Spain

Several countries partly fund universities on the basis of scientific performance. In this context, peer review is often used to assess the quality of scientific products. In large research assessment exercises peer review may become costly, and bibliometric indicators have been suggested as an alternative. A recurrent question is whether peer review and metrics tend to yield similar outcomes. In this paper, we study the agreement between bibliometric indicators and peer review at the institutional level. Additionally, we also quantify the internal agreement between two peer reviewers. We find that the level of agreement is generally higher at the institutional level than at the publication level. Overall, the agreement between metrics and peer review is on par with the internal agreement among two peer reviewers. This suggests that bibliometric indicators may indeed be considered as an alternative to peer review for large institutional research assessment exercises.

PEERE

Metrics and peer review agreement at the institutional level

national evaluation exercise

performance-based research funding

agreement

peer review

metrics

Following the success of the 1st PEERE International Conference on Peer Review, the PEERE COST Action follow-up community organised the second edition of the PEERE International Conference on Peer Review  and was celebrated virtually on 29 September – 1 October 2020.

The conference welcomed any contribution (e.g., original research, position papers, literature reviews) that used any method (e.g., quantitative, experimental or qualitative) to investigate peer review in a variety of scientific domains (e.g., scholarly journals, funding agencies, research assessment). Interdisciplinary research is especially welcome.

The program is published here: http://tiny.cc/peere2020 


The second edition of the PEERE International Conference on Peer Review aims to provide a forum for scholars, practitioners and science stakeholders to share research, models, tools and experience on peer review in different fields, e.g., medicine, computer science, social sciences and humanities. It aims to stimulate the use of evidence-based research in the design and implementation of peer review in a variety of fields and encourage more systematic research.

technical paper

A Content Exploration of Reviewers' Comments in FP7 Marie Curie ITN evaluation reports 
Darko Hren1, David G. Pina2, Cristopher Norman3, Ana Marušić4

1University of Split Faculty of Humanities and Social Sciences, Split, Croatia
2Research Executive Agency, European Commission, Brussels, Belgium
3LIMSI, CNRS, Université Paris Saclay, Paris, France
4University of Split School of Medicine, Split, Croatia

Objectives:
To identify themes in the content of reviewers' evaluation reports about proposals submitted to the FP7 Marie Curie Initial Training Networks calls and to explore the content areas in reviewers' comments that differentiate successful from unsuccessful proposals.

Methods:
We qualitatively analyzed the contents of reviewers' comments to obtain the most common themes within evaluation reports. Given a large amount of textual data, we used a two-step approach.
The first step was automated to help reduce the data to the manageable size. We used K-Means clustering to identify the clusters of statements within sections of the evaluation reports. Clusters were visualized in 2-dimensional space using TensorBoard Embedding Projector (https://projector.tensorflow.org/). Principal components analysis (PCA) was used for the visualization of clusters and cosine distances between clusters as a reference to cluster distances.
In the second step, the clusters were further grouped into themes using empirical estimates from PCA in step 1 and qualitative analysis based on cluster terms and content. To identify clusters that could potentially be grouped into themes, we used cosine distances lower than 1 as an arbitrary indicator of cluster proximity. Next, we inspected the content of the obtained clusters (terms and example statements) to verify whether and how groups of clusters could be interpreted as broader themes. Inspection of content had precedence over empirical data, i.e. if there was no meaningful overlap in cluster terms and content, even clusters that had cosine distance below 1 were not combined.
Finally, we registered the frequencies of obtained themes within each evaluation report (i.e. a number of sentences containing terms belonging to a specific theme) to inspect potential differences between proposals with different decision status.

Results:
We analyzed the content of reviews of 3666 grant proposals submitted to Marie Curie Initial Training Networks from the FP7 PEOPLE programme in 2008 (n=886), 2010 (n=857), 2011 (n=909), and 2012 (n=1013).
Each review had four sections of criteria (C1-Science & Technology Quality, C2-Training, C3-Implementation, and C4-Impact), and each section was divided into two parts (Strengths and Weaknesses). Therefore, 12 separate K-Means cluster analyses were performed in the first step, and we identified between 9 and 12 clusters explaining between 32.7 and 51.6% of the variance. 
In the second step, we identified between 29 themes within different sections of the evaluation reports (Figure 1A-D). Evaluation reports of proposals from the main and reserve lists included more themes related to strengths and fewer themes related to weaknesses than those of rejected proposals. Themes related to strengths were generally more frequent than those related to weaknesses. Most prominent themes related to strengths were: Research programme and objectives and Multidisciplinary context (C1), Consistency of the research and training programme and Career development (C2), Capacities of research and training (C3), and Outreach activities and Career prospects (C4). Most prominent themes related to weaknesses were: Research programme and objectives and Research methodology (C1), Engagement of ESRs and Training characteristics (C2), Capacities of research and training (C3), and Exploitation and dissemination (C4).

