pith. sign in

arxiv: 2605.04926 · v1 · submitted 2026-05-06 · 💻 cs.CL

Unintended Negative Impacts of Promotional Language in Patent Evaluation

Bingkun Zhao , Chenwei Zhang , Hao Peng This is my paper

Pith reviewed 2026-05-08 16:28 UTC · model grok-4.3

classification 💻 cs.CL
keywords promotional languagepatent evaluationUSPTO applicationsgrant probabilityownership transferappeal successcombinatorial noveltyexaminer characteristics
0
0 comments X

The pith

Higher promotional language in patent applications links to lower grant, transfer, and appeal success rates.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper analyzes 2.7 million USPTO patent applications using a 135-word lexicon to measure promotional language. It establishes that greater use of such words correlates with reduced probabilities of patent grants, ownership transfers, and successful appeals, even after statistical controls and across tech fields. This runs counter to promotional language's helpful role in science communication. The work further shows that promotional phrasing tracks objective measures of novelty and future citations rather than masking weak inventions. Examiner traits such as gender and experience moderate how much this language is tolerated.

Core claim

A higher frequency of words from a validated 135-term promotional lexicon is negatively associated with the probability that a patent application is granted, ownership is transferred, or a rejection is successfully appealed. The association persists after matching and regression controls for confounders and holds across technological areas. In matched samples the success gap between lowest and highest promotional-density quintiles is 5.5, 5.9, and 5.3 percentage points for the three outcomes. Promotional language nevertheless correlates positively with combinatorial novelty and forward citation impact, indicating it signals rather than conceals technological strength. Tolerance for the same

What carries the argument

A 135-word promotional lexicon combined with large-scale matching and regression on 2.7 million USPTO applications to isolate language effects on grant, transfer, and appeal outcomes.

Load-bearing premise

The 135-word lexicon captures promotional framing without systematic bias or omission in the legal-technical language of patents, and the matching plus regression setup accounts for all relevant confounders.

What would settle it

Re-running the analysis with an alternative promotional word list or additional controls for unmeasured technology quality that eliminates or reverses the negative association with grant probability.

Figures

Figures reproduced from arXiv: 2605.04926 by Bingkun Zhao, Chenwei Zhang, Hao Peng.

Figure 1
Figure 1. Figure 1: Promotional language is negatively associated with patent outcomes. (A–C) Binned scatter plots showing the weighted Pearson correlation between the density of promotional words and the probability of an application being (A) granted a patent, (B) transferred ownership, and (C) successfully appealed. Each dot represents a CPC Class (124 in total), with size proportional to its number of applications and col… view at source ↗
Figure 2
Figure 2. Figure 2: The negative association between promotional language and patent success is largely robust across different technological areas. This figure shows the predicted marginal effects of the percentage of promotional words on (A) the probability of an application being granted a patent and (B) the probability of a patent application being transferred ownership. We fit a logistic regression model to predict each … view at source ↗
Figure 3
Figure 3. Figure 3: Promotional language is linked with a higher degree of novelty and citation rate. Predicted marginal effects of the density of promotional words on (A) percentile ranking of novelty score for patent applications (N = 2,748,927) and (B) percentile ranking of 3-year citation count for granted patents (N = 1,151,644). The predictions are based on an OLS regression fitted to each dependent variable using the p… view at source ↗
Figure 4
Figure 4. Figure 4: Men and experienced examiners are more receptive to a higher frequency of pro￾motional language than women and novice examiners in patent approval. (A) The predicted gender difference (men - women) in the probability of an application being granted as a function of the percentage of promotional words. The upward trend indicates that male examiners have a higher tolerance than comparable female examiners fo… view at source ↗
read the original abstract

