Content uploaded by Chelsea K Quinlan
Author content
All content in this area was uploaded by Chelsea K Quinlan
Content may be subject to copyright.
Content uploaded by Sherry H Stewart
Author content
All content in this area was uploaded by Sherry H Stewart
Content may be subject to copyright.
ORIGINAL ARTICLE
The Impact of Attention Style on Directed Forgetting
Among High Anxiety Sensitive Individuals
Melanie Noel •Tracy L. Taylor •Chelsea K. Quinlan •
Sherry H. Stewart
!Springer Science+Business Media, LLC 2011
Abstract Results of research investigating the link
between anxiety sensitivity (AS) and memory biases
toward threat have been inconsistent. There may be sub-
groups of high AS individuals who differ in their preferred
mode of attending to threat-related information, thereby
impacting memory. The impact of individual attention
style on intentional forgetting of words varying in emo-
tional valence was examined among individuals with
varying levels of AS. By incorporating an inhibition of
return (IOR) task (to yield a proxy of attentional allocation)
within the study phase of the item-method directed for-
getting paradigm, we categorized high, moderate, and low
AS individuals according to their attention style in
response to threat stimuli: ‘threat attenders’ (small IOR
effect) and ‘threat avoiders’ (large IOR effect). Among
high AS individuals only, ‘threat avoiders’ showed greater
intentional forgetting of threat-related words than ‘threat
attenders’. High AS ‘threat avoiders’ also had higher levels
of anxiety-related psychopathology (AS and health anxi-
ety) than high AS ‘threat attenders’.
Keywords Anxiety sensitivity !Directed forgetting !
Memory !Attention
Introduction
Anxiety sensitivity (AS) is an individual difference vari-
able wherein individuals show differential tendencies to
misinterpret bodily symptoms as signs of impending
physical, psychological, or social calamity (Reiss et al.
1986; Taylor 1999). For example, an individual high in AS
may fear that heart palpitations are symptomatic of an
imminent heart attack; fear that difficulties concentrating
signify a mental illness; and/or fear trembling in anticipa-
tion of social rejection or ridicule by others (Reiss 1991).
Not surprisingly, research has consistently revealed a link
between AS and anxiety disorders (Cox et al. 1999). In
fact, there is compelling evidence for the role of AS as a
risk factor for the development and maintenance of panic
disorder (Cox et al. 1999; Ehlers 1995; Maller and Reiss
1992; Schmidt et al. 1997) as well as hypochondriasis (Otto
et al. 1998; Stewart and Watt 2000).
To further elucidate the role of AS as a cognitive risk
factor for the development of anxiety disorders, research
has investigated cognitive biases including those involving
attention and memory, among high AS individuals prior to
the onset of anxious psychopathology (Keogh et al. 2001;
Lees et al. 2005; McCabe 1999; McNally et al. 1999a;
Stewart et al. 1998). Generally, cognitive theories suggest
the presence of biases toward anxiety-relevant information
that contribute to both the development and maintenance of
anxiety disorders (McNally 1995). Given that cognitive
biases might precede the development of anxious psycho-
pathology, identifying the nature of these biases among
high AS individuals is critical.
There has been some evidence to suggest that both
individuals with panic disorder (e.g., Becker et al. 1994;
Cloitre and Liebowitz 1991; Cloitre et al. 1994; McNally
et al. 1989), and those with high levels of AS (McCabe
M. Noel (&)!T. L. Taylor !C. K. Quinlan !S. H. Stewart
Department of Psychology, Dalhousie University,
Halifax, NS B3H 4J1, Canada
e-mail: melanie.noel@dal.ca
S. H. Stewart
Departments of Psychiatry, Psychology, and Community Health
and Epidemiology, Dalhousie University, Halifax, NS, Canada
123
Cogn Ther Res
DOI 10.1007/s10608-011-9366-y
1999), display a proclivity to selectively remember threat-
related stimuli. Nevertheless, there are several studies
demonstrating that these individuals show no memory bias
to selectively remember threat-related stimuli (e.g., Otto
et al. 1994; Pickles and van den Broek 1988; Rapee 1994).
Despite these inconsistent findings for remembering, some
researchers have investigated the forgetting—particularly,
the intentional forgetting, of threat-related information
among high AS individuals.
Intentional forgetting can be studied in the laboratory
using the item-method directed forgetting paradigm (see
Golding and MacLeod 1998 for a review). In this para-
digm, participants are presented with a series of items, one
at a time, each followed with equal probability by an
instruction to remember or an instruction to forget. Fol-
lowing the presentation of all study items, memory per-
formance is tested using either recall or recognition.
Regardless of the memory test employed, typically a
directed forgetting effect occurs, which is defined as
greater memory performance for remember-cued items
compared to forget-cued items. Importantly, the directed
forgetting effect cannot be explained by demand charac-
teristics: If participants are offered monetary compensation
for the accurate report of additional forget-cued words,
memory performance remains unchanged (MacLeod
1999).
McNally et al. (1999b) used the item-method directed
forgetting paradigm to determine whether individuals with
panic disorder would show a greater directed forgetting
effect (and thus better intentional forgetting) for threat-
related words compared to positive and neutral words. In
this study, the authors suggested that forget instructions
would converge with any tendencies to cognitively avoid
threat-related information among patients with panic dis-
order. As such, they predicted that compared to normal
control participants, panic disordered patients would recall
fewer threat words (compared to positive and neutral
words) that they were instructed to forget. Neither group
(the panic patients nor the controls) was expected to show
an effect of word valence when instructed to remember.
Thus, the prediction was for a larger directed forgetting
effect for threat compared to positive and neutral words
and only in the patients with panic disorder. Contrary to
this prediction, however, the results did not reveal a larger
directed forgetting effect for threat words among individ-
uals with panic disorder; instead, the directed forgetting
effect occurred equally for all word types and was similar
for both panic disordered patients as well as control
participants.
A few explanations have been proposed to elucidate
why some researchers have found a memory bias toward
threat among high AS individuals, whereas others have not
(see McNally et al. 1999b; Teachman 2005). Nevertheless,
these explanations fail to fully account for the mixed
findings in the literature on AS and memory biases (both
remembering and forgetting). As a result, researchers
have highlighted the importance of identifying individual
difference variables that might vary within groups of high
AS individuals to influence memory for threat (McNally
1999). One potential individual difference variable that
might vary within high AS groups is individual attention
style. On the one hand, AS is thought to increase one’s
alertness to anxiety-provoking information (Reiss and
McNally 1985). Indeed, attentional biases favoring the
processing of threatening information have been found
among individuals with panic disorder (Asmundson et al.
1992; Beck et al. 1992; Ehlers et al. 1988; McNally et al.
1992) and those with high levels of AS (Hunt et al. 2006;
Keogh et al. 2001; Koven et al. 2003; Stewart et al.
1998). On the other hand, AS is thought to increase one’s
avoidance of anxiety-provoking information. For example,
AS has been found to be associated with agoraphobic
avoidance among individuals with panic disorder
(McNally and Lorenz 1987; Reiss et al. 1986, White et al.
2006) and emotional avoidance among individuals with
no known history of psychopathology (Stewart et al.
2002; Zvolensky and Forsyth 2002). Similarly, AS has
been shown to prospectively predict behavioral avoidance
among non-clinical samples of adolescents (Wilson and
Hayward 2006).
Because some studies have shown AS to be associated
with attentional biases toward threat, whereas others have
shown AS to be associated with avoidance of threat, there
may be subgroups of high AS individuals who vary in their
preferred (and perhaps automatic) mode of attending to
threatening information. Such a distinction has been made
among individuals with hypochondriasis, such that some
individuals avoid any reminder of illness, whereas other
individuals engage in repetitive hypervigilant behaviours
such as frequently visiting doctors’ offices or continual
body checking (APA 2000; Stewart and Watt 2001). In a
recent meta-analysis on memory bias for threat-related
stimuli in anxiety and anxiety disorders (including panic
disorder), Mitte (2008) concluded that while some high-
anxious individuals show greater memory performance and
processing of threatening information, other individuals
show lower memory performance and avoidant processing
of threatening information.
Similarly, subgroups of anxious individuals who either
selectively attend to, or avoid, threatening information
have previously been identified. For example, in his model
outlining the processes of attention orientation in the
presence of threat, Krohne (1993) distinguished between
groups of high-anxious individuals based on their defensive
coping styles: ‘‘sensitizers’’ (high vigilance, low cognitive
avoidance) and ‘‘repressors’’ (low vigilance, high cognitive
Cogn Ther Res
123
avoidance).
