{"id":"b9e4b6f0-2adc-44c7-999a-739e50053a32","arxiv_id":"2508.19768","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Routing posts from small trusted teams into larger channels through member votes, with thresholds that scale with channel size, produced active collaborative curation and reduced posting hesitancy in a 36-person, ten-day field study.","lead":"Burst is a social media design where posts start in a user's small trusted circle, then the circle votes to push posts into larger channels, with bigger channels requiring more votes. A ten-day test with 36 people suggests participants actively curated each other's content and reported less fear about posting.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Mandatory onboarding and homogeneous campus sample make it impossible to attribute observed bursting to the design; the field study lacks any baseline or control to rule out protocol-driven engagement.","rationale":"The reader identifies the protocol-mandated onboarding and the homogeneous campus sample as the weakest assumption, and I agree: this is the most load-bearing concern for the central claim. The paper's own Figure 4 caption asserts that bursting itself was not required, which partially mitigates the concern; however, the required team-building and channel-joining still create the social structure and notification-driven context in which bursts occur, and the single-university sample means the trust loop may not generalize. The design concept is coherent, the qualitative analysis is used fairly, and the authors are unusually candid about limitations; this is not a rejection. The field study is a reasonable feasibility demonstration, but the abstract's 'demonstrate' should be softened to 'suggest' or 'provide initial evidence for' unless log reanalysis or a follow-up comparison rules out the protocol account. Since the reader already issued CONDITIONAL, my read does not change the verdict. I also note the reported numeric inconsistencies (Figure 4 caption vs Table 2, 'ten days' vs 'approximately one week'); they are secondary but should be corrected before publication.","tokens_in":823,"tokens_out":886,"duration_ms":111895,"concrete_test":"Obtain the de-identified activity logs and recompute the Figure 4 / Table 2 burst statistics. Then build a per-user daily burst time series for days 1-10 and mark the day each user completed mandatory onboarding (three team invites accepted plus three channel joins). Test whether burst rate per user-day after onboarding is significantly lower than during onboarding and whether there is a decreasing day-level trend. If burst activity persists after onboarding with no steep decay, the protocol-driven explanation is weakened; if bursts collapse once requirements are satisfied, the central claim fails. A secondary split by whether users exceeded the minimum team/channel counts would further separate voluntary engagement from compliance.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that Burst enabled a participatory curation culture requires that observed burst activity and felt safety are attributable to the burst mechanism rather than to the study protocol. Section 4.2 makes that attribution insecure: each participant was required to invite at least three users as team curators and to join at least three channels. The trusted-team-plus-multi-channel structure was therefore manufactured by the study, not voluntarily adopted, and the paper reports no baseline or control arm that would separate the design's effect from mandated onboarding, compensation, novelty, or demand characteristics. Section 6 further concedes that the sample was one university, culturally homogeneous, often with pre-existing ties ('everything was close to home'); the trust loop that makes bursting feel safe may be an artifact of that familiarity. If mandatory onboarding and campus cohesion rather than thresholded peer curation produced the 75% burst participation and the positive safety ratings, the abstract's 'demonstrate' outruns the evidence. The limitations discussion acknowledges the generalization gap but does not test it.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces Burst, a social media design in which posts are initially visible only to a user's small trusted team and are then routed to progressively larger channels only after enough team members 'burst' the post. A mobile app implements the design, and a ten-day field study (N=36) collects activity logs, pre/post surveys, and 16 interviews. The reported behavioral findings are that 75% of participants burst at least once, 297 burst actions occurred, and only 36% of posts remained solely within the poster's team; interview and survey data are used to argue that Burst lowered posting anxiety, supported audience-appropriate sharing, and created a meaningful middle layer between private groups and public squares. The paper frames this as demonstrating a participatory curation culture and discusses limitations around scale, duration, and sample homogeneity.","tokens_in":27816,"tokens_out":4339,"duration_ms":56921,"significance":"The design contribution is genuinely interesting: Burst operationalizes a middle ground between small private groups and large public feeds, and the multi-channel threshold mechanism is a concrete, implementable alternative to retweet/algorithmic distribution. The study is a genuine field deployment rather than a lab simulation, and the behavioral logs are useful evidence that the interaction is learnable and voluntarily used: 75% of participants burst despite no requirement to do so, and the distribution of actions in Table 2 shows bursting sits between reacting and replying. The qualitative analysis is extensive and well-integrated. However, the paper's central claim that Burst 'enabled' a participatory curation