Promotional language has been increasingly used to aid the communication of innovative ideas in science. Yet, less is known about its role in the context of technological innovation. Here, we use a validated and domain-diagnosed lexicon of 135 promotional words to study the association between promotional language and patent evaluation outcomes among 2.7 million USPTO patent applications. Our large-scale study reveals three unexpected findings. First, in contrast to scientific evaluation, we find that a higher frequency of promotional words is negatively associated with the probability of an application being (i) granted a patent, (ii) transferred ownership, and (iii) successfully appealed. This promotional penalty holds even after accounting for a range of confounding factors and is largely robust across different technological areas. Among matched samples, the difference in the success rate between the lowest and highest promotional density quintile is 5.5, 5.9, and 5.3 percentage points for patentability, transferability, and rejection reversal. Second, contrary to institutional skepticism, we show that promotional language is not a mask of weak technology, but objectively reflects the degree of combinatorial novelty and future citation impact. Third, digging into the mechanisms, we find that the tolerance to promotional framing is strongly moderated by human factors, with men and experienced examiners showing a higher acceptance of promotional narratives than women and novice examiners. By revealing an emerging paradox in the patent system, our study offers theoretical and practical implications for improving patent evaluation through more objective scrutiny of linguistic patterns in patent filings.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 2 minor

Summary. The manuscript analyzes 2.7 million USPTO patent applications using a 135-word promotional lexicon (originally validated in science and described as domain-diagnosed) to examine associations with evaluation outcomes. It reports that higher promotional density is negatively linked to grant probability, ownership transfer, and successful appeal reversal (differences of 5.5, 5.9, and 5.3 percentage points across quintiles in matched samples), even after controls and matching; that promotional language correlates with combinatorial novelty and future citations rather than masking weakness; and that effects are moderated by examiner gender and experience, with robustness across technological fields.

Significance. If the associations hold after addressing measurement concerns, the study identifies an unintended penalty in the patent system against language that objectively signals innovation, with implications for examiner training and evaluation design. The large sample, matched-sample design, and cross-field robustness checks provide a solid empirical base for these claims.

major comments (3)
  1. [Methods] Methods: The 135-word lexicon is applied to patent texts without patent-specific validation (e.g., no expert annotation of context, inter-rater reliability on excerpts, or ablation of terms like 'novel' or 'improved' that fulfill statutory requirements). This risks conflating promotional tone with claim breadth or legal phrasing, which could bias the density measure and undermine the reported negative associations with grant/transfer/appeal outcomes.
  2. [Results] Results: Exact regression specifications (full controls, fixed effects, quintile boundary definitions, and sensitivity to alternative word lists) are not detailed, despite claims of robustness; this makes it difficult to evaluate whether the 5.5–5.9 pp differences in matched samples are sensitive to specification choices or measurement error correlated with application quality.
  3. [Results] Results/Discussion: The claim that promotional language 'objectively reflects' novelty and citation impact requires clearer separation of these measures from the lexicon itself, plus checks on whether associations persist after additional controls for technological complexity or filing ambition.
minor comments (2)
  1. [Abstract] Abstract and Methods: Clarify the precise 'domain-diagnosis' process applied to patents, including any adjustments to the original science-validated list.
  2. [Tables] Tables/Figures: Report standard errors or confidence intervals alongside the quintile success-rate differences and ensure all robustness checks include error bars.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for their constructive and detailed feedback. We address each major comment point by point below, indicating the revisions we will incorporate to strengthen the manuscript.

read point-by-point responses

  1. Referee: [Methods] Methods: The 135-word lexicon is applied to patent texts without patent-specific validation (e.g., no expert annotation of context, inter-rater reliability on excerpts, or ablation of terms like 'novel' or 'improved' that fulfill statutory requirements). This risks conflating promotional tone with claim breadth or legal phrasing, which could bias the density measure and undermine the reported negative associations with grant/transfer/appeal outcomes.