1
Krohne (1993) posited that these individual
coping modes would result in distinct attentional biases
that are elicited when an individual is confronted with
threat cues. Whereas vigilance is characterized by the
intensified processing and intake of threat-related stimuli,
cognitive avoidance is characterized by the diversion of
attention away from threat-related stimuli. As such, the
model holds that individuals differ in the degree to which
these attention styles are activated in anxiety-evoking sit-
uations. One way researchers have empirically investigated
the links between anxiety and attentional biases is by cat-
egorizing individual coping modes using self-report mea-
sures (e.g., the Mainz Coping Inventory; MCI; Krohne
1989; Krohne et al. 2000). However, given that attentional
biases may be relatively automatic
2
(i.e., involuntary) and
outside of the individual’s awareness (Mathews 1993), the
utility of self-report measures (in contrast to cognitive
tasks) to characterize individual attention styles is ques-
tionable (for a discussion, see De Houwer 2006). Further-
more, although Krohne’s (1993) model was developed to
explain individual differences in coping with trait and state
anxiety, similar individual differences could exist among
high AS individuals.
By highlighting the importance of identifying individ-
uals who engage in preferential versus avoidant processing
of threatening information, this could potentially account
for the inconsistent findings on memory biases and AS.
Because it has been suggested that anxiety-related differ-
ences during encoding may contribute to memory biases in
individuals with anxiety (Mathews 1993), differences in
attention style (the tendency to attend vs. avoid) are of
particular importance. Furthermore, the link between
attention and memory is well established such that memory
is improved by attending more closely to information,
whereas, memory is impaired by withdrawing attention
from information (Cowan 1995). Although the role of
attention as an influential factor in the relationship between
anxiety and memory is at the core of many theories used to
account for memory biases in anxiety (e.g., Beck et al.
1985; Mogg and Bradley 1998; Eysenck et al. 2007),
measures of attention are rarely included in conjunction
with memory tasks (although see McNally et al. 1999a;
Teachman 2005; Teachman et al. 2007). Moreover, to our
knowledge, a study using a cognitive task to identify
individuals who attend to or avoid threatening information
in the context of a memory task has yet to be conducted.
The current study used the methods of Taylor (2005)
and embedded an inhibition of return (IOR) task into the
study phase of the item-method directed forgetting para-
digm. This paradigm was presented to high, moderate, and
low AS individuals. The IOR task provided a proxy for
identifying different attention styles as they were mani-
fested during the instantiation of instructions to remember
and forget; ‘threat avoiders’ and ‘threat attenders’ were
defined operationally based on their IOR scores. IOR refers
to slowed responding to onset targets that appear at a
previously cued peripheral location as opposed to a novel
location (Posner and Cohen 1984). Although IOR can co-
occur with attention (e.g., Berlucchi et al. 2000; Chica
2008; Chica et al. 2006; Lupia
´n
˜ez et al. 2004) it is normally
revealed when attention is withdrawn from the peripherally
cued location (Danziger and Kingstone 1999). If attention
continues to dwell at the peripherally cued location, the
resulting concurrent facilitatory effect (Dorris et al. 2002;
Klein 2000; Ro and Rafal 1999; Tipper et al. 1997) can
mask the IOR effect (e.g., Danziger and Kingstone 1999).
In this way, the occurrence of an IOR effect can be used to
infer the withdrawal of attention from a spatial location.
Because IOR can provide an index of the dwell and/or
withdrawal of attention, it was used in the current study to
differentiate between high AS individuals whose attention
tends to dwell on threatening information (‘threat attend-
ers’; as evidenced by a low or reduced IOR effect fol-
lowing presentation of threat words) and those whose
attention tends to withdraw from threatening information
(‘threat avoiders’; as evidenced by a high or enhanced IOR
effect following presentation of threat words).
Given that there may be individual differences in
attention style among high AS individuals, and attention is
thought to influence the relationship between anxiety and
memory, we hypothesized that there would be differences
between high AS attention style subgroups in their ability
to intentionally forget threat-related stimuli. More specifi-
cally, high AS ‘threat avoiders’ would display a greater
magnitude directed forgetting effect for threat words as
compared to high AS ‘threat attenders’. It was expected
that this difference in the magnitude of the directed for-
getting effect as a function of attention style would be
specific to threat words and would not be evident for
positive and neutral words. Also, because there are no
known differences in attention style among individuals
with moderate or low AS levels, no effect of attention style
on the directed forgetting effect for any of the word types
(threat, positive, neutral) was expected among individuals
with moderate or low AS levels. In addition, similar to
McNally et al. (1999b), we hypothesized no subgroup
differences in recognition of any remember-cued words.
1
In addition to ‘‘sensitizers’’ and ‘‘repressors’’, Krohne’s conceptu-
alization also allowed individuals to be categorized as being high
(‘‘high anxiety’’) or low (‘‘non-defensives’’) on both vigilance and
avoidance, respectively.
2
The processing of threat cues in pathological anxiety has been
shown to be automatic in the sense that it occurs involuntarily, but not
automatic in the sense that it is not capacity free. See McNally (1995)
for a discussion on automaticity and anxiety disorders.
Cogn Ther Res
123
While we presumed that high AS ‘threat avoiders’ have a
preference for avoiding threat-related information, we had
no reason to believe that this tendency would be suffi-
ciently strong to override the intention to remember in this
non-clinical sample. A tendency to avoid threat-related
information would enable ‘threat avoiders’ to easily dis-
regard threat-related words when instructed to forget, but
would not necessarily prevent them from committing
threat-related words to memory when instructed to
remember.
Additionally, we predicted that the two high AS atten-
tion style subgroups would differ on anxiety-related mea-
sures, given that cognitive biases (McNally 1995) and in
particular, avoidance biases (Buller et al. 1986; Ehlers
1995) are thought to lead to the development and mainte-
nance of anxious psychopathology among individuals with
panic disorder. Specifically, we hypothesized that high AS
individuals who exhibited an avoidance of threat words
would have significantly higher levels of AS and health
anxiety, as well as be more likely to report a history of
panic attacks, than high AS individuals who did not exhibit
an avoidance of threat words. In other words, the two high
AS attention style subgroups would differ in their level of
anxiety-related psychopathology. No subgroup differences
on anxiety-related variables were expected among the
moderate or low AS groups.
Method
Participants
Participants were 82 (60 Females, 22 Males; M
age
=
20.73 years, SD =4.62 years) undergraduate students who
volunteered to participate in exchange for credit towards
their grade in an eligible Psychology class at Dalhousie
University.
Participants completed the Anxiety Sensitivity Index
(ASI; Peterson and Reiss 1992) as part of an online mass
screening. This yielded an ASI score that was used to
allocate individuals into one of three AS groups (high,
moderate, and low). To maximize the likelihood that high
and low AS groups differed from one another, extreme
groups as opposed to median splits were used to determine
group membership. High and low AS groups were based on
one standard deviation above (Females: [26.8; Males:
[23.3) or below (Females: \8.0; Males: \5.9) the gender-
specific published ASI means for Dalhousie University
undergraduate students (Stewart et al. 1997). The moderate
AS group consisted of individuals with ASI scores between
.5 SD above (Females: [22.1; Males: [18.95) and below
(Females: \12.7; Males: \10.25) these gender-specific
published ASI means. Only those individuals with ASI
scores within these ranges were invited to take part in the
present study. This method of AS group allocation has been
used in previous research (Shostak and Peterson 1990) and
gender specific means were used in light of previous
research that revealed gender differences in ASI scores
(Stewart et al. 1997). Finally, although AS studies do not
always include moderate AS groups in addition to high and
low AS groups, the present study included a moderate AS
group in light of research suggesting that low AS indi-
viduals are an extreme group that may behave differently
than average AS individuals (Shostak and Peterson 1990).
Given the novelty of this research, low and moderate AS
groups were included to demonstrate that attention style
subgroup differences were unique to high AS individuals.
All participants reported having normal or corrected-to-
normal vision; no gross hearing, visual, or motor impair-
ments; the physical ability to use a computer keyboard; and
a good understanding of the English language. Given that
attentional subgroup differences in panic history were
investigated, individuals who reported having ever expe-
rienced panic attacks were included. The experiment was
conducted in one session lasting less than 1 h.
Measures
Demographic Questionnaire
Information about the participant’s age, gender, level of
education, ethnicity, and visual and hearing impairments
was collected using a demographic questionnaire created
by the primary investigator.
Anxiety Sensitivity Index (ASI; Peterson and Reiss 1992)
The ASI consists of 16 items that assess beliefs that anxiety
sensations are associated with catastrophic physical, psy-
chological, or social consequences. The respondent rates
his/her level of agreement/disagreement with each item on
a five-point Likert scale (0 =very little; 4 =very much).
The ASI has excellent reliability and validity for both
clinical and non-clinical populations (Peterson and Reiss
1992).
