{"id":"63840f6a-10ef-4e69-bb92-53da1e926f52","arxiv_id":"2507.14846","paper_version":1,"verdict":"REJECT","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"A small AR-learning study shows that inserting 4-second gaps in progressive sentence display improves recall during walking, but the claimed benefit of progressive over full-sentence presentation lacks a formal test.","lead":"The paper tests a 'progressive sentence' presentation on AR smart glasses, revealing sentence parts one at a time. It finds that adding timed gaps between words helps recall while walking, but the key comparison against full-sentence and word-only formats was only piloted and never formally tested.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The headline claim that progressive presentation enhances recall over full-sentence or word-only formats has no significant formal support: Study 1 varied only word-gap and mobility within the Progressive condition, and the format comparison rests on Pilot 2 (N=6) with non-significant differences.","rationale":"The reader's weakest assumption and my concern coincide: the claimed superiority of Progressive over Full and Word formats has no formal baseline, because the only controlled study varies factors inside the Progressive condition. I do not see a more fundamental internal flaw; Study 1 uses pseudowords, within-subject counterbalancing, and appears to apply appropriate repeated-measures/ART analyses. The problem is specifically that the evidence collected does not match the claim made in the abstract. Pilot 2 N=6 is too small to support a headline assertion, and self-reported preference is not a recall measure. The time-matching assertion in §4.4.5 is also unverified, but even if it were correct, the missing comparison would remain. I therefore agree with the reader's rejection: the paper would need a powered, time-matched comparison against Full and Word, ideally under walking, before the central claim can stand. My recommendation keeps the reader's verdict unchanged rather than moving it, because my concern supports the same conclusion rather than introducing a new one.","tokens_in":14303,"tokens_out":3562,"duration_ms":43853,"concrete_test":"Run a preregistered within-subjects experiment comparing three presentation formats—Progressive (with 4-s gaps, per Study 1), Full sentence, and random Word list—under both Sitting and Walking, with sample size determined a priori from the effect size Pilot 2 would require (N≈36–50 given the small pilot; use sequential testing if desired). Match total exposure time across formats exactly, per §4.4.5, and record timing logs from the GitHub implementation to verify equality. Primary outcomes: immediate Free-Recall and 7-day delayed Free-Recall. Pre-specify the contrasts Progressive > Full and Progressive > Word in the Walking condition, and report Sitting separately. If either contrast is not significant or the effect is negligible, the Abstract's central claim as written is unsupported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central claim—stated in the Abstract and Introduction—is that progressive presentation enhances recall compared with word-only and full-sentence learning, particularly while walking. Study 1 (§8) does not test that claim: its 2×2 design varies only Mobility and Word-Gap within the Progressive condition. The only formal comparisons are between Sitting and Walking and between Gap and NoGap; the significant finding is that a 4-s word gap improves immediate free recall during Walking (p_bonf < .05, d = 0.44), not that Progressive beats Full or Word. The format comparison rests entirely on Pilot 2 (§6), N=6, where 'differences were not statistically significant,' plus self-reported preference. Preference rankings and qualitative comments are not recall evidence. Section 4.4.5 asserts that 'all modalities and gaps were time-matched across conditions,' but the paper does not report the actual durations for the Full and Word conditions in Pilot 2, so even the descriptive trend could reflect unequal exposure. Pilot 4 (§9, N=4) repeats the same internal design with Norwegian and likewise lacks a format baseline. Thus the load-bearing assumption—that the pilot's descriptive advantage would reach significance in a powered comparison with time-matched controls—is untested. If that assumption fails, the headline claim has no controlled support. This is not an internal inconsistency in Study 1; it is a mismatch between the evidence collected and the claim made.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a 'progressive presentation' method for second-language vocabulary learning on augmented-reality smart glasses, in which a sentence is revealed component-by-component (subject, verb, object) while prior context is retained, and reports a sequence of pilots and one formal study. Study 1 (N=12) uses a 2x2 within-subject design (Mobility x Word-Gap) inside the Progressive format and finds a significant immediate free-recall benefit of a 4-second word gap in the Walking condition, along with overall higher recall when Sitting. The manuscript claims in the Abstract and Introduction that progressive presentation enhances recall over full-sentence and word-only formats, but that claim rests on Pilot 2 (N=6), whose format differences were not statistically significant. Pilot 4 (N=4) provides no format baseline.","tokens_in":14606,"tokens_out":5288,"duration_ms":52477,"significance":"The topic is timely and the apparatus is appropriate for studying AR microlearning. The authors use Wuggy pseudowords to control prior knowledge, standard recall measures with partial-credit