    Authors: We acknowledge the importance of domain-specific validation. The lexicon originates from prior work validated in scientific contexts and applied here as a measure of promotional tone. In the revision, we will add an ablation analysis excluding terms such as 'novel' and 'improved' to test sensitivity of the main results. We will also report inter-rater reliability from manual annotation of a random sample of 500 patent excerpts by two independent coders, focusing on whether promotional usage aligns with or diverges from legal phrasing. These steps will directly address potential conflation with claim breadth. revision: yes

  2. Referee: [Results] Results: Exact regression specifications (full controls, fixed effects, quintile boundary definitions, and sensitivity to alternative word lists) are not detailed, despite claims of robustness; this makes it difficult to evaluate whether the 5.5–5.9 pp differences in matched samples are sensitive to specification choices or measurement error correlated with application quality.

    Authors: We agree that full transparency on specifications is essential for evaluating robustness. The revised manuscript will include a new appendix section providing the complete regression equations, all control variables and fixed effects (year, technology class, examiner), exact quintile boundaries derived from the promotional density distribution, and results from alternative word lists. This documentation will allow readers to confirm that the reported percentage-point differences in matched samples are not driven by specification choices or correlated measurement error. revision: yes

  3. Referee: [Results] Results/Discussion: The claim that promotional language 'objectively reflects' novelty and citation impact requires clearer separation of these measures from the lexicon itself, plus checks on whether associations persist after additional controls for technological complexity or filing ambition.

    Authors: We will clarify the independence of measures in the revision: combinatorial novelty is computed from patent classification co-occurrence patterns and citation impact from forward citations, neither of which depends on the promotional lexicon. We will also add explicit robustness checks controlling for technological complexity (number of claims and IPC subclasses) and filing ambition (inventor team size and priority claims). These additions will demonstrate that the positive associations with novelty and citations, as well as the negative associations with evaluation outcomes, persist after accounting for these factors. revision: yes

Circularity Check

0 steps flagged

Empirical regression analysis on external data with pre-validated lexicon; no derivation reduces to inputs by construction

full rationale

The paper reports observational associations between word counts from a pre-existing 135-word lexicon and USPTO outcomes (grant, transfer, appeal) via matching and regression on 2.7 million applications. No equations, fitted parameters, or predictions are presented as independent results; the lexicon is described as already validated and domain-diagnosed prior to this study. Claims about novelty and citations are treated as separate objective measures rather than derived from the promotional counts. No self-citation chain or self-definitional step is load-bearing for the central findings.

Axiom & Free-Parameter Ledger

1 free parameters · 2 axioms · 0 invented entities

The central claim rests on the accuracy of the 135-word lexicon for patents and the assumption that statistical controls and matching fully address confounding; no new entities are postulated.

free parameters (1)
  • promotional density quintile boundaries
    Data-driven cutoffs used to compare lowest vs highest groups; boundaries chosen from the observed distribution.
axioms (2)
  • domain assumption The pre-validated 135-word lexicon captures promotional language without domain-specific bias when applied to patent text.
    Abstract states the lexicon is validated and domain-diagnosed, but provides no patent-specific validation details.
  • domain assumption All relevant confounders for patent success are captured by the included controls and matching procedure.
    Abstract claims robustness after accounting for confounders, but does not list the full set of controls.

pith-pipeline@v0.9.0 · 5569 in / 1419 out tokens · 21161 ms · 2026-05-08T16:28:24.536209+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

74 extracted references · 2 canonical work pages

  1. [1]

    We calculated the final density of promotional words as the percentage of remaining 135 words relative to the total number of words in each application’s full text

    Further removal offers little improvement in reliability and could risk excluding words that are often used in a promotional tone. We calculated the final density of promotional words as the percentage of remaining 135 words relative to the total number of words in each application’s full text. The full list of 135 promotional words used in our analysis i...

