Lesson Report:
Title
Algorithmic Polarization in Practice: Recreating Cho et al.’s YouTube Study via “Walkable Cities�
Synopsis: After a brief check-in and housekeeping (including a deadline extension and onboarding for new students), the class shifted from last week’s “data buyers/monetization� lens to the citizen’s lived experience of algorithmic feeds. Students learned about affective polarization from Cho et al. and then attempted a mini-replication of Cho’s YouTube experiment, comparing “self-generated� vs. “socially generated� recommendation paths using the polarized topic of 15-minute/walkable cities. Groups seeded the algorithm with pro or con content, documented how recommendations evolved, and began mapping their trajectories for comparison next session.
Attendance
– Absent students mentioned: 0
Topics Covered (chronological, with activity names and detailed instructions)
1) Housekeeping and Onboarding
– Reflection Journal deadline: extended to Thursday (midnight Bishkek time). About half the class already has feedback on Journal 1; students were invited to email with questions.
– New students welcomed. Instructor offered a separate onboarding session next week to review course foundations.
– eCourse access: some students still lack credentials (especially OSUN students). Students without access were asked to email the instructor so PDFs/readings can be forwarded directly.
– Note: several students reported connectivity issues; instructor accommodated re-entry and breakout reassignment as needed.
2) From Data Points to Lived Experience (Goal-Setting and Recap)
– Recap of last class: How individuals are turned into data points; how those data points are sold to political actors and businesses.
– Today’s pivot: How algorithmic systems (reels/shorts/recommendations) shape the citizen’s viewpoint—comfort vs. challenge, and implications for political perception.
3) Warm-Up Poll: “Comforted or Challenged?�
– Prompt: When was the last time you used Instagram Reels/TikTok, and did your feed make you feel more “comfortedâ€� or “challengedâ€�?
– Results pattern: Most students use algorithmically-driven short-form video; responses leaned toward “comforted,â€� though several reported feeling “challengedâ€� or “curious/motivated.â€�
– Framing: Sets up the hypothesis that algorithmic feeds often deliver comfort/confirmation, priming affective polarization.
4) Key Reading Anchor: Cho et al. and Affective Polarization
– Concept: Affective polarization—when repeated exposure to confirming content shifts feelings about out-groups from mere disagreement to moralized disdain (e.g., seeing the other side as “wrong,â€� “bad,â€� even “evilâ€�).
– Platform focus: YouTube. Cho et al. examined when and how the algorithm amplifies polarization.
– Crucial distinction:
– Self-generated recommendations: user searches for a topic, then clicks results.
– Socially generated recommendations: user receives/opens a specific video link shared from outside the platform.
– Scientific principle: Replicability. The class will emulate core conditions of Cho’s design to compare how recommendation pathways differ.
5) Polarizing Topic Selection: “Walkable Cities� (15-Minute Cities)
– Quick definitions:
– Walkable/15-minute city: most daily essentials (school, hospital, groceries) reachable on foot within ~15 minutes; public transit and bike infrastructure prioritized.
– Student perspectives surfaced:
– Pro (e.g., Banu): better air quality, less traffic, strong public transit; cars retained for special needs/longer distances.
– Con (Elijah’s framing): risk of control—fear that limiting car ownership/mobility could restrict freedom of movement; government overreach concerns.
– Clarifications (Aya and Gavin): practical descriptions of benefits and what “15 minutesâ€� means in daily life.
– Rationale for topic choice: Widely debated in the US/EU; capable of triggering opposing feeds and echo chambers.
6) Mini-Replication Experiment: YouTube Recommendation Pathways
– Group structure:
– 8 breakout rooms labeled Pro or Con (stance on walkable cities).
– Odd-numbered rooms = Self-generated polarization condition.
– Even-numbered rooms = Socially generated polarization condition.
– Seeding rules and search/link inputs:
– Self-generated groups (odd rooms):
– Pro search terms (examples): “why Dutch cities are better,â€� “case for walkable cities,â€� “ending car dependency,â€� “benefits of bike lanes.â€�
– Con search terms (examples): “problem with 15-minute cities,â€� “the war on cars is real,â€� “why we need cars for freedom,â€� “problems with public transportation,â€� “in defense of suburbs.â€�
– Procedure:
1) Use YouTube search to find a video consistent with assigned stance.
2) Watch 30–45 seconds; pause; refresh page.
3) Click a sidebar recommendation on the same topic/stance; again watch ~30–45 seconds; refresh.
4) Repeat to a third video.
5) Take a screenshot of the recommended videos sidebar after the third watch (this is the first “feed snapshot�).
– Socially generated groups (even rooms):
– Instructor provided a seed link aligned with the assigned stance (one link for Pro, one for Con).
– Procedure mirrors above after opening the seeded link:
1) Watch 30–45 seconds; pause; refresh.
2) Select a sidebar recommendation on the same topic/stance; watch; refresh.
