CoChat

Target mechanisms of mindfulness-based programmes and practices: a scoping review

Open access
Maloney, Shannon, Kock, Merle, Slaghekke, Yasmijn, Radley, Lucy, Lopez-Montoyo, Alba, Montero-Marin, Jesus, Kuyken, Willem · 2024 · BMJ Mental Health

The DOI matches its Crossref record; 40 of the 41 references it cites were found in a registry or an index; 13 of the 30 sources its claims cite were checked for support.

Abstract

Question Mindfulness-based programmes (MBPs) and practices have demonstrated effects in mental health and well-being, yet questions regarding the target mechanisms that drive change across the population remain unresolved. Study selection and analysis Five databases were searched for randomised controlled trials that evaluate the indirect effects (IEs) of an MBP or mindfulness practice in relation to mental health and well-being outcomes through psychological mechanisms. Findings 27 eligible studies were identified, with only four exploring mechanisms in the context of specific mindfulness practices. Significant IEs were reported for mindfulness skills, decentering and attitudes of mindfulness (eg, self-compassion) across different outcomes, population samples, mental health strategies and active comparators. Evidence gap maps and requirements for testing and reporting IEs are provided to help guide future work. Conclusions Mindfulness skills, decentering and attitudes of mindfulness may be key intervention targets for addressing the mental health of whole populations. However, future work needs to address significant knowledge gaps regarding the evidence for alternative mechanisms (eg, attention and awareness) in relation to unique outcomes (eg, well-being), mental health strategies (ie, promotion) and active comparators. High-quality trials, with powered multivariate mediation analyses that meet key requirements, will be needed to advance this area of work. Trial registration number 10.17605/OSF.IO/NY2AH.

Overview

CoChat Summary

TL;DR

This scoping review of 27 randomised controlled trials found that mindfulness skills, decentering and attitudes of mindfulness (eg, self-compassion) are the most consistently supported indirect-effect mechanisms of mindfulness-based programmes and practices, with the strongest evidence in mental languishing outcomes and prevention strategies. Evidence for other mechanisms (eg, attention and awareness), for well-being outcomes, promotion strategies and active comparators remains preliminary or mixed.

Key findings

  • 27 eligible studies were identified; only four evaluated a formal mindfulness practice (1 body scan, 1 mindfulness of emotions, 1 breath, 1 body scan and breath), while 23 evaluated an MBP.
  • Significant indirect effects were most consistently reported for mindfulness skills, decentering and attitudes of mindfulness, with the largest proportion of significant to non-significant/opposite-direction IEs in mental languishing outcomes, mixed findings for mental health conditions and preliminary evidence for well-being.
  • By mental health strategy, the largest proportion of significant IEs was in prevention (16 studies), with mixed evidence in treatment (11 studies) and preliminary evidence in promotion (9 studies).
  • Of the 27 studies, only 13 fully reported the direct effect and the signs of the a×b and c paths needed to classify mediation: 5 reported complementary mediation, 10 indirect-only, 3 direct-only effects and 1 no effect; none reported competitive mediation. Quality ratings (0–30) were high for reporting the IE (22) and plausibility/coherence (25.5), intermediate for timeline (12.5) and low for gradient (6).

Methods

  • Scoping review following PRISMA-ScR and Arksey/O'Malley and JBI frameworks, preregistered on OSF (DOI: 10.17605/OSF.IO/XJDSU; addendum 10.17605/OSF.IO/NY2AH), with searches run up to 7 March 2024.
  • Five databases (PubMed, PsycINFO, Embase, Scopus, Cochrane Central) plus grey literature and snowballing; 43 392 records identified, 25 394 duplicates removed, 18 004 title/abstract screened, 1289 full texts screened, 30 records included, 3 sharing samples, leaving 27 studies.
  • Inclusion required adult samples (18+), an MBP or formal mindfulness practice, an RCT design with a comparator having active ingredients (more than passage of time), and evaluation of indirect effects via psychological mediators; data extraction was piloted and 15% cross-checked, with narrative synthesis and evidence gap maps by outcome, mental health strategy and active comparator.

