Full text of DOI:10.1348/135532500168029

Legal and Criminological Psychology (2000), 5, 123–133 Printed in Great Britain
Ó 2000 The British Psychological Society
123
Interrogative suggestibility: The role of
interviewer behaviour
Stella A. Bain
James S. Baxter*
and
University of Strathclyd e, UK
Purpose. Interrogative suggestibility may vary as a function of interviewer
behaviour. The present study assessed the e ect of two interviewer styles on
V
measures of interrogative suggestibility obtained using the Ž rst of the Gudjonsson
Suggestibility Scales (GSS1). It was hypothesized that a generally abrupt demean-
our adopted by the interviewer would produce greater psychological distance, and
therefore higher GSS1 scores, than a friendly demeanour.
Methods. The study had a single factor between participants design. Participants
were tested on the GSS1 by an interviewer whose behaviour was either ‘friendly’
or ‘abrupt’. One female experimenter conducted all of the interviews. Fifty-Ž ve
participants took part in the study. Most participants were Ž rst year undergraduate
psychology students. Others were university administrative sta .
V
Results. Two of the GSS1 measures appeared to be biased signiŽ cantly by
interviewer style. Participants tested in the ‘abrupt’ condition gained higher scores
for Shift and Total Suggestibility than those in the ‘friendly’ condition.
Conclusions. These results are consistent with the view that the GSS1 provides
measures of two di erent types of suggestibility. However, this Ž nding may
V
also mean that whilst initial responses to leading questions are mediated by more
stable cognitive factors that are relatively una ected by interviewer demeanour,
V
post-feedback scores may be more sensitive to the social aspects of suggest-
ibility. Implications of the results for the objectivity and administration of the
Gudjonsson Suggestibility Scales are discussed.
Much research on the reliability of eyewitnesses has focused on in uences present
during questioning which may distort testimony (cf. Loftus, 1979). Gudjonsson
has termed susceptibility to such in uences ‘interrogative suggestibility’ (e.g.
Gudjonsson, 1984a). Gudjonsson and Clark (1986) presented a model which
attempted to identify these in uences and the interactions between them. Accord-
ing to Gudjonsson and Clark, interrogative suggestibility depends on the coping
strategies witnesses use during an interrogation. All witnesses, victims and suspects
enter an interrogation with a general cognitive set regarding the situation.
This cognitive set is in uenced by uncertainty about the subject-matter of the
*Requests for reprints should be addressed to Dr James S. Baxter, Department of Psychology, University of
Strathclyde, 40 George St, Glasgow G1 1QE, UK.

124
Stella A. Bain and James S. Baxter
interrogation, the degree of interpersonal trust witnesses feel towards the interro-
gator, and their expectations regarding what is about to happen. This general
cognitive set can facilitate either a resistant or suggestible behavioural response to
the interrogation.
Also important is any form of negative feedback communicated to, or perceived
by, the witness. Usually this will be some form of disapproval or criticism
of the witness and may be overt or implicit. Gudjonsson (1984a, 1984b) argued
that negative feedback can have two distinct e ects. First, it can make
V
interviewees change or shift previous responses. Secondly, it can increase an
individual’s susceptibility to further leading questions. Negative feedback can
either be accepted or rejected. If the interviewee rejects the negative feedback
then there will be no major e ect on further susceptibility to suggestions.
V
Gudjonsson and Clark (1986) argued that, occasionally, negative feedback can
make resistant responders even more resistant to subsequent suggestions because it
makes them even more suspicious of the interrogator and the situation than
they were before. Accepting negative feedback is thought to increase uncertainty,
which increases susceptibility to suggestions. Accepting negative feedback is also
likely to diminish an individual’s self-esteem and increase anxiety, if only
temporarily, making him or her more likely to attend to external cues rather
than relying on his or her own internal frame of reference (Gudjonsson,
1992), making him or her more suggestible. The model does not assume that
accepting negative feedback necessarily leads to an increased suggestible cognitive
set, though it commonly does. For some, negative feedback may be construed
as a challenge to improve, making them more critical of the situation and so
less suggestible.
Responses to the GSS procedure may be further a ected by compliance.
