.92.}. Exfififl. .

I 3
fi <1

. A 3 . 2
A A A , t «a»

A, . in r. x

"a. .A ,A .25..

31%....33: ,

é :Y . A .
.. $3.... 1..

, I
fisfiuumf
.-

I
3...
{Is

.. .3. A; .
x . , .

.305

3.:
5x

. x a. z 1
32:21:13.
s r .

2:. gawk: .5,
guns.” 1;. I
.2 +24

- I'VfiXIDQ
a: .3!
1“ ' p.
, z ,
. 31.5.. . .
.32....» . . , , . A A hum: t
. r . . . ‘ . r , 1- 55 you .Yd:
' . . . . PW O‘: iii“. It: I A I.
. A . , at... tug L t...)
V . 3- .32...) $.15:-
, , 4.; .V
\ . .hfi. am.
3.91% , A A A , A 1:. «3.5... a».
1 .. i. A A . . . A mutanhvm.
. N. ntt!.11aa1
. r...

z...
I

grassy: 3.!
5.. z .5.

. fi
..
u, . «x. . lfl

) 13

inf-1W .‘1
?.ufii21.. . .
, .

3i. .

l3...
3: 5.233
I i. I: .
‘r A 5...».
itla‘:
..:.$:15..-.

 

 

 

 

“Anvmrnflymu, . A . , V f ,. . . . . , {.71. .
, . . t . A, , _, V wgwé i,

 

 

 

,2 ' LIBRARY
2007 Michigan State
University

 

 

This is to certify that the
dissertation entitled

THE IMPACT OF A'ITACHMENT ACTIVATION ON MEMORY IN
WOMEN

presented by

Ellen O’Toole

has been accepted towards fulfillment
of the requirements for the

Ph.D. degree in Psychologl

 

 

Major Professor’s Signature
87/” / 2006

Date

MSU is an Affinnative Action/Equal Opportunity Institution

- — —--.--I-Q-I-I-I-v-I-I-0-'-C-I-O-fl-t-0-o-D-

-l_-.—'-v---|-O-I-I-l-I-l-D-I-o-l—a.

 

PLACE IN RETURN Box to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 p:/ClRC/DateDue.indd-p.1

 

THE IMPACT OF ATTACHMENT ACTIVATION ON MEMORY IN WOMEN
By
Ellen O’Toole

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Department of Psychology

2006

ABSTRACT
THE IMPACT OF ATTACHMENT ACTIVATION ON MEMORY IN WOMEN
By

Ellen O’Toole

Priming associate cognitive networks and level of anxiety were examined as
mechanisms underlying the impact of attachment activation on word list recall. A
block random experiment was conducted with 323 female undergraduate
students (mean age 20 years) with two independent factors: level of anxious
attachment (high or low) and memory task outcomes after engaging in one of
four conditions (Attachment-Death Story, Spider Story, Attachment-Embedded
Story, and Neutral Story). It was hypothesized that different cognitive associative
networks could be activated by using relevant primes, and that this would be
seen via differential processing of word types in a recall task. Another
hypothesis was that exposing persons with high vigilance to a specific
domain/network to an intense activation of that domain would result in anxiety
strong enough to disrupt memory processing. It was also posited that previous
patterns of word recall found in the literature would be replicated or extended.
Results did not support these hypotheses. Another postulation was that an overt
and intense method of activating the attachment network would decrease
memory performance for women when compared a covert and non-intense
activation method. While a trend was noted in support of this hypothesis, the

result was not significant. The implications of these findings are discussed.

Dedicated to my partner, Holly Oppenheimer, whose support, encouragement,

and love sustained me throughout my graduate studies.

ACKNOWLEDGMENTS

I would like to acknowledge the efforts of the many people who enabled
me to accomplish this project, and to complete my degree. First of all, I want to
thank my dissertation chair, Bob Caldwell, for his guidance and generosity of
spirit in working with me. I deeply appreciate his support of my success in
bringing this project to completion, and for all of his encouragement over the
years. I would also like to thank my other committee members, Alytia
Levendosky, Chris Larson, and Norman Abeles. Alytia Levendosky has been a
notably enthusiastic mentor and wonderful clinical supervisor who has taught me
the importance of celebrating my accomplishments, small and large, throughout
graduate school. I thank Chris Larson for all of the stimulating discussions on the
prefrontal lobes, as well as her cheerful ability to provide feedback that is both
encouraging and challenging. I am very grateful to Norman Abeles. who has
been an invaluable mentor in the area of neuropsychology and has always
shown his support of my career goals. I would also like to thank Anne Bogat,
who contributed many years to both this project and to my development as a
psychologist.

I would also like to recognize the research assistants who worked hard to
interview and enter data for this project: Cara Crawford, Krista Snoblen, Megan
Ford, and Meghan Hanlon. They were a delightful and responsible team to work
with. I also thank the women who consented to be interviewed for this project.

I also wish to acknowledge those who supported me in other ways. The

good friends I met through the program are people I truly appreciate and enjoy:
Iv

Carolyn Dayton, En'ka DeJonghe, Mark Ettenhofer, Greg Nicholson, Juliette
Rederstorff, Jen Sachek, Gillian Stavro, and Mai Mai Tun. I also want to thank
my friends Dia Day, Lisa Harris, and Jules VanderGalien for years of friendship
and support. My two mentors Carol George and Marilyn Charles have remained
in my life long after their official duties were over and have now become my
friends, as well. I thank everyone for their kindness and support throughout days
and nights of intense work, laughter, sleep deprivation, and, at last, achievement.
Finally, I would like to thank my family. My parents, Rosemary and Jim,
and my siblings (Jim, Dennis, Donna, Kathy, Mike, and Mary) along with their
children and partners, have always been there to say “you can do it” whenever I
needed encouragement. They have lovingly supported me in so many ways over
the years. My sister Mary, in particular, has acted as my role model as well as
my collaborator in mischief. Most of all, I would like to thank my partner, Holly
Oppenheimer. She has provided the core of love, kindness, and practical
support which allowed me to take on my graduate studies and earn a doctoral

degree.

TABLE OF CONTENTS

LIST OF TABLES ........................................................................... viii
INTRODUCTION ............................................................................. 1
Attachment ............................................................................ 3
Theories of Anxiety and Cognition ............................................... 6
Impact of Attachment on Cognitive Processes ............................... 8
Attachment and Intelligence ................................................ 9
Attachment Priming and Reaction Time Studies ...................... 10
Attachment as a Cognitive Schema ...................................... 14
Attachment Priming and Memory ......................................... 17
Synthesizing Predictions of Cognition and Emotion .................. 21
RATIONALE AND HYPOTHESES ..................................................... 24
METHOD ..................................................................................... 31
Participants ............................................................................ 31
Measures .............................................................................. 32
Demographics .................................................................. 32
Attachment ...................................................................... 32

Anxiety and Depression ..................................................... 33

Attention and Concentration ................................................ 34

Memory .......................................................................... 35

Spider Fear ..................................................................... 37
Procedure .............................................................................. 37
RESULTS .................................................................................... 41
Power Analysis ....................................................................... 42
Tests of Hypotheses ................................................................ 43
Differentiation Attachment System Activation .......................... 43

Anxiety Effect ................................................................... 48
Supplemental Analyses ...................................................... 51

vi

Differentiating Methods of Activating Attachment ..................... 52

Replication of Patterns of Recall .......................................... 53
Supplemental Analyses ...................................................... 54
DISCUSSION ................................................................................ 61
Discussion of Results ............................................................... 61
Differentiating Attachment System Activation .......................... 61

Anxiety Effect ................................................................... 66
Differentiating Methods of Activating Attachment ..................... 70
Replication of Patterns of Recall .......................................... 71

Clinical Implications ................................................................. 75
Strengths and Limitations .......................................................... 76
Future Directions ..................................................................... 79
REFERENCES .............................................................................. 82
APPENDICES ............................................................................... 89
Appendix A: Online Consent Form ............................................... 89
Appendix B: Demographics Questionnaire .................................... 90
Appendix C: Reciprocal Attachment Questionnaire ......................... 92
Appendix D: BDI ..................................................................... 95
Appendix E: State-Trait Anxiety Inventory ..................................... 98
Appendix F: Trailmaking Test ..................................................... 100
Appendix G: Spider Story ......................................................... 101
Appendix H: Attachment-Death Story .......................................... 103
Appendix I: Attachment-Embedded Words Study ........................... 105
Appendix J: Neutral Story ......................................................... 107
Appendix K: Letter-Number Sequencing ....................................... 109
Appendix L: Word List Task ....................................................... 110
Appendix M: Fear of Spiders Questionnaire .................................. 112
Appendix N: Debriefing lnforrnation ............................................. 1 13

vii

LIST OF TABLES

Table 1: Group by Experimental Condition with Projected Sample
Size ...........................................................................

Table 2: Skew and Kurtosis Values for Transformed Variables. .........

Table 3: Descriptive Statistics for Analysis Covariates. .

Table 4: Means and Standard Deviations for Dependent Variables
by Experimental Condition and Level of Anxious Attachment

Table 5: Means and Standard Deviations for Independent Variables

and Covariates by Experimental Condition and Level of
Anxious Attachment. .............................................................

Table 6: Summary of One-way ANCOVA Testing Hypothesis 1 .........

Table 7: Estimated Marginal Means and Standard Errors Tested in
Hypothesis 1 ...............................................................

Table 8: Summary of One-way ANCOVA Testing Hypothesis 2.

Table 9: Estimated Marginal Means and Standard Errors Tested in
Hypothesis 2 ...............................................................

Table 10: Hypothesis 3: Pearson Correlations Between Post-Story
Anxiety and Cognitive Variables for High-Anxious Attachment
Subjects in Death Story Condition .....................................

Table 11: Hypothesis 4: Pearson Correlations Between Post-Story
Anxiety and Cognitive Variables for High-Spider Fear
Subjects in Spider Story Condition ..................................

