{"id":"1fefca43-ad39-4654-9999-c3e4466e9de9","arxiv_id":"2506.03687","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Visually impaired tram passengers in Linz experience accessibility as an emergent outcome of socio-technical interactions, with directional two-sense warning signals and varied technology comfort shaping their independence.","lead":"A mixed-method study of seven visually impaired tram riders in Linz, Austria, finds that accessibility arises from interactions between people, apps, infrastructure, and staff, not from the tram system alone. The paper is a compact, readable input for designers of inclusive public transport and the EU-funded OptiPex project.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Central finding is largely a restatement of the ANT framework; empirical support is not independent because the same lens guided data collection and coding.","rationale":"The reader identified the small, single-city sample as the weakest assumption and the theory-laden nature of the conclusion as a lesser concern (circularity burden 3). I agree that the sample limits generalizability, but the more load-bearing problem is that the central claim is effectively guaranteed by the method: the coding categories were derived from ANT, so the finding that accessibility emerges from socio-technical networks substantially restates the theoretical lens. This would persist even with a larger sample, and it undermines the paper's strongest claim more directly than the sample size alone. The unsupported 'strong correlations' and the unfulfilled shadowing analysis are concrete supporting evidence of the same underlying issue: the empirical record is thinner than the interpretive apparatus implies. The paper is still an honest, well-structured qualitative study whose specific user-level preferences (directional audio, tactile feedback, staff mediation) are useful for design. The concerns can be addressed by publishing the coding materials, reporting shadowing findings, softening the statistical language, and explicitly acknowledging the circularity risk in a limitations section. Under the AHFE conference format these revisions may be tight, but they are feasible; the verdict should remain CONDITIONAL pending those changes.","tokens_in":6093,"tokens_out":3389,"duration_ms":36894,"concrete_test":"Obtain the anonymized focus group transcript and shadowing field notes. Have two independent researchers who are blind to the research question and to Actor-Network Theory code the data using open, inductive thematic analysis. If the theme 'accessibility as relational/emergent' does not arise spontaneously—or if a simpler 'multiple independent barriers' theme dominates—then the central claim is an artifact of the pre-imposed ANT codebook. As a secondary check, tabulate the participant-level data behind the claimed 'strong correlations' and run a basic non-parametric test (e.g., Kendall's tau, n=7); if the correlation is not significant, remove the 'strong' language from Sections 3.2 and 3.3.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's principal claim—that accessibility emerges from relations among human and non-human actants rather than being an inherent property of transport systems—is built into the research design. In the Qualitative Data Analysis section, the authors state that initial coding categories were derived from ANT concepts, and Table 1 includes an explicit 'ANT analysis' code family with nodes such as 'Distributed Action and Networks,' 'Actants,' 'Delegation,' and 'Script and Inscription.' The analysis was therefore structured to rediscover the framework's assumptions, making the headline finding a near-tautological output of the coding scheme. The focus-group quotes supplied (e.g., preference for directional warning signals) are plausible user preferences, but they do not independently demonstrate that accessibility is inherently relational; they could equally support a conventional 'multiple barriers' interpretation. The paper also asserts 'strong correlations' between frequency of use and independence and between technology use and independence (Sections 3.2–3.3), yet no quantitative data, effect sizes, or statistical tests are provided for a sample of seven participants. In addition, shadowing is listed as a core method, but no analysis of shadowing observations appears in the results, so the promised mixed-method triangulation is not actually evidenced. These issues together mean that the central claim currently rests on a self-confirming analytic frame rather than on independently verifiable empirical support.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports a qualitative study of visually impaired tram passengers in Linz, Austria, using Actor-Network Theory (ANT) as a sensitizing framework. The authors conducted a shadowing exercise and a focus group with seven visually impaired participants and one companion, analyzed the material with qualitative content analysis, and report findings on network configuration, mobility patterns, technology integration, warning signals, and additional dimensions such as risk management and spatial navigation. The central claim is that accessibility is not an inherent property of transport infrastructure but emerges from the interactions of human and non-human actants, so interventions should target network interaction points such as platform-vehicle transitions and should follow a 'two-sense principle' combining directional audio and tactile feedback for warning signals.","tokens_in":6268,"tokens_out":3363,"duration_ms":37890,"significance":"If the empirical evidence supported it, the paper's relational framing of accessibility would be a useful corrective to purely infrastructural or purely technological design