V
Compliance occurs, for example, when interviewees give in to what they perceive
is required of them in an attempt to appease interviewers and avoid confrontation,
so that they yield to suggestions and change their responses during the procedure,
even if they know privately that their answers are wrong. A tendency to compliance
may also forestall memory search and retrieval processes. Some compliant
interviewees may not be aware of any contradiction between what they say and the
truth because they attend to situational demands rather than searching memory (cf.
Baxter & Boon, in press). It is at this point that compliance and suggestibility
overlap (see Gudjonsson, 1992, 1997 for further analysis of the relationship
between interrogative suggestibility and compliance).
Gudjonsson developed two scales designed to measure an individual’s level of
interrogative suggestibility, known as the Gudjonsson Suggestibility Scales 1 and 2
(GSS1 and GSS2; Gudjonsson 1984a, 1987). These scales were primarily designed
to be used as clinical or forensic tools to help assess the reliability of confessions
that have been retracted, or to identify particularly vulnerable individuals who may
require extra care during interviewing (Gudjonsson, 1992). The scales aim to
measure responses to the two principal types of suggestive in uence thought to
underlie interrogative suggestibility: leading questions and interrogative pressure
(Gudjonsson, 1983). Both types of suggestive in uence may compromise the
accuracy and reliability of testimony.

Interrogative suggestibility
125
Each scale consists of a spoken narrative and 20 questions about the narrative.
Fifteen of these questions are leading in that they suggest certain details that were
not a part of the original narrative. Five are ‘true’ questions which do not contain
misleading suggestions. The number of suggestions accepted by the interviewee
provides an initial score termed ‘Yield 1’. Negative feedback is then administered.
This is done by telling interviewees, ‘Ž rmly’, that ‘You have made a number of
errors. It is therefore necessary to go through the questions once again, and this
time try to be more accurate’. The questions are then repeated. Three further scores
are then calculated: Yield 2, Shift, and Total Suggestibility. Yield 2 is a measure of
the number of suggestions that are accepted following the negative feedback. Shift
is a measure of the number of responses an interviewee changes subsequent to the
negative feedback and includes all 20 of the questions. Total Suggestibility is the
sum of Yield 1 and Shift.
The administration of negative feedback is an important aspect of admin-
istration of the scales. It is assumed that negative feedback applies a certain level
of interrogative pressure and that variations in style should be avoided.
Gudjonsson (1984a) raised concern over the consistency with which negative
feedback is administered. If, for example, an interviewee has made no errors it
may be awkward or embarrassing for the interviewer to administer negative
feedback and this may a ect scoring of Shift. Haraldsson (1985) found that
V
interviewer embarrassment when administering the negative feedback reduced
Shift, although the e ect was not signiŽ cant. Gudjonsson and Lister (1984) noted
V
a non-signiŽ cant tendency for one of their experimenters to gain higher Shift
scores than the other.
Baxter and Boon (in press) argued that one e ect of being Ž rm rather than
V
friendly in administering negative feedback may be to maximize psychological
distance, giving the interrogator a ‘tactical advantage’ over the interviewee (cf.
Gudjonsson & Lister, 1984). Baxter and Boon argued that Ž rm negative feedback
may also have a related e ect in that it precipitates compliance. They conducted a
V
study to assess the impact of varying only the negative feedback component of the
GSS2. Participants were tested under one of three conditions which varied the
Ž rmness of negative feedback. The three conditions were deŽ ned as ‘friendly’,
‘Ž rm’ and ‘stern’ negative feedback demeanour. Baxter and Boon found Yield 2 and
Shift scores to increase as interviewer demeanour changed from ‘friendly’ through
to ‘stern’ when administering negative feedback, although Total Suggestibility
scores did not di er signiŽ cantly across their conditions. They argued, in support of
V
Gudjonsson (1983), that Yield 1 is primarily related to cognitive factors of the
individual, whereas Yield 2 may better represent interpersonal in uences present at
negative feedback. They also argued that, because Total Suggestibility is a balanced
measure of pre- and post-feedback interrogative suggestibility, it may be relatively
insensitive to the e ects of variations in interviewer manner in delivering negative
V
feedback.