    2010

  2. [2]

    & Jones, B

    Ahmadpoor, M. & Jones, B. F. The dual frontier: Patented inventions and prior scientific advance.Science357, 583–587 (2017)

    2017

  3. [3]

    Kortum, S. S. Research, patenting, and technological change.Econometrica: Journal of the Econometric Society1389–1419 (1997)

    1997

  4. [4]

    Patent statistics as economic indicators: a survey

    Griliches, Z. Patent statistics as economic indicators: a survey. InR&D and productivity: the econometric evidence, 287–343 (University of Chicago Press, 1998). 24

    1998

  5. [5]

    & Mart ´ınez, C

    Encaoua, D., Guellec, D. & Mart ´ınez, C. Patent systems for encouraging innovation: Lessons from economic analysis.Research Policy35, 1423–1440 (2006)

    2006

  6. [6]

    Patents and innovation: evidence from economic history.Journal of Economic Perspectives27, 23–44 (2013)

    Moser, P. Patents and innovation: evidence from economic history.Journal of Economic Perspectives27, 23–44 (2013)

    2013

  7. [7]

    Spulber, D. F. How patents provide the foundation of the market for inventions.Journal of Competition Law & Economics11, 271–316 (2015)

    2015

  8. [8]

    Lemley, M. A. & Sampat, B. Examiner characteristics and patent office outcomes.Review of Economics and Statistics94, 817–827 (2012)

    2012

  9. [9]

    Frakes, M. D. & Wasserman, M. F. Is the time allocated to review patent applications inducing examiners to grant invalid patents? evidence from microlevel application data.Review of Economics and Statistics99, 550–563 (2017)

    2017

  10. [10]

    & Moreira, S

    Argente, D., Baslandze, S., Hanley, D. & Moreira, S. Patents to products: Product innovation and firm dynamics. Tech. Rep., National Bureau of Economic Research (2025)

    2025

  11. [11]

    & Schankerman, M

    Galasso, A. & Schankerman, M. Patents and cumulative innovation: Causal evidence from the courts.The Quarterly Journal of Economics130, 317–369 (2015)

    2015

  12. [12]

    C., Sarnoff, J

    Marco, A. C., Sarnoff, J. D. & Charles, A. Patent claims and patent scope.Research Policy 48, 103790 (2019)

    2019

  13. [13]

    Wagner, R. P. Understanding patent-quality mechanisms.University of Pennsylvania Law Review157, 2135–2173 (2009)

    2009

  14. [14]

    & Marco, A

    Kuhn, J., Younge, K. & Marco, A. Patent citations reexamined.The RAND Journal of Eco- nomics51, 109–132 (2020)

    2020

  15. [15]

    B., Sun, S

    Kong, N., Dulleck, U., Jaffe, A. B., Sun, S. & Vajjala, S. Linguistic metrics for patent disclo- sure: Evidence from university versus corporate patents.Research Policy52, 104670 (2023)

    2023

  16. [16]

    & Ferriani, S

    Falchetti, D., Cattani, G. & Ferriani, S. Start with “why,” but only if you have to: The strategic framing of novel ideas across different audiences.Strategic Management Journal43, 130–159 (2022)

    2022

  17. [17]

    & Evans, J

    Shi, F. & Evans, J. Surprising combinations of research contents and contexts are related to impact and emerge with scientific outsiders from distant disciplines.Nature Communications 14, 1641 (2023). 25

    2023

  18. [18]

    Peng, H., Ke, Q., Budak, C., Romero, D. M. & Ahn, Y .-Y . Neural embeddings of scholarly periodicals reveal complex disciplinary organizations.Science Advances7, eabb9004 (2021)

    2021

  19. [19]

    & Wezel, F

    Kov ´acs, B., Carnabuci, G. & Wezel, F. C. Categories, attention, and the impact of inventions. Strategic Management Journal42, 992–1023 (2021)

    2021

  20. [20]

    & Stam, D

    Beretta, M., Deichmann, D., Frederiksen, L. & Stam, D. Do you see what i see? how expertise and a decision-maker role influence the recognition and selection of novel ideas.Research Policy54, 105139 (2025)

    2025

  21. [21]