scoring, a within-subject Latin-square design, and a 7-day delayed recall; these are strengths. However, the central comparative claim about progressive versus full-sentence and word-only learning is not supported by the data as reported: the only formal study that is statistically analyzed (Study 1) does not include those formats, and the only format comparison (Pilot 2) is underpowered and non-significant. The narrower finding that a 4-second gap improves immediate free recall during walking is plausible but depends on a gap duration chosen in a two-participant pilot. As submitted, the paper's headline overclaims its evidence.","major_comments":[{"comment":"The headline claim that progressive presentation enhances recall relative to full-sentence and word-only formats is not tested by Study 1. Study 1's 2x2 design varies only Mobility (Sitting/Walking) and Word-Gap (NoGap/Gap) while all conditions use the Progressive format; there is no Full or Word condition. The only formal comparison across formats is Pilot 2 (Section 6), N=6, which reports that 'differences were not statistically significant.' The abstract's statement that 'Pilot and formal studies revealed that progressive presentation enhances recall' therefore overstates the evidence, and the preference and qualitative feedback in Section 8 do not constitute recall evidence. This is a load-bearing mismatch between claim and design.","section":"Abstract; Section 8; Section 6"},{"comment":"Section 4.4.5 asserts that 'all modalities and gaps were time-matched across conditions,' but no exposure durations are reported for the Full and Word conditions in Pilot 2. If the Full or Word condition presented the same sentence components for more or less total time, the descriptive advantage for Progressive could be an artifact of exposure time rather than format. The manuscript should report the actual per-item and per-condition durations for each format and, ideally, verify that the Word condition presented the same number of words for the same total duration before treating the Pilot 2 trend as meaningful.","section":"Section 4.4.5; Section 6"},{"comment":"The 4-second gap that is central to the significant interaction in Study 1 was itself selected from Pilot 3 (N=2) and then tested in the same research program, so Study 1 is a confirmation of that pilot's choice rather than an independent test of gap duration. In addition, the significant gap benefit appears only in immediate free recall during Walking (p_bonf < .05, d = 0.44); no gap effect is reported for delayed recall, and the gap had no effect in Sitting. The abstract's claim that 'incorporating timed gaps between word presentations further improved learning effectiveness' should be qualified to immediate recall under multitasking and treated as exploratory given the parameter-selection dependence.","section":"Section 7; Section 8"},{"comment":"Pilot 4 (N=4) repeats the same within-Progressive design with Norwegian materials and, like Study 1, does not include Full or Word conditions. It therefore cannot provide the format baseline needed to support the central claim. Moreover, Section 10's discussion of 'why progressive presentation is effective' relies heavily on subjective preference and interview responses, which are not measures of learning outcome. The manuscript should either confine its conclusions to the within-format gap and mobility effects or collect a powered, time-matched format comparison.","section":"Section 9; Section 10"}],"minor_comments":[{"comment":"The participant count is inconsistent: Section 4.1 says 'Participants (total of 27),' but the individual studies sum to 5+6+2+12+4=29, and Section 4.1 also states that no participant took part in more than one study. Please reconcile the total.","section":"Section 4.1"},{"comment":"The caption of Figure 4(b) says the maximum score for each measure is 8 for most measures, but Pilot 4 has 7 new words per condition, so the maximum for Free-Recall and Word-Recall should be 7; the caption in Figure 3 is correct for Study 1 but not for Pilot 4.","section":"Figure 4 caption"},{"comment":"The presentation schedule is described only qualitatively ('at least 6 seconds,' '8 seconds,' '12 seconds,' 'shown twice'); a concrete per-condition timing table for the NoGap and Gap variants of Progressive, and for Full and Word, would make the time-matching claim verifiable.","section":"Section 4.4.5"},{"comment":"The main effect of Mobility is reported only as p < 0.05 without test statistics for most recall measures; please report exact F values, degrees of freedom, and effect sizes for all significant main effects and interactions.","section":"Section 8"}],"recommendation":"reject","confidential_remarks":"The central comparative claim is not supported by the experimental design; the only statistically analyzed study omits the format factor entirely, and the format comparison is an underpowered, non-significant pilot. This cannot be fixed by rewriting alone and would require a new, adequately powered experiment. If the venue is open to preliminary results, a revised manuscript restricted to the within-format gap and mobility effects might be reconsidered, but as submitted the abstract overclaims."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Hi [Colleague],\n\nQuick take on arXiv:2507.14846. The one solid, new empirical finding is that inserting a 4-second gap between words in a progressive sentence display improves free recall during walking (d=0.44, p_bonf<.05) and not when sitting. That's a clean interaction, the direction matches the qualitative feedback, and it's the kind of concrete design guideline AR learning apps can use.\n\nThe paper's bigger claim, that the progressive format itself beats full-sentence and word-only learning, is not supported by the evidence they collected. Study 1 only varies mobility and gap inside the Progressive condition; there is no formal baseline with Full or Word. The only formal comparison between formats is Pilot 2, N=6, where they candidly report no significant differences, plus self-reported preference. The abstract and introduction assert the stronger claim anyway.\n\nOther soft spots: the 4-second gap was selected from a 2-person pilot and then confirmed in Study 1, so there's mild circularity in the parameter choice. Section 4.4.5 says all conditions were time-matched but doesn't report actual durations for Full and Word, so the reader can't fully verify the descriptive trend in Pilot 2. The delayed recall results are non-significant, and the authors acknowledge this.\n\nNone of this is fatal to the paper's actual content. The gap-by-mobility interaction is a legitimate, falsifiable finding with a moderate effect size. The paper is honest about its limitations in Section 10.1, and the materials and procedure are more carefully controlled than much of the microlearning literature (pseudowords, matched sentence lengths, counterbalancing). The problem is a classic claim-evidence mismatch: they had a modest result and dressed it as a format comparison.\n\nWho's it for? Designers of AR or mobile microlearning apps, and researchers working on presentation timing under multitasking. The gap finding is worth knowing even if the format claim should be ignored until a proper powered comparison is run.\n\nFor peer review: I'd send it, but with clear instructions that the abstract and discussion need to be scaled back to what Study 1 actually tests, or a format-comparison experiment needs to be added. It's a short adjunct paper; a revision is feasible.\n\nBest,\n[Your name]","headline":"The gap effect during walking is real and worth reporting, but the abstract's claim that progressive beats word and full-sentence learning is not backed by any formal comparison.","tokens_in":15125,"tokens_out":2799,"would_cite":false,"duration_ms":26357,"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":"A progressive sentence-reveal format with timed pauses improves recall in mobile AR language learning.","keywords":["progressive presentation","second-language acquisition","augmented reality smart glasses","mobile microlearning","word-gap timing","multimodal learning","cognitive load","free recall"],"falsifier":"A preregistered, adequately powered study (for example, $N \\geq 60$) comparing progressive, full-sentence, and word-only formats during walking with equal total exposure time would settle the claim: if the progressive format does not produce significantly higher free recall than both alternatives, the headline claim fails. A second check is to remove the final full-sentence display from the progressive sequence; if recall then drops to word-only levels, the benefit would come from the appended sentence rather than from stepwise disclosure.","tokens_in":14127,"feed_emoji":"🥽","tokens_out":7991,"duration_ms":81192,"temperature":0.7,"pith_summary":"This paper tries to establish that progressive presentation is a better way to teach second-language words on AR smart glasses than showing the whole sentence at once or showing words in isolation. In progressive presentation, a sentence appears one component at a time (subject, then verb, then object) while the earlier components stay visible and the new one is highlighted, after which the full sentence is shown. The intended practical payoff is that short, distracted moments like walking can become effective vocabulary-learning time. The formal study shows that inside this format, inserting a four-second pause between components significantly raises immediate free recall while walking, with no such benefit while sitting; the comparison against full-sentence and word-only formats comes from pilots and user preference rather than from the formal experiment.","feed_headline":"Paced sentence reveals boost word recall while walking","feed_subtitle":"AR-glasses study: 4-second pauses between sentence parts improved recall during walking.","key_machinery":"The central mechanism is progressive disclosure applied to sentence structure: a sentence is segmented into subject, verb, and object components that appear sequentially, earlier components remain visible, and the newest component is signaled by a transparency cue, followed by a final full-sentence display. This follows the multimedia-learning principles of segmenting, breaking content into small progressive units, and signaling, highlighting new elements, which are meant to manage cognitive load. All modalities (L2 text, L1 text, images, synchronized audio) were time-matched across conditions so that total exposure time was equal, isolating the format and the word-gap timing as the manipulated variables.","core_discovery":"Progressive presentation combines the advantages of word and sentence learning: it gives novice learners the one-to-one word mapping of isolated vocabulary while preserving the grammatical context of a full sentence. In the 2×2 formal study, walking participants who saw sentence components separated by a 4-second gap recalled more L2–L1 word pairs freely ($M = 4.7$, $SD = 2.2$) than those without a gap ($M = 3.1$, $SD = 1.8$; $p_{\\mathrm{bonf}} < 0.05$, $d = 0.44$); sitting participants showed no gap benefit. The authors interpret the gap as giving learners time to consolidate each word before the next appears, which matters most when attention is divided between walking and learning. Most participants also preferred progressive over full-sentence and random-word formats, reporting that preserving sentence context helped them connect L2 words to their L1 meanings.","pith_inferences":["A direct powered comparison of progressive, full-sentence, and word-only formats, matched for exposure and with the progressive condition's final full-sentence display removed, would test whether the benefit comes from stepwise disclosure itself or from the appended sentence replay.","The gap effect predicts a testable interaction with task load: making the walking task harder should shift the optimal gap duration, suggesting an adaptive-pacing design space beyond the fixed 4 seconds used here.","The same progressive-disclosure pattern could be applied to other sequential learning content such as step-by-step instructions or short narratives, and a similar walking-versus-sitting comparison would show whether the effect is general or specific to subject-verb-object sentences."],"forward_implications":["Designers of AR language apps can expect that inserting short pauses between sentence components will improve immediate recall specifically for learners who are mobile, such as walking.","The progressive format is best positioned as a scaffolding technique: it should be offered to novice learners and gradually replaced by full-sentence presentation as word familiarity grows.","Pacing should be treated as adaptive rather than fixed, since participants reported that the ideal gap duration depends on how demanding the concurrent task is.","Because the 7-day delayed-recall measures showed no significant effects, the demonstrated advantage is in short-term retention rather than long-term memory.","Time-matched multimodal presentation with text, image, and audio together was preferred and performed at least as well as text-only, supporting the multimedia principle on see-through glasses."],"supporting_citations":[{"why":"Supplies the segmenting and signaling principles from multimedia learning theory that motivate revealing sentence components one at a time.","marker":"[31]"},{"why":"Provides perceptual load theory, used to explain why attention is taxed during walking and why pauses help consolidation.","marker":"[26]"},{"why":"Shows that persistent sequential delivery on head-mounted displays improves recall in mobile video learning, the pattern progressive presentation extends to sentences.","marker":"[34]"},{"why":"Grounds the choice of a 4-second gap by analogy to notification resumption gaps during multitasking.","marker":"[22]"},{"why":"Establishes the AR vocabulary-learning baseline that the paper extends from single words to sentence-level presentation.","marker":"[19]"},{"why":"Provides the pseudoword generator used to construct controlled L2 materials.","marker":"[24]"}],"fun_headline_variants":["Paced sentence parts boost recall while walking","Gap-timed sentence chunks improve word recall on the go","Progressive sentence display aids memory for walkers","Walking learners recall more with sentence context","AR study: spaced sentence pieces enhance mobile learning"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The headline claim that progressive presentation beats full-sentence and word-only learning rests on a six-person pilot with no significant differences plus self-reported preference, because the formal experiment only varied mobility and word-gap within the progressive format itself.","fun_headline_variants_meta":{"raw":{"variants":["Paced sentence parts boost recall while walking","Gap-timed sentence chunks improve word recall on the go","Progressive sentence display aids memory for walkers","Walking learners recall more with sentence context","AR study: spaced sentence pieces enhance mobile learning"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000222,"raw_usage":{"total_tokens":1420,"prompt_tokens":881,"completion_tokens":539,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":497,"completion_tokens_details":{"reasoning_tokens":468}},"tokens_in":497,"tokens_out":539,"duration_ms":7039,"temperature":1.0,"reasoning_tokens":468,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T15:45:59.311834+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A preregistered, adequately powered study (for example, $N \\geq 60$) comparing progressive, full-sentence, and word-only formats during walking with equal total exposure time would settle the claim: if the progressive format does not produce significantly higher free recall than both alternatives, the headline claim fails. A second check is to remove the final full-sentence display from the progressive sequence; if recall then drops to word-only levels, the benefit would come from the appended sentence rather than from stepwise disclosure.","supporting_citations":[{"cited_title":"de Fockert, and Essi Viding","cited_arxiv_id":null,"evidence_quote":"Provides perceptual load theory, used to explain why attention is taxed during walking and why pauses help consolidation."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Shows that persistent sequential delivery on head-mounted displays improves recall in mobile video learning, the pattern progressive presentation extends to sentences."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes the AR vocabulary-learning baseline that the paper extends from single words to sentence-level presentation."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the pseudoword generator used to construct controlled L2 materials."}],"review_version":1}