Conclusions:
We used a novel approach which combined machine learning and qualitative inquiry to identify the most common themes of the evaluation reports of proposals submitted to the FP7 Marie Curie Initial Training Networks calls. The themes reflected general evaluation criteria applied by the European Commission, but also specific themes that distinguished successful from unsuccessful proposals thus demonstrating the validity of the selection process.
The obtained results may be used by researchers preparing to apply for grants.



A Content Exploration of Reviewers' Comments in FP7 Marie Curie ITN evaluation reports

eLife launched a trial peer-review process in June 2018, designed to give authors greater control over the decision to publish. Authors opted into the trial process when submitting their manuscript for consideration and, once the eLife editors decided that a submission would be sent for in-depth peer review, the journal was committed to publishing the paper. Authors chose how to address the concerns raised during review and, upon receiving the revised submission, editors rated whether these concerns had been addressed. This editorial rating was published with the paper, as were the decision letter after peer review, the separate reviews, and the accompanying author responses. The trial process was available until more than 300 authors opted in (in total, 313 opted in) and we compare these with 616 papers submitted into the regular process during the same period of time. Overall, the trial review process resulted in a significantly higher acceptance rate (20.8% compared with 14.9% in the regular process), with all the issues raised by reviewers addressed by the authors in the vast majority (88%) of revised submissions. We observed no significant difference in acceptance rates for male and female last authors in the trial process (21% and 20.6% respectively). However, we observed marginally higher acceptance rates for late-career researchers (32%) in the trial compared to their early- (22%) and mid-career (19%) colleagues and significantly higher acceptance rates for last authors based in North America (29.2%) and Europe (18.1%) compared to last authors based in Asia (11.9%). Trial submissions took significantly longer to be accepted (median = 187 days, IQR = 130.5-235.5) compared to the regular process (median = 138 days, IQR =114-170).

Comparison of two types of collaborative peer review at eLife

A comparative secondary data analysis was conducted on reviewers‘ answers to peer review questionnaires from two different journals from the life sciences (5382 reports in total). The correspondence of the answering patterns to peer review criteria sets with the reviewers’ recommendations was investigated, as well as the peer reviewers’ disagreement about manuscript quality in the respective dimensions. The analyses show a high internal consistency for the individual opinion questionnaires with the resulting recommendation. Furthermore, high reviewers’ disagreement was observed on the given criteria as well as the overall recommendation. Taken all together, the results indicate, that the norm of disinterestedness is reflected in peer reviewers' behavior and further depends on editorial decisions.

Consistency and disagreement in journal peer review and the norm of disinterestedness

Peer-review through our viewpoint is understood to be characterised by heterogeneity. In particular, reviewers are generally seen to be moved by a variety of reasons. We argue that some may have a burning quest for better science. In such cases, if it is seen worth doing so, they might try to convince the authors to steer the paper in the directions they deem necessary, with their reviews. In a quest for ever-better science they may, for example, produce extremely detailed and long reviews where they attempt at showing the “right” way to science to the authors. In this way, we may say, they attempt to ‘hack’ the paper through their review. This work studies the effect of this hacker-reviewer phenomenon using a modified version of the PRIF agent-based model (Secchi & Cowley, 2019). The aim of the model is that of exploring the effects of reviewer-triggered influence on submitted journal articles when the reviewer acts as a hacker.


Hacking one's way to peer-review

The Horizon 2020 Marie Skłodowska-Curie Actions (MSCA) Individual Fellowships (IF) funding programme is one of the largest and most competitive EU research programme (around 15% success rate). It would be expected that the applications of funded proposals have detected ethical issues well. The aim of this study was to describe the ethics evaluation process in the MSCA IF funding scheme . We first assessed the ethical issues identified by the applicants when preparing their proposals and compared these to the ethics issues identified by ethics experts during the ethics evaluation performed for those proposals retained for funding. In total, we analyzed 3074 proposals from three MSCA IF calls ( 2016, 2017 and 2018). Applicants identified almost two times fewer ethical issues compared to the experts performing the ethics review. Most of the ethics issues were identified in proposals from the Life Sciences and Social Sciences and Humanities research panels, which represented more than 60% of all proposals. Out of all 3074 proposals which were main listed and accepted for funding, 2225 (72.4%) received conditional ethics clearance, pending satisfaction of requirements defined by the ethics experts. The most prevalent requirements were within the Protection of personal data, Environment, Health and Safety, Humans, and non-EU countries categories. Our results emphasize that successful applicants to the MSCA IF programme, consisting mostly of postdoctoral researchers, are not adequately trained or knowledgeable about the ethics issues related to their proposed research, and that problem persists over time. Training in research ethics and proper guidance by the host institutions are paramount for narrowing the gap between the applicants’ ethics self-assessments and the results of the ethics review procedure. 