Panic Attack Questionnaire-Revised
(PAQ-R; Cox et al. 1992)
The PAQ-R assesses history of panic attacks as defined in
the Diagnostic and Statistical Manual of Mental Disorders,
3rd edition—Revised (American Psychiatric Association
[APA] 1987). Participants were provided with a definition
of a panic attack according to the DSM-III-R (APA 1987)
definition, and indicated whether they had experienced one
or more panic attacks. Panic history (whether or not
Cogn Ther Res
123
participants endorsed having ever experienced a panic
attack) was used as the measure of panic history in the
present study.
Health Anxiety Questionnaire (HAQ; Lucock and Morley
1996)
The HAQ is a 21-item measure of health anxiety and ill-
ness beliefs, and it reflects enduring features that are con-
sistent with a cognitive-behavioural model of health
anxiety. Responses reflect how often respondents experi-
ence each symptom of health anxiety (e.g., worry and
health preoccupation, fear of illness and death, reassur-
ance-seeking behaviour, and the extent to which symptoms
interfere with an individual’s life). Responses are rated on
a four-point Likert scale (0 =not at all or rarely; 3 =most
of the time). The HAQ has good internal consistency and
appropriate discriminate validity (Lucock and Morley
1996).
Stimuli and Apparatus
PsyScope 5.1.2 (Cohen et al. 1993) was used to run the
experiment on a G4-400 Macintosh computer equipped
with a 17’’ Macintosh Studio Display Colour monitor or a
17’’ ViewSonic PT775 monitor, a standard Macintosh
Universal Serial Bus keyboard, and Sony MDR-XD100
stereo headphones. Three rectangular stimulus boxes
drawn in a one-point thick line were aligned along the
horizontal axis of the computer screen. From a viewing
distance of 57 cm, each box measured 5.5"of visual angle
along the horizontal axis and 3.0"of visual angle along the
vertical axis. The middle box was centered on the monitor;
the peripheral boxes were separated from the centre box by
6.2"of visual angle, measured centre-to-centre. The fixa-
tion and target stimuli consisted of a black solid-fill circle,
0.5"in diameter that was centred in the middle box or else
one of the peripheral boxes, respectively. The memory
instruction was a high- (1,170 Hz) or low-frequency
(260 Hz) tone that was played through both channels of the
stereo headphones. Study words were presented in size-24
Arial font, centered in one of the peripheral stimulus boxes.
At recognition, words were also presented in size-24 Arial
font. All stimuli were black, presented on a uniform white
background.
Word lists were created from lists that have been used in
previous research investigating cognitive biases among
individuals with high AS (McCabe 1999; McNally et al.
1999a; Stewart et al. 1998). This resulted in three word lists
that differed in emotional valence and included threat,
positive and neutral words (Appendix). Prior to each test-
ing session, custom software was used to randomly dis-
tribute the words without replacement from each word list
(threat, positive, neutral) to eight study lists (n=3) and
one foil list (n=24)
3
. All word lists were balanced for
word frequency (Carroll et al. 1971). Although threat
words did not differ from positive (t(136) =-1.97,
p[.05) or neutral words (t(146) =1.87, p[.05) in
terms of length, neutral words were greater in length than
positive words (t(156) =4.02, p\.001). Given that
positive and neutral words only served as control stimuli
for the threat words in the current study, this difference in
word length was not deemed to be an issue.
Procedure
After providing their written informed consent, participants
received verbal instructions from the experimenter that
were reiterated onscreen at the start of the experiment.
Participants were instructed that a word would appear with
equal probability to the left or right stimulus box and that
after the word disappeared, they would hear a tone
instructing them to Remember that word for a later memory
test or instructing them that they could Forget that word.
For half of the participants, the high tone served as the
instruction to Remember and the low tone served as the
instruction to Forget, whereas this designation was
reversed for the other half of the participants.
Following the memory instruction on each trial, partic-
ipants were told that a dot would appear with equal prob-
ability to the left or the right. They were asked to rest their
left index finger on the ‘f’ key and their right index finger
on the ‘j’ key and to press the corresponding key to indicate
the target location. They were asked to make their
responses as quickly, but also as accurately, as possible.
Tone Familiarization
Participants received 10 tone familiarization trials prior to
beginning the experiment. On these trials, a verbal
descriptor of the tone designation was presented for
3,000 ms at the centre of the computer monitor (e.g., ‘High
tone—Remember’). The corresponding tone was played for
400 ms, beginning 2,000 ms after the first appearance of
the verbal descriptor. Equal numbers of low- and high-
frequency tones were presented randomly in these famil-
iarization trials.
Study Phase
Following the tone familiarization trials, participants
completed the demographic questionnaire on the computer.
Participants then began the experiment. Each trial started
with the appearance of the three stimulus boxes. After
3
n= cell size.
Cogn Ther Res
123
500 ms, the fixation dot appeared in the middle stimulus
box and remained visible throughout the duration of the
trial. Following an 800 ms delay, a word was presented for
400 ms, with equal probability in the left or right stimulus
box. Beginning 600 ms after the disappearance of the
peripheral word, the tone that served as the memory
instruction was played over the headphones for 400 ms. At
200 ms after the end of the tone, a target dot appeared with
equal probability in the left or right stimulus box and
remained on the screen for 400 ms. Following the pre-
sentation of the target dot, participants were allowed up to
1,000 ms to make a speeded button-press to identify the
location of the target. At the end of each trial, there was a
200 ms inter-trial interval.
The study lists created from each valenced word list
(Threat, Positive, Neutral) were each associated with one
cell of the 2 (Word location: Left,Right)92 (Memory
instruction: Remember,Forget)92 (Target location: Left,
Right) design. With three trials on each of these eight study
lists, this made for a total of 24 study trials per word type
and thus, a grand total of 72 study trials. For the purpose of
analyses, the factors of word and target location were
collapsed into Same word-target location (Left–Left,Right–
Right) and Different word-target location (Left–Right,
Right–Left).
Recognition Phase
The recognition phase occurred immediately after the final
trial in the study phase. Instructions explaining the task
appeared at the top of the computer screen and remained
visible throughout all of the recognition trials. Words were
drawn randomly without replacement from the valenced
study lists and from the valenced foil lists. This made for a
total of 144 words (72 study; 72 foil) that were presented
one at a time on the computer monitor. Participants were
asked to press the ‘y’ key if they recognized the word from
any of the study trials—regardless of whether a Remember
or a Forget instruction had been presented, and to press the
‘n’ key if they did not recognize the word from any of the
study trials. Keyboard responses appeared within the bor-
ders of a 6-point outline rectangle and could be self-cor-
rected until submitted by a press of the space bar. The task
was self-paced.
Upon completion of the recognition phase, participants
completed the self-report measures in the following order:
ASI, PAQ-R, and HAQ. The ASI was completed at 2 time
points (prescreening and post-test) to confirm the stability
of the construct over time and to ensure that group mem-
bership was accurate. Following measure completion,
participants were fully debriefed and given information
about mental health services that they could access in their
community, if needed.
Results
Data Preparation
For the IOR task, if participants failed to execute a correct
response within 80–1,000 ms of target onset, mislocalized
the target, pressed a key other than ‘f’ or ‘j’, or made more
than one response, the trial was considered an error and
excluded from the analysis of reaction times (RTs). Across
all participants in all conditions, an average of only 2% of
trials were thereby excluded.
Using only those target trials on which a correct local-
ization response was made within 80–1,000 ms of target
onset, the IOR effect (Same RTs–Different RTs) was cal-
culated separately for each word type (Threat, Neutral,
Positive) and for each memory instruction (Remember,
Forget). Then, this value was collapsed across memory
instruction (Remember, Forget) to produce a mean IOR
value for each word type. This provided an IOR value
which could serve as a relative measure of each groups’
overall tendency to dwell on (small IOR) or withdraw
(large IOR) their attention from each word type (Threat,
Positive, Neutral). It also provided an empirically derived
behavioral definition of individuals’ relative tendency to
attend to or avoid threat-related information in the context
of the memory task in which performance was measured.
For the recognition memory task, prior to conducting the
primary analyses, false alarm rates for the foil words were
analysed in one-way analyses of variance (ANOVAs) with
word type (Threat, Positive, Neutral) as a within-subjects
factor for each of the three AS groups (High, Moderate,
Low). These analyses revealed that for each AS group,
there was a significant effect of word type (high AS: F(2,
50) =24.63, p\.001; moderate AS: F(2, 62) =46.18,
p\.001; low AS: F(2, 40) =12.45, p\.001). Specifi-
cally, participants in each AS group responded ‘yes’ to
threat foils significantly more than to positive foils
[(high AS: (t(26) =4.12, p\.001); moderate AS:
(t(32) =5.06, p\.001); low AS: (t(21) =2.22,
p\.05)] and neutral foils [(high AS: (t(26) =6.56,
p\.001); moderate AS: (t(32) =9.63, p\.001; low AS:
(t(21) =5.78, p\.001)]. Similarly, participants in each
AS group responded ‘yes’ to positive foils significantly
more than to neutral foils [(high AS: (t(26) =3.09,
p\.01); moderate AS: (t(32) =4.66, p\.001); low AS:
(t(21) =2.45, p\.05)].