3) Repeat a third time.
4) Screenshot the recommended videos sidebar after the third watch (first “feed snapshot�).
– Documentation and diagramming (the “algorithm traceâ€�):
– Build a visual diagram showing:
– Starting point: search term used (self-generated) OR the seed link (socially generated).
– Video 1 → Video 2 → Video 3 (titles/thumbnails preferred).
– Feed Snapshot 1: screenshot of recommendations after Video 3.
– Then click one more on-topic recommendation (Video 4), refresh, and capture Feed Snapshot 2 to show evolution.
– Format: whiteboard, Google Doc, or slides; use arrows to show the path and paste images/titles of videos and feed snapshots.
– Timing:
– Initial breakout: ~10 minutes. Follow-up consolidation: ~12 minutes to assemble diagrams and capture additional feed snapshots.
– Instructor handled intermittent disconnections and reassignments to keep groups on track.
7) Wrap-Up and Next Steps
– Time ran short; full group presentations deferred to Thursday.
– Immediate request: Each group to drop their diagram (even if partial) into the chat.
– Preview of Thursday’s class:
– Compare self-generated vs. socially generated pathways and Pro vs. Con feeds.
– Conduct an activity exploring each student’s “algorithmic self-profileâ€� to see what platforms likely infer about them.
Actionable Items
Urgent (before Thursday)
– Submit Reflection Journal: Due Thursday by midnight Bishkek time. Students with feedback on Journal 1 may email questions.
– No eCourse access? Email the instructor so readings/PDFs can be forwarded; report when credentials arrive.
– Groups: Finalize your YouTube pathway diagram and ensure you have both Feed Snapshot 1 and Feed Snapshot 2. Be ready to share.
– Save artifacts: Keep your screenshots and note which account/device state you used so your follow-up clicks remain consistent next class.
Next Class (Thursday)
– Present diagrams: Be prepared to walk through your starting point, Videos 1–3 (and 4), and how your recommended feed evolved.
– Analysis focus: Differences between self-generated vs. socially generated conditions and between Pro vs. Con seeding.
Administrative/Onboarding
– Schedule an optional onboarding session for new/late-adding students (confirm time via email).
– Maintain a list of students still lacking eCourse access; continue forwarding PDFs until all are onboarded.
Technology/Logistics
– Prepare and archive the seed links used for socially generated groups (Pro and Con) for reference during analysis.
– Encourage students to use the same YouTube context (logged in/out, same browser) for continuity when resuming the experiment.
Homework Instructions:
ASSIGNMENT #1: Video Reflection Journal (deadline extended to Thursday)
You will record a short video reflection that connects this week’s reading (Cho et al. on affective polarization) with our in-class discussion and experiment on algorithmic recommendation systems (Instagram Reels/TikTok/YouTube) and the “walkable cities/15-minute city� debate. The purpose is to help you synthesize how platforms can comfort or challenge your views and how that may contribute to polarization.
Instructions:
1) Revisit the key ideas from today:
– Your own recent experience with algorithmic feeds (Reels/TikTok/YouTube) and whether they made you feel comforted or challenged.
– Cho et al.’s concept of affective polarization: how repeated exposure to agreeable content can harden attitudes toward opposing views.
– Our class experiment distinguishing “self-generatedâ€� vs. “socially generatedâ€� recommendations and how quickly YouTube began narrowing recommendations after 2–3 clicks.
– The walkable cities/15-minute cities topic as a live example of a polarizing debate.
2) Plan your reflection:
– Briefly summarize one takeaway from the reading (in your own words).
– Describe one concrete moment from your own feed use (e.g., last time you scrolled Reels/TikTok) and how it felt (comforting vs. challenging).
– Connect that experience to Cho et al.: How might recommendation systems contribute to affective polarization in your case?
– Use the walkable cities debate to illustrate: How might “proâ€� or “conâ€� content streams lead to different perceptions of the other side?
– If you participated in the YouTube mini-experiment today, mention what you saw change in your recommendations after 2–3 clicks. If you didn’t finish, you may try a quick replication on your own (watch ~30–45 seconds of a relevant video, refresh, click a recommended video on the same topic, repeat 2–3 times) and describe what the sidebar started to look like.
3) Record your video reflection:
– Keep it focused and clear. Aim to cover the points in Step 2.
– Speak in your own voice; prioritize insight over summary.
– Optional: Show or reference a screenshot from your recommendations if you tried the mini-experiment on your own.
4) Title and submission:
– Title your file or post with your name and “Reflection Journal – Affective Polarization.â€�
– Submit it to the Reflection Journal assignment page for this week.
5) Deadline:
– Due by Thursday at midnight (Bishkek time).
– If you already submitted on Sunday, you are all set and do not need to resubmit.
6) After you submit:
– Check back for feedback. If anything in the comments is unclear, email with your questions.