A few weeks ago the European Commission published a list with all the names of evaluators they used during the first year of H2020. The list contains the names of all evaluators and the H2020 specific work programme in which they judged proposals (but not the topics!). In addition it shows the “Skills & Competences” of each evaluator as he/she wrote it into the EC’s Expert Database.
It’s not the first time the EC publishes this type of list, of course, but the publication comes at a poignant moment. As success rates for submitted proposals are rapidly declining and many researchers are wondering whether it is still worth developing anything for H2020 at all. In previous articles I already wrote about some of the main reasons behind this decline and possible solutions.
In this article I want to focus on the actual H2020 evaluation process, as for many researchers it is becoming very frustrated to write top-level proposals (as evidenced by scores well over the threshold and oftentimes higher than 13.5 points) and then see their efforts go under in short, sometimes seemingly standardised, comments and descriptions. Is it just the brutal trade-off between the increasing number of applications versus a limited budget, or do the evaluation comments and points hide another side to the process? Why are some people starting to call the H2020 evaluation process a ‘lottery’? Is there some truth to their criticism? How can the European Commission counter this perception?
I am sure most researchers have nothing against the EC’s basic principle of using a “science-beauty contest” to allocate H2020 funds. There is a lot of discussion however – also on social media – on selection process. Part of that discussion concerns the level of actual experience and specific expertise of those who do the selection. Simply put: Are the evaluators truly the most senior qualified experts in the H2020 domains they are judging proposals on, or is a – possibly significant – significant part of the panels comprised of a mix of younger researchers – using the evaluation process to learn about good proposal writing – and mid-level scientists who are good in their specific field, but also have to evaluate proposals that are (partly) outside their core competence?
Before I continue, please accept that this article is not intended in any way to slack off the expertise of people who have registered on the FP7 and H2020 Expert Database and have made valuable time available to read and judge funding proposals. I myself am in no position to make statements about what constitutes sufficient research expertise to be acceptable for peers when judging their proposals. The only point I will be making below is that if the EC wants to keep the H2020 evaluation process credible (and not stigmatised as a ‘lottery’), it needs to demonstrate to the community that it is selecting evaluators not just on availability but on their real understanding of what top-class research in a given research field means.
So let me continue with my argument: H2020 specifically states that it is looking for the most innovative ideas from our brightest researchers and developed by the best possible consortia. If that is the case, knowing that you cannot be an expert-evaluator if you are part of a consortium in that same funding Call, then already quite a few of the “best possible experts” will by default have disqualified themselves from participation in the evaluation process. It is also widely known that not all top-level experts want to involve themselves in evaluation either because to them it is not sufficiently important or because of other (time-)constraints. As a result, groups of typically 3-5 evaluators will typically consist a combination of real experts in that specific research domain and other evaluators who come from adjacent or even (very) different disciplines. So part of a given evaluation committee may therefore consist of – for lack of a better word – what I call ‘best effort amateurs’. Again: I have no doubt these are good researchers in their own field, but at this point they may be asked to judge projects outside of their core-competence.
Now you might think I am making these observations because I want to build a quick case against the evaluation process as it is. That’s not true. What I said above are views and comments that have been made by many researchers, on and off-stage, both from academia and inside industry. It’s thát feeling of unease and not knowing, that contributes to the more general and growing perception that H2020 is turning into a somewhat lottery-type process. For that reason alone, it is very important that the Commission now shows that their choice of combinations of evaluators is based on specific merit and not on general availability.
One way to refute the perception of a lottery, is – in my mind – for the EC to perform a rigorous analysis (and subsequent publication!) of the scientific quality of the people it used as evaluators. How to do that? A start would be to assess each evaluator against the no. of publications ánd the no. of citations he/she has in a given research science field. You can include some form of weighting if required to allow for the type and relative standing of the journals in which the publications featured. The assessment could also include the number of relevant patents of researchers. This should give a fairly clear indication on the average level of research seniority among the evaluation panels. One issue when doing this, will be that very experienced evaluators from industry may have fewer published articles, as the no. of publications is often considered less important (or even less desired) by the companies they work for. So the criterium of publications and citations should be applied primarily to academic evaluators. The same thing essentially also holds for patents, as ‘industry patents’ often are registered and owned by the company the researchers work for (or have worked for in the past). In other words: the analysis will probably not deliver a perfect picture, but as academics make up the majority of the evaluation panels, it should still give a fairly good indication.
Next step could be to check if the evaluators were actually tasked to assess only projects in their core-research domain or if they also judged projects as ‘best effort amateurs’. The EC could do that my matching the mentioned publications/citations/patents overview against the list of “Skills & Competences” the researchers listed in the Expert Database and the specific H2020 topics they were asked to be an evaluation panel-member in. Once you know that, you can also check if the use of ‘best effort amateurs’ happened because there just were not sufficient available domain-specific top-experts, or if the choice was based on prior experience with or availability of particular evaluators. I would find it strange, should that be the case. After all: the EC database of experts exceeds 25.000 names. The final step – in my view – would then be for the Commission to publish the findings – the statistical data should of course be anonimised – so that the wide research community can see for itself whether the selection of proposals is based on senior research quality linked to the H2020 proposal domain, or not.
I am sure that I am not complete (or even fully scientifically correct) in my suggestions on how to analyse the H2020 evaluation process. Things like the number of proposals to score and the average time spent on reading individual proposals will propoably also have an effect on scores. So what is the time-pressure the evaluators are under when they do their assessment and are there possibilities to reduce that pressure?
Please take this article as an effort to trigger further discussion that will lead to appropriate action from the EC in avoiding that our top researchers start dismissing H2020 as a worthwhile route to facilitating scientific excellence.
So European Commission: are you up to the challenge? If not, I am sure that there will be somebody out there to pick up the glove… I will be most interested in the results. Undoubtedly to be continued…