Conference PaperPDF Available

The effects of alcohol on false memories



Content may be subject to copyright.
The effect of alcohol and repetition at encoding on implicit
and explicit false memories
S. N. Garfinkel &Z. Dienes &T. Duka
Received: 22 February 2006 /Accepted: 14 June 2006 / Published online: 11 August 2006
#Springer-Verlag 2006
Rationale Alcohol impairs explicit memory, whilst leaving
implicit memory relatively intact. Less is known about its
effects on false memories.
Aim The present study examines the effects of alcohol on
explicit and implicit false memories using study list
repetition as a tool for modulating learning at encoding.
Methods Thirty-two participants were given either an
alcohol (0.6 g/kg) or placebo beverage before undergoing
an encoding phase consisting of 10 lists of nine associated
words (veridical items). Each list was associated to a word,
which was not presented at encoding (semantically associ-
ated non-studied lure; critical item), serving as the measure
for false memory. Half of the lists were presented once, and
half were repeated three times. The next day, participants
underwent an implicit (stem completion and post hoc
awareness measurements), and an explicit (free recall) task.
Results Alcohol decreased veridical and false explicit
memory for singularly presented lists compared to placebo;
no group difference existed for repeated lists. Implicit
veridical memory was not affected by alcohol. Awareness
memory measures demonstrated in placebo participants an
increased ability with repetition in rejecting false memories.
The reverse was found in intoxicated participants who with
repetition accepted more false memories.
Conclusion Alcohol appears to decrease semantic activa-
tion leading to a decline in false memories. Increased
learning with repetition, which increases the rejection of
false memories under placebo, is reversed under alcohol
leading to a decrease in rejection of false memories. The
latter effect of alcohol may be due to its ability to impair
monitoring processes established at encoding.
Keywords Learning .Free recall .Stem completion .
DRM paradigm .Awareness
False memories constitute a process involving either
remembering events that never happened, or remembering
them quite differently from the way they happened
(Roediger and McDermott 1995). Deeses(1959) memory
paradigm revived and restructured by Roediger and
McDermott (1995) has provided the basis for much recent
false memory research (DRM paradigm). Its popularity can
in part be attributed to the robustness with which it
generates false memories, in both recall (e.g. Deese 1959;
Roediger and McDermott 1995) and recognition (e.g.
Roediger and McDermott 1995). The DRM paradigm is
composed of lists of associated (veridical) words (e.g. hill,
valley, climb, summit, etc.), with each list related to a
semantically associated non-studied lure(e.g. mountain)
termed the criticalitem. The probability that this non-
presented critical item will subsequently be erroneously
recalled or recognised as having been previously presented
serves as the measure for false memory.
To date, only two published studies have investigated the
effect of alcohol on false memories using the DRM
paradigm (Milani and Curran 2000; Mintzer and Griffiths
2001), with contrasting findings. Milani and Curran (2000)
did not detect an effect of alcohol on false recall rates,
though a tendency for alcohol to increase false recognition
Psychopharmacology (2006) 188:498508
DOI 10.1007/s00213-006-0480-6
S. N. Garfinkel :Z. Dienes :T. Duka (*)
Department of Psychology School of Life Sciences,
University of Sussex, Falmer,
Brighton BN1 9QG, UK
was obtained. In contrast, Mintzer and Griffiths (2001) did
not find an increase in false recognition rates under alcohol
relative to a placebo. These equivocal findings warrant
further investigation into the effect of alcohol on false
memories. Research into the effect of alcohol on false
memories is not only interesting in its own right, but can
also provide an insight into the cognitive processes
underlying memory judgements. This follows a tradition
where researchers have used pharmacological agents to
provide insight into underlying mechanisms (Duka et al.
1996). To date, the effect of alcohol on false memories has
not been accounted for using contemporary false memory
The Activation Monitoring Framework (AMF) was
proposed by Roediger and colleagues (Gallo and Roediger
2002; Roediger et al. 2001). It postulates dual opponent
processes, which determine false memory levels: an
activation component and a monitoring component. False
memories are thought to be elicited via their activation, and
false remembering is thought to subsequently occur when
monitoring fails and individuals misattribute the source of
this activation to prior presentation.
As the AMF is formed of opponent processes, it can
accommodate both an increase and decrease in false
memory levels dependent upon the extent to which
factors differentially affect activation and monitoring
potential. Fundamentally, one can predict that any
process which selectively decreases activation should
lead to a decrease in false memory, whilst any manipu-
lation that decreases monitoring should cause an increase
in false memory. A primary prediction is that alcohol will
decrease false memories because it impairs semantic
processing (Craik 1977;Birnbaumetal.1980; Weissenborn
and Duka 2000), decreasing activation of false memory
The amnesic effects of alcohol consumed before encod-
ing are well documented, with studies consistently demon-
strating an impairment of free recall (e.g. Birnbaum et al.
1978) and cued recall (Duka et al. 2001) relative to a
placebo control. In 1997, Craik (1977) published an article
which accounted for the deleterious effect of alcohol on
memory performance using the levels of processing (LOP)
framework. According to his argument, alcohol reduces the
depth at which items are processed, with depth being
defined as a continuum of processing running from
shallow sensory analyses requiring little attention to deeper
semantic processes through which the stimulus is identified,
interpreted and enriched by associations with stored
knowledge. Alcohol has been shown to reduce attentional
and processing resources (Schweizer et al. 2005), which are
deemed necessary for deep encoding (Craik and Lockhart
1972). Because studies which have reduced depth of
processing found a decrease in false memory levels
(Rhodes and Anastasi 2000; Thapar and McDermott
2001;Togliaetal.1999), superficial encoding when
intoxicated would reduce false memories.
Parker et al. (1974) hypothesised that the more
demanding the task, the greater the impairment from
alcohol. Johnson (1977) qualified this claim by defining
demandingas being the extent to which a task neces-
sitates the finding or generating [of] associations, inter-
relationships or structures. Parker et al. (1976)foundthat
a high dose of alcohol (1.0 ml/kg [0.8 g/kg]), impaired
paired associative learning. Alcohol was also found
to reduce selectively memory for high associations
(Weissenborn and Duka 2000) and Birnbaum et al.
(1980) have shown an impairment in the semantic
processing of incoming stimuli. The impairment of such
processes has implications for the forming of conscious
associations between veridical items, a process, which is
thought to be a potential route to increasing false memories
(Libby and Neisser 2001;Roedigeretal.2001). Repetition
as a means of increasing learning has been applied with the
DRM paradigm as an experimental manipulation to
investigate the mechanisms underlying false memories
(Benjamin 2001). Repetition at encoding has been shown
to increase false memories (Benjamin 2001), decrease false
memories (Tussing and Greene 1997), have no effect on
false memories (Tussing and Greene 1999) and affect false
memories in an inverted U-shaped function (Seamon et al.
2002). As repetition may both increase false memories via
increased activation of semantically associated non-studied
items, and decrease false memories through increased
monitoring, observing the effect of repetition on false
memory levels can provide insight into whether activation
or monitoring processes are prevailing in memory judge-
ments. We, thus, used in the present study repetition as a
tool to gain insight into the effect of alcohol on activation
and monitoring processes relevant for false memory
Implicit and explicit measures of memory will be
takenthis is because alcohol has been shown to
differentially affect these memory measures, by impairing
explicit memory, such as free recall, but leaving implicit
memory relatively intact (Duka et al. 2001). Priming
measures will be taken for both the veridical (studied
items) and critical (semantically associated non-studied)
items based on previous methodologies used in the
alcohol memory literature (e.g. Duka et al. 2001;
Weissenborn and Duka 2000). Testing will take place
next day without the influence of alcohol so that the
effects of alcohol on encoding will be tested indepen-
dently from its effects on retrieval. Previous studies
testing the effects of alcohol on false memories did not
fully separate the encoding and retrieval phases of the
experiment in terms of alcohol intoxication (Curran and
Psychopharmacology (2006) 188:498508 499
Hildebrandt 1999; Mintzer and Griffiths 2001). Testing
after a 24-h interval will, in addition, provide information
about long-term priming for false memories because no
study using implicit measures has previously attempted a
24-h retention interval. Finding ways in which veridical
memory operates in the same way and differs from false
memory provides insight into mechanisms underlying
these two types of memory.
Measurements of drinking habits and history were
evaluated to make sure that groups did not differ on
aspects, which might influence alcohol effects.
Materials and methods
Thirty-two volunteers (16 men and 16 women) were
recruited from the undergraduate and postgraduate popu-
lation at the University of Sussex. They received either
cash or course credits as payment. All participants were
native English speakers, not dyslexic and aged 18
34 years. Participants were randomly assigned to either
an alcohol or a placebo condition, with the constraint that
groups were matched for gender (eight men and eight
women). The groups had very similar mean age (mean age
21.30 years in the alcohol group and 20.96 years in the
placebo group).
The ethics committee of the University of Sussex
approved the experiment. After giving informed consent,
participants were screened on the basis of their medical
history. Exclusion criteria included: a history of severe
mental illness, a history of drug or alcohol misuse, an
altered metabolism of alcohol (as determined by impaired
liver function or gastroenteritis), anyone displaying current
symptoms of mental illness or neurological disease.
Participants were asked to abstain from illicit recreational
drugs for a minimum of 7 days, from sleeping tablets or hay
fever medication for 48 h, and from drinking tea or coffee
immediately before the commencement of the experiment
and were required to have a low fat breakfast the morning
of the experiment. In addition, participants were told to
abstain from drinking for 24 h before the beginning of the
experimental session, and were breathalysed on entering the
lab as a means to determine whether they had complied
with this requirement.
The experiment was double blind and took place over two
consecutive days. The drink was consumed on day 1, and
was followed by the learning phase. Participants returned
the next day to undergo two memory tests: an implicit test
and free recall.
Memory task
Eleven 9-word lists were taken from Stadler et al. (1999)
when lists conformed to the specifications made by
McEvoy et al. (1999) deemed necessary for stem comple-
tion tests. To investigate false memories under implicit stem
completion instructions, critical items had to adhere to the
specifications outlined by Graf and Schacter (1985), which
had been followed by other researchers using the DRM
paradigm (McKone and Murphy 2000). These were: critical
items were a minimum of five letters long, all 3-letter
stems had a minimum of eight different word completions,
stems were required to not form words in themselves,
all stems used were distinct and not the same as the opening
three letters of any veridical words and that baseline stem
completion rates were not higher than a 50% probability of
completion, as a means to avoid ceiling effects on priming
scores (McKone and Murphy 2000).
Only 11 lists with the norms of Stadler et al. (1999)
fulfilled the above criteria. Nine new lists were created with
the aid of the Edinburgh Associative Thesaurus (EAT:
association norms, displaying the counts of word associa-
tion elicited in association to target words, as collected
empirically from participants following the methodology
employed by Kiss et al. (1973). Following the procedure
used by Roediger and McDermott (1995), lists were formed
of the top veridical associates of nine new critical words. In
a pilot study, the new lists were evaluated and analyses
determined that the newly created DRM lists did not
significantly differ from those used by Stadler et al.
(1999) in terms of backwards associative strength from
the veridical items to the critical items [t(9)=1.51,
p=0.166], baseline stem completion rates for critical items
[t(9)=0.39, p=0.708] and probability that lists would elicit a
false memory [t(15)=1.19, p=0.252].
The 20 9-word lists were separated into two sets
consisting of 10 lists each (set A and set B). Each
participant viewed either set A or set B in the encoding
phase. Each set was also divided into two sections, section
1 and section 2 (resulting in five lists in each section). The
difference between the two sections was the frequency of
presentation (once or three times). Lists in each section
were matched for word frequency (Kucera and Francis
1967) and stem completion rate. List repetition was
counterbalanced across sections. The presentation order of
lists was fully randomised, with the constraint that
repetitions were consecutive.
500 Psychopharmacology (2006) 188:498508
The implicit test consisted of 80 stems: two critical items
and two veridical items from each list. Twenty studied
veridical and 20 non-studied critical items corresponded to
lists viewed in encoding. Ten of these veridical items came
from lists presented once, whilst the remaining 10 were
from repeated lists. Similarly, 10 critical items were from
repeated lists, whilst 10 were from singularly presented
lists. The remaining 40 stems were from the set not viewed
in the encoding session, and served as baseline measures.
These baseline stems were comprised of 20 veridical items
not viewed in the encoding session, and their associated 20
critical items.
After the implicit test, a free recall test was introduced.
Participants were given instructions to write down all the
words they remembered from the previous days encoding
Alcohol administration
Alcohol was administered at the dose of 0.6 g/kg, with 90%
v/walcohol diluted with tonic water to make up a 300-ml
beverage. The drink was divided into 10×30 ml portions
and each portion was mixed with four drops of Angostura®
bitters. The drink was consumed at a rate of one portion
every 3 min over a 30-min period in the presence of the
experimenter. The placebo beverage consisted of the
respective volume of Angostura® bitters and tonic water
Alcohol use questionnaire
Participantsweekly alcohol consumption was assessed
using the alcohol use questionnaire (AUQ) (Mehrabian and
Russell 1978). The AUQ consists of 12 questions designed
to evaluate consumption of beer/cider, wine and spirits in
terms of frequency and quantity. Due to the recent changes
in drinking trends, an updated version was used which
incorporated questions about the consumption of Alco pops
(Knowles and Duka 2004). The dependent variables
produced by the AUQ are: 8 g UK units of alcohol units
drunk each week and an overall AUQ score which
represents speed and frequency of drinking and intoxica-
tion. In addition, a separate binge drinking score based just
on speed of drinking and frequency of intoxication
developed by Townshend and Duka (2002) was also
determined. The questionnaire was administered to ensure
that participants met the specifications for participating
that they were social drinkers who consumed on average
between 5 to 50 U a week. In addition, group analyses
could ensure that the alcohol and placebo participants were
matched in terms of units consumed per week, AUQ score
and binge drinking score.
Visual analogue scales
To assess mood, both in terms of baseline measures and the
acute effects of alcohol on mood, participants were asked to
complete a set of 100 mm visual analogue scales (VASs)
(Duka et al. 1998). These provided measures of how
participants were feeling at that particular moment. Depen-
dent variables were 0100 scores for the following
adjectives: contented, lightheaded and relaxed.
Assessment of baseline memory
Baseline memory was assessed to ensure that no differences
existed between alcohol and placebo groups. Fifteen-
member word listslists A and B from the Rey Audito-
ryVerbal Learning Test Word Lists which were validated
for frequency and complexity by Lezak (1983)were
administered before drink consumption. The dependent
variable was the mean number of words recalled from the
two lists (5 min after presentation).
Blood alcohol concentrations
A standard breathalyser with a detection limit equivalent
to 0.01 g/l of alcohol in the bloodstream was used to
measure blood alcohol concentrations (BAC) levels (Alcol-
meter S-D3M, Loborservice GmbH, Bonn, Germany).
The dependent variable was BAC measured in gram per
Participants were tested individually. Upon entering the lab,
participants signed a consent form and read a brief
description about the experiment, which stated that the
effect of alcohol on learning was to be tested. They were
not informed that the experiment was concerned with false
memory. Participants underwent a medical interview and
filled out a medical questionnaire to ensure they were
medically fit to take part, and the AUQ to ensure they were
moderate social drinkers.
All participants were breathalysed to ensure their
baseline breath alcohol concentrations (BACs) were 0.
Participants height and weight were taken, they con-
sumed a standardised lunch (a small roll and a glass of
water) and completed the AUQ and a verbal memory task
(Lezak 1983), as a means to obtain baseline memory
scores. They were then taken to a medical room to
consume their drinks over a 30-min period. Their BAC
Psychopharmacology (2006) 188:498508 501
level was measured 10 min after the final drink was
provided (40 min after the initiation of drinking) when
participants completed also a series of visual analogue
scales, which took approximately 5 min. Participants were
then taken to the experimental rooms where they under-
went the encoding phase. The encoding phase lasted for
15 min.
Upon completion of the encoding phase, participants
were sent to sit in the waiting room and were
subsequently breathalysed at 30-min intervals. Once
consent that they would not drink, ride a bike or operate
any kind of machinery for 4 h and were released from
the laboratory.
On day 2, participants again completed the VAS and
then completed the implicit test following the instruc-
tions. Instructions stated that the stems were to be
completed with the first word that came to mind.
Testing of free recall followed the stem completion.
Participants were asked to generate as many words as
possible from the ones, which they viewed the day
before. No time limit was enforced. Participants then
returned to the words they had generated under implicit
test instructions; they were instructed to circle all the
words they were awareof having viewed in the
encoding phase. The words that participants had
completed to form words viewed at the encoding
phase, which were not circled were classified as
Statistical analyses
Following a mode of analysis employed by previous
researchers using pharmacological manipulations in con-
junction with the DRM paradigm (Huron et al. 2001), all
analyses were performed separately in veridical and critical
items. This allowed an assessment of how repetition and
alcohol selectively modulated false and veridical memory
levels. All analyses were performed on percentages of
words recalled of the words presented at encoding (free
recall) or percentages of words completed of the words to
be completed (implicit task). The main analysis is divided
into three parts. Firstly, the free recall data (number of
words recalled) is addressed. The data generated under
implicit instructions (number words completed) are ana-
lysed with regard to priming (number of words completed
studied vs non-studied), followed by an analysis which is
concerned with items completed and labelled unaware
(rejected as not presented previously at encoding) vs
aware(erroneously accepted as being presented previous-
ly). Repetition is also included as a within subject factor in
each analysis. Details of individual ANOVAs preformed
precede each analysis.
Group characteristics
Analyses of the alcohol and placebo groups revealed that
participants were matched for units consumed per week
[t(30)=1.60, p=0.120], AUQ score [t(30)=0.26, p=0.799],
binge drinking score [t(30)=0.46, p=0.625] and baseline
memory measures (t(30)=0.29, p=0.772) (see Table 1).
VAS self-ratings
The alcohol and placebo groups did not differ on self-
ratings of relaxedness on either day 1 [t(30)=0.88,
p=0.383] or day 2 [t(30)=0.83, p=0.412]. Nor did they
differ on ratings of contentedness on either day 1 [t(30)=
1.12, p=0.261] or day 2 [t(30)=0.068, p=0.946]. A
significant effect of alcohol was displayed in subjective
ratings of lightheadedness. On day 1, post consumption of
the drink, the alcohol group rated themselves as significantly
more lightheaded than the placebo group [t(30)=5.39,
p<0.001]. On day 2, when no alcohol was administered, no
significant difference existed in lightheadedness ratings
between the two groups [t(30)=0.00, p=1.000] (see Table 2).
Free recall
To determine whether it was necessary to correct memory
scores for the recall of intrusions (recall of non-presented
items, which were not critical items), between subjects ttests
were administered to determine the extent to which intrusion
rates differed between the alcohol and placebo groups. The
amount of non-critical intrusions recalled between the two
groups approached significance [t(30)=1.86, p=0.072], the
mean value of non-critical intrusions recalled was higher in
the placebo group (4.81) than the alcohol group (2.81). As
placebo participants recalled more veridical and critical items
than the alcohol group (see analyses below), two new
variables were formed which looked at the ratio of non-
critical intrusions to total veridical memory and total critical
memory recalled; these new variables resulted in non-
Table 1 Units per week consumed, AUQ score, binge score and
baseline memory measures [mean (SEM)] for the alcohol and placebo
Placebo Alcohol
Units per week
37.11 (5.40) 26.96 (3.32)
AUQ score 54.09 (5.85) 51.57 (7.92)
Binge score 21.46 (2.91) 24.61 (5.65)
Baseline memory score
(number of words)
17.56 (1.08) 17.19 (1.09)
1 U=8 g of alcohol
502 Psychopharmacology (2006) 188:498508
significant differences in intrusions as a function of total
veridical memory recalled [t(30)=.04, p=0.91] and total
critical memory recalled [t(30)=0.95, p=0.349]. Consequent-
ly, no corrections were administered before analyses and,
thus, results were analysed using raw data only.
Veridical items
To explore the effect of alcohol on veridical items, a mixed
2(drink: alcohol vs placebo)×2(repetition: lists presented
once vs lists presented three times) ANOVA was performed
on the percentage of veridical words recalled, with drink
serving as a between subjects variable, and repetition as a
within subjects variable.
A main effect of repetition was found [F(1, 30)=67.14,
p<0.001; see Fig. 1] demonstrating that repetition of lists
resulted in a greater probability of later recall. The
repetition × drink interaction was not found to be
significant [F(1, 30)=1.19, p=0.284], nor was a main effect
of drink obtained [F(1, 30)=1.77, p=0.193; Fig. 1].
Due to homogeneity of variance problem, separate
between subjects ttests between alcohol and placebo
groups were run in the single and repeated conditions. It
was found that participants in the alcohol group recalled
significantly less items than placebo participants for
singularly presented lists [t(30)=2.95, p=0.006]. No differ-
ence existed between the two groups for repeated lists
[t(30)=0.42, p=0.679].
Critical items
To explore the effect of alcohol and repetition on critical
items a mixed 2(drink: alcohol vs placebo)×2(repetition:
lists presented once vs lists presented three times) ANOVA
was performed on the percentage of critical words recalled,
with repetition as a within subjects factor, and drink as a
between subjects factor.
A main effect of repetition [F(1, 30)=3.93, p=0.057]
bordered on significance, with repetition increasing critical
recall. Whilst there was no main effect of the drink [F(1,
30)=1.38, p=0.249], a repetition × drink interaction [F(1,
30)=3.93, p=0.057], approached significance. Further
exploration revealed a non-significant effect of repetition
on the percentage of critical items recalled in placebo
subjects [t(1, 15)=0.00, p=1.00]. In contrast, repetition was
found to increase significantly the percentage of critical
items recalled in participants who had consumed alcohol
[t(15)=2.79, p=0.014; Fig. 2].
Implicit test of memory
A mixed 2(status: studied vs non-studied)×2(drink: placebo
vs alcohol) ANOVA was performed on the percentage of
words completed for the veridical and critical items
separately. The drink served as the between subjects
Free Recall (veridical items)
le Re
40 Placebo
Percentage of words
Fig. 1 Percentage of veridical items (mean±SEM) recalled from lists
presented singularly (single) or three times (repeat) after drinking
alcohol or placebo beverage. Free recall was tested 24 h after list
presentation. *p<0.05 compared to placebo in the single presentation
Table 2 VAS self-ratings [mean (SEM)] for alcohol and placebo
groups on day 1 measured after participants completed drinking the
beverage (shortly before list presentation), and on day 2 shortly before
Day 1 Day 2
Placebo Alcohol Placebo Alcohol
Relaxed 5.51 (.46) 6.1 (.47) 5.32 (.41) 4.77 (.52)
Content 5.65 (.39) 6.27 (.37) 5.65 (.39) 5.69 (.51)
Lightheaded 1.83 (.56) 5.96 (.52)* 0.53 (.09) 0.53 (.09)
*p<0.05 vs placebo on the same day of testing
Free Recall (critical items)
le Repeat
30 Placebo
Percentage of wor ds
Fig. 2 Percentage of critical items (mean±SEM) recalled associated
with lists presented singularly (single) or three times (repeat) after
drinking alcohol or placebo beverage. Free recall was tested 24 h after
list presentation. *p<0.05 compared to repeated presentation in the
alcohol condition
Psychopharmacology (2006) 188:498508 503
variable, whist word type and status were within subjects
There was an effect of status for veridical [F(1, 30)=
10.17, p=0.003], but not for critical items [F(1, 30)=1.83,
p=0.186] reflecting the finding that previously studying
veridical words increased the likelihood of completing a
stem to form that word. There was no status × drink
interaction found either in veridical [F(1, 30)=0.91,
p=0.346], or in critical [F(1, 30)=0.36, p=0.552], items
(see Fig. 3a,b). A further analysis was also performed on
the veridical and critical items from the studied words
(excluding baseline scores) to compare stem completion for
different degrees of repetition. A mixed 2(repetition: single
vs repeated)×2(drink: placebo vs alcohol) ANOVA was
performed on the percentage of words completed for the
veridical and critical items separately. An effect of repetition
was found for the veridical items [F(1, 30)=5.98, p=0.021]
but not for the critical items [F(1, 30)=0.12, p=0.729]
indicating that completion rates increased with repetition for
the items studied at encoding (data not shown). There were
no interactions found between drink and repetition [Fs(1, 30)
<0.66, Ps>0.422].
As a consequence of the non-significant priming effect in
critical items, conclusions regarding implicit memory can only
relate to veridical items, but the critical priming data can still
be used to generate measures of aware and unaware items.
Awareness measures
Analyses were performed to compare the relative rates of
aware and unaware items, and how they differed as a
function of alcohol and repetition.
Veridical items A mixed 2(drink: placebo vs alcohol)×2
(awareness: aware vs unaware)×2(repetition: presented once
vs repeated three times) ANOVA was performed on the
percentages of stems completed of the veridical items or of
their semantic associates (critical items) encountered in the
encoding phase that were presented during the implicit task.
The drink served as a between subjects variable, whilst
repetition and awareness were within subjects variables.
A main effect of awareness was found, signifying more
aware than unaware items [F(1, 30)=8.82, p=0.006]; thus
once participants had completed a stem to correspond to an
item viewed at encoding, they were more likely to endorse
it as an item they were aware of, than fail to do so. A main
effect of repetition was also found [F(1, 30)=5.98,
p=0.021], demonstrating that repetition served to increase
the probability that a stem would be completed to form
veridical items viewed in the encoding phase. An aware-
ness × repetition interaction [F(1, 30)=14.26, p=0.001],
signified that repetition significantly increased the amount
of aware items [t(31)=3.54, p=0.001], whilst had no effect
on the amount of unaware items [t(31)=1.22, p=0.234]. An
awareness × repetition × drink interaction [F(1, 30)=4.71,
p=0.038], was further explored using two-way ANOVAs
separately for the alcohol and placebo groups. In the
placebo group, the repetition × awareness interaction was
non-significant [F(1, 15)=1.24, p=0.283; Fig. 4a]. In
contrast, analysis in the alcohol group found a significant
repetition × awareness interaction [F(1, 15)=18.48,
p=0.001]. This interaction was further explored to reveal
that repetition significantly decreased the amount of unaware
veridical items [t(15)=2.18, p=0.046] but increased the
amount of aware veridical items [t(15)=3.39, p=0.004]
within the alcohol group (see Fig. 4b).
Critical items A mixed 2(drink: placebo vs alcohol)×2
(awareness: aware vs unaware)×2(repetition: presented
once vs repeated three times) ANOVA was performed on
the amount of stems completed to correspond to critical
items semantically related to veridical items encountered in
the encoding phase. The drink was a between subjects
variable, whilst awareness and repetition were both within
subjects variables.
a. Priming effect-veridical items
Baseline Studied
30 Baseline
Percentage of stems
b. Priming effect-critical items
Baseline Studied
Percentage of stems
Fig. 3 Percentage of stems completed (mean±SEM) to form veridical
(a) or critical (b) items corresponding to those viewed in the encoding
phase (studied) vs the percentages of stems completed to form words
not encountered in the encoding phase (baseline) for the alcohol and
placebo groups. *p<0.05 compared to studied words
504 Psychopharmacology (2006) 188:498508
A main effect of awareness was found [F(1, 30)=4.53,
p=0.042], reflecting the finding that participants were more
aware than unaware of critical items. There was no
significant main effect of repetition [F(1, 30)=0.12,
p=0.729], and neither the awareness × repetition [F(1, 30)
=0.35, p=0.558], nor the drink × repetition [F(1, 30)=0.66,
p=0.442] interactions were found to be significant. A
significant awareness × repetition × drink interaction [F(1,
30)=8.75, p=0.006] was further explored using separate
two-way ANOVAs for the alcohol and placebo groups. In
the placebo group, a significant repetition × awareness
interaction F(1, 15)=5.44, p=0.034], was further explored
to reveal that repetition increased the amount of unaware
critical words [t(15)=2.24, p=0.041], but did not affect
critical aware words [t(30)=1.31, p=0.211]. As being
unaware of critical items constitutes accurate memory,
being unaware of them, thus, amounts to being able to
correctly reject them, and this was enhanced in the placebo
group as a function of repetition. In the alcohol group, a
repetition × awareness interaction approached significance
[F(1, 15)=4.24, p=0.057]. This was further explored to
reveal that repetition had a borderline significant tendency to
decrease the amount of unaware critical items [t(15)=2.09,
p=0.055]. In contrast, repetition did not affect the amount of
aware items [t(15)=1.37, p=0.191]. As depicted in Fig. 5a,
b, a double dissociation was obtained, as repetition was
found to have opposite effects on levels of aware and
unaware critical items, dependent on the drink consumed.
Repetition was thus found to increase the accuracy of
memory for participants in the placebo group, and decrease
memory accuracy for participants in the alcohol group.
The present study investigated how alcohol and repetition
given at encoding modulated veridical and false memory
levels tested 24 h later, and whether these two distinct ways
of manipulating encoding interacted. Repetition was found
to effectively modulate the degree of information encoding
for both alcohol and placebo participants, as a main effect
of repetition signified that viewing lists multiple times
increased the probability of veridical recall. In addition, and
in accordance with the established anterograde impairments
of alcohol on explicit memory (e.g. Birnbaum and Parker
1977; Birnbaum et al. 1978; Duka et al. 2001; Weissenborn
and Duka 2000), participants in the alcohol group recalled
significantly less singularly presented veridical items than
placebo participants. This deleterious effect of alcohol was
rendered non-significant through increased learning, as
recall levels for repeated veridical items did not differ as a
function of the drink.
Participants in the alcohol group also recalled signifi-
cantly less false memory items from singularly presented
lists than the placebo group. The drink was not found to
affect false recall for items from repeated lists. Levels of
false recall were thus found to mirror patterns of veridical
memory in the alcohol and placebo groups. A meta-analysis
conducted by Roediger et al. (2001) demonstrated that
levels of veridical recall significantly predicted the degree
of false memory levels. Consequently, the decrease in false
memories in the alcohol group relative to the placebo group
for false memory items from singularly presented lists
could be the direct result of decreased veridical memory for
singularly presented lists. Craik (1977) hypothesised that
alcohol impaired veridical memory via a reduction of
encoding depth, leading to superficial encoding and a
reduction in semantic processing. Depth of processing has
been shown to predict the degree of semantic activation,
affecting levels of false memory (Rhodes and Anastasi
2000; Thapar and McDermott 2001; Toglia et al. 1999).
Superficial encoding under an alcohol challenge could
account for the effects of alcohol to decrease both veridical
and false memory. The effect of alcohol to decrease false
memory recall differs from previous findings (Mintzer and
Griffiths 2001; Milani and Curran 2000) showing no effect
Placebo group
Awareness measurements
Unaware Aware
30 Single
Percentage of Stems
(veridical it ems)
Alcohol group
Awareness measurements
Unaware Aware
30 Single
Percentage of Stems
(veridical items)
Fig. 4 Number of veridical stems completed (mean±SEM) and
rejected (unaware) or accepted (aware) as previously seen in placebo
(a) and in alcohol (b) participants for both the singularly (single) and
three times (repeat) presented lists. *p<0.05, **p<0.01 compared to
single presentation in aware and unaware judgement
Psychopharmacology (2006) 188:498508 505
or a tendency for alcohol to increase the false recognition
rates, respectively. The previous studies, however, differed
from the present study in that they did not test recall but
recognition and testing took place under the influence of
The implicit instructions for the stem completion task
(complete the stems to form the first word to come to
mind) resulted in a priming effect for veridical items, but
not critical items. The priming found with veridical items
did not differ as a function of the drink in accordance with
previous data showing that priming (measured also 24 h
after encoding) was unaffected by the drink consumed
(Duka et al. 2001). It is also in accordance with empirical
research that has demonstrated the preservation of auto-
matic components of memory under alcohol (Kirchner and
Sayette 2003; Tracy and Bates 1999). In contrast, no
priming effect was obtained in critical items. This appears
contrary to researchers who demonstrated a significant
effect of priming in critical items using stem completion
(Hicks and Starns 2005; McKone and Murphy 2000).
These researchers, however, used immediate stem comple-
tion, as opposed to the 24-h retention interval employed in
the current experiment. This differential effect of priming
for critical and veridical items over an extended retention
interval thus provides an interesting insight into the relative
durability of traces underlying true and false memory items.
Repetition was found to increase rates of completion in the
veridical items, whilst left the rates of completion with
critical items unaffected. This finding suggests that dura-
bility of traces underlying true memory can be further
supported by repetition.
Participants were required to make awareness judge-
ments for the words they had completed under implicit
instructions, following a procedure used by Duka et al.
(2001). By circling a word, participants were endorsing it as
having been presented at encoding, thus proclaiming
awarenessfor its prior presentation. Awareness of
veridical words, thus, constituted accurate memory. The
reverse was true, however, for false memory items. For, as
they were never presented, proclaiming awarenessof their
presentation was thus incorrect. As a priming effect was
only achieved in veridical items, a conclusion regarding the
automatic influences of memory as indicated by levels of
unaware items is restricted to veridical items (Duka et al.
2001). Whilst aware and unaware measures can still be
obtained for critical items, conclusions cannot be derived
from these measures in terms of automatic and controlled
memory processes. This is because stem completion for
critical items was at chance and consequently one cannot
infer that automatic memory processes aided their
For the veridical items, repetition was found to increase
stem completion under implicit instructions. It is interesting
to note that research into the effect of multiple repetitions
on implicit memory is unclear (Jacoby and Dallas 1991;
Roediger and Challis 1992). Whilst research has demon-
strated no effect of study list repetition on implicit memory
(Challis and Sidhu 1993; Parkin et al. 1990), the present
finding is consistent with more recent research that has
found a significant priming effect for multiply presented
words over singly presented words (Erickson and Reder,
unpublished manuscript). This finding can be accounted for
using the Source of Activation Confusion theory (e.g.
Reder and Schunn 1996), which implies that multiple
repetitions of study items should enhance priming relative
to single presentation of items due to an increase in
activation of corresponding concept (word) nodes. Multiple
presentations, hence, increase the potential for super
threshold activation of the word in question, which leads
into an enhanced priming effect. Such a finding substan-
tiates the use of repetition as a modulator to enhance
encoding and demonstrates that this effect of repetition
to enhance encoding can be obtained using implicit
Awareness status for veridical words differed as a
function of repetition in alcohol participants. This was
Placebo group
Awareness measurements
Unaware Aware
20 Single
Percent age of Stems
(criti cal i tems)
Alcohol group
Awareness measurements
Unaware Aware
20 Single
Percentage of Stems
(criti cal i tems)
Fig. 5 Number of critical stems completed (mean±SEM) and rejected
(unaware) or erroneously accepted (aware) as previously seen in
placebo (a) and in alcohol (b) participants for both the singularly
(single) and three times (repeat) presented lists. *p<0.05 compared to
single presentation in unaware judgement
506 Psychopharmacology (2006) 188:498508
not true for placebo participants who were more aware of
presented words than unaware of them irrespective of how
many times the words were presented (Fig. 4a). Specifi-
cally, in the alcohol group, increased repetitions served to
bring veridical items into conscious awareness; for words
viewed just once, participants were significantly more
likely to be unaware of their presentation relative to words
viewed repeated times (Fig. 4b). This finding supports
empirical research which documents the differential im-
pairment of alcohol on automatic and controlled memory
processes (Duka et al. 2001; Kirchner and Sayette 2003;
Lister et al. 1991; Tracy and Bates 1999). As priming was
not affected by the drink, automatic memory influences, as
quantified using degree of priming for veridical items,
were not found to be affected by the drink consumed.
Explicit awareness of previous presentation of words
generated under implicit instructions, however, was found
to be affected by the drink. In addition, explicit awareness
of previous presentation of veridical items was mediated
by repetition in the alcohol group only, as increased
learning brought these words into conscious awareness
counteracting the deleterious effect of alcohol on encoding.
Awareness measures taken for critical items initially
completed under implicit instructions demonstrated that
repetition had opposite effects in placebo and alcohol
participants. Awareness, and thus endorsement, of critical
items increased with repetition in alcohol participants,
whilst no such effect was present in the placebo group. In
addition, placebo participants were significantly better at
rejecting false memory items from repeated lists than from
lists presented once, whilst the reverse was true for
alcohol participants. Thus, regarding memory accuracy
as indexed by false memory endorsement, placebo
participantsmemory got more accurate with repetition,
whilst accuracy for participants in the alcohol group
This finding can be accounted for using the AMF
(Roediger et al. 2001). Under the AMF, activation of the
critical item can occur via direct activation of semantic
networks as a result of the formation of explicit associa-
tions between veridical items during list presentation
(Underwood 1965). Critical items are also thought to be
activated by the automatic spread of activation within
semantic networks. Under this latter route, conscious
thought of the critical item is not deemed necessary, as
the mere processing of veridical items is considered
sufficient for activation to spread to the critical item. The
effect of alcohol to increase false memories during free
recall (as discussed above) can be explained by its ability
to block activation of explicit associations during encod-
ing. Alcohol should leave unaffected the automatic spread
of activation within semantic networks as previous work
would suggest (Kirchner and Sayette 2003).
The monitoring component within the AMF functions to
ascertain the source of the semantic activation with
erroneous attribution of familiarity thought to derive from
prior presentation leading to false memory. Correctly
identifying the source of the activation reflects a successful
reality monitoring process(Johnson et al. 1993), as it
supposedly demonstrates the ability of individuals to
successfully attribute the activation of the critical items to
internal (e.g. thinking of the critical items) as opposed
to external (e.g. prior presentation) factors. It has been
argued that such a process is dependent upon intact
recollective memory (Benjamin 2001).
It is possible that due to impaired learning under alcohol,
activation levels are initially (under single presentation)
reduced. Repeated presentation, however, may lead to
increased activation of semantic networks and, hence, to
an increase in false memories. In contrast, placebo
participants are able to use repetition to increase the
richness of their memory traces. Such a richness would
mean when making awareness judgments for false memory
items, they are better able to differentiate between which
items were presented and which were notresulting in an
enhanced ability to reject critical false memory items with
repetition, something that alcohol participants are unable to
do because of their shallow processing. This differential
effect of repetition on false memory endorsement is
consistent with experiments that have shown increased
false memories as a function of repetition in populations
with impaired recollective processes such as the elderly
(Benjamin 2001).
Thus, in summary, the present study has shown a
differential effect of alcohol and repetition on the genera-
tion of false memories, and on accuracy judgements in
recognition of false memories in comparison to placebo.
Alcohol relative to placebo decreased the generation of
false memories for items presented once; repeated presen-
tations of items counteracted this effect of alcohol. On the
other hand, alcohol decreased accuracy judgements for
repeated items whilst an increase in accuracy judgement
was seen with repeated presentation of items in the placebo
participants. The activation and monitoring framework is
used to explain these effects of repetition and alcohol.
Benjamin AS (2001) On the dual effects of repetition on false
recognition. J Exp Psychol Learn Mem Cogn 27:941947
Birnbaum IM, Parker ES (1977) Alcohol and human memory.
Erlbaum, New Jersey
Birnbaum IM, Parker ES, Hartley JT, Noble EP (1978) Alcohol and
memoryretrieval processes. J Verbal Learn Verbal Behav
Psychopharmacology (2006) 188:498508 507
Birnbaum IM, Johnson MK, Hartley JT, Taylor TH (1980) Alcohol
and elaborative schemas for sentences. J Exp Psychol [Hum
Learn] 6:293300
Challis BH, Sidhu R (1993) Dissociative effect of massed repetition
on implicit and explicit measures of memory. J Exp Psychol
Learn Mem Cogn 19:115127
Craik F (1977) Similarities between the effects of ageing and
alcoholic intoxication on memory performance, construed within
alevels of processingframework. In: Birnbaum IM, Parker ES
(eds) Alcohol and human memory. Erlbaum, Hillsdale, NJ,
pp 921
Craik FIM, Lockhart RS (1972) Levels of processingframework for
memory research. J Verbal Learn Verbal Behav 11:671684
Curran HV, Hildebrandt M (1999) Dissociative effects of alcohol on
recollective experience. Conscious Cogn 8:497509
Deese J (1959) On the prediction of occurrence of particular verbal
intrusions in immediate recall. J Exp Psychol 58:1722
Duka T, Curran HV, Rusted JM, Weingartner HJ (1996) Perspectives
on cognitive psychopharmacology research. Behav Pharmacol
Duka T, Stephens DN, Russell C, Tasker R (1998) Discriminative
stimulus properties of low doses of ethanol in humans.
Psychopharmacology (Berl) 136:379389
Duka T, Weissenborn R, Dienes Z (2001) State-dependent effects of
alcohol on recollective experience, familiarity and awareness of
memories. Psychopharmacology (Berl) 153:295306
Gallo DA, Roediger HL 3rd (2002) Variability among word lists in
evoking memory illusions: evidence for associative activation
and monitoring. J Mem Lang 47:469497
Graf P, Schacter DL (1985) Implicit and explicit memory for new
associations in normal and amnesic subjects. J Exp Psychol
Learn Mem Cogn 11:501518
Hicks JL, Starns JJ (2005) False memories lack perceptual detail:
evidence from implicit word-stem completion and perceptual
identification tests. J Mem Lang 52:309321
Huron C, Servais C, Danion JM (2001) Lorazepam and diazepam
impair true, but not false, recognition in healthy volunteers.
Psychopharmacology (Berl) 155:204209
Jacoby LL, Dallas M (1991) On the relationshop between autobio-
graphical memory and perceptual learning. J Exp Psychol Learn
Mem Cogn 110:306340
Johnson MK (1977) What is bring counted none the less? Erlbaum,
New Jersey
Johnson MK, Hashtroudi S, Lindsay DS (1993) Source monitoring.
Psychol Bull 114:328
Kirchner TR, Sayette MA (2003) Effects of alcohol on controlled and
automatic memory processes. Exp Clin Psychopharmacol
Kiss GR, Armstrong C, Milroy R, Piper J (1973) An associative
thesaurus of English and its computer analysis. University Press,
Knowles SK, Duka T (2004) Does alcohol affect memory for
emotional and non-emotional experiences in different ways?
Behav Pharmacol 15:111121
Kucera H, Francis WN (1967) Computational analysis of present-day
American English. Brown University Press, Rhode Island
Lezak MD (1983) Neuropsychological assessment. Oxford University
Press, New York
Libby LK, Neisser U (2001) Structure and strategy in the associative
false memory paradigm. Memory 9:145163
Lister RG, Gorenstein C, Fisher-Flowers D, Weingartner HJ, Eckardt
MJ (1991) Dissociation of the acute effects of alcohol on
implicit and explicit memory processes. Neuropsychologia
McEvoy CL, Nelson DL, Komatsu T (1999) What is the connection
between true and false memories? The differential roles of
interitem associations in recall and recognition. J Exp Psychol
Learn Mem Cogn 25:11771194
McKone E, Murphy B (2000) Implicit false memory: effects of
modality and multiple study presentations on long-lived semantic
priming. J Mem Lang 43:89109
Mehrabian A, Russell JA (1978) A questionnaire measure of habitual
alcohol use. Psychol Rep 43:803806
Milani R, Curran HV (2000) Effects of a low dose of alcohol on
recollective experience of illusory memory. Psychopharmacology
(Berl) 147:397402
Mintzer MZ, Griffiths RR (2001) Alcohol and false recognition: a
doseeffect study. Psychopharmacology (Berl) 159:5157
Parker ES, Alkana RL, Birnbaum IM, Hartley JT, Noble EP (1974)
Alcohol and the disruption of cognitive processes. Arch Gen
Psychiatry 31:824828
Parker ES, Birnbaum IM, Noble EP (1976) Alcohol and memory
storage and state dependency. J Verbal Learn Verbal Behav
Parkin AJ, Reid TK, Russo R (1990) On the differential nature of
implicit and explicit memory. Mem Cognit 18:514
Reder LM, Schunn CD (1996) Metacognition does not imply
awareness: strategy choice is governed by implicit learning and
memory. In: Reder LM (ed) Implicit memory and metacognition.
Erlbaum, Hillsdale, pp 4577
Rhodes MG, Anastasi JS (2000) The effects of a levels-of-processing
manipulation on false recall. Psychon Bull Rev 7:158162
Roediger HL 3rd, McDermott KB (1995) Creating false memories:
remember words not presented in list. J Exp Psychol Learn Mem
Cogn 21:803814
Roediger HL 3rd, Watson JM, McDermott KB, Gallo DA (2001)
Factors that determine false recall: a multiple regression analysis.