approaches in inclusive public transport, and the concrete recommendations about directional warning signals and platform-vehicle transitions are actionable. The manuscript has strengths: the code system is fully tabled, the ethical protocol is explicit, participant quotes are provided, and the qualitative methodology is standard for a short conference paper. However, as the study stands, the headline conclusion largely restates the ANT lens that was used to derive the initial coding categories, and the supporting evidence is weaker than the claims, so the contribution is currently more an interpretive application of ANT than an empirical demonstration of the framework's explanatory power.","major_comments":[{"comment":"The core conclusion that accessibility emerges from socio-technical networks is at risk of being circular. The text states that 'initial coding categories were derived from ANT concepts,' and Table 1 includes an explicit code family 'ANT analysis' with nodes such as 'Distributed Action and Networks' and 'Actants.' Because the analytic categories already embody the relational ontology of ANT, the empirical material cannot independently confirm that accessibility is relational; it can only illustrate the framework. The authors should either present the finding as an interpretive application of ANT rather than an empirical discovery, or show which inductive codes, if any, emerged independently of ANT and how those codes were validated. This is load-bearing because the paper's main contribution rests on the claim of empirical support.","section":"Qualitative Data Analysis; Table 1"},{"comment":"Sections 3.2 and 3.3 twice assert 'strong correlations' between frequency of use and enhanced independence and between technology use and enhanced independence. No quantitative data, correlation coefficients, effect sizes, or statistical tests are reported, and the sample of seven participants (with one companion) is far too small to support such claims. The authors should replace the word 'correlations' with qualitative patterns, or provide actual quantitative analysis if such data exist.","section":"Mobility Patterns and Transportation Usage; Technology Integration and Digital Solutions"},{"comment":"The methodology describes shadowing as a core component that 'complemented the focus group findings,' but the results section contains no analysis of the shadowing observations; the only related elements are the two photographs in Figure 1 with captions that are not coded or interpreted. The promised mixed-method triangulation is therefore absent. The authors need to report what the shadowing data showed, how they were analyzed, or explicitly state that the shadowing served only as orientation and that the results rest exclusively on the focus group.","section":"Shadowing"},{"comment":"The sample consists of seven visually impaired individuals and one companion, all recruited through a single advocacy association in a single city. The paper does not state this as a limitation, yet Sections 3 and 4 make general design and policy recommendations for public transport. The authors should add an explicit limitations paragraph that acknowledges the sample size, recruitment channel, and geographic scope, and should temper the generalizability claims accordingly.","section":"Participant Recruitment and Ethical Considerations; Conclusions"}],"minor_comments":[{"comment":"The sentence 'Coding was done using Mayring's (Mayring & Fenzl, 2019) qualitative content analysis methodology' is redundant; it should be 'Mayring and Fenzl's (2019) methodology.'","section":"Qualitative Data Analysis"},{"comment":"The in-text citation '(Boadi-Kusi et al., 2023)' does not match the reference list entry, which is dated 2024; please reconcile the year.","section":"References / Introduction"},{"comment":"The 'two-sense principle' is introduced in the results but not defined; a brief definition, such as 'redundant information via two different sensory modalities,' would improve clarity.","section":"Warning Signals"},{"comment":"The notation '(S03, pos. 43)' is used without explaining what 'pos.' refers to; please define it in the table notes or main text.","section":"Table 2"},{"comment":"The photographs in Figure 1 are not explicitly referenced in the body of the paper; please add a sentence that refers to them and states what they are intended to illustrate.","section":"Figure 1"}],"recommendation":"major_revision","confidential_remarks":"For the editor: the manuscript is a compact conference paper and the topic fits an HCI/accessibility venue, but the central claim is currently too close to the analytic framework from which the coding scheme was built. I recommend asking the authors to address the circularity, the unsupported 'correlation' language, and the missing shadowing analysis. These issues are fixable within the manuscript's scope, so I do not recommend rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a modest but genuine qualitative study. The useful part is the concrete user testimony from seven visually impaired tram riders in Linz — directional warning signals that indicate the origin of a hazard, the smart-cane idea, human assistance remaining necessary even with good apps, and the wide variation in comfort with technology. Those are practical inputs, and the authors present them with participant quotes and a fully tabled code system. The ethics and recruitment process look careful.\n\nThe main soft spot is exactly what the stress-test flags: the headline conclusion — accessibility emerges from socio-technical networks — is written into the coding scheme from the beginning. Table 1 includes an explicit 'ANT analysis' code family, and the methods say initial categories were derived from ANT concepts. So the results inevitably rediscover the framework. That doesn't make the individual findings wrong, but it does mean the central claim is a lens application rather than an independent empirical discovery. The reader's circularity burden of 3 seems right.