Unpacking validation

The term ‘valid’, derived from the Latin validus, has long been used as a general synonym for ‘good’, ‘accurate’ and, more specifically, ‘legally sound’.Footnote ⁸ Early in the twentieth century, the meaning and use of evaluative concepts became more differentiated: in psychology, ‘validity’ became a technical term of art for judging whether a test was informative about the thing that one wanted to test.Footnote ⁹ This meaning of ‘validity’ was tied to modern articulations of not directly observable traits that could only be captured with the help of tests: in 1904, the British psychologist Charles Spearman famously inscribed Francis Galton’s mathematical theory of association between two things into experimental psychology, conceptualizing ‘general intelligence’ (‘g’) as a ‘latent variable’.Footnote ¹⁰ The rapid uptake and harsh contestation of psychological tests for mental abilities were propelled by their abundant use in many different sectors, such as in assessments for job selection, in school entry exams and in recruitment to military programs, as well as in psychological research.Footnote ¹¹ By the mid-century, practical applications of testing and managing ‘human resources’ in industrialized societies coalesced with theoretical research in the new field of psychometrics and advances in statistical methods.Footnote ¹² The sophisticated differentiation of sub-forms of validity concepts were deeply inflected by the US American psychology profession’s concerns about the costs and benefits of their published tests’ commercialization.Footnote ¹³

Institutions removed from the human sciences were struggling with analogous problems of safeguarding their scientific authority while expanding their economic and societal impact. What measurements – and whose measurements – would certify the quality of agricultural products, the efficacy of an immune serum or the safety of synthetic chemicals?Footnote ¹⁴ The term ‘validation’ began cropping up in these other regulatory contexts, as individuals and regulatory bodies sought ways to legitimate the tests they relied upon. Over the twentieth century, validation became a shorthand for a certified process that included assumptions about a test’s informativeness, reliability and applicability, tasking committees and agencies to negotiate compromises between these evaluative dimensions. Our collection takes care to unpack the respective understandings of concepts and methods in different times, places and disciplines, being aware that ‘valid’ and ‘validated’ were and continue to be associated with distinct, though partly overlapping, vernacular, technical and epistemic genealogies and meanings.

Despite the polysemy of ‘validation’, the contributions to Testing Knowledge interrogate, collectively, a prevailing assumption of the twentieth century, namely that validated tests, however understood in detail, constitute actionable knowledge. Is a pharmaceutical safe and effective to administer? Should a chemical be banned because it may cause cancer? Is a veteran’s mental-health diagnosis attributable to their war experience, entitling them to public benefits?Footnote ¹⁵ Throughout the twentieth century, experts and regulatory authorities turned to testing to provide answers to such difficult questions. In turn, whether people accepted the answers reflected, at least in part, their trust in the institutions and actors that vouched for test methods and data. The compiled case studies direct our attention towards a diverse set of practices and regulations that have been mobilized in the past 150 years to guard institutional trust even in the absence of successful test performance, for instance by distinguishing a test’s scientific validity from the ethics of misuse or, more mundanely, the ability to use the right test in the right way for decision making. At the most basic level, the fact that a test has been validated can become a shortcut for arguing for its scientific status, and thereby for legitimizing its role in adjudicating individual, institutional and political decisions.

Standardized tests as forms and objects of regulation

While testing is certainly not specific to modernity, our collection suggests that standardized tests became a typical feature of twentieth-century bureaucratic rationality.Footnote ¹⁶ In part this reflected ambitions to standardize technologies and products to facilitate international trade as well as the ‘trust in numbers’ in public life.Footnote ¹⁷ Along these lines, and more broadly, this volume complements existing work on standardization by historians, philosophers and sociologists of technology and medicine.Footnote ¹⁸ Scholars from various disciplines have pointed to how standardization serves as a form of regulation, whether formal or informal.Footnote ¹⁹ For instance, in their typology of standardization, Stefan Timmermann and Steven Epstein discuss ‘procedural standards’, which ‘delineate the steps that are to be taken when specified conditions are met’.Footnote ²⁰ In many cases, such ‘specified conditions’ arise from tests, which tie scientific procedures into regulation via standards.