4
There were no significant dif-
ferences between attention style subgroups (‘threat
attenders’ vs. ‘threat avoiders’) on the number of ‘yes’
responses to foils of any valenced word type (high AS:
4
Given that the positive and neutral word lists differed in overall
word length, it is possible that differences in recognition for positive
and neutral foils were secondary to differences in overall word length.
Cogn Ther Res
123
F(2, 50) =.84, p[.05; moderate AS: F(2, 62) =.58,
p[.05; low AS: F(2, 40) =1.70, p[.05).
To take into account the differential false alarm rates by
valence, hit rates (i.e., the percentage of ‘yes’ responses to
Remember and Forget words of each valence) were cor-
rected for their respective false alarm rates (i.e., the per-
centage of ‘yes’ responses to words that were never
presented during the study trials [foils]; see the discussion
of corrections for guessing in Baddeley 2004). In other
words, the false alarm rate for each word type (Threat,
Positive, Neutral) and memory instruction (Remember,
Forget) was subtracted from the hit rate for each word type
(Threat, Positive, Neutral) and memory instruction
(Remember, Forget) resulting in six corrected hit rates.
Mean false alarm rates, hit rates, and corrected hit rates as
a function of AS group (High, Moderate, Low), attention
style subgroup (‘Threat-avoiders’, ‘Threat-attenders’) and
memory instruction (Remember, Forget) are shown in
Table 1.
Given that all hypotheses pertained to differences
between attention style subgroups for the high AS group,
all analyses were conducted separately for each AS group.
Thus, analyses conducted with the moderate and low AS
groups served as control analyses.
Attention Style Subgroups
High, moderate, and low AS individuals were categorized
into one of two attention style subgroups based on the
median split of each group’s IOR value for threat words
only.
5
That is, for the high AS group, individuals who
scored below their group-specific median for IOR threat
(Mdn =22.33, SD =61.45) were categorized as ‘threat
attenders’, whereas those who scored above the median
were categorized as ‘threat avoiders’. For the moderate AS
group, individuals who scored below their group specific
median for IOR threat (Mdn =29.50, SD =52.46) were
categorized as ‘threat attenders’, whereas those who scored
above the median were categorized as ‘threat avoiders’.
Similarly, for the low AS group, individuals who scored
below their group specific median for IOR threat
(Mdn =-3.00, SD =54.17) were categorized as ‘threat
attenders’, whereas those who scored above the median
were categorized as ‘threat avoiders’. Thus, for each AS
Table 1 Means (and SDs) for hit rates and corrected hit rates as a function of AS group, attention style subgroup, word type and memory instruction
Recognition uncorrected scores Recognition corrected scores for guessing
Threat Positive Neutral Threat Positive Neutral
RFFoil RFFoil RFFoil RFFoil RFFoil RFFoil
High AS
Avoiders 88.68
(12.24)
58.35
(20.40)
27.30
(16.80)
63.65
(15.94)
47.04
(17.60)
15.00
(11.13)
66.33
(19.84)
40.97
(26.46)
7.71
(7.38)
61.37
(15.11)
31.05
(20.23)
27.30
(16.80)
48.65
(16.68)
32.04
(20.81)
15.00
(11.13)
58.62
(20.00)
33.26
(26.37)
7.71
(7.38)
Attenders 80.16
(15.47)
66.63
(11.43)
24.22
(10.33)
69.16
(21.40)
45.54
(18.43)
17.12
(9.99)
75.81
(17.15)
48.89
(19.72)
10.21
(14.40)
55.94
(14.05)
42.41
(11.45)
24.22
(10.33)
52.04
(20.76)
28.42
(15.97)
17.12
(9.99)
65.60
(17.58)
38.67
(14.70)
10.21
(14.40)
Moderate AS
Avoiders 80.06
(12.47)
68.49
(19.90)
27.74
(14.47)
70.94
(15.81)
54.78
(21.02)
16.33
(12.75)
72.80
(20.73)
43.71
(17.76)
8.91
(10.18)
52.32
(22.48)
40.74
(22.59)
27.74
(14.47)
54.61
(16.50)
38.46
(20.35)
16.33
(12.75)
63.89
(26.06)
34.80
(18.28)
8.91
(10.18)
Attenders 80.04
(13.90)
66.32
(21.86)
27.82
(10.10)
64.44
(21.00)
41.47
(12.34)
13.67
(7.35)
68.03
(16.50)
36.38
(18.83)
4.22
(5.72)
52.22
(20.34)
38.50
(18.38)
27.82
(10.10)
50.77
(21.16)
27.80
(14.35)
13.67
(7.35)
63.81
(15.64)
32.16
(18.78)
4.22
(5.72)
Low AS
Avoiders 68.81
(27.03)
57.30
(23.63)
17.92
(15.93)
62.32
(21.60)
56.17
(19.54)
14.93
(10.74)
73.35
(21.93)
47.52
(20.37)
9.79
(7.36)
50.89
(25.98)
39.38
(26.38)
17.92
(15.93)
47.39
(15.95)
41.25
(13.25)
14.93
(10.74)
63.56
(23.05)
37.73
(21.58)
9.79
(7.36)
Attenders 81.14
(22.01)
60.49
(22.07)
29.05
(15.15)
68.94
(24.46)
50.41
(27.92)
18.69
(16.61)
70.47
(19.76)
46.86
(27.94)
12.59
(13.51)
52.09
(19.00)
31.45
(17.80)
29.05
(15.15)
50.25
(23.17)
31.72
(28.06)
18.69
(16.61)
57.88
(20.18)
34.27
(28.66)
12.59
(13.51)
RRemember instruction, FForget instruction
5
Analyses were also conducted using attention style subgroups that
were categorized based on the overall median split of the IOR value
for threat words for all AS groups combined (Mdn =19.125,
SD =57.52). Overall, the results remained the same. However, low
AS ‘threat attenders’ (n=14) had a greater magnitude of the directed
forgetting effect for threat words than low AS ‘threat avoiders’
(n=8) (M=21.38, SD =11.72; M=6.80, SD =17.14, respec-
tively; t(20) =2.37, p\.05) which could have been due to the small
number of low AS ‘threat avoiders’.
Cogn Ther Res
123
group, two attention style subgroups (‘threat attenders’ and
‘threat avoiders’) were created and entered as a between
subjects variable in all subsequent analyses. Preliminary
analyses revealed that the AS groups (High, Moderate,
Low) and attention style subgroups did not differ on
demographic variables (age, gender, years of education).
However, high AS ‘threat avoiders’ included more females
(100% vs. 57.1%) and fewer males (0% vs. 42.9%) than
high AS ‘threat attenders’ (X
corr
2
(1, N=27) =4.90,
p\.05).
Directed Forgetting
The magnitude of the directed forgetting effect was cal-
culated by taking the difference in the percentage of ‘y’
(yes) responses to Remember words and Forget words
(minus false alarm rate) of each valence. We decomposed
the complete 3 (AS group: High, Moderate, Low)92
(attention style: ‘Threat attenders’, ‘Threat avoiders’)92
(memory instruction: Remember, Forget) table of corrected
recognition hits (i.e., hits minus false alarms) into a series
of a priori planned comparisons (cf. Birch et al. 2008),
which allowed us to examine the most relevant compari-
sons with conventional alpha levels (Tabachnick and Fidell
2001). Specifically, we examined differences in the mag-
nitude of the directed forgetting effect between attention
style subgroups for each AS group separately, with inde-
pendent samples ttests. Given that we hypothesized a pri-
ori that high AS ‘threat avoiders’ would forget a greater
number of threat words than high AS ‘threat attenders’, we
conducted a directional one-tailed independent samples
ttest on corrected recognition hits (i.e., hits minus false
alarms) for forget-cued words. Given that we hypothesized
a priori that high AS subgroups would not differ on threat
words that were remember-cued, we conducted a non-
directional two-tailed independent samples ttest on cor-
rected recognition hits (i.e., hits minus false alarms) for
remember-cued words.
6
Similarly, given that no differ-
ences in the directed forgetting effect for any word type
were expected between attention style subgroups for
moderate and low AS groups, non-directional two-tailed
independent samples ttests were conducted on the mag-
nitude of the directed forgetting effect for these groups.
High AS
As shown in Fig. 1panel a, as expected, the magnitude of
the directed forgetting effect for threat words was signifi-
cantly greater for high AS ‘threat avoiders’ (M=30.33,
SD =15.07) than ‘threat attenders’ (M=13.52,
SD =12.45; t(25) =3.17, p\.01, g
p
2
=.29). There were
no significant differences between high AS ‘threat avoid-
ers’ and ‘threat attenders’ on the magnitude of the directed
forgetting effect for positive (M=16.61, SD =15.90;
M=23.61, SD =18.42, respectively; t(25) =-1.05,
p[.05, g
p
2
=.04) or neutral (M=25.36, SD =10.44;
M=26.93, SD =18.30, respectively; t(25) =-.27,
p[.05, g
p
2
=.00) words.
