All content in this area was uploaded by Thomas C Ormerod on Aug 03, 2018
Content may be subject to copyright.
All content in this area was uploaded by Thomas C Ormerod on Aug 03, 2018
Content may be subject to copyright.
RUNNING HEAD: Convergent thinking and false memory
Convergent, but not divergent, thinking predicts susceptibility to associative memory illusions
Stephen A. Dewhurst1, Craig Thorley2, Emily R. Hammond2, and Thomas C. Ormerod3
1Department of Psychology, University of Hull
2Department of Social and Psychological Sciences, Edge Hill University
3Department of Psychology, Lancaster University
Word count = 3587
Address for correspondence:
Stephen A. Dewhurst
Department of Psychology
University of Hull
Hull HU6 7RX
Phone +44 1482 465931
Fax +44 1482 465599
The relationship between creativity and susceptibility to associative memory illusions in the
Deese/Roediger-McDermott procedure was investigated using a multiple regression analysis.
Susceptibility to false recognition was significantly predicted by performance on a measure of
convergent thinking (the Remote Associates Task) but not by performance on a measure of
divergent thinking (the Alternative Uses Task). These findings suggest that the ability to
engage in convergent (but not divergent) thinking underlies some of the individual variation in
susceptibility to associative memory illusions by influencing the automaticity with which
critical lures are activated at encoding.
Key words: False memory; individual differences; creativity
Roediger and McDermott (1995) showed that illusions of memory can be created when
participants study lists of associated words. In the Deese/Reodiger-McDermott (DRM)
procedure (Deese, 1959, Roediger & McDermott, 1995), participants study lists of associates
of a nonpresented “critical lure”. For example, participants study words such as sour, candy,
and sugar, which are associates of the critical lure sweet. When memory for the lists is tested,
participants frequently claim to remember the critical lures, with levels of false memory
equaling or even exceeding levels of correct memory. The DRM illusion has been explained in
terms of an activation-monitoring account (Roediger, Watson, McDermott, & Gallo, 2001)
whereby participants spontaneously generate associates of the studied words. The critical lures
are then subject to errors of source monitoring (Johnson, Hashtroudi, & Lindsay, 1993) and
falsely endorsed as having been studied. An alternative explanation is provided by fuzzy-trace
theory (FTT, see Reyna & Brainerd, 1998) whereby critical lures are falsely remembered
because they match the “gist” of the related items presented at study.
Although Roediger and McDermott’s findings have been replicated many times (see
Gallo, 2006, for a review), one phenomenon that has yet to be explained is the considerable
individual variation in susceptibility to the DRM illusion. Elevated levels of false memory
have been reported in elderly adults (Balota et al., 1999) and patients with frontal lobe damage
(Melo, Winocur, & Moscovitch, 1999), while reduced levels of false memory have been
observed in children (e.g., Brainerd, Reyna, & Forrest, 2002). Other studies have attempted to
identify the causes of individual variation within the general adult population. For example,
elevated levels of false memory have been observed in individuals who reported high levels of
dissociative experiences and vivid imagery (Winograd, Peluso, & Glover, 1998), individuals
with low working memory capacity (Watson, Bunting, Poole, & Conway, 2005), individuals
high in need-for-cognition (Graham, 2007), and extroverts (Sanford & Fisk, 2009).
Given the extensive use of the DRM procedure in the study of false memories, it is
important to identify other cognitive and personality factors that influence susceptibility to the
illusion. The aim of the current research was to investigate whether susceptibility to the DRM
illusion is predicted by creativity. A number of previous studies have shown that creative
individuals are particularly susceptible to false autobiographical memories. For example,
Hyman and Billings (1998) found that creativity (as measured by the Creative Imagination
Scale) was positively related to the creation of false childhood memories. However, to the best
of our knowledge, no studies have as yet investigated the influence of creativity on
susceptibility to the DRM illusion.
Although creativity is a complex mental faculty that encompasses a variety of cognitive
abilities (see Dietrich, 2004, for a review), a number of measurable components have been
identified. It is possible that some, but not all, aspects of creativity may predict susceptibility
to the DRM illusion. The aspects of creativity that were the focus of the current study are
commonly referred to as convergent and divergent thinking (Guilford & Hoepfner, 1971).
Convergent thinking requires the production of the best single answer to a problem or set of
problems and can be measured by the Remote Associates Task (RAT; Mednick, 1962). In the
RAT, participants are presented with three words (e.g., food/forward/break) and asked to
generate a semantic associate that can be paired with each of the three to form a compound
word or phrase (e.g., fast). Divergent thinking requires the generation of multiple answers to a
single problem and can be measured by the Alternative Uses Task (AUT; Guilford, 1967) in
which participants are asked to generate alternative uses for a common object (e.g., a brick).
Mednick (1962) developed the RAT on the basis of his theory that creative individuals
generate more and broader associations to a given stimulus. Our hypothesis, therefore, was
that the false recognition of critical lures would be predicted by performance on the RAT, as
both involve the generation of semantic associations. In contrast, the AUT measures the ability
to generate novel or atypical ideas, which has less overlap with the processes that underlie the
DRM illusion; therefore we did not expect the false recognition of critical lures to be predicted
by performance on the AUT.
Participants were 55 undergraduate students (41 females) who took part for course
credit. Mean age was 21 years (SD=5.29). They were tested at individual workstations in
groups of up to 12 and participated for course credit. The research was carried out in
accordance with The Code of Ethics of the World Medical Association (Declaration of
Helsinki) for experiments involving humans.
2.2. Stimuli and design
Study items consisted of 16 DRM lists rated by Stadler, Roediger, and McDermott
(1999) as producing high levels of false recognition. Each list comprised 12 associates of a
nonpresented critical lure. The lists were divided into two sets of 8. Each set was studied by
half the participants and the other set provided the distractor items for the recognition test. The
recognition test consisted of a printed sheet containing 8 studied words (one from each list), the
8 critical lures of the studied lists, plus 8 list items and the 8 critical lures from the unstudied
lists. The stimuli for the RAT consisted of 24 three-item problems taken from Bowden and
Jung-Beeman (2003) presented on a two-sided response sheet with two columns of six items
on each side. The items in each problem were presented one above the other with a line to the
right for participants to record their responses. All participants saw the same stimuli in the
same order. The AUT (Guildford, 1967) required participants to list alternative uses for a
The DRM lists were presented one at a time on PCs at a rate of 2 seconds per word with
a 1 second interval. Each list was preceded by the list number (List 1, List 2, etc) displayed for
2 seconds. After the presentation of the final list, participants were given a letter cancelation
task for 1 minute. They were then given the recognition test, which they completed at their
own pace. Participants were then allowed 8 minutes to complete the AUT, followed by a
further 8 minutes to complete the RAT (these times were based on the results of pilot studies).
As an example of the RAT, participants were shown that the word pin could be paired with
safety, cushion, and point to make safety pin, pincushion, and pinpoint.
Multiple regression was used to assess the ability of convergent and divergent thinking
to predict critical lure, studied word, and distractor item recognition rates. Preliminary analyses
were conducted to ensure no violation of the assumptions of normality, linearity,
muticollinearity, and homoscedasticity. The sample size (n = 55) was also sufficient for this
procedure according to the guidelines of Stevens (2009).
The AUT responses were rated for creativity on a scale of 0-4. Impossible uses (e.g., a
time machine) were given a score of 0, standard uses (e.g., to build a wall) were given a score
of 1 (with no additional scores for repetition of uses), and alternative uses were given scores of
2, 3, or 4 depending on the rated creativity. Initial ratings were made by the third author, and
20% were blind double-rated by the second author. The initial inter-rater reliability score was
92%, with all disagreements resolved through discussion.
3.1. Critical Lure Results: Hierarchical Multiple Regression
A 61% false recognition rate for critical lures was observed (M = 4.90, SD = 1.72),
indicating that the DRM effect was successfully replicated. It was expected that convergent
thinking (M = 7.14, SD = 2.43) would be a significant predictor of critical lure recognition
whereas divergent thinking (M = 26.91, SD = 11.77) would not. Given these strong
predictions, a Hierarchical Multiple Regression was conducted with the convergent thinking
scores entered at Step 1 and the divergent thinking scores at Step 2 (see Table 1).