Limitations

  • Stated: evidence gap maps captured multiple findings from the same study and many studies spanned multiple categories, which could inflate the evidence; consensus is still needed on operational definitions of mindfulness and overlapping processes (attention and awareness, attitudes of mindfulness, decentering).
  • Stated: including only RCTs with active comparators may have skewed results toward clinical samples and high-risk approaches (treatment, indicated prevention); most studies used TAU controls, which can be flawed by resentful demoralisation, and TAU was often embedded within the mindfulness condition, complicating isolation of mindfulness-specific effects.
  • Stated: qualitative studies and alternative designs (eg, quasi-experimental) were excluded to isolate the highest-quality mediation evidence, though these may identify additional mechanisms; evident from the text, only 12 studies met the timeline criterion and only 5 met the gradient criterion, limiting confidence in temporality and dose–response.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1136/bmjment-2023-300955

Cites

40 of 40 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Figure 1 PRISMA flow chart. The PRISMA flow diagram of the records initially identified from the database search before and after deduplication is illustrated. The chart also outlines the records screened at the title and abstract stage as well as the full-text stage, including the number of records removed after full-text screening with the reasons specified. The additional records found outside of the search (n=6) were identified from either forward or backward citation searching of the included records and relevant reviews, from grey literature sources or from corresponding authors who shared additional records. Out of the 30 records included, 3 records shared the same sample with another record and therefore there were 27 studies included in total. Note that both English and Spanish publications were considered, given the primary languages of the research team. However, all included studies were published in English and, therefore, no Spanish publications met the inclusion criteria. RCT, randomised controlled trial.
Figure 1 PRISMA flow chart. The PRISMA flow diagram of the records initially identified from the database search before and after deduplication is illustrated. The chart also outlines the records screened at the title and abstract stage as well as the full-text stage, including the number of records removed after full-text screening with the reasons specified. The additional records found outside of the search (n=6) were identified from either forward or backward citation searching of the included records and relevant reviews, from grey literature sources or from corresponding authors who shared additional records. Out of the 30 records included, 3 records shared the same sample with another record and therefore there were 27 studies included in total. Note that both English and Spanish publications were considered, given the primary languages of the research team. However, all included studies were published in English and, therefore, no Spanish publications met the inclusion criteria. RCT, randomised controlled trial.CC BY · PubMed Central
Figure 2 The evidence gap map for mechanisms of MBP and practices split by outcomes (mental health conditions: anxious, depressive and adjustment disorders; psychosomatic disorders; behavioural disorders; mental languishing: mental health, and stress and burnout; and well-being: positive affect, positive states of mind, quality of life and well-being) is illustrated. This evidence map reflects the key findings (related to the indirect effects in univariate mediation models) reported in online supplemental materials. Indirect effects (IEs) in multivariate mediation models are included in online supplemental materials. The IEs in the context of univariate models are prioritised in the evidence maps for conceptual clarity as multivariate models demonstrate the extent to which the mediated effect is dependent on the interaction between multiple mediators and the extent to which some make significant independent contributions, which are nuances that could not be captured here but are provided in online supplemental materials. For studies that reported total scores and subscale scores, total scores were prioritised. For studies whereby the proposed mechanisms were explored pre–post and pre follow-up, the pre–post results were prioritised. Online supplemental materials show the list of mechanism categories (mindfulness skills, attention and awareness, decentering, attitudes of mindfulness, emotional regulation, and dysfunctional thoughts, attitudes, and behaviours) and outcome categories and provides a working definition for each category and the full list of measures (with item examples) from the included papers that map onto each category. The total number of non-significant and significant indirect effects are reported, along with significant opposite direction findings. The size of the circle corresponds to the number of findings with larger circles signifying more findings. The opposite direction findings are indicated when the results provided support for the active comparator (eg, cognitive–behavioural therapy) rather than for the mindfulness condition. To help isolate mechanisms of change, and in light of the limited evidence base for individual mindfulness practices, the evidence gap maps combined the evidence across MBPs and practices. Note that multiple findings can be reported from one study. Studies therefore had multiple aims or addressed multiple outcomes that fell under different categories. =Significant findings; =Non-significant findings; =Significant findings in opposite direction; =Less evidence; =more evidence. N, number of studies.
