Content uploaded by Leilani B Goodmon
Author content
All content in this area was uploaded by Leilani B Goodmon
Content may be subject to copyright.
Theoretical Implications of Extralist Probes for Directed Forgetting
Lili Sahakyan and Leilani B. Goodmon
University of North Carolina at Greensboro
In 5 experiments, the authors examined the influence of associative information in list-method
directed forgetting, using the extralist cuing procedure (Nelson & McEvoy, 2005). Targets were
studied in the absence of cues, but during retrieval, related cues were used to test their memory.
Experiment 1 manipulated the degree of resonant connections from associates of the target back to
the target. Experiment 2 varied the degree of connectivity of associates of the target. Experiment 3
varied the size of the associative neighborhood of the target. Experiment 4 varied the direct
target-to-cue strength, and Experiment 5 varied the indirect strength between the cue and the target.
Reliable directed forgetting impairment emerged in all experiments. Furthermore, directed forgetting
reduced the effects of the associates contributing to the target activation strength (Experiments 1–2),
and it also reduced the effects of the associates contributing to the cue–target intersection strength
(Experiments 3–5). Together, these results support the context account and challenge the inhibitory
interpretation of directed forgetting.
Keywords: directed forgetting, context change, cued recall, independent probes
Memory researchers have always been interested in the funda-
mental act of forgetting and the mechanisms by which memories
become attenuated. Traditionally, the investigation of the forget-
ting processes has emphasized passive forms of forgetting, which
arise unintentionally in response to changes in the environment or
because of the accumulation of more traces in memory (for a
review, see M. C. Anderson & Neely, 1996; Postman, 1971). In the
past decade, however, there has been growing interest in the ability
to retrieve appropriate memories and exclude inappropriate ones.
Any system that uses memory must be able to identify some
memories as more relevant than others and must be able to reduce
the accessibility of unwanted information. Research on intentional
forms of forgetting has shifted focus from viewing forgetting as a
passive process that happens to people to viewing forgetting as a
process over which people can exert control.
In the laboratory, intentional forgetting has been studied with a
variety of paradigms, including the think/no-think (e.g., M. C.
Anderson & Green, 2001), retrieval practice (e.g., M. C. Anderson,
Bjork, & Bjork, 1994), and directed forgetting paradigms (e.g.,
R. A. Bjork, LaBerge, & LeGrand, 1968). In this article, we used
the directed forgetting paradigm. However, instead of the more
established free recall or recognition tests, we assessed directed
forgetting in a novel way that involved the use of unstudied cues
to test memory for the information that participants were instructed
to forget. Such an approach allowed us to evaluate the leading
theories of directed forgetting.
In directed forgetting studies, participants study some informa-
tion and are subsequently told to forget certain portions of it. In the
list-method version of the procedure, participants are told to forget
an entire block of the items they studied earlier, whereas in the
item-method version, participants are told to forget or remember on
an item-by-item basis (e.g., Basden, Basden, & Gargano, 1993;
MacLeod, 1999;). This article utilizes the list-method design, which
typically involves presenting two lists of items to study for a later
memory test. Following the first list, some participants are told to
forget the first list because it was the practice list, whereas others are
told to remember it because it was the first half of the study list.
Everyone encodes the second list, which is followed by a memory test
for both lists. Directed forgetting instructions impair recall of List 1
items and enhance recall of List 2 items in the forget group compared
with the remember group; these are termed the costs and the benefits
of directed forgetting, respectively (for reviews, see E. L. Bjork,
Bjork, & Anderson, 1998; Johnson, 1994; MacLeod, 1998).
The aim of this article was to evaluate the two major theoretical
explanations of directed forgetting—the inhibitory account and the
context-change account. According to the inhibitory account, the
forget instruction inhibits List 1 items, impairing their recall without
altering their availability in memory (e.g., E. L. Bjork & Bjork, 1996,
2003; R. A. Bjork, 1989; Geiselman, Bjork, & Fishman, 1983).
“Inhibition merely limits retrieval by reducing activation of unwanted
items” (M. C. Anderson, 2009, p. 224). As List 1 items become
inhibited, they interfere less with List 2 items, thereby producing the
benefits of directed forgetting. The proponents of the inhibitory view
have argued that inhibition can be released with the provision of
appropriate cues. For example, the absence of directed forgetting on
recognition tests or implicit fragment completion tests was interpreted
as evidence of release of inhibition resulting from the re-exposure of
to-be-forgotten (TBF) items on these tests (Basden et al., 1993; E. L.
Bjork & Bjork, 1996). Also, in some studies, a free recall test was
Lili Sahakyan and Leilani B. Goodmon, Department of Psychology,
University of North Carolina at Greensboro.
We are grateful to Alyson Gurr, Alexandra Kueider, Steven Minor, Cory
Viklund, Catherine Patteson, and Nathaniel Foster for their help in data
collection and scoring.
Correspondence concerning this article should be addressed to Lili
Sahakyan, Department of Psychology, University of North Carolina at
Greensboro, 296 Eberhart Building, P.O. Box 26170, Greensboro, NC
27402-6170. E-mail: l_sahaky@uncg.edu
Journal of Experimental Psychology: © 2010 American Psychological Association
Learning, Memory, and Cognition
2010, Vol. 36, No. 4, 920–937
0278-7393/10/$12.00 DOI: 10.1037/a0019338
920
administered both before and after some intervening activities, and
depending on the nature of the intervening task, directed forgetting
was occasionally released on the final test compared with the initial
test (e.g., Basden et al., 1993; E. L. Bjork & Bjork, 1996). For
example, an intervening recognition test that included TBF items as
distractors released directed forgetting on the subsequent recall test
(Basden et al., 1993; E. L. Bjork & Bjork, 1996). In contrast, when
recall was delayed by intervening arithmetic (Basden et al., 1993;
E. L. Bjork & Bjork, 1996) or by an implicit fragment completion task
that included TBF items (E. L. Bjork & Bjork, 1996), then directed
forgetting was not released. E. L. Bjork and Bjork (1996) noted that
it is not the mere exposure to TBF items that releases them from
inhibition but rather the processes initiated by the intervening task;
these processes reverse directed forgetting if they redirect attention to
the original study episode. Thus, a recognition test is an explicit test
that directs attention to the study episode as opposed to the implicit
fragment completion test, which does not invoke such processes.
These observations suggest that the reinstatement of contextual infor-
mation associated with encoding of TBF items might be the key factor
underlying the release phenomena.
Indeed, another theoretical explanation of directed forgetting
invokes contextual factors rather than inhibition as the basis of
directed forgetting. According to the context-change account of
directed forgetting, the forget instruction encourages participants
to change their mental context between the lists, thereby segregat-
ing the two lists as separate events (Sahakyan & Kelley, 2002).
The impairment arises because at the time of test, context better
matches the context that was prevalent during the encoding of List
2 than List 1, producing forgetting of List 1 items in the forget
group. The benefits of directed forgetting arise because contextual
differentiation reduces interference between the lists. Although
many studies have not obtained directed forgetting in recognition,
recently there have been several reports of significant effects in
recognition (e.g., Benjamin, 2006; Sahakyan & Delaney, 2005),
particularly when the conditions of recognition promoted the im-
portance of contextual information (e.g., Lehman & Malmberg,
2009; Sahakyan, Waldum, Benjamin, & Bickett, 2009). In con-
trast, when contextual information is irrelevant to the task, as is the
case with implicit memory tests, then directed forgetting is not
obtained (e.g., Basden et al., 1993; E. L. Bjork & Bjork, 1996).
Thus, the presence or absence of directed forgetting in recognition
or implicit tests can also be explained in terms of the conditions
that encourage utilization of contextual information. Likewise, the
release of directed forgetting from certain types of intervening
tests can be explained by reinstatement of the context associated
with encoding of TBF items. In prior research, mentally reinstating
the episodic context of TBF items before the final test reduced
directed forgetting even in the absence of any re-exposure to TBF
items (e.g., Sahakyan & Kelley, 2002).
In this article, we utilized a novel procedure for examining the
directed forgetting theories, by using the extralist cuing technique,
pioneered by Nelson, McEvoy, and their colleagues (for a review,
see Nelson & McEvoy, 2005). The technique involves presenting
target words for study and later testing people with cue words that
are related to the targets but were not studied with them— hence
the name extralist. For example, participants may study the target
word dinner and receive the test cue lunch to help them retrieve
dinner. Performance on this task is influenced both by the target
characteristics and the cue characteristics (for reviews, see Nelson,
McKinney, Gee, & Janczura, 1998; Nelson & Zhang, 2000).
Extralist cuing allows us to independently manipulate the cue
characteristics while holding the target characteristics constant (or
vice versa). This aspect of the procedure is appealing because it
enables us to contrast the inhibitory account and the context
account of directed forgetting. The two accounts make opposing
predictions for our experiments, and we describe them in greater
detail in the sections that follow.
To date, little is known about how directed forgetting might
behave in extralist cued recall. Only one study so far has used an
extralist cuing procedure in directed forgetting (Basden et al.,
1993). Participants studied the second word of associatively re-
lated word pairs and later received the first word as a cue to help
them retrieve the second word. Basden et al. (1993) reported better
recall for to-be-remembered (List 2) compared with TBF (List 1)
items in the forget group; there was no remember group included
in the design. Although in the past some investigators have as-
sessed directed forgetting by comparing recall between the lists in
the forget group, this method confounds directed forgetting with
numerous other factors, including recency (for a review, see M. C.
Anderson, 2005), and it does not allow for the estimation of
directed forgetting costs and benefits because assessing these
effects requires comparisons with the remember group. In other
words, because the costs and the benefits cannot be evaluated
without the remember group, it remains an empirical question
whether they can be obtained with extralist cued recall. In the
sections that follow, we briefly review the research on the implicit
variables that we manipulated in this article. Some of the variables
describe the properties of the targets, whereas others refer to the
properties of the test cues. To set the stage for the predictions of
directed forgetting theories, we first discuss the processes impli-
cated in extralist cued recall, including how disruptions of episodic
context influence performance on this task.