Table 12: Summary of Two-way ANCOVA Testing Hypothesis 5 .......

Table 13: Estimated Marginal Means and Standard Errors Tested in
Hypothesis 5 ...............................................................

Table 14: Total Word Recall List Scores for the Attachment
Conditions .................................................................

viii

26

41

42

44

46

46

46

47

47

48

49

50

50

52

Table 15: Mean Word Recall for Combined Attachment Conditions...

Table 16: Mean Word Recall for All Subjects ................................

Table 17: Summary of Two-way ANCOVA Testing Hypothesis 8:
Positive Attachment Words ...........................................

Table 18: Estimated Marginal Means and Standard Errors Tested in
Hypothesis 8 ............................................................

Table 19: Summary of Two-way ANCOVA Testing Hypothesis 8:
Negative Attachment Words ..........................................

Table 20: Estimated Marginal Means and Standard Errors Tested in
Hypothesis 8 ...............................................................

Table 21: Sample Size per Cell Using High/Low Compulsive Care-
Seeking .....................................................................

Table 22: Summary of Two-way ANCOVA Testing Hypothesis 8
using Compulsive Care-Seeking: Positive Attachment
Words .......................................................................

Table 23: Estimated Marginal Means and Standard Errors Tested in
Hypothesis 8: Compulsive Care-Seeking ........................

53

53

55

55

56

56

57

59

60

INTRODUCTION

Attachment theory has generally been utilized by researchers examining
relational and emotional responses. It has only been in recent years that interest
has grown in the explanatory power of attachment theory as it interacts with
cognitive processes. This area of research has mainly focused on those mental
processes such as attention, reaction time (speed of processing), and memory
(both episodic and semantic). A notable attribute of these studies is the
utilization of cognitive theory (i.e., associative networks), to examine the
influence of attachment on cognition. The theory of associative networks is
based on the concept of a semantic mental network where associated ideas or
concepts are linked together. When the network is primed or activated (by a
word or phrase associated with the network), a spreading activation throughout
the network occurs, and related components of that network are thought to be
processed more quickly and efficiently (Solso, 1998). This concept has been
applied to attachment as well, with exposure to relevant words, phrases, or other
material priming an “attachment associative network.” A variety of paradigms
have been used to activate the attachment system, such as subliminal priming
(Maier et al., 2004; Mikulincer, Bimbaum, Woddis, & Nachmias, 2000; Mikulincer,
Gillath, & Shaver, 2002), overt priming with attachment-relevant sentences
(Poole, 1998), writing about an attachment relationship (Rowe & Camelley,
2003), conditioned pairing of rejection/acceptance with tones (Baldwin & Kay,
2003; Baldwin & Meunier, 1999), and reading attachment-relevant vignettes
(Pereg & Mikulincer, 2004). Unfortunately, such paradigms are not consistent in

format or application and researchers have not employed them to answer the
1

questions of 1) are core attachment elements specifically being activated, as
predicted by theory, versus a general anxiety response? and 2) during
attachment activation, how is the emotional aspect of attachment varying in

intensity, and thus altering the cognitive outcome?

Theories of anxiety predict a cycle of physiological arousal and emotion
that impacts cognitive processing (Eysenck, 1982; Eysenck, 1985; and Spence &
Spence, 1966). Attachment theory is specific to intimate protective relationships,
and also predicts an anxious response, one that may be expected to impact
cognition, as well. The two theories do not contradict one another, but Eysenck’s
theory of anxiety is much broader and refers to physiological arousal generated
by a wide variety of threats and negative anticipatory thoughts. The current
literature on attachment and cognition has failed to examine the emotional
response so important to attachment theory. One example of the importance of
considering the emotional response (conceived of as an attachment-specific
anxiety) concurrent with attachment activation is found in the literature on
anxiety. It is a well-established fact that very high levels of anxiety consume
cognitive resources, thus negatively impacting cognitive functioning (Eysenck,
1982). Research in attachment has found that different attachment groups
respond with varying levels of psychological anxiety and cortisol secretion
(Goldberg, 1995) in response to having the attachment system activated. Those
persons who are anxiously attached may reach such high levels of anxiety when
the attachment system is intensely activated. This is predicted to result in
impaired cognitive processing of demanding tasks. This study thus attempts to

refine the theoretical predictive power of attachment in cognitive studies.
2

Attachment

Bowlby identified attachment as a protective system developed in infancy
as a survival mechanism (Ainsworth, Blehar, Waters, & Wall, 1978, and Bowlby,
1969, 1973, 1980). It is a dually informed system, with both infant and caregiver
acting and reacting within the relationship, that revolves around the caregiver-
infant bond in times of stress or threat (Ainsworth et al., 1978; George &
Solomon, 1999; Goldberg, 1995). This perceived threat begins a cycle (or
activation of the attachment system) whereby the child seeks contact with a
primary caregiver or expresses distress, and a timely, sensitive response from
the caregiver soothes the child (Ainsworth et al., 1978; Goldberg, 1995). Such
comfort affirms for the child the sensitivity of the caregiver to such expressed
needs, and a feeling of safety is therefore associated with proximity to this
caregiver. It follows that the attachment behaviors of the child are no longer
necessary when an appropriate response is provided; the attachment system
shuts down, or deactivates, and the child’s behavior and physiological arousal
returns to a normal state (Fox & Card, 1999; Hofer, 1995).

Researchers have identified some typical patterns of response to the
attachment system being activated. The optimal, or secure, response develops
from caregiving interactions whereby the infant learns that his or her attachment
bids will be met by a caregiver who is available and responsive in an appropriate
and timely manner (Ainsworth et al., 1978; Goldberg, 1995). They engage in
exploration in the mother’s presence, maintain proximity checks, and show a
variety of distress responses upon her leaving, yet her return allows these

Children to calm down relatively quickly and deactivate the attachment system.
3

Nonoptimal patterns of responding have been identified as well over the last
three decades of attachment research. One of these patterns, the anxious-
insecure style, develops in response to caregiving that is unpredictable, and at
times, inappropriate (Ainsworth et al., 1978; Goldberg, 1995). Those in the
anxious-insecure category tend to cling to the mother, avoiding exploration, and
show great distress when the mother leaves and returns. Upon reunion, these
children may show both angry and contact seeking behavior, and will take much
longer to be comforted than secure children (Ainsworth et al., 1978; George &
Solomon, 1999; Goldberg, 1995). Different terms have been used for such
patterns of attachment responses, and the terms used for childhood have adult
equivalents. This pattern has commonly been called ambivalent/resistant or
anxious-ambivalent for infants and children, and anxious, preoccupied, or
ambivalent for adults. Here it will be referred to as anxious attachment.

The attachment system is thought to extend into adulthood, functioning in
a similar way as in childhood, but with some important differences. Attachment
researchers have identified several criteria that differentiate adult attachment
relationships from other social relationships: seeking contact with the attachment
figure when contending with stress; decreased anxiety through use of the
attachment figure as a haven of safety (either through contact or by use of
mental representation); a substantial increase in anxiety and searching behavior
if threatened by separation from the attachment figure; expectations that the
relationship will last over time; and a reciprocity appropriate to adult
relationships, based on a mutual sharing of comfort and protection (George et al.,

1999; West, Sheldon, & Reiffer, 1987). Feeney (1999) described the events that
4

arouse the adult attachment system: “stressful conditions in the social or physical
environment; conditions that appear to threaten the future of the attachment
relationship; and conditions of the individual, such as ill health" (Feeney, 1999:
p.371 ).

Secure adults give evidence of valuing attachment relationships while
being able to appropriately resolve or down-regulate the attachment system (for
example, seeking proximity to an attachment figure during a time of stress).
Anxiously attached adults may present as angry, passive, or fearful in the face of
an activated attachment system. Narrative research interviews reveal that while
secure adults are coherent and consistent in discussing attachment relationships,
the interviews of anxiously attached adults contain long, rambling sentences with
repeated or excessive detail.

Relevant to the area of research on attachment and cognition is the
concept of lntemal working models. lntemal working models are mental
representations of self and attachment figures that develop during childhood.
Such representations of self may include a sense of being valued and deserving,
as well as secure and confident in the attachment figure's responsivity.
Representations of the attachment figure include the level of availability,
responsiveness, and sensitivity regarding attachment needs (Bowlby, 1973).
Bretherton and Munholland (1999) described internal working models as “a
multiply connected schema network” (p. 98) in which repeated specific
attachment-relevant incidents (e.g., “mom came home from work and comforted
me when I hurt myself') inform the development of overall meaning constructs

(“Mom makes herself available when I need her”). Such an infonnation-
5

processing network provides a framework for understanding how attachment
functions, especially in adults.
Theories of Anxiety and Cognition

Prior research has shown a curvilinear relationship between anxiety and
resulting cognitive performance, with moderate anxiety producing increased
effort on mental tasks (Spence & Spence, 1966), and higher levels of anxiety
showing decrease in cognitive performance (Eysenck, 1982). Known as the
Yerkes-Dodson theory, this indicates that arousal related to anxiety assists
mental efforts, while high levels lower performance (Eysenck, 1985). More
recent research suggests the existence of a “threshold effect,” whereby deficits in
cognitive processing occur as anxiety reaches clinically significant levels

(Waldstein et al., 1997).

Eysenck and Calvo’s (1992) processing efficiency theory adds the
prospect of limited mental processing resources, with performance on mental
tasks thought to be impacted by availability of such resources. Additionally,
anxiety is postulated to decrease cognitive capacity as resources are diverted to
threat-vigilance. This explains the phenomena of high anxiety/low performance
via the increased level of intrusive worry which is thought to accompany higher
levels of anxiety. This intrusive worry is thought to overwhelm cognitive
processing by increasing task demands of the moment, while cognitive resources
remain the same. In Barlow’s (2000) discussion of this phenomenon, he
concurs, stating that “[a]t sufficient intensity, this process results in disruption of

concentration and performance.”

The extant literature appears to provide mixed evidence linking anxiety to
cognitive performance. Yovel and Mineka (2004) reported failing to find a link
between anxiety and attentional bias for negative information. These findings,
however, provide some support for a threshold effect: studies which examine
nonpsychiatric populations have found little impact of anxiety on cognitive
functioning (Hoffman & al’Absi, 2003; Palav et al. 2000; and Waldstein et al.
1997), while clinical subjects with anxiety disorders often perform worse on
neuropsychological measures than nonclinical controls (Dibartolo, Brown, &
Barlow, 1997; Kizilbash, Vanderploeg, & Curtiss, 2002; and Livingston, Stark,
Haak, & Jennings, 1996). However, exceptions have been noted as well, with
Clinical populations not always showing such a performance (Zalewski,
Thompson, & Gottesman, 1994). One reason that significant differences have
not been consistently found might be due to not controlling for gender, which has
been linked to both levels of anxiety and differential performance in certain
neuropsychological domains.

Gender is a notable factor in examining the effects of anxiety and
cognitive testing. Women are more likely to be diagnosed with anxiety disorders
than men, as well as showing significantly higher levels of anxiety on self-report
measures (Barlow, 2000). Furthermore, gender effects in neuropsychological
functioning have been found, with female subjects performing better on verbal
memory and word fluency tasks (Orsini et al., 1982) and male subjects showing
higher results on motor and visuospatial tasks (Buckelew & Hannay, 1986). In
Martin and Franzen’s (1989) nonclinical student sample, cognitive performance

varied by gender, with males doing more poorly on memory and executive
7

functioning tasks with mild levels of anxiety and females improving or staying
stable. These findings indicate the need to control for gender by cognitive
domain in anxiety studies.

Using this knowledge of the impact of anxiety on cognition enhances our
understanding of those studies that have endeavored to measure the effect of
attachment on cognitive processes.

Impact of Attachment on Cognitive Processes

Over the past decade, a small body of research has been developed
examining the impact of attachment styles on cognitive processes. Early reports
in this area focused on a potential relationship between attachment and general
intelligence. More recently, researchers have employed experimental
methodology to test the effect of attachment on attention and memory; some
studies have made use of overt or unconscious activation of attachment, while
others have simply measured attachment styles without attempts at activation
prior to the cognitive tests. Researchers using the priming paradigms are
utilizing the concept of associative networks. Once a semantic or conceptual
associative network is primed, a spreading activation throughout the network
begins, and related items from that network should be processed more quickly
and efficiently (Solso, 1998). In using the attachment system in this manner, the
different attachment styles predict differential processing of attachment
information. For example, anxious attachment is related to hypervigilance for
rejection once the system is activated, and thus leads to a prediction that
attachment primes will facilitate the processing of threatening attachment

information. Other researchers in this area have either employed different
8

theoretical conceptualizations (such as cognitive schemas) or have examined a
different cognitive process (such as memory). The extant research has been
critically examined in the following sections, which are grouped by outcome
variables or cognitive theory employed.
Attachment and Intelligence

Research on attachment and intelligence in children has provided
inconsistent results. A meta-analysis by van lJzendoom and colleagues (1995)
reported a weak pooled effect size for attachment and various cognitive tests.
However, 20 out of the 25 studies examined reported nonsignificant results.
Indeed, the authors concluded “for all practical purposes the association between
attachment and DQ/IQ is too weak to recommend the routine inclusion of IQ
[Intelligence Quotient] or DQ [Developmental Quotient] tests in order to control

for this type of cognitive difference” (p. 125).

Adult attachment and cognitive measures have been examined together
mainly in the context of a potential confound in narrative measures, but have not
shown a significant relationship between attachment and intelligence (Abraham

et al. 1995, and Bakennans-Kranenberg et al. 1993).

The van lJzendoom and colleagues’ meta-analysis (1995) reported a
salient result: a larger proportion of anxious/ambivalent individuals per study was
related to increased effect size for the attachment-intelligence relationship. The
authors explained this by noting that ambivalent children tend to be “upset in any
strange environment and by any stranger who is trying to interact, for example, in

the context of testing cognitive abilities.” The clear implication here is that

activation of the attachment system is interfering with cognitive processing for
ambivalent children. Since the attachment system is not always activated, this
may be a transient effect, thus depressing IQ scores only due to state anxiety.
However, this has never been tested. Regarding the adult research, adults have
a higher threshold for activation of attachment; in the two studies cited
previously, the AAI was given 2 and 3 months prior to cognitive testing
(Bakermans-Kranenberg et al., 1993; Abraham et al., 1995). It may be that a
different set of results would have been reported had the cognitive tests directly

followed a task intense enough to activate the attachment system.

It is important to note here that IQ tests measure a range of cognitive
processes, from crystallized semantic knowledge to higher-level visual
abstraction, and of course encompassing such discrete processes as attention
and working memory. However, when considering such a vast array of abilities,
the theoretical basis tends to be less integrated, with the result that predictive
power declines. The following sections examine more discrete processes and

theofies.

Attachment Priming and Reaction Time Studies

Measuring reaction time examines the ability to attend to information and
the speed of mental processing during tasks. Common outcome measures in
this research have been Stroop tests and lexical-decision making tasks. In these
studies, researchers have employed either subliminal priming of attachment or
conditioned pairing of attachment-relevant information with tones as ways to

activate the attachment system.

10

Subliminal priming paradigms of threat versus neutral words have been
used in two published articles covering several studies (Mikulincer et al., 2000;
Mikulincer et al., 2002). Mikulincer et al. (2000) used a lexical decision making
task in which a threat or a neutral word prime was subliminally presented directly
before a word or nonword target (some of which were attachment-relevant).
Subjects were instructed to quickly determine whether the targets were words or
nonwords, with the outcome measure being their reaction time. The results
suggested that secure subjects experience attachment system activation via
threat primes (but not neutral primes) and tend to respond to this activation with
associations to positive relationship traits, such as love and companionship.
Additionally, subjects with high levels of anxious attachment were more likely to
be hypervigilant With respect to threat cues or other signals that initiate
attachment activation, and such vigilance regarding attachment appeared to be
actively present even In non-threat conditions. These responses suggest that
secure individuals may be differentiated from those who are anxiously attached.
via their expectations regarding attachment figures: the secure person expects
to be comforted and protected, and the anxiously attached individual carries
expectations of uncertainty or rejection. Mikulincer and colleagues (2002)
employed a similar priming paradigm, but used a lexical decision task with five
different kinds of names or nonwords as targets; and, in a separate analysis,
substituted a Stroop task, but used the same target variables. Subjects were
asked to rapidly judge whether these targets were words or nonwords. The
results suggested that quicker accessibility to names of attachment figures was

related to having high levels of anxious attachment, or when threat cues were
11

presented. Again, this is indicative of a link between anxious attachment and
some degree of vigilancy for attachment information even in non-threat
conditions.

What is problematic with these two studies is that threat primes are
employed and the distinction is not discussed between activating a specific
attachment associative network, as opposed to one composed of general threat.
As will be seen, this is a major criticism of many of the studies in the literature. If
one is not determining differences between the two networks, then questions are
left unanswered as to whether one is activating general threat or attachment.
This same criticism applies to the failure of Mikulincer et al. (2000) to use target
words that are generally positive or negative in tone, but not specific to
attachment. When the dependent variables are indistinguishable in terms of
concept, it is difficult to determine whether broad negative and positive biases
are being utilized or whether associative networks are indeed being activated.

Maier and colleagues (2004) also made use of a subliminal priming
paradigm in their study. They used a neutral prime versus a maternal rejection
prime (“my mom rejects me”). Their outcome measure was a series of self-
evaluative sentences to which subjects were told to answer “yes” or “no” as
quickly as possible. Two examples of the sentences used are “my mom loves
me,” and “I am lovable.” This material was developed to encourage a bias
towards positive responses and mean reaction times were recorded. Results
showed secure subjects with significantly faster reactions to positive self-
statements and self-efficacy in the maternal rejection priming condition, but not

with the neutral prime. This finding is consistent with Mikulincer et al.’s (2000)
12

conclusion that attachment security is related to positive attachment expectations
or beliefs in the face of threat. Maier and colleagues did not find significant
results for the anxious-preoccupied dimension, explaining that the variance of the
dimension may have been too low. However, theory predicts that anxious-
preoccupied subjects have a high vigilance for negative attachment threats, thus
faster reaction times for that material. Therefore, the attachment target content
used by the authors may not have been negatively charged enough to see an
effect.

Baldwin and his colleagues (Baldwin & Kay, 2003; Baldwin & Meunier,
1999) reported on the use of conditioning attachment-relevant material with tones
as a priming paradigm. Baldwin and Kay (2003) conditioned specific tones to
either social acceptance or social rejection, and presented the tones before a
lexical decision task. Subjects were asked to rapidly determine whether the
targets (including various words associated with either attachment or affect) were
words or nonwords. Low levels of anxious attachment were related to slower
reactions to rejection targets, when primed with either acceptance or rejection
tones. It is possible that this outcome reflects an infonnation-inhibition strategy
used in the face of rejection cues by either securely or avoidantly attached
individuals (those attachment styles that correspond to low levels of anxious
attachment) in order to maintain positive expectations of attachment figures
(secure) or to deactivate attachment (avoidant).

In an earlier study, Baldwin and Meunier (1999) asked subjects to identify
and then visualize either an unconditionally accepting or conditionally accepting

person known to them. This visualization was then paired with tones in a
13

conditioning exercise, and a lexical decision task similar to that used by Baldwin
and Kay (2003) was employed, with the addition of word primes (success, failure,
or neutral categories) before targets were presented. Higher levels of security
meant that subjects were more likely to associate to positive expectations of
attachment responses, with preoccupied attachment tending to predict vigilance
for rejection. These findings provided evidence consistent with the previous
research discussed, predicting differential expectations when attachment is
aroused for secure versus anxiously attached individuals.

These two latter studies (Baldwin & Kay, 2003; and Baldwin & Meunier,
1999) provide more specific evidence for the predictive power of attachment
theory in relation to cognition. This is due to employing attachment-relevant
primes or outcome measures. However, since the same lab reported these
results, it invites replication by others for confirmation.
Attachment as a Cognitive Schema

Some researchers using experimental methods to examine attachment
and cognition have not relied explicitly on use of primes leading to a spreading
activation within an associative network. Instead, mental representations
associated with the attachment system, or lntemal working models, have been
conceptualized as cognitive schemas that guide the mental processing of
attachment-relevant information (Miller, 1999; van Emmichoven, van lJzendoom,
De Ruiter, & Brosschot, 2003). Thus, the processing of attachment words has
been predicted to vary by attachment group or dimension, as regulated by
internal working models. For example, a Stroop task was employed by van

Emmichoven and colleagues (2003) in a study using both clinically anxious and
14

nonclinical control groups. The authors postulated that the Stroop task
responses would slow when words were “associated with participants’ concerns
and thus distract[ing] them from their task” (p. 221 ), but that when material did
not fit attachment working model expectancies, defensive exclusion of the
information would occur. Target word groups in the Stroop task consisted of
positive, neutral or threatening categories. A pattern emerged, with the secure-
Clinical and insecure-nonclinical groups processing threatening words more
slowly, thus showing better schema processing, and quicker responses by the
secure-nonclinical and insecure-clinical groups. These findings are supportive of
an overall effect of anxiety on performance. Furthermore, quicker reaction times
of the insecure-clinical group to this task can be explained by insecure
deactivating strategies, while the faster reaction time speed of the secure-non-
clinical may be explained by defensive exclusion of negative and threatening
attachment material in the face of expecting positive information. In a separate
memory task, van Emmichoven et al. (2003) found that insecure clinical subjects
performed more poorly on recall of threat related words compared to secure
subjects. This is in contrast to the previous research cited in the area, and may
be due to the authors combining the (not comparable) groups of dismissing and
anxious-preoccupied subjects for analyses. Unfortunately, the grouping of
insecure versus secure may create a heterogeneous attachment group made up
of members who react to attachment stimuli in widely varying ways, or, the
grouping may be more heavily slated towards certain subclassifications.
Avoidant subjects are more prevalent in the population, and are known for using

such deactivating strategies, as opposed to anxiously attached subjects. Also,
15

as mentioned above, the use of general threats as targets does not allow the
researchers to make claims based on cognitive schemas specific to attachment.
Again, this brings into question whether broad negative schemas or affect are
being tapped here.

Miller (1999) examined attachment and cognition, using the same
framework of attachment as a cognitive schema. Using story recall, the author
reported that both secure and fearful classifications were related to memory for
negative friendship stories, with secure persons recalling friends’ joint activities
and fearful individuals remembering details of activities engaged in separately for
the friends. (The fearful dimension is another insecure attachment construct, that
while not equivalent to anxious-preoccupied attachment, it does provide relevant
information pertaining to attachment-related anxiety.) The author concluded that
attachment-schematic processing occurred despite nonactivation of the
attachment system. An alternative explanation, however, is that these negative
stories naturally elicited some activation of the attachment system due to the
component of a relationship threat. Indeed, other researchers have found that
merely presenting subjects with an attachment-relevant word (such as
separation) provides enough activation of the attachment system to see
differences in accessibility of names of attachment figures (Mikulincer et al.,
2002) and recall of attachment-relevant words (Rowe & Camelley, 2003).

In an unpublished dissertation, Poole (1998) also employed the concept of
attachment as a cognitive schema and presented subjects with a recall task for
sentences describing attachment-relevant and -—irrelevant events. Poole

reported “schema-consistent processing” for both high security (a significantly
16

higher recall of happy attachment sentence events), and fearful attachment
(significantly lower recall for happy attachment events). Inconsistent with other
studies, anxious-preoccupied attachment was significantly related to poor recall
of high anxiety attachment events. Along with this inconsistency with other
studies, the relationship between fearful subjects and poorer recall for happy
events does not speak to a level of vigilance for negative or threatening
attachment stimuli that is predicted by theory.

Attachment Priming and Memory

Studies employing memory outcomes also test the strength of associative
networks with priming paradigms. The work in this area has shown that
attachment style influences the ability to recall attachment-related information
(Hahn, 1995; Pereg & Mikulincer, 2004; and Rowe & Camelley, 2003).

Pereg and Mikulincer (2004) studied the effect of attachment style on
recall of type of information. This study employed an affect-induction task where -
subjects were assigned to one of two conditions: reading an article about kites or
reading an article about a car accident that led to a child’s death. Although this
task was used with the intention of examining the effects of a neutral versus a
negative emotion induction, the nature of the latter article works as a device to
activate the attachment system, and thus the study may be considered to have
employed an overt priming paradigm. After the presentation of the article,
subjects read several headlines (positive, neutral, or negative), which they were
later asked to recall. Subjects with high levels of secure attachment recalled
significantly more positive headlines and fewer negative headlines; the inverse

was found for participants high in anxious attachment (lower recall of positive
17

headlines and higher recall of negative ones). Again, these results are
consistent with research finding different expectations by secure and anxious
groups when presented with attachment events. Pereg and Mikulincer also
noted that the headlines were not specific to attachment, and that this study
contributed the knowledge that attachment associations could Impact memory
processing of non-attachment information. However, such induction of negative
emotion and the resulting outcomes may be due to higher levels of depression
that have been noted in insecure attachment groups. Again, this brings to
question whether a negative bias is the actual mechanism underlying the pattern
of recall, or if it is attachment.

In a different study of attachment and memory, Rowe and Camelley
(2003) primed subjects with different attachment style schemas. These schemas
represented actual relationships reported for subjects, but were not always
congruent with subjects’ overall or global attachment style. Subjects then
engaged in a recall task for a word list which contained positive and negative
attachment and non-attachment words. Participants primed with a secure
schema showed significantly higher recall of positive attachment words, and
those primed with an anxious attachment schema had significantly lower recall of
both positive attachment and positive non-attachment words. The relationship
previously seen between anxious attachment and a strong recall or association
to negative attachment information did not appear here. One explanation for this
could be based on the fact that the attachment schemas primed were not always

the global attachment orientation for subjects.

18

An unpublished dissertation also reported findings on attachment and
memory. Hahn (1995) investigated both actual Childhood memories and word list
recall. Using a priming paradigm with both positive and negative attachment
words, Hahn employed two tasks: one measuring the speed of retrieval of both
negative and positive memories of childhood and parents, and another requiring
free recall of a word list describing positive, negative, and neutral relationships
with mothers and fathers. Anxious-preoccupied attachment was significantly
related to quicker recall of negative childhood and parental memories. Secure
attachment was associated with both significantly higher recall of the positive
maternal descriptors and lower recall of the negative maternal descriptors, when
compared to the anxious-preoccupied group.

In sum, the limited research on attachment and cognitive processes
provides relatively consistent results. Experiments on attachment and reaction
times or mental processing speed have shown high levels of anxious attachment
associated with quicker responses to attachment words after both threat and
neutral primes (Mikulincer et al., 2000; Mikulincer et al., 2002) and rapid
identification of rejection targets with dual primes of failure and conditional
acceptance (Baldwin & Meunier, 1999), and low levels of anxious attachment
related to slower reaction times to rejection targets when primed with acceptance
or rejection contexts (Baldwin & Kay, 2003) and when primed dually with failure
and unconditional acceptance (Baldwin & Meunier, 1999). Secure attachment
has been found significantly related to rapid responses to positive attachment
words when primed with a threat word (Mikulincer et al., 2000), faster reaction

times to positive statements about the self and self-efficacy when primed with
19

maternal rejection (Maier et al., 2004), and quicker recognition of acceptance
targets when dually primed with conditional acceptance and a success context
(Baldwin & Meunier, 1999).

Studies examining attachment and memory have shown insecure and/or
highly anxiously attached subjects with significantly better recall of negative
headlines (Pereg & Mikulincer 2004), negative stories (Miller, 1999), negative
memories of childhood and parents (Hahn, 1995), and negative attachment-
related words (Rowe & Camelley, 2003) when primed with negative affect,
attachment schemas, attachment-relevant words, or negative friendship stories.
Secure subjects have also been shown to have a better recall of happy
attachment events (Poole, 1998), positive maternal descriptors from a word list
(Hahn, 1995), and positive attachment words (Rowe & Camelley, 2003). In
addition, Hahn (1995) also found that secure individuals tend to have a
significantly lower recall of negative maternal descriptors when compared to
anxious-preoccupied persons. These findings suggest that secure and anxiously
attached persons have different expectations when faced with attachment events
or information. Secure persons appear to expect a positive and nurturing
response from an attachment figure and anxiously attached individuals are either
uncertain or expecting rejection. Anxiously attached persons may also
experience a lower threshold for activation of the attachment system. Thus,
there appears to be vulnerability for individuals with high levels of anxious
attachment, due to these negative expectations. The literature on attachment
also notes a link between anxiously attached individuals and varying degrees of

anxiety and depression with the potential for clinical diagnoses (Dozier, Stovall, &
20

Albus, 1999). Therefore, this study focuses on this subgroup of attachment in
order to better understand the mechanisms which may be a part of such
vulnerability.
Synthesizing Predictions of Cognition and Emotion

Researchers employing cognitive paradigms with attachment are relying
on theories of either associative networks or cognitive schemas to explain the
underlying mechanisms. However, both research and theory support predictions
that both cognition and emotion are significant aspects of the functioning of the
attachment system. Here it is hypothesized that associative networks function
Clearly at low levels of attachment anxiety. That is, priming with attachment-
relevant words will activate a mental network or schema devoted to the
attachment system; this, In turn, allows subjects to process attachment-relevant
material more quickly and deeply. However, it is also predicted that associative
networks play a less straightfonrvard role when very high levels of anxiety are
reached. Two factors play into this: first, when the intensity of the attachment
activation is increased; and secondly, when the level of anxious attachment
predisposes a person to reach that threshold of anxiety much more quickly than
those low in anxious attachment. When such an effect is achieved and high
levels of anxiety are broached, it is predicted that cognitive resources are
overloaded and processing of information will slow; also, that vigilance to all
threat cues increases (as opposed to a domain specific vigilance for attachment-
relevant threat). The literature also reveals relationships between subgroups of
attachment and the tendency to process attachment content differently when

attachment is activated. For example, anxiously attached individuals tend to
21

show an increase in vigilance for negative attachment information, resulting In
more efficient processing of and memory for such stimuli. This is not to say that
anxious attachment increases cognitive capacity overall, but that mental
resources appear to be concentrated on particular subsets of information within

the environment.