Ethics assessment of H2020 Marie Skłodowska-CUrie individual fellowships

ARCADIA (Assessment of Review reports with a Checklist Available to eDItors and Authors) constitutes the first tool that has been systematically developed to assess the quality of peer review reports in biomedical research. It is a checklist which consists of five domains covering 14 items. In the presentation, we will show the psychometric properties of the newly developed ARCADIA tool.

Psychometric testing of ARCADIA, a tool for assessing peer review report quality in biomedical research

Effect of an editorial intervention to improve the completeness of reporting of randomised trials: a randomised controlled trial

A peer review programme for mental health service users
Objective
Service user involvement should be an essential component of medical research, and many funders are requiring evidence in funding applications that expertise by experience is included. As editors, while we highly value peer review by researchers with lived experience of mental illness, we are concerned that individuals without formal training in research methods are not given the opportunity to express their opinion on the work we publish. Our objective was to conduct a pilot programme to explore how people with lived experience but without research qualifications or an employment contract at a university can contribute to peer review.
Methods
The Lancet Psychiatry launched in 2018 a peer review programme for mental health service users in collaboration with the McPin Foundation, a charity based in London, UK that champions expertise from experience in mental health research. Participants who have used mental health services in the UK were recruited via on-line advertisements, The Lancet Psychiatry and service user news bulletins. Applications were reviewed by a panel from the McPin Foundation and Lancet Psychiatry. Selected individuals received training in peer review in London, UK, and online from The Lancet Psychiatry editorial team supported by McPin staff. Travel costs for an annual meeting are funded by the McPin Foundation, and The Lancet Psychiatry pays for each review. 
Results
48 people applied. McPin staff produced a short list of 20 candidates. Lancet Psychiatry editors and McPin staff then selected ten participants: eight women and two men. 9 participants attended a training day at the Lancet office in December 2018. All participants were encouraged to complete the online training in peer review provided by Publons with a Lancet Psychiatry editor as their mentor. 7 participants (6 women, 1 man) had completed this by mid-August 2019 and have started reviewing manuscripts for The Lancet Psychiatry. They are being invited to review manuscripts in areas such as clinical trials, service delivery, observational studies, qualitative research, health policy, systematic reviews, and opinion pieces. 


A peer review programme for mental health service users

We explore which characteristics of a reviewer affect disagreement with the other reviewers in research proposal evaluation, studying 52,488 proposals submitted to the Horizon 2020 Marie Skłodowska-Curie Actions (MSCA) and 1,939 proposals to the European Cooperation in Science and Technology Actions (COST). We find that experience in previous calls significantly reduces disagreement, and in MSCA - IF, an inverted U relationship between the number of proposals evaluated in a given call and disagreement, with a remarkable decrease of disagreement above thirteen proposals evaluated. The results suggest that reviewing experience improve reliability, curbing unwarranted disagreement by gradually fine-tuning reviewers’ evaluation.

Does reviewers' traits affect disagreement in proposal evaluation?

The organization of peer review processes in grant proposals evaluation is rarely studied. We analysed 77 thousand proposals from the Marie Curie programme, submitted over a period of 12 years, with the aim to get insights on the reliability of the process. Mean scores, Average Deviation (AD) indices, used as a measure of inter-rater agreement, and interrupted time series analysis were calculated to study changes in reviewers’ scores and level of agreement. Our results indicate that scores given by and agreement among reviewers may be more dependent on the type of grants or scientific panels than on the way the evaluation process is organised, with important economic and environmental implications for policy-making in research grants evaluation.

Metrics and peer review agreement at the institutional level

Downloads

Next from PEERE

A Content Exploration of Reviewers' Comments in FP7 Marie Curie ITN evaluation reports

Similar lecture

Feedback to improve agreement in grant peer review

Metrics and peer review agreement at the institutional level

.css-70qvj9{display:-webkit-box;display:-webkit-flex;display:-ms-flexbox;display:flex;-webkit-align-items:center;-webkit-box-align:center;-ms-flex-align:center;align-items:center;}Downloads

Next from PEERE

A Content Exploration of Reviewers' Comments in FP7 Marie Curie ITN evaluation reports

Similar lecture

Feedback to improve agreement in grant peer review

Downloads