Why do classic biostatistical studies, alleged to provide causal explanations of effects, often fail? This article argues that in statistics-relevant areas of biology—such as epidemiology, population biology, toxicology, and vector ecology—scientists often misunderstand epistemic constraints on use of the statistical-significance rule (SSR). As a result, biologists often make faulty causal inferences. The paper (1) provides several examples of faulty causal inferences that rely on tests of statistical significance; (2) uncovers the flawed theoretical assumptions, especially those related to randomization, that likely (...) contribute to flawed biostatistics; (3) re-assesses the three classic (SSR-warrant, avoiding-selection-bias, and avoiding-confounders) arguments for using SSR only with randomization; and (4) offers five new reasons for biologists to use SSR only with randomized experiments. (shrink)

During the twentieth century statistical methods have transformed research in the experimental and social sciences. Qualitative evidence has largely been replaced by quantitative results and the tools of statistical inference have helped foster a new ideal of objectivity in scientific knowledge. The paper will investigate this transformation by considering the genesis of analysis of variance and experimental design, statistical methods nowadays taught in every elementary course of statistics for the experimental and social sciences. These methods were developed by the mathematician (...) and geneticist R. A. Fisher during the 1920s, while he was working at Rothamsted Experimental Station, where agricultural research was in turn reshaped by Fisher’s methods. Analysis of variance and experimental design required new practices and instruments in field and laboratory research, and imposed a redistribution of expertise among statisticians, experimental scientists and the farm staff. On the other hand the use of statistical methods in agricultural science called for a systematization of information management and made computing an activity integral to the experimental research done at Rothamsted, permanently integrating the statisticians’ tools and expertise into the station research programme. Fisher’s statistical methods did not remain confined within agricultural research and by the end of the 1950s they had come to stay in psychology, sociology, education, chemistry, medicine, engineering, economics, quality control, just to mention a few of the disciplines which adopted them. (shrink)

Since the beginning of the twentieth century statistics has reshaped the experimental cultures of agricultural research taking part in the subtle dialectic between the epistemic and the material that is proper to experimental systems. This transformation has become especially relevant in field trials and the paper will examine the British agricultural institution, Rothamsted Experimental Station, where statistical methods nowadays popular in the planning and analysis of field experiments were developed in the 1920s. At Rothamsted statistics promoted randomisation over systematic arrangements, (...) factorisation over one-question trials, and emphasised the importance of the experimental error in assessing field trials. These changes in methodology transformed also the material culture of agricultural science, and a new body, the Field Plots Committee, was created to manage the field research of the agricultural institution. Although successful, the vision of field experimentation proposed by the Rothamsted statisticians was not unproblematic. Experimental scientists closely linked to the farming community questioned it in favour of a field research that could be more easily understood by farmers. The clash between the two agendas reveals how the role attributed to statistics in field experimentation defined different pursuits of agricultural research, alternately conceived of as a scientists’ science or as a farmers’ science. (shrink)

The correspondence between Edgar Anderson and Ernst Mayr leading into their 1941 Jesup Lectures on “Systematics and the Origin of Species” addressed population thinking, the nature of species, the relationship of microevolution to macroevolution, and the evolutionary dynamics of plants and animals, all central issues in what came to be known as the Evolutionary Synthesis. On some points, they found ready agreement; for others they forged only a short term consensus. They brought two different working styles to this project reflecting (...) their different appreciations of what was possible at this point in evolutionary studies. For Mayr, it was a focused project with definitive short term conclusions imminent while Anderson viewed it as an episode in an ongoing historical process that, while exciting and suggestive, remained openended. Thus, Mayr and Anderson represent two distinct perspectives on the Evolutionary Synthesis in formation; by understanding both of their points of view, we can grasp more fully the state of evolutionary theory at this key moment. (shrink)

