Professional Documents
Culture Documents
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/
info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content
in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship.
For more information about JSTOR, please contact support@jstor.org.
University of Illinois Press is collaborating with JSTOR to digitize, preserve and extend access to The American Journal of
Psychology.
http://www.jstor.org
This content downloaded from 130.237.29.138 on Sat, 27 Feb 2016 16:02:46 UTC
All use subject to JSTOR Terms and Conditions
Are subjective memory problems related to
suggestibility, compliance, false memories,
and objectivememoryperformance?
SASKIAVANBERGEN, MARKOJELICIC,
and HARALD
MERCKELBACH
MaastrichtUniversity
compliance,falsememo
between subjectivememorybeliefsand suggestibility,
The relationship
was studied ina communitysample of young and
ries,and objectivememoryperformance
=
middle-agedpeople (N 142).We hypothesizedthatpeoplewith subjectivememoryproblems
and compliance levelsand would be more susceptibleto
would exhibithighersuggestibility
thanthosewho are optimisticabout theirmemory.Inaddition,we expected
false recollections
a discrepancybetween subjectivememoryjudgmentsand objectivememoryperformance. We
with compliance,with more
foundthatsubjectivememoryjudgmentscorrelatedsignificantly
negativememoryjudgmentsaccompanyinghigherlevelsof compliance.Contraryto our expec
tation,subjectivememoryproblemsdid not correlate or false recollections.
with suggestibility
Furthermore, were accurate inestimatingtheirobjectivememoryperformance.
participants
There are large individual differences inhow people far-reaching consequences. Given theweight that
evaluate theirown memory. Although most of us tend triersof fact attach to confidence (Leippe, Manion,
tobe quite optimistic about thepower of ourmemory, & Romanczyk, 1992), people with such pessimistic
some people believe theirmemory ismuch poorer ideasmay erroneously be treated as less credible eye
than thatof others from theirown age group (Crom witnesses or
suspects.
are hints in the literature
bag, Merckelbach,
8c Elffers, 2000; Magnussen et There suggesting that
ideas about ideas about one's own memory are associat
al., 2006). However, subjective memory negative
do not always correspond to objective memory per edwith elevated suggestibility levels and an enhanced
formance (Ponds 8c Jolles, 1996; Ponds, Van Boxtel, susceptibility to false recollections. People who judge
8c Jolles, 2000). For example, healthy older people their own memory as very poor because they suffer
who rate theirmemory from the "memory distrust syndrome" are thought to
(more than 55 years old)
as very poor often exhibit normal performance on be especially prone tomemory distortions (Gudjons
standard memory tasks. In a forensic setting, pes son 8cMacKeith, 1982). According toGudjonsson
simistic ideas about one's own memory might have (2003, p. 196),memory distrust is "a condition where
Summer 2009, Vol. 122,No. 2 pp. 249-257 ? 2009 by the Board ofTrustees of theUniversity of Illinois
This content downloaded from 130.237.29.138 on Sat, 27 Feb 2016 16:02:46 UTC
All use subject to JSTOR Terms and Conditions
people develop profound distrust of theirmemory plywith the suggestive cues given by others. Further
recollections, as a result ofwhich theyare particularly more, based on thework of Crombag et al. (2000),
susceptible to relying on external cues and sugges we expected that
people would overestimate their
tions."This author described a number of court cases own memory
functioning. This hypothesis is con
inwhich defendants sufferingfrommemory distrust sistentwith research demonstrating that in healthy
developed false memories, eventually resulting in older adults, subjective evaluations ofmemory do not
false confessions (Gudjonsson, 2003; Gudjonsson, appear to be associated with objective memory per
Kopelman, 8cMacKeith, 1999). Note that such iso formance (Ponds & Jolles, 1996; Ponds et al, 2000).
lated single cases do not provide hard evidence for We decided to select a sample of young and middle
an intimateconnection between
pessimistic opinions aged participants so as to be sure that the subjective
about one's own memory and suggestibility.Further memory judgments we measured were not linked to
more, we believe that a distinction should be drawn age-related memory problems.
between state and traitmemory distrust, with the for
mer referringto the cases described in the literature
EXPERIMENT
and the lattermanifesting itselfas a personality trait.
In this article,we will focus on trait
memory distrust, METHOD
which can be seen as a negative subjective memory
judgment.