If the manner in which negative feedback is administered is important, then there
may be more general problems associated with experimenter e ect in the
V
administration of the scales. It may be that a particular style of interviewing results
in higher suggestibility scores even where Ž rmness of negative feedback is held

126
Stella A. Bain and James S. Baxter
constant. Gudjonsson and Lister (1984) examined the role of self-concept and
locus of control in the extent to which individuals are susceptible to suggestion. In
this study all participants were placed under the same interrogative pressure. The
major Ž nding was that perception of distance between the participant and the
experimenter was highly correlated with suggestibility. Gudjonsson and Lister
identiŽ ed the variables of perceived lack of conŽ dence and control in coping with
the interrogation as being most clearly related to suggestibility. Feelings of anxiety
and powerlessness were also identiŽ ed as important variables. Psychological
distance between an interviewee and the interrogator may create a certain pressure
that makes the interviewee more susceptible to suggestion (cf. Gudjonsson &
MacKeith, 1982; Irving, 1980).
These results suggest that certain interrogation techniques could be used to
increase psychological distance in the interrogative context. By manipulating an
individual’s level of self-esteem and perception of power and control, such
techniques could perhaps make individuals more susceptible to suggestion
than they would otherwise be if such techniques were not used. Baxter and Boon
(in press) manipulated interviewer demeanour only during the presentation of
negative feedback. The present study tested the hypothesis that the greater the
psychological distance created between the interviewer and the interviewee by a
general demeanour adopted throughout the interview, the higher GSS1 scores will
be. It was predicted that a generally abrupt demeanour would produce greater
psychological distance, and hence higher GSS1 scores, than a friendly demeanour.
Method
Design
The study had a single factor between participants design. The independent variable was interviewer
behaviour which was varied such that there were two conditions: ‘Friendly’ and ‘Abrupt’ interviewer
behaviour.
Participants
Most participants were Ž rst year undergraduate psychology students. Others were university
administrative sta (mean age = 25.18 years, SD = 13.57, range 17–74). In all, 29 females and 26 males
V
took part in the study. All participants were told on recruitment that they would take part in memory
research. Of the participants, 29 were assigned to the Friendly condition and 26 to the Abrupt
condition. Participants were assigned to conditions at random. Numbers of females and males were
approximately equal in each condition.
Procedure
The GSS1 (see Gudjonsson, 1997 for a full description of the procedure) was administered under two
di erent conditions of interviewer behaviour. The same female experimenter conducted all interviews
V
in both conditions.
In the Friendly condition the experimenter smiled when participants entered the test room,
thanking them for taking part. The experimenter attempted to respond in a friendly manner to any
conversation initiated by the participant prior to testing, and to maintain this manner in explaining the
procedure to participants. The experimenter smiled frequently and always smiled back if participants

Interrogative suggestibility
127
smiled. A body position of leaning back, away from the table and the participant, was adopted for the
entire period of the experiment. Eye contact was maintained throughout testing, except when
consulting the scoring sheet.
In the Abrupt condition no attempt was made to build rapport or be friendly when the participant
entered the o ce. The experimenter gave minimal responses to any attempts at conversation by the
Y
participant and limited speaking to issuing instructions in an abrupt manner. The experimenter did not
smile or make any facial response to anything the participant said. Instead, an expression intended to
convey mild annoyance was maintained throughout the experiment. The experimenter adopted a body
position of leaning forward across the table towards the participant. Again, eye contact was
maintained by the experimenter throughout the period of testing, except when it was necessary to
consult the scoring sheet.
Apart from interviewer manner, the only departure from standard GSS procedure was to present
the narrative on audiotape in an attempt to standardize conditions at the time of encoding.
Participants completed an unrelated Ž ller task in the 50 minutes between testing of immediate and
delayed recall.
On completion of GSS1 testing participants completed a questionnaire which asked for 5-point
Likert scale ratings on 18 aspects of the interviewer’s manner. These aspects were: nervous, severe,
friendly, understanding, assertive, conŽ dent, professional, Ž rm, respectful, positive, formal, warm,
stern, organized, e ective, authoritative, competent, and negative, with a high score being more
V
nervous, etc.
Before leaving all participants were debriefed. Those participants who had been tested under the
Abrupt condition were told that the experimenter’s behaviour was a necessary part of the experiment.
Scoring the GSS1
Scoring of the GSS1 was done in accordance to the guidelines provided by Gudjonsson (1997).
(i) Memory recall. Both immediate and delayed recall are scored for each correct idea recalled. Ideas are
scored as correct if the meaning is the same as the original item in the narrative. Each correct idea
earns one point, with the maximum score being 40. The scoring of memory recall is not included in
the scoring of suggestibility.
(ii) Suggestibility. The scale provides four scores:
(1) Yield 1. Every suggestive question that is answered a rmatively, or in the case of false alternatives
Y
where one alternative is chosen, in the Ž rst period of questioning is scored as one yield point. The
range of possible scores is 0 to 15.
(2) Yield 2. This measure is scored in the same manner as Yield 1 following the administration of the
negative feedback. Again, possible range of scores is 0 to 15.
(3) Shift. Any distinct change in response to all 20 questions in the second period of testing is scored
as a shift. Thus, possible shift scores range from 0 to 20.
(4) Total Suggestibility. Total Suggestibility is the sum of Yield 1 and Shift. Therefore, the range of
scores is 0 to 35.
Results
A one-way ANOVA revealed no signiŽ cant di erences for Yield 1 or Yield 2
V
between the two conditions. SigniŽ cant di erences for Shift (F(1,54) = 6.2,
V
p = .016) and Total Suggestibility (F(1,54) = 4.5, p = .039) were found between the
two conditions. Participants in the Abrupt condition scored signiŽcantly higher on
Shift and Total Suggestibility than those in the Friendly condition. Associated
descriptive statistics are shown in Table 1.

128
Stella A. Bain and James S. Baxter
Table 1.
Mean suggestibility scores3 condition and GSS1 norms
Condition
Friendly
Abrupt
GSS1 norms
M
SD
M
SD
M
SD
Yield 1
Yield 2
Shift
4.1
5.4
2.2
6.6
2.7
3.2
1.4
3.1
4.9
5.9
4.0
8.9
2.4
3.1
2.9
4.8
4.6
5.6
2.9
7.5
3.0
3.8
2.5
4.6
Total suggestibility
Note: Norms derived from Gudjonsson (1997).
Table 2.
Mean recall scores3 condition and GSS1 norms
Condition
Friendly
Abrupt
GSS1 norms
M
SD
M
SD
M
SD
Immediate recall
Delayed recall
23.7
21.2
5.5
5.9
24.1
21.7
5.8
6.8
21.3
19.5
7.1
7.5
Note: Norms derived from Gudjonsson (1997).
Two one-way ANOVAs conducted on the data for Immediate and Delayed
Recall 3 Condition found no signiŽ cant di erences. Means and standard
V
deviations for Memory Recall (Immediate and Delayed) are shown in Table 2.
A one-way ANOVA conducted on the ratings of interviewer behaviour showed
signiŽ cant di erences between the two conditions for participants’ ratings on 12 of
V
the adjectives included on the form. The signiŽ cant results were as follows: severe
(F(1,54) = 21.7, p< .001); friendly (F(1,54) = 48.6, p< .001); understanding
(F(1,54) = 52.7, p< .001); assertive (F(1,54) = 5.6, p = .022); Ž rm (F(1,54) = 15.1,
p< .001); respectful (F(1,54) = 14.4, p< .001); positive (F(1,54) = 11.1, p = .002);
formal (F(1,54) = 15.6, p< .001); warm (F(1,54) = 47.3, p< .001); stern
(F(1,54) = 49.6, p< .001); authoritative (F(1,54) = 24.6, p< .001); and negative
(F(1,54) = 6.8, p = .012). Ratings for nervous, conŽ dent, professional, organized,
e ective and competent did not di er between the conditions. Participants rating
V
V
the behaviour of the Friendly interviewer were more likely to rate the interviewer
as friendly, understanding, respectful, positive and warm. Participants rating the
behaviour of the Abrupt interviewer were more likely to rate the interviewer
as severe, assertive, Ž rm, formal, stern, authoritative and negative. Associated
descriptive statistics are shown in Table 3.

Interrogative suggestibility
129
Table 3.
Interviewer behaviour ratings3 condition
Condition
Friendly
Abrupt
M
SD
M
SD
Nervous
Severe
1.3
1.3
4.3
4.1
3.5
4.2
4.3
3.0
4.3
4.2
3.0
3.8
1.4
4.5
4.2
2.5
4.4
1.2
0.6
0.6
0.8
0.9
1.0
0.8
0.9
1.2
0.7
0.8
1.2
0.9
0.8
0.9
0.9
1.1
0.8
0.7
1.4
2.4
2.6
2.3
4.1
4.1
4.3
4.1
3.4
3.3
4.1
2.1
3.3
4.5
4.2
3.9
4.4
1.8
0.7
1.0
0.9
0.8
0.9
1.1
0.9
0.9
1.1
1.1
1.0
0.9
1.1
0.9
1.0
1.0
0.9
0.9
Friendly
Understanding
Assertive
ConŽ dent
Professional
Firm
Respectful
Positive
Formal
Warm
Stern
Organized
E ective
V
Authoritative
Competent
Negative
In sum, the results indicate that interviewer behaviour had a signiŽcant e ect on
V
participants’ scores on the GSS1. This e ect operated on Shift and Total
V
Suggestibility scores only. No e ect of variations in interviewer behaviour was
V
found on Yield 1 or Yield 2 scores.
Discussion
That the attempt to vary interviewer behaviour was successful seems broadly
conŽ rmed by the results obtained from the interviewer rating forms. In the
Friendly condition the interviewer tended to be rated as friendly, understanding,
respectful, positive and warm. In the Abrupt condition the interviewer tended to be
rated as severe, assertive, Ž rm, formal, stern, authoritative and negative (see Table
3). The scores for memory recall did not vary as a function of experimental
condition and were closely comparable to norms for the general population (see
Table 2), suggesting that each group had comparable memories for the GSS1
narrative against which to compare con icting information contained in the
questions.
SigniŽ cant di erences between the experimental conditions were found for Shift
V
and Total Suggestibility scores with participants in the Abrupt condition gaining

130
higher scores on these measures than those in the Friendly condition. The Yield 1
scores shown in Table 1 indicate that the di erence in Total Suggestibility was due
Stella A. Bain and James S. Baxter
V
primarily to the di erence in the Shift component on this measure between the
V
conditions. Shift is assumed to measure the e ects of interrogative pressure on an
V
individual (Gudjonsson, 1983), pressure which is applied overtly by administering
negative feedback and implicitly by requestioning in the standard GSS procedure.
These procedures did not di er, however, between the conditions in the present
V
study: negative feedback was administered ‘Ž rmly’ in both conditions and all
participants were questioned twice. The tendency of participants in the Abrupt
condition to be more likely to change their responses at requestioning seems most
likely, therefore, to be because of the generally negative manner of the interviewer
in that condition.
The Yield 1 and Yield 2 scores did not di er signiŽ cantly between the conditions
V
(see Table 1), although both tended to be higher in the Abrupt condition. These
scores departed only slightly from the general population norms for the scales. It
does not seem, therefore, that interviewer manner as manipulated in the present
study signiŽ cantly a ected either the capacity to detect misleading information or a
V
tendency to comply with the demands of leading questions. One reason for this
may be that an overall interviewer manner which interviewees have no reason to
link to their own behaviour communicates no particular expectancy, but increases
interviewee uncertainty above the levels created by negative feedback alone in the
standard GSS procedure. That this additional pressure operated primarily on Shift
rather than Yield 2 of the two post-feedback scores may indicate that its e ect is to
V
increase the likelihood that interviewees will examine what was ‘wrong’ with their
previous answers in an attempt to identify what can be changed. That Yield 2 was
not found to be signiŽ cantly di erent between conditions may be because Yield 2
V
provides a speciŽ c measure of the e ect of negative feedback and the valence
V
of negative feedback was held constant between conditions (cf. Baxter & Boon,
in press; Gudjonsson, 1983, 1992, 1997).