culture goes beyond what the design of the study can support, because the trusted-team-plus-multiple-channel structure was mandated by the onboarding protocol and no baseline or control condition isolates the burst mechanic. The authors candidly acknowledge the homogeneity of the sample ('everything was close to home', Section 6), but they do not test whether the trust loop depends on pre-existing ties. These issues are load-bearing for the abstract's 'demonstrate' but ar","major_comments":[{"comment":"The central evaluative claim is underdetermined by the study design. Section 4.2 states: 'we required participants to invite at least three users to join their team as curators and to join at least three channels based on their interests.' This mandate manufactures the exact structural precondition for bursting (a small trusted team that can route content into multiple channels), and there is no baseline or control arm in which users could post directly to channels or in which team formation was not required. The 75% burst participation reported in Section 5.1 is therefore evidence that the protocol can produce burst usage, not that the burst mechanic, rather than the mandated structure, novelty, or compensation, enabled a participatory curation culture. I recommend either (a) reframing the contribution as a feasibility and usage study of a design concept, (b) adding a minimal control co","section":"Section 4.2 and Section 5.1"},{"comment":"The attitudinal claims rest on inconsistently reported denominators. Section 4.1 says 36 participants completed both pre- and post-study surveys, but percentages such as 72.41%, 68.97%, and 65.52% in Section 5.1 imply a denominator of 29. Sections 5.2 and 5.3 report percentages such as 63%, 47%, 53%, and 79% without stating the number of respondents or whether these are item-level response counts. This makes the survey-based claims about comfort, audience match, and satisfaction impossible to verify. Please report item-level Ns, missing-data patterns, and exact survey items (e.g., in an appendix), and treat the percentages as estimates with appropriate uncertainty.","section":"Sections 5.1-5.3 (survey percentages)"},{"comment":"The paper explicitly concedes that the participant pool was culturally homogeneous, drawn from a single institution, and that team members often had pre-existing ties ('everything was close to home'). This is not merely a generalizability caveat: the core mechanism of Burst is that users feel safe posting to a trusted team and that the team's curation is prosocial. If the trust loop is produced by shared campus context and swift trust rather than by the thresholded burst design, the qualitative evidence of 'safe posting' (Section 5.2) does not transfer to the broader settings the paper motivates (Section 1). The limitation is acknowledged but not tested. At minimum, the paper should include an analysis comparing participant behavior or attitudes for teams composed of known peers versus those formed purely through the onboarding process, or explicitly restrict the contribution claim to th","section":"Section 6 (limitations)"}],"minor_comments":[{"comment":"The caption says 'On average, participants generated a maximum of 40 bursts, with an average of 15.79 bursts per day.' Given 297 total burst actions over ten days and 36 participants, the aggregate daily rate is about 0.8 bursts per participant per day. Please clarify which population and time unit this number refers to, or correct it.","section":"Figure 4 caption"},{"comment":"The thematic analysis is described as performed by one author on the first five transcripts, with codes then refined. Please report whether other authors coded transcripts independently, and if so, provide agreement metrics. If the analysis was single-coded throughout, say so explicitly.","section":"Section 4.3"},{"comment":"The first paragraph says the evaluation focused on 'approximately one week,' but the study is described elsewhere as ten days. Align the wording.","section":"Section 6"},{"comment":"The figure shows a burst progress of '2/1' for a channel that requires one burst; this is visually confusing because progress exceeds the threshold. Use an example where the numerator is below or at the threshold unless intentionally illustrating an already-burst state.","section":"Figure 3(e)"},{"comment":"Reference [21] (Dzimianski et al. on ISG15) appears unrelated to the sentence about WhatsApp/Discord/private communication. Please verify the citation.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The paper fits CSCW well and the design idea is worth publishing, but the evaluation's causal claim needs reining in. The missing baseline is common in social computing systems research, but the mandatory onboarding makes the attribution problem acute. I would push the authors to either add a control condition/analysis or explicitly reframe the contribution as a feasibility demonstration. The survey denominator issue should also be fixed before acceptance."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick verdict: Burst is a real design idea, clearly described, and the paper deserves peer review. The synthesis—private team seeding plus anonymous thresholded voting to route content outward and inward—is genuinely not in the prior work they cite, and the ten-day study shows people can learn and use the mechanic. The interview material gives a plausible account of why it felt safe, and the paper is unusually candid about its limits. I trust that.\n\nThe soft spots are real but mostly about the gap between claim and evidence. The abstract says \"demonstrate... participatory curation culture,\" but the study is a single-university, 36-person field test with no baseline or control. Mandatory onboarding (invite three team curators, join three channels) means the trusted-team structure was manufactured, not voluntarily adopted. That's not fatal for a design paper, but it does undercut the \"culture\" language. The authors acknowledge the homogeneity explicitly, which helps.