Participants
Compliance is a concept related to suggestibility. Our sample consisted of 142 young and middle
It can be defined as "a tendency of the individual to
aged research participants (37men), who were re
go along with propositions, requests or instructions, cruited in a regional news
through advertisements
for some immediate instrumental =
gain" (Gudjons paper. Their mean age
was 34.23
years (SD 8.06,
son, 2003, p. 370). Whereas suggestibility assumes =
range 17-46). The participants were told that
that people accept the information provided, this would be administered severalmemory tasks
they
does not apply for compliance. Nonetheless, the and questionnaires. They were given financial
consequences of complying can be far reaching in compensation ( 25) forparticipating in our study.
some The researchwas approved by the standing ethical
settings.For example, thinkof complying with
statements about amurder committeeof theFaculty ofPsychology,Maastricht
expressed by a police of
ficer or complying with a
therapist's suggestion that University.
This content downloaded from 130.237.29.138 on Sat, 27 Feb 2016 16:02:46 UTC
All use subject to JSTOR Terms and Conditions
functioning:theDutch translationof theCognitive second set ofquestions (maximum = 15). Shift is the
FailuresQuestionnaire (Broadbent,Cooper, Fitzger number of answers that the participant changes
as
ald,& Parkes, 1982;Merckelbach, Muris, Nijman, & a result of thenegative feedback (maximum = 20).
= The total suggestibility score is the sum ofYieldi
Dejong, 1996),with Cronbach's a .92.The Cogni
tiveFailures Questionnaire consists of 25 items that and Shift,with higher scores indicatinghigher sug
measure
self-reported frequency of everyday lapses gestibilitylevels.Research has shown thattheDutch
and errors inmemory, perception
and attention, and version of theGudjonsson Suggestibility Scale is a
action. Illustrative items are "Do you forget where psychometrically sound instrument(Merckelbach,
you put something like a newspaper or a book?", "Do Muris, Wessel, 8c Van Koppen, 1998). More spe
you fail tonotice signposts on the road?", and "Do Merckelbach et al. reported sufficientinter
cifically,
you drop things?"Participants are asked to indicate nal consistency,with Cronbach's alphas being .79
on a
5-point scale how often they have experienced (Yieldi), .75 (Shift), and .82 (total suggestibility).
each cognitive
failure in recent months (o
=
never, Test-retest stabilitywas modest (r = .55).They also
= Scores are summed to obtain a total found indications for thepredictive validity of the
4 very often).
sure suggestibility.The Gudjonsson Suggestibility The items have a true-false format. After recoding,
Scale consists of a story thatis read out loud by the scores are summed
(range 0-20),
with higher scores
memory questions. After theyhave answered the ediger-McDermott (DRM) paradigm (Roediger 8c
questions, participants
receive
negative feedback McDermott, 1995).The 10word lists thatwere used
irrespective
of their performance (i.e.,
"You have comprised 15Dutch words each. Each listcontained
made a number of errors. It is therefore necessary words bed, dream) thatwere associated
(e.g., drowsy,
to go through the questions once more, and this with a nonpresented themeword (sleep).This word
time, try to be more accurate"). Next, the questions is called the critical lure.Extensive pilot studies by
are
repeated. Peters,Jelicic, andMerckelbach (2008) showed that
The Gudjonsson SuggestibilityScale yields sev proportions of recall and recognitionof theseword
eral parameters. First, with regard
to the free recall, lists are comparable to those reported by Roediger
thenumber of correctlyreproduced storyelements and McDermott. Participants were told that they
= for both would hear several word lists and would be tested
40). This
is counted is done
(maximum
to
immediate and delayed recall. The corresponding immediatelyafterhearing each listby being asked
interrater reliability parameters
were .95 and .93, write down all thewords theycould remember. They
<
.001).Also, four suggestibil lists were read out
were instructed not to guess. The
respectively(bothps
can be derived from theGudjonsson loud by the experimenter at a pace of 1word per sec
ityparameters
Suggestibility Scale. Yieldi is thenumber ofmislead ond, with an interstimulus interval of 1 s. After each
ing elements that the participants accept during the list,participantswere given 2min towrite down all
firstset of questions (maximum = 15).Yield2 is the thewords theycould remember.Split-half reliability
number of accepted misleading elementsduring the was excellent for correctly recalled words (Spearman
This content downloaded from 130.237.29.138 on Sat, 27 Feb 2016 16:02:46 UTC
All use subject to JSTOR Terms and Conditions
Brown coefficient= .92) andmoderate forthecritical delayed recall followed by recognition,participants
=
lures coefficient several unrelated filler
(Spearman-Brown .54), completed questionnaires.