Sociologists of knowledge were among the first to point to testing itself as a promising subject. In 1989, Donald Mackenzie argued that the testing of technologies was ‘a major locus of the construction of facts’.Footnote ²¹ A few years later, Trevor Pinch singled out testing as a key site for sociological inquiry, ‘because in testing something is at stake’.Footnote ²² Both of them viewed the analysis of testing as a fruitful way to extend the sociology of science into technology, where, as Mackenzie put it, ‘it will often be found that test procedures rest on the assumption of the validity of certain scientific theories’.Footnote ²³ He argued that test validity remains open to challenge, a point echoed by many of the cases in this collection.

Contemporary with Mackenzie and Pinch, some sociologists and anthropologists of medicine focused on testing, with more of an emphasis on its social implications. They worked against a backdrop of increasing medical testing generally, as well as the impact of new diagnostics and screens based on genetics.Footnote ²⁴ In 1989, Dorothy Nelkin and Laurence Tancredi warned that the conjunction of these new kinds of tests with information technologies meant that employers, government authorities and insurers could effectively use personal data as a means of social control.Footnote ²⁵ They drew examples from disease screening, workplace testing and standardized exams in schools, where more and more students were being diagnosed with learning or behavioural disorders. Rayna Rapp’s Testing Women, Testing the Fetus offered a much more nuanced representation of the new world of genetic testing by looking at how women experienced prenatal screening.Footnote ²⁶ She differentiated responses to amniocentesis among US American women from different ethnic backgrounds, religions and classes, while being sensitive to the tension between reproductive rights and disability rights when it comes to medical abortion.

Since the turn of the century, testing has featured in scholarly analyses of medical systems and markets. Shobita Parthasarathy’s analysis of genetic testing for breast cancer in the US and the UK pointed to how different health care systems led to testing regimes of different kinds. As she argues, a testing system’s structure ‘articulates the rights, responsibilities, and authority of its users’.Footnote ²⁷ Likewise, historians of science and medicine have highlighted the manifold interactions between users, producers and regulators, in particular regarding the research and development of pharmaceuticals, and the advertisement of drugs and diagnostic devices to doctors and patient–consumers.Footnote ²⁸ Especially after the Second World War, the predominant political economy of pharmaceutical companies in the global North has promulgated a testing regime based on screening innovations and gatekeeping intellectual property.Footnote ²⁹

In addition, and more closely aligned with our interests, the realm of medical research and diagnosis lends itself to examining the socio-epistemological dimensions of testing. As Ilana Löwy’s work makes clear, this approach is indebted to Ludwik Fleck’s 1935 analysis of the Wassermann test for syphilis, and his understanding of diagnostic stabilization.Footnote ³⁰ While Fleck emphasized how the Wassermann test allowed for the generation of new questions and a new understanding of syphilis in the first decades of the twentieth century, Löwy adds two further layers to this story. For one, she shows how the test was incisive in shaping the concepts of a diagnostic test’s specificity and sensitivity, despite its wrong assumptions about how exactly the serological test functioned. Tests not only developed into key scientific tools but also became objects of analysis. Many contributions in our volume can be read in this regard: the abundance of testing paved the way for elaborating a terminology and methodological toolkit of evaluative concepts and practices to test the test and guide the user. For another, Löwy reminds us in her more recent work on the diagnosis and management of the infectious diseases of syphilis and Zika in Brazil that if a test is not incorporated into an infrastructure that makes tests and treatments available for those who really need them, it will not succeed. Löwy shows that the failure to get a clear epidemiological picture of the 2015–17 Zika epidemic was due not only to the methodological challenges of testing (particularly those related to the heterogeneity of disease manifestation and the cross-reactivity of antibodies to dengue) but also to the lack of economic and political interests in solving the problem once it lost its allure as an international public-health threat.Footnote ³¹ Testing in the context of scarcity and injustice reminds us that we should not reduce the socio-epistemic realm to researchers and professionals alone.