Implicit Associations and Their Influence in Extralist
Cued Recall
Much evidence shows that encoding a familiar word implicitly
activates its related concepts from past experiences (e.g., J. R.
Anderson, 1983; Kintsch, 1988; Nelson, Schreiber, & McEvoy,
1992). Although the associates of the target are not consciously
experienced, they produce systematic effects in memory (for re-
views, see Nelson & McEvoy, 2005; Nelson et al., 1998; Nelson &
Zhang, 2000). Nelson et al. (1998) used free association to mea-
sure the associative structures of many words and have shown that
words differ in how many associates they have and in the patterns
of connections among them. The probability that any word is
produced in response to another word in free association is taken
as a measure of relative strength between the words.
Free association measurements may or may not be representa-
tive of the associative structure of a single individual (e.g.,
Bilodeau & Howell, 1968; Fox, 1968; Simpson & Voss, 1967).
Clearly, there is variability both between people because of dif-
fering experiences (e.g., heavy drinkers have stronger associations
to the word alcohol than light drinkers; Reich & Goldman, 2005)
and within a given individual as a result of recent experience.
However, the norms are not used to predict individual participant
data but rather performance of similar groups of participants.
Research suggests that free association measures capture the key
921
DIRECTED FORGETTING IN EXTRALIST CUING
aspects of preexisting lexical experience with high reliability (Nel-
son, Dyrdal, & Goodmon, 2005; Nelson, McEvoy, & Dennis,
2000) and that they are useful in predicting free recall (Deese,
1965), cued recall (Bahrick, 1970), recognition (Nelson et al.,
1998), and false memories (Deese, 1959; McEvoy, Nelson, &
Komatsu, 1999).
Implicit Variables
Several robust findings have emerged from research on pre-
existing associations. We limit the discussion to five variables that
were examined in relationship to directed forgetting (for additional
variables, see Nelson et al., 1998). The number of associates that the
target produces in free association (known as the size of the network)
affects extralist cued recall, with targets with smaller networks having
arecalladvantagecomparedwithtargetswithlargernetworksof
associates (Nelson & Friedrich, 1980; Nelson et al., 1998; Nelson &
Schreiber, 1992). Also, targets whose associates are more connected
to each other are more likely to be recalled than targets whose
associates are less connected to each other (Nelson, Bennett, Gee,
Schreiber, & McKinney, 1993). Finally, targets whose associates
produce them as a response in free association are better recalled than
targets whose associates are less likely to produce them (Nelson et al.,
1998). In addition to the target characteristics, cue characteristics also
affect extralist cued recall. For instance, associates that are strongly
activated by the target during encoding are more effective retrieval
cues for that target compared with weakly activated associates (e.g.,
Nelson & McEvoy, 1979). Furthermore, indirect connections between
the test cue and the target also affect cued recall (e.g., Nelson et al.,
1998).
The most critical finding that motivated the experiments in this
article concerns the effect of contextual disruptions on implicitly
activated information. In particular, conditions that block or dis-
rupt the retrieval of episodic contextual information were shown to
reduce the influence of implicit variables on memory. For exam-
ple, set size effects and connectivity effects are reduced after
contextual disruptions (Nelson, Goodmon, & Akirmak, 2007; Nel-
son, McEvoy, Janczura, & Xu, 1993; Nelson et al., 1998) and so
are the effects of direct and indirect connections between the target
and the cue (Nelson & Goodmon, 2002). In some studies, access
to context was manipulated by giving the test in a new physical
location (Nelson & Goodmon, 2002; Nelson, Goodmon, & Ceo
2007) or by introducing test delays of increasing durations (Nel-
son, Goodmon, & Akirmak, 2007). Other studies have used tests
that discourage the use of context during the test, such as primed
free association (Goodmon & Nelson, 2004; Nelson, Goodmon, &
Akirmak, 2007). In all of these studies, the effects associated with
implicit variables were reduced.
Explaining Extralist Cued Recall
To explain these findings, Nelson and colleagues developed a
theoretical model, according to which two key processes are in-
volved in extralist cued recall—the target activation process and
the cue–target intersection process (for computational details, see
Nelson, Goodmon, & Ceo, 2007). Target activation is an integrat-
ing process that involves parallel activation of the target’s associ-
ates and the links that connect the associates to one another when
the target is initially processed. It is described by adding the
strengths of all the links in the associative network (i.e., the
strengths of target-to-associates connections, associates-to-target
connections, and associate-to-associate connections). Higher val-
ues indicate stronger target activation and lead to better memory.
This explains the memorial advantage of targets that have more
links between their associates as well as the advantage of targets
that have more links from their associates— known as the connec-
tivity effect and the resonance effect. Both of these effects are
thought to emerge during the learning stage, and they are found
also in recognition tests (Nelson, Zhang, & McKinney, 2001).
When the test cue is presented at the time of test, it activates its
own associates, just as the target did during encoding. During the
retrieval stage, a separating process selects the target from the
associates activated by the cue and the associates activated by
the target. The success of the intersection process is influenced by
the connections that bind the target and the cue together as well as
by the connections that fail to join them. Figure 1 shows examples
of linking and nonlinking connections between a hypothetical
target and its test cue. Whereas the linking connections between
the cue and the target facilitate memory, the nonlinking connec-
tions become a source of interference that impairs memory. As the
size of the associative neighborhood increases, so does the number
of the nonlinking connections in that network, which accounts for
the negative effects of the set size (Schreiber, 1998; Schreiber &
Nelson, 1998). The model describes the intersection process with
a ratio rule, which is a function of contrast between the strength of
Target
A1 A2
A4
A5
A6
A7
A8
A3
cue
Figure 1. This figure shows the associative network of a hypothetical
target, which produces eight associates (A
1
–A
8
) in free association. The
length of links and the distances between the associates are drawn arbi-
trarily. The dotted lines represent the associative links involving the hy-
pothetical test cue (A
3
). Note that associates outside of the target’s network
can also serve as test cues. Two types of direct connections link the cue and
the target together—the target-to-cue association and the cue-to-target
association. Increases in strength in both directions independently facilitate
recall (Nelson, Fisher, & Akirmak, 2007; Nelson & McEvoy, 1979; Nelson
et al., 1998). There are also indirect connections between the cue and the
target that can facilitate recall. For example, associates A
1
,A
5
, and A
6
are
shared associates because both the target and the test cue produce them.
The more indirect associations there are between the cue and the target, the
more successful the cue is in retrieving the target (Nelson & Goodmon,
2002, 2003; Nelson & McEvoy, 2002; Nelson et al., 1998). Finally, there
are nonlinking connections that do not bind the test cue and the target
together. For example, associates A
2
,A
4
,A
7
, and A
8
are unique to the target
because they are not linked to the cue. Such associates are competitors, and
they lower the probability that the target will be chosen in the presence of
the test cue (Nelson et al., 1998).
922 SAHAKYAN AND GOODMON
the linking connections relative to the strength of the nonlinking
connections.
Finally, to explain the effects of contextual disruptions, the
model adopts an interactive cuing assumption. It proposes that the
extralist cue and the episodic context cues combine together in a
multiplicative way to retrieve the target. Episodic contextual in-
formation encoded along with the target is critical in the retrieval
process in part because it helps to differentiate the target from
other associates activated by the cue.
Current Studies and Predictions of Directed
Forgetting Theories
We report five experiments that crossed directed forgetting with
several associative variables. The first two experiments manipu-
lated variables that influence the target activation strength, includ-
ing target resonance (Experiment 1) and target connectivity (Ex-
periment 2). The next three experiments manipulated variables that
influence the cue–target intersection strength. These involved
varying target set size (Experiment 3), direct target-to-cue strength
(Experiment 4), and indirect strength between the target and the
cue (Experiment 5). We did not manipulate the explicit encoding
strength, because explicit encoding strength (varied through levels-
of-processing, study time, or spaced repetitions) does not interact
with implicit variables and has only an additive contribution to
recall (Nelson, Bennett, et al., 1993; Nelson, Bennett, & Leibert,
1997; Nelson, Fisher, & Akirmak, 2007; Nelson & Goodmon,
2002; Nelson, McEvoy, et al., 1993; Nelson, McEvoy, & Schre-
iber, 1990; Nelson et al., 1992).
Because episodic contextual disruptions were shown to reduce
the effects of implicit variables, these findings have implications
for the contextual account of directed forgetting. The latter predicts
effects similar to those produced by the disruption of context found
in prior studies. That is, implicit associates should show reduced
effects in the forget group compared with the remember group,
producing an interaction between the manipulated implicit variable
and directed forgetting. Furthermore, directed forgetting should
interact with all implicit variables regardless of whether the study
manipulates the variables involved in the target activation or the
cue–target intersection process. This prediction results from the
interactive cuing assumption of Nelson, Goodmon, and Ceo’s
(2007) model, which explains how disruptions in context affect
extralist cued recall.
The inhibitory account does not make the same predictions, and
one could even argue that it makes the opposite predictions in
some experiments, considering its assumptions. For Experiments
1–2, which manipulated the target activation level, it is not obvious
why highly activated targets should suffer more from directed
forgetting than weakly activated targets (as predicted by the con-
text account). If anything, one would expect the opposite, because
highly activated targets may be harder to inhibit. For Experiments
3–5, which manipulated the intersection strength variables, the
inhibitory account would predict the opposite of the context ac-
count. Whereas the context account predicts a reduced effect of the
cue strength in the forget condition compared with the remember
condition, the inhibitory account predicts an enhanced effect of
the cue strength in the forget condition. This is because the
inhibitory account assumes that items can be released from inhi-
bition with appropriate cues. If inhibition can be released, then one
would expect that stronger cues would be more effective at releas-
ing the items from inhibition than weaker cues, resulting in a larger
effect of the cue strength in the forget than the remember condi-
tion. Overall, unlike the context account, the inhibitory account
does not predict reduced effects of implicit variables in the forget
group in any of the experiments.
Section 1: Target Activation Variables
Experiment 1: Resonance
Words vary in terms of the likelihood that their associates
produce them in turn in free association. For some items, most of
their associates produce them as a response in free association,
whereas for other words, few of their associates produce them as
a response. This property of the words has been termed resonance
(Nelson et al., 1998). Figure 2 shows two hypothetical targets with
similar associative set sizes but different degrees of resonant
Figure 2. Associative map of targets with high and low resonant connections from the associates back to the
target. Resonant links are shown with bold arrows.
923
DIRECTED FORGETTING IN EXTRALIST CUING
connections. The word CHALK is a low resonance item because
only two of its associates produce it in free association (e.g., board
and eraser), whereas the word HAMMER is a high resonance item
because four of its associates produce it in free association (e.g.,
tool,nail,pound, and saw). Items with more extreme resonance
characteristics also exist, such that either none of their associates
produce them in free association (e.g., airport) or all of their
associates produce them in free association (e.g., circle).
Studies have shown that targets with more and stronger resonant
links are more likely to be recalled in extralist cued recall (e.g.,
Nelson et al., 1998). The memorial advantage of items with high
resonance is known as the resonance effect. Resonance represents
a source of target activation strength, and its effect is independent
of connectivity of associates in the network of the target (Nelson,
McEvoy, & Pointer, 2003).
To the best of our knowledge, there have been no studies
examining the effect of contextual disruption on the resonance
effect. Thus, it remains an empirical question whether directed
forgetting is affected by resonance. On the basis of the context
account and the prior research on disruptions of context in the
extralist cued recall, we expected an interaction between these
variables, with directed forgetting reducing the magnitude of the
resonance effect.
Method.
Participants and design. Participants were 72 undergraduates
at the University of North Carolina at Greensboro who took part in
the study in exchange for course credit. The experimental design
was a 2 !2 mixed-factorial design, with resonance (high, low)
manipulated within subjects and instruction (forget, remember)
manipulated between subjects.
Materials. We selected 32 words from the University of South
Florida Free Association Norms (Nelson, McEvoy, & Schreiber,
1999) to serve as targets for two unrelated 16-item study lists. An
additional 32 words served as test cues, one per target item.
Targets were unrelated to each other, and cues were unrelated to
each other. Each cue word was related only to a single target. In
selecting cue–target pairs, we controlled the variables that are
known to influence recall except for the associations involved in
target resonance. This was accomplished with the help of a com-
puter program called ListChecker. When target resonance was
high, the probability that the associates in the target’s network
produced the target in free association averaged .70 (SD ".09),
with the sum of associate-to-target strengths averaging 2.49 (SD "
0.48). When target resonance was low, the probability of resonant
connections averaged .24 (SD ".06), with the sum of associate-
to-target strengths averaging 1.16 (SD "0.15). Whereas target
resonance was varied, other variables were controlled across the
resonance conditions. These included target set size (M"15.16,
SD "2.67), target competitor strength (M".61, SD ".14), target
connectivity as indexed by the number of associate-to-associate
connections (M"1.63, SD "0.52) and the sum of those connec-
tion strengths (M"2.96, SD "1.17), target concreteness on a 1–7
scale (M"5.02, SD "1.38), and target printed frequency (M"
57.00, SD "40.03; Kucˇera & Francis, 1967). We also controlled
cue set size (M"15.34, SD "2.69), cue competitor strength
(M".51, SD ".14), cue concreteness on a 1–7 scale (M"5.08,
SD "1.26), and cue printed frequency (M"43.44, SD "49.09).
Finally, direct and indirect associations between the cue and the
target were also held constant across the resonance conditions.
Direct cue-to-target strength averaged .05 (SD ".02), and direct
target-to-cue strength averaged .04 (SD ".03). Indirect strength
involving shared associates and mediators was also controlled.
Shared associate strength averaged .02 (SD ".02) and mediated
associate strength averaged .01 (SD ".02).
Procedure. Participants were tested individually. Each partic-
ipant studied two lists of items, with half the items in each list
being high-resonance targets and the remaining items being low-
resonance targets. The position of the two lists was counterbal-
anced during presentation. Participants were told to read each word
and to try to remember the words for a later unspecified memory
test. The words were presented one at a time on a computer screen
at a rate of 4 s. The presentation order of high- and low-resonance
targets within each list was randomized. After the first list, partic-
ipants in the forget condition were informed that the first list was
only for practice, that there was no need to remember the items,
and that they should try to forget them. Participants in the remem-
ber condition were told that the list they studied was only the first
half of the items and that they should remember them for a later
test. Both forget and remember groups then studied the second list
of targets. Afterward, all participants received an extralist cued
recall test for items from both lists. They were told that meaning-
fully related cues were going to be presented to help them remem-
ber the studied words. They were instructed to use each cue to
recall a related item from either the first list or the second list. The
experimenter further clarified to the forget group that the first list
referred to the practice list and that they should try to retrieve both
the practice items and the second list items in response to the test
cues. Participants were told that if no word from either list came to
mind, they could guess. Guessing instructions are typical in ex-
tralist cuing studies because prior research has shown that the
magnitude of the implicit variable effects is unaffected by whether
the test instructions require, merely encourage, or forbid guessing
(e.g., Nelson et al., 1992). Cues for both lists were randomly
combined and presented one at a time on the computer screen.
Participants made their responses aloud, and the experimenter
recorded them. Testing was self-paced. We tested both lists simul-
taneously rather than separately because the testing procedure was
under the control of the experimenter, reducing the potential for
output interference at the time of test. In contrast, in free recall,
participants could start their recall from List 2 items, thereby
confounding directed forgetting with output interference. There-
fore, in free recall, each list is usually tested separately, whereas
simultaneous testing is more common with other memory tests,
such as recognition (e.g., Benjamin, 2006; MacLeod, 1999;
Sahakyan & Delaney, 2005), or implicit tests (e.g., Basden et al.,
1993; E. L. Bjork & Bjork, 1996).
Results.
Directed forgetting costs. To analyze the costs of directed
forgetting, a Resonance (high, low) !Instruction (forget, remem-
ber) mixed-factorial analysis of variance (ANOVA) was con-
ducted on the proportion of List 1 recall, with instruction as the
between-subjects variable. The results are shown in Figure 3 (top
panel). There was a main effect of resonance, F(1, 70) "9.27,
MSE "0.015, p#.01, $
2
".12, indicating better List 1 memory
for high-resonance targets (.31) than low-resonance targets (.25).
There was also a main effect of instruction, F(1, 70) "9.89,
MSE "0.016, p#.01, $
2
".12, indicating that the remember
group participants recalled more List 1 items (.32) than the forget
924 SAHAKYAN AND GOODMON
group participants (.25). The Instruction !Resonance interaction
was also significant, F(1, 70) "4.12, p#.05, $
2
".06. There was
an 11% resonance effect in the remember condition, t(35) "3.21,
p#.01, and a substantially reduced resonance effect (3%) in the
forget condition (t#1).
Directed forgetting benefits. To analyze the benefits of di-
rected forgetting, similar analyses were carried out on the propor-
tion of List 2 recall. The results are shown in Figure 3 (bottom
panel). There was a main effect of resonance, F(1, 70) "6.41,
MSE "0.022, p#.05, $
2
".08, indicating better List 2 memory
for high-resonance targets (.38) than low-resonance targets (.32).
There was no main effect of instruction, F(1, 70) "1.83, p".18
(.38 in the forget and .32 in the remember group). There was also
no interaction (F#1). These results suggest that regardless of the
resonance condition, there were no significant directed forgetting
benefits. Although numerically the effect was in the right direc-
tion, it was not significant.
Experiment 2: Connectivity
Words vary in terms of the density of connections among the
associates in their associative network—that is, the extent to which
the associates of a target produce each other when they are inde-
pendently normed. This property of the word has been termed
connectivity (Nelson et al., 1998). Figure 4 shows two hypothetical
targets with similar associative set size. However, the associates of
the target word POLITE are more densely connected to each other
than the associates of the target word CORK. Research has shown
List 1 Costs
.00
.10
.20
.30
.40
.50
puorg rebmemeRpuorg tegroF
Proportion List 1 Recall
high resonance
low resonance
List 2 Benefits
.00
.10
.20
.30
.40
.50
puorg rebmemeRpuorg tegroF
Proportion List 2 Recall
high resonance
low resonance
Figure 3. List 1 recall (top) and List 2 recall (bottom) as a function of associative resonance in Experiment 1.
Error bars represent standard error of the means.
925
DIRECTED FORGETTING IN EXTRALIST CUING
that targets with a densely connected network of associates are
more likely to be recalled in extralist cued recall than targets with
sparsely connected sets (Nelson, Bennett, et al., 1993; Nelson et
al., 1998). Connectivity represents another source of target activa-
tion strength, and its effect is independent of resonance (Nelson et
al., 2003), word concreteness, frequency, or target set size (Gee,
Nelson, & Krawczyk, 1999; Nelson, Bennett, et al., 1993). Of
importance, the connectivity effect is reduced by manipulations
that reduce accessibility of the episodic context cues (Nelson,
Goodmon, & Akirmak, 2007; Nelson et al., 1998); therefore, we
expected reduced connectivity effects in the forget group.
Method.
Participants and design. Participants were 48 University of
North Carolina at Greensboro undergraduates who took part in the
experiment in exchange for course credit. None of them had
participated in the previous experiment. The experimental design
was a mixed-factorial design, with connectivity (high, low) ma-
nipulated within subjects and instruction (forget, remember) varied
between subjects.
Materials. We selected 32 words from the University of South
Florida Free Association Norms to serve as targets for two unre-
lated 16-item study lists. An additional 32 words served as test
cues, one per target item. Targets were unrelated to each other, and
cues were unrelated to each other. Each cue word was related only
to a single target. When target connectivity was high, each asso-
ciate in the target’s associative network was connected to an
average of 2.36 (SD "0.38) associates, with the sum of associate-
to-associate link strengths averaging 4.20 (SD "1.00). When
target connectivity was low, each associate was connected to an
average of only 0.70 (SD "0.14) associates, with the sum of
associate-to-associate link strengths averaging 1.50 (SD "0.28).