7
In line with our hypotheses, high AS ‘threat avoiders’
correctly recognized fewer threat words (M=31.05,
SD =20.23) that they were instructed to forget than high AS
‘threat attenders’ (M=42.41, SD =11.45; t(25) =-1.81,
p\.05, g
p
2
=.12). As expected, there was not a significant
difference between ‘threat avoiders’ (M=61.37, SD =
15.11) and ‘threat attenders’ (M=55.94, SD =14.05) on
their recognition of threat words that they were instructed to
remember (t(25) =.97, p[.05, g
p
2
=.04).
Moderate and Low AS
As expected, there were no significant differences between
moderate AS ‘threat attenders’ and ‘threat avoiders’ on the
magnitude of the directed forgetting effect for threat
(t(31) =.25, p[.05, g
p
2
=.00), positive (t(31) =.89,
p[.05, g
p
2
=.03), or neutral (t(31) =.30, p[.05,
g
p
2
=.00) words. Similarly, there were no significant dif-
ferences between low AS ‘threat attenders’ and ‘threat
avoiders’ on the magnitude of the directed forgetting effect
for threat (t(20) =1.43, p[.05, g
p
2
=.09), positive
(t(20) =1.92, p[.05, g
p
2
=.16), or neutral (t(20) =
-.33, p[.05, g
p
2
=.01) words. Because there were no
effects of attention style on the magnitude of the directed
forgetting effect in these AS groups, differences between
attention style subgroups for remember- versus forget-cued
threat words were not investigated.
Attention Style and Anxious Psychopathology
The means (and standard deviations) for anxiety-related
variables for the total sample as a function of AS group
(High,Moderate,Low) and attentional style subgroup
(‘Threat Attender’, ‘Threat Avoider’) are shown in Table 2.
Given that we hypothesized a priori that high AS ‘threat
avoiders’ would have higher levels of AS and health anx-
iety than high AS ‘threat attenders’, we conducted direc-
tional one-tailed independent samples ttests on their ASI
and HAQ scores.
8
Given that we hypothesized a priori that
6
Conclusions did not differ when directional one-tailed ttests were
conducted.
7
ANCOVA analyses revealed that the overall directed forgetting
results remain the same when ASI scores, HAQ scores and gender are
co-varied out.
8
Conclusions did not differ when non-directional two-tailed ttests
were conducted.
Cogn Ther Res
123
there would be no differences between attention style
subgroups for the moderate and low AS groups on AS or
health anxiety, non-directional two-tailed independent
samples ttests were conducted on ASI and HAQ scores for
these groups.
9
High AS
As expected, analyses revealed that high AS ‘threat
avoiders’ had significantly higher levels of AS
(t(25) =2.85, p\.01, g
p
2
=.25) and health anxiety
(t(24) =2.33, p\.05, g
p
2
=.18) than high AS ‘threat
attenders’. A greater proportion of high AS ‘threat avoid-
ers’ (61.5%) reported a history of panic attacks than high
AS ‘threat attenders’ (42.9%); however, chi square analy-
ses revealed that this difference was not statistically sig-
nificant (X
2
(1, N=27) =.94, p[.05).
Moderate and Low AS
As expected, there were no significant differences between
moderate AS ‘threat attenders’ and ‘threat avoiders’ or low
AS ‘threat attenders’ and ‘threat avoiders’ on measures of
AS (t(31) =.15, p[.05, g
p
2
=.00; t(20) =-1.36,
p[.05, g
p
2
=.09, respectively) or health anxiety
(t(31) =1.0, p[.05, g
p
2
=.03; t(20) =-1.10, p[.05,
g
p
2
=.06, respectively). Similarly, ‘threat attenders’ and
‘threat avoiders’ did not differ in self-reported history
of panic attacks for either the moderate (X
corr
2
(1,
N=33) =.00, p[.05) or low (X
corr
2
(1, N=22) =.00,
p[.05) AS groups.
Discussion
The present study suggests that there are differences
between attention style subgroups of high AS individuals in
their ability to intentionally forget threat-related informa-
tion and their degree of anxiety-related psychopathology.
Consistent with our hypotheses, high AS individuals who
had a tendency to avoid threat-related information were
found to have a greater magnitude of directed forgetting for
threat-related information as compared to high AS indi-
viduals who had a tendency to attend to threat-related
information. This difference was specific to high AS ‘threat
avoiders’’ greater ability to intentionally forget threat-
related words that they were instructed to forget and as
expected, there was no difference in memory for words that
they were instructed to remember. Moreover, high AS
‘threat attenders’ and ‘threat avoiders’ differed in their
Fig. 1 The mean directed forgetting effect as a function of attention
style subgroup (‘Threat Attenders’, ‘Threat Avoiders’) and word type
among high (a), moderate (b) and low (c) AS individuals. Note. Error
bars represent standard error of the mean (SEM)
9
Conclusions did not differ when directional one-tailed ttests were
conducted.
Cogn Ther Res
123
degree of psychopathology related to anxiety: High AS
‘threat avoiders’ had significantly higher levels of AS and
health anxiety than high AS ‘threat attenders’. In fact, high
AS individuals who had a tendency to avoid threat-related
information had ASI scores ([30) in the range found
among individuals with panic disorder (Cox et al. 1999).
Given that the present study investigated a non-clinical
sample of young adults, this may suggest that high AS
individuals who have an avoidant attention style in
response to threat, and a superior ability to intentionally
forget threat-related information, may be at increased risk
of developing anxiety-related psychopathology.
The finding that high AS individuals who avoided
threat-related information exhibited a greater magnitude of
directed forgetting than high AS individuals who attended
to threat-related information is a novel contribution to the
field. To date, only one other study has investigated
intentional forgetting among high AS individuals with
panic disorder (McNally et al. 1999b) and it found no
differences between AS groups in their intentional forget-
ting of threat-related information. However, unlike previ-
ous research (e.g., McNally et al. 1999b), the present study
investigated intentional forgetting of threat-related infor-
mation between subgroups of high AS individuals based on
their preferred (and perhaps automatic) mode of attending
to this information, as measured during their study of this
material. The results of previous research on AS and
memory (remembering and forgetting) may have been
affected by not considering attentional style, and researchers
may have been inaccurately conceptualizing heterogeneous
high AS groups as being one homogeneous AS group.
In contrast to previous research that has revealed mem-
ory biases toward threat-related information among indi-
viduals with panic disorder (e.g., Becker et al. 1994; Cloitre
and Liebowitz 1991; Cloitre et al. 1994; McNally et al.
1989) and high levels of AS (McCabe 1999), no such dif-
ferences in memory for remember-cued words were found
between attention style subgroups of high AS individuals in
the present study. This is similar to other research failing to
reveal memory biases toward threat among high AS indi-
viduals (e.g., Otto et al. 1994; Pickles and van den Broek
1988; Rapee 1994) and is supported by the results of a
recent meta-analysis showing that anxiety (including panic)
had no significant impact on recognition of threat (Mitte
2008). Similar to McNally et al. (1999b), we expected that
our non-clinical sample of high AS individuals would be
capable of instantiating an instruction to remember threat-
related words; we did not conceive of their automatic ten-
dency to avoid threat-related information as sufficiently
strong to override top–down intentions to commit such
items to memory. The findings of the present study suggest
that memory differences among high AS individuals may
emerge only when individuals are instructed to forget
threat-related words. Under these circumstances, high AS
individuals who tend to avoid threatening information may
be better able to forget this information (and likely avoid
processing it) when they are given instruction to do so. This
suggests that the ability to effectively forget threat-related
information might be under high AS ‘threat avoiders’’
strategic control; in other words, cognitive avoidance might
be subject to ‘‘strategic override’’ (Williams et al. 1996,
p. 19). More specifically, when given explicit instructions to
remember, high AS ‘threat avoiders’ may be able to sup-
press or override their natural tendency to avoid threat-
related information. However, in daily life, without explicit
instructions to remember, high AS ‘threat avoiders’ and
‘threat attenders’ might consistently demonstrate their nat-
ural or preferred tendencies to avoid or attend to threat by
being less likely to exert cognitive control over their pre-
ferred attention styles. Regardless, the results of the current
study highlight the utility of a directed forgetting paradigm
in research investigating the link between AS and memory
to assess both remembering and forgetting among high AS
individuals. In this paradigm, instructions to forget threat-
related words appear to intensify pre-existing cognitive
avoidance tendencies (McNally et al. 1999b) among high
AS ‘threat avoiders’.
Investigation of differences between high AS attention
style subgroups revealed that high AS individuals who had a
tendency to avoid threatening information had higher levels
of AS and health anxiety than high AS individuals who had a
tendency to preferentially process threatening information.
