Julian Boon, Lynsey Gozna and Stephen Hall have a paper forthcoming in the journal Personality and Individual Differences exploring whether it’s possible to ‘fake bad’ on the Gudjonsson Suggestibility Scales (GSS). These tests measure ‘Interrogative Suggestibility’ (IS), which is defined as “the extent to which, within a closed social interaction, people come to accept messages communicated during formal questioning, as a result of which their subsequent behavioural response is affected” (Gudjonsson & Clark, 1986, p. 84). People who are high in IS are more susceptible to making false confessions under interrogative pressure, in a police or military interrogation scenario, for instance. However, as the authors point out, some offenders might be motivated to appear suggestible or vulnerable even if they are not. For instance, if an offender wanted to retract a statement or confession, or “in circumstances where the successful demonstration of vulnerability may lead to a reduction in a fine or sentence or even to escaping a custodial sentence”.
Gudjonsson’s tests for suggestibility are now quite widely known and it’s relatively easy to find information about the procedure on the internet. This presents a problem: the reliability of the GSS depends on test takers being unaware of the purpose of the test. Boon et al. thus explored whether knowledge of the purpose of the test influenced performance, as well as examining the performance of those who deliberately tried to fake bad.
The ‘test aware’ group in this study performed differently on the suggestibility measures compared to the fakers and to a control group who had just been given the standard test. The fakers also produced a different pattern of results. Comparing the fakers’ results on suggestibility measures to the norms for individuals who are mentally impaired revealed that the results were almost identical. However, fakers could be discriminated from genuinely mentally impaired people because they performed better on a test of memory. The authors suggest that this unusual combination of results could be a ‘red flag’ for faking bad:
Specifically, this red-flag could be where individuals’ scoring profiles revealed near identical scores on the principal suggestibility measures to those of the intellectually disabled norms, while simultaneously they were scoring significantly higher on the free recall measures.
The authors report that the interviewer administering the tests didn’t know which conditions participants had been allocated to, but tried to guess. She was only correct 58% of the time, suggesting that participants were good at fooling the interviewer. This shows the value of being able to detect faking via measures in the tests rather than simply relying on the judgement of the interviewer.
The limitations of this study are the usual ones – participants were undergraduate students whose motivation for faking bad is probably rather less than that of real criminals trying to escape a prison sentence.
- Julian Boon, Lynsey Gozna and Stephen Hall (in press). Detecting ‘faking bad’ on the Gudjonsson Suggestibility Scales. Personality and Individual Differences
Abstract below the fold.
Individuals who fake bad on psychological tests can potentially distort the validity of the test scores which are obtained. Furthermore in so doing those individuals may derive instrumental gain in certain forensic and clinical contexts. Little is known of the ability of interviewees to fake bad on the Gudjonsson Suggestibility Scales (GSS’s) and this study sought to investigate the degree to which this could be achieved. Participants were randomly allocated to one of three groups – Standard Procedure, Test Aware and Faking Bad. Performance levels were compared both among groups and with established population norms. The findings support the view that the participants who were attempting to fake bad on the GSS were successful in doing so on the principal suggestibility measures of the test. However they also indicate that there may be potential in coding for additional information which can reveal ‘red-flags’ with which to unmask the interviewees attempting to fake-bad.