{"id":"7ebe00f5-9885-402b-b228-c211360226eb","arxiv_id":"2411.15045","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"About 54.8% of Indian Web of Science publications from 2011-2022 acknowledge funding, led by DST, CSIR, UGC, SERB, and DBT among Indian agencies.","lead":"This paper counts which funding agencies are acknowledged in Indian research papers indexed by Web of Science from 2011 to 2022, finding that about 55% of them mention a funder. It matters because India is setting up a new national research foundation and the breakdown of public versus private funding is central to that policy debate.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Central percentages rest on WoS funding-metadata completeness and an unreleased manual disambiguation; the decisive missing check is a full-text audit of records classified as non-funded.","rationale":"The reader's weakest assumption—that presence or absence of WoS funding metadata faithfully indicates fundedness after manual cleaning—is exactly the load-bearing point. My review sharpens it by noting two concrete consequences: (a) the missing-metadata problem is directly acknowledged in the paper and is testable by full-text audit; (b) the agency-level analysis has an internal accounting gap, since the paper does not report how many of the 504,268 funded publications were matched to at least one of the 1,407 retained agencies, so the denominator used for the Table 2 percentages is not auditable. These are not internal contradictions that warrant rejection; they are unverified empirical assumptions. The proposed audit would settle the first and most fundamental issue. If a full-text audit finds negligible acknowledgements among 'non-funded' records, the central percentage survives; if it finds substantial acknowledgements, the headline and all derived shares shift. This does not change the reader's conditional verdict: the paper should be accepted or promoted only with the release of the disambiguation mapping, the processing scripts, and the validation evidence. The tone of the paper is appropriately cautious, and the limitations section is honest, so the concern is about verifiability and metadata bias, not about the authors' intent or a mathematical flaw.","tokens_in":10491,"tokens_out":6028,"duration_ms":65182,"concrete_test":"Conduct a full-text audit of a stratified random sample of about 400 WoS records classified as non-funded, drawn across years (e.g., 50 each from 2011, 2014, 2017, 2020, 2022), and search the publisher versions/PDFs for explicit funding acknowledgement statements. If more than about 5% of sampled 'non-funded' records contain such statements, the headline 54.79% is materially understated and the year-over-year trend in Table 1 is unreliable. In the same pass, if the authors release their disambiguation mapping and Python scripts, independently reapply them to the raw Funding Orgs/Funding Text fields and verify that the Table 2 counts reproduce; if the mapping is not released, the conditional verdict remains.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central claim—54.79% of Indian WoS articles are funded, with DST, CSIR, UGC, SERB, and DBT as the top agencies—requires two conditions: (1) absence of a WoS 'Funding Orgs' entry means the paper did not acknowledge funding, and (2) the manual cleaning of agency name variants is correct and complete. Both are insecure. First, WoS funding metadata is publisher-supplied and known to be incomplete; the paper says so in Limitations but never quantifies the gap. If missing metadata is correlated with year, publisher, or field, then both the overall 54.79% figure and the year-wise trend in Table 1 are biased. Second, the agency-level pipeline keeps only 1,407 agencies with at least 50 supported publications, but the paper never reports how many of the 504,268 funded publications were actually matched to at least one of these agencies. Publications acknowledging only a below-threshold funder, or a funder whose name was mis-cleaned, remain in the denominator but contribute to no agency's count, distorting the percentage shares in Table 2. The manual disambiguation mapping is not released and has no inter-annotator validation, so the top-agency ranking cannot be independently checked. These are correctness risks in the central quantitative claim, not merely missing documentation.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper quantifies acknowledged research funding in Indian publications indexed in the Web of Science core collection (articles and reviews) during 2011–2022. Using the WoS 'Funding Orgs' field, it classifies 504,268 of 920,284 publications (54.79%) as funded and the rest as non-funded. It then identifies the most frequently acknowledged agencies, reporting DST (22.44% of funded publications), CSIR (16.32%), UGC (14.43%), SERB (8.28%), and DBT (6.23%) as the leading Indian agencies, and NIH, UKRI, NSF, and NSFC as notable foreign agencies. The paper also categorizes agencies by national versus international and public versus private status, compares India's funded share with selected countries, and discusses implications for the Anusandhan National Research