After all listshad been presented, a recognition
testwas given.This testconsisted of 30 old words RESULTS
and 30 new words. Three items per list correspond
Table l shows themean scores, standard deviations,
ingwith serialpositions 1,8, and 10were selected as
old words. Ten of thenew words were critical lures; and ranges of the self-reportquestionnaires, theGud
the other 20 words were derived fromotherword jonsson Suggestibility Scale, the falsememory task
lists that were not used during this (DRM), and theAuditory Verbal Learning Test.
experiment.
20 questions about the story,and aftertheyhad been nificant positive correlation was found between old
false feedback on their were words that were remembered as old words
given performance, they correctly
instructed to answer the 20 questionsand the SSMQ, r = .ig,p < .05. Such a relationship
again. Next,
theAuditoryVerbal LearningTest was administered. was absent for the SSMQ and critical lures thatwere
In the time intervalbetween immediate recall and remembered as old r = -.12, =
.16.
words, p
This content downloaded from 130.237.29.138 on Sat, 27 Feb 2016 16:02:46 UTC
All use subject to JSTOR Terms and Conditions
TABLE 1. Mean scores,standarddeviations,and rangesof thequestionnairesand tests ina
=
communitysample (N 142)
Instrument M
SD Range
Deese-Roediger-McDermott Task
Recall
91.07
Correct 16.17 53-128
False:criticallures 4.97 2.23 0-10
Recognition
Correct:oldwords 23.94 3.69 10-30
False: criticallures 8.35
2.21 1-10
Verbal LearningTest
Auditory
Immediaterecall
46.46
Correct 11.1021-69
1.27
Incorrect 1.97
0-10
Delayed recall
9.75
Correct 3.411-15
Incorrect 0.37 0-3
0.67
Recognition
Truepositive 13.97 4-15
1.69
Falsepositive 0.27 0.72 0-6
False negative 1.04 1.69 0-11
This content downloaded from 130.237.29.138 on Sat, 27 Feb 2016 16:02:46 UTC
All use subject to JSTOR Terms and Conditions
r = -.25, <
.05. No correla
negatives, p significant
TABLE 2. Pearson's product-moment correlations tion emerged between SSMQand falsepositives, that
between subjectivememory(indexedby theSquire
is, falsely recognizing words thatwere not studied,
SubjectiveMemoryQuestionnaire)and othertests ina r =
-.og,p
=
.30.
=
community sample (N 142)a
Extreme values
Instruments r Because correlational analyses assume linearity,and
this assumption might not be true forour sample,we
CognitiveFailuresQuestionnaire -.72*
conducted additional analyses forwhich we created
GudjonssonComplianceScale -.22*
two groups consisting of participants with extreme
Suggestibility
SSMQ values: a poor memory group, with the 25%
Immediaterecall .30* lowest SSMQ scores (n = 35), and an excellentmemo
=
Delayed recall .30* rygroup,with the25% highest SSMQ scores (n 38).
Yieldl .13 Independent sample t tests revealed a similar pat
Yield2 .01 ternas was obtained with Pearson product-moment
Delayed recall
Correct .29*
Incorrect .00
TABLE 3. Correlationmatrixof questionnairesand
testscores ina community =
sample (N 142)
Recognition
Truepositive .24*
Tests CFQ GSS GCS DRM AVLT
False positive -.09
? -.14 .12 -.01 -.14
CFQ
False negative -.25*
GSS ? -.06 .19* -.17
aTheapplication of a Bonferronicorrectiondid not change the
?
general pattern. GCS .05 -.10
*p < .05. ?
DRM -34*
AVLT ?
Note. AVLT= correct immediate recall of theAuditoryVerbal
=
Learning Test;CFQ Cognitive Failures Questionnaire total score;
DRM = Deese-Roediger-McDermott recall of critical lures;GCS =
=
Gudjonsson Compliance Scale total score; GSS Gudjonsson Sug
gestibilityScale total score.
*p < .05.