As in Experiment 1, for each cue–target pair in the high- and
low-connectivity conditions, we controlled the variables known to
affect recall probability. The controlled variables along with their
values are shown in Table 1.
Procedure. The procedures were identical to those used in
Experiment 1, except that the manipulated implicit variable was
target connectivity instead of resonance.
Results.
Directed forgetting costs. To analyze the costs of directed
forgetting, a Connectivity (high, low) !Instruction (forget, re-
member) mixed-factorial ANOVA was conducted on the propor-
tion of List 1 recall. The results are shown in Figure 5 (top panel).
There was a main effect of connectivity, F(1, 46) "116.94,
MSE "0.010, p#.001, $
2
".72, indicating better List 1 memory
for high-connectivity targets (.40) than low connectivity targets
(.18). There was also a main effect of instruction, F(1, 46) "
12.58, MSE "0.031, p#.01, $
2
".22, indicating that the
remember group remembered more List 1 items (.35) than the
forget group (.23). These main effects were qualified by a signif-
icant Instruction !Connectivity interaction, F(1, 46) "10.61,
MSE "0.010, p#.01, $
2
".19. The connectivity effect was
much larger in the remember condition (28%) t(23) "9.27, p#
.001, than in the forget condition (15%), t(23) "5.80, p#.001.
Directed forgetting benefits. To analyze the benefits of di-
rected forgetting, a similar analysis was conducted on the propor-
tion of List 2 recall. The results are shown in Figure 5 (bottom
panel). There was a main effect of connectivity, F(1, 46) "34.53,
MSE "0.022, p#.001, $
2
".43, indicating better List 2 memory
for high-connectivity items (.38) than low-connectivity items (.20).
There was no main effect of instruction (F#1) and no significant
interaction (F#1). In other words, there were no directed forget-
ting benefits regardless of the connectivity condition.
Discussion of Experiments 1 and 2
The first two experiments demonstrated that directed forgetting
impairment can be reliably obtained with the extralist cuing pro-
cedure. These experiments also confirmed that the manipulation of
two independent sources of target activation strength was success-
ful. There was an advantage for targets with high resonance in
Experiment 1 and an advantage for targets with high connectivity
in Experiment 2 in the remember conditions of the experiments. Of
importance, the magnitude of the resonance effect and of the
connectivity effect was substantially reduced by the directed for-
getting manipulation. Thus, an instructional variable (e.g., “forget
the words”) produced effects similar to those found in prior studies
with extralist cued recall, which manipulated the environmental
context or the delay interval. These results were predicted by the
contextual account of directed forgetting costs, and they are hard
to reconcile with the inhibitory account without additional assump-
tions.
In contrast to the reliable directed forgetting costs, we did not
obtain directed forgetting benefits. Although numerically the
Figure 4. Associative map of targets with high and low connectivity among their associates.
926 SAHAKYAN AND GOODMON
means were in the correct direction, the effect was not reliable in
either experiment. This null effect needs to be treated with caution,
and given the new methodology used in these studies, it needs to
be replicated.
Section 2: Cue–Target Intersection Variables
Whereas the previous two experiments manipulated implicit
variables that affect the target activation strength during encoding,
in the next three experiments we introduced implicit variables, the
effects of which emerged during the test stage. Namely, we ma-
nipulated three variables implicated in the cue–target intersection
process while controlling for the target activation levels. In other
words, the recall differences were expected to be driven by the
strength of the test cues rather than the targets. We varied target set
size (Experiment 3), direct target-to-cue strength (Experiment 4), and
indirect strength between the target and the cue (Experiment 5).
As described earlier, the two theories of directed forgetting
make opposite predictions for this set of experiments. The context
account predicts a smaller effect of the cue strength in the forget
condition than in the remember condition, because directed for-
getting disrupts the context, making the strong cues less effective.
In contrast, the inhibitory account predicts a larger effect of the cue
strength in the forget condition than in the remember condition,
because stronger cues can release the items from inhibition.
Experiment 3: Set Size
Targets that produce fewer associates are recalled better in
extralist cued recall than targets with larger associative sets (Nel-
son & Friedrich, 1980; Nelson et al., 1998; Nelson & Schreiber,
1992). The memorial advantage of small set size items is known as
the set size effect. Set size effects are obtained regardless of word
ambiguity (Gee, 1997), concreteness (Nelson & Schreiber, 1992),
or frequency (Nelson & Xu, 1995) and regardless of how well the
target has been processed explicitly (e.g., Nelson et al., 1992).
They are also independent of connectivity and resonance (Nelson
& Zhang, 2000).
Although set size is a property of the target, its effect is driven
completely by the selected test cue. In other words, it represents a
retrieval phenomenon that arises when one word is used to search
the memory for a related word. Set size effects are not found in
single item recognition, where the target is used as a cue for itself
(e.g., Fisher & Nelson, 2006), but they do have a robust effect in
extralist cued recall when the test cue is used to search for the
target.
The negative effect of set size is attributed to the increasing
number of competitors in the large networks (Schreiber, 1998;
Schreiber & Nelson, 1998). In other words, the set size effect is a
competitor effect in disguise, with increases in competitor strength
decreasing the probability of recall, because they lower the cue–
target intersection strength. (For the illustration of competitors, see
Figure 1.) Of importance, set size effects are reduced by manipu-
lations that block or disrupt access to contextual cues (Nelson,
McEvoy, et al., 1993; Nelson et al., 1998); therefore, we predicted
a smaller set size effect in the forget group compared with the
remember group.
Method.
Participants and design. Participants were 72 University of
North Carolina at Greensboro undergraduates who took part in the
study in exchange for course credit. None of them had participated
in previous experiments. The experimental design was a 2 !2
mixed-factorial design, with set size (small, large) manipulated
within subjects and instruction (forget, remember) manipulated
between subjects.
Materials. We selected 32 words from the University of South
Florida Free Association Norms to serve as targets for two unre-
lated 16-item study lists. An additional 32 words served as test
cues, one per target item. Targets in the small set condition had an
Table 1
Mean Strengths and Standard Deviations of Controlled Variables for Items in Experiments 2–5
Controlled variable
Experiment 2:
Connectivity
Experiment 3:
Target set size
Experiment 4:
Target-to-cue
strength
Experiment 5:
Shared associate
strength
M SD M SD M SD M SD
Direct target-to-cue strength 0.05 0.03 0.02 0.02
—
— 0.02 0.02
Direct cue-to-target strength 0.08 0.02 0.06 0.03 0.04 0.02 0.06 0.04
Shared associate strength 0.01 0.01 0.06 0.07 0.05 0.07 — —
Mediated strength 0.01 0.02 0.04 0.05 0.02 0.03 0.07 0.10
Target set size 15.58 3.45 — — 13.84 3.97 12.59 3.66
Target competitor strength 0.54 0.15 — — 0.46 0.20 0.53 0.14
Cue set size 15.50 5.94 13.28 5.36 14.47 4.34 10.59 2.41
Cue competitor strength 0.49 0.17 0.41 0.24 0.46 0.19 0.38 0.18
No. of associate-to-associate connections — — 1.36 0.65 1.36 0.55 1.41 0.59
Sum of associate-to-associate link strengths — — 2.37 0.84 2.57 0.80 2.68 1.23
No. of associates-to-target connections 0.44 0.23 0.49 0.22 0.41 0.19 0.49 0.25
Sum of associates-to-target link strengths 1.86 0.82 1.78 0.69 1.44 0.42 1.90 0.66
Target printed frequency (Kucˇera & Francis) 74.10 58.36 89.94 79.22 33.94 29.81 78.75 72.77
Target concreteness on a scale from 1 to 7 4.71 1.32 5.00 1.19 5.08 1.38 4.89 1.43
Cue printed frequency (Kucˇera & Francis) 50.95 59.28 38.63 45.94 72.59 95.86 53.63 92.19
Cue concreteness on a scale from 1 to 7 4.65 1.24 4.88 1.18 4.74 1.96 4.74 1.56
Note. Dashes indicate that a variable was manipulated, and the specific values can be found in the Methods section of corresponding experiments.
927
DIRECTED FORGETTING IN EXTRALIST CUING
average of 6.56 associates (SD "1.59) in their set, whereas targets
in the large set condition had an average of 19.88 associates (SD "
3.05). There were more competitors in the large set condition
(M"16.31, SD "3.89) than in the small set condition (M"4.25,
SD "1.77). Hence, competitor strength was higher in the large set,
averaging .64 (SD ".13), whereas in the small set it averaged .31
(SD ".28). The remaining variables were controlled across the set
size conditions (see Table 1).
Procedure. The procedure was identical to that of Experiment
1, except that the implicit variable manipulated within subjects
involved set size.
Results.
Directed forgetting costs. To analyze the costs of directed
forgetting, a Set Size (small, large) !Instruction (forget, remem-
ber) mixed-factorial ANOVA was conducted on proportion List 1
recall. The results are shown in Figure 6 (top panel). There was a
main effect of set size, F(1, 70) "41.80, MSE "0.020, p#.001,
$
2
".37, indicating better List 1 memory for targets in small sets
(.44) than large sets (.29). There was also a main effect of instruc-
tion, F(1, 70) "7.48, MSE "0.047, p#.01, $
2
".10, indicating
that the remember group recalled more List 1 items (.41) than the
forget group (.31). These effects were qualified by an Instruc-
tion !Set Size interaction, F(1, 70) "4.63, MSE "0.020, p#
.05, $
2
".06. The set size effect was larger in the remember
condition (20%), t(35) "5.70, p#.001, than in the forget
condition (10%), t(35) "3.29, p#.01.
Directed forgetting benefits. To analyze the benefits of di-
rected forgetting, a similar analysis was conducted on proportion
List 1 Costs
.00
.10
.20
.30
.40
.50
.60
puorg rebmemeRpuorg tegroF
Proportion List 1 Recall
high connectivity
low connectivity
List 2 Benefits
.00
.10
.20
.30
.40
.50
.60
puorg rebmemeRpuorg tegroF
Proportion List 2 Recall
high connectivity
low connectivity
Figure 5. List 1 recall (top) and List 2 recall (bottom) as a function of associative connectivity in Experiment
2. Error bars represent standard error of the means.
928 SAHAKYAN AND GOODMON
List 2 recall (see Figure 6, bottom panel). There was a main effect
of set size, F(1, 70) "66.05, MSE "0.026, p#.001, $
2
".49,
indicating better memory for targets in the small set condition (.53)
than the large set condition (.31). There was neither a main effect
of instruction nor an interaction (Fs!1), indicating no directed
forgetting benefits in either set size condition.
