A new approach to grant review assessments: score, then rank

BACKGROUND

In many grant review settings, proposals are selected for funding on the basis of summary statistics of review ratings. Challenges of this approach (including the presence of ties and unclear ordering of funding preference for proposals) could be mitigated if rankings such as top-k preferences or paired comparisons, which are local evaluations that enforce ordering across proposals, were also collected and incorporated in the analysis of review ratings. However, analyzing ratings and rankings simultaneously has not been done until recently. This paper describes a practical method for integrating rankings and scores and demonstrates its usefulness for making funding decisions in real-world applications.

METHODS

We first present the application of our existing joint model for rankings and ratings, the Mallows-Binomial, in obtaining an integrated score for each proposal and generating the induced preference ordering. We then apply this methodology to several theoretical "toy" examples of rating and ranking data, designed to demonstrate specific properties of the model. We then describe an innovative protocol for collecting rankings of the top-six proposals as an add-on to the typical peer review scoring procedures and provide a case study using actual peer review data to exemplify the output and how the model can appropriately resolve judges' evaluations.

RESULTS

For the theoretical examples, we show how the model can provide a preference order to equally rated proposals by incorporating rankings, to proposals using ratings and only partial rankings (and how they differ from a ratings-only approach) and to proposals where judges provide internally inconsistent ratings/rankings and outlier scoring. Finally, we discuss how, using real world panel data, this method can provide information about funding priority with a level of accuracy in a well-suited format for research funding decisions.

CONCLUSIONS

A methodology is provided to collect and employ both rating and ranking data in peer review assessments of proposal submission quality, highlighting several advantages over methods relying on ratings alone. This method leverages information to most accurately distill reviewer opinion into a useful output to make an informed funding decision and is general enough to be applied to settings such as in the NIH panel review process.

Do peers share the same criteria for assessing grant applications?

This study examines a basic assumption of peer review, namely, the idea that there is a consensus on evaluation criteria among peers, which is a necessary condition for the reliability of peer judgements. Empirical evidence indicating that there is no consensus or more than one consensus would offer an explanation for the disagreement effect, the low inter-rater reliability consistently observed in peer review. To investigate this basic assumption, we have surveyed all humanities scholars in Switzerland on 23 grant review criteria. We have employed latent class tree modelling to identify subgroups in which scholars rated criteria similarly (i.e. latent classes) and to explore covariates predicting class membership. We have identified two consensus classes, two consensus-close classes, and a consensus-far class. The consensus classes contain a core consensus (10 criteria related to knowledge gaps; feasibility; rigour; comprehensibility and argumentation; academic relevance; competence and experience of the applicant) and a broad consensus that include the core consensus plus eight contribution-related criteria, such as originality. These results provide a possible explanation for the disagreement effect. Moreover, the results are consistent with the notion of conservatism, which holds that original research is undervalued in peer review, while other aspects, such as methodology and feasibility, are overweighted. The covariate analysis indicated that age and having tenure increase from the consensus far to the consensus close to the consensus classes. This suggests that the more academic experience scholars accumulate, the more their understanding of review criteria conforms to the social norm.

Individual versus general structured feedback to improve agreement in grant peer review: a randomized controlled trial

BACKGROUND

Vast sums are distributed based on grant peer review, but studies show that interrater reliability is often low. In this study, we tested the effect of receiving two short individual feedback reports compared to one short general feedback report on the agreement between reviewers.

METHODS

A total of 42 reviewers at the Norwegian Foundation Dam were randomly assigned to receive either a general feedback report or an individual feedback report. The general feedback group received one report before the start of the reviews that contained general information about the previous call in which the reviewers participated. In the individual feedback group, the reviewers received two reports, one before the review period (based on the previous call) and one during the period (based on the current call). In the individual feedback group, the reviewers were presented with detailed information on their scoring compared with the review committee as a whole, both before and during the review period. The main outcomes were the proportion of agreement in the eligibility assessment and the average difference in scores between pairs of reviewers assessing the same proposal. The outcomes were measured in 2017 and after the feedback was provided in 2018.

RESULTS

A total of 2398 paired reviews were included in the analysis. There was a significant difference between the two groups in the proportion of absolute agreement on whether the proposal was eligible for the funding programme, with the general feedback group demonstrating a higher rate of agreement. There was no difference between the two groups in terms of the average score difference. However, the agreement regarding the proposal score remained critically low for both groups.

CONCLUSIONS

We did not observe changes in proposal score agreement between 2017 and 2018 in reviewers receiving different feedback. The low levels of agreement remain a major concern in grant peer review, and research to identify contributing factors as well as the development and testing of interventions to increase agreement rates are still needed.

TRIAL REGISTRATION

The study was preregistered at OSF.io/n4fq3 .