Читать книгу Philosophy of Psychology - Lisa Bortolotti - Страница 18

Psychological studies, just like other scientific studies, need to be replicated in order to be credible. It is possible that some interesting results fail to be replicated. In fact, psychology, social psychology in particular, is now facing the so-called ‘replication crisis’ (e.g., Earp & Trafimow 2015): the replicability of social psychological studies, including some famous ones, seems to be remarkably low.

A recent controversy, for example, is about a series of studies of priming effects by John Bargh and colleagues. Priming effects are the unconscious effects that the exposure to a stimulus has on the responses to subsequent stimuli. One famous study (Bargh, Chen, & Burrows 1996, which has over 5,500 citations on Google Scholar as of October 2020) reports the surprising result that the participants who had completed the simple task of making a sentence out of some given words, including words related to stereotypes of the elderly (e.g., ‘grey’, ‘wise’, or ‘wrinkle’), walked away more slowly than other participants who had completed the same task without elderly-related words. However, several researchers (e.g., Doyen et al. 2012) reported their failure to replicate the elderly stereotype study and suggested that the popularity of priming studies by Bargh and colleagues is disproportional to their scientific credibility. Daniel Kahneman, who had once mentioned the elderly stereotype study in his best-selling book Thinking, Fast and Slow (2011), admitted that he had ‘placed too much faith in underpowered studies’ in response to a blog post on the controversy (Schimmack, Heene, & Kesavan 2017).

The problem is compounded by the fact that replication failure tends not to be reported in journals (unless, just like in the case of priming studies by Bargh and colleagues, a serious controversy arises). Paul Bloom says that when the project of replicating a study turns out to be unsuccessful, ‘[u]sually, the project is just abandoned, though sometimes the word gets out in an informal way – in seminars, lab meetings, conferences – that some findings are vaporware (“Oh, nobody can replicate that one”)’ and ‘[m]any psychologists now have an attitude that if a finding seems really implausible, just wait a while and it will go away’ (Bloom 2017, 224).

We do not necessarily endorse radical pessimism about social psychology; after all, many important studies have been replicated. But, as a general rule, replication should always be kept in mind when evaluating psychological studies, and perhaps some extra care is needed when evaluating studies in social psychology with surprising results.

Replication really matters, but what about the observations of rare conditions about which we cannot expect statistical analysis of data or replication? Despite the lack of statistical analysis or replication, the study of some unusual behaviours can be informative and can help us to understand how the mind works. In the book Phantoms in the Brain, V. S. Ramachandran defends the usefulness of observing rare cases as opposed to the statistical study of normal individuals:

[I]n neurology, most of the major discoveries that have withstood the test of time were, in fact, based initially on single studies and demonstrations. More was learned about memory from a few days of studying a patient called H.M. than was gleaned from previous decades of research averaging data on many subjects. The same thing can be said about hemispheric specialization (organization of the brain into a left brain and a right brain, which are specialized for different functions) and the experiments carried out on two patients with so-called split brains (in whom the left and right hemispheres were disconnected by cutting the fibers between them). (Ramachandran & Blakeslee 1998, xiii)

We agree with Ramachandran, especially with his idea that we can gain useful insights by observing some rare cases. This is why we will include the split-brain cases (which Ramachandran mentions in this passage) in our discussion of self-knowledge in Chapter 2, and Capgras syndrome (which is a rare condition that Ramachandran discusses in his book) in our exploration of delusion in Chapter 7. Ramachandran’s idea is also perfectly compatible with the importance of statistical analysis and replication when it comes to the phase in which we test the insights that we initially gain in observing rare cases. In fact, he only says that major discoveries in neurology were ‘initially’ based on single studies and demonstrations; his recommendation is ‘to begin with experiments on single cases and then to confirm the findings through studies of additional patients’ (Ramachandran & Blakeslee 1998, xiii).