Experiment 4: Direct Target-to-Cue Strength
In this experiment, we investigated another variable that affects
the cue–target intersection by varying the direct target-to-cue
strength. The latter describes the probability that the target acti-
vates the test cue during learning. This link facilitates recall
because it provides direct access from the target to the test cue
(Tulving & Thomson, 1973). Prior research shows that associates
that are more strongly activated by the target during encoding are
more successful as retrieval cues for that target compared with
associates that are weakly activated by the target (Humphreys &
Galbraith, 1975; Nelson & Goodmon, 2003; Nelson & McEvoy,
1979). The effects of target-to-cue strength are not contingent on
how well the target is encoded explicitly (Nelson, Fisher, &
Akirmak, 2007; Nelson & Goodmon, 2002). Furthermore, these
effects are reduced by manipulations that block or disrupt access to
the contextual information encoded about the target during learn-
ing (Nelson & Goodmon, 2003; Nelson, Goodmon, & Ceo, 2007).
Therefore, we expected that directed forgetting would reduce the
target-to-cue strength effect.
Method.
Participants and design. Participants were 72 University of
North Carolina at Greensboro undergraduates who took part in the
experiment in exchange for extra course credit. None of the par-
ticipants took part in previous experiments. The experimental
design formed a 2 !2 mixed-factorial design, with target-to-cue
List 1 Costs
.00
.10
.20
.30
.40
.50
.60
.70
puorg rebmemeRpuorg tegroF
Proportion List 1 Recall
small set
large set
List 2 Benefits
.00
.10
.20
.30
.40
.50
.60
.70
puorg rebmemeRpuorg tegroF
Proportion List 2 Recall
small set
large set
Figure 6. List 1 recall (top) and List 2 recall (bottom) as a function of target set size in Experiment 3. Error
bars represent standard error of the means.
929
DIRECTED FORGETTING IN EXTRALIST CUING
strength (high, low) varied within subjects and instruction (forget,
remember) varied between subjects.
Materials. We selected 32 words from the University of
South Florida Free Association Norms to serve as targets for
two unrelated 16-item study lists. An additional 32 words
served as test cues, one per target item. Targets were unrelated
to each other, and cues were unrelated to each other. During the
test, participants received a mixture of high- and low-strength
cues, such that some targets were tested with strong cues,
whereas the remaining targets were tested with weak cues.
Strong cues had an average target-to-cue strength of .15 (SD "
.02), whereas weak cues had an average target-to-cue strength
of .04 (SD ".03). The remaining variables that could affect
cued recall were controlled across the cue strength conditions
(see Table 1).
Procedure. The procedure was identical to that of Experiment
1, except that the within-subjects implicit variable was the direct
target-to-cue strength.
Results.
Directed forgetting costs. To analyze the costs of directed for-
getting, a Target-to-Cue Strength (high, low) !Instruction (forget,
remember) mixed-factorial ANOVA was conducted on proportion
List 1 recall. The results are shown in Figure 7 (top panel). There was
amaineffectoftarget-to-cuestrength,F(1, 70) "18.97, MSE "
0.025, p#.001, $
2
".213, indicating that target recovery was better
when target-to-cue strength was high (.41) than when it was low (.30).
There was also a main effect of instruction, F(1, 70) "11.86, MSE "
0.042, p#.01, $
2
".060, indicating that the remember group
recalled more List 1 items (.42) than the forget group (.30). These
effects were qualified by a significant Instruction !Target-to-Cue
List 1 Costs
.00
.10
.20
.30
.40
.50
.60
puorg rebmemeRpuorg tegroF
Proportion List 1 Recall
high direct strength
low direct strength
List 2 Benefits
.00
.10
.20
.30
.40
.50
.60
puorg rebmemeRpuorg tegroF
Proportion List 2 Recall
high direct strength
low direct strength
Figure 7. List 1 recall (top) and List 2 recall (bottom) as a function of direct strength from the target to the test
cue in Experiment 4. Error bars represent standard error of the means.
930 SAHAKYAN AND GOODMON
Strength interaction, F(1, 70) "4.46, MSE "0.025, p#.05, $
2
"
.184. The magnitude of the target-to-cue strength effect was larger in
the remember group (17%), t(35) "4.92, p#.001, than in the forget
group (6%), t(35) "1.49, p".15.
Directed forgetting benefits. To analyze the benefits of di-
rected forgetting, the same analysis was conducted on proportion
List 2 recall (see Figure 7, bottom panel). There was a main effect
of target-to-cue strength, F(1, 70) "17.83, MSE "0.034, p#
.001, $
2
".203, indicating better recall of targets when target-to-
cue strength was high (.45) than when it was low (.32). There was
no main effect of instruction (F#1) and no significant interaction
(F#1), indicating no directed forgetting benefits in either of the
target-to-cue strength conditions.
Experiment 5: Shared Associate Strength
In this experiment, we manipulated the indirect strength between
the target and the cue. If the test cue and the target both independently
produce one or more common associates (e.g., CHORUS 3music 4
SONG,whereCHORUS is the target, SONG is the test cue, and music
is the associate that they share), then the probability of retrieving the
target during the test is greater when there are more shared associates
between the cue and the target than when there are fewer shared
associates between them (Nelson & Goodmon, 2002, 2003; Nelson et
al., 1998; Nelson & McEvoy, 2002). The indirect connections be-
tween the test cue and the target enhance recall by enhancing the
cue–target intersection strength (if not offset by the strength of com-
petitors). Their effects, however, are reduced by contextual disrup-
tions (Nelson & Goodmon, 2003); therefore, we expected reduced
effects of the shared associate strength in the forget condition com-
pared with the remember condition.
Method.
Participants and design. Participants were 72 University of
North Carolina undergraduates who took part in exchange for extra
course credit. None of them had participated in previous experi-
ments. The experimental design formed a 2 !2 mixed-factorial
design, with shared associate strength (high, low) varied within
subjects and cue (forget, remember) varied between subjects.
Materials. Thirty-two words were selected from the University
of South Florida Free Association Norms to serve as targets for two
unrelated 16-item study lists. An additional 32 words served as test
cues, one per target item. Targets were unrelated to each other, and
cues were unrelated to each other. During the test, participants re-
ceived a mixture of cues that had high and low shared associate
strength. When shared associate strength was high, there was an
average of 2.50 (SD "0.83) associates between the target and the cue,
with an average shared associate strength of .19 (SD ".08). When
shared associate strength was low, there was an average of 0.81
(SD "1.00) associates between the target and the cue, with an
average shared associate strength of .00 (SD ".01). The remaining
variables were controlled across the shared associate strength condi-
tions (see Table 1).
Procedure. The procedure was identical that of to Experiment
1, except that the within-subjects implicit variable was shared
associate strength.
Results.
Directed forgetting costs. To analyze the costs of directed
forgetting, a Shared Associate Strength (high, low) !Instruction
(forget, remember) mixed-factorial ANOVA was conducted on the
proportion of List 1 recall. The results are shown in Figure 8 (top
panel). There was a main effect of shared associate strength, F(1,
70) "67.53, MSE "0.014, p#.001, $
2
".49, indicating better
List 1 memory for targets that were tested with cues that had high
shared associate strength with the targets (.39) than for those tested
with cues that had low shared associate strength (.23). There was
also a significant main effect of instruction, F(1, 70) "4.60,
MSE "0.042, p#.05, $
2
".06, indicating that the remember
group recalled more List 1 items (.35) than the forget group (.27).
These effects were qualified by an Instruction !Shared Associate
Strength interaction, F(1, 70) "8.85, MSE "0.014, p#.01, $
2
"
.11. As predicted, the shared associate strength effect was larger in
the remember group (22%), t(70) "8.57, p#.001, than in the
forget group (11%), t(70) "3.46, p#.01.
Directed forgetting benefits. To analyze the benefits of di-
rected forgetting, similar analyses were performed on the propor-
tion of List 2 recall (see Figure 8, bottom panel). There was a main
effect of shared associate strength, F(1, 70) "44.55, MSE "
0.020, p#.001, $
2
".39, indicating better List 2 recall of targets
that were tested with high shared associate strength cues (.38) than
for those tested with low shared associate strength cues (.23).
There was no main effect of instruction, F(1, 70) "1.36, p".25,
and no significant interaction, F(1, 70) "1.01, p"30. In other
words, there were no directed forgetting benefits in either condi-
tion of the shared associate strength.
General Discussion
Across five experiments, we found that implicitly activated
associates had systematic effects on memory. Targets with high
resonance from their associates, high connectivity amongst their
associates, and smaller associative neighborhoods were better re-
called than targets with low resonance, low connectivity, or larger
sets. Also, cues that had stronger direct or indirect associations
with the targets were more successful at retrieving the targets than
cues that had weaker associative links with the target. These
findings confirm the results of prior research demonstrating that
the associates of the target, albeit not consciously experienced,
impact memory in systematic ways. In addition, we found that
implicitly activated associates influence directed forgetting. Spe-
cifically, the forget cue reduced the effect of target resonance and
connectivity as well as the effects of target set size, direct target-
to-cue strength, and indirect strength between the target and the
cue. In other words, the memory advantage produced by stronger
associative links of every type was reduced by directed forgetting.
This was obtained regardless of whether the source of the memory
advantage was driven by the implicit target activation strength
(Experiments 1 and 2) or by the cue–target intersection strength
(Experiments 3–5).