Figure 2 The evidence gap map for mechanisms of MBP and practices split by outcomes (mental health conditions: anxious, depressive and adjustment disorders; psychosomatic disorders; behavioural disorders; mental languishing: mental health, and stress and burnout; and well-being: positive affect, positive states of mind, quality of life and well-being) is illustrated. This evidence map reflects the key findings (related to the indirect effects in univariate mediation models) reported in online supplemental materials. Indirect effects (IEs) in multivariate mediation models are included in online supplemental materials. The IEs in the context of univariate models are prioritised in the evidence maps for conceptual clarity as multivariate models demonstrate the extent to which the mediated effect is dependent on the interaction between multiple mediators and the extent to which some make significant independent contributions, which are nuances that could not be captured here but are provided in online supplemental materials. For studies that reported total scores and subscale scores, total scores were prioritised. For studies whereby the proposed mechanisms were explored pre–post and pre follow-up, the pre–post results were prioritised. Online supplemental materials show the list of mechanism categories (mindfulness skills, attention and awareness, decentering, attitudes of mindfulness, emotional regulation, and dysfunctional thoughts, attitudes, and behaviours) and outcome categories and provides a working definition for each category and the full list of measures (with item examples) from the included papers that map onto each category. The total number of non-significant and significant indirect effects are reported, along with significant opposite direction findings. The size of the circle corresponds to the number of findings with larger circles signifying more findings. The opposite direction findings are indicated when the results provided support for the active comparator (eg, cognitive–behavioural therapy) rather than for the mindfulness condition. To help isolate mechanisms of change, and in light of the limited evidence base for individual mindfulness practices, the evidence gap maps combined the evidence across MBPs and practices. Note that multiple findings can be reported from one study. Studies therefore had multiple aims or addressed multiple outcomes that fell under different categories. =Significant findings; =Non-significant findings; =Significant findings in opposite direction; =Less evidence; =more evidence. N, number of studies.CC BY · PubMed Central
Figure 3 Evidence map for mechanisms of MBP and practices split by mental health strategy (treatment, prevention and promotion) is provided. The different types of mental health strategies are reported in online supplemental materials. This evidence map reflects the key findings (related to the indirect effects in univariate mediation models) reported in online supplemental materials. Indirect effects (IEs) in multivariate mediation models are included in online supplemental materials. The IEs in the context of univariate models are prioritised in the evidence maps for conceptual clarity as multivariate models demonstrate the extent to which the mediated effect is dependent on the interaction between multiple mediators and the extent to which some make significant independent contributions, which are nuances that could not be captured here but are provided in online supplemental materials. For studies that reported total scores and subscale scores, total scores were prioritised. For studies whereby the proposed mechanisms were explored pre–post and pre follow-up, the pre–post results were prioritised. Online supplemental materials shows the list of mechanism and outcome categories and also provides a working definition for each category and the full list of measures (with item examples) from the included papers that map onto each category. The total number of non-significant and significant indirect effects are reported, along with significant opposite direction findings. The size of the circle corresponds to the number of findings with larger circles signifying more findings. The opposite direction findings are indicated when the results provided support for the active comparator (eg, cognitive–behavioural therapy) rather than for the mindfulness condition. To help isolate mechanisms of change, and in light of the limited evidence base for individual mindfulness practices, the evidence gap maps combined the evidence across MBPs and practices. Note that multiple findings can be reported from one study. =Significant findings; =Non-significant findings; =Significant findings in opposite direction; =Less evidence; =more evidence. N, number of studies.
