{"id":"d9bac283-3e13-42d5-a99b-52e9271d1717","arxiv_id":"2507.09988","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"A new poll of 40 physicists at four Czech institutions finds the Copenhagen interpretation remains the most preferred, similar to earlier polls.","lead":"This paper reports a new survey of physicists at four Czech institutions asking which interpretation of quantum mechanics they prefer. It finds Copenhagen remains the most popular choice and argues that preferences have been fairly stable since the 1990s.","discovery_kind":"replication","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The stability claim rests on comparing a 40-person Czech volunteer sample to earlier polls, but the paper's own no-preference rate (53% in Q9) diverges sharply from the 36% in Sivasundaram & Nielsen, so sample comparability is not established and the 8-minute cutoff is untested.","rationale":"The reader identified sample comparability as the weakest assumption; I agree and sharpen the concern with a specific internal discrepancy (the 53% no-preference rate in Q9 versus 36% in Sivasundaram & Nielsen). This strengthens the need for sensitivity analysis, but it does not overturn the descriptive snapshot or clearly demand rejection. The reader's CONDITIONAL verdict remains appropriate: the paper should either provide uncertainty estimates and justify the 8-minute cutoff, or temper the cross-decade stability claim. My agreement is partial because I add the Q7/Q9 consistency check and the cutoff sensitivity test as concrete ways to resolve the concern.","tokens_in":10235,"tokens_out":4577,"duration_ms":52005,"concrete_test":"Using the GitLab repository dataset, recompute the Q7 (favorite interpretation) distribution in three ways: (1) as reported (>=8 min, N=30); (2) including all completions regardless of time; (3) restricted to respondents who answered Q9 with 'Never', 'Once', or 'Several times' (i.e., those with a stated preference). Also compute the cross-tab of Q7 with Q9 to see how many Q7 'Other' or 'Copenhagen' respondents are actually in the no-preference group. If the Copenhagen share moves by more than ~10 percentage points between (1) and (2) or (3), the 'among those who picked an interpretation' claim is sensitive to arbitrary preprocessing and sample definition; if it remains stable near 60%, the main stability conclusion is more robust.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim (Sec. 5) that Copenhagen support has been 'quite consistent over a 25-year period' requires that the 2023/24 Czech sample be drawn from the same population of 'physicists in general' as the Sivasundaram & Nielsen (2016) online poll, the only other poll whose authors claim representativeness. That comparability is not established and is internally challenged by the paper's own data: Fig. 9 (Q9) shows 53% of respondents report having no preferred interpretation, versus 36% in Sivasundaram & Nielsen. This 17-point gap suggests either different sampling frames or different question functioning; either way, a raw comparison of the 60% Copenhagen share in Q7 (N=30) with earlier decider-only shares is not valid without adjustment. The problem is compounded by the arbitrary 8-minute completion cutoff (Sec. 3.2): no analysis shows whether excluded fast respondents would change the Q7 distribution. The authors acknowledge sampling biases but do not quantify them or bound their effect on the key comparison, so the stability conclusion is not yet supported.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript reports a new survey on interpretational preferences in quantum mechanics, administered by email to physicists at four Czech research institutions, yielding 40 valid questionnaires after an 8-minute completion-time cutoff. The questionnaire adapts 10 of the 16 questions from Schlosshauer, Kofler, and Zeilinger (2013); the results are presented as descriptive distributions (Figs. 1-10), analyzed with Bayesian-network learning and jackknifed chi-square and Spearman tests (Figs. 11-13), and compared with four earlier polls (Tegmark 1998; Schlosshauer et al. 2013; Norsen and Nelson 2013; Sivasundaram and Nielsen 2016). The central claim, stated in Sec. 5, is that the 60% Copenhagen share among the 30 respondents who expressed a preference confirms a 'quite consistent pattern over a 25-year period' of relatively strong Copenhagen support among physicists who hold a preferred interpretation, and that observed fluctuations are mostly due to the biased audiences of specialized quantum-foundations events.","tokens_in":10566,"tokens_out":11043,"duration_ms":118819,"significance":"The paper is a descriptive contribution to the sociology of quantum foundations. Its genuine strengths are that the questionnaire, dataset, and analysis code are made available (Sec. 3.3), the descriptive statistics are presented clearly, and the authors explicitly acknowledge the informal and biased character of the earlier polls they compare against. If the stability claim survives scrutiny, the paper would supply a useful aggregation point: four polls across 25 years would indicate that Copenhagen's dominance among those who declare an interpretation is robust to the rise of quantum information, decoherence, and the modern Everett revival. The contribution is modest in size, however: N=40 overall (N=30 for the headline question), a self-selected volunteer sample recruited with a bookstore