Against this background, we might also revisit Peter Keating and Alberto Cambrosio’s concept of a biomedical platform to describe how instrumentation and skilled practice are combined in the realm of biology-based medical diagnosis. They argue that biomedical platforms, such as the one they examine for immunophenotyping, require extensive regulation, much of it by professionals.Footnote ³² Immunologists must agree on common classification systems to standardize results and physicians must agree on how to interpret the readout of immune cell types in terms of a disease diagnosis. Processes of validation are woven into these systems of testing and classification at each stage. Our collection makes use of this analysis but puts more emphasis on the gaps and failures of coordination, and the alliances between scientists and political and economic stakeholders. Some contributions detail the personal and sometimes existential impact that tests can have on the tested individuals. In this regard, we want to also point readers to a book by Katrin Solhdju that shares the same title as our collection.Footnote ³³ Solhdju stresses the life-changing consequences of genetic testing in presymptomatic carriers of Huntington’s disease and outlines alternative ways of meaning-making that are in conversation with, but also contest, the scientific authority of the test.

Lastly, Testing Knowledge can be read as a historical commentary on current complaints that much scientific knowledge itself does not stand the test of scrutiny, particularly in the health and human sciences. The controversial methodologist John Ioannidis has gone so far as to claim that ‘most published research findings are false’.Footnote ³⁴ We would situate these methodological concerns with the reproducibility of research, coming from psychologists and statisticians concerned with p-hacking on the one hand and pharmaceutical companies concerned with the irreproducibility of preclinical findings on the other, in a historical trajectory. In the US, earlier demands for accountability in research surfaced in the 1970s and 1980s at the same time as regulatory agencies developed procedures for certifying reliable data for their decision making.Footnote ³⁵ Mark Rushefsky, Sheila Jasanoff and Sarah Vogel have referred to the valorization of such procedures in terms of ‘regulatory science’, which differs from basic research in its emphasis on standardized protocols and materials, strict record-keeping and laboratory inspections.Footnote ³⁶ Our collection offers a genealogy of how these audit mechanisms for science emerged over time as part of an apparatus to establish validated knowledge in regulatory contexts, and puts the recent prioritization of traceability as the measure of good science into critical perspective.

Three phases of validation

The history of validating measurements, models and tests is closely connected to longer trajectories of the roles of quantification for the making of the modern state. However, the case studies in this special issue show that the problem of validity also reflects more specific developments within the twentieth-century history of the science–market–policy interfaces. Taking this perspective, we differentiate three interlocking historical phases that have given shape to validation. These phases should be seen as overlapping layers, not as clearly demarcated epistemes; the testing related to standardization and market coordination of the late nineteenth century continued, even as methodological and conceptual innovations regarding validating tests on humans became signifiers of professional differentiation in postwar societies, to which was added the institutional entrenchment of testing regimes in late modern regulatory science.

First, the increasing use of more or less standardized testing for scientific, medical, military and industrial purposes in the late nineteenth century introduced a set of methodological and organizational sociological problems: who should evaluate the tests’ performance, and how should this be done? We argue that the establishment and social acceptance of committees and licensing bodies – many of which were forerunners of national or international regulatory agencies – were tightly connected to this new need for evaluating testing methods. Industrialization provided the context for the spread of new kinds of standards, which went hand in hand with technologies for measurement and testing, many from engineering, but also drawing on methods from analytical chemistry and bacteriology.Footnote ³⁷ Although standards for unit measurement were associated with national governments, many industrial standards (like that for the ohm) were set by private bodies, either scientific or industrial.Footnote ³⁸ In the US, a non-profit Underwriters Laboratory (UL) certified the electrical safety of products and materials, which received a UL listing and label.Footnote ³⁹ This was emblematic of industrialized nations’ reliance from 1870 to the 1920s on industry associations to set building-code, fire protection and other safety standards. Testing was a key part of this system of voluntary standard-setting, as evidenced by the establishment of an American Society for Testing and Materials in 1898.Footnote ⁴⁰ Through the twentieth century, the expansion of government regulation, particularly over the safety of new technological systems, did not supplant the role of private bodies in regulatory activities. Legal scholar Timothy Lytton discusses examples from fire safety and kosher food to argue that private certification plays an important regulatory role and that it ‘compare[s] favorably with government regulation’.Footnote ⁴¹