Shift is assumed to measure the overall e ects of interrogative pressure on an
V
individual (Gudjonsson, 1983) and primary sources of interrogative pressure in the
standard GSS procedure are negative feedback and the further negative feedback
implicit in requestioning. However, the present results have identiŽ ed a further
factor which can contribute to Shift. Interviewees can evidently be pressured to
shift their responses more frequently if an interrogator has an abrupt manner than
will be the case if an interrogator is friendly but nonetheless delivers ‘Ž rm’ negative
feedback. One reason for this, which further work might assess speciŽ cally, may be
that a generally negative interviewer manner does not necessarily bias interviewees’
responses by causing them to attend to external cues at the expense of internal
cues, as the standard GSS negative feedback procedure is assumed to do
(Gudjonsson, 1992). Interviewees may continue to attend to internal cues, but may
devalue them. A further possibility is that participants may have attempted to
appease the severe interviewer by simply complying with her perceived demands,
knowingly shifting their initial responses while still believing them to have been
correct. A limitation of the present study is that it cannot identify the relative
contributions which devaluation of internal cues, compliance and other in uences

Interrogative suggestibility
131
may have made to the levels of interrogative suggestibility found: these and other
in uences may have operated in isolation or interactively to various degrees
between and within interviewees throughout the procedure or at di erent stages of
V
the procedure (Baxter & Boon, in press; Gudjonsson, 1992, 1997; Gudjonsson &
Clark, 1986).
These Ž ndings are consistent with Gudjonsson’s argument that interrogative
suggestibility consists of two distinct types of suggestibility (Gudjonsson, 1984a,
1992, 1997). Yield 1 assesses the e ect of leading questions or suggestive stimuli on
V
accuracy of testimony and the present results indicate that this factor was not
a ected by interpersonal factors, although Shift and Total Suggestibility were found
V
to be. This Ž nding may mean that initial responses to leading questions are
mediated by more stable cognitive factors, perhaps involving a capacity for source
monitoring or discrepancy detection, that are relatively una ected by the manner of
V
the interrogator, whereas the post-feedback GSS measures may be more sensitive
to social aspects of suggestibility.
Note that, although the Shift and Total Suggestibility di erences found in the
V
present study were statistically signiŽ cant, they were relatively modest and all scores
obtained in the present study fell within one standard deviation of general
population norms. It should also be remembered that most participants were
undergraduates and the interrogative suggestibility of a student sample may not be
representative of the general population, even though, as noted above, pre-
feedback and recall scores obtained in the present study were comparable to general
population norms for the scale used (Gudjonsson, 1997). More vulnerable
interviewees may be more sensitive to variations in interviewer manner and future
research might usefully assess this possibility.
In practical terms the present Ž ndings suggest, in line with the conclusions of
Haraldsson (1985) and Baxter and Boon (in press), that prospective users of the
Gudjonsson scales should be aware that variations in their demeanour may bias
their results with the consequent danger that they will fail to identify vulnerable
witnesses if their demeanour is too mild, or may falsely identify acceptable
witnesses as vulnerable if their demeanour is too severe. A comparison of the
present Ž ndings with those of Baxter and Boon may provide proŽ les of two
di erent types of problem interviewers. (That the comparability of these two
V
studies should not be a ected by the use in the present study of the GSS1, rather
V
than the GSS2, is evident from the norms for the scales provided by Gudjonsson,
1997.)
Baxter and Boon (in press) varied the manner in which negative feedback was
delivered when testing with the GSS2, but kept interviewer demeanour ‘neutral’
throughout the other stages of the procedure. They reported that an interviewer
whose manner was ‘stern’ or ‘severe’ rather than ‘Ž rm’ when delivering negative
feedback produced higher Yield 2 and Shift scores when using the GSS2 than an
interviewer whose manner was generally ‘friendly’ when delivering negative
feedback. They found no signiŽcant di erences in Total Suggestibility as a function
V
of this manipulation. The method used in the present study essentially inverted
the procedure used by Baxter and Boon: interviewer demeanour at negative
feedback was ‘Ž rm’ in both conditions (the closely comparable Yield 2 scores

132
Stella A. Bain and James S. Baxter
appear to conŽ rm this), but general demeanour throughout the other stages of the
GSS procedure was varied. This may mean that interviewers who misinterpret
Gudjonsson’s (1992, 1997) instruction to be ‘Ž rm’ in delivering negative feedback,
such that they are either too severe or too familiar, will tend to produce Yield 2 and
Shift scores which are, on average, artiŽ cially too high or too low respectively.