This content downloaded from 130.237.29.138 on Sat, 27 Feb 2016 16:02:46 UTC
All use subject to JSTOR Terms and Conditions
were negatively correlated, r = -.34,^ < .05.The other the SSMQ and suggestibility.However, it should be
correlations did not reach significance. noted that levels of interrogativepressure during the
ing frommemory distrust (i.e., who evaluate their ing their own memory than older adults with self
memory as poor) aremore susceptible to the author reportedmemory problems. Interestingly,our results
ityof others. As a result, theyperceive the of fitnicely with those of Brewin and Stokou
opinion (2002),
others as more important than theirown opinions. who found thatpeople who judge themselves tohave
However, because this relationship was based only poor childhood memory perform worse on a stan
on correlations, the causal
relationship between these dardized autobiographical memory test than those
two concepts cannot be established. One could also who report normal childhood memories. It is diffi
argue thatpeople start to distrust theirmemory be cult to saywhy some of the older people believe
they
cause theyaremore easily intimidated
by authorities have poor memory capabilities when there is
nothing
who might have played an active role inundermining wrong with theirmemory. Perhaps these people had
the confidence inone's memory abilities.This points been confrontedwith cases of dementia in their rela
out that the relationship between memory distrust tives and friends. Such an experience may lead one
and compliance could go both ways. to interpretnormal age-related memory decline as a
In this study,we failed to find a significant cor form of pathological aging (Commissaris, Ponds, 8c
relation between the SSMQ and false recollections Jolles, 1998). Our younger and middle-aged partici
(i.e., critical lures) elicited by theDRM task.This pants may not have these experiences, and this may
finding could be explained by themoderate reliability help to explain why in thisgroup subjective and ob
of this task. Furthermore, following Gudjonsson's jective memory parameters are not dissociated to the
line of reasoning, one would also expect high suggest extent often seen inolder adults. Our sample may also
ibility levels inpeople with negative memory beliefs have more realisticmemory beliefs because
they are
(Gudjonsson, 2003). Contrary to our expectation, more often exposed to feedback about theirmemory
we did not find evidence to support thishypothesis. performance (e.g., in the context of theirwork or edu
Thus, no significantcorrelationswere found between cation), which could promote a better calibration.
This content downloaded from 130.237.29.138 on Sat, 27 Feb 2016 16:02:46 UTC
All use subject to JSTOR Terms and Conditions
In addition, we found thatpoor memory perfor For the forensic setting, our results imply that
mance (indexed
by immediate recall on theAuditory people who have optimistic ideas about theirmem
Verbal Learning Test) was related tomore false recol ory generally show bettermemory performance and
lections (i.e.,more critical lures).However, thiswas to lower compliance scores than thosewith pessimistic
be expected because "(partial) amnesia is a necessary views. However, these correlations are modest. Fur
condition for the development of a full-blown false thermore,people with optimistic beliefs about their
autobiographical memory" (Smeets, Merckelbach, memory are not less susceptible to suggestions and
Horselenberg, 8c Jelicic, 2005, p. 925). false recollections than people with pessimistic be
There are several limitations in thepresent study liefsabout theirown memory. Therefore, one should
thatdeserve some comment. First, although the cor not regard their testimonies as more reliable.
relations between subjectivememory judgments and
thefact thatthe sample did not involvepatients or older Organization for Scientific Research (400.04.048). Thanks
are due toHeleen van
adults (i.e., participants with more extreme opinions Hoynck Papendrecht and Francine
Schneider for data collection.
about theirmemory). Furthermore, themodest cor
Address correspondence about this article to Saskia
relationsmay have todo with the fact that the testwe
van of Clinical Science,
Bergen, Department Psychological
used tomeasure objective memory performance, the
Faculty of Psychology and Neuroscience, Maastricht Univer
a
Auditory Verbal Learning Test, taps only specific sity,P.O. Box 616,6200 MD Maastricht, The Netherlands
type ofmemory (i.e., learning of new verbal informa (e-mail: s.vanbergen@psychology.unimaas.nl). Received for
tion). Second, anotherweakness of thepresent study publication January 18,2008; revision received July 16,2008.
is thatfor technical reasons, scales and taskswere given Action Editor: Donalson E.