In our collection, the contributions by Colleen Lanier-Christensen and Axel Hüntelmann study the establishment of national and international committees at the intersection of academic science and industrial stakeholders in this period. Lanier-Christensen looks at the success of the Association of Official Agricultural Chemists (AOAC) in the US in validating tests for agricultural products since the 1880s, ranging from the amount of nitrogen and phosphorus in fertilizer to the content of butterfat in milk. This de facto regulatory body took up similar problems in toxicological testing in the 1970s, without success. Hüntelmann’s article underlines the challenges of method validation when working on biological materials, which are notoriously variable. As he shows, the serum evaluation method of the Royal Prussian Institute for Serum Research and Serum Testing enabled the international certification of biologicals, but only by standardizing and regulating a long chain of procedures, tests and subjects, including bacteria, guinea pigs, horses and technicians. Hüntelmann highlights that validation of methods involved moral authority and trust, which in his case came to be grounded in an institutionalized internationalism rather than in national bodies.

In Jingwen Li’s article, the variable test subjects are humans themselves. She focuses on a contested glasses prescription in 1926 in Shanghai where the authority of a new machine for measuring nearsightedness could not override the ultimate arbiter of effective vision correction: consumer experience. Eram Alam also deals with the complexities of tests involving human subjects, moving beyond agriculture and health to the labour market. She traces the development of US American certification policies for assessing the equivalence of immigrants’ medical education throughout the twentieth century. Here the standardization had to do with human capital, and the validity of licensing tests encountered the racial politics of post-1965 immigration. These four authors show how political, economic and scientific interests – and the tensions among them – could complicate efforts to set standards and quality assurance mechanisms. When these efforts succeeded, testing was valued for facilitating market coordination.Footnote ⁴²

Testing in the health and human sciences gained a new theoretical momentum in the postwar period. The second phase of validation is marked by an increasing theorization of evaluative concepts, and their strategic employment in a time of increasing professional differentiation and scaling up of research infrastructures in the mid-century.Footnote ⁴³ Human bodies and capacities had not been excluded from the materials testing of the nineteenth century (as emblematized by Francis Galton’s attempts to measure and study human variation), but the emergence of psychology as an experimental discipline focused greater attention on testing people.Footnote ⁴⁴ In a pragmatist move, unsettled questions about measurement of qualitative states and mental abilities had been transferred to a subsidiary question in the first decades of the twentieth century: were the tests fit for their purpose in assisting the selection of humans in military and educational contexts?Footnote ⁴⁵ This implied that the epistemological question of the informativeness of a test was coupled to political ideology and societal structures that constrained the identification of purposes that testing should serve.

Alex Kindel’s paper on psychological test standards examines how academic psychologists sought to protect their field’s reputation against the misuse and misinterpretation of tests. The commercial success of psychological tests, which were published and in widespread use by the mid-twentieth century, prompted the American Psychological Association to form a committee to set recommendations on test validity. Contrary to the examples discussed in the first phase, this committee did not mean to authorize standards. Rather, it used the spectre of poor practices outside the academy to make the question of a test’s validity methodologically demanding, so much so that it even annoyed a doyen of the field. This strategy ultimately helped to safeguard the exclusive role of professional psychological and psychometrics expertise in identifying malpractice in the lucrative world of educational testing. The work of this committee resulted in a paper authored by Lee Cronbach and Paul Meehl that became a heavily cited milestone not only in the human sciences but more broadly among practitioners theorizing validity. They offered a refined technical vocabulary including the new concept of construct validity that resonated with deep theoretical and epistemological problems that had already been raised by nineteenth-century attempts to make abstract human qualities measurable.Footnote ⁴⁶

