Trowbridge Foundation: Current Article in Full

Interrogative Suggestibility in a U.S. Context: Some Preliminary Data on Normal and FAS/FAE Subjects

By R Pollard1, B Trowbridge1, PD Slade2, AP Streissguth3,
A Laktonen3, and BD Townes3

This article supported by The Trowbridge Foundation

Presented at the Psychology and Law International Interdisciplinary Conference, Edinburgh, Scotland, July 7-12, 2003

Published in the Journal: Personality and Individual Differences,
Volume 37/5, pp 1101-1108

Introduction

The United States Supreme Court Miranda Ruling of 1966 established the ‘Right to Remain Silent’ of a suspect during police interrogation. Despite this ruling, very few subjects apparently exercise their ‘right’ – only between 4–16% according to a review by Gudjonsson (Gudjonnson, 1992); on the other hand, between 42–76% make admissions of guilt or a full confession. Moreover, the confession of the suspect is often the major piece of evidence in the case and can be the only piece of evidence. The suspect’s confession – oral, written or both – can be the basis for the prosecution of the individual and the main reason for a guilty verdict.

Various theories have been put forward as to why suspects confess to having committed crimes, particularly when they are subsequently proved innocent, although few empirical data are available on this subject. A notable exception is the questionnaire study of Gudjonnson & Petursson (1991), who found that three separate factors may be at work. These were described as: (1) internal pressure to confess – that is, subjects of police interrogations confess because they feel ‘guilty’ about what they have done and want to ‘get it off their chest’, (2) external pressure to confess – subjects confess due to social pressure applied by the interviewing police officers, and (3) perception of ‘proof’- that is, the subjects confess because they believe the police have overwhelming proof of their guilt and think there is no point in denying the accusations against them. Gudjonnson & Petursson carried out their study on 74 convicted Icelandic prisoners who were serving sentences for violent offences (17%), sexual offences (11%), property offences (38%) and other offences (34%). They found of the three reasons for confessing, perception of the weight of ‘proof’ was the most frequently given overall. Internal pressure to confess due to feelings of guilt was found to be higher in the sexual and violent offender groups, while external pressure from police officers was associated with most dissatisfaction on the part of the prisoners. The authors speculate that “confessions that result from perceived police pressure are viewed as psychologically coerced, which makes offenders subsequently disgruntled about the confession. It would be important in future research to study the behavioral and psychological sequelae of such disgruntlement”. One consequence may be the subjects’ desire to retract their confession at a later date.

Many convicted prisoners (12% in a study by Sigurdsson & Gudjonsson, 1996) subsequently retract the statements they made during police interrogation and claim these were ‘false confessions’. However, it is only when the individual is subsequently cleared by DNA or other evidence that the fact of a previous false confession can be established beyond doubt. Thus, the initial making of a false confession usually has profound consequences for the person concerned.

Gudjonnson (2001) has suggested that there may be four basic types of false confession:

•Voluntary False: elicited without any pressure – the person simply goes to the police after a reported crime and confesses. The voluntary false confession is commonly due to a morbid desire for notoriety, which may temporarily improve low self-esteem; or an inability to distinguish facts from fantasy. Mental illness and personality disorder are the two most important factors affecting voluntary false confessions. (Gudjonsson, 1999).

•Coerced-compliant: commonly elicited during persuasive interrogation where the person perceives there is some immediate gain from confessing falsely (e.g. escaping from a stressful interrogation, or release from custody). The person is fully aware of not having committed the crime (Gudjonsson & MacKeith, 1990).

•Coerced-internalised: elicited by persuasive questioning, but using more subtle and psychologically manipulative techniques. Suspects are gradually persuaded that they have committed a crime of which they have no memory, or they have become so confused that they no longer trust their own memory and accept a false scenario suggested by the police (Gudjonsson & MacKeith, 1982). This ‘memory distrust syndrome’ is a common phenomenon in alcoholism, but can also occur for other psychological reasons (Gudjonsson, 1995, 1999).

