstract concepts that emerged from individual interviews. Potential abstract concepts could include users’ interactions with relative comparison to the content they see, working identity and sense of self, and aesthetic trends.
THREATS TO VALIDITY
One limitation of semi-structured interviews is that social-desirability bias and recall bias can play into invalidating some of the responses we receive, particularly if participants understand the scope of our study. Another restriction includes how our theory is shaped through our data therefore our study develops as we gather information from our participant interviews. In seeking to develop theory in an area that has received fairly minimal academic exploration, there is a degree of fine-tuning our team will have to do as the study progresses.
The rising prominence of TikTok in everyday life and culture has introduced a new research chapter. Its presence can not go unnoticed by many, which signifies the importance of our research as it would provide insight into the inner workings of an algorithm that heavily dictates why the platform has become so successful. Using information from our interviews, we can begin to generate theories about the long-lasting impacts of an algorithm distinct from many of its social networking peers.
As a research proposal that utilizes grounded theory, it is unclear exactly what our expected results would entail. However, given our experiences on the app, the knowledge of our peers, and the methods we are utilizing, we believe our interviews will uncover common threads about addiction and the larger implications of becoming categorized through an algorithm. Unlike other social media, “[the] experience of using TikTok is one of repeatedly engaging with one’s own self: intra rather than interpersonal connection,“ (Bhandari & Bimo, 2020). The user is constantly interacting with an externalized version of themselves on TikTok–one that is iterative, and possibly reductive–while continuing to use the app because of traditional dark patterns that perpetuate continued use. In this way, the algorithm becomes a dark pattern that impacts what the user takes away from the app and potentially influences their wellbeing. By constantly seeing content hyper-categorized to the user, set along the backbone of harmful UI practices, the user continues to see content that resonates with them but only through what the algorithm can understand–sometimes missing key elements of an individual. It opens avenues for them to create toxic conclusions about themselves with an obsession to self-categorize into the algorithm.
"Unlike other social media, '[the] experience of using TikTok is one of repeatedly engaging with one’s own self: intra rather than interpersonal connection,' (Bhandari & Bimo, 2020)."
Understanding the scope of our research, its findings can be applied to further more discourse about TikTok’s algorithm, particularly with UX/UI designers who can uphold our findings to understand the negative ramifications of algorithmic dark patterns and the material impacts they can have on impressionable teenagers (those who often navigate social media the most).
Having said this, our research runs into certain limitations that must be understood and acknowledged. As a research proposal that is focused on algorithmic dark patterns, our study remains unable to fully explore the breadth of the culture the algorithm creates. This includes larger conversations about body image, microfame, racial biases, and more. While these are topics worthy of scholarship, this discussion cannot be had without the foundational exploration of TikTok’s algorithm.
Elaborating on possible ethical issues in our study, it investigates the well-being of an individual yet we retain the ability to be intrusive and potentially cover digital habits that might be too personal for individuals to answer. Moreover, because our study finds its data through interviews, there is always the potential for information about what the participants are saying to be misconstrued. Overall, as literature on dark patterns develops, there remains the possibility that all we, as researchers, can do is raise awareness while companies continue to perpetuate the use of dark patterns for the sake of profit. Despite this, we see an opportunity to recentralize the user in our design practices.
With this proposed research, we seek to begin discourse around the design of algorithms. As we construct our digital experiences, we must also begin to grapple with the effect of becoming algorithmized. If these algorithms must craft content through their understanding of our identity, what does that communicate to us? In TikTok’s case, with an algorithm much different than other social media, we believe this is a space worthy of investigation–a space that must be understood within the literature of dark patterns and harmful design practices. Given the scale of TikTok’s reach, the absence of studies that establish the app’s algorithmic dark pattern raises several questions about the future of social technology. As a result, our research illuminates the qualitative impact this cultivates in its consumers and provides room for this literature to be created. Using these findings, we strongly believe we will be able to understand growing impacts and better challenge the principles of UX/UI designers that have perpetuated the continued use of dark patterns to retain young users.
"As we construct our digital experiences, we must also begin to grapple with the effect of becoming algorithmized. If these algorithms must craft content through their understanding of our identity, what does that communicate to us?"
Anderson, K. E. (2020). Getting acquainted with social networks and apps: it is time to talk about TikTok. Library Hi Tech News, 37(4), 7-12. https://doi.org/10.1108/lhtn-01-2020-0001
Bhandari, A., & Bimo, S. (2020). TikTok and the ‘Algorithmized Self”: a New Model of Online Interaction. AoIR Selected Papers of Internet Research. Published. https://doi.org/10.5210/spir.v2020i0.11172
Bongard-Blanchy, K., Rossi, A., Rivas, S., Doublet, S., Koenig, V., & Lenzini, G. (2021). “I am Definitely Manipulated, Even When I am Aware of it. It’s Ridiculous!” - Dark Patterns from the End-User Perspective. Proceedings of ACM DIS Conference on Designing Interactive Systems. Published. http://hdl.handle.net/10993/47008
Bossen, C. B., & Kottasz, R. (2020). Uses and gratifications sought by pre-adolescent and adolescent TikTok consumers. Young Consumers, 21(4), 463-478. https://doi.org/10.1108/yc-07-2020-1186
Charmaz, K. (1996). Grounded Theory. Rethinking Methods in Psychology, 27-49. https://doi.org/10.4135/9781446221792.n3
Collie, N., & Wilson-Barnao, C. (2020). Playing with TikTok: algorithmic culture and the future of creative work. The Future of Creative Work, 172-188. https://doi.org/10.4337/9781839101106.00020
Gray, C. M., Kou, Y., Battles, B., Hoggatt, J., & Toombs, A. L. (2018). The Dark (Patterns) Side of UX Design. Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems, 1–14. https://doi.org/10.1145/3173574.3174108
Grison, S., & Gazzaniga, M. S. (2019). Self and Personality. Psychology in Real Life (3rd ed., pp. 502–543). W. W. Norton & Company.
Guinaudeau, B., Votta, F., & Munger, K. (2021). Fifteen Seconds of Fame: TikTok and the SupplySide of Social Video. Published. https://osf.io/aeqcw/
Haenlein, M., Anadol, E., Farnsworth, T., Hugo, H., Hunichen, J., & Welte, D. (2020). Navigating the New Era of Influencer Marketing: How to be Successful on Instagram, TikTok, & Co. California Management Review, 63(1), 5-25. https://doi.org/10.1177/0008125620958166
Mathur, A., Kshirsagar, M., & Mayer, J. (2021). What Makes a Dark Pattern. . . Dark? Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems. Published. https://doi.org/10.1145/3411764.3445610
Mou, J. B. (2020, May). Study on social media marketing campaign strategy TikTok and Instagram [Master’s thesis, Massachusetts Institute of Technology]. MIT Libraries DSpace@MIT. https://dspace.mit.edu/handle/1721.1/127010
Simpson, E., & Semaan, B. (2021). For You, or For"You"?: Everyday LGBTQ+ encounters with TikTok. Proceedings of the ACM on Human-Computer Interaction, 4(CSCW3), 1–34. https://doi.org/10.1145/3432951
PROPOSED SURVEY QUESTIONS
This appendix features a general format of the proposed recruitment screener survey.
This survey will require you to share openly about personal preferences. Do you agree to share honestly about these subjects? (Agree or Disagree)
How old are you?
What is your race? What is your ethnicity?
What are your preferred pronouns?
What is your sexuality?
What is your gender identity?
Have you participated in a consumer research study in the past 3 months? (Please specify if yes)
Please tell us how often you use the TikTok app in a week: (Select one)
- I do not use TikTok
- 1-2 times a week
- 3-4 times a week
- 5+ times a week
- I don’t know
- I do not use TikTok
- 1-2 times a week
- 3-4 times a week
- 5+ times a week
- I don’t know
Please approximate the amount of time spent on the app each time you use it: (Select one)
- I do not use TikTok
- 30 minutes-1 hour or less
- 2-3 hours
- 4-5+ hours
- I don’t know
- I do not use TikTok
- 30 minutes-1 hour or less
- 2-3 hours
- 4-5+ hours
- I don’t know
PROPOSED SURVEY QUESTIONS
This appendix features the full list of the proposed interview questions.
When did you first get TikTok?
How many times do you go on TikTok in a week? Each time you use the app, approximately how long do you stay on it? (minutes)
Central I Questions:
Why did you first start using TikTok?
Why do you still use it today?
What keeps you engaged on the app?
Central II Questions:
What content does the For You Page show you? Is it reflective of who you are?
How does the content on your For You Page make you feel?
What impact has TikTok had on you since you first started using the app?
Are you aware of how the TikTok algorithm works?
How does the algorithm impact what you take away from the app?