In postwar science, the high stakes and many possibilities of misuse and error provided a justification for cultivating and differentiating professional testing expertise, which in turn led to the proliferation of profession-specific concepts. For example, in psychiatry, ‘diagnostic validity’ referred at the same time to theories about what a diagnosis should capture and to theories about how to measure and evaluate the degree to which a given psychiatric classification fulfils these ideals.Footnote ⁴⁷ In the case of schizophrenia diagnosis, this was actually intended to test the psychiatrists (rather than the patients) and regulate how they diagnose – turning an old controversy about psychiatric schools into a problem of incoherent epidemiological data.Footnote ⁴⁸ The advent of computers, digitalization and big data came with its own set of assumptions regarding data quality, which had repercussions for evaluating psychiatric diagnosis.Footnote ⁴⁹ Yet, in the courtroom, the usefulness of a diagnostic category could be assessed quite differently, as Alfred Freeborn’s contribution to our collection shows. It looks at a British law that gave veterans the benefit of the doubt in their war-related disability pension appeals and put pressure on courts to develop procedures for granting claims despite unsettled questions regarding causality and the nature of psychiatric diagnoses. In several cases in the late 1960s, the new medico-legal diagnostic category of stress-induced schizophrenia was put to use, allowing psychiatric experts to flag uncertainty (rather than reliability in the case of psychiatric epidemiology), while allowing judges to base their decisions on psychiatric expertise.

Several contributions in this volume probe the co-development of test methods, the theorization of evaluative concepts, and the contestation of what this implied about the measured constructs. Michael McGovern’s contribution looks at how scientific testing was explicitly called upon by civil rights activists to identify unjustness and racial discrimination in US employment. He highlights how the emerging practice and technical vocabulary of statistical validation were mobilized in important ways. Eventually, the activists’ grounds of evidence were rejected on the basis that one could not infer motive from statistical correlation: racial discrimination, the judges argued, was a category restricted to cases of intent. In disrupting the connection between test results that demonstrated unequal treatment and the normative definition of the concept of discrimination, the government could disregard scientific test data as being statistically significant but legally uninformative.

In contrast, Christopher Phillips’s and Sarah Marks’s contributions highlight how statistical validation became deeply embedded in conceptualizing, testing and interpreting a treatment’s effectiveness, the assessment of costs and benefits, and evidence-based allocation of health care expenditures in the second half of the twentieth century. At the same time, both authors stress that statistics served a cause rather than dictating what to do. Sarah Marks shows, for instance, that the postwar British institutionalization of clinical trials that were randomized was not a mere methodological issue, but a pragmatic solution that was compatible with socialist ideals of fair treatment of all patients. Likewise, in the 1990s the ‘what-works’ policy of the New Labour years was perfectly suited to reform the management of the National Health Service (NHS) in a similar spirit, and established the National Institute for Health and Care Excellence (NICE) as the British iteration of evidence-based medicine. Phillips shows how statisticians at the US National Institutes of Health (NIH) developed methods to forge regulatory consensus about extrapolating toxicology experiments with laboratory animals to low-dose exposures of humans. Marks’s and Phillips’s case studies form a nexus to the third phase of validation as they are situated in two newly founded institutions, the NHS in the UK and the NIH in the US.

Postwar politics contributed to the successive transformation of validation into a means for regulation. From the 1960s onwards regulatory sciences and agencies gained importance as gatekeepers of test validity and application. This is emblematized in the emergence of environmental regulation in the US. Laws such as the Clean Air Act and Clean Water Act mandated environmental sampling and monitoring for pollutants.Footnote ⁵⁰ In 1962, new safety requirements for pharmaceuticals went into effect in the US, followed by more stringent requirements for pesticides and oversight of all commercial chemicals.Footnote ⁵¹ These regulatory regimes rested on extensive toxicological testing, including new kinds of in vitro laboratory tests that sought to obtain information without more costly animal experiments. Validation – showing that certain tests are reliable and predict select end points reasonably well – was (and remains) an essential component of this kind of regulatory science. The heavy emphasis on safety testing transmuted public concerns about health and pollution from new chemicals into manageable technopolitical procedures.Footnote ⁵² Fleshing out this historical trajectory concerning the control of chemicals, Colleen Lanier-Christensen, Valentin Thomas and Angela Creager each examine how different tests were validated, and expose the challenges with stabilizing a test for regulatory use. If the first phase of validation was about standard-setting for marketable products, and the second phase about professional differentiation and negotiating expertise, the third phase is characterized by a preoccupation with regulatory procedure and managing risk.