•Coerced-reactive: occurs when a person is pressured to confess by somebody other than the police (e.g. a peer; a spouse). McCann (1998) gives the example of a violent man who intimidated his wife into taking responsibility for the murder of one of her children.

The above typology of false confessions is based on psychological vulnerabilities or mental disorder. One type of vulnerability that has been extensively explored by Gudjonsson and colleagues in the United Kingdom is Interrogative Suggestibility.

The Gudjonsson Suggestibility Scales

Interrogative Suggestibility has been defined by Gudjonsson & Clark (1986) 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 behavioral response is affected". Against the background of this general definition, Gudjonsson developed tests of Interrogative Suggestibility which would mimic the types of social pressure applied to suspects during police interrogation (Gudjonsson, 1984: 1987). These tests are known as the Gudjonsson Suggestibilty Scales (GSS 1 and GSS 2). These scales measure two specific aspects of interrogative suggestibility – ‘Yield’, which is the extent to which the individual yields to leading questions; and ‘Shift’, which is the extent to which individuals are prepared to shift their answers following negative feedback/interpersonal pressure. These measures have been shown to be factorially separate and seem to reflect different but allied processes.

Considerable normative data have been collected on the GSS scales in the UK and in Iceland. These have included studies on subjects with intellectual difficulties (mental retardation) who tend to score highly on Interrogative Suggestibilty, particularly on the ‘Yield’ measure (Clare & Gudjonsson, 1993). In another series of studies, Interrogative Suggestibility was found to be common among delinquent boys (Gudjonsson & Singh, 1984) and juvenile offenders (Richardson, Gudjonsson & Kelly, 1995). Both of these groups obtain higher scores on average on ‘Shift’, that is, being particularly responsive to negative feedback/interpersonal pressure. Thus the type of Interrogative Suggestibility exhibited seems to vary for different groups of vulnerable persons.

FAS/FAE Individuals

Although it has long been observed that maternal alcohol consumption during pregnancy affects the developing fetus, it was not until 1973 that Jones & Smith, dysmorphologists, first coined the term “fetal alcohol syndrome” (FAS). This term was used to describe individuals having a specific pattern of physical anomalies, growth deficiency, and central nervous system dysfunction. The term ‘fetal alcohol effects’ (FAE) has subsequently been used to describe individuals who were clearly exposed but do not have enough features to fulfill the criteria for a FAS diagnosis (Streissguth & O’Malley, 2000). Fetal Alcohol Spectrum Defects (FASD) is a relatively new term that describes the whole spectrum of defects attributable to the effects of alcohol consumption during pregnancy (Streissguth & O’Malley, 2000).

FAS/FAE has been extensively studied by Ann Streissguth and her colleagues at the Fetal Alcohol and Drug Unit (FADU), at the University of Washington, over the past twenty five years. A cohort of 416 patients with a diagnosis of FAS or FAE has been closely studied over this period. More recently their adolescent adjustment problems have been a particular focus of study (Streissguth, Barr, Kogan & Bookstein, 1996) the results of which revealed, among other things, a high rate of ‘problems with the law’ among subjects with FAS/FAE. Sixty one percent of adolescents and 58 % of adults studied had such problems and, in the case of males, 38% had been convicted of criminal offences. These included, in descending order of occurrence, shoplifting, theft, burglary, assault, domestic violence and child molestation. Those involved with the well-being of FAS/FAE persons have become increasingly concerned about the problem of ‘false confessions’ in this highly vulnerable group.

Aims of the Present Study

The present study had a number of aims, the emphasis of which changed during the course of the study, as will be explained. The main aims of the study were as follows: (1) To collect normative data on the Gudjonnson Suggestibility Scale (GSS 2) on a control sample of US Citizens. (As far as we are aware, this will be the first reported data on US subjects.), (2) To examine differences on the GSS 2 scales with respect to gender, occupational status, age and estimated IQ, (3) To compare a sample of FAS/FAE individuals with the sample of control subjects, and (4) To explore the possible mechanisms underlying Interrogative Suggestibility and establish provisional cutoff scores.