coupon, an arbitrary response-time cutoff, and cross-poll comparisons made without confidence intervals or formal tests. The paper's value therefore depends entirely on whether the cross-poll comparison can be placed on a defensible statistical footing; the major comments below identify the load-bearing points that need work.","major_comments":[{"comment":"The 8-minute completion cutoff is a free parameter that the paper justifies only by 'we considered this duration to be the minimum to meaningfully respond to the entire questionnaire.' The manuscript reports neither the number of responses excluded by this rule nor any sensitivity analysis. Since the headline result rests on N=30, even a handful of excluded fast responses could move the 60% figure materially, and the authors should report the exclusion count and re-run the Q7 analysis under a plausible range of cutoffs (e.g., 5, 8, and 12 minutes). A related design issue is that Q7 (Sec. 3.1) offers no 'no preferred interpretation' option, while Q9 (Fig. 9) records 53% of responders as having no preferred interpretation. The paper never reconciles these facts: if roughly half of the valid sample is undecided, the 30 Q7 responders are not automatically 'those who have picked an interpretation to begin with' as the conclusion in Sec. 5 phrases it, because a forced-choice question can inflate the apparent decider share. A cross-tabulation of Q9 by Q7 would clarify whether the no-preference respondents are realistically excluded from the headline number.","section":"Secs. 3.2 and 5"},{"comment":"The stability claim requires that the 40 Czech volunteers be comparable to the populations surveyed by Tegmark, Schlosshauer et al., Norsen and Nelson, and Sivasundaram and Nielsen. The paper asserts this comparability ('our poll together with the Sivasundaram and Nielsen poll, were likely the most representative ones,' Sec. 5) but does not establish it, and its own data point in the opposite direction: the no-preference rate in Q9 is 53%, versus 36% in Sivasundaram and Nielsen as reported in Sec. 2. A 17-point gap of this size suggests either different sampling frames or different question functioning; in either case, the raw comparison of the 60% Q7 share with earlier figures is not yet valid. The authors concede in Sec. 5 that such biases 'inadvertently impacted our poll' but do not quantify them or bound their possible effect on the key comparison. A sensitivity or bounding analysis, or at minimum a formal comparison of the demographic and no-preference profiles across polls, is needed before the representativeness claim can be maintained.","section":"Secs. 2, 3.2, and 5"},{"comment":"The central comparison mixes denominators. The present 60% is the share among the 30 respondents who answered Q7, whereas the comparative figures cited from earlier polls are shares of all poll respondents (e.g., 27% in Tegmark, 39% Copenhagen and 36% no-preference in Sivasundaram and Nielsen, 4% in Norsen and Nelson, all in Sec. 2). For a claim about 'those who have picked an interpretation to begin with,' the appropriate comparator for Sivasundaram and Nielsen would be the decider-only share, approximately 39%/(1-36%) = 61%, which is close to the present 60% and would support the authors' conclusion; however, this renormalization is never computed. Moreover, no confidence intervals or formal tests accompany the comparisons: the Wilson 95% confidence interval for 18/30 is roughly 42-75%, which overlaps the range of the earlier polls' reported shares, so the raw numbers alone cannot sustain the claim of a 'quite consistent pattern over a 25-year period' in either direction. The authors should report intervals for all Q7-derived estimates and either formally test the cross-poll differences or explicitly bound them.","section":"Secs. 4.1 and 5"}],"minor_comments":[{"comment":"The option label 'A severe diffculty threatening quantum mechanics' contains a typo ('diffculty'), and the wording differs from the questionnaire text in Sec. 3.1, which reads 'A serious difficulty threatening quantum mechanics (Serious)'; the two should be harmonized.","section":"Fig. 4"},{"comment":"The sentence 'See R Core Team (2021) for details' is attached to the explanation of CPDAGs, but the R Core Team reference documents the R software environment, not graphical-model theory; the CPDAG explanation should cite the graphical-model literature already listed (e.g., Pearl 1988 or Koller and Friedman 2009).","section":"Sec. 3.3"},{"comment":"The Data availability statement says that 'The data that support the findings of this study are available within the article,' but the article contains only aggregate percentages; the raw data appear to reside in the GitLab repository mentioned in Sec. 3.3, and the statement should point readers there explicitly.","section":"Statements and Declarations"},{"comment":"The manuscript nowhere states when the survey was administered, which is essential for a claim about stability 'over a 25-year period'; the data-collection dates should be reported.","section":"Sec. 3.2"},{"comment":"Q7 is phrased 'Which interpretation of quantum mechanics do I prefer?', presumably a template remnant that should read 'do you prefer'; Q6 also lacks closing punctuation.","section":"Sec. 3.1"},{"comment":"The statement that the questionnaire was piloted and that 'No further validation was performed' is transparent, but the manuscript should also report the