\n\nThe other issues are also real but minor. 'Strong correlations' between frequency of use and independence, and between technology use and independence, appear twice with no quantitative support and n=7; that phrase should be toned down to 'patterns' or 'tendencies.' Shadowing is listed as a method, but the results section reads entirely like focus group output — no shadowing observations are analyzed, so the mixed-method triangulation isn't actually on the page. The sample is seven people plus one companion in one city, recruited through one association; that's fine for an exploratory study but should be stated as such. No transcripts or coding reliability checks are available, though for a conference paper that's more of a constraint than a flaw.\n\nThe circularity concern is the load-bearing one. The empirical content does support a conventional 'multiple barriers' reading just as well as an ANT reading. The authors should either present the ANT framing as an interpretive lens rather than a finding, or add some analysis that distinguishes the relational account from the barriers account.\n\nBottom line: it's a small, honest study worth knowing about if you work on accessible transport, but it's not a landmark. It deserves a serious referee — the data is real and the authors engage with the literature — though a good referee should push them to soften the correlation claims, report the shadowing if they have it, and reframe the conclusion as an interpretation rather than a discovery.","headline":"Small, honest qualitative study with useful specifics; the ANT-framed conclusion is partly built into the coding, so treat it as an interpretation rather than a discovery.","tokens_in":6899,"tokens_out":2438,"would_cite":false,"duration_ms":25368,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Accessibility of a tram system is an outcome of interactions among riders, devices, staff, and infrastructure, not a built-in property.","keywords":["actor-network theory","public transport accessibility","visual impairment","tram systems","socio-technical networks","two-sense principle","warning signals","assistive technology"],"falsifier":"A multi-city study using a larger, more diverse sample that finds no difference in boarding difficulty between stops with two-sense directional warnings and stops without them would falsify the design recommendation; evidence that accessibility is better predicted by an individual's degree of impairment than by network interaction points would undercut the central claim.","tokens_in":5833,"feed_emoji":"🚋","tokens_out":7700,"duration_ms":75092,"temperature":0.7,"pith_summary":"The paper sets out to show that tram accessibility for visually impaired passengers is not a fixed property of vehicles or stations, but an emergent effect of the relations among riders, staff, assistive devices, and infrastructure. The authors study the tram network in Linz, Austria, through one focus group and shadowing observations, and interpret the results with Actor-Network Theory. They identify four dimensions that shape accessibility: network configuration, mobility patterns, technology integration, and warning systems. Their central recommendation is to intervene at network interaction points, especially platform-vehicle transitions, and to design warning signals around a two-sense principle that combines directional audio with tactile feedback. If the claim holds, transport providers should shift from installing isolated accessibility features to reworking the moments where human and technical actors must align.","feed_headline":"Tram access is built in interactions, not infrastructure","feed_subtitle":"For visually impaired riders, access depends on staff, apps, and two-sense warnings at handoffs.","key_machinery":"The central machinery is Actor-Network Theory, a framework that treats people, devices, and infrastructure symmetrically as actants whose alignments produce outcomes. Within it, translation and obligatory passage points identify where a network must come together, while the concepts of script and mental model explain how technologies carry assumptions about users and how users learn around them. The two-sense principle is the paper's concrete design output: a warning should combine directional audio with tactile feedback so that hazard location is perceptible without sight.","core_discovery":"The study claims that accessibility emerges from the alignment of human and non-human actants in a socio-technical network rather than being an inherent quality of infrastructure. In the Linz tram setting, obligatory passage points such as ticket validation, boarding, and the platform-vehicle transition are where the network must successfully converge for a visually impaired person to travel. Digital apps translate visual information into audio, staff mediate between system rules and individual needs, and regular users build mental models through repeated practice, so accessibility is continuously produced and can fail at specific convergence points. On this basis the paper argues that warning signals should engage at least two senses, directional sound and touch, so riders can tell where a hazard is coming from.","pith_inferences":["Going beyond the paper's data, a direct test would compare boarding success for visually impaired riders in cities that install two-sense directional warnings at platform-vehicle transitions versus cities that keep conventional single-sense warnings.","The same interaction-point