# About engaging and governing strategies > About Engaging and Governing Strategies: A Thematic Analysis of Dark patterns in Social Networking Services. About Engaging and Governing Strategies: A Thematic Analysis of [Dark patterns](https://wiki.g15e.com/en/pages/Deceptive%20patterns.txt) in Social Networking Services. ## Abstract > Research in has shown a growing interest in unethical design practices across numerous domains, often referred to as "[dark patterns](https://wiki.g15e.com/en/pages/Deceptive%20patterns.txt)". There is, however, a gap in related literature regarding social networking services (SNSs). In this context, studies emphasize a lack of users' self-determination regarding control over personal data and time spent on SNSs. We collected over 16 hours of screen recordings from 's, 's, 's, and 's mobile applications to understand how [dark patterns](https://wiki.g15e.com/en/pages/Deceptive%20patterns.txt) manifest in these SNSs. For this task, we turned towards HCI experts to mitigate possible difficulties of non-expert participants in recognizing dark patterns, as prior studies have noticed. Supported by the recordings, two authors of this paper conducted a thematic analysis based on previously described taxonomies, manually classifying the recorded material while delivering two key findings: We observed which instances occur in SNSs and identified two strategies - engaging and governing - with five dark patterns undiscovered before. ## Conclusion > In recent years, the dark pattern landscape has expanded into various different domains. In this paper, we contribute to this body of research through the application and expansion of a comprehensive taxonomy of dark patterns in the context of SNSs. Supported by thematic analysis, we investigate Facebook, Instagram, TikTok, and Twitter and confirm that the platforms all deployed a variety of dark patterns and design strategies aimed at limiting users' agency, steering their decision-making. Findings suggest that dark patterns with higher grades of abstraction are easier to be applied in multiple contexts compared to those given narrower definitions. Lastly, our results yield evidence for two high-level strategies - engaging and governing - containing five types of dark patterns previously not described. ## Contents ### Abstract - Existing research has mainly focused on e-commerce sites, and there is a lack of research focusing on SNS. - To analyze [dark patterns](https://wiki.g15e.com/en/pages/Deceptive%20patterns.txt) applied on SNS, experts recorded their screens while using , , , and (total of 16 hours). Complements the limitations of existing non-expert research (failure to recognize dark patterns) - Through thematic analysis (: an analysis method that finds patterns through qualitative analysis and classifies them by topic), two strategies for intervening in user decision-making (engaging and governing) and five new dark patterns were discovered. ### 1. Introduction - Existing research has mainly focused on e-commerce sites, and there is a lack of research focusing on SNS. - Research on SNS is needed for effective regulation. , , etc. focus on data and privacy protection. RQs: - RQ1. What types of dark patterns are currently used in the four SNSs Facebook, Instagram, TikTok, and Twitter? - RQ2. Do SNSs contain dark patterns currently unique to their domain? ### 2. Related Work (Contents about history omitted) **2.4. Design Strategies on SNSs** - Existing studies also partially cover SNS, but analysis specific to SNS is lacking. - Example: [Defining and identifying attention capture deceptive designs in digital interfaces](https://wiki.g15e.com/en/pages/Defining%20and%20identifying%20attention%20capture%20deceptive%20designs%20in%20digital%20interfaces.txt) focuses only on attention - We analyzed strategies specific to the SNS domain. ### 3. Expert Review & Data Collection - Six experts (3 men and 3 women in Germany, mean age 28.33, SD=1.63) used the four most popular SNS apps, were given 10 tasks, and had their screens recorded. They were asked to [think aloud](https://wiki.g15e.com/en/pages/Think%20aloud.txt). - Since one person used two apps, there are three recordings for each app (6 people * 2 videos / 4 apps = 3 videos/app) - To eliminate bias due to , each participant was given one and one device. They were asked to use both devices simultaneously and attend to differences between them. - Only SNS apps were installed on the devices, and a new phone number and new email account were created for all participants. *I think that if you do this, it will no longer be [Contextual inquiry](https://wiki.g15e.com/en/pages/Contextual%20inquiry.txt) and its meaning will be greatly diminished. An SNS study that begins by severing all the social networks. --ak* Tasks: 1. Turn on screen recording on each device. 