Table 2 Means (SDs) for anxiety-related variables for the total sample as a function of AS group and attention style subgroup
High AS Moderate AS Low AS
Threat avoiders
(n=13)
Threat attenders
(n=14)
Threat avoiders
(n=16)
Threat attenders
(n=17)
Threat avoiders
(n=11)
Threat attenders
(n=11)
Anxiety sensitivity (ASI) 34.54* (7.40) 25.71 (8.57) 15.88 (7.59) 16.24 (6.57) 11.81 (6.54) 8.73 (3.72)
Health anxiety (HAQ) 26.38* (10.90) 17.07 (9.84) 11.31 (10.01) 14.47 (8.11) 9.82 (11.98) 5.73 (3.04)
Panic history (% panickers)
(PAQ-R)
61.5% 42.9% 25.0% 23.5% 18.2% 27.3%
*p\.05
Cogn Ther Res
123
These differences were expected based on research showing
the deleterious effects of avoidance on the development and
maintenance of anxious psychopathology among individuals
with panic disorder (Buller et al. 1986; Ehlers 1995). Simi-
larly, high levels of AS have been found to place individuals
at increased risk for developing and maintaining panic (Cox
et al. 1999; Ehlers 1995; Maller and Reiss 1992; Schmidt
et al. 1997) as well as hypochondriasis (Otto et al. 1998;
Stewart and Watt 2000). This finding is clinically relevant
because it is thought that cognitive avoidance and disen-
gagement from threatening information may limit the
effectiveness of exposure therapy (Foa and Kozak 1986).
Accordingly, it is possible that high AS ‘threat avoiders’
have higher levels of psychopathology than ‘threat attenders’
because of the lack of exposure to threat experienced in their
daily lives, thereby limiting important experiential learning
opportunities that could decrease anxiety levels. Of rele-
vance to the present study, high AS ‘threat avoiders’ who
suppress memories of threat (through enhanced forgetting)
may be effectively reducing their exposure to threat, thereby
preventing habituation. Future research should explore the
potential mediating role of cognitive biases such as directed
forgetting of threat in explaining the increased risk of anxi-
ety-related psychopathology among high AS individuals,
and in particular, ‘threat avoiders’. The present study also
revealed that high AS ‘threat avoiders’ included more
women than high AS ‘threat attenders’. This is consistent
with previous research showing that women tend to have
higher levels of both AS (Peterson and Reiss 1992; Stewart
et al. 1997) and health anxiety (Marcus and Church 2003)
than men.
The present research assessed avoidance and vigilance of
threat stimuli (words) by measuring IOR (as a proxy for
attentional allocation) during the study trials of an item-
method directed forgetting paradigm. This is in contrast to
past research that investigated avoidance and vigilance
through the use of self-report measures (e.g., the Mainz
Coping Inventory; MCI; Krohne 1989) and gaze duration in
social interactions (Hock 1993). Given that the IOR task
was embedded within the memory task, it allowed for the
assessment of high AS individuals’ tendency to withdraw
and dwell attention ‘‘online’’ (e.g., during the experiment as
opposed to before or after the experiment) and in response
to the threat stimuli that were later assessed by the recog-
nition memory task. It is unclear whether results would be
similar if individual attention style was measured using
gaze duration and/or self-report questionnaires. Moreover,
it is important to investigate the concepts of attentional
vigilance and avoidance in the context of threat experienced
by high AS individuals in their daily lives. Whereas certain
threat-related words might have evoked more avoidance
among particular individuals than others (e.g., the word
‘‘cancer’’ among high AS individuals with health anxiety
pertaining to this illness), it is possible that the ecological
validity of attentional sub-typing may have been higher if
the threat stimuli ideographically relevant to each individ-
ual’s anxiety were used (see Riemann and McNally 1995).
Assessing attention style subgroups among high AS
individuals is a novel approach that requires further empir-
ical investigation using a variety of paradigms and tasks.
Future longitudinal research should assess the impact of
attention style on a broad range of information processing
biases (e.g., attention, interpretation, memory) that vary in
reliance on strategic/explicit and automatic/implicit pro-
cesses given that such biases have been shown to uniquely
predict panic symptoms (Teachman et al. 2007). Addition-
ally, there is a need to examine attention style subgroup
differences using AS groups categorized based on new
empirically derived and clinically meaningful cut-points
(Bernstein et al. 2010). Recent research suggests that using a
dimensional approach to AS classification based on the
newer ASI-3 might be able to meaningfully identify ‘‘high’’
or ‘‘at risk’’ AS individuals (Bernstein et al. 2010). Future
longitudinal research should investigate attentional sub-
group differences among individuals using this classification
approach to determine whether cognitive biases are stable
within individuals across situations and time. It is possible
that ‘‘real avoiders’’ may be more likely or uniquely found
among high-AS (relative to low-AS) individuals. Although
attentional subgroups among moderate and low AS groups
were also included in the present investigation for compar-
ison purposes, these might not be meaningful groups.
In summary, the present research extends the literature on
AS and memory biases in several ways. It is the only study to
examine memory among subgroups of high AS individuals
created based on their tendency to preferentially attend to or
avoid threat-related information. This extends the coping
literature on vigilant and avoidant subgroups of high trait
anxious individuals (e.g., Krohne 1993) to literature on AS
and memory biases. Also, these findings address the heter-
ogeneity inherent within high AS groups, which might
account for the lack of consensus in the literature on AS and
memory biases. Moreover, the present research highlights
the importance of simultaneously assessing both attention
and memory when examining the role that attention might
play in selecting items for commitment to memory. Given
the theoretical importance of attention in influencing the
relationship between high levels of anxiety and memory
(e.g., Beck et al. 1985; Mogg and Bradley 1998; Eysenck
et al. 2007), it is surprising that only a few studies (e.g.,
McNally et al. 1999a; Teachman 2005; Teachman et al.
2007) have assessed both attention and memory in the same
study. The present research not only investigated attention
and memory among individuals with varying levels of AS in
a single study, but it also embedded a task that is able to yield
a proxy measure of attentional dwell and withdrawal
Cogn Ther Res
123
‘‘online’’ within the context of the study trials of an item-
method directed forgetting paradigm. Therefore, this task
captured the moment-to-moment fluctuations in attention
that impact memory performance. This provides us with a
measure that more traditional questionnaire techniques
cannot. It also enabled more direct investigation of the
impact of attention style in response to the same threat
stimuli assessed at recognition. To date, a directed forgetting
paradigm has only been used with high AS individuals in one
previous study (McNally et al. 1999b) and has never before
been used to investigate differences between high AS indi-
viduals as a function of their attention style in response to
threat. Directed forgetting tasks offer advantages over other
tasks because they allow for an examination of both
remembering and forgetting. AS is thought to increase one’s
alertness to, as well as avoidance of, anxiety-provoking
information (Reiss and McNally 1985). Given that high AS
individuals might differ in their tendency to avoid (versus
attend to) threat, and avoidance might facilitate the ability to
strategically forget this information (McNally et al. 1999b),
directed forgetting tasks might be particularly valuable in
research on AS and memory. Future longitudinal investiga-
tions using these paradigms with clinical populations could
provide insight into the nature of the relationship between
AS and memory biases which continues to perplex
researchers and theorists alike.
Acknowledgments This research was supported by a grant from the
Natural Sciences and Engineering Research Council (NSERC) of
Canada awarded to Dr. Taylor. Noel was supported by a Canadian
Institutes of Health Research (CIHR) CGS Doctoral Research Award,
a CIHR Team in Children’s Pain Fellowship, a Nova Scotia Health
Research Foundation Doctoral Student Research Award, a Killam
Predoctoral Scholarship, and a stipend from the CIHR strategic
training initiative on Pain in Child Health at the time this research was
conducted. Quinlan was supported by an NSERC Julie Payette
scholarship and by a Killam Predoctoral scholarship. Dr. Stewart was
supported through a Killam Research Professorship from the
Dalhousie University Faculty of Science.
Appendix
See Table 3.
Table 3 Directed forgetting task word list
Threat Positive Neutral
breathless optimistic armchair
disoriented admired bench
flushed playful blender
numbness adored bookcase
palpitation pleasant buffet
shaking brilliant bureau
Table 3 continued
Threat Positive Neutral
sweating celebration cabinet
trembling praised washer
fainting cheerful carpet
choked relaxed clock
collapse delighted couch
crazy satisfied curtains
stroke elation desk
insane successful drawers
suffocate enjoyable dresser
worried superb lounge
trapped enthusiastic oven
fearful excited refrigerator
frightened thrilled rocker
helpless exhilaration shelf
nervous happiness stool
panic triumph stereo
scared incredible television
terrified laughter vanity
afraid optimistic banister
anxious admired basement
attack playful bedroom
cardiac adored cellar
coronary pleasant chair
despair charming closet
dizzy clever counter
dying confident cupboard
horror contented dishwasher
phobic easygoing doorknob
suffering ecstasy fan
tense friendly fork
terror generous freezer
tumor healthy mailbox
fatal joy mattress
deathbed loyal microwave
seizure outgoing mirror
cancer polite paneling
assault protected porch
illness reassured sheets
hospitals safety stairs
paralyzed secure table
hazard sincere tub
injure sociable windows
accident steady metals
lethal helpful sleepy
wound kindness raisin
destruction applause motel
coffin overjoyed lamb
poison serene vase
Cogn Ther Res
123
References
American Psychiatric Association. (1987). Diagnostic and statistical
manual of mental disorders: DSM-III-R (3rd rev. ed.). Wash-
ington, DC: Author.
American Psychiatric Association. (2000). Diagnostic and statistical
manual of mental disorders: DSM-IV-TR (4th ed., text rev.).
Washington, DC: Author.
Asmundson, G. J., Sandler, L. S., Wilson, K. G., & Walker, J. R.
(1992). Selective attention toward physical threat in patients with
panic disorder. Journal of Anxiety Disorders, 61, 295–303.
Baddeley, A. D. (2004). The psychology of memory. In A.
D. Baddeley, M. D. Kopelman, & B. A. Wilson (Eds.), The
essential handbook of memory disorders for clinicians (pp.
1–13). Chichester, UK: Wiley.
Beck, A. T., Emery, G., & Greenberg, R. I. (1985). Anxiety disorders
and phobias. New York: Basic Books.
Beck, J. G., Stanley, M. A., Averill, P. M., Baldwin, L. E., & Deagle,
E. A. (1992). Attention and memory for threat in panic disorder.
Behaviour Research and Therapy, 30, 619–629.
Becker, E., Rinck, M., & Margraf, J. (1994). Memory bias in panic
disorder. Journal of Abnormal Psychology, 103, 396–399.
Berlucchi, G., Chelazzi, L., & Tassinari, G. (2000). Volitional covert
orienting to a peripheral cue does not suppress cue-induced inhi-
bition of return. Journal of Cognitive Neuroscience, 12, 648–663.
Bernstein, A., Stickle, T. R., Zvolensky, M. J., Taylor, S., Abramo-
witz, J., & Stewart, S. (2010). Dimensional, categorical, or
dimensional-categories: testing the latent structure of anxiety
sensitivity among adults using factor-mixture modeling. Behav-
ior Therapy, 41(4), 515–529.
Birch, C. D., Stewart, S. H., Wiers, R. W., Klein, R. M., MacLean, A.
D., & Berish, M. J. (2008). The mood-induced activation of
implicit alcohol cognition in enhancement and coping motivated
drinkers. Addictive Behaviors, 33, 565–581.
Buller, R., Maier, W., & Benkert, O. (1986). Clinical subtypes of
panic disorder: Their descriptive and prospective validity.
Journal of Affective Disorders, 11, 105–114.
Carroll, J. B., Davies, P., & Richman, B. (1971). Word frequency
book. Boston, MA: Houghton Mifflin.
Chica, A. B. (2008). Differential effects of endogenous and exoge-
nous attention on information processing. Ph.D. dissertation.
Chica, A. B., Lupia
´n
˜ez, J., & Bartolomeo, P. (2006). Dissociating
inhibition of return from endogenous orienting of spatial
attention: Evidence from detection and discrimination tasks.
Cognitive Neuropsychology, 23, 1015–1034.
Cloitre, M., & Liebowitz, M. R. (1991). Memory bias in panic
disorder: An investigation of the cognitive avoidance hypothesis.
Cognitive Therapy and Research, 15, 371–386.
Cloitre, M., Shear, M. K., Cancienne, J., & Zeitlin, S. B. (1994).
Implicit and explicit memory for catastrophic associations to
bodily sensations words in panic disorder. Cognitive Therapy
and Research, 18, 225–240.
Cohen, J. D., MacWhinney, B., Flatt, M., & Provost, J. (1993).
PsyScope: An interactive graphic system for designing and
controlling experiments in the psychology laboratory using
Macintosh computers. Behavior Research Methods, Instruments,
and Computers, 25, 257–271.
Cowan, N. (1995). Attention and memory: An integrated framework.
New York: Oxford University Press.
Cox, B. J., Borger, S. C., & Enns, M. W. (1999). Anxiety sensitivity
and emotional disorders: Psychometric studies and their theo-
retical implications. In S. Taylor (Ed.), Anxiety sensitivity:
Theory, research and treatment of the fear of anxiety (pp.
115–148). Mahwah, NJ: Erlbaum.
Cox, B. J., Norton, G. R., & Swinton, R. P. (1992). The panic attack
questionnaire—revised. Unpublished scale, Centre for Addiction
and Mental Health, Toronto, Canada.
Danziger, S., & Kingstone, A. (1999). Unmasking the inhibition of
return phenomenon. Perception and Psychophysics, 61,
1024–1037.
De Houwer, J. (2006). What are implicit measures and why are we
using them? In R. W. Wiers & A. W. Stacy (Eds.), Handbook of
implicit cognition and addiction (pp. 1–8). London, UK: Sage
Publications Ltd.
Dorris, M. C., Klein, R. M., Everling, S., & Munoz, D. P. (2002).
Contribution of the superior colliculus to inhibition of return.
Journal of Cognitive Neuroscience, 14, 1256–1263.
Ehlers, A. (1995). A one-year prospective study of panic attacks:
Clinical course and factors associated with maintenance. Journal
of Abnormal Psychology, 104, 164–172.
Ehlers, A., Margraf, J., Davies, S., & Roth, W. T. (1988). Selective
processing of threat cues in participants with panic attacks.
Cognition and Emotion, 2, 201–219.
Eysenck, M. W., Derakshan, N., Santos, R., & Calvo, M. G. (2007).
Anxiety and cognitive performance: Attentional control theory.
Emotion, 7, 336–353.
Table 3 continued
Threat Positive Neutral
catastrophe gratified detergent
disease millionaire towel
emergencies heroic wardrobe
physician prize identical
ambulance proficient button
pain merriment reported
heartbeat jollily pillow
violence holiday anchor
tingling intelligent booklet
surgery terrific furnished
mutilated honour bleach
capable democratic
harmony sports
excellent cereals
devoted occasion
comfort groceries
achieves firelight
miracle moderate
fortunate implements
prestigious snowshoe
fingertips
wing
peasants
switch
chimney
shower
mantelpiece
robe
advertise
broom
Cogn Ther Res
123
Foa, E. B., & Kozak, M. J. (1986). Emotional processing of fear:
Exposure to corrective information. Psychological Bulletin, 99,
20–35.
Golding, J. M., & MacLeod, C. M. (Eds.). (1998). Intentional
forgetting: Interdisciplinary approaches. Mahwah, NJ: Erlbaum.
Hock, M. (1993). Coping dispositions, attentional direction, and
anxiety states. In H. W. Krohne (Ed.), Attention and avoidance:
Strategies in coping with aversiveness (pp. 139–169). Toronto,
Canada: Hogrefe & Huber.
Hunt, C., Keogh, E., & French, C. C. (2006). Anxiety sensitivity: The
role of conscious awareness and selective attentional bias to
physical threat. Emotion, 6, 418–428.
Keogh, E., Dillon, C., Georgiou, G., & Hunt, C. (2001). Selective
attentional biases for physical threat in physical anxiety sensi-
tivity. Journal of Anxiety Disorders, 15, 299–315.
Klein, R. M. (2000). Inhibition of return. Trends in Cognitive
Sciences, 4, 138–147.
Koven, N. S., Heller, W., Banich, M. T., & Miller, G. A. (2003).
Relationships of distinctive affective dimensions to performance
on an emotional Stroop task. Cognitive Therapy and Research,
27, 671–680.
Krohne, H. W. (1989). The concept of coping modes: Relating
cognitive person variables to actual coping behavior. Advances
in Behaviour Research and Therapy, 11, 235–248.
Krohne, H. W. (1993). Vigilance and cognitive avoidance as concepts
in coping research. In H. W. Krohne (Ed.), Attention and
avoidance: Strategies in coping with aversiveness (pp. 19–50).
Toronto, Canada: Hogrefe & Huber.
Krohne, H. W., Egloff, B., Varner, L. J., Burns, L. R., Weidner, G., &
Ellis, H. C. (2000). The assessment of dispositional vigilance and
cognitive avoidance: Factorial structure, psychometric proper-
ties, and validity of the Mainz Coping Inventory. Cognitive
Therapy and Research, 24, 297–311.
Lees, A., Mogg, K., & Bradley, B. P. (2005). Health anxiety,
anxiety sensitivity and attentional biases for pictorial and
linguistic health-threat cues. Cognition and Emotion, 19,
453–462.
Lucock, M., & Morley, S. (1996). The health anxiety questionnaire.
British Journal of Health Psychology, 1, 137–150.
Lupia
´n
˜ez, J., Decaix, C., Sie
´roff, E., Chokron, S., Milliken, B., &
Bartolomeo, P. (2004). Independent effects of endogenous and
exogenous spatial cueing: Inhibition of return at endogenously
attended target locations. Experimental Brain Research, 159,
447–457.
MacLeod, C. M. (1999). The item and list methods of directed
forgetting: Test differences and the role of demand character-
istics. Psychonomic Bulletin and Review, 6, 123–129.
Maller, R. G., & Reiss, S. (1992). Anxiety sensitivity in 1984 and
panic attacks in 1987. Journal of Anxiety Disorders, 6, 241–247.
Marcus, D., & Church, S. (2003). Are dysfunctional beliefs about
illness unique to hypochondriasis? Journal of Psychosomatic
Research, 54, 543–547.
Mathews, A. (1993). Attention and memory for threat in anxiety. In
H. W. Krohne (Ed.), Attention and avoidance: Strategies in
coping with aversiveness (pp. 119–135). Toronto, Canada:
Hogrefe & Huber.
McCabe, R. E. (1999). Implicit and explicit memory for threat words
in high- and low-anxiety-sensitive participants. Cognitive Ther-
apy and Research, 23, 21–38.
McNally, R. J. (1995). Automaticity and the anxiety disorders.
Behaviour Research and Therapy, 33, 747–754.
McNally, R. J. (1999). Anxiety sensitivity and information-processing
biases for threat. In S. Taylor (Ed.), Anxiety sensitivity: Theory,
research, and treatment of the fear of anxiety (pp. 183–197).
Mahwah, NJ: Erlbaum.
McNally, R. J., Foa, E. B., & Donnell, C. D. (1989). Memory bias for
anxiety information in patients with panic disorder. Cognition
and Emotion, 3, 27–44.
McNally, R. J., Hornig, C. D., Hoffman, E. C., & Han, E. M. (1999a).
Anxiety sensitivity and cognitive biases for threat. Behavior
Therapy, 30, 51–61.
McNally, R. J., & Lorenz, M. (1987). Anxiety sensitivity in
agoraphobics. Journal of Behavior Therapy and Experimental
Psychiatry, 18, 3–11.
McNally, R. J., Otto, M. W., Yap, L., Pollack, M. H., & Hornig, C. D.
(1999b). Is panic disorder linked to cognitive avoidance of
threatening information? Journal of Anxiety Disorders, 13,
335–348.
McNally, R. J., Riemann, B. C., Louro, C. E., Luckach, B. M., &
Kim, E. (1992). Cognitive processing of emotional information
in panic disorder. Behaviour Research and Therapy, 30,
143–149.
Mitte, K. (2008). Memory bias for threatening information in anxiety
and anxiety disorders: A meta-analytic review. Psychological
Bulletin, 134, 886–911.
Mogg, K., & Bradley, B. P. (1998). A cognitive-motivational analysis
of anxiety. Behaviour Research and Therapy, 36, 809–848.
Otto, M. W., Demopulos, C. M., McLean, N. E., Pollack, M. H., &
Fava, M. (1998). Additional findings on the association between
anxiety sensitivity and hypochondriacal concerns: Examination
of patients with major depression. Journal of Anxiety Disorders,
12, 225–232.
Otto, M. W., McNally, R. J., Pollack, M. H., Chen, E., & Rosenbaum,
J. F. (1994). Hemispheric laterality and memory bias for threat in
anxiety disorder. Journal of Abnormal Psychology, 103,
828–831.
Peterson, R. A., & Reiss, S. (1992). Anxiety sensitivity index manual
(2nd ed.). Worthington, OH: International Diagnostics Systems.
Pickles, A. J., & van den Broek, M. D. (1988). Failure to replicate
evidence for phobic schemata in agoraphobic patients. British
Journal of Clinical Psychology, 27, 271–272.
Posner, M. I., & Cohen, Y. (1984). Components of visual orienting. In
H. Bouma & D. Bouwhuis (Eds.), Attention and performance X
(pp. 531–556). London, UK: Erlbaum.
Rapee, R. M. (1994). Failure to replicate a memory bias in panic
disorder. Journal of Anxiety Disorders, 8, 291–300.
Reiss, S. (1991). Expectancy model of fear, anxiety, and panic.
Clinical Psychology Review, 11, 141–153.
Reiss, S., & McNally, R. J. (1985). Expectancy model of fear. In S.
Reiss & R. R. Bootzin (Eds.), Theoretical issues in behavior
therapy (pp. 107–121). New York: Academic Press.
Reiss, S., Peterson, R. A., Gursky, D. M., & McNally, R. J. (1986).
Anxiety sensitivity, anxiety frequency and the prediction of
fearfulness. Behaviour Research and Therapy, 24, 1–8.
Riemann, B. C., & McNally, R. J. (1995). Cognitive processing of
personally relevant information. Cognition and Emotion, 9,
325–340.
Ro, T., & Rafal, R. D. (1999). Components of reflexive visual
orienting to moving objects. Perception and Psychophysics, 61,
826–836.
Schmidt, N. B., Lerew, D. R., & Jackson, R. J. (1997). The role of
anxiety sensitivity in the pathogenesis of panic: Prospective
evaluation of spontaneous panic attacks during acute stress.
Journal of Abnormal Psychology, 106, 355–364.
Shostak, B. B., & Peterson, R. A. (1990). Effects of anxiety sensitivity
on emotional response to a stress task. Behaviour Research and
Therapy, 28, 513–521.
Stewart, S. H., Conrod, P. J., Gignac, M. L., & Pihl, R. O. (1998).
Selective processing biases in anxiety-sensitive men and women.
Cognition and Emotion, 12, 105–133.
Cogn Ther Res
123
Stewart, S. H., Taylor, S., & Baker, J. M. (1997). Gender differences
in dimensions of anxiety sensitivity. Journal of Anxiety Disor-
ders, 11, 179–200.
Stewart, S. H., & Watt, M. C. (2000). Illness attitudes scale
dimensions and their associations with anxiety-related constructs
in a nonclinical sample. Behaviour Research and Therapy, 38,
83–99.
Stewart, S. H., & Watt, M. C. (2001). Assessment of health anxiety.
In G. J. G. Asmundson, S. Taylor, & B. J. Cox (Eds.), Health
anxiety: Clinical and research perspectives on hypochondriasis
and related disorders (pp. 95–131). Sussex, UK: Wiley.
Stewart, S. H., Zvolensky, M. J., & Eifert, G. H. (2002). The
relationship of anxiety sensitivity, experiential avoidance, and
alexithymic coping to young adult’s motivations for drinking.
Behavior Modification, 26, 274–296.
Tabachnick, B. G., & Fidell, L. S. (2001). Using multivariate
statistics (4th ed.). Needham Heights, MA: Allyn and Bacon.
Taylor, S. (1999). Anxiety sensitivity: Theory, research, and treatment
of the fear of anxiety. Mahwah, NJ: Erlbaum.
Taylor, T. L. (2005). Inhibition of return following instructions to
remember and forget. Quarterly Journal of Experimental
Psychology: Human Experimental Psychology, 58A, 613–629.
Teachman, B. A. (2005). Information processing and anxiety
sensitivity: Cognitive vulnerability to panic reflected in
interpretation and memory biases. Cognitive Therapy and
Research, 29, 479–499.
Teachman, B. A., Smith-Janik, S. B., & Saporito, J. (2007).
Information processing biases and panic disorder: Relationships
among cognitive and symptom measures. Behaviour Research
and Therapy, 45, 1791–1811.
Tipper, S. P., Rafal, R. D., Reuter-Lorenz, P., Staarveldt, Y., Ro, T.,
Egly, R., et al. (1997). Object based facilitation and inhibition
from visual orienting in the human split-brain. Journal of
Experimental Psychology: Human Perception and Performance,
23, 1522–1532.
White, K. S., Brown, T. A., Somers, T. J., & Barlow, D. H. (2006).
Avoidance behavior in panic disorder: The moderating influence
of perceived control. Behaviour Research and Therapy, 44,
147–157.
Williams, J. M. G., Mathews, A., & MacLeod, C. (1996). The
emotional Stroop task and psychopathology. Psychological
Bulletin, 120, 3–24.
Wilson, K. A., & Hayward, C. (2006). Unique contributions of
anxiety sensitivity to avoidance: A prospective study in adoles-
cents. Behaviour Research and Therapy, 44, 601–609.
Zvolensky, M. J., & Forsyth, J. P. (2002). Anxiety sensitivity
dimensions in the prediction of body vigilance and emotional
avoidance. Cognitive Therapy and Research, 26, 449–460.
Cogn Ther Res
123