Recruitment of Subjects

Data were collected on a total of 80 subjects as follows: 60 control subjects were tested by Jim Chantler at Harborview Medical Center (HMC); a further 12 control subjects were tested at FADU, together with 8 FAS/FAE subjects, by Alberta Laktonen and Marie Baldwin. All three testers were trained in the administration of the tests and satisfactory inter-rater reliability established. Control subjects were recruited by fliers at HMC and, in the case of FADU, at University of Washington Medical Center. These fliers invited subjects to take part in a study involving recall of a short story, and in return for their participation offered them a $20 reward. FAS/FAE subjects were recruited by letters sent to individuals enrolled in the UW Fetal Alcohol Follow-Up Study, and followed up by telephone where possible. They were also offered $20 for their participation.

It had been hoped to recruit a far larger sample of FAS/FAE subjects. However, of the 24 subjects initially approached by letter and telephone, only eight agreed to take part. Of a further 20 approached none was willing to participate. As a consequence of this recruitment problem, the primary emphasis of the study was changed from a FAS/FAE comparison study (although a small comparison study was still possible) to one primarily establishing some US normative data on Interrogative Suggestibility.

Testing Procedure

All subjects were administered two tests: the Shipley Institute of Living Scale (Zachary, 1986), which assesses general intellectual functioning in adults and adolescents, and the Gudjonsson Suggestibility Scale Version 2 (Gudjonsson, 1997), which measures Interrogative Suggestibility.

The Shipley Institute of Living Scale involves two subtests; a ten minute multiple-choice following completion of the Shipley Scale vocabulary test, and a ten minute abstraction test. Subjects’ responses were hand scored and resulting scores were translated into estimated IQ’s. This test is used routinely in a variety of settings to obtain a quick, reliable and valid estimate of intelligence (Zachary, 1986).

Following completion of the Shipley Scale, the Gudjonsson Suggestibilty Scale (GSS 2) was administered to all subjects in accordance with the procedure outlined by Gudjonsson (1997). The subject was told, “I want you to listen to a short story. Listen carefully because, when I am finished, I want you to tell me everything you remember.” Then a tape of the GSS 2 was played. After the tape was finished, the subject was told, “Now tell me everything you remember about the story.” Each of 40 specific ideas recalled counts as one point towards the ‘Free Recall’ score (max = 40).

Next the subject was asked to answer 20 questions, 15 that suggested details that were not present in the story and 5 that were (to conceal the true purpose of the scale). The 15 leading questions are designed to measure how much the subject is influenced by suggestive/leading questions. Each suggestive question that is answered affirmatively scores one point towards the ‘Yield’ score (max =15).

After the subject answered all 20 questions, he/she was firmly told “You have made a number of errors. It is therefore necessary to go through the questions once more and this time try to be more accurate”. The questions were then repeated, with each distinct change in the nature of the reply to any of the 20 questions counting as one point towards the ‘Shift’ score (max = 20). The ‘Total Suggestibility’ score was then calculated by simply adding the ‘Yield” and ‘Shift” scores (max = 35).

After the testing was completed, the subject was debriefed in the following manner:

“Thank you for participating in this study. The purpose of the study is to determine if there are certain factors that determine how easily led, or suggestible, a person is. The test you took required that all subjects be told that they answered questions incorrectly. All subjects were then asked to answer the same questions again, regardless of their answers. In other words, your performance was not the reason you had to answer the original questions a second time. In fact, most of the questions you were asked had no right or wrong answers. All subjects had to be told that they had wrong answers in order to see if they would change their answers after receiving this feedback. We regret that it was necessary to mislead you to conduct this research. We appreciate your taking the time to complete this study. Do you have any questions?”