The final paper extends the third phase by showing how the boundary between regulatory science and academic research has corroded over time. Nicole Nelson and Lara Keuck argue that a ‘regulatory ethos’ has been extended into mainstream science, as illustrated by the introduction of ‘preregistration’ schemes into academic research. This regulatory device, in which scientists must declare at the outset how they will collect and analyse their data, was instituted in clinical trials to ensure the integrity of drug efficacy reports. As with other audit mechanisms that are increasingly used in the public sector, there are questions as to who benefits from this kind of accountability, and what kinds of costs they entail for basic research.

The structure of the issue

We have introduced the papers in this issue along a three-phase but overlapping chronology. Here we want to highlight some themes that cut across these three time periods. The articles in this issue are organized to highlight these three areas of dispute around validity.

Papers in the first section, ‘Validating “good enough”’, feature cases in which testing was used for establishing equivalence among products, professionals or psychological assessments. In three of these historical cases, the demand for validated tests came directly from the economic sector, to enable agricultural products to be graded (Lanier-Christensen), biologicals to be evaluated for potency (Hüntelmann), and physicians to be certified as licensable to practise in another country (Alam). In Kindel’s paper on psychological test standards, psychologists sought to protect their reputation against those administering tests in commercial settings. As the work of Joanne Yates and Craig Murphy illustrates, the role of standards in facilitating economic integration and coordinating markets is well studied, but there is less attention to how stakeholders and scientists come to agreement on what constitutes a good measure.Footnote ⁵³

Our second set of papers, ‘The politics of statistical validation’, takes up a different challenge in testing, namely how government agencies and courts could use test data as evidence for efficient or fair decisions.Footnote ⁵⁴ The authors of the papers in this section study the tensions and co-dependencies of the scientific or medical ways of assessing evidence with political ideology. Here we include the emergence of evidence-based medicine (Marks) and the process for ‘invaliding’ veterans seeking war disability pensions (Freeborn) in Britain. Picking up a key case in US civil rights law, McGovern probes how debates over the validity of standardized tests were deployed in lawsuits about racial discrimination in employment. Phillips focuses on the American regulatory system, tracing the impact of the framework that government statisticians developed for extending dose–response curves for carcinogens to low-dose exposures, an important extrapolation for Food and Drug Administration decisions about the safety of food additives and drugs. While the statisticians were aware of the limitations of their methods, both bureaucrats and industry representatives embraced these tools as part of the expanding world of risk assessment for governing chemical safety. By the 1980s, risk assessment became the dominant framework for using toxicology tests in regulatory decision making. As Phillips shows, one element of the politics of risk assessment was the black-boxing of statistical methods to try to prevent scientific uncertainty from destabilizing regulatory legitimacy.

The third set of papers, ‘Contesting validity’, focuses on the critics of validation who call into question tests and evidence-based standards. Li takes us to a Shanghai courtroom, where a medical malpractice case over a prescription for glasses challenged the reliability of a new refraction-determining machine (the retinoscope) and the trustworthiness of the optician against the subjective experience of the patient. Her article points to the inescapability of subjectivity in diagnostic tests on humans, particularly tests of sensory perception. Angela Creager and Valentin Thomas take us into the heart of the world of regulatory science in the 1970s and 1980s, as national agencies and international bodies struggled to establish toxicology standards to be used in regulating products of the chemical industry. The title of Thomas’s essay, ‘In the absence of adequate data on humans’, explains the demand for laboratory tests that could supply regulators with reliable data on chemical safety. The challenges with validating those tests for regulatory use, such as at the US Environmental Protection Agency (EPA), opened up technical issues that could be exploited by stakeholders, particularly in the private sector. As both Thomas and Creager observe, raising questions about the scientific legitimacy of the rodent carcinogenicity test could lead government agencies to make policy changes favouring toxicology tests that industry preferred. We have already mentioned the rodent test that is at the centre of Han’s essay. The ‘forced swim test’ was developed for use in identifying promising candidates for treating human depression. Its validity was contested not only by sceptical psychologists but also by animal welfare groups. In each of these articles, historical actors contested the validity of a test in the service of a particular (sometimes explicitly political) agenda.