A variety of methods have been used to prime or activate attachment in
order to measure cognitive outcomes. These range from subliminal priming to
extended immersion in attachment related narratives (Mikulincer et al., 2000;
Mikulincer, Gillath et al., 2001; Mikulincer, Gillath, & Shaver, 2002; Mikulincer,
Hirschberger, Nachmias, 8. Gillath, 2001; Pereg & Mikulincer, 2004). While these
studies have uncovered, for example, that activation of attachment impacts the
cognitive processing of differently valenced emotional content (Pereg &
Mikulincer, 2004), they have not differentiated whether they were activating

attachment-relevant anxiety, another specific anxiety, or broad negative affect.

Another relevant concern in making claims regarding discriminant validity
is a choice of dependent variables which further affirrn that the concept of
attachment is being tapped into. An example would be asking subjects to recall
words related to attachment (comforting, rejecting) as opposed to words related
more to a general sense of threat (failure). It is reasonable to postulate that a
more general threat cue would be “picked up” by the selective attentional sweep
of an activated attachment system, the theory of associative mental networks
predicts that attachment-coded material will be processed more quickly and

efficiently. However, prior research has allowed overlap between attachment-

22

relevant stimuli and general threat stimuli in dependent measures. Therefore, it
is important to choose dependent measures which are clearly related to either

attachment or general threat.

Attachment and general threat material may be difficult to differentiate,
given that threats of many types will activate the attachment system, and that
once the attachment system is activated, there is evidence of psychological and
physiological processes closely related to anxiety (Hofer, 1995). Comparison
with other situation-specific anxieties is one way to begin differentiating
attachment processes from other anxiety processes, however. One such
situation-specific anxiety is spider anxiety. Some level of anxious reaction to
spiders is common; researchers have found that even non-phobic subjects
respond to spiders with significant levels of distress (Vernon & Berenbaum,
2002). Although induction of spider fear naturally leads to some level of threat, a
reaction to such a natural phenomena is not as closely imbued with attachment-
relevant traits such as an embedded association with relationships. Spider
anxiety induction has also been differentiated from a more general anxious
arousal. When cued, both phobic and non-phobic subjects show a selective
attention for spider-relevant information versus a general bias toward any
information containing an element of threat (Kindt & Brosschot, 1998). Use of a
spider-anxiety paradigm as a comparison condition (with both attachment-
relevant and spider-relevant information as outcome measures) is thus predicted

to confirm the discriminant validity of attachment paradigms.

23

RATIONALE AND HYPOTHESES

Attachment theory has been employed to explain and predict behavioral
and emotional responses. In order to explain and predict mental processes as
well during attachment activation, the theory must be further refined. Prior
' research suggests that the attachment system can also be conceptualized as a
unique semantic associative network. By using the commonly employed
cognitive method of priming, it is predicted that the attachment associative
network will be activated, and thus that attachment relevant material will be more
quickly and efficiently processed. Use of mixed stimuli (a list of words containing
both attachment and spider concepts) can help to understand if this extension of
attachment theory is useful and has predictive power.

Related to this are questions of methodology. Different methods of
activation of attachment may tap such emotional intensity, as mentioned above.
If the level of resulting anxiety varies widely among methods, this may help
explain inconsistencies in the literature. Over the past decade, the attention of
attachment researchers has turned Increasingly towards the use of priming and
activating methods or paradigms to examine the effects of attachment on
cognitive processes. These studies have provided an important move forward in
the field of attachment research. Currently, there exists a lack of discriminant
validity regarding the methods used to stimulate and measure attachment
activation. As a result, we do not know that such paradigms are limited to
tapping attachment-relevant working models. Theories of attachment predict that
upon activation of the system, not only are cognitive associative networks

initiated, but also attachment-relevant anxiety. Thus, there is a need to
24

determine to what extent such paradigms are indeed activating the attachment
system in particular, or whether these conditions are the result of a general
arousal of anxiety or negative affect. A parallel concern here is regarding the use
of dependent variables in which an attachment-relevant bias is assumed, but
which may be better understood as inducing broad negative affect instead.

Thus, two issues which need to be concurrently addressed in order to begin
establishing discriminant validity for attachment-activation paradigms are: 1) use
of comparison conditions to test and confirm that attachment paradigms may be
distinguished from methods which activate other specific anxieties, and 2) the
use of dependent variables which specifically tap into the attachment system.

In sum, for cognitive research relying on attachment activation, the choice
of both paradigms and dependent variables is relevant when making claims
based on attachment dimensions. If we are, in fact, are only arousing a general
anxiety or eliciting a recall/reaction bias based on general anxiety or negative
affect, there is little that may be stated unequivocally about the function of the
attachment system.

The following hypotheses are based on the use of four experimental
conditions with equal numbers of subjects of either high or low anxious

attachment assigned to each (see Table 1):

1. Spider Story (to activate a spider associative network)

2. Attachment-Death Story (to overtly activate attachment associative
network)

3. Attachment-Embedded Story (to covertly activate attachment
associative network)

4. Neutral Story (control group condition)

25

Table 1
Group By Experimental Condition With Projected Sample Size

 

 

 

Spider Death Embedded Control
High Anxious
Attachment 40 40 40 40
Low Anxious
Attachment 4O 40 4O 4O

 

Based on the extant research, four overall hypotheses are projected here

to guide the specific hypotheses:
Differentiating Attachment System Activation

It is predicted that attachment system arousal can be differentiated from the
activation of other domains, and that cognitive functioning during suchactivation
can be predicted by the use of semantic associative networks. The theory of
associative networks assumes that particular words or phrases tap into specific
networks and activate them; other words/concepts within that network should
subsequently be easier to process and remember. Therefore, attachment cues
should activate an attachment associative network and spider cues should
activate a spider associative network. If groups are cued with different stories,
yet all are given both attachment- and spider-associated words in a memory task,
we should see differences between groups on recall.

Hypothesis 1: Subjects in the two attachment-relevant conditions will show

a significantly higher recall of attachment words than subjects in the other two
conditions.

26

Hypothesis 2: Subjects in the spider-relevant condition will have significantly
higher recall of the spider words than subjects in the other conditions.
Anxiety Effect

Another extension of theory in exploring cognitive processes post-
activation of attachment is to examine how the emotional response interacts with
cognitive performance. It has been shown that high levels of anxiety related to a
particular domain leads to a higher bias to seeking out information in that area.
For example, the ambivalent attachment group has been shown to be more
highly vigilant for attachment information in the environment. These high levels
of situation-specific anxiety may be thought of as traits delineating groups
vulnerable to a particular domain, or network of stimulus cues. An additional
hypothesis is that an intensity of activation will prove to be more important to
these vulnerable groups than to nonvulnerable groups; after all, one might expect
that a spider-phobic is more emotional and fearful at the sight of many large
spiders than a nonphobic. It is predicted that for such a vulnerable group, a more
emotionally “intense” activation of that domain or information will result in very
high levels of anxiety. In accordance with well-established theories of anxiety
and cognition, we then expect that such high anxiety will interfere with cognitive
performance on resource-heavy tasks.

The activation of both attachment and spider associative networks
stimulates not only cognitive associations, but anxiety specific to each domain,
as well. An effect of anxiety on the relationship between anxious-attachment and
cognitive performance will occur when two factors are present together: (1)

persons with high vigilance to a particular domain are (2) exposed to intense

 

27

activations of that domain. When these two factors are present together this will
result in very high levels of anxiety, increased to the point of interfering in
cognitive processing for these subjects. Therefore, the presence of high
vigilance to either attachment threat cues (high levels of Anxious Attachment) or
spider cues (high levels of Spider Fear) combined with the intense activation of
the respective domains (the Attachment-Death Story or Spider Story) should
result in high anxiety levels. In turn, these particularly high levels of anxiety will
decrease cognitive functions. Thus, when both factors are present, anxiety will
mediate the relationship between the cued domain and memory functioning. If
either of these factors (high vigilance or intense activation) is not present, then
anxiety will not reach this threshold.

It is also predicted that subjects with high attachment vigilance assigned to
the Attachment-Death Story condition will show high levels of anxiety, which will
in turn significantly disrupt their cognitive performance compared to other
subjects. This disruption is expected to extend beyond attachment-relevant
information.

Hypothesis 3: High Anxious Attachment subjects in the Death Story

condition will have a significant relationship between post-experimental

anxiety and both working memory and memory tasks (Letter-Number

Sequencing and total words recalled on the Word List). The higher the

level of anxiety, the poorer the cognitive functioning will be.

Hypothesis 4: High Spider Fear subjects in the Spider Story condition

will have a significant relationship between post-experimental anxiety and

both working memory and memory tasks (Letter-Number Sequencing and
total words recalled on the Word List). The higher the level of anxiety, the

poorer the cognitive functioning will be.

Hypothesis 5: High Anxious Attachment subjects in the Attachment-
Death Story (high intensity) condition will perform significantly more poorly

28

on non-attachment-relevant information processing (Letter-Number

Sequencing) than under the other conditions.
Differentiating Methods of Activating Attachment

Activating the attachment system at different intensities will result in
different outcomes on memory tasks. It is predicted that more intense activations
of attachment will also access enough attachment-relevant anxiety that subjects
in this group will have a lower overall performance on memory recall than
subjects in the less intense attachment activation condition.