2. Open the app and create an account to log in and then out. 3. Close and reopen the app. 4. Create any kind of content, post it, and delete it. 5. Follow and unfollow other accounts. 6. Visit the personal settings. 7. Visit the ad-related settings. 8. Use the application for its intended use (minimum of five minutes) 1. Describe the natural flow of the app - what did you use it for? 2. Could you use the app as you wanted, or did some features 'guide' your interactions? 3. How easy was it to get distracted, and if so, what distracted you? 9. Delete your account 10. Turn off screen recording and save the recording. Procedure: - One hour of online training in advance. Explains dark patterns and provides classification methods. - After the training is over, training materials are distributed. - Randomly assign two SNSs to each reviewer. (Each app was independently evaluated three times by different experts) ### 4. Findings **4.1. Analysis of the Data** - Perform thematic analysis on videos and recordings - Simultaneous use of inductive coding () and deductive coding () - The deductive coding book is derived from the existing dark pattern classification (). Two coders coded independently. - If a pattern is found that does not fit the existing classification, a new code is generated and added to the inductive codebook. - After all coding sessions were completed, the two coders discussed their findings, created an inductive codebook containing 22 patterns, and classified the material accordingly. (Using technique) - **Because they could replay important parts of the videos and obtain additional information from think-aloud recordings, they could identify further patterns hidden across multiple screens.** **4.2. Deductive Codebook Analysis: Findings** (Omitted. Contents that many existing known patterns were found) **4.3. Inductive Codebook Analysis: Findings** - Discovering the "persuasive language" pattern, which is the mirror image of "confirm shaming". Drive users toward specific funnels through positive language. - Finding two strategies and five patterns specific to SNS - Engaging strategies: interactive hooks, social brokering - Governing strategies: decision uncertainty, labyrinthine navigation, redirective conditions **4.4. Engaging Strategies** Strategies to keep people on social media as long as possible - **Interactive hooks**: Reward devices to make you feel good during your stay. Example: Gamification that rewards you for entering more information or adding more friends. - **Social brokering**: Creating more connections between people, encouraging them to connect with new people. If users are asked to upload their contacts, information from people who do not yet use SNS can be obtained without their consent. **4.5. Governing Strategies** A strategy to encourage users to take action that suits the purpose of the SNS provider during their stay - **Decision uncertainty**: Making it unclear which options are available and which option was chosen. Example: When you sign up for TikTok and set up advertising-related settings, [Attention](https://wiki.g15e.com/en/pages/Attention.txt) is stolen by playing a video in the background. When you try to delete your Twitter account, you are blocked in the middle of the path to account deletion by first suggesting 'Deactivate account'. - **Labyrinthine navigation**: Mainly used in SNS settings menu. Making people give up because they can't find the settings they want. - **Redirective conditions**: Includes patterns similar to Forced action. Ensuring that a desired feature is activated only after you 'accept' a seemingly completely unrelated feature. *One of the things I did around <2010> exactly meets this standard. Making players play the game longer (engaging strategies), making them use more paid items while playing the game (governing strategies) --ak* ### 5. Discussions & Implications - As a result of analyzing four SNSs, 44 out of 80 patterns classified in existing literature were being used. - Inductively discovered 22 new patterns and classified them into 5 themes - We were able to discover patterns specific to SNS. ### 6. Limitations & Future Work - We only focused on four SNSs. It would be nice to expand to other SNS such as . - Access to SNS was limited to mobile apps. Need to expand to other modalities. *I suspect that behind Facebook's announcement of its renaming to and focus on the metaverse lay a vision of extending social media's addictiveness into other modalities. --ak* - The study was conducted remotely because of . - The study analyzed HCI experts. - When creating the codebook, we limited it to the classification (What Makes…) used in academia. ### 7. Conclusion - Expanding the dark pattern research of the past few years in the context of SNS - It was revealed that existing SNS widely uses dark patterns. - Discovering new dark patterns specialized for SNS - Identifying two major strategic categories: engaging and governing.