Psychology: it’s not dead yet. But Chris Chambers makes a stark case for its having engaged in sins that call its validity into question, risking it ending up as we now view phrenology – as science’s embarrassing cousin. Given the recent spotlight on cases of non-replicability in social psychology (although the problems are not confined to social psychology), would it be so surprising if we were no longer seen as scientists? Chambers argues that cases of non-replicability, bias and data dishonesty result from flaws in the way we tally our outputs and hoard our data and publications. Paywalls make research inaccessible to those who seek to use it – including teachers, policymakers and science journalists – and leave scientists in an echo chamber.
In our search for novel results, we interrogate data like a lawyer when they are non-significant, even as we shelter significant data. We want data to be our best friend and confirm what we want to hear. The search for positive results is pernicious. I recently heard of a PhD student who had spent three years studying ego depletion, which is one of the social psychology fields that failed to replicate. The student now worries about defending successfully, having begun their studies before the scandal. For me, after shopping a manuscript around to nine journals, I capitulated to reviewers who said I should remove all mention of having assessed a non-significant measure. This felt disingenuous, but I was browbeaten. People want a simple story that includes a positive effect. The reader wants to see some action.
Pride and precipice: there is a slippery slope of face-saving strategies when replications fail. My own experience when asking about replications is that scientists and editors argue that it is “rude” to the original authors to call for replication. It is strange that we prioritise authors’ pride over pursuing scientific truths. I’d argue that the way psychological scientists display greed, covetousness and pride is consistent with people with psychopathy, the cohort I have studied for the past few decades. Psychopaths are selfish, deceptive and driven by incentives, awards and financial gain without feeling guilt. Chambers briefly suggests screening people for psychopathy to prevent fraudulent behaviour in our next generation of scientists, but cut-off scores are non-existent and measures rely on self-report. Additionally, there may be good qualities inherent in psychopathic traits, such as cold ambition.
Intent is important for these sins. Chambers admits that he has played the game: used p-hacking and changed his hypothesis after a study, making it appear a priori. But we should strive to be better versions of ourselves, and Chambers offers suggestions for redemption. One involves using Bayesian statistics, preregistration of research and open access publication, perhaps with a new system where journals bid to have the best research in their pages. I like the fact that Chambers uses diagnostic analogies to the probability of someone having Alzheimer’s. All measures in psychology have some wiggle room, yet, states Chambers, many doctors overestimate the probability of truth of a positive test. However, this is not a problem confined to Bayesian statistics. We are not natural statisticians. But we can redeem ourselves before our continuing sins lead us to be unmasked, like the Wizard of Oz, as false prognosticators.
Luna C. M. Centifanti is senior lecturer in the department of clinical psychology, University of Liverpool.
The 7 Deadly Sins of Psychology: A Manifesto for Reforming the Culture of Scientific Practice
By Chris Chambers
Princeton University Press, 288pp, £24.95
ISBN 9780691158907 and 9781400884940 (e-book)
Published 17 May 2017