These findings were predicted by the context account of directed
forgetting, and they were motivated by prior research that inves-
tigated how disruptions of episodic context influence performance
in extralist cued recall. According to Nelson’s model that explains
those findings, the extralist cue combines together with the context
cue to elicit the episodically primed target (e.g., Nelson, Goodmon,
& Ceo, 2007). Providing a specific extralist cue for the target does
not diminish the importance of recovering the contextual informa-
tion encoded about that target. If context cues are unavailable
during the test, they diminish the power of the extralist cue to
931
DIRECTED FORGETTING IN EXTRALIST CUING
differentiate the target from other activated associates of the cue,
and the extralist cue loses its ability to restore the former target
activation levels. Because the context account uses the mental
context change taking place between the two lists to explain
directed forgetting impairment, it predicts that effects of the forget
cue on implicit variable effects will be similar to those of other
forms of disruption of context (e.g., Nelson, Bennett, et al., 1993;
Nelson & Goodmon, 2003; Nelson, Goodmon, & Akirmak, 2007
Nelson, McEvoy, et al., 1993; Nelson et al., 1998). Importantly,
both the context account and Nelson’s model made a strong
prediction that directed forgetting would interact with all types of
implicit variables, regardless of whether they were implicated in
the target activation or in the cue–target intersection process.
Indeed, implicit variables interacted with directed forgetting in all
experiments.
These findings support the contextual account of directed for-
getting costs, and they are problematic for the inhibitory account.
It is unclear, for example, why in Experiments 1 and 2, the strongly
activated targets were inhibited more than the weakly activated
targets. To explain these findings, an item-level inhibitory mech-
anism is needed, as in the retrieval-induced forgetting or think/no-
think research (e.g., M. C. Anderson et al., 1994; M. C. Anderson
& Green, 2001), along with some untested assumptions. In
retrieval-induced forgetting studies, for example, high-frequency
exemplars of a category were inhibited more than low-frequency
exemplars, presumably because they caused more competition
during the retrieval practice stage (e.g., M. C. Anderson et al.,
1994). Therefore, it needs to be assumed that strongly activated
targets across the lists caused more competition with each other
and thus were inhibited to a greater extent than weakly activated
targets. We do not know of any empirical evidence that suggests
this is the case. Furthermore, a related line of work from our lab
has indicated that higher episodic strength by itself is insufficient
to cause variable degrees of directed forgetting across the strong
List 1 Costs
.00
.10
.20
.30
.40
.50
puorg rebmemeRpuorg tegroF
Proportion List 1 Recall
high indirect strength
low indirect strength
List 2 Benefits
.00
.10
.20
.30
.40
.50
puorg rebmemeRpuorg tegroF
Proportion List 1 Recall
high indirect strength
low indirect strength
Figure 8. List 1 recall (top) and List 2 recall (bottom) as a function of indirect shared strength between the
target and the test cue in Experiment 5. Error bars represent standard error of the means.
932 SAHAKYAN AND GOODMON
and weak items, unless that strength refers to the contextual
strength in the memory trace rather than the item strength (e.g.,
Sahakyan, Delaney, & Waldum, 2008). Thus, it seems unlikely
that higher implicit strength of the items could cause greater
competition and lead to greater inhibition.
Even more problematic for the inhibitory view are the findings
of Experiments 3–5, because those results were driven entirely by
the test cue. In fact, as the cue effectiveness was manipulated in
those experiments, the target characteristics were fully controlled.
Nevertheless, there was greater impairment in the conditions
where targets were tested with stronger cues than in those tested
with weaker cues. Because inhibition must be aimed at targets
before any test cues are presented, it is unclear why inhibition
differentially impaired some cue–target pairs more than others. If
anything, given the assumption of the release from inhibition, one
would expect that strong cues would be more effective at releasing
the targets from inhibition and would lead to less forgetting rather
than more forgetting in those conditions. To reconcile these find-
ings with the inhibition view, one might try to argue that when the
target was inhibited, its entire implicit representation was inhib-
ited, and because of the association with the target, the cues were
inhibited along with the target, which would reduce their effec-
tiveness during the test. However, even if the cues were inhibited
with the targets, one would expect them to be released from
inhibition to their full strength when they were presented during
the test. Moreover, this still does not explain why stronger cue–
target relationships suffered more from inhibition than weak ones.
Of interest, further examination of the results at the level of
individual pairs in Experiments 3 through 5 confirmed the inter-
actions of cue strength with directed forgetting for pairs where the
cues were members of the target’s set as well as for pairs where the
cues were not members of the target’s set. Note that for any
cue–target pair, the cue may be a member of the target’s set, but it
may also be any associate outside of the target’s network that has
a direct or indirect association with the target. We used the
target-to-cue strength of .000 as a definition that the cue was not
a member of the target’s set. These association values show the
probability of the target activating the cue in the absence of a study
trial. With this definition in mind, we discovered that 40% of the
cues used in Experiments 3 through 5 were not members of the
target’s set. Nevertheless, regardless of the cue status, they all
produced interactive effects in directed forgetting, F(1, 184) "
4.90, p#.05 (see Figure 9). Overall, these results are inconsistent
with the inhibitory explanation.
Some researchers have proposed that the context-change ac-
count is not inconsistent with the retrieval inhibition account if one
assumes that mental context change induced by the forget cue is
accomplished by inhibiting the unwanted context (e.g., M. C.
Anderson, 2005). Others have argued that directed forgetting is
caused by inhibition that induces some form of contextual isolation
of List 1 items (e.g., Ba¨uml, Hanslmayr, Pastotter, & Klimesch,
2008). In other words, both of these proposals attribute the recall
impairment to diminished access to the List 1 context along with
a tacit assumption about the importance of context in retrieval—
claims directly made by the context-change account. However,
unlike the context-change account, which makes no claims about
the inhibition of context and suggests that the two lists are simply
contextually segregated, these interpretations of the inhibition ac-
count suggest that List 1 context is inhibited. To the best of our
knowledge, there is no empirical evidence demonstrating the in-
hibition of context in directed forgetting studies, and in the absence
of such evidence, the context-change account and the inhibition-
of-context account are virtually indistinguishable.
Finally, although the earlier proposed selective rehearsal expla-
nation of directed forgetting (e.g., R. A. Bjork, 1970, 1972) has
been dismissed because directed forgetting emerges in incidental
.00
.05
.10
.15
.20
.25
.30
.35
.40
.45
.50
.55
Forget Remember Forget Remember
cue is a member of the target's set cue is NOT a member of the target's set
Proportion Recalled
implicit advantage
implicit disadvantage
Figure 9. List 1 recall in the remember and the forget groups as a function of implicit variable level (recall
advantage vs. disadvantage) and status of the test cue (member vs. nonmember of the target set) in Experiments
3 through 5.
933
DIRECTED FORGETTING IN EXTRALIST CUING
learning as well (e.g., Geiselman et al., 1983; Sahakyan &
Delaney, 2005; Sahakyan et al., 2008), some researchers have
suggested that it needs to be entertained as a candidate explanation
in intentional learning situations (e.g., Benjamin, 2006; Sheard &
MacLeod, 2005). We therefore consider our findings from the
perspective of the selective rehearsal account, according to which
the forget cue encourages participants to terminate List 1 rehearsal
and therefore devote extra rehearsal to List 2 items, leading to the
costs and benefits, respectively. In all of our experiments, the
selective rehearsal account would predict equivalent directed for-
getting for items with an implicit advantage and items with an
implicit disadvantage. This is because terminating rehearsal should
be more detrimental to the explicit encoding strength of the items
rather than to their implicit strength, and explicit strength does not
interact with implicit variables (Nelson, Bennett, et al., 1993;
Nelson et al., 1997; Nelson, Fisher, & Akirmak, 2007; Nelson &
Goodmon, 2002; Nelson, McEvoy, et al., 1993; Nelson et al.,
1990, 1992). Furthermore, in some experiments, we introduced the
implicit variable during testing (Experiments 3–5), whereas in
other experiments, we introduced the implicit variable during
learning (Experiments 1–2). The rehearsal explanation would pre-
dict different results for variables introduced at testing as opposed
to learning, yet they all produced the same effect. Therefore, the
results are inconsistent with the selective rehearsal account.
Caveats
Despite obtaining directed forgetting impairment in List 1 recall
with extralist cuing, we did not observe directed forgetting en-
hancement in List 2 recall in any of the experiments. The lack of
directed forgetting benefits is inconsistent with all existing ac-
counts of directed forgetting, including the dual-factor accounts,
which invoke encoding-based mechanisms to explain the benefits
(e.g., Ba¨uml et al., 2008; Sahakyan & Delaney, 2005). When we
combined the data across all the experiments, we observed the
benefits in List 2 recall, although the effect was modest in mag-
nitude, F(1, 326) "3.86, p".05. Unlike the directed forgetting
costs, the benefits did not interact with implicit condition in the
combined analyses (F#1). The lack of benefits seemed to be a
reliable finding because it emerged repeatedly in all of the exper-
iments. It could be driven by the new testing procedure used in the
experiments for as yet undiscovered reasons. In general, the num-
ber of reports where the costs and the benefits were not observed
together is growing in the literature (e.g., Benjamin, 2006; Con-
way, Harries, Noyes, Racsma’ny, & Frankish, 2000; Delaney &
Sahakyan, 2007; Macrae, Bodenhausen, Milne, & Ford, 1997;
Minnema & Knowlton, 2008; Pastötter & Ba¨uml, 2010; Sahakyan
& Delaney, 2003; Sahakyan & Foster, 2009; Sahakyan & Good-
mon, 2007; Sahakyan et al., 2009; Whetstone, Cross, & Whet-
stone, 1996; Zellner & Ba¨uml, 2006). Our understanding of the
factors that produce these dissociations is incomplete, and this
inconsistency should become a priority for future research. New
research from our lab (Sahakyan & Delaney, 2010) has suggested
that even when the recall rates do not signify directed forgetting
benefits, there are other diagnostic markers that imply the benefits
(e.g., reduced intrusions). Thus, in future investigations, research-
ers should consider variables beyond the recall rates to infer about
the presence or absence of directed forgetting benefits.