Subjects were then given a sheet of paper with this debriefing statement on it and a telephone contact number if they had any further questions or concerns

Study Findings

Control Sample. The two groups of control subjects, tested at Harborview and FADU, were combined to form one US normative control sample with an N of 72. There were 34 males and 38 females in this sample. The mean age was 26.54 (SD = 5.27, Range = 18 – 35). The sample had a mean of 14.8 years of education (SD = 2.05, Range = 9 – 20 years). The occupational breakdown was as follows: Professional – 4, Administrative – 10; Students – 15; Skilled – 24; Unskilled -12, Unemployed - 6, Homemaker –1. In other words, this sample was composed of young adults, with education ranging from High School to Ph.D., and major occupational categories included.

Comparison with UK General Population Sample. The present US sample of Control Subjects was compared directly with the UK sample reported by Gudjonsson in the GSS Manual (1997). The comparison is presented in Table 1. The two samples are of similar size, with mixed gender composition, although the UK sample has a higher representation of males. The UK sample also has a wider age range leading to a significantly higher mean age. Interesting similarities and differences were found between the two samples on the GSS 2 measures.

Table 1. Comparison of US with UK General Population Samples on GSS 2.

Variable	US Sample	UK Sample	'F' Significance
N	72	83
Gender	34 m, 38 f	53 m, 30 f
Age: Mean	26.54	30.00	61.8	.000
(SD)	(5.3)	(8.8)
Range	18-35	16-69
GSS 2: Means (SD's):
Free Recall	18.9 (5.2)	19.7 (6.1)	1.89	NS
Yield	3.4 (2.5)	4.5 (3.6)	4.74	.05
Shift	4.0 (3.2)	3.0 (3.0)	4.03	.05
Total Suggestibility	7.4 (5.1)	7.5 (5.3)	<1.0	NS
NS = not statistically significant

The two samples scored similarly on the Free Recall measure and on Total Suggestibility; but they differed significantly on the two component parts of interrogative suggestibility. The US subjects scored significantly lower on the ‘Yield’ measure (suggesting less yielding to leading questions), but scored significantly higher on the ‘Shift’ measure (suggesting a greater shift in answers in response to negative feedback/social pressure). We will return to a consideration of these possible cultural differences later.

Gender and Occupational Differences on GSS 2 Measures

Gender and occupational group comparisons on the GSS are presented in Table 2. Although the females scored significantly higher than the males on Free Recall, there were no significant gender differences on any of the three suggestibility measures. This is in line with the absence of gender differences reported in UK samples (Gudjonsson, 1997). Similarly, no significant differences were found between the four occupational groupings on any of the GSS measures. This suggests that Interrogative Suggestibility is unaffected by gender and occupational groupings.

Table 2. Gender and Occupational Differences on GSS 2 Suggestibility Measures.

	N	Free Recall	Yield	Shift	Total Suggestibility
Gender
Males	34	17.0 (5.5)	3.7 (2.4)	4.3 (3.3)	8.0 (5.1)
Females	38	20.5 (4.4)	3.2 (2.6)	3.8 (3.1)	7.0 (5.0)
'F'		8.83	0.78	0.45	0.73
Significance		.004	NS	NS	NS
Occupational Groupings
Prof., Admin.	14	20.1 (4.4)	3.1 (2.5)	4.5 (3.6)	7.6 (5.5)
Students	15	21.1 (4.4	2.9 (4.4)	3.2 (2.6)	6.1 (5.2)
Skilled	24	18.6 (5.6)	3.8 (2.4)	4.9 (3.7)	8.7 (5.3)
Unskilled	19	16.6 (5.3)	3.6 (2.4)	3.2 (2.1)	6.7 (4.2)
'F'		2.54	0.44	1.53	0.93
Significance		NS	NS	NS	NS

Differences between FAS/FAE and Control Subjects

The demographic characteristics of FAS/FAE and control subjects, together with GSS 2 measures, are presented in Table 3. Although the two groups do not differ in age, they differ on all other variables. The FAS/FAE subjects have significantly less education (none of them got beyond the 12th grade) and, as might be expected, their mean estimated IQ of 83.8 (range 58-106) is much lower than that of controls (mean IQ of 108.2 with a range 81-125). On the GSS 2 they had significantly poorer Free Recall scores and higher mean scores on all three measures of Interrogative Suggestibility. Indeed, their mean Total Suggestibility score of 15.5 is higher than that of any of the mean scores quoted by Gudjonsson in the GSS manual. The nearest comparative UK group is that of 68 non-forensic subjects with intellectual disabilities (IQ scores between 57 – 75) who had a mean score of 14.6 (see Table 5.9 in Gudjonsson, 1997). Interestingly, the United States FAS/FAE subjects scored much lower on average on the ‘Yield’ measure than their UK comparisons (7.0 compared to 9.8) but much higher on the ‘Shift’ score (8.5 compared to 4.8). This mirrors the pattern found for the US/UK Control comparison reported on above and in Table 1.

Table 3. Differences Between FAS/FAE and Control.

	FAS/FAE	Controls	F	Significance
N	8	72
Age	24.3 (5.2)	26.5 (5.3)	1.37	NS
Years of Education	12.0 (0.0)	14.8 (2.1)	14.77	.000
Estimated IQ	83.8 (18.6)	108.2 (8.2)	46.65	.000
GSS2: Means (SD's)
Free Recall	10.8 (4.7)	18.9 (5.2)	17.69	.000
Yield	7.0 (2.7)	3.4 (2.5)	14.50	.000
Shift	8.5 (3.6)	4.0 (3.2)	14.22	.000
Total Suggestibility	15.5 (5.2)	7.4 (5.1)	18.27	.000
NS = not statistically significant

Correlations between GSS 2 Measures and Other Variables

The correlations between the three Suggestibility measures and Age, Years of Education, and Estimated IQ are presented in Table 4. There are no significant correlations between Suggestibility and either age or years of education in this US control sample. This is consistent with the absence of either gender or occupational group differences in Suggestibility. There are, however, significant correlations between all three suggestibility measures and both Estimated IQ and Free Recall Scores. The size of these correlations and their interrelationships are in line with the UK studies reported by Gudjonsson and colleagues (Gudjonsson, 1990; Clare & Gudjonsson, 1993; Gudjonsson & Clare, 1995). The pattern of the correlations between IQ and Recall scores, and Suggestibility measures, is very similar to those found in the UK studies.

Table 4. Correlations Between Measures (N = 72 controls).

	GSS Total	GSS Yield	GSS Shift	Education	Estimated IQ
Yield	.86**
Shift	.91**	.59**
Age	.11	.02	.16
Education	-.03	-.19	.10
Estimated IQ	-.32**	-.32**	-.25*	.26*
Recall	-.32**	-.39**	-.20*	.35**	.56**
=<.05 *=<.01

Possible Mechanisms underlying Interrogative Suggestibility

One possible explanation for High Interrogative Suggestibility scores is low intelligence. That is, subjects who are of low intelligence are more susceptible to misleading questions /misinformation and more vulnerable to negative feedback/social pressure to respond in a desired manner. This hypothesis is consistent with the observed Suggestibility Scores of the FAS/FAE subjects in this sample. However, the data presented in Table 4 suggest that an alternative hypothesis; namely that high scores on Suggestibility measures are linked to low scores on Free Recall, which are in turn linked to low intelligence.

It would be hypothesized that it is not low intelligence per se but rather poor recall for events or story details that renders subjects more susceptible to misleading information and pressure to alter their recollections. That is, not being able to remember story details (but wanting to please the examiner) leads subjects to adopt incorrect information presented in the suggestive questions. To investigate this possibility, the distribution of Free Recall scores was inspected for the 72 control subjects and a cutoff score of 10 or less was established. (The 6th percentile was chosen as it represented the closest whole number to the 5th percentile). This cutoff score was then applied to the 8 FAS/FAE subjects. Of the eight, half had a Free Recall score of 10 or less. The high Total Suggestibility Scores recorded by these four FAS/FAE subjects (i.e., Ss 5, 8, 16,19) is hypothesized to be related to poor immediate Free Recall of the story. We will refer to this as Poor Recall Suggestibility.

Predicting Interrogative Suggestibility Scores

Given the significant negative correlation between Recall and Total Suggestibility scores in the control sample shown in Table 4, it is possible to derive a simple regression equation for predicting the latter from the former.

Predicted Total Suggestibility Score = 13.24 – (.31 x Recall Score).
This has a Standard Error of Prediction of 4.82.

We can then calculate an Abnormally High Total Suggestibility Score by the following:
Abnormal Suggestibility = Total Sugg. - Predicted / (1.68 x Standard Error of Prediction).

Using this formula Predicted Suggestibility scores were calculated for the 72 control subjects and the 8 FAS/FAE subjects. Abnormally High Total Suggestibility scores were then identified as those which were more than 8.10 (1.68 x standard error of prediction) above the predicted score. Three of the 72 control subjects (4%) and two of the eight FAS/FAE subjects (25%) were identified as having abnormally high Total Suggestibility scores by this method.

We will refer to this form of abnormality as Hypersuggestibility as it reflects Interrogative Suggestibility which cannot be accounted for by poor recall/ memory.

Table 5 summarizes the relevant individual data on the eight FAS/FAE subjects without providing any individual identifying information. Subjects 2, 3, 6, and 8 were those identified as having Poor Recall Suggestibility; while subjects 1and 5 were those who were identified as showing Hypersuggestibility. Two subjects, 4 and 7, fell into neither of these groupings.

Table 5. Summary Data on FAS/FAE Subjects.

Subject No.	IQ	Recall	Total Suggestibility	Predicted Suggestibility	Difference OBS.-Pred.
1	61	13	24	9.21	14.79 ###
2	94	3 ***	15	12.31	2.69
3	89	10 ***	8	10.14	-2.14
4	106	19	13	7.35	5.65
5	58	13	19	9.21	9.79 ###
6	95	9 ***	16	10.45	5.55
7	99	12	10	9.52	0.48
8	68	07 ***	19	11.07	7.93
*** = Poor Recall Suggestibility ### = Hypersuggestibility

Discussion

The first issue concerns the data on Interrogative Suggestibility, which were collected on a normative US sample of 72 control subjects. As far as we are aware this is the first data set of its kind collected on a sample of US citizens. Interrogative Suggestibility was found to be unrelated to gender, socioeconomic groupings as defined by occupational status, age, or years of education, although it was related to estimated IQ. This pattern of relationships (or lack of them) is very much in keeping with the published data on UK and Icelandic subjects (ref. Gudjonsson, 1997).

Secondly, when this normative US control sample was compared with the UK general population sample provided by Gudjonsson (1997), both similarities and differences were observed (see Table 1). Immediate Free Recall scores and Total Suggestibility scores were similar; but there were significant differences between the two national samples on ‘Yield’ and ‘Shift’. The UK sample scored significantly higher on ‘Yield’(susceptibility to leading questions) while the US sample scored significantly higher on ‘Shift’ (responsiveness/change due to social pressure). A similar pattern emerged when the FAS/FAE sample was compared with a sample of UK mentally retarded individuals – the UK subjects scored higher on ‘Yield’ while the US subjects scored higher on ‘Shift’.

A careful scrutiny of the GSS Manual showed that all UK samples scored higher on ‘Yield’ than on ‘Shift’; while the opposite was true for the US samples tested by us. Indeed, when the present US sample was divided into two groups using the median split on estimated IQ, both the less intelligent and the more intelligent subgroups had higher ‘Shift’ than ‘Yield’ scores. Although further data are clearly necessary, it does appear that there may be cross-national differences in the pattern of Interrogative Suggestibility. One potential avenue of investigation may be to study Yield/Shift differences as a consequence of cross-cultural differences in Acquiescence (Asch, 1951) and Obedience (Milgram, 1974). In the former, the pressure is more implicit and the subjects may be unaware that they are being influenced; whereas in the latter the subjects obey the instructions of the interviewers (interrogators) because of their perceived greater knowledge and expertise (see Gudjonsson, 1992; Blass, 2000). The measure of ‘Yield’ could be argued to be more like that of ‘Acquiescence’ whereas that of ‘Shift’ is more like that of ‘Obedience’ to seemingly legitimate and expert power.

A third point of note concerns the differences between the small sample of FAS/FAE subjects and the control sample. The FAS/FAE group was significantly different from controls on all the GSS measures and in terms of years of education and estimated IQ (Table 3). Potential mechanisms underlying individual differences in Interrogative Suggestibility were therefore examined.

In conclusion, a study of the pattern of correlations between GSS Suggestibility and other measures showed that estimated IQ and immediate Free Recall scores were both significantly and equally correlated with suggestibility. It was therefore hypothesised that it is poor recall of details rather than low IQ per se which can lead to increased interrogative suggestibility. This led us to do two things using the data from the 72 normative control subjects.
First we established a cutoff score for Free Recall at the 6th percentile of the normative control group (i.e. only 6% of the control group scored below the cutoff of 10 or less). Fifty percent of the FAS/FAE were found to have scores below this cutoff. We have referred to this as Poor Recall Suggestibility. And secondly, we calculated a simple regression equation for predicting Total Suggestibility scores from Free Recall scores. A difference of greater than 8.10 between the observed Total Suggestibility score and the Predicted score can be viewed as abnormal. We refer to this as Hypersuggestibility as it reflects a level of Interrogative Suggestibility which cannot be accounted for by poor recall scores. Twenty-five percent of the FAS/FAE subjects were classified as Hypersuggestible by this method.

Thus, we have identified two simple methods for classifying and evaluating the Interrogative Suggestibility performance of subjects using the normative data on 72 control subjects collected in this study – one a classification based on Free Recall score on the GSS (Poor Recall Suggestibilty) and the other based on a significant difference between Observed and Predicted Total Suggestibility score (Hypersuggestibility). Until more US data become available we would suggest that practitioners use these two methods for evaluating the performance of subjects who are referred for assessment of Interrogative Suggestibility.

To return to the Miranda issue with which we began, if Miranda were to be rescinded as is currently being suggested in some quarters, there would be an even greater need for the assessment of Interrogative Suggestibility and the evaluation of possible false confessions.

Acknowledgements: the authors wish to thank the subjects who volunteered their time; Jim Chantler, Alberta Laktonen and Marie Baldwin who tested the subjects; and the Trowbridge Foundation for their encouragement and support. This project was supported by a grant from the Trowbridge Foundation.

1 The Trowbridge Foundation, Olympia, USA

2 Formerly of the University of Liverpool, UK

3 The University of Washington, Seattle, USA

References

Asch, SE (1951). Effects of group pressure upon the modification and distortion of judgments. Chapter in H. Guetzkow (Ed): Groups, Leadership, and Men. Pitsburgh: Carnegie Press

Blass, T (2000). The Milgram paradigm after 35 years: Some things we now know about obedience to authority. Chapter in T. Blass (Ed): Obedience to Authority: Current Perspectives on the Milgram Paradigm. Mahwah, NJ: Lawrence Erlbaum Associates, Inc.

Clare, ICH & Gudjonsson, GH (1993). Interrogative suggestibility, confabulation, and acquience in people with mild learning disabilities (mental handicap): Implications for reliability during police interrogations. British Journal of Clinical Psychology, 32, 295-301.

Gudjonsson, GH (1984). A new scale of interrogative suggestibility. Personality and Individual Differences, 5, 303-313.

Gudjonsson, GH (1987). A parallel form of the Gudjonsson Suggestibility Scales. British Journal of Clinical Psychology, 26, 215-221.

Gudjonsson, GH (1990). The relation of intellectual skills to suggestibility, compliance and acquiescence. Personality and Individual Differences, 11, 227-231.

Gudjonsson, GH (1992). The Psychology of Interrogations, Confessions and Testimony (chapter 4). Chichester: John Wiley & Sons.

Gudjonsson, Gh (1995). “I’ll help you boys as much as I can” – How eagerness to please can result in a false confession. Journal of Forensic Psychiatry, 6, 333-342.

Gudjonsson, GH (1997). The Gudjonsson Suggestibility Scales Manual. UK: Psychology Press.

Gudjonsson, GH (1999). Police interviewing and disputed confession. Chapter in A. Menon & R. Bull (Eds): The Psychology of Interviewing (pp. 327-341). Chichester: Wiley.

Gudjonsson, GH (2001). False Confession. The Psychologist, 14, (No 11), 588-592.

Gudjonsson, GH & Clare, CH (1995). The relation between confabulation and intellectual ability, memory, interrogative suggestibility and acquiescence. Personality and individual Differences, 19, 333-338.

Gudjonsson, GH & Clark, NK (1986). Suggestibility in Police Interrogation: A social psychological model. Social Behaviour, 1, 83-104.

Gudjonsson, GH & MacKeith, JAC (1982). False confessions. Psychological effects of interrogation. A discussion paper. In A. Trankell (Ed): Reconstructing the past: The role of psychologists in criminal trials (pp. 253-269).Deventer, The Netherlands: Kluwer.

Gudjonsson, GH & MacKeith, JAC (1988). Retracted confessions: Legal, psychological and psychiatric aspects. Medicine, Science and the Law, 28, 187-194.

Gudjonsson, GH & Petursson, H (1991). Custodial interrogation: Why do suspects confess and how does it relate to their crime, attitude and personality? Personality and Individual Differences, 12, 295-306.

Gudjonsson, GH & Singh, KK (1984). Interrogative suggestibility and delinquent boys: an empirical validation study. Personality and Individual Differences, 5, 425-430.

Jones, KL & Smith, DW (1973). Recognition of the fetal alcohol syndrome in early infancy. Lancet, 2, 999-1001.

McCann, J (1998). A conceptual framework for identifying various types of confessions. Behavioral sciences and the Law, 16, 441-453.

Milgram, S (!974). Obedience to Authority. London: Tavistock Publications.

Richardson, G, Gudjonsson, GH & Kelly, TP (1995). Interrogative suggestibility in an adolescent forensic population. Journal of Adolescence, 18, 211-216.

Sigurdsson, JF & Gudjonsson, GH (1996). Psychological characteristics of ‘false confessors’. A study among Icelandic prison inmates and juvenile offenders. Personality and Individual Differences, 20, 321-329.

Streissguth, AP, Barr, HM, Kogan, J and Bookstein, FL (1996). Understanding the occurrence of secondary disabilities in clients with fetal alcohol syndrome (FAS) and fetal alcohol effects (FAE). Final Report to the Center for Disease Control and Prevention (CDC), August 1996, University of Washington, Fetal Alcohol and Drug Unit Tech. Rep. 96-06.

Streissguth, AP & O’Malley, K (2000). Neuropsychiatric implications and long-term consequences of fetal alcohol spectrum disorders. Seminars in Clinical Neuropsychiatry, 5, 177-190.

Zachary, RA (1986). Shipley Institue of Living Scale: Revised Manual. Los Angeles: Western Psychological Services.