\n\nThere are also reporting inconsistencies: Figure 4's caption and Table 2 disagree on burst counts, the study is \"ten days\" in the abstract and \"approximately one week\" in the discussion, and survey denominators shift between 36 and 29. Minor, but a referee should ask for a pass that reconciles them. No code, data, or instruments are shipped; for a design paper in this venue that's a disappointment, not a disqualifier.\n\nThe stress-test note that attribution is \"impossible\" is close to fair, but I'd soften it to \"hard.\" The engagement data are consistent with the design being usable; they just can't show the design caused the engagement. The paper's own discussion does not overclaim much, so I'd call it a soundness problem in the abstract, not a load-bearing flaw.\n\nBottom line: this is work for social computing design researchers, especially those looking for alternatives to the public-square/private-group binary. It deserves peer review, with the expectation of a major revision that tightens claims and stats. I'd cite it if I were working on similar mechanisms, and it would make a good reading-group discussion about design evidence.","headline":"A genuine design contribution that deserves a serious referee, but the field study is a feasibility demo rather than a demonstration: the abstract overreaches, the stats are sloppy, and the protocol itself manufactured the trust structure.","tokens_in":28461,"tokens_out":1937,"would_cite":true,"duration_ms":23762,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Burst: let trusted friends route posts to the right audience","keywords":["social media design","collaborative curation","bursting","trusted teams","audience selection","field study","online communities","psychological safety"],"falsifier":"Run a replication where team formation and channel-joining are optional rather than required, with participants recruited from different universities or from an anonymous online population, and measure whether the 75% burst rate and 36%-stays-in-team figure reproduce. If burst actions drop to near-zero once the requirement to invite three curators is removed, or if posts almost never leave the starting team when teammates are strangers, the central claim that the design creates participatory curation would be contradicted.","tokens_in":27486,"feed_emoji":"📣","tokens_out":5761,"duration_ms":56740,"temperature":0.7,"pith_summary":"This paper argues that social media's split into private group chats and public squares is not inevitable, and introduces Burst, a design where every post first appears only to a small trusted \"team.\" The team—and later other members—can \"burst\" the post, each burst counting as an endorsement toward a channel-specific threshold; a post appears in a larger channel only after enough endorsements. A ten-day field study with 36 participants found that 75% of participants burst at least once, an average of 8.14 bursts per person, and only 36% of posts stayed inside their starting team. Survey and interview data report that the trusted-team start lowered hesitancy to post and that participants treated bursting as a meaningful middle tier of contribution between reacting and replying. The authors' claim is that this structure creates a participatory curation culture, with content flowing between many semi-public communities rather than only between private and fully public extremes.","feed_headline":"Burst: let trusted friends route posts to the right audience","feed_subtitle":"A 10-day, 36-person field study suggests peer-curated 'bursting' can fill the gap between private chats and public squares.","key_machinery":"The core mechanism is the burst threshold: each channel has a size-proportional threshold, and a post appears in that channel only after that many distinct users have endorsed it through the burst action. Around it sit four supporting features: \"Your Team,\" a private reviewer group where all posts originate; anonymous bursts, which reduce social pressure on curators; poster-controlled channel suggestions and blocks, which give the author influence over where content may travel; and a #everyone channel with the highest threshold, representing full public reach. The threshold turns cross-posting from a unilateral, norm-violating act into a collective vote, and it is the piece that carries the","core_discovery":"The central claim is that curation can be a first-class social media action rather than an invisible algorithmic process. In Burst, content is not posted directly to a global feed: it begins visible only to the author's trusted team, and viewers nominate it outward by bursting, with each channel requiring a threshold of endorsements before it will show the post; larger channels require more endorsements. The field study's headline numbers are that 27 of 36 participants burst at least one post, participants performed 297 burst actions over ten days, and 64% of posts burst into at least one channel beyond the starting team. Qualitative data adds the mechanism's intended effect: posters said th","pith_inferences":["One test the paper leaves open is whether the trust loop survives cold-start conditions: in this study participants came from a single institution and often knew each other, so a replication with anonymous strangers would tell whether the burst mechanic itself, rather than offline familiarity, drives curatorial participation.","The anonymity of bursts could hide reciprocal exchange: several participants described bursting as \"I help yours, you help mine,\" but the paper does not analyze whether bursts form