The initial correlations revealed a significant relationship between convergent thinking
and critical lure recognition (r = .33, p<.01), but no significant relationship between divergent
thinking and critical lure recognition (r = .15, p = .13), or between convergent thinking and
divergent thinking (r = .02, p = .44). The regression analysis revealed that convergent thinking
accounted for 11% (R2 = .11) of the variance in false recognition F(1, 53) = 6.29, p<.05. The
addition of the divergent thinking scores in Step 2 resulted in a non-significant 2% increase in
the explained variance, ∆F (1, 53) = 1.25, p = .27. Convergent thinking therefore appears to be
a significant predictor of critical lure false recognition (β = .32, p<.05), whereas divergent
thinking does not (β = .14, p = .27).
3.2. Distractor Items: Simultaneous Multiple Regression
As no relationship was expected between either convergent or divergent thinking and
the false recognition of distractor items (M = 2.16, SD = 2.15), a Simultaneous Multiple
Regression was used for this second analysis (see Table 3). Initial correlations revealed a
significant relationship between convergent thinking and distractor item recognition (r = .40,
p<.01), but no significant relationship between divergent thinking and distractor item
recognition (r = -.15, p = .14). The regression analysis revealed that convergent and divergent
thinking together accounted for 18% (R2 = .18) of the variance in distractor item recognition,
F(2, 52) = 5.91, p<.01. However, distracter item recognition was significantly predicted only
by convergent thinking (β = .40, p<.01) and not by divergent thinking (β = -.15, p =.22).
3.3. Convergent Thinking: Simultaneous Multiple Regression for Critical Lure and Distractor
The above results demonstrate that convergent thinking predicts both critical lure false
recognition and distractor item recognition. A Simultaneous Multiple Regression was
therefore used to determine whether critical lure false recognition or distractor item false
recognition is the strongest predictor of convergent thinking (see Table 2). Initial correlations
revealed no significant relationship between critical lure false recognition and distractor item
recognition (r = .10, p<.23). The two predictors together accounted for 24% (R2 = .24) of the
variance in convergent thinking, F(2, 52) = 8.34, p = <.01. Both critical lure false recognition
(β = .29, p<.05) and distractor item recognition (β = .37, p<.01) were significant predictors of
convergent thinking, with β values indicating that distractor item recognition is a slightly
3.4. Studied Words: Simultaneous Multiple Regression
Neither convergent nor divergent thinking were expected to predict studied word
recognition (M = 5.85, SD = 1.58), therefore a Simultaneous Multiple Regression was used to
assess this (see Table 3). Initial correlations revealed no significant relationship between
convergent thinking and studied word recognition (r = .06, p<.32), or divergent thinking and
studied word recognition (r = -.02, p = .44). The regression analysis revealed that the two
predictors together accounted for less than 1% (R2 = .004) of the variance in studied word
recognition, F(2, 52) = .12, p = .89, confirming that neither convergent (β = .06, p = .65) nor
divergent thinking (β = -.02, p = .88) were significant predictors of correct recognition.
The main finding from the current study was that susceptibility to the DRM illusion
was significantly predicted by convergent thinking (as measured by the RAT) but not by
divergent thinking (as measured by the AUT). This pattern is consistent with the activation-
monitoring account of the DRM procedure (Roediger et al., 2001), as both the RAT and the
DRM illusion rely on the activation of semantic associates. In contrast, the AUT requires
participants to generate novel uses for a common object, which has less in common with the
processes underlying the DRM procedure. In terms of fuzzy-trace theory (Reyna & Brainerd,
1998), it is possible that participants’ ability to connect the gist of semantically related words
underlies performance on the RAT, which relies on semantic connections, but not on the AUT,
which does not rely on such processes.
The current findings are consistent with the proposal by Mednick (1962) that creative
individuals generate more and broader associations to a given stimulus. Recent findings by
Rossmann and Fink (2010) also support this view by showing that the rated associative
distance between unrelated words was lower for creative individuals than for individuals rated
as less creative. According to Howe, Wimmer, Gagnon, and Plumpton (2009), it is the
automaticity with which critical lures are activated that determines the likelihood that they will
be falsely remembered. It is possible that the shorter associative pathways used by creative
individuals increase the automaticity with which the critical lures are generated in the DRM
It is, perhaps, surprising that the ability to engage in convergent thinking (considered a
positive trait in terms of creativity) has the negative consequence of increasing susceptibility to
false memory. However, previous research has shown that other ostensibly positive traits can
increase susceptibility to false memory. For example, Castel, McCabe, Roediger, and Heitman
(2007) reported that expertise in a given domain has what they termed a “dark side”, whereby
experts were more prone than novices to domain-relevant intrusions. Castel et al. suggested
that experts’ superior organizational processes, which usually enhance memory for domain-
relevant information, also support the associations that give rise to memory illusions. More
broadly, the DRM illusion itself is a negative corollary of spreading activation processes that
support normal memory functions (see Roediger & McDermott, 1995). The increased
susceptibility to false memory as a function of convergent thinking can also be seen as the
“dark side” of an otherwise adaptive process.
An unexpected finding from the current study was that convergent thinking
significantly predicted levels of false recognition of the unrelated distracters. Although we did
not anticipate such an effect, it is consistent with Mednick’s (1962) proposal that creative
individuals make broader associations to a given stimulus. It is possible that participants who
scored high on the RAT in the current study generated associations that went beyond the
themes of the DRM lists, leading to the partial priming of words not directly associated with
the list items. One way to test the influence of the breadth of associative pathways would be to
manipulate the backwards associative strength of the critical lures. If increased susceptibility
to the DRM illusion is the result of creative people using broader associative pathways, then it
is likely that participants who score high on convergent thinking will falsely recall lures of low
BAS to a greater degree than participants who score low on convergent thinking.
Although both critical lure false recognition and distractor false recognition were
significant predictors of performance on the RAT, comparison of the β values in Table 2
suggests that distractor items were a slightly stronger predictor than critical lures. Although
this finding was unexpected, one possible explanation is that the greater predictive power of
distractors relative to critical lures reflects the differential use of a recall-to-reject strategy.
Previous research has shown that participants can use a recall-to-reject strategy to avoid false
recognition in the DRM procedure, whereby they reject critical lures because they can recall
the associated items that were presented at study (e.g., Gallo, 2004). It is likely that the
effectiveness of such a strategy will depend on the strength of association between the critical
lure and the associated study item. A recall-to-reject strategy will, therefore, be less effective
at reducing the false recognition of unrelated distracters. We acknowledge, however, that this
is a post hoc explanation and that the effect observed with the unrelated distracters warrants
To summarize, the current study found that levels of false recognition in the DRM
procedure were significantly predicted by performance on a test of convergent thinking but not
by performance on a test of divergent thinking. These findings suggest that individuals who
perform well on tests of creativity may be at increased risk of false memories, at least the false
memories produced by the DRM procedure. This effect is, however, highly specific and
depends on the overlap between the processes that underlie the creativity test and those that
give rise to the DRM illusion. Although the current findings do not arbitrate between the two
main theories of the DRM illusion (activation-monitoring and fuzzy-trace theory), they suggest
one possible source of the individual variation in susceptibility to the DRM illusion.
Balota, D.A., Cortese, M.J., Duchek, J.M., Adams, D., Roediger, H.L. III, McDermott, K.B.,
& Yerys, B.E. (1999). Veridical and false memories in healthy older adults and in
dementia of the Alzheimer’s type. Cognitive Neuropsychology, 16, 361-384.
Bowden, E.M., & Jung-Beeman, M. (2003). Normative Data for 144 Compound Remote
Associate Problems. Behavior Research Methods, Instruments, & Computers, 35, 634-
Brainerd, C.J., Reyna, V.F., & Forrest, T.J. (2002). Are young children susceptible to the false-
memory illusion? Child Development, 73, 1363-1377.
Castel, A.D., McCabe, D.P., Roediger, H.L., & Heitman, J.L. (2007). The dark side of
expertise: Domain-specific memory errors. Psychological Science, 18, 3-5.