Foundation.","tokens_in":10733,"tokens_out":4426,"duration_ms":47827,"significance":"If the measurement is reliable, this is a useful first comprehensive description of the funding-acknowledgment landscape in Indian WoS-indexed publications, with direct policy relevance. The paper is careful to describe the WoS query, the document-type restriction, and the broad steps of the analysis, and it explicitly acknowledges the limitation that missing funding metadata leads to publications being counted as non-funded. However, the central quantitative claim rests on two unvalidated assumptions: that absence of a WoS 'Funding Orgs' entry reliably indicates the absence of an acknowledged funder, and that the manual agency-name disambiguation is complete and correct. The paper does not report the degree to which funded publications were matched to the 1,407 retained agencies, and it does not release the disambiguation mapping or any validation sample. These issues affect the headline percentage, the year-wise trend, and the agency rankings, so they are load-bearing for the paper's contribution rather than merely presentation problems.","major_comments":[{"comment":"The central figure of 54.79% funded publications treats a missing WoS 'Funding Orgs' entry as equivalent to a publication not acknowledging funding. The Limitations section correctly concedes that 'funding metadata is missing for some other reason, the publication is not counted as an instance of funded publication even if it may have received funding support,' but the paper never quantifies this risk. Since Table 1 and the cross-country comparison in Fig. 2 depend on the completeness of this field, the authors should provide a sensitivity check, for example a random-sample audit of full-text funding acknowledgments in records with an empty 'Funding Orgs' field, and report whether missingness varies by year, publisher, or subject category. Without such a check, the year-wise trend and the comparison to other countries could reflect metadata-coverage differences rather than actual funding differences.","section":"Data and Methodology (funded/non-funded classification) and Limitations"},{"comment":"The agency-level shares in Table 2 depend on a manual disambiguation of agency-name variants and a threshold of at least 50 supported publications, but the paper does not report how many of the 504,268 funded records were actually matched to at least one of the 1,407 retained agencies. Publications acknowledging only a below-threshold funder, or a funder whose name was mis-cleaned, remain in the denominator of funded publications yet contribute to no agency's count, which distorts the percentage shares. The authors should report the coverage of the matching step, release the disambiguation mapping (or a substantial sample of it), and validate the mapping through independent annotation or a random-sample check against full-text acknowledgments. Without these, the ranking of DST, CSIR, UGC, SERB, and DBT cannot be independently verified.","section":"Data and Methodology (agency extraction and manual disambiguation) and Table 2"}],"minor_comments":[{"comment":"The text reports 441,988 publications funded by Indian agencies (87.63% of funded publications) and 371,163 funded by foreign agencies (73.60%), which together exceed the 504,268 total; the authors should explicitly report the number of publications acknowledging both Indian and foreign agencies and clarify whether these percentages use total funded publications or matched publications as the denominator.","section":"Results (paragraph on Indian and foreign agency counts)"},{"comment":"The sentence in the Conclusion, 'To study analysed the research publications,' contains a grammatical error and should read 'The study analysed...'.","section":"Discussion and Conclusion"},{"comment":"Reference 46 says 'Assessed on Aug 21, 2024' but should say 'Accessed'; also, Figure 2 and the surrounding text use 'England' for what may be the United Kingdom in WoS country-level data, so the authors should clarify the WoS country-field definition.","section":"References and Figure 2"},{"comment":"In the discussion of private agencies, 'Merk' should presumably be 'Merck'; also, the abbreviations in Table 2 such as 'KSCSTE' and 'DOST' should be expanded or defined in a note for readers outside India.","section":"Results (private funding agencies)"}],"recommendation":"major_revision","confidential_remarks":"The paper's topic fits the journal and the descriptive results are potentially valuable, but the absence of a released agency-name mapping and the lack of any validation of the 'non-funded' classification are substantial reproducibility gaps. If the authors can provide the requested coverage statistics, a validation sample, and the disambiguation mapping as supplementary material, the paper would be much closer to acceptable. I would also encourage the editor to consider whether the discussion's comparison of DST expenditure amounts with publication counts is sufficiently cautious; the paper currently makes a loose correspondence without directly measuring the funding amount per publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a descriptive bibliometric paper, and as such it mostly works. It gives the first full-country numbers for acknowledged funding in Indian WoS-indexed articles and reviews (54.79% funded over 2011-2022; DST, CSIR, UGC, SERB, and DBT as the top Indian agencies). That is genuinely new and policy-relevant given ANRF planning. The method is simple and reproducible in principle: one WoS query, Python processing, and year-wise and agency-level counts. The comparisons with China, the US, and others are useful context. The limitations section is honest about WoS coverage and missing funding metadata.