No one will disagree that research is supposed to benefit society, and that research and researchers should be evaluated for their contribution to this societal mission. It therefore comes as an inconvenient truth to those impact evaluation proponents that a fair and robust evaluation of societal impact needs to balance a celebration of positive impacts, while acknowledging those that do not “benefit” society.
Impact is different to everyone. As we have increasingly seen lately, a public swing towards new, populist political movements results in a public that is prone to confirmation bias,
distrusts public experts, is incapable of distinguishing between fake news and evidence, and therefore only values research that conforms to existing ideological conceptions.
In other words, research can have impact, but also can have Grimpact.
This research endeavour of ours started nearly 2 years ago on Twitter. Angered by the distortion of science to fulfil political aims, three of us started exchanging tweets about our knowledge of impact pathways, and our hypotheses about the nature of negative impact. Together we developed Grimpact, an exotic take on impact that acknowledges that research is mostly introduced with good intentions but also can, like a weed, spiral out of control.
If you want to know what Grimpact looks like, take three extreme examples. First, the publication of a paper in the Lancet in 2005, later retracted, which led to rejection of the MMR vaccine and resulting deaths from Measles. In addition, along this grimpact pathways exists the generation of anti-vaccine advocacy groups; political campaigns based on fear; a rejection of expertise over anecdotal experience; and a movie screened at the Cannes Film Festival promoted by some of the most prominent Hollywood actors. This is the most recognised example of grimpact, but other examples include the generation of a tool that was capable of discerning emotions of users on social media platforms. In the case of Cambridge Analytica, the grimpact started when ethics boundaries were violated by both researchers and users, who then used the tool for political gain as opposed to public benefit. Finally, a more nuanced example is how the deregulation of financial services and the resulting behaviour of banks underpinning the 2008 financial crisis was validated by an overarching belief in economists and economic theory that promoted free-market, laissez-faire approaches.
These are research grimpacts that the academy does not acknowledge nor promote because there is a political advantage in accepting that research only benefitssociety. Research gains the majority of support and funding based on the assumption that it can add value to the public. This is also reflected in the majority of impact definitions from countries such as the UK, The Netherlands, Norway, Australia and research organisations committed to assessing societal excellence of research through impact. In the UK specifically, the prominence of words in the REF2014 (and possibility REF2021) definitions that emphasis “benefits”, and “change”, inherently embed an impression that impact is, and only is positive. In addition, the public too are conditioned to regard valuable research as research that only makes a positive change. We see this in the celebration of how medical research has saved lives, how jobs were created, technologies invented or how policies were improved. These impacts provide a stark comparison to the above aforementioned grimpacts.
What a focus on impact-only fails to grasp (and assess) is that a pathway to impact can include mistakes, misdirected tangents, dishonest users and unforeseen circumstances. In other words, pathways to impact also include grimpact and although impact-assessment does recognise that research’s journey to impact is not linear, it fails to acknowledge that within these non-linear pathways, or serendipitous tangents lies grimpact that may act against society’s best interests. While acknowledging it would be detrimental for the case of public research for the public’s sake, it is essential for a more balanced, and robust impact-evaluation. By studying extreme grimpact examples, we hope to initiate a discussion that specifies it dimensions and characteristics towards a more reflective and robust framework for evaluating societal impact.
What we have found is that grimpact is enabled by a violation and breakdown of normal feedback mechanisms between researchers and users.
In addition, if promoting accountability is the aim of impact evaluation, then arguably the same ideal should apply to grimpact, but attribution is more difficult for grimpact. For grimpact it is less clear to attribute “fault” (the flip side to impact’s “duty”). Is it the incidence of research misconduct or, more complexly, is when research was co-opted by stakeholders for purposes beyond the original intentions of the researcher? Grimpact is also more contagious than impact as the loss of control of the research’s original narrative leads to the message being co-opted by multiple partners who excessively frame or overly-promote politically desirable results, while silencing conflicting evidence. In fact, for grimpact, there is such a complex web of interactions with multiple users beyond what is considered healthy science-society partnerships, that go beyond conventional capabilities of capturing impact, that is makes it almost impossible to assign any blame.
So why does this matter, and why, as a community can’t we just focus on our successes? Because doing so affects how trustworthy we appear to the public to govern ourselves through evaluation, and how valuable our advice and expertise appears to a public that is currently prone to doubt.
For the future of societal impact assessment at least, Grimpact matters.
The research is part of the paper “Derrick GE, Faria R, Benneworth B, Budtz-Petersen G & Sivertsen G. (2018) Towards characterising negative impact: Introducing Grimpact”, presented at the 23rd International Conference on Science and Technology Indicators (STI 2018), ‘Science, Technology and Innovation indicators in transition’, 12-14 September 2018, Leiden, The Netherlands”. http://sti2018.cwts.nl/