reciprocal pairs; a network-level test could check whether curation becomes a gift economy or a one-way broadcast.","Threshold policy is a tunable control knob the study only samples at one setting; varying thresholds experimentally could map the trade-off between reach (low thresholds) and content filtering (high thresholds), and might reveal a tipping point where channels either starve or flood.","Ten days is short enough that novelty effects could inflate engagement; a longer deployment would be needed to see whether participatory curation persists after the initial excitement of a research app wears off."],"forward_implications":["Social media platforms can support a middle layer of semi-public communities, letting content graduate from trusted groups to bounded larger audiences without jumping straight to viral public exposure.","Curation becomes a mainstream user behavior with an intermediate effort level: more involved than a reaction, less involved than a reply, and available to people who do not create content themselves.","Cross-posting is reframed as a positive, expected act, potentially reducing the stigma and apology culture around sharing content across communities.","The burst mechanic is portable: the same threshold-and-endorsement structure can be embedded in existing space-based platforms such as Slack, so the design does not require a full new network to be tested.","If deployed at scale, burst thresholds need to be hardened against coordinated manipulation, for example through diversity-aware threshold algorithms, a direction the paper explicitly proposes."],"supporting_citations":[{"why":"supplies the Grudin's Paradox framing—diffuse benefits, immediate personal costs—that Burst's design is meant to overcome","marker":"[30]"},{"why":"provides the Cura algorithm that the paper adapts to make burst thresholds resistant to coordinated manipulation","marker":"[33]"},{"why":"contributes the context-collapse and imagined-audience concepts that motivate starting posts in a trusted team","marker":"[52]"},{"why":"supports the assumption that content consumers are also motivated to curate and share","marker":"[8]"},{"why":"defines the threaded-space design space and community metaphors that position Burst among social media architectures","marker":"[77]"},{"why":"justifies the field-study methodology as an accepted evaluation strategy for social computing systems","marker":"[7]"},{"why":"documents the small-group versus public-square polarity in current social media that Burst responds to","marker":"[57]"},{"why":"grounds the expectation that small online communities motivate participation, which the trusted-team design builds on","marker":"[34]"}],"fun_headline_variants":["Burst: friends decide what spreads","Peer-curated posts from private to public","Curation by community, not algorithm","Burst: from trusted circles to wider audiences"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"The load-bearing premise is that the engagement measured in this ten-day, single-campus study—where everyone was required to invite at least three people to their team and to join three channels, and where many participants already knew each other—reflects the burst design itself rather than the mandatory setup or pre-existing trust.","fun_headline_variants_meta":{"raw":{"variants":["Burst: friends decide what spreads","Peer-curated posts from private to public","Curation by community, not algorithm","Burst: from trusted circles to wider audiences"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001078,"raw_usage":{"total_tokens":4302,"prompt_tokens":654,"completion_tokens":3648,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":398,"completion_tokens_details":{"reasoning_tokens":3594}},"tokens_in":398,"tokens_out":3648,"duration_ms":32594,"temperature":1.0,"reasoning_tokens":3594,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T15:28:47.570889+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run a replication where team formation and channel-joining are optional rather than required, with participants recruited from different universities or from an anonymous online population, and measure whether the 75% burst rate and 36%-stays-in-team figure reproduce. If burst actions drop to near-zero once the requirement to invite three curators is removed, or if posts almost never leave the starting team when teammates are strangers, the central claim that the design creates participatory curation would be contradicted.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"supplies the Grudin's Paradox framing—diffuse benefits, immediate personal costs—that Burst's design is meant to overcome"},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"provides the Cura algorithm that the paper adapts to make burst thresholds resistant to coordinated manipulation"},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"contributes the context-collapse and imagined-audience concepts that motivate starting posts in a trusted team"},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"supports the assumption that content consumers are also motivated to curate and share"},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"defines the threaded-space design space and community metaphors that position Burst among social media architectures"},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"justifies the field-study methodology as an accepted evaluation strategy for social computing systems"},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"documents the small-group versus public-square polarity in current social media that Burst responds to"},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"grounds the expectation that small online communities motivate participation, which the trusted-team design builds on"}],"review_version":1}