Another finding that requires further attention is the absence of
directed forgetting for items with an implicit disadvantage. In all
experiments, we found greater effects of directed forgetting for
items with implicit advantage than for those with disadvantage as
predicted by the context account. However, directed forgetting
impairment was significant in the strong conditions of the exper-
iments, but it was not significant in the weak conditions of the
experiments, although numerically it was in the right direction.
When all experiments are combined together, there was significant
forgetting in both weak, t(334) "2.34, p#.05, and strong
conditions, t(334) "7.83, p#.001, although the effect is not quite
robust in the weak condition. This was not necessarily a predicted
finding; nevertheless, it seems reliable given that it was found in
all experiments. The null effect of directed forgetting on weak
items is reminiscent of findings of prior research on contextual
disruptions and set size, which affect the small set items but not the
large set items (Nelson et al., 1998). Why this is the case is
currently unknown and needs further investigation.
In all experiments, we predicted and observed significant inter-
actions between implicit variables and directed forgetting. Because
this interaction is critical for our predictions, one could criticize it
by invoking the scaling argument (e.g., Loftus, 1978). Because the
recall rates of items in the implicit advantage (IA) and implicit
disadvantage (ID) conditions were not the same in the remember
condition, the IA items had more room to suffer from directed
forgetting than ID items. In other words, the interaction may
reflect a measurement problem caused by the underlying function
relating implicit strength to the probability of recall. To address
this potential concern, we identified the IA and ID items that were
recalled at approximately the same level in the remember condi-
tion of each experiment and examined the recall of the same items
in the forget condition. We determined the median recall values of
IA and ID items in each experiment (by pooling the data over
participants in the remember condition) and selected the bottom
half of the IA items and the top half of the ID items. This selection
approximately equated the recall of IA and ID items in the remem-
ber condition. (Note that the items were still different normatively
on the relevant implicit characteristics.) Next, we evaluated the
recall of this subset of items with a factorial ANOVA, using cue
(forget vs. remember) and implicit condition (IA vs. ID). The
results revealed a significant interaction, F(1, 156) "4.36, p#.05
(see the top portion of Table 2). For the analysis, the data from all
experiments were combined because individual experiments did
not provide enough items to capture a potential interaction. How-
ever, the pattern was present numerically in all the experiments
(see the bottom portion of Table 2). This analysis suggests that the
interaction of directed forgetting with implicit condition was
driven by the factors underlying the manipulation of implicit
strength.
Conclusions
The results of five experiments demonstrated that directed for-
getting can be successfully obtained with extralist cued recall. This
is an important first step toward investigating directed forgetting
with methods other than free recall and recognition. The extralist
cuing procedure not only allowed testing of TBF items with
unstudied cues but also enabled independent manipulation of the
characteristics of the TBF items and the characteristics of the test
934 SAHAKYAN AND GOODMON
cues. This approach allowed us to contrast the leading theories of
directed forgetting. The results of these five experiments under-
score that the mechanism behind directed forgetting is contextual
rather than inhibitory in nature.
References
Anderson, J. R. (1983). The architecture of cognition. Cambridge, MA:
Harvard University Press.
Anderson, M. C. (2005). The role of inhibitory control in forgetting
unwanted memories: A consideration of three methods. In C. MacLeod
& B. Uttl (Eds.), Dynamic cognitive processes (pp. 159 –190). Tokyo,
Japan: Springer-Verlag.
Anderson, M. C. (2009). Motivated forgetting. In A. Baddeley, M. W.
Eysenck, & M. C. Anderson (Eds.), Memory (pp. 217–243). London,
England: Psychology Press.
Anderson, M. C., Bjork, R. A., & Bjork, E. L. (1994). Remembering can
cause forgetting: Retrieval dynamics in long-term memory. Journal of
Experimental Psychology: Learning, Memory, and Cognition, 20, 1063–
1087.
Anderson, M. C., & Green, C. (2001). Suppressing unwanted memories by
executive control. Nature, 410, 131–134.
Anderson, M. C., & Neely, J. H. (1996). Interference and inhibition in
memory retrieval. In E. L. Bjork & R. A. Bjork (Eds.), Memory:
Handbook of perception and cognition (2nd ed., pp. 237–313). San
Diego, CA: Academic Press.
Bahrick, H. P. (1970). Two-phase model for prompted recall. Psycholog-
ical Review, 77, 215–222.
Basden, B. H., Basden, D. R., & Gargano, G. J. (1993). Directed forgetting
in implicit and explicit memory tests: A comparison of methods. Journal
of Experimental Psychology: Learning, Memory, and Cognition, 19,
603– 616.
Ba¨uml, K.-H., Hanslmayr, S., Pastotter, B., & Klimesch, W. (2008).
Oscillatory correlates of intentional updating in episodic memory. Neu-
roImage, 41, 596 – 604.
Benjamin, A. S. (2006). The effects of list-method directed forgetting on
recognition memory. Psychonomic Bulletin and Review, 13, 831– 836.
Bilodeau, E. A., & Howell, D. C. (1968). Stimulated recall of idiosyncratic
and cultural free associates. Journal of Verbal Learning and Verbal
Behavior, 7, 348 –350.
Bjork, E. L., & Bjork, R. A. (1996). Continuing influences of to-be-
forgotten information. Consciousness and Cognition: An International
Journal, 5, 176 –196.
Bjork, E. L., & Bjork, R. A. (2003). Intentional forgetting can increase, not
decrease, residual influences of to-be-forgotten information. Journal of
Experimental Psychology: Learning, Memory, and Cognition, 29, 524 –
531.
Bjork, E. L., Bjork, R. A., & Anderson, M. A. (1998). Varieties of
goal-directed forgetting. In J. M. Golding & C. M. MacLeod (Eds.),
Intentional forgetting: Interdisciplinary approaches (pp. 103–137).
Mahwah, NJ: Erlbaum.
Bjork, R. A. (1970). Positive forgetting: The noninterference of items
intentionally forgotten. Journal of Verbal Learning and Verbal Behav-
ior, 9, 255–268.
Bjork, R. A. (1972). Theoretical implications of directed forgetting. In
A. W. Melton & E. Martin (Eds.), Coding processes in human memory
(pp. 217–235). Washington, DC: Winston.
Bjork, R. A. (1989). Retrieval inhibition as an adaptive mechanism in
human memory. In H. L. Roediger III & F. I. M. Craik (Eds.), Varieties
of memory and consciousness: Essays in honour of Endel Tulving (pp.
309 –330). Hillsdale, NJ: Erlbaum.
Bjork, R. A., LaBerge, D., & LeGrand, R. (1968). The modification of
short-term memory through instructions to forget. Psychonomic Science,
10, 55–56.
Conway, M. A., Harries, K., Noyes, J., Racsma’ny, M., & Frankish, C. R.
(2000). The disruption and dissolution of directed forgetting: Inhibitory
control of memory. Journal of Memory and Language, 43, 409 – 430.
Deese, J. (1959). On the prediction of occurrence of particular verbal
intrusions in immediate recall. Journal of Experimental Psychology, 58,
17–22.
Deese, J. (1965). The structure of associations in language and thought.
Baltimore, MD: Johns Hopkins University Press.
Delaney, P. F., & Sahakyan, L. (2007). Effects of working memory
capacity on context maintenance: Adding individual differences to the
contextual-change account of directed forgetting. Memory & Cognition,
35, 1074 –1082.
Fisher, S. L., & Nelson, D. L. (2006). Recursive reminding: Effects of
repetition, printed frequency, connectivity, and set size on recognition
and judgments of frequency. Memory & Cognition, 34, 295–306.
Fox, P. W. (1968). Recall and misrecall as a function of cultural and
Table 2
Proportion Recall of the Subset of Items Included in the
Item-Level Analyses as a Function of Cue and Implicit
Condition Across Experiments
Study, implicit condition, and cue M SE
Overall across all experiments
Advantage
Remember .34 .02
Forget .24 .02
Disadvantage
Remember .33 .02
Forget .30 .02
Experiment 1: Resonance
Advantage
Remember .29 .02
Forget .20 .04
Disadvantage
Remember .31 .02
Forget .28 .02
Experiment 2: Connectivity
Advantage
Remember .34 .02
Forget .24 .04
Disadvantage
Remember .34 .03
Forget .31 .03
Experiment 3: Set size
Advantage
Remember .34 .03
Forget .25 .04
Disadvantage
Remember .30 .02
Forget .31 .05
Experiment 4: Target-to-cue strength
Advantage
Remember .36 .04
Forget .25 .06
Disadvantage
Remember .38 .02
Forget .32 .05
Experiment 5: Shared associate strength
Advantage
Remember .36 .06
Forget .26 .03
Disadvantage
Remember .33 .02
Forget .28 .05
935
DIRECTED FORGETTING IN EXTRALIST CUING
individual word association habits and regulation of the recall environ-
ment. Journal of Verbal Learning and Verbal Behavior, 7, 632– 637.
Gee, N. R. (1997). Implicit memory and word ambiguity. Journal of
Memory and Language, 36, 253–275.
Gee, N. R., Nelson, D. L., & Krawczyk, D. (1999). Is the concreteness
effect a result of underlying network interconnectivity? Journal of
Memory and Language, 40, 479 – 497.
Geiselman, R. E., Bjork, R. A., & Fishman, D. L. (1983). Disrupted
retrieval in directed forgetting: A link with posthypnotic amnesia. Jour-
nal of Experimental Psychology: General, 112, 58 –72.
Goodmon, L. B., & Nelson, D. L. (2004). Strengthening the activation of
unconsciously activated memories. Memory & Cognition, 32, 804 – 818.
Humphreys, M. S., & Galbraith, R. C. (1975). Forward and backward
associations in cued recall: Predictions from the encoding specificity
principle. Journal of Experimental Psychology: Human Learning and
Memory, 1, 702–710.
Johnson, H. (1994). Processes of successful intentional forgetting. Psycho-
logical Bulletin, 116, 274 –292.
Kintsch, W. (1988). The role of knowledge in discourse comprehension: A
construction–integration model. Psychological Review, 95, 163–182.
Kucˇera, H., & Francis, W. (1967). Computational analysis of present-day
American English. Providence, RI: Brown University Press.