\n\nThe soft spots are real but not fatal. The funded share depends on treating 'no Funding Orgs entry' as 'not funded,' which WoS metadata completeness makes shaky. The paper admits this but never quantifies the gap, so the 54.79% figure could shift, and any shift could be correlated with year or field. The bigger issue is the manual disambiguation of agency name variants. Only 1,407 agencies with at least 50 publications are kept, and the paper never reports how many of the 504,268 funded publications actually matched at least one of those agencies. Publications acknowledging a below-threshold or mis-cleaned funder remain in the denominator but in no agency's numerator, which would distort the percentage shares in Table 2. The mapping itself is not released and has no inter-annotator check, so the top-agency ranking cannot be independently verified.\n\nNone of this makes the paper a bad paper. It is a measurement study, and the right response is to ask for the data and a validation audit, not to reject the approach. A full-text audit of a sample of records classified as non-funded, plus a release of the agency mapping, would settle the main uncertainty. The central findings are plausible and they fit what is known about Indian R&D funding, but 'plausible' is not 'verified.'\n\nWho it is for: science-policy researchers and bibliometricians who need a baseline for Indian funding acknowledgements. It deserves a serious referee. I would send it to review with requests for supplementary data, and I would cite it as the first national baseline, with caveats.","headline":"A useful first national baseline for Indian funding acknowledgements, but the headline percentages rest on WoS metadata and an unreleased manual disambiguation that need validation before the numbers are treated as hard policy facts.","tokens_in":11250,"tokens_out":1666,"would_cite":true,"duration_ms":16417,"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":"This paper quantifies funded Indian research, finding 54.79% of publications acknowledge support, led by DST at 22.44%.","keywords":["research funding","India","funding acknowledgement","Web of Science","bibliometrics","DST","CSIR","ANRF"],"falsifier":"Compare the same 2011–2022 Indian article set in Scopus, which has its own funding-acknowledgement field, against the Web of Science-derived list; if the funded-publication share differs by more than a few points, the Web of Science metadata misses a systematic set of acknowledgements.","tokens_in":10309,"feed_emoji":"📊","tokens_out":4683,"duration_ms":41323,"temperature":0.7,"pith_summary":"This paper sets out to measure how much of India's research output is visibly supported by funding agencies, using Web of Science publication records from 2011 to 2022 as a proxy. It reports that 54.79% of 9,20,284 Indian articles and reviews acknowledge funding, and identifies the Department of Science and Technology, CSIR, and UGC as the dominant supporters. The findings matter because they provide a first quantitative baseline for which agencies actually underwrite Indian science, and they suggest that government, not private, funding dominates. The authors connect these patterns to the upcoming Anusandhan National Research Foundation, which intends to shift funding toward private investment.","feed_headline":"Over half of India's research papers name a funder","feed_subtitle":"New bibliometric study of 9,20,284 Web of Science papers finds DST, CSIR, and UGC dominate, with private funding scarce.","key_machinery":"The analytical engine is the Web of Science 'Funding Orgs' metadata field, treated as a proxy for whether a publication received funding. The paper tabulates presence of this field to classify publications as funded or non-funded, then extracts and manually disambiguates agency-name variants (e.g., 'DST', 'DST India', 'DST Delhi' all become 'DST') to attribute each funded paper to one or more agencies. That step converts noisy acknowledgement strings into a clean agency-level ledger, from which the share computations and rankings follow.","core_discovery":"The paper finds that 5,04,268 of 9,20,284 (54.79%) Indian articles and reviews in the Web of Science core collection for 2011–2022 carry funding acknowledgement data. Among funded publications, Indian agencies account for 87.63% of the total, with DST (22.44%), CSIR (16.32%), UGC (14.43%), SERB (8.28%), and DBT (6.23%) leading. Foreign funders, chiefly NIH, UKRI, and NSF, support a smaller but substantial share (73.60% of funded publications, counting overlaps), indicating widespread international collaboration. The paper also reports that India's funded-publication share (about 55%) sits