|Photo: Jerome Ravetz|
|Is science in a mess? |
Should that occur, and ‘science’ ceases to be a key cultural symbol of both truth and probity, along with material beneficence, then the consequences could be far-reaching. To that end, we should consider what lies behind the malpractices whose exposure has triggered the crisis over the last decade.
It is clear that a combination of circumstances can go far to explain what has gone wrong. Systems of controls and rewards that had evolved under earlier conditions have in many ways become counterproductive, producing perverse incentives that become increasingly difficult for scientists to withstand. Our present problems can be explained partly by the transformation from the ‘little science’ of the past to the ‘big science’ or ‘industrialised science’ of the present. But this explanation raises a problem: if the corrupting pressures are the result of the structural conditions of contemporary science, can they be nullified in the absence of a significant change in those conditions?
We should explore how these new conditions lead to these new pressures. There are two familiar qualitative aspects of the steady quantitative growth of the scientific enterprise. The first is the loss of ‘Gemeinschaft’, where all communities and sub-communities have become so large that personal acquaintance no longer dominates in the professional relationships. The old informal systems of rewards and sanctions are no longer effective. Under the new ‘Gesellschaft’ conditions, such intimate tasks of governance must be made ‘objective’. Ironically, applying a ‘scientific’ methodology to the tasks of governance of science leads directly to corruption, since any such system can be gamed. Allied to that development is a second one, the hugely increased capital-intensity of science, so that the typical context of discovery is no longer the scientist with his test-tube, but a large lab with division of labour on an industrial scale. In the absence of the discipline of customers for a product (however corrupted that might be), there is nothing to ensure quality control except those informal systems that are already obsolete.
Just as this new system was becoming dominant, by a cruel accident of fate a third element has intruded: stasis. The social subsystem of science whereby it reproduces itself, namely the training and certification of postgraduates, depends on the possibility of recruitment of at least a significant minority. This will necessarily be small, as even the traditional steady growth rate of science allows only a few new recruits in the course of a scientist’s career.
But when even that prospect vanishes, recruitment stalls, and the existing corps of researchers is squeezed, many pathologies inevitably ensue. The obvious one is the proletarianisation of research work. Recruits (and teachers) face the prospect of a lifetime sequence of short-term jobs on contracts, lacking any rights of security and whose renewal depends on the favour of the principal investigator. Maintaining the lofty ideals of independence and integrity becomes increasingly difficult.
Under these harsh conditions, quality becomes instrumentalised. To strive for ‘excellence’ may be impractical; ‘impact’ is the name of the game. The self-sacrificing quest for scientific rigour is displaced by the need to jockey among journals, and perhaps also engage in p-hacking to obtain interesting results. Such conditions can go far to explain the distressing results that John Ionnidis first found a decade ago. But there is a deeper cause at work. Perhaps those who engage in what we might call ‘shoddy science’ or even ‘sleazy science’ don’t even know that it is sub-standard. The problem may have been building up for decades in the past, when standards gradually slipped and the basic skills of rigorous scientific work were allowed to atrophy. As evidence we have the current state of statistical practice, of which the best is as sophisticated and self-critical as possible, but where there is also much that is an insult.
Source: The Guardian