Hypothesis 6: Subjects in the condition which intensely activates the

attachment system (Attachment-Death Story) will have a significantly

lower overall recall on the word list task than those in the less intensely

activated condition (Attachment-Embedded Story).
Replication of Patterns of Word Recall

Patterns of recall for the Attachment-Death Story and Attachment-
Embedded Story conditions will be replicated or extended (same patterns seen
for memory tasks as seen in prior research for reaction time tasks and memory
research). Prior research has shown particular patterns for subjects with high
levels of anxious attachment with a variety of paradigms and outcome measures.
However, most priming studies have used reaction time outcomes. The memory
recall studies use the same general theories, show some of the same patterns
consistently, and thus we predict that the use of priming should show similar
results.

Hypothesis 7: All subjects in the two attachment-relevant conditions will

recall significantly more attachment words than non-attachment—related words
(that is, spider words).

29

Hypothesis 8: High and Low levels of Anxious Attachment will differentially
impact word recall for negative and positive attachment words for those
subjects exposed to attachment cues in the experimental condition.
Specifically, flgh Anxious Attachment subjects in the two attachment-relevant
experimental conditions are expected to show a significantly higher recall of
negative attachment words than all other subjects, including the Neutral Story
and Spider Story conditions. Q3! Anxious Attachment subjects in the two
attachment-relevant experimental conditions are expected to have a
significantly higher recall of msitive attachment words than all other subjects,
including the Control and Spider Anxiety conditions.

30

METHOD

Participants

Three hundred and twenty-three female undergraduate subjects were
reanited using the MSU psychology subject pool. Gender differences have been
established for both tests of verbal memory (Orsini et al., 1982) and levels of
anxiety (Barlow, year), indicating the need to control for gender with these
variables. Martin and Franzen’s (1989) findings conclude that male performance
on memory tasks is more sensitive to mild levels of anxiety. These results
indicate that using a female sample is a more rigorous test of the anxiety effect
hypothesis, and at the very least, if using both male and female subjects,
comparisons by gender are necessary. In the interest of controlling the sample

size, a female undergraduate sample of participants was examined for this study.

Subjects ranged in age from 18 to 56 years (M = 19.67, SD =2.73).
Ethnicity of subjects was mainly Caucasian (79%), with a minority identifying as
African American (9%), Asian American (7%), other ethnic group (4%), or
Hispanic (1%). All subjects were undergraduates, reporting their grade level as
freshman (45%), sophomore (25%), junior (21%) or senior (9%). Median annual

household income reported was $80,000 to $89,999.

Attentional processes have been found to be significantly related to
memory tasks as an initial processing system which discriminates essential from
inessential information as a precursor to encoding it in memory; thus, memory
processes rely on attentional ability (Lezak, 1995). The Trails A task, a measure

of simple attention, was added to the design as a covariate. Depression has also

31

been found to decrease memory (Lezak, 1995), and subjects with significant
depression have been shown to have decreased volume of the hippocampus, a
brain structure localized as an important memory processing area (Chronister &
Hardy, 2002).. The Beck Depression Inventory (BDI; Beck, 1987) was added to
use as an additional covariate in case the groups differed significantly for level of

depression.

Upon recruitment via the online surveys, subjects were placed into High
Anxious Attachment or Low Anxious Attachment groups based on their score
from the Reciprocal Attachment Questionnaire (see below). They were then
assigned to one of four experimental conditions. In a 2 (group) X 4 (condition)
design, forty subjects were projected to be assigned to each cell; due to uneven
recruitment, the actual number of subjects per cell varied slightly, with 40
subjects in all cells, except for Low Anxious X Death Story, which ended up with

43 total subjects.

Measures

Demographics. Subjects completed a demographics questionnaire
gathering information pertaining to age, ethnicity/race, family income level, and

years of education completed. See Appendix B.

Attachment. The Reciprocal Attachment Questionnaire (RAQ; West at al.,
1987) measures nine subscales tapping into attachment dimensions (use of the
attachment figure, angry withdrawal, proximity seeking, compulsive care-giving,
availability of the attachment figure, separation protest, feared loss, compulsive

self-reliance, and compulsive care-seeking). A dimension of anxious attachment
32

is obtained by summing the scores of three of these subscales: Separation
Protest, Proximity Seeking and Feared Loss (West et al., 1987). Subjects start
the questionnaire by identifying an attachment figure (defined as “the person you
have been most likely to turn to or depend on for comfort or help when facing
stress”) and are directed to refer to this person while answering the 43
relationship items (6.9., “I feel it is best not to depend on my attachment figure”)
using a 5-point Likert scale (1 =strongly agree to 5=strongly disagree). The three
subscales used to measure anxious attachment have shown good internal
consistency (Proximity Seeking: a=.74; Separation Protest: a=.89; and Feared
Loss: a=.86). The items from these three subscales were combined to form an
Anxious Attachment scale. In the current study, the lntemal consistency score
for this Anxious Attachment scale was good: a=0.83. Scores ranged from 9 to
39, with higher scores indicating higher levels of anxious attachment (M=18.73,

SD=5.38). See Appendix C.

Anxiety and depression. Depression has been found to significantly
impact attention and memory task performance (Kizilbash et al., 2002; Palav et
al., 2000). Thus, it was measured in subjects to be used as a covariate in the
analyses. Level of depression was measured using the BDI (Beck, 1987), a 21-
item self-report questionnaire. Respondents were instructed to choose one of
four self-evaluative statements presented for each item that best reflects how
they had been feeling over the past week. Examples of statements presented for
a single item include “I do not feel sad,” “I feel sad,” “I am sad all the time and

can’t snap out of it," or “I am so sad or unhappy that I can’t stand it.” The BDI’s

33

test-retest reliability has been reported ranging from 0.74 to 0.93, with validity
coefficients of 0.66 against clinical ratings of depression (Lezak, 1995). In the
current study, lntemal consistency was high (Cronbach’s alpha=0.89). Scores
ranged from 0 to 36 (M=7.53, SD=6.99), with a higher score indicating higher

levels of depression. See Appendix D.

Anxiety was measured using the State-Trait Anxiety Inventory (STAI;
Spielberger, Gorsuch, & Lushene, 1970). The STAI contains 20 items regarding
state anxiety and 20 items pertaining to trait anxiety. These self-evaluative
statements are answered using a 1-4 Likert-type scale, ranging from 1, not at all,
to 4, very much so. Examples of items include “I feel jittery" and “I am relaxed.”
State anxiety tends to fluctuate according to circumstance, and trait anxiety is
considered a stable characteristic or dimension of personality. Finney (1985)
reports that test-retest reliability is high for the trait anxiety scale (ranging from
.73 to .86), but low for state anxiety (0.16 to 0.54), which corresponds to the
expectations based on the given definitions. In the current study, Cronbach’s
alpha was strong for Trait Anxiety (a=0.94), as well as for the State Anxiety scale
employed during the experiment (a=0.92). Scores for Trait Anxiety ranged from
20 to 73 (M=39.27, _SQ=11.06), and scores for the post-experimental State

Anxiety ranged from 20 to 69 (M=34.24, SD=9.24). See Appendix E.

Attention and concentration. The Trailmaking Test-Part A from the
Halstead-Reitan Neuropsychological Battery (Reitan & Wolfson, 1985) was used
as a baseline measure of attention and concentration before giving the

experimental condition tasks; this baseline was used as a covariate in the

34

analyses. Good test-retest reliability has been reported for this test (0t=0.79;
Dikmen, Heaton, Grant, & Temkin, 1999). This task requires the subject to scan
a sheet of paper and connect dots in a numerical order, from 1 to 25. Subjects
are instructed to perform this test as quickly as possible, and are asked to stop
and correct any mistakes made during the course of testing. Time to completion
in seconds is used as the measure, then converted into a standard T-Score norm
(controlling for gender, age and education of the subject). In this study, T-Scores
on the Trails A task ranged from 25 (0.6th percentile) to 84 (<99.9th percentile),

with a median score of 55 (69th percentile). See Appendix F.

Memory. Letter-Number Sequencing, a measure of working memory, was
employed to test the hypothesis that strong activation of the attachment system
would significantly impact tasks which increase cognitive demand for subjects
with high levels of anxious attachment (Hypothesis 2b). This task is a
population-nomed subtest of the Wechsler Memory Scale-Third Edition
(Wechsler, 1997). It is a task of mental manipulation and altemate sequencing of
numbers and letters during immediate recall. Subjects are verbally presented
with a series of increasingly longer strings of mixed letters and numbers which
they are instructed to repeat back to the examiner, presenting first the numbers
(in numerical order) and then the letters (in alphabetical order). Subjects are
given 1 point for every trial repeated in the correct sequence. Administration is
discontinued when subjects make mistakes on all 3 trials within one block, or at
the end of the last block. Test-retest reliability for this measure is good

(alpha=.82) (Wechsler, 1997). Scores are the sum of all correct trials, which are

35

then converted into a normative scaled score (controlling for gender, age and
education level). For this study, the scaled scores ranged from 5 (4th percentile)
to 18 (99.6th percentile), with a median score of 9 (37th percentile). See Appendix

K.

A list of positive and negative attachment words was used to measure
differential attachment-related memory processing. The original list, containing
30 positive and 30 negative attachment words, was developed by Rowe and
Camelley (2003). For this study, the list was modified and decreased to 10
positive and 10 negative words. A list of 20 spider-relevant words was
administered interspersed throughout the attachment word list. This set of words
was adapted from Kindt and Brosschot (1998b). This list was employed to test
whether the Spider Story condition could be differentiated from the two
attachment conditions. All subjects received the attachment and spider-relevant
combined list (total of 40 words) as presented in Appendix L. Subjects were
presented with the word list from an audio recording, in order to control for
examiner bias. The subjects were instructed by the examiner to listen carefully,
and told they would be asked to recall the word list after hearing it. Immediately
after the administration, the subject was asked to recall as many of the words
from the list as she could recall, and the examiner marked the recalled items on a
checklist. Scores were derived based on the total number of all words recalled,
total number of all attachment words, total number of spider words, total number

of positive attachment words and total number of negative attachment words.