Figure 3 Evidence map for mechanisms of MBP and practices split by mental health strategy (treatment, prevention and promotion) is provided. The different types of mental health strategies are reported in online supplemental materials. This evidence map reflects the key findings (related to the indirect effects in univariate mediation models) reported in online supplemental materials. Indirect effects (IEs) in multivariate mediation models are included in online supplemental materials. The IEs in the context of univariate models are prioritised in the evidence maps for conceptual clarity as multivariate models demonstrate the extent to which the mediated effect is dependent on the interaction between multiple mediators and the extent to which some make significant independent contributions, which are nuances that could not be captured here but are provided in online supplemental materials. For studies that reported total scores and subscale scores, total scores were prioritised. For studies whereby the proposed mechanisms were explored pre–post and pre follow-up, the pre–post results were prioritised. Online supplemental materials shows the list of mechanism and outcome categories and also provides a working definition for each category and the full list of measures (with item examples) from the included papers that map onto each category. The total number of non-significant and significant indirect effects are reported, along with significant opposite direction findings. The size of the circle corresponds to the number of findings with larger circles signifying more findings. The opposite direction findings are indicated when the results provided support for the active comparator (eg, cognitive–behavioural therapy) rather than for the mindfulness condition. To help isolate mechanisms of change, and in light of the limited evidence base for individual mindfulness practices, the evidence gap maps combined the evidence across MBPs and practices. Note that multiple findings can be reported from one study. =Significant findings; =Non-significant findings; =Significant findings in opposite direction; =Less evidence; =more evidence. N, number of studies.CC BY · PubMed Central
Figure 4 The evidence gap map for mechanisms of MBPs and practices split by type of active comparator is illustrated. Five categories for different types of active comparators were identified: treatment as usual (TAU), stress management and relaxation (SM/R), cognitive behavioural therapy/cognitive interventions (CBT/CI), attentional control (AC) and self-help mindfulness (SH-M). This evidence map reflects the key findings (related to the indirect effects in univariate mediation models) reported in online supplemental materials. Indirect effects (IEs) in multivariate mediation models are included in online supplemental materials. The IEs in the context of univariate models are prioritised in the evidence maps for conceptual clarity as multivariate models demonstrate the extent to which the mediated effect is dependent on the interaction between multiple mediators and the extent to which some make significant independent contributions, which are nuances that could not be captured here but are provided in online supplemental materials. For studies that reported total scores and subscale scores, total scores were prioritised. For studies whereby the proposed mechanisms were explored pre–post and pre follow-up, the pre–post results were prioritised. Online supplemental materials show the list of mechanism and outcome categories and also provides a working definition for each category and the full list of measures (with item examples) from the included papers that map onto each category. The total number of non-significant and significant indirect effects are reported, along with significant opposite direction findings. The size of the circle corresponds to the number of findings with larger circles signifying more findings. The opposite direction findings are indicated when the results provided support for the active comparator (eg, cognitive–behavioural therapy) rather than for the mindfulness condition. To help isolate mechanisms of change, and in light of the limited evidence base for individual mindfulness practices, the evidence gap maps combined the evidence across MBPs and practices. Note that multiple findings can be reported from one study. =Significant findings; =Non-significant findings; =Significant findings in opposite direction; =Less evidence; =Less evidence. N, number of studies.
Figure 4 The evidence gap map for mechanisms of MBPs and practices split by type of active comparator is illustrated. Five categories for different types of active comparators were identified: treatment as usual (TAU), stress management and relaxation (SM/R), cognitive behavioural therapy/cognitive interventions (CBT/CI), attentional control (AC) and self-help mindfulness (SH-M). This evidence map reflects the key findings (related to the indirect effects in univariate mediation models) reported in online supplemental materials. Indirect effects (IEs) in multivariate mediation models are included in online supplemental materials. The IEs in the context of univariate models are prioritised in the evidence maps for conceptual clarity as multivariate models demonstrate the extent to which the mediated effect is dependent on the interaction between multiple mediators and the extent to which some make significant independent contributions, which are nuances that could not be captured here but are provided in online supplemental materials. For studies that reported total scores and subscale scores, total scores were prioritised. For studies whereby the proposed mechanisms were explored pre–post and pre follow-up, the pre–post results were prioritised. Online supplemental materials show the list of mechanism and outcome categories and also provides a working definition for each category and the full list of measures (with item examples) from the included papers that map onto each category. The total number of non-significant and significant indirect effects are reported, along with significant opposite direction findings. The size of the circle corresponds to the number of findings with larger circles signifying more findings. The opposite direction findings are indicated when the results provided support for the active comparator (eg, cognitive–behavioural therapy) rather than for the mindfulness condition. To help isolate mechanisms of change, and in light of the limited evidence base for individual mindfulness practices, the evidence gap maps combined the evidence across MBPs and practices. Note that multiple findings can be reported from one study. =Significant findings; =Non-significant findings; =Significant findings in opposite direction; =Less evidence; =Less evidence. N, number of studies.CC BY · PubMed Central