number of pilot participants and the nature of the wording adjustments, since question-functioning differences (especially the forced-choice format of Q7) are directly relevant to the cross-poll comparisons.","section":"Sec. 3.1"}],"recommendation":"major_revision","confidential_remarks":"This is a small descriptive study whose evidentiary weight is almost entirely carried by the cross-poll comparison; if the target journal is a core physics venue, the novelty is thin, whereas a foundations-of-physics or philosophy-of-science venue would be a more natural fit. The editor may wish to verify that the GitLab repository (gitlab.cesnet.cz/utia/public/quantum-interpretations) is publicly accessible and that the raw responses match the reported percentages, since the data-availability statement currently claims the data are in the article, which they are not. The authors' assertion that their poll and Sivasundaram and Nielsen's are the 'most representative' is a judgment that should be either defended with demographic or sensitivity evidence or softened."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a transparent, honest snapshot of interpretational preferences from a small Czech sample, and the descriptive figures are worth having. The cross-decade stability claim is the soft spot: the paper's own Q9 no-preference rate (53%) is far above Sivasundaram & Nielsen's 36%, which signals the samples are not measuring the same population, and the 8-minute cutoff is untested. That doesn't kill the paper, but it means the '25-year consistency' conclusion is not yet supported.\n\nWhat's new: a fresh 2025 dataset from four Czech institutions, a direct replication of selected SKZ questions, and a Bayesian network analysis plus chi-square/Spearman with jackknife frequencies. The authors put code and data on GitLab, and they are explicit about the response-rate-per-question and the one-option limitation. That is real work. The descriptive results--Copenhagen at 60% of deciders, randomness as fundamental, superpositions possible--line up with earlier broad polls, which is a useful confirmation.\n\nWhere it gets shaky: the central comparison. To claim stability over 25 years, the 2023/24 Czech volunteers need to be comparable to Tegmark's workshop, SKZ's conference, Norsen's invitation-only realist crowd, and S&N's online sample. The paper acknowledges sampling biases in Section 5 but never quantifies them. The Q9 gap is the internal tell: 53% no-preference versus 36% in S&N. Either the framing differs or the populations differ, and both undermine a raw comparison of the 60% Copenhagen share among deciders. The 8-minute cutoff is asserted, not tested. A quick sensitivity check--reporting the excluded fast respondents' Q7 distribution--would settle whether it changes anything.\n\nThe Bayesian network is fine as exploratory, but with 40 respondents and mostly weak edges, it adds little weight to the main claim.\n\nBottom line: worth a serious referee, because the dataset is a legitimate new measurement and the paper is honest about its limitations. But the authors should be asked to either temper the stability claim or provide a comparability analysis. I'd cite it as a data point, not as proof of stability.","headline":"Useful new 2025 poll data, but the cross-decade stability claim rests on a comparability assumption the paper's own no-preference numbers undermine.","tokens_in":10980,"tokens_out":2282,"would_cite":true,"duration_ms":25825,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["03.65.Ta"],"model":"deepseek-v4-flash","headline":"New poll: physicists still favor Copenhagen after 25 years.","keywords":["quantum foundations","Copenhagen interpretation","many-worlds interpretation","de Broglie-Bohm theory","interpretational preferences","scientific polling","Bayesian network analysis","replication study"],"falsifier":"A fresh probability-based survey of a large, representative sample of working physicists that showed, say, many-worlds or pilot-wave interpretations ahead of Copenhagen, or a steady year-by-year decline in Copenhagen's share, would overturn the claim of 25-year stability; the observation needed is simply a different preference ordering in an unbiased sample.","tokens_in":10047,"feed_emoji":"⚛️","tokens_out":6831,"duration_ms":74797,"temperature":0.7,"pith_summary":"The paper reports a new, small poll of physicists at four Czech research institutions, run with questions adapted from the 2011 conference survey, and compares answers to four earlier polls spanning 1997–2016. Its central claim is that the Copenhagen interpretation has remained the most preferred interpretation among physicists who state a preference, and that the large swings seen in past polls come mostly from the self-selected audiences at specialist conferences rather than from a genuine change in community attitudes over time. The authors therefore read their results as evidence against an often-quoted claim from the 1997 workshop poll that interpretational preferences were 'gradually changing.' A sympathetic reader would care because it suggests that theoretical and experimental advances in quantum foundations have not, so far, shifted where most physicists stand on what quantum theory means.","feed_headline":"New poll: physicists still favor Copenhagen after 25 years","feed_subtitle":"A 40-person replication poll finds the same split as the 1990s, with swings driven by conference crowds.","key_machinery":"The instrument that carries the argument is the repeated forced-choice poll question on preferred interpretation, kept comparable across polls by adapting the 2011 survey's questionnaire. Around it, the paper builds a comparison table of four historical polls and adds a Bayesian-network analysis with leave-one-out resampling to identify stable associations between answers. This machinery turns a single 40-person snapshot into a claim about a 25-year trend.","core_discovery":"On the paper's own terms, the discovery is the cross-decade stability of interpretational preference: in the new poll, 60% of the 30 respondents who answered the favorite-interpretation question chose Copenhagen, and this pattern matches the ordering in the 1997 workshop poll, the 2011 conference poll, and the 2016 online poll, while the strongly pro-pilot-wave result of the 2013 conference poll is treated as an artifact of that event's recruitment. The paper also finds that Copenhagen-minded respondents tend to view quantum randomness as fundamental and quantum states as purely statistical, and that physicists holding the completeness view that the wave function description is complete rarely switch interpretations. These correlations, the paper argues, fit the picture of Copenhagen as the default taught to students and as a pragmatically minimal position that avoids extra metaphysical baggage.","pith_inferences":["If the pedagogical-default explanation is correct, a curriculum that taught many-worlds or Bohmian mechanics earlier and more prominently would predictably erode Copenhagen's lead; this could be tested by polling students before and after such courses.","Because 53% of respondents in this poll said they have no preferred interpretation, the headline support for Copenhagen applies only to the minority who commit; an alternative framing that counts all physicists would put 'no preference' in first place.","The stability pattern could be checked quantitatively by re-analyzing the published response tables of all four earlier polls with the same Bayesian-network and correlation method, testing whether the Copenhagen-randomness-statistical cluster appears in each era."],"forward_implications":["If the stability claim is right, interpretational preferences are governed more by education and pragmatic attitudes than by new experiments or theorems, so no near-term consensus should be expected.","The widely cited 'gradually changing views' conclusion from the 1997 poll would be wrong; apparent shifts like the 2013 pro-pilot-wave spike would be explained by who happened to attend.","Future polls should recruit from the general physics community rather than from quantum-foundation meetings, and should report the substantial 'no preferred interpretation' share.","The data link Copenhagen to a specific cluster of views (fundamental randomness, statistical quantum states, observer as formal role), so future studies could track this cluster as an indicator of Copenhagen's standing."],"supporting_citations":[{"why":"Supplies the first 1997 workshop poll and the 'gradually changing' claim that this paper argues against.","marker":"Tegmark (1998)"},{"why":"Supplies the 16-question survey instrument and the 2011 conference baseline from which the new questionnaire is adapted.","marker":"Schlosshauer, Kofler, and Zeilinger (2013)"},{"why":"Supplies the outlier 2013 poll whose pro-pilot-wave result the paper attributes to self-selected attendance.","marker":"Norsen and Nelson (2013)"},{"why":"Supplies the broad online poll of 149 physicists that anchors the 'representative' comparison and the stability conclusion.","marker":"Sivasundaram and Nielsen (2016)"}],"fun_headline_variants":["60% of physicists still choose Copenhagen","Copenhagen remains top quantum interpretation after decades","Quantum favorites: Copenhagen still wins, poll finds","Physicists' quantum stance unchanged for quarter century"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that 40 volunteers, self-selected from four Czech institutions by an email invitation with a bookstore coupon and filtered by an 8-minute completion cutoff, speak for the same underlying population as the earlier conference and online polls, so the comparison across decades means what the paper says it means.","fun_headline_variants_meta":{"raw":{"variants":["60% of physicists still choose Copenhagen","Copenhagen remains top quantum interpretation after decades","Quantum favorites: Copenhagen still wins, poll finds","Physicists' quantum stance unchanged for quarter century"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00092,"raw_usage":{"total_tokens":3899,"prompt_tokens":853,"completion_tokens":3046,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":469,"completion_tokens_details":{"reasoning_tokens":2989}},"tokens_in":469,"tokens_out":3046,"duration_ms":26917,"temperature":1.0,"reasoning_tokens":2989,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T17:41:50.617618+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A fresh probability-based survey of a large, representative sample of working physicists that showed, say, many-worlds or pilot-wave interpretations ahead of Copenhagen, or a steady year-by-year decline in Copenhagen's share, would overturn the claim of 25-year stability; the observation needed is simply a different preference ordering in an unbiased sample.","supporting_citations":[],"review_version":1}