36

Spider fear. Level of spider fear was measured using The Fear of Spiders
Questionnaire (FSQ; Szymanski & O’Donohue, 1995), an 18-item measure of
pre-existing spider fear. It employs a seven point Likert scale (ranging from 1,
“not at all like you,” to 7, “definitely like you”) to rate items such as “Spiders are
one of my worst fears,” and “I would feel very nervous if I saw a spider now.”
Test-retest reliability has been reported ranging from 0.63 to 0.97, and good
internal consistency for the scale has been determined (a=0.92; Szymanski &
O’Donohue, 1995). In the current study, lntemal consistency of this scale was
high (Cronbach’s alpha=0.96). Scores are a sum of all item ratings. Scores for

this study ranged from 18 to 126 (M=45.11, SD=26.13). See Appendix M.

Procedure

Participants were recruited using the MSU psychology subject pool, and
received course credit for their participation. Subjects participated in two
separate phases of data collection, an initial online phase of answering self-
report questionnaires, and an in-person interview in whichthe experimental
conditions were employed and the dependent variables measured. The initial
online data collection consisted of submitting the consent form (see Appendix A)
and completing five questionnaires in this order: a demographics questionnaire
(see Appendix B), an attachment questionnaire (Reciprocal Attachment
Questionnaire, RAQ; see Appendix C), measures of depression and anxiety
(Beck Depression Inventory, BDI; State-Trait Anxiety Inventory, STAI; see
Appendices D and E), and a spider fear questionnaire (Fear of Spiders

Questionnaire, FSQ; see Appendix M). Subjects then signed up for a time to be

37

interviewed in person on the university campus. The online segment of the study
lasted approximately 30 minutes.

Before the in-person interview, RAQ anxious attachment scores were
calculated. The mean score of 18.5 reported by West and colleagues (1999)
with undergraduate females was used as a high/low cutoff and each participant
was assigned to either the ‘High Anxious Attachment’ (scores 19 and above) or
'Low Anxious Attachment’ group (scores 18 and below) based on her score. The
current study’s mean score of 18.7 on the anxious attachment scale was not
significantly different from West et al.’s (1999) mean [t(322)=.78, p>.05)].
Subjects were then randomly assigned to one of four experimental conditions
(Attachment-Death Story, Attachment-Embedded Story, Spider Story, or Neutral
Story; see Appendices G, H, I, and J), using random block design.

The in-person interview was conducted administering these tasks in this
order: a measure of attention (Trailmaking Test Part A; see Appendix F), the
experimental condition (one of four vignettes will be read and questions
answered; see Appendices G, H, l, and J), a measure of state anxiety (the State
Anxiety items from the STAI; see Appendix E), a working memory task (Letter-
Number Sequencing; see Appendix K), and the word list recall task (a combined
list of 20 attachment words and 20 spider-related words; see Appendix L). At the
end of the interview, participants were thanked for their time, and read a
debriefing statement about the nature of the experiment (see Appendix N).

During the in-person interview, subjects were first given the Trailmaking
Test Part A, in order to measure a baseline of attentional processing as a

covariate for the statistical analyses. The experimenter read the instructions of
38

this task to the subject, then told her when to begin; the task was timed using a
stopwatch. Subjects were then presented with the task corresponding to the
experimental condition they were assigned to. Each condition consisted of a
vignette of comparable length, which subjects were asked to read. The four
vignettes each related an experience of a college-age woman. For the
Attachment-Death Story condition, subjects read about the woman finding out
about the death of her mother and her subsequent grief; for the Spider Story
condition, the vignette recounted the woman’s nightmares about spiders; for the
Neutral Story condition, the vignette was about shopping for a new computer; for
the Attachment-Embedded Story condition, the vignette was the same as the
Neutral Story condition, except that attachment-relevant words (such as
“divorced” and “death”) were embedded in the narrative. These attachment-
relevant words were inserted throughout the vignette using alternative, non-
attachment-related meanings. For example, in the Neutral Story condition
vignette, the sentence “Shayla heard the prices Jared quoted and thought he
must be out of contact with reality,” was changed in the Attachment-Embedded
Story condition to “Shayla heard the prices Jared quoted and thought he must be
completely divorced from reality.” This is intended to act as a non-conscious
priming of attachment. Subjects were then asked to write answers to eight
questions about the vignette, which ask the subject to imagine herself
experiencing what the main character in the reading did.

Subjects then completed the State Anxiety section of the STAI as a
measure of their immediate post-experiment anxiety. Next, subjects were given

the Letter-Number Sequencing task, requiring them to hold an increasingly
39

longer set of items in working memory while repeating numbers and letters
separately, in sequential order, to the examiner. This task was immediately
followed by the word list recall task consisting of both attachment words and
spider-relevant words. This word list was presented in an audio recording (to
avoid examiner bias) to the subject, who was asked to recall as many of the
words as possible. At the end of the presentation of the word list, the examiner
marked the words recalled by the participant on a checklist. A short debriefing
concerning the nature of the experiment was then conducted (see Appendix N for
the script). The entire in-person interview took approximately 20-30 minutes.

Four undergraduate student volunteers and the graduate student
investigator conducted the interviews. Training on the interview protocol
consisted of two weeks of supervised practice administration, including meetings
devoted to issues such as interviewer bias and standardized test administration.
Interviewers then administered the protocol with an 85% accuracy rate or better
on two practice and two supervised interviews, consisting of accuracy on these
items: timing the Trails A task; scoring on the LNS and Word List Recall tasks;
and reading all directions correctly. Subjective feedback about overall interview
administration, including handling questions from subjects, was provided as well.
Weekly meetings and feedback continued throughout the interviewing process.
Furthermore, after interviewers had completed five interviews on their own, they
were observed again, with each eaming an 85% accuracy rate or better on

interview administration.

40

RESULTS
Missing Data, Kurtosis and Skew, and Covariates

Interviews were conducted with 323 subjects, with each experimental cell
having 40 participants, with the exception of the Low Anxious Attachment X
Attachment-Death Story group, which had 43. All subjects interviewed were
included In the data analyses. Missing data points were found for four different
participants on four different items: question #27 on the RAQ, question #21 on
the BDI, question #11 on the SFQ, and the demographic item on income.
Missing data was estimated by substituting sample means for each of the items.
Kurtosis and skewness were examined for all study variables. Significant
kurtosis and/or skew were found on two variables, level of Spider Fear and
Depression. A square root transformation was performed on these two variables,
which brought them into an acceptable range for kurtosis and skewness (see
Table 2).

Table 2
Skew and Kurtosis Values for Transformed Variables

 

 

Variable Pre-Transforrnation Post-Transformation
Skew (SE) Kurtosis (SE) Skew (SE) Kurtosis (SE)

Depression 1.37(.14) 1.67(.27) .17(.14) -.30(.27)

Spider Fear 1.01(.14) .05(.27) .65(.14) -.68(.27)

 

In assigning subjects to conditions, only high/low anxious attachment
status was controlled. An initial 2 (High/Low Anxious Attachment) X 4
(Experimental condition) ANOVA was thus used to determine whether there
would be significantly different levels of Spider Fear among the eight cells.
41

Although the analysis showed that the overall model was not significant, F (7,
315)=1.87, p>.05, post hoc painrvise comparisons indicated that subjects in the
Spider Condition had higher levels of Spider Fear than those in the Attachment-
Death Story Condition [t(161)=2.01, p<.05] as well as subjects in the Attachment-
Embedded Story Condition [t(158)=2.65, p<.01]. Although the experimental
conditions were randomly assigned to subjects, these differences indicate that
the level of Spider Fear was unevenly distributed across experimental condition
groups. This difference in initial levels of Spider Fear was a potentially
confounding influence on subjects’ recall of spider words. Therefore, the level of
Spider Fear was accounted for in the univariate and multivariate analyses by
using it as a covariate. Other covariates used were depression (BDI) and
attentional ability (Trails A). The empirical relevance of using these two variables
as covariates is noted in the Hypotheses section. See Table 3 for means and
standard deviations of the three covariates. Table 4 presents the means and
standard deviations for all of the dependent variables by experimental condition
and level of Anxious Attachment. See Table 5 for means and standard
deviations for independent variables and covariates by experimental condition
and level of Anxious Attachment.

 

 

Table 3
Descriptive Statistics for Analysis Covariates
Variable Mean SD Range Possible
Rang_e_
Trails A 55.11 10.77 25-84 0-100
Depression 7.53 6.99 0-36 0-63
Spider Fear 45.11 26.13 18-126 18-126

 

Note: Score given for Trails A is a T-score

Power Analysis

A power analysis was conducted using the Power and Precision program

(Borenstein, Rothstein, Cohen, Schoenfeld, & Berlin, 2000), based on the sample
42

size, an effect size drawn from previous research, and an alpha of 0.05. An
effect size estimate of 0.19 (approximately a medium effect size) was employed
based on Mikulincer et al.’s (2002) effect size found for the main effect of priming
for the anxious attachment group (eta-squared=0.19). The use of covariates is
one way to increase power without increasing sample size, thus three
appropriate covariates (depression, attention, and spider fear) were added as
part of the study design and as parameters in the power analysis. The expected
power of 0.80 was yielded based on a final sample size of 323 as well as the
statistical analysis that was employed which demands the most power (2 X 4
ANCOVA).

Tests of Hypotheses

Differentiating Attachment System Activation

For Hypothesis 1, a one-way ANCOVA was used to test the prediction that
the two attachment-relevant experimental conditions (Attachment-Death Story
and Attachment-Embedded Story) would show a significantly higher mean score
on the total attachment word list than the other two groups (Spider Story and
Neutral Story groups). A main effect was not found for experimental condition
[F(3, 316)=1.17, p>.05]. See Tables 6 and 7.