Psychon Bull Rev 8:385407
Roediger HL, Challis BH (1992) Effects of exact repetition and
conceptual repetition on free recall and primed word-fragment
completion. J Exp Psychol Learn Mem Cogn 18:314
Schweizer TA, Vogel-Sprott M, Dixon MJ, Jolicoeur P (2005) The
stage-specific effect of alcohol on human information processing.
Psychopharmacology (Berl) 178:5257
Seamon JG, Luo CR, Schwartz MA, Jones LJ, Lee DM, Jones SJ
(2002) Repetition can have similar or different effects on accurate
and false recognition. J Mem Lang 46:323340
Stadler MA, Roediger HL 3rd, McDermott KB (1999) Norms for
word lists that create false memories. Mem Cognit 27:494500
Thapar A, McDermott KB (2001) False recall and false recognition
induced by presentation of associated words: effects of retention
interval and level of processing. Mem Cognit 29:424432
Toglia MP, Neuschatz JS, Goodwin KA (1999) Recall accuracy and
illusory memories: when more is less. Memory 7:233256
Townshend JM, Duka T (2002) Patterns of alcohol drinking in a
population of young social drinkers: a comparison of question-
naire and diary measures. Alcohol Alcohol 37:187192
Tracy JI, Bates ME (1999) The selective effects of alcohol on automatic
and effortful memory processes. Neuropsychology 13:282290
Tussing AA, Greene RL (1997) False recognition of associates: how
robust is the effect? Psychon Bull Rev 4:572576
Tussing AA, Greene RL (1999) Differential effects of repetition on
true and false recognition. J Mem Lang 40:520533
Underwood BJ (1965) False recognition produced by implicit verbal
responses. J Exp Psychol 70:122129
Weissenborn R, Duka T (2000) State-dependent effects of alcohol
on explicit memory: the role of semantic associations.
Psychopharmacology (Berl) 149:98106
508 Psychopharmacology (2006) 188:498508
ResearchGate has not been able to resolve any citations for this publication.
Full-text available
Roediger and McDermott (1995) induced false recall and false recognition for words that were not presented in lists. They had subjects study 24 lists of 15 words that were associates of a common word (called the critical target or critical lure) that was not presented in the list. False recall and false recognition of the critical target occurred frequently in response to these lists. The purpose of the current work was to provide a set of normative data for the lists Roediger and McDermott used and for 12 others developed more recently. We tested false recall and false recognition for critical targets from 36 lists. Despite the fact that all lists were constructed to produce false remembering, the diversity in their effectiveness was large—60% or more of subjects falsely recalled window and sleep following the appropriate lists, and false recognition for these items was greater than 80%. However, the list generated from king led to 10% false recall and 27% false recognition. Possible reasons for these wide differences in effectiveness of the lists are discussed. These norms serve as a useful benchmark for designing experiments about false recall and false recognition in this paradigm.
Full-text available
Two experiments (modeled after J. Deese's 1959 study) revealed remarkable levels of false recall and false recognition in a list learning paradigm. In Experiment 1, subjects studied lists of 12 words (e.g., bed, rest, awake ); each list was composed of associates of 1 nonpresented word (e.g., sleep). On immediate free recall tests, the nonpresented associates were recalled 40% of the time and were later recognized with high confidence. In Experiment 2, a false recall rate of 55% was obtained with an expanded set of lists, and on a later recognition test, subjects produced false alarms to these items at a rate comparable to the hit rate. The act of recall enhanced later remembering of both studied and nonstudied material. The results reveal a powerful illusion of memory: People remember events that never happened.
The effects of study-list repetition on false recognition of semantic associates were examined using aging (Experiment 1) and recognition time pressure (Experiment 2). Participants studied word lists, each of which was composed of high associates to a single, unstudied word (the critical lure). Under normal testing circumstances, young adult participants (ages 19-26) falsely endorsed fewer critical lures associated with lists that had been presented multiple times than lists presented only once. However, young participants tested under time pressure and older participants (ages 67-85) endorsed a greater number of critical items associated with lists presented thrice than with lists presented once. The results suggest dual bases for the recognition decision, one of which is based on the rapid spread of activation within domains of semantic similarity and the other of which functions to attribute that activation to likely sources and set appropriate decision criteria. The latter capacity is compromised both under conditions of time pressure and in the elderly.
In the Deese paradigm, studying a list of semantic associates (holiday, beach, etc.) produces false explicit memory for a nonpresented lure (vacation). Here, we examine false memory with implicit retrieval. Experiment 1 tested memory for lures and matched on-list targets, using stem completion (implicit) and stem-cued recall (explicit). We replicated McDermott's (1997) finding of implicit false memory at a 10-min delay, using better controls for explicit contamination. In Experiment 2, we show that this semantic priming is modality-specific (on the visual test, lure priming was reduced with auditory study of the semantic associates), consistent with the perceptual nature of stem completion. In Experiment 3, additional study presentations reduced false memory with explicit, but not implicit, retrieval, consistent with suppression of gist-based responses by veridical information only when explicit retrieval is required. Results also dissociate implicit and explicit false memory and demonstrate similarities between false and true memories with implicit retrieval.
Many studies of recognition memory have shown that repeating a word on a study list increases true recognition (i.e., the hit rate) of that word. Less clear is the extent to which repeating a word on the study list increases false alarms to a semantically related distractor. Five experiments examined the effect of repetition on true and false recognition. In the first two experiments, false recognition of a distractor was only found when five distinct words related to it appeared on the list. In Experiments 3 and 4, presenting a list word either 5 or 10 times failed to increase false recognition of a related distractor, relative to when the list word occurred once. Experiment 5 revealed that when five words related to a distractor were blocked together in the study list, repeating them caused a significant decrease in false recognition of the distractor. Repetition appears to be one variable that differentially affects true and false recognition.
False recognition of nonpresented words that were strong associates of 12 words in a study list was examined. Six lists were read to subjects; each list contained the 12 strongest associates to a critical nonpresented word. False-alarm rates to the 6 critical nonpresented words were obtained under several different conditions. The manipulations included varying the level of processing done to the study lists, varying the recognition-test procedure, repeating each of the study lists three times, and mixing the words from the six study lists together. A reliable false-recognition effect for critical nonpresented words was obtained in all conditions. However, the effect was not impervious to all of the manipulations. Significantly lower false recognition was obtained when learning was incidental as well as when the words on the six lists were mixed together. Neither level of processing nor repetition significantly influenced false recognition. This last result is inconsistent with Hintzman’s (1988) MINERVA 2 global memory model, but agrees with predictions from Shiffrin, Ratcliff, and Clark’s (1990) SAM model.