Concluding reflections

Overall, this collection shows that the validation of tests for regulation has been pertinent in a wide range of health and human sciences, from serology and psychiatry to civil rights law and environmental protection, from the late nineteenth century to the present. Many of our papers refer to classic works of Theodore Porter, Sheila Jasanoff and Lorraine Daston and Peter Galison to discuss test validation in relation to their histories of quantification and coordinating trust, of technocracy and regulatory science, of epistemic fears and forms of objectivity, including regulatory objectivity.Footnote ⁵⁵ While these frameworks remain useful points of departure, our authors also complicate and nuance them. Our collection shows that test validation is not just a passive outcome of broader characteristics of modernity such as standardization and institutionalization, but that relevant ideals, assumptions, practices and regulations – as well as a new class of technical specialists – have shaped the role of testing in late modern science, policy, legal debates and everyday life.

The interweaving of normative, methodological and socio-epistemological issues in notions of validity epitomizes the modern effort to both express and corral concerns about uncertainty, legitimacy and relevance within the technical realm. For instance, ethics concerning human subjects or protection were negotiated in and through setting up committees that enforce technical guidelines (which may be remote from considerations of actual well-being).Footnote ⁵⁶ Committees are granted the power to limit or enforce testing. As our contributors make clear, such implicit normativity of validation plays a regulatory role on several levels: first, it is referred to when assessing the soundness of specific test procedures: is a test good enough? Second, the implicit normativity of validation is called upon in cost–benefit assessments of testing: is the testing worth its costs and side effects? Third, the normativity becomes foregrounded in meta-level discussions on distinguishing good from bad; that is to say, when validation is deployed in disputes over standards dividing ‘good’ research from ‘misconduct’.Footnote ⁵⁷ While situated mainly in twentieth-century North America and Western Europe, the contributions offer a critical perspective on the hegemony of Western actors and imperial thought in setting the methods that certify knowledge for international policymaking – or market.Footnote ⁵⁸

When we started to work on this collection, the COVID pandemic was in full swing, and the connection between trust in institutions and trust in science, diagnostic tests and medical interventions received much public attention. Our articles do not take up the development of COVID testing or distrust in vaccination, though the vaccine trials and validation of diagnostic tests we all witnessed were deeply indebted to the processes and institutions discussed in this issue. Perhaps more to the point, our contributions, being more historical, do not address the ways in which scientific and regulatory institutions backed by state authority have been put under new pressures in the 2020s.Footnote ⁵⁹ Nonetheless, we cannot help but comment on the timing of our endeavour. During the five years that we spent on this project, from its inception in 2021 to publication in 2026, the present has changed. We are witnessing the mainstreaming of pejorative stances towards institutions and distrust in scientific authority in a way that made us ask at times whether we had reached a tipping point in our positionality to the past: is our collection of papers still a genealogy of the contemporary?

In some cases, surely it is – for instance with respect to understanding the reproducibility crisis as an outcome of the century of testing. Even the Trump administration’s EPA appeals to validation to authorize their intention to eliminate animal testing in toxicology. A recent EPA press release explained how clarity ‘drives innovation’ with the following assertion: ‘A robust toolbox of validated methods gives industry regulatory certainty, rewards safer product design, and strengthens American scientific leadership’.Footnote ⁶⁰ (Unmentioned is that the new methods save industry a lot of money.) On a similar track, Trump’s Health and Human Services trumpets its reliance on ‘gold standard science’ in its funding and regulatory policies, even as many scientists argue that Robert Kennedy’s decisions on vaccination and other issues are at odds with medical consensus.Footnote ⁶¹ But to the degree that technical certification no longer persuades a wary public, we might have already flipped from a genealogy into a historical reconstruction that gives us access to a past world view differing radically from our social reality today. Just as we begin to see the granularity of our black-and-white cover image of the rat in the glass vessel as being antiquated, a world in which a combination of bureaucracy and science was imagined to achieve social equality and fairness might need more explanation today.