below China (~86%), Japan (~71%), the US (~68%), Germany (~69%), and England (~69%). On the public–private axis, almost all significant funders are public bodies; private foundations and firms appear in small numbers, mostly foreign.","pith_inferences":["If the missing-funding-metadata rate is systematically higher for certain fields or institution types, the 54.79% overall share could substantially understate the true funded fraction; a next step would be to validate against Scopus or national datasets.","The paper's reliance on the WoS 'Funding Orgs' field means papers funded by agencies but not acknowledged by authors are invisible; the observed ranking of DST, CSIR, and UGC could partly be an artifact of acknowledgement norms rather than the actual distribution of funding.","The 73.60% foreign-funder share overlapping with 87.63% Indian-funder share implies high co-funding; analyzing the co-funding matrix could reveal which Indian agencies typically pair with NIH, UKRI, or NSF, offering a testable extension.","If ANRF shifts funding toward private entities, a natural follow-up is whether the share of funded publications acknowledging private agencies rises; the paper's baseline enables that measurement."],"forward_implications":["India's funded-publication share staying near 52–58% across 2011–2022, while its total output grows at a 9.30% CAGR, implies funding acknowledgements are keeping pace with output rather than driving a disproportionate expansion.","Since 87.63% of funded publications acknowledge Indian agencies, the bulk of Indian research support is domestic and public; this contradicts any assumption that foreign or private money underwrites most Indian science.","The pattern of private funders being rare and mostly foreign (BMGF, Sloan, pharmaceutical firms) suggests that policies relying on private-sector R&D investment, like ANRF, start from a low base.","DST's share (22.44%) being the highest aligns with its extramural R&D expenditure being the largest among Indian agencies, suggesting a loose link between money spent and papers acknowledged.","The year-wise dip after 2018 to 52.68% in 2022 could signal changes in acknowledgement behavior or funding intake, but the paper does not model why."],"supporting_citations":[{"why":"Supplies the official R&D expenditure statistics that the paper uses to check whether agency publication shares track extramural grant amounts.","marker":"[4]"},{"why":"Documents the incomplete journal coverage of Web of Science, which the paper acknowledges as a boundary on its funded-publication counts.","marker":"[47]"},{"why":"Provides the government's description of ANRF's intended public–private funding split, which frames the paper's policy discussion.","marker":"[46]"}],"fun_headline_variants":["Public agencies fund 88% of credited Indian research","DST, CSIR, UGC top India's research funding","Half of Indian papers name funders, mostly public","Private funding negligible in Indian research","India's funded research share trails global peers"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The analysis assumes that the presence or absence of funding information in a Web of Science record faithfully indicates whether and by whom a paper was funded, and the paper itself concedes that missing metadata can make funded work appear unfunded.","fun_headline_variants_meta":{"raw":{"variants":["Public agencies fund 88% of credited Indian research","DST, CSIR, UGC top India's research funding","Half of Indian papers name funders, mostly public","Private funding negligible in Indian research","India's funded research share trails global peers"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000257,"raw_usage":{"total_tokens":1572,"prompt_tokens":932,"completion_tokens":640,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":548,"completion_tokens_details":{"reasoning_tokens":568}},"tokens_in":548,"tokens_out":640,"duration_ms":5905,"temperature":1.0,"reasoning_tokens":568,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T14:32:43.410128+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compare the same 2011–2022 Indian article set in Scopus, which has its own funding-acknowledgement field, against the Web of Science-derived list; if the funded-publication share differs by more than a few points, the Web of Science metadata misses a systematic set of acknowledgements.","supporting_citations":[{"cited_title":"Research & Development Statistics at a Glance 2022-23","cited_arxiv_id":null,"evidence_quote":"Supplies the official R&D expenditure statistics that the paper uses to check whether agency publication shares track extramural grant amounts."},{"cited_title":"and Mayr, P., The journal coverage of Web of Science, Scopus and Dimensions: A comparative analysis","cited_arxiv_id":null,"evidence_quote":"Documents the incomplete journal coverage of Web of Science, which the paper acknowledges as a boundary on its funded-publication counts."},{"cited_title":"Ministry of Science and Technology, New Delhi","cited_arxiv_id":null,"evidence_quote":"Provides the government's description of ANRF's intended public–private funding split, which frames the paper's policy discussion."}],"review_version":1}