Science,numbers and politics in a “post-truth” world

1. Science, evidence and “post-truth” politics2

Science and politics have long been in a close yet intricate relationship. Especially since the nineteenth century, scientific evidence has gained in importance for and within policy making, and a “scientization” of political debate and discourse can be witnessed. In this context, numbers have become a paramount tool for making scientific results “workable”. This is manifest both in the increasing production of numbers – Ian Hacking talks of an “avalanche of printed numbers” [2] – and the development of statistical methods. Still, the assumption of a consistently accelerated use of scientific evidence, and numerically-based and -expressed forms of evidence in particular, within the political realm must be qualified: despite an overarching trend towards more science and, indeed, more numbers, it would be misleading to regard this development as linear, or even inevitable.

Time-specific contexts and constraints have always determined the relationship between science and politics, and the actual role of science (and numbers) for and in politics has differed over time and in different parts of the world. Questions as to the very nature of the relationship of science and politics can thus not be easily answered in general terms. What can be argued, however, is that such questions, which have long been the subject of public, professional and academic discussion and debate, have gained in urgency and importance over the last few year; years in which democratic institutions and the model of liberal democracy as such have become fundamentally challenged across the world, and with it scientific evidence, on which especially democratic systems are relying.

In such systems, scientific evidence performs all sorts of work, in part because democracies are answerable to the people, and the people need to be able to examine and assess the workings of government, even if in practice the reliance on scientific evidence may hardly be a guarantee of transparency. But what was previously mainly seen as routine procedures (though never without some degree of contention and controversy) is now facing a new kind of opposition: one that is accompanied by a distinct populist re-ideologization of politics and perhaps best captured by the popular expression of “post-truth politics” [3]. The very notion of scientific and evidence-informed policymaking is coming under pressure, and likewise the role of “expert knowledge” and factual evidence within the political process more generally.

In a “post-truth” world, “facts” – including numbers – have tended to take on a very different status. Not only have older arguments about “objectivity” come under attack, but also more modest claims of facts as matters of general, informed consensus. Today, claims for facticity are often perceived as negative or pessimistic, and “facts” have even become portrayed as a tool that “elites”, “mainstream media” or “the establishment” use to consciously falsify “the truth”. Existing discontent or hostility towards the political status quo within public opinion can easily extend to polemics against “experts” and “technocrats” or scepticism towards all sorts of research- and fact-based evidence, for which the Brexit debate or the US presidential elections of 2016 are just two examples of the recent past. During the run-up to the Brexit referendum, for example, the British Conservative Michael Gove responded to critics who, using analyses and studies, warned of the likely negative economic consequences of the UK’s withdrawal from the European Union that “people in this country have had enough of experts [4].” In the US, Kellyanne Conway, senior aide to US President Donald Trump, responded in 2017 to questions about the White House press secretary’s false statements on attendance numbers at the inauguration ceremony by making the (in)famous claim that they merely reflected “alternative facts [5]”. Yet talking about “post-truth” – the Oxford Dictionaries’ “word of the year” 2016, considered to be one of the “defining words of our time” [6] – requires attention be paid to the ambiguities of the term. “Post-truth” has been primarily coined by those in favour of evidence- and expert-based politics, while the criticism currently put forward against “rationalised politics” by advocates of “post-truth” is closely linked with growing political populism. This is not all-too surprising, since among the key features of populism is the claim to be able to “speak to the people” in a way that is actually comprehensible for everybody – something for which expert cultures, science or statistics do not necessarily offer the most adequate toolset. This helps to explain the appeal for many politicians of neglecting or opposing “facts” in favour of appeals to “values” and “emotions”.

But does that mean that “post-truth politics” is populism? This would be too simplistic an equation, since “post-truth” expresses itself in very different forms and usages, ranging from mere ignorance of facts to their intentional manipulation. Rarely, “post-truth” is about a fundamental rejection of facts and rational argumentation per se. Rather, what is characteristic of “post-truth politics” is the attempt to create an alternative reality by making use of intentionally partial and quasi-neutral arguments that look like facts – which demonstrates that in the end, even a post-truth politician cannot avoid confronting the challenges of fact-based argument and policy entirely.3 This is since facts and evidence are part of the lexicon of politics and continue to “work” as reference and argument points across the political spectrum, even for those who define themselves in opposition to whatever is understood as “mainstream”.