Deese, J. (1959). On the prediction of occurrence of particular verbal intrusions in immediate
recall. Journal of Experimental Psychology, 58, 17-22.
Dietrich, A. (2004). The cognitive neuroscience of creativity. Psychonomic Bulletin & Review,
Gallo, D.A. (2004). Using recall to reduce false recognition: Diagnostic and disqualifying
monitoring. Journal of Experimental Psychology: Learning, Memory, and Cognition,
Gallo, D.A. (2006). Associative illusions of memory. New York: Psychology Press.
Graham, L.M. (2007). Need for cognition and false memory in the Deese-Roediger-
McDermott paradigm. Personality and Individual Differences, 42, 409-418.
Guilford, J. P. (1967). The nature of human intelligence. New York: McGraw-Hill.
Guilford, J.P., & Hoepfner, R. (1971). The analysis of intelligence. New York: McGraw-Hill.
Howe, M. L., Wimmer, M. C., Gagnon, N., & Plumpton, S. (2009). An associative-activation
theory of children's and adults' memory illusions. Journal of Memory and Language,
Hyman, I.E., & Billings, F.J. (1998). Individual differences and the creation of false childhood
memories. Memory, 6, 1-20.
Johnson, M.K., Hashtroudi, S., & Lindsay, D.S. (1993). Source monitoring. Psychological
Bulletin, 114, 3-28.
Mednick, S.A. (1962). The associative basis of the creative process. Psychological Review, 69,
Melo, B., Winocur, G., and Moscovitch, M. (1999). False recall and false recognition: An
examination of the effects of selective and combined lesions to the medial temporal
lobe/diencephalon and frontal lobe structures. Cognitive Neuropsychology, 16, 343-
Reyna, V.F., & Brainerd, C.J. (1998). Fuzzy-trace theory and false memory: New frontiers.
Journal of Experimental Child Psychology, 71, 194-209.
Roediger, H.L., III, & McDermott, K.B. (1995). Creating false memories: Remembering
words not presented in lists. Journal of Experimental Psychology: Learning, Memory,
and Cognition, 21, 803-814.
Roediger, H. L., III, Watson, J. M., McDermott, K. B., & Gallo, D. A. (2001). Factors that
determine false recall: a multiple regression analysis. Psychonomic Bulletin and
Review, 8, 385-407.
Rossmann, E., & Fink, A. (2010). Do creative people use shorter associative pathways?
Personality & Individual Differences, 49, 891-895.
Sanford, L.C., & Fisk, J.E. (2009). How does the extraversion personality trait influence false
recall with the Deese-Roediger-McDermott paradigm? Journal of Research in
Personality, 43, 972-977.
Stadler, M.A., Roediger, H.L., & McDermott, K.B. (1999). Norms for word lists that create
false memories. Memory & Cognition, 27, 494-500.
Stevens, J. (2009). Applied Multivariate Statistics for the Social Sciences (5th Ed.). New York:
Watson, J.M., Bunting, M.F., Poole. B.J., Conway, A.R.A. (2005). Individual differences in
susceptibility to false memory in the Deese–Roediger–McDermott paradigm. Journal
of Experimental Psychology: Learning, Memory, & Cognition, 31, 75-85.
Winograd, E., Peluso, J. P., & Glover, T. A. (1998). Individual differences in susceptibility to
memory illusions. Applied Cognitive Psychology, 12, S5–S27.
Table 1: Summary of the Hierarchical Multiple Regression Analysis for Convergent and
Divergent Thinking in Relation to Critical Lure False Recognition
CT = Convergent Thinking, DT = Divergent Thinking, * = p<.05
Table 2: Summary of the Simultaneous Multiple Regression Analysis for Critical Lures and
Distractor Items in Relation to Convergent Thinking
CL = Critical Lures, DI = Distractor Items, * = p<.05, ** = p<.01
Table 3: Summary of the Simultaneous Multiple Regression Analysis for Convergent and
Divergent Thinking in Relation to Studied Word and Non-Studied Distractor Item Recognition
CT = Convergent Thinking, DT = Divergent Thinking, * = p<.01