Figure 5 Flow chart of included studies by mediation type and reporting requirements. The flow chart of the included studies is illustrated based on meeting key steps for reporting and categorising mediation: (1) the significance of path a×b, (2) the significance of path c and (3) the sign of paths a×b and c. All included studies reported the significance of the a×b path as this was a key inclusion criteria. From there, the number of included studies (n) that reported the significance of path c was recorded, followed by the number of included studies that reported the sign of the a×b and c paths. At the bottom of the flow chart, we have the total number of studies for each category of mediation. Note that for studies that partially demonstrated the significance of path c and partially demonstrated the signs of a×b and c, this meant that these papers met the criteria for some findings but not for all. Only 13 studies fully met these criteria, and therefore mediation categories were determined. This figure reflects key findings in online supplemental 13 and was adapted from Zhao et al.13 .
Figure 5 Flow chart of included studies by mediation type and reporting requirements. The flow chart of the included studies is illustrated based on meeting key steps for reporting and categorising mediation: (1) the significance of path a×b, (2) the significance of path c and (3) the sign of paths a×b and c. All included studies reported the significance of the a×b path as this was a key inclusion criteria. From there, the number of included studies (n) that reported the significance of path c was recorded, followed by the number of included studies that reported the sign of the a×b and c paths. At the bottom of the flow chart, we have the total number of studies for each category of mediation. Note that for studies that partially demonstrated the significance of path c and partially demonstrated the signs of a×b and c, this meant that these papers met the criteria for some findings but not for all. Only 13 studies fully met these criteria, and therefore mediation categories were determined. This figure reflects key findings in online supplemental 13 and was adapted from Zhao et al.13 .CC BY · PubMed Central
Figure 6 Conceptual framework of target mechanisms. A conceptual framework of target mechanisms that may be responsible for shifting change across the entire population distribution (eg, from mental health conditions to mental languishing to well-being) is illustrated. In this figure, the ‘red’ region on the left-hand side represents outcomes pertaining to mental health conditions (eg, anxious, depressive and adjustment disorders; psychosomatic disorders; behavioural disorders, etc). The ‘orange-to-yellow’ region represents outcomes pertaining to mental languishing (eg, mental health symptoms, stress, burnout, etc). The ‘blue’ region represents outcomes pertaining to well-being (eg, mental well-being, quality of life, positive affect, positive states of mind, etc). The grey arrow represents the proposed mechanisms of change (ie, mindfulness skills, decentering and attitudes of mindfulness). This conceptual framework was informed by the results of the scoping review and summarises key hypotheses regarding the ‘how’ and ‘why’ mindfulness-based programmes and practices work across the whole population distribution. However, future research will need to further interrogate this conceptual framework in different contexts (eg, mental health strategies, ie, promotion), outcomes (eg, well-being) and active comparators (eg, self-help mindfulness, attentional control). Illustration by Delphine Perrot.
Figure 6 Conceptual framework of target mechanisms. A conceptual framework of target mechanisms that may be responsible for shifting change across the entire population distribution (eg, from mental health conditions to mental languishing to well-being) is illustrated. In this figure, the ‘red’ region on the left-hand side represents outcomes pertaining to mental health conditions (eg, anxious, depressive and adjustment disorders; psychosomatic disorders; behavioural disorders, etc). The ‘orange-to-yellow’ region represents outcomes pertaining to mental languishing (eg, mental health symptoms, stress, burnout, etc). The ‘blue’ region represents outcomes pertaining to well-being (eg, mental well-being, quality of life, positive affect, positive states of mind, etc). The grey arrow represents the proposed mechanisms of change (ie, mindfulness skills, decentering and attitudes of mindfulness). This conceptual framework was informed by the results of the scoping review and summarises key hypotheses regarding the ‘how’ and ‘why’ mindfulness-based programmes and practices work across the whole population distribution. However, future research will need to further interrogate this conceptual framework in different contexts (eg, mental health strategies, ie, promotion), outcomes (eg, well-being) and active comparators (eg, self-help mindfulness, attentional control). Illustration by Delphine Perrot.CC BY · PubMed Central

Verification

Claims the paper makes

9 Valid · 0 Not valid · 9 Not checked · 45 Own results and interpretations

29 more claims

Create a free account to see all 63 claims and to verify the 3 references still open.

See all claims — free account
  • It may state two separate claims as one. 21
  • Probably not a substantial claim. 10

Create a free account to see these claims.

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

1 of 1

Citations

Cites

Show all 40 references

Research works printed but not linked: 1

  1. Call D . Effectiveness of mindfulness techniques in reducing symptoms of anxiety-related worry [PhD Thesis]. Northern Illinois University, 2014

Cited by

See all 22

Metadata

Authors
Maloney, S., Kock, M., Slaghekke, Y., Radley, L., Lopez-Montoyo, A., Montero-Marin, J., & Kuyken, W.
Affiliations
  • University of Oxford (GB)
  • KU Leuven (BE)
  • Universitat Jaume I, Castello de la Plana, Comunitat Valenciana, Spain
  • Department of Psychiatry, University of Oxford, Oxford, Oxfordshire, UK
Corresponding author
Maloney, Shannon
Venue
BMJ Mental Health
Publisher
BMJ
Published
2024-01
Volume
27
Issue
1
Pages
e300955
ISSN
2755-9734 (ISSN-L)
OpenAlex
W4401849308
Cited by
22
References
41
Citations by year
2026: 10 · 2025: 7 · 2024: 3
FWCI
1.77
Type
journal article
Language
en
Open access
Locations
  • BMJ Mental Health (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • Lirias (submittedVersion, cc-by)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
  • Wellcome Trust (WT104908/Z/14/Z and WT107496/Z/15/Z)
  • National Institute for Health and Care Research (NIHR) Oxford Health Biomedical Research Centre (NIHR203316)
Topics
  • Social Sciences › Psychology › Clinical Psychology › Mindfulness and Compassion Interventions
  • Social Sciences › Psychology › Social Psychology › Psychological Well-being and Life Satisfaction
  • Social Sciences › Psychology › Clinical Psychology › Psychotherapy Techniques and Applications
Keywords
mindfulness practice, indirect effects, mental health, well-being, decentering, self-compassion, mindfulness skills, attention and awareness, evidence gaps
MeSH terms
Humans, Mental Health, Randomized Controlled Trials as Topic, Mindfulness

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-03, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-03 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (qwen3.8-omni-flash and gpt-6-sol).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Target mechanisms of mindfulness-based programmes and practices: a scoping review. https://app.cochat.ai/papers/10.1136/bmjment-2023-300955
The paper
Maloney, S., Kock, M., Slaghekke, Y., Radley, L., Lopez-Montoyo, A., Montero-Marin, J., & Kuyken, W. (2024). Target mechanisms of mindfulness-based programmes and practices: a scoping review. BMJ Mental Health. https://doi.org/10.1136/bmjment-2023-300955

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.