    S., Fosse, H

    Peng, H., Qiu, H. S., Fosse, H. B. & Uzzi, B. Promotional language and the adoption of inno- vative ideas in science.Proceedings of the National Academy of Sciences121, e2320066121 (2024)

    2024

  22. [22]

    S., Peng, H., Fosse, H

    Qiu, H. S., Peng, H., Fosse, H. B., Woodruff, T. K. & Uzzi, B. Use of promotional language in grant applications and grant success.JAMA Network Open7, e2448696–e2448696 (2024)

    2024

  23. [23]

    Stavrova, O., Kleinberg, B., Evans, A. M. & Ivanovi´c, M. Scientific publications that use pro- motional language in the abstract receive more citations and public attention.Communications Psychology3, 118 (2025)

    2025

  24. [24]

    J., Sorenson, O

    Lerchenmueller, M. J., Sorenson, O. & Jena, A. B. Gender differences in how scientists present the importance of their research: observational study.BMJ367(2019)

    2019

  25. [25]

    & Hua, X

    Bromham, L., Dinnage, R. & Hua, X. Interdisciplinary research has consistently lower funding success.Nature534, 684–687 (2016)

    2016

  26. [26]

    & Stephan, P

    Wang, J., Veugelers, R. & Stephan, P. Bias against novelty in science: A cautionary tale for users of bibliometric indicators.Research Policy46, 1416–1436 (2017)

    2017

  27. [27]

    & Lakhani, K

    Teplitskiy, M., Peng, H., Blasco, A. & Lakhani, K. R. Is novel research worth doing? evi- dence from peer review at 49 journals.Proceedings of the National Academy of Sciences119, e2118046119 (2022)

    2022

  28. [28]

    Allison, J. R. & Ouellette, L. L. How courts adjudicate patent definiteness and disclosure. Duke LJ65, 609 (2015)

    2015

  29. [29]

    A., Lemley, M

    Cotropia, C. A., Lemley, M. A. & Sampat, B. Do applicant patent citations matter?Research Policy42, 844–854 (2013). 26

    2013

  30. [30]

    & Simcoe, T

    Righi, C. & Simcoe, T. Patent examiner specialization.Research Policy48, 137–148 (2019)

    2019

  31. [31]

    Harhoff, D., Narin, F., Scherer, F. M. & V opel, K. Citation frequency and the value of patented inventions.Review of Economics and Statistics81, 511–515 (1999)

    1999

  32. [32]

    Cumulative innovation and market value: Evidence from patent citations.The Economic Journal122, 265–285 (2012)

    Belenzon, S. Cumulative innovation and market value: Evidence from patent citations.The Economic Journal122, 265–285 (2012)

    2012

  33. [33]

    & Baruffaldi, S

    Poege, F., Harhoff, D., Gaessler, F. & Baruffaldi, S. Science quality and the value of inven- tions.Science Advances5, eaay7323 (2019)

    2019

  34. [34]

    Kim, Y . K. & Oh, J. B. Examination workloads, grant decision bias and examination quality of patent office.Research Policy46, 1005–1019 (2017)

    2017

  35. [35]

    Lim, H.et al.Panorama: A dataset and benchmarks capturing decision trails and rationales in patent examination.arXiv preprint arXiv:2510.24774(2025)

    work page arXiv 2025

  36. [36]

    Natural language analysis of patent claims

    Sheremetyeva, S. Natural language analysis of patent claims. InProceedings of the ACL-2003 Workshop on Patent Corpus Processing, 66–73 (2003)

    2003

  37. [37]

    & Budgell, B

    Millar, N., Batalo, B. & Budgell, B. Trends in the use of promotional language (hype) in ab- stracts of successful national institutes of health grant applications, 1985-2020.JAMA Network Open5, e2228676–e2228676 (2022)

    1985

  38. [38]

    & Rogers, M

    Helmers, C. & Rogers, M. Does patenting help high-tech start-ups?Research Policy40, 1016–1027 (2011)