Dulany.
in a fixed order, and thismay have introduced carry
over effects.Finally and most importandy,inour study, REFERENCES
people with low SSMQ scores might not necessarily Brewin, C. R., 8c Stokou, L. (2002). Validating reports of
sufferfroma full-blownmemory distrust syndrome. It poor childhood memory. Applied Cognitive Psychology,
results cannot be easily generalized tomemory distrust cation and Counseling, 34, 25-32.
cases. Therefore, futurestudies should look atwhether Crombag, H., Merckelbach, H., 8c Elffers, H. (2000). Other
memory. Psychology, Crime &Law, 6, 251-265.
people sufferingfrom statememory distrust are likely
people's
Deelman, B. G., Brouwer, W. H., Van Zomeren, A. H., 8c
to create falsememories and have high suggestibility
Saan, R. J. (1980). Functiestoornissen na trauma
capi
levels. Itwould be interestingto examine how memory tis [Cognitive dysfunctions after head trauma]. In A.
distrust is related to other typesof tasks,notably tasks Jennekens-Schinkel,J. J. Diamant, H. R A. Diesfeldt 8c
inwhich external pressure is high. An example is the R. Haaxma (Eds.), Neuropsychologie inNederland [Neu
This content downloaded from 130.237.29.138 on Sat, 27 Feb 2016 16:02:46 UTC
All use subject to JSTOR Terms and Conditions
Gudjonsson, G. H. (1997). The Gudjonsson Suggestibility (1998).The Gudjonsson Suggestibility
Scale (GSS): Fur
Scales manual. Hove, UK: Psychology Press. ther data on its reliability, validity, and metacognition cor
G. H. (2003). The psychology of interrogations relates. Social Behavior and
Gudjonsson, Personality, 26, 203-210.
and confessions: A handbook. New York: Wiley. Peters, M.J. V, Jelicic, M., & Merckelbach, H. (2008).
G. H., Hannesdottir, K., Petursson, H., 8c A Dutch
Gudjonsson, Inducing false memories: version of theDeese/
try, 13, 53-67 plaints in elderly people: The role of memory abilities,
Horselenberg, R., Merckelbach, H., & Josephs, S. (2003). memories: Remembering words not presented in lists.
Individual differences and false confessions: A concep ofExperimental
Journal Psychology: Learning, Memory,
tual replication of Kassin and Kiechel and
(1996). Psychology, Cognition, 21, 803-814.
Crime &Law, g, 1-8. Smeets, T. (2008). Gudjonsson Suggestibility Scales
en Gud
Horselenberg, R., Merckelbach, H., Smeets, T., Frans jonsson Compliance Scale [Gudjonsson Suggestibility
sens, D., Peters, G. J.Y, 8c Zeles, G.
(2006). False confes Scales and Gudjonsson Compliance Scale]. In T. Gies
sions in the lab: Do mat
plausibility and consequences brecht, C. de Ruiter 8c M. Jelicic
(Eds,), Forensisch psy
ter?Psychology, Crime &Law, 12, 61-75. en
chodiagnostisch gereedschap. Malingering, psychopathie
Kassin, S. M., 8c Kiechel, K. L. (1996). The social psychol andere persoonlijkheidstrekken [Forensic psychodiagnos
ogy of false confessions: Compliance, internalization, and tic tools. Malingering, psychopathy and other personality
confabulation. Psychological Science, j, 125-128. traits] (pp. 63-72). Amsterdam: Harcourt.
Leippe, M. R., Manion, A. P., 8c Romanczyk, A. (1992). Eyewit Smeets, T, Merckelbach, H., Horselenberg, R., 8c Jelicic, M.
ness
persuasion: How and how well do fact finders judge (2005). Trying to recollect past events: Confidence,
the accuracy of adults' and children's memory reports? beliefs, and memories. Clinical 25,
Psychology Review,
Journal ofPersonality and Social Psychology, 63,181-197. 917-934
Magnussen, S., Andersson,J., Cornoldi, C, De Beni, R., Squire, L. R., Wetzel, C, 8c Slater, P. C. (1979). Memory
G. S., et al. (2006). What complaint after electroconvulsive
Endestad, T, Goodman, people therapy: Assessment
believe about memory. Memory, with a new
14, 595-613. self-rating instrument. Biological Psychiatry,
Merckelbach, H., Muris, P., Nijman, H., 8c De Jong, P. J. 14, 791-801.
(1996). Self-reported cognitive failures and neurotic Van Bergen, S., Brands, I., Jelicic, M., 8c Merckelbach, H.
and Individual
symptomatology. Personality Differences, (2009). Assessing trait memory distrust: Psychometric
20, 715-724. properties of the Squire Subjective Memory Questionnaire.
Merckelbach, H., Muris, P.,Wessel, I., 8c Van Koppen, P. J. Manuscript submitted for publication.
This content downloaded from 130.237.29.138 on Sat, 27 Feb 2016 16:02:46 UTC
All use subject to JSTOR Terms and Conditions