Sludge Content: What the Attention Evidence Actually Measures

Does unrelated gameplay make an explanation easier to follow, or simply add another thing to watch? The available studies ask narrower questions than the sweeping claims attached to this format. Start with what each experiment actually measured.

Name the format before judging it

Use concurrent unrelated panels for a narration or primary clip paired with gameplay, slime, cutting, or another moving image that does not explain the message. That is different from a supporting visual: a labeled diagram, demonstration, caption, example, or screen recording selected because it helps the viewer understand the same claim. Motion alone does not make a visual relevant. This distinction matters because a test of unrelated gameplay cannot establish that every split screen is harmful, and a test of captions cannot validate decorative footage.

Keep three outcomes separate

Immediate comprehension asks what a viewer understood just after one clip. Sustained attention research asks how someone performs on a separate vigilance task after exposure. Instagram retention describes viewing behavior on a particular Reel, such as watch time or a retention curve when the account exposes one. These are not interchangeable outcomes. A format could accompany acceptable recall in a classroom experiment yet lose viewers at a commercial handoff; a post could hold views while leaving the message misunderstood. None of those observations proves long-term attention change.

Read the direct sludge experiment narrowly

A 2026 peer-reviewed research article exposed 48 adults aged 22–25 to ten minutes of media or no media, then compared pre- and post-tests (Annals of Neurosciences). The 45 medium-to-high media users were randomized across sludge, regular short-video, and no-media groups of 15; only three low-media users formed a separate sludge group. The sludge group had the largest decline on the continuous-performance attention task, while the analysis did not detect a between-group working-memory difference. Crucially, the adjusted pairwise sludge-versus-regular comparison was not statistically significant. The study measured an immediate laboratory task after a short exposure—not Reel retention, addiction, ADHD, dopamine, brain damage, or lasting impairment.

Use the thesis as a comprehension clue

A 2026 Stockholm University master’s thesis, not a peer-reviewed journal article, randomized 54 Swedish students aged 15–16 to audio-only, captioned, or unrelated-gameplay versions of a three-minute English listening exercise (Stockholm University DiVA). Both visual conditions produced higher immediate multiple-choice comprehension than audio-only, and students reported better focus; the abstract reports no significant difference between the two visual conditions. That result is useful because it resists a simple “extra motion always ruins comprehension” story. It still concerns one school, second-language listening, one short lesson, and immediate testing—not Instagram behavior or durable learning.

Treat context switching as adjacent evidence

A 60-person experiment published at CHI and available through an arXiv author copy compared TikTok, Twitter, YouTube, and no activity during a prospective-memory task (Chiossi and colleagues). Performance was worse in the TikTok condition, which the authors connected to short video and rapid context switching. It did not compare a split-screen Reel with a single-panel version, so it cannot isolate gameplay, slime, captions, or a borrowed opener. Use it to frame a question about switching demands, not as direct proof about sludge layouts.

Choose visuals by the message job

Before publishing, label each panel evidence, orientation, access, or unrelated stimulus. Keep a supporting panel when it shows the claim, locates the next step, or makes speech accessible. Test an unrelated panel only against a clean version with the same narration, duration, caption, and call to action. Review available retention alongside one comprehension check and one relevant action; do not let views stand in for understanding. The present evidence supplies different short-term outcomes and small samples, not a guarantee that either layout will win.