In science and technology studies today, there is a troubling tendency to portray actors in the biosciences as “cultural dopes” and technology as having monolithic qualities with predetermined outcomes. To remedy this analytical impasse, this article introduces the concept styles of valuation to analyze how actors struggle with valuing technology in practice. Empirically, this article examines how actors in a bioscientific laboratory struggle with valuing the properties and qualities of algorithms in a high-throughput setting and identifies the copresence of several (...) different styles. The question that the actors struggle with is what different configurations of algorithms, devices, and humans are “good bioscience,” that is, what do the actors perform as a good distribution of agency between algorithms and humans? A key finding is that algorithms, robots, and humans are valued in multiple ways in the same setting. For the actors, it is not apparent which configuration of agency and devices is more authoritative nor is it obvious which skills and functions should be redistributed to the algorithms. Thus, rather than tying algorithms to one set of values, such as “speed,” “precision,” or “automation,” this article demonstrates the broad utility of attending to the multivalence of algorithms and technology in practice. (shrink)

Since wild badgers were first connected with outbreaks of bovine TB in UK cattle herds in the early 1970s, the question of whether to cull them to control infections in cattle has been the subject of a protracted public and policy controversy. Following the recommendation of Prof. John Krebs that a “scientifically based experimental trial” be carried out to test the effectiveness of badger culling, the Randomised Badger Culling Trial was commissioned by Government in 1998. One of the largest field (...) experiments ever conducted in the UK, the RBCT sought to recreate the conditions of a randomised controlled trial across approximately 3000 km2 of the South West of England. Despite widespread expectations that the RBCT would provide the necessary evidence to resolve the controversy, its findings have instead been widely contested and reinterpreted, while arguments over badger culling have become increasingly polarised. This paper will investigate the complexities of field experimental knowledge by following the story of the RBCT from this initial proposal, through processes of research design, implementation, analysis, interpretation and reinterpretation of the findings by multiple actors. It asks what kind of experiment the RBCT actually was, and examines how it has contributed to the protracted controversy over whether to cull badgers in order to control bTB in cattle. Finally, it will explore the wider implications of this case for contemporary debates over the contribution that RCTs can make to formulating public policy. (shrink)

The most conspicuous form of agricultural experiment is the field trial, and within the history of such trials, the arrival of the randomised control trial is considered revolutionary. Originating with R.A. Fisher within British agricultural science in the 1920s and 30s, the RCT has since become one of the most prodigiously used experimental techniques throughout the natural and social sciences. Philosophers of science have already scrutinised the epistemological uniqueness of RCTs, undermining their status as the ‘gold standard’ in experimental design. (...) The present paper introduces a historical case study from the origins of the RCT, uncovering the initially cool reception given to this method by agricultural scientists at the University of Cambridge and the National Institute of Agricultural Botany. Rather than giving further attention to the RCT, the paper focuses instead on a competitor method – the half-drill strip – which both predated the RCT and remained in wide use for at least a decade beyond the latter’s arrival. In telling this history, John Pickstone’s Ways of Knowing is adopted, as the most flexible, and productive way to write the history of science, particularly when sciences and scientists have to work across a number of different kinds of place. It is shown that those who resisted the RCT did so in order to preserve epistemic and social goals that randomisation would have otherwise run a tractor through. (shrink)

A key role in inference is played by randomization, which has been extensively used in clinical trials designs. Randomization is primarily intended to prevent the source of bias in treatment allocation by producing comparable groups. In the frequentist framework of inference, randomization allows also for the use of probability theory to express the likelihood of chance as a source for the difference of end outcome. In the Bayesian framework, its role is more nuanced. The Bayesian analysis of clinical trials can (...) afford a valid rationale for selective controls, pointing out a more limited role for randomization than it is generally accorded. This paper is aimed to offer a view of randomization from the perspective of both frequentist and Bayesian statistics and discussing the role of randomization also in theoretical decision models. (shrink)

Experimental balance is usually understood as the control for the value of the conditions, other than the one under study, which are liable to affect the result of a test. We will discuss three different approaches to balance. ‘Millean balance’ requires to identify and equalize ex ante the value of these conditions in order to conduct solid causal inferences. ‘Fisherian balance’ measures ex post the influence of uncontrolled conditions through the analysis of variance. In ‘efficiency balance’ the value of the (...) antecedent conditions is decided ex ante according to the efficiency they yield in the estimation of the treatment outcome. Against some old arguments by John Worrall, we will show that in both Fisherian and efficiency balance there are good reasons to randomize the allocation of treatments, in particular when there is no agreement among experimenters as to the antecedent conditions to be controlled for. (shrink)