For Hypothesis 2, a one-way ANCOVA was used to test the hypothesis
that subjects in the Spider Story experimental condition would have a
significantly higher mean score on spider-relevant words when compared to each
of the other three groups. There was no main effect found for condition [F(3,

316)=1.12, p>.05]. See Tables 8 and 9.

43

Mum—<3 50% 9.00m VENEER—w m m_ $24 50* £020 950% ”902

 

 

 

838.2 a. :28 c. :88 a. :82 8:28 6.38.2 26..

a. :88 3.38.8 a. :86 c. :28 a. :88 838.2 :9... 62w .3562
to
3383 3388 a. :84 c. :28 8. :28 8.38.2 26.. 8an:%
a 38 m 8 38 m a :8 v 8 :2 N 8 :8 N C 38 2 :9... 35585.
3.33.6 8388 8:8... 2:»: 8:88 3.38.2 26.. 69w 525
838.6 8. :88 3. :28 c. :82 2. :88 338.8 :9... 3:68:85.
8. :82 3.38.6 8.32.4 e. :82 8. :8.~ 2.38.2 26.. 62m 698
3.38.6 338.8 3.. :8... a. :2; 3:88 6.32.2 c9: .
3.25 363 mEo>>

mEo>> EoEcomu< EmEcomt< EcEcomt/q. mEo>> EcEcomtc.

wz... 52am .30... m>zmmm2 m>Emom =30». m:o.§:<
oBmtm> coEccoO
_mEoEtcaxm

 

EcEcomuc‘ 2525‘ .6 36> 3 new 3.56200 BEoEthxm 3 moBmtm> 3626qu Le 6:26.53 6.66265 ham acme:

v 2an

8.363 838 88 8.3863 88.8 8.25 2-3 .8253 2-6 88 3.8. .2 838 825.6 883898 882

 

 

 

 

. :8 :88 8.38.8 8.38.8 8.38.8 8.38.3 26..
88:88 8838.8 8.38.» 8.38.8 838.8 :9: 32m .3882
to
8.2 :28 8.838“ S 3388 8.38. 8 8.38.3 26.. 888cm
8 38 8 8 838 8 8 38 2 C :8 8 8 :8 8 :9: 386885.
88:88 8.32.8 8.38.4 8.28.8 8.38.3 26.. 69m 5me
88:88 8. 38.8 8.38.8 88:88 8.38.8 :9... 385832
88:88 8.38.3. 8.38.8 8.32.8 8.38.2 26.. 29m 89%
88:88 8.38.8 8.38.8 8.38.8 8.38.8 c9: .
900m
3mm". bo_xc< EoEcoEE EcEcomtc‘
8:02:82. 82am «cemmeaco 808.383 m:o_xc< w:o..x:<
camtm> coaucoo
_SccEtoaxm
EmEcomfi< m:o..x=< Co

:93 bcm 20.56200 :mEmEtqum 3 823860 new chmtm> Embccqmbs Lox m:§m.§cQ b.8265 Em 8662

m 838

45

Table 6

Summary of One-way ANCOVA Testing Hypothesis 1

 

 

F Values df p 772 Power
Experimental Condition 1 .17 3 .32 .01 .31
Trails A (C) 1.39 1 .24 <01 .22
Depression (C) .03 1 .87 <01 .05
Spider Fear (C) 4.20 1 .04 .01 .53
Corrected Model .57

 

Note. C = Covariate

Table 7

Estimated Marginal Means and Standard Errors Tested in Hypothesis 1

 

 

 

52:21:23“

Total Attachment Word List Recall
Spider Story 4.28 (.19)
Attachment-Death Story 3.86 (.19)
Attachment-Embedded Story 4.29 (.19)
Neutral Story 4.17 (.19)

 

Table 8

Summary of One-way ANCO VA Testing Hypothesis 2

 

 

F Values df p 772 Power
Experimental Condition 1.1 1 3 .34 .01 .30
Trails A (C) 2.26 1 .13 <01 .32
Depression (C) .64 1 .42 <01 .13
Spider Fear (C) .31 1 .58 <01 .09
Corrected Model .39

 

Note. C = Covariate

Table 9

Estimated Marginal Means and Standard Errors Tested in Hypothesis 2

 

Experimental

 

 

Condition Variable
Spider Word Recall
Spider Story 6.01 (.23)
Attachment-Death Story 6.10 (.23)
Attachment-Embedded Story 6.57 (.23)
Neutral Story 6.16 (.23)

 

47

Anxiety Effect

For Hypothesis 3, Pearson correlations were used to test whether post-
experimental State Anxiety impacted performance on memory and working
memory tasks (the total number of words recalled on the Word List Recall Task
and the Letter-Number Sequencing task) for High Anxious Attachment subjects
in the Attachment-Death Story condition. There was no significant relationship
between level of state anxiety and cognitive measures for these subjects (see
Table 10).

For Hypothesis 4, Pearson correlations were used to test whether post-
experimental State Anxiety impacted performance on memory and working
memory tasks (the total number of words recalled on the Word List Recall Task
and the Letter-Number Sequencing task) for High Spider Fear subjects in the
Spider Story condition. A median split on the variable level of Spider Fear was
used to create High and Low Spider Fear groups for the analyses. There was no
significant relationship between level of state anxiety and cognitive measures for

these subjects (see Table 11).

Table 10
Hypothesis 3: Pearson Correlations Between Post-Story Anxiety and Cognitive
Variables for High-Anxious Attachment Subjects in Death Story Condition

 

 

Total Words
LNS Recalled
Post-Story Anxiety .04 -.19

 

48

Table 11
Hypothesis 4: Pearson Correlations Between Post-Story Anxiety and Cognitive
Variables for High—Spider Fear Subjects in Spider Story Condition

 

 

Total Words
LNS Recalled
Post-Story Anxiety -.05 .10

 

For Hypothesis 5, highly anxiously attached subjects in the Attachment-
Death Story condition were expected to perform significantly more poorly on the
Letter-Number Sequencing task than all other subjects. A 2 X 4 ANCOVA was
used to test this hypothesis. The results of the analysis did not support the
hypothesis. The interaction between High/Low Anxious Attachment and
Experimental Condition was not significant [ F(3, 312)=1.08, p>.05] nor were
main effects found for condition [F(3,312)=.38, p>.05] or level of anxious

attachment [F(1,312)=.33, p>.05]. See Tables 12 and 13.

49

Table 12

Summary of Two-wa y ANCO VA Testing Hypothesis 5

 

 

Factors F Values df p 772 Power
Experimental Condition (EC) .38 3 .77 <01 .13
Anxious Attachment (AA) .33 1 .57 <01 .09
EC*AA 1.08 3 .36 .01 .29
Trails A (C) 9.06 1 <01 .03 .85
Depression (C) .42 1 .52 <01 .10
Spider Fear (C) 1.08 1 .30 <01 .18
Corrected Model .77

 

Note. C = Covariate

 

 

 

Table 13
Estimated Marginal Means and Standard Errors Tested in Hypothesis 5
Experimental .
Condition Vanable
Anxious .
Attachment Letter—Number Sequencmg
. High 9.49(.35)
Sp'de' St°ry Low 9.98(.35)
Attachment-Death High 9.86(.35)
Story Low 9.45(.34)
Attachment-Embedded High 10.02(.35)
Story Low 9.87(.35)
High 9.62(.35)
Nem'a' Stay Low 10.27(.35)

 

50

Supplemental analyses. In order to further address the question of
whether an anxiety threshold effect exists, additional analyses were conducted.
Reported levels of post-experimental anxiety and amount of change from
baseline levels were examined. When experimental conditions were grouped
according to “intense activation” (spider and death stories) or “non-intense
activation” (embedded attachment words and neutral computer stories),
significant group differences were found for both reported levels of post
experimental anxiety [t(321)=3.05, p<.01)], and the amount of change in anxiety
from baseline [t(321)=4.99, p<.001)]. Intense condition groups reported
comparatively higher anxiety after reading the stories. Both groups showed a
drop in anxiety from pre- to post-test, with the nonintense group showing a
significantly greater decline. In examining the group means by condition
separately, only the High Anxious Attachment subjects in the Attachment-Death
Story condition showed a mean increase in reported anxiety. Effects of
vulnerability impacted anxiety reports, as well. For subjects in the Attachment-
Death Story condition, High Anxious Attachment persons reported significantly
higher levels of anxiety than the Low Anxious Attachment group [t(81)=-2.30,
p<.05)]. Interestingly, the effect of vulnerability was not observed for High/Low
Spider Fear groups in the Spider Story condition [t(78)=-.40, p>.05)].

Another aspect of an anxiety threshold effect is the degree to which
heightened anxiety impacts cognitive functioning. Correlations did not reveal a
significant relationship between either post-experimental anxiety and total word

recall scores (r=.04, p>.05) or the amount of change in anxiety and total word

51

recall (r=.07, p>.05). A final observation about levels of anxiety is that the group
with the highest reported post-experimental anxiety, the High Anxious
Attachment group given the Attachment-Death Story condition, had a mean
percentile of 58.9, normed by age and gender against a general population. This
percentile rank falls within normal limits, with only scores at the 93rd percentile or
higher generally considered abnormally high.
Differentiating Methods of Activating Attachment

For Hypothesis 6, an independent samples t-test was used to test the
mean score of the total number of words recalled on the Word List Recall Task
for subjects in the Attachment-Death Story condition compared to those in the
Attachment-Embedded Story condition. Subjects in the Attachment-Death Story
condition were predicted to have a significantly lower mean score recall of all
words when compared to the Attachment-Embedded Story group. The
difference between group means was not significant, although it approached
significance [t(161)= -1.96, p=.052], in the direction predicted. See Table 14.

Table 14
Total Word Recall List Scores for the