Clarification of the memory artefact in the assessment of suggestibility.
Yes/no or photo recognition questions slash false-high suggestibility scores in adults with ID.
01Research in Context
What this study did
One adult with intellectual disability took three versions of the Gudjonsson Suggestibility Scale. The standard test asks, 'Tell me everything you remember.' The two new versions showed photos or asked yes/no questions instead.
The researcher compared how many leading questions the man changed his answers to under each format.
What they found
Recognition-memory formats cut measured suggestibility by more than half. The yes/no version called ASS3 worked best.
Free recall made the man look far more suggestible than he really was.
How this fits with other research
McKenzie et al. (2015) also checked an ID screener against a gold-standard test. Both studies show you must validate your tool before trusting the score.
Lerman et al. (1995) proved the Sexual Knowledge Interview is reliable. Willner (2008) now shows the same need for psychometric care applies to suggestibility tests.
Chiviacowsky et al. (2013) found poor item-level agreement on the MAS and QABF. Together the papers warn: poor items can mislabel an adult with ID.
Why it matters
If you assess witness capacity or confession reliability, switch from open questions to photo or yes/no recognition formats. You will get a truer picture of the client's vulnerability and avoid court challenges to your findings.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Replace free-recall questions with yes/no or photo arrays when you test suggestibility.
02At a glance
03Original abstract
AIM: The Gudjonsson Suggestibility Scale (GSS) assesses suggestibility by asking respondents to recall a short story, followed by exposure to leading questions and pressure to change their responses. Suggestibility, as assessed by the GSS, appears to be elevated in people with intellectual disabilities (ID). This has been shown to reflect to some extent the fact that people with ID have poor recall of the story; however, there are discrepancies in this relationship. The aim of the present study was to investigate whether a closer match between memory and suggestibility would be found using a measure of recognition memory rather than free recall. METHOD: Three modifications to the procedure were presented to users of a learning disabilities day service. In all three experiments, a measure of forced-choice recognition memory was built into the suggestibility test. In experiments 1 and 2, the GSS was presented using either divided presentation (splitting the story into two halves, with memory and suggestibility tests after each half) or multiple presentation (the story was presented three times before presentation of the memory and suggestibility tests). Participants were tested twice, once with the standard version of the test and once with one of the modified versions. In experiment 3, an alternative suggestibility scale (ASS3) was created, based on real events in a learning disabilities day service. The ASS3 was presented to one group of participants who had been present at the events, and a second group who attended a different day service, to whom the events were unfamiliar. RESULTS: As observed previously, suggestibility was not closely related to free recall performance: recall was increased equally by all three manipulations, but they produced, respectively, no effect, a modest effect and a large effect on suggestibility. However, the effects on suggestibility were closely related to performance on the forced-choice recognition memory task: divided presentation of the GSS2 had no effect on either of these measures; multiple presentation of the GSS2 produced a modest increase in recognition memory and a modest decrease in suggestibility; and replacing the GSS with the ASS3 produced a large increase in recognition memory and a large decrease in suggestibility. IMPLICATIONS: The results support earlier findings that the GSS is likely to overestimate how suggestible a person will be in relation to a personally significant event. This reflects poor recognition memory for the material being tested, rather than increased suggestibility per se. People with ID may in fact be relatively non-suggestible for well-remembered events, which would include personally significant events, particularly those witnessed recently.
Journal of intellectual disability research : JIDR, 2008 · doi:10.1111/j.1365-2788.2007.01022.x