Their Total Suggestibility scores, however, may be less seriously a ected by such
V
variability. The present study shows, by contrast, that interviewers who deliver
negative feedback correctly but whose demeanour is generally too severe or too
familiar will produce normal Yield 2 scores but will also produce Shift and Total
Suggestibility scores which are, on average, artiŽ cially too high or too low
respectively. This may be the more serious problem than that identiŽ ed by Baxter
and Boon, given that it a ects the key Total Suggestibility measure. Interviewers
V
showing this constellation of scores may have a more general and less tractable
diYculty than simply adopting the correct manner at negative feedback and may
require careful monitoring.
Taken together, these two sets of Ž ndings suggest a further function of the
Gudjonsson Suggestibility Scales. If Yield 1 is indeed relatively independent of
interviewer manner, then GSS results which show normal Yield 1 scores but
lowered or raised post-feedback scores may identify interviewees who are suggest-
ible because they are particularly vulnerable to interpersonal pressure. However,
such a pattern of scores may also identify interviewers whose manner is either
especially informal or especially overbearing. It would seem therefore that anybody
intending to use these scales for research, clinical or forensic purposes should Ž rst
establish in practice sessions that their standard technique produces results closely
comparable to the established norms for the scales (Baxter & Boon, in press).
References
Baxter, J. S., & Boon, J. C. W. (in press). Interrogative suggestibility: The importance of being earnest.
Personality and Individual Di erences.
V
Gudjonsson, G. H. (1983). Suggestibility, intelligence, memory recall and personality: An experimental
study. British Journal of Psychiatry, 142, 35–37.
Gudjonsson, G. H. (1984a). A new scale of interrogative suggestibility. Personality and Individual
Di erences, 5, 303–314.
V
Gudjonsson, G. H. (1984b). Interrogative suggestibility: Comparison between ‘false confessors’ and
‘deniers’ in criminal trials. Medicine, Science and Law, 24, 56–60.
Gudjonsson, G. H. (1987). A parallel form of the Gudjonsson Suggestibility Scale. British Journal of
Clinical Psychology, 26, 215–221.
Gudjonsson, G. H. (1988). Interrogative suggestibility: Its relationship with assertiveness, social
evaluative anxiety, state anxiety and method of coping. British Journal of Clinical Psychology, 27,
159–166.
Gudjonsson, G. H. (1992). The psychology of interrogations, confessions and testimony. Chichester: Wiley.
Gudjonsson, G. H. (1997). The Gudjonsson Suggestibility Scales Manual. Hove: Psychology Press.
Gudjonsson, G. H., & Clark, N. K. (1986). Suggestibility in police interrogation: A social
psychological model. Social Behaviour, 1, 83–104.
Gudjonsson, G. H., & Lister, S. (1984). Interrogative suggestibility and its relationship with
perceptions of self-concept and control. Journal of the Forensic Science Society, 24, 99–110.
Gudjonsson, G. H., & MacKeith, J. A. C. (1982). False confessions, psychological e ects of
V
interrogation. A discussion paper. In A. Trankell (Ed.), Reconstructing the past: The role of psychologists
in criminal trials (pp. 253–269). Deventer: Kluwer.

Interrogative suggestibility
133
Haraldsson, E. (1985). Interrogative suggestibility and its relationship with personality, perceptual
defensiveness and extraordinary beliefs. Personality and Individual Di erences, 6, 765–767.
V
Irving, B. (1980). Police interrogation. A case study of current practice. Research studies No. 2. London: HMSO.
Loftus, E. F. (1979). Eyewitness testimony. Cambridge, MA: Harvard University Press.
Received 15 January 1999; revised version received 17 May 1999