    2011

  39. [39]

    & Lerner, J

    Webb, M., Short, N., Bloom, N. & Lerner, J. Some facts of high-tech patenting. Tech. Rep., National Bureau of Economic Research (2018)

    2018

  40. [40]

    & Windeler, A

    Sydow, J. & Windeler, A. Organizing and evaluating interfirm networks: A structurationist perspective on network processes and effectiveness.Organization Science9, 265–284 (1998)

    1998

  41. [41]

    Arousal increases social transmission of information.Psychological Science22, 891–893 (2011)

    Berger, J. Arousal increases social transmission of information.Psychological Science22, 891–893 (2011)

    2011

  42. [42]

    & Packard, G

    Boghrati, R., Berger, J. & Packard, G. Style, content, and the success of ideas.Journal of Consumer Psychology33, 688–700 (2023). 27

    2023

  43. [43]

    & Park, Y

    Han, Y .-j. & Park, Y . Patent network analysis of inter-industrial knowledge flows: The case of korea between traditional and emerging industries.World Patent Information28, 235–247 (2006)

    2006

  44. [44]

    B., Jeong, H

    Kim, D., Cerigo, D. B., Jeong, H. & Youn, H. Technological novelty profile and invention’s future impact.EPJ Data Science5, 8 (2016)

    2016

  45. [45]

    Kaltenberg, M., Jaffe, A. B. & Lachman, M. E. Invention and the life course: Age differences in patenting.Research Policy52, 104629 (2023)

    2023

  46. [46]

    Peng, H., Teplitskiy, M., Romero, D. M. & Horv ´at, E.- ´A. The gender gap in scholarly self- promotion on social media.Nature Communications16, 5552 (2025)

    2025

  47. [47]

    M., Kortum, S

    Cockburn, I. M., Kortum, S. S. & Stern, S. Are all patent examiners equal? the impact of examiner characteristics (2002)

    2002

  48. [48]

    Experience effects of patent examiners: an empirical study of the career length and citation patterns on triadic patents.Scientometrics129, 6333–6348 (2024)

    Wada, T. Experience effects of patent examiners: an empirical study of the career length and citation patterns on triadic patents.Scientometrics129, 6333–6348 (2024)

    2024

  49. [49]

    From discovery to invention: The writing and rewriting of two patents.Social Studies of Science25, 57–105 (1995)

    Myers, G. From discovery to invention: The writing and rewriting of two patents.Social Studies of Science25, 57–105 (1995)

    1995

  50. [50]

    A., Spangler, W

    Hasan, M. A., Spangler, W. S., Griffin, T. & Alba, A. Coa: Finding novel patents through text analysis. InProceedings of the 15th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 1175–1184 (2009)

    2009

  51. [51]

    & Gomez, J

    Arts, S., Cassiman, B. & Gomez, J. C. Text matching to measure patent similarity.Strategic Management Journal39, 62–84 (2018)

    2018

  52. [52]

    Linguistics and patent claim construction.Rutgers LJ38, 61 (2006)

    Osenga, K. Linguistics and patent claim construction.Rutgers LJ38, 61 (2006)

    2006

  53. [53]

    Patent signals.The University of Chicago Law Review625–679 (2002)

    Long, C. Patent signals.The University of Chicago Law Review625–679 (2002)

    2002

  54. [54]

    Jiang, L., Scherz, P. A. & Goetz, S. Towards better evaluation for generated patent claims. InProceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), 3775–3788 (2025)

    2025

  55. [55]

    & Courtis, J

    Leung, S., Parker, L. & Courtis, J. Impression management through minimal narrative disclo- sure in annual reports.The British Accounting Review47, 275–289 (2015). 28

    2015

  56. [56]

    & McDonald, B

    Loughran, T. & McDonald, B. When is a liability not a liability? textual analysis, dictionaries, and 10-ks.The Journal of Finance66, 35–65 (2011)