lens could be applied to other disability groups and transit modes, predicting that exclusion concentrates at gates, ticket machines, and vehicle-door handoffs rather than inside vehicles.","Accessibility audits informed by this view would map translation failures over time, such as announcements not matching a real-time board, rather than measuring physical compliance alone.","The paper suggests but does not test that mental models built through repeated travel can compensate for poor network alignment, which would explain why familiar users succeed where first-time users fail."],"forward_implications":["Redesigning platform-vehicle transitions should produce larger accessibility gains than isolated single-point fixes, because these are the obligatory passage points where the whole network must converge.","Warning signals at tram stops and crossings should be directional and use both acoustic and tactile modalities, so riders can locate the source of danger.","Digital departure and navigation apps should be treated as access tools that also support social inclusion, while designers account for wide variation in users' comfort and skill.","Human assistance should be designed into the system alongside technology, since verbal interaction with staff and other passengers remains a load-bearing part of the accessibility network.","Transport providers can use network analysis to find modest changes at convergence points that improve boarding and safety for all passengers."],"supporting_citations":[{"why":"Supplies the Actor-Network Theory concepts of actants, translation, and distributed agency that structure the analysis.","marker":"Latour (2005)"},{"why":"Introduces the notion of obligatory passage points used to identify critical convergence nodes such as platform-vehicle transitions.","marker":"Callon (1984)"},{"why":"Supplies the concept of script to explain how technologies inscribe assumptions about users and create inclusion or exclusion.","marker":"Akrich (1992)"},{"why":"Supplies the notion of mental models used to explain how repeated tram use builds navigation skills.","marker":"Norman (1988)"},{"why":"Frames disability as an outcome of relations between bodies and technologies, supporting the claim that accessibility is emergent.","marker":"Moser (2006)"},{"why":"Supports viewing accessibility as an emergent property of networked interactions rather than a fixed infrastructure characteristic.","marker":"Galis & Lee (2014)"},{"why":"Reconsiders infrastructures through a disability-studies lens, reinforcing the network-based account of access.","marker":"Velho (2021)"}],"fun_headline_variants":["Blind tram access is a socio-technical outcome, not a fixed feature","For visually impaired, tram access is made at every handoff","Two-sense warnings help blind tram riders navigate hazards","Tram accessibility for the blind is co-produced by humans and apps","Visually impaired tram passengers need aligned people, tech, and warnings"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The findings rest on the assumption that one focus group of seven visually impaired participants plus one companion, recruited through a single advocacy association in one city, represents the wider population of visually impaired tram users.","fun_headline_variants_meta":{"raw":{"variants":["Blind tram access is a socio-technical outcome, not a fixed feature","For visually impaired, tram access is made at every handoff","Two-sense warnings help blind tram riders navigate hazards","Tram accessibility for the blind is co-produced by humans and apps","Visually impaired tram passengers need aligned people, tech, and warnings"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000858,"raw_usage":{"total_tokens":3681,"prompt_tokens":860,"completion_tokens":2821,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":476,"completion_tokens_details":{"reasoning_tokens":2732}},"tokens_in":476,"tokens_out":2821,"duration_ms":24058,"temperature":1.0,"reasoning_tokens":2732,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T10:57:11.206603+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A multi-city study using a larger, more diverse sample that finds no difference in boarding difficulty between stops with two-sense directional warnings and stops without them would falsify the design recommendation; evidence that accessibility is better predicted by an individual's degree of impairment than by network interaction points would undercut the central claim.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the Actor-Network Theory concepts of actants, translation, and distributed agency that structure the analysis."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces the notion of obligatory passage points used to identify critical convergence nodes such as platform-vehicle transitions."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the concept of script to explain how technologies inscribe assumptions about users and create inclusion or exclusion."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the notion of mental models used to explain how repeated tram use builds navigation skills."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Frames disability as an outcome of relations between bodies and technologies, supporting the claim that accessibility is emergent."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supports viewing accessibility as an emergent property of networked interactions rather than a fixed infrastructure characteristic."},{"cited_title":"They’re changing the network just by being there","cited_arxiv_id":null,"evidence_quote":"Reconsiders infrastructures through a disability-studies lens, reinforcing the network-based account of access."}],"review_version":1}