Lehman, M., & Malmberg, K. J. (2009). A global theory of remembering
and forgetting from multiple lists. Journal of Experimental Psychology:
Learning, Memory, and Cognition, 35, 970 –988.
Loftus, G. R. (1978). On interpretation of interactions. Memory & Cogni-
tion, 6, 312–319.
MacLeod, C. M. (1998). Directed forgetting. In J. M. Golding & C. M.
MacLeod (Eds.), Intentional forgetting: Interdisciplinary approaches
(pp. 1–57). Mahwah, NJ: Erlbaum.
MacLeod, C. M. (1999). The item and list methods of directed forgetting:
Test differences and the role of demand characteristics. Psychonomic
Bulletin and Review, 6, 123–129.
Macrae, C. N., Bodenhausen, G. V., Milne, A. B., & Ford, R. L. (1997). On
regulation of recollection: The intentional forgetting of stereotypical
memories. Journal of Personality and Social Psychology, 72, 709 –719.
McEvoy, C. L., Nelson, D. L., & Komatsu, T. (1999). What is the
connection between true and false memories? The differential roles of
interitem associations in recall and recognition. Journal of Experimental
Psychology: Learning, Memory, and Cognition, 25, 1177–1194.
Minnema, M. T., & Knowlton, B. J. (2008). Directed forgetting of emo-
tional words. Emotion, 8, 643– 652.
Nelson, D. L., Bennett, D. J., Gee, N. R., Schreiber, T. A., & McKinney,
V. (1993). Implicit memory: Effects of network size and interconnec-
tivity on cued recall. Journal of Experimental Psychology: Learning,
Memory, and Cognition, 19, 747–764.
Nelson, D. L., Bennett, D. J., & Leibert, T. W. (1997). One step is not
enough: Making better use of association norms to predict cued recall.
Memory & Cognition, 25, 785–796.
Nelson, D. L., Dyrdal, G. M., & Goodmon, L. B. (2005). What is preex-
isting strength? Predicting free association probabilities, similarity rat-
ings, and cued recall probabilities. Psychonomic Bulletin & Review, 12,
711–719.
Nelson, D. L., Fisher, S. L., & Akirmak, U. (2007). How implicitly
activated and explicitly acquired knowledge contribute to the effective-
ness of retrieval cues. Memory & Cognition, 35, 1892–1904.
Nelson, D. L., & Friedrich, M. A. (1980). Encoding and cuing sounds and
senses. Journal of Experimental Psychology: Human Learning and
Memory, 6, 717–731.
Nelson, D. L., & Goodmon, L. B. (2002). Experiencing a word can prime
its accessibility and its associative connections to related words. Memory
& Cognition, 30, 380 –398.
Nelson, D. L., & Goodmon, L. B. (2003). Disrupting attention: The need
for retrieval cues in working memory theories. Memory & Cognition, 31,
65–76.
Nelson, D. L., Goodmon, L. B., & Akirmak, U. (2007). Implicitly activated
memories get connected to context just like episodic memories. Memory
& Cognition, 35, 1878 –1891.
Nelson, D. L., Goodmon, L. B., & Ceo, D. (2007). How does delayed
testing reduce effects of implicit memory: Context infusion or cuing
with context? Memory & Cognition, 35, 1014 –1023.
Nelson, D. L., & McEvoy, C. L. (1979). Encoding context and set size.
Journal of Experimental Psychology: Human Learning and Memory, 5,
292–314.
Nelson, D. L., & McEvoy, C. L. (2002). How can the same type of prior
knowledge both help and hinder recall? Journal of Memory and Lan-
guage, 46, 652– 663.
Nelson, D. L., & McEvoy, C. L. (2005). Implicitly activated memories:
The missing links of remembering. In C. Izawa & N. Ohta (Eds.),
Learning and memory: Advances in theory and applications (pp. 177–
198). Mahwah, NJ: Erlbaum.
Nelson, D. L., McEvoy, C. L., & Dennis, S. (2000). What is free associ-
ation and what does it measure? Memory & Cognition, 28, 887– 899.
Nelson, D. L., McEvoy, C. L., Janczura, G. A., & Xu, J. (1993). Implicit
memory and inhibition. Journal of Memory and Language, 32, 667– 691.
Nelson, D. L., McEvoy, C. L., & Pointer, L. (2003). Spreading activation
or spooky action at a distance? Journal of Experimental Psychology:
Learning, Memory and Cognition, 29, 42–52.
Nelson, D. L., McEvoy, C. L., & Schreiber, T. A. (1990). Encoding context
and retrieval conditions as determinants of the effects of natural category
size. Journal of Experimental Psychology: Learning, Memory, and Cog-
nition, 16, 34 – 41.
Nelson, D. L., McEvoy, C. L., & Schreiber, T. A. (1999). The University
of South Florida word association, rhyme and word fragment norms.
Retrieved from http://www.usf.edu/FreeAssociation/
Nelson, D. L., McKinney, V. M., Gee, N. R., & Janczura, G. A. (1998).
Interpreting the influence of implicitly activated memories on recall and
recognition. Psychological Review, 105, 299 –324.
Nelson, D. L., & Schreiber, T. A. (1992). Word concreteness and word
structure as independent determinants of recall. Journal of Memory and
Language, 31, 237–260.
Nelson, D. L., Schreiber, T. A., & McEvoy, C. L. (1992). Processing
implicit and explicit representations. Psychological Review, 99, 322–
348.
Nelson, D. L., & Xu, J. (1995). Effects of implicit memory on explicit
recall: Set size and word frequency effects. Psychological Research, 57,
203–214.
Nelson, D. L., & Zhang, N. (2000). The ties that bind what is known to the
recall of what is new. Psychonomic Bulletin and Review, 7, 604 – 617.
Nelson, D. L., Zhang, N., & McKinney, V. M. (2001). The ties that bind
what is known to the recognition of what is new. Journal of Experimen-
tal Psychology: Learning, Memory, and Cognition, 27, 1147–1159.
Pastötter, B., & Ba¨uml, K.-H. (2010). Amount of postcue encoding predicts
amount of directed forgetting. Journal of Experimental Psychology:
Learning, Memory, and Cognition, 36, 54 – 65.
Postman, L. (1971). Transfer, interference, and forgetting. In J. W. Kling
& L. A. Riggs (Eds.), Woodworth and Schlosberg’s experimental psy-
chology (3rd ed., pp. 1019 –1132). New York, NY: Holt, Rinehart, &
Winston.
Reich, R. R., & Goldman, M. S. (2005). Exploring the alcohol expectancy
memory network: The utility of free associates. Psychology of Addictive
Behaviors, 19, 317–325.
Sahakyan, L., & Delaney, P. F. (2003). Can encoding differences explain
the benefits of directed forgetting in the list-method paradigm? Journal
of Memory and Language, 48, 195–201.
Sahakyan, L., & Delaney, P. F. (2005). Directed forgetting in incidental
learning and recognition testing: Support for a two-factor account.
936 SAHAKYAN AND GOODMON
Journal of Experimental Psychology: Learning, Memory, and Cogni-
tion, 31, 789 – 801.
Sahakyan, L., & Delaney, P. F. (in press). Item-specific encoding produces an
additional benefit of directed forgetting: Evidence from intrusion errors.
Journal of Experimental Psychology: Learning, Memory, and Cognition.
Sahakyan, L., Delaney, P. F., & Waldum, E. R. (2008). Intentional forget-
ting is easier after two “shots” than after one. Journal of Experimental
Psychology: Learning, Memory, and Cognition, 34, 408 – 414.
Sahakyan, L., & Foster, N. L. (2009). Intentional forgetting of actions:
Comparison of list-method and item-method directed forgetting. Journal
of Memory and Language, 61, 134 –152.
Sahakyan, L., & Goodmon, L. B. (2007). The influence of directional
associations on directed forgetting and interference. Journal of Experi-
mental Psychology: Learning, Memory, and Cognition, 33, 1035–1049.
Sahakyan, L., & Kelley, C. (2002). A contextual change account of the
directed forgetting effect. Journal of Experimental Psychology: Learn-
ing, Memory, and Cognition, 28, 1064 –1072.
Sahakyan, L., Waldum, E. R., Benjamin, A. S., & Bickett, S. P. (2009).
Where is the forgetting with list-method directed forgetting in recogni-
tion? Memory & Cognition, 37, 464 – 476.
Schreiber, T. A. (1998). Effects of target set size on feelings of knowing
and cued recall: Implications for the cue effectiveness and partial-
retrieval hypotheses. Memory & Cognition, 26, 553–571.
Schreiber, T. A., & Nelson, D. L. (1998). The relation between feelings of
knowing and the number of neighboring concepts linked to the test cue.
Memory & Cognition, 26, 869 – 884.
Sheard, E. D., & MacLeod, C. M. (2005). List method directed forgetting:
Return of the selective rehearsal account. In N. Ohta, C. M. MacLeod,
& B. Uttl (Eds.), Dynamic cognitive processes (pp. 219 –248). Tokyo,
Japan: Springer-Verlag.
Simpson, P., & Voss, J. F. (1967). Stability of response hierarchies.
Journal of Experimental Psychology, 7, 170 –174.
Tulving, E., & Thomson, D. M. (1973). Encoding specificity and retrieval
processes in episodic memory. Psychological Review, 80, 352–373.
Whetstone, T., Cross, M. D., & Whetstone, L. M. (1996). Inhibition,
contextual segregation, and subject strategies in list method directed
forgetting. Consciousness and Cognition, 5, 395– 417.
Zellner, M., & Ba¨uml, K. H. (2006). Inhibitory deficits in older adults:
List-method directed forgetting revisited. Journal of Experimental Psy-
chology: Learning, Memory, and Cognition, 32, 290 –300.
Received September 26, 2009
Revision received February 8, 2010
Accepted February 17, 2010 !
937
DIRECTED FORGETTING IN EXTRALIST CUING
A preview of this full-text is provided by American Psychological Association.
Content available from Journal of Experimental Psychology Learning Memory and Cognition
This content is subject to copyright. Terms and conditions apply.