    2011

  57. [57]

    Lemley, M. A. Rational ignorance at the patent office.Nw. UL Rev.95, 1495 (2000)

    2000

  58. [58]

    & Sampat, B

    Alc ´acer, J., Gittelman, M. & Sampat, B. Applicant and examiner citations in us patents: An overview and analysis.Research Policy38, 415–427 (2009)

    2009

  59. [59]

    Frakes, M. D. & Wasserman, M. F. Irrational ignorance at the patent office.Vand. L. Rev.72, 975 (2019)

    2019

  60. [60]

    B., Uzzi, B

    Peng, L., Fosse, H. B., Uzzi, B. & Peng, H. The power of “we”’ in science funding and publication.SocArXiv(2025)

    2025

  61. [61]

    & Stephan, P

    Sauermann, H. & Stephan, P. Conflicting logics? a multidimensional view of industrial and academic science.Organization Science24, 889–909 (2013)

    2013

  62. [62]

    & Wright, B

    Lei, Z. & Wright, B. D. Why weak patents? testing the examiner ignorance hypothesis. Journal of Public Economics148, 43–56 (2017)

    2017

  63. [63]

    Burke, P. F. & Reitzig, M. Measuring patent assessment quality—analyzing the degree and kind of (in) consistency in patent offices’ decision making.Research Policy36, 1404–1430 (2007)

    2007

  64. [64]

    & Lee, S

    Kim, J. & Lee, S. Patent databases for innovation studies: A comparative analysis of uspto, epo, jpo and kipo.Technological Forecasting and Social Change92, 332–345 (2015)

    2015

  65. [65]

    & Funk, R

    Park, M., Leahey, E. & Funk, R. J. Papers and patents are becoming less disruptive over time. Nature613, 138–144 (2023)

    2023

  66. [66]

    J., Marco, A

    Graham, S. J., Marco, A. C. & Myers, A. F. Patent transactions in the marketplace: Lessons from the uspto patent assignment dataset.Journal of Economics & Management Strategy27, 343–371 (2018)

    2018

  67. [67]

    & Jones, B

    Uzzi, B., Mukherjee, S., Stringer, M. & Jones, B. Atypical combinations and scientific impact. Science342, 468–472 (2013)

    2013

  68. [68]

    & Tripathi, R

    Sharma, P. & Tripathi, R. Patent citation: A technique for measuring the knowledge flow of information and innovation.World Patent Information51, 31–42 (2017). 29

    2017

  69. [69]

    Nunnally, J. C. An overview of psychological measurement.Clinical Diagnosis of Mental Disorders: A Handbook97–146 (1978)

    1978

  70. [70]

    Cronbach, L. J. Coefficient alpha and the internal structure of tests.Psychometrika16, 297– 334 (1951)

    1951

  71. [71]

    Bland, J. M. & Altman, D. G. Statistics notes: Cronbach’s alpha.BMJ314, 572 (1997)

    1997

  72. [72]

    A new readability yardstick.Journal of Applied Psychology32, 221 (1948)

    Flesch, R. A new readability yardstick.Journal of Applied Psychology32, 221 (1948)

    1948

  73. [73]

    A concrete example of construct construction in natural language.Organiza- tional Behavior and Human Decision Processes162, 81–94 (2021)

    Yeomans, M. A concrete example of construct construction in natural language.Organiza- tional Behavior and Human Decision Processes162, 81–94 (2021)

    2021

  74. [74]

    & Acuna, D

    Liang, L. & Acuna, D. demographicx: A python package for estimating gender and ethnicity using deep learning transformers.Zenodo https://doi. org/10.5281/zenodo4898367(2021). AcknowledgementsWe thank Brian Uzzi, Tara Sowrirajan, and Ching Jin for helpful discussion. We thank Wenhui Chen and Tara Sowrirajan for initial assistance on the data cleaning and m...

    work page doi:10.5281/zenodo4898367(2021 2021