Post-truth politics has triggered counter-reactions and strategies, such as deliberate“fact checking” by various institutions and websites to assess the accuracy of political actors’ statements. The Pulitzer-Prize-winning website Politifact, for example, analyses and illustrates truthfulness by means of its “Truth-O-Meter” – a visual indicator, with doubtful statements and claims classified as either “mostly false”, “false” or “pants on fire” [8]. Still, measuring truth is and continues to be a difficult task, not least since there are no satisfactory answers to the fundamental questions as to who should have the ultimate authority to distinguish truth from falsehood, and whether “fake news” may ever justify the restriction of freedom of expression, no matter how good the intentions may be.

In short: how societies can or should deal with the challenges of a “post-truth world” is not clear. This is especially true in view of the realities of the new media ecosystem, in which messages spread at tremendous speed with little or no scrutiny and “everything is true and nothing is true”, given that no immediate or structural distinctions between the reasoned and the bizarre are made.4 And some truth can even be admitted to critics of current cultures of expertise, which are often perceived as too self-centric, elitist and missing any clear public purpose. Accordingly, experts, too, have to meet their responsibility when it comes to ensuring successful and meaningful communication of knowledge also outside their professional circles or to the public at large.

In summary, it can be argued that “post-truth” is the latest expression of the complicated nature of dealing with facts and evidence in (democratic) political contexts – something that has been a challenge for at least as long as the achievements of science were put to the uses of governments. Stating that the relationship between science and politics has always been fraught with tensions refers to the fundamentally different “logics” of science and politics, respectively.

2. The diverging logics of science and politics

Pursuant to the conventional point of view of their respective practitioners, but also many outsiders, science and politics operate according to clearly different underlying logics. Whereas politicians are primarily concerned with the extension and preservation of power and political legitimacy, for scientists the production and advancement of systemic knowledge and “scientific legitimacy” remains a central. To lean on Max Weber’s notion of the ideal type for heuristic purposes, the differences between science and politics can be formulated in terms of the oppositional pairing “power vs. truth”, which underlines the serious challenges of science’s integration into politics and vice versa.

That science and politics have become considerably intertwined is undeniable. Two opposing, but interrelated, developments can be observed in this regard: on the one hand, a scientization of politics (captured also by the term “evidence-based policymaking”), with science and experts increasingly relied upon when political and social challenges are being addressed; on the other hand, a politicization of science, manifest, for example, in the growing influence of political decision-makers on the objects, methods and processes of scientific research [10]. Both developments are accompanied by clear risks. A full-scale scientization of politics bears the risk of ending up with a hyper-technocratic regime within which all kinds of individual and/or value judgments would become obsolete. On the other hand, the politicization of science can easily diminish the reputation of scientific policy advice and undermine the credibility of science per se. It can also severely damage public trust in the political system, offering support for the view that political decision-makers are merely interested in pushing their agendas by making instrumental use of everything at their disposal, including science. In other words: the view that politicians are less striving for “evidence-based policy” than “policy-based evidence”. Maintaining the legitimacy and integrity of the political and the scientific realms and ensuring the highest possible standards in their relationship is therefore an issue of vital interest for both.

There are diverging ideas about how further erosion of public trust in either science or politics might be avoided. Advocates of a rationalized “evidence-based policymaking”, for example, have actively pushed the idea of further intensifying the use of science as the only adequate basis of modern political decision-making – also and especially in a “post-truth” world. One of their central pleas is “to increase government effectiveness through the use of rigorous evidence about what works” [11]. Critics, on the other hand, maintain that policymaking on the basis of scientific evidence has always been more of a myth than reality, and actively advocate a “re-politicization” of politics, with values serving as an antidote to the fetishization of facts – though not necessarily to the extent represented by radical “post-truth politics”.

In the twenty-first century, it is difficult to envisage policymaking without at least some sort of (scientific) expertise and advice, and so a future in which scientific findings, data and facts would no longer have any relevance is most unlikely. Nevertheless, to assume that simply more (scientific) evidence and more data and facts necessarily means better policymaking is also ideal naive. Instead of capitulating to either extreme, open debate is needed about what scientific evidence can realistically do in and for politics, which in turn requires a more in-depth understanding of how science and politics interact in practice. This includes the need for a critical analysis of the connecting devices within the relationship between science and politics, the perhaps most central of these devices being “numbers”. They serve as a specific means by which (scientific) knowledge is expressed and through which that knowledge can be transferred into the political realm. Hence functioning as a crucial interface between science and politics, numbers merit closer attention.

3. Numbers as a connecting device between science and politics

It is stating the obvious that numbers are an important instrument in both science and politics: in as much as researchers and scientists of many disciplines use quantitative methods and numbers in their work, the intrinsic appeal of and a trend towards numbers is also apparent within politics, where decisions are commonly informed by quantitative forms of evidence. Numerically-based probability analyses and impact assessments are used to determine the possible effects of a particular decision in measurable terms, and efforts to “quantify” are likewise reflected in the desire to formulate number-based policy objectives (e.g., binding quotas, growth rates, benchmarks) whose degree of attainment serves as an indicator for political success. The 2 ∘ C target in international climate change policies or the European Union’s educational policy objective to reduce the share of early school leavers from 15% to less than 10% by 2020 are just two concrete examples in this regard.

Despite their ubiquity, however, the use of numbers for the formulation of policy goals or the assessments of a specific policy’s impact is anything but uncontroversial. Proponents of number-based policymaking emphasize the increased transparency and improved information available for decision-makers, stressing potential gains in objectivity and measurability. Some would go even further and claim that decision-making that is not based on numbers and statistical evidence is virtually impossible: “If you can’t measure it, you can’t manage it and you can’t fix it” [12], stated Michael Bloomberg, a former mayor of New York City, in 2014.5 He thus expressed a widely held view that only regular and vigorous measuring and statistical analysis were adequate means for taking political action in the modern age.

Conversely, critics stress that numbers and statistics cannot be the only – nor should they necessarily be the major – determinant of political decision-making. Apart from the fact that statistics or numerical evidence are as fallible as any other form of (scientific) evidence, policies focused on numbers and “hard facts” risk neglecting qualitative aspects that are not – or at least not fully and directly – measurable. Another frequent criticism is that the assumed neutrality of statistics and numerical evidence can be deeply deceptive: those who produce and use statistics do so in a particular cultural or social context with vested political, organizational or institutional interests. A degree of “subjectivity” is thus an ever-present companion of numbers.

Having said that, acknowledging that numbers can never claim to be completely objective does not imply that there are no means available to enhance the legitimacy of statistical findings and analyses both at an international and national level. The Code of Practice of the European Statistical System (ESS) can be mentioned as an example in this regard, as can the IMF Standards for Data Dissemination or the UN Fundamental Principles of Official Statistics. At the same time, many countries have been strengthening their national statistical legislation to emphasise impartiality and independence. Taken together, these increasingly common quality-assurance measures and instruments allow today for the distinction between what is commonly referred to as “official statistics” and other statistical sources and producers.

Still, even if there might be a consensus over what is and is not “official” – and thus perhaps: more reliable and “legitimate” – statistical evidence, another major point of criticism can hardly be discarded: namely that statisticians (and experts more generally) may wield considerable political power without any direct democratic legitimacy.

Such concerns underpin the potential risk mentioned earlier of both a far-reaching “scientization” of politics, which would ultimately reduce policymaking to a mere extension of scientific and technocratic rationality, and a “politicization” of science. Such “politicization” – which can become outright discrediting, as discussed above – takes place when the independence of science and experts from politics is implicitly or explicitly questioned, thus making scientific findings appear unreliable as a basis for possible consensus or action. Moreover, the ever-present degree of uncertainty and fallibility of scientific estimates can serve to undermine the credibility of the scientific community in the eyes of the general public, especially when new results emerge that contradict previous findings that had been central to a particular political discourse.6 It is therefore the responsibility and also in the interest of experts not to present their findings as absolute truths or to give the impression of certainty when this is not possible. Similarly, experts must make every effort that their research is not (ab)used in a way that does not stand up to scrutiny.

In summary, it can be stated that numbers are a very present, but also controversial instrument for making scientific results “workable” for actual policymaking. What applies for science in general thus also goes for numbers in particular: they are surrounded by contestation when entering and being used in a political context.

With this in mind, what role can scientific evidence then realistically play for political decision-makers? Again, a closer look at the specific case of numbers as transmitters of scientific evidence might help to address this question.

4. Context and dynamism

A fundamental prerequisite for assessing the actual role of numbers in policymaking is their contextualization. That is, numbers always need to be interpreted in their respective and particular contexts, with general assertions being only possible (and useful) to a limited degree.

No less important than contextualization is to understand numbers’ entering the political realm as an active and dynamic process. Three sequential stages can be distinguished to capture the dynamic relationship between numbers and politics: production, transfer (also in the form of translation or migration) and use of numbers.

Based on the assumption that numbers and their origins are neither neutral nor objective, a strong plea can be made to consider the normative aspects of the production of numbers. Numbers are inherently linked to classification systems and concepts such as population, gender or nationality. Furthermore, in the political realm there is no linear sequence of (conscious) production of numbers being followed by their use. Rather, there is an interplay of “supply and demand”: on the one hand, political decision-makers ask for expert knowledge and fall back on producers of numbers (e.g., representatives of scientific institutions, statistical offices or interest groups). On the other hand, scientists produce politically relevant quantifications, even if a political significance was never initially intended. A clear norm of how a number is produced and subsequently becomes relevant in politics is absent. While one does not need to go so far and view the production of numbers as a “black box”, it can be understood as taking place on a multi-dimensional playing field, where even number producers are not always able to oversee all conditions of the entire production process. Similarly, consumers of numbers not necessarily dispose of any in-depth knowledge about why and how the numbers they actively use were generated.

A frequently underestimated aspect of the science-numbers-politics nexus is the intermediary step between the production of numbers (as expression of scientific evidence) and their actual political use, which can take at least three different forms: “transfer”, “translation” and “migration”. “Transfer” can be understood as an intentional process: a number is transmitted via a clearly defined channel from producer to user. Such transfers may be initiated and managed by researchers, but also NGOs, policy advisers, interest groups or even politicians themselves. The term “transfer” implies that an actor has at least some authoritative control, though that control must be put into perspective given that in most cases a multitude of actors are involved in the mediation process. In contrast to “transfer”, “translation” can be understood as a deliberative process: the translator (whoever that might be) renders a number from the language of science into the language of politics, and is ideally “fluent” in both languages, though translation errors might occur. A characteristic feature of the translation of number-based expertise for political action is that it can easily become disputed and may lead to a politicization of the translation process itself. A third possibility is the “migration” of produced numbers into politics. In contrast to the concept of “transfer” and “translation”, “migration” reminds us of the fact that the transition of numbers into politics is not always controlled and can take place even unconsciously, for example through media discourse. Yet whatever form the step between the production of numbers and their actual political use may take, the importance of “numerical literacy” for policy makers on the one hand, that of “political literacy” for scientists on the other, is evident: the better the literacy to understand and communicate with the respective other, the more likely it is that effective interaction can be ensured.

Once numbers arrive in the political realm – be it more as cognitive numbers (referring to a rational process in which a number is factually transferred or translated into politics for measuring a particular phenomenon) or normative numbers (which serve a predominantly symbolic function, for example by being used to designate the need for political action) –, their actual usage may take on many different forms. This also depends on changing political and societal challenges that may have considerable repercussions on the use of numbers in politics. Generally speaking, a distinction can be made between the argumentative use of a number in political discourse on the one hand ( = political use), and the use of a number to formulate and concretize political decisions in legislation on the other ( = policy use), even though in practice considerable conflation between the two is possible.

5. Outlook

When it comes to lessons to be learnt regarding the use of numbers as transmitters of scientific evidence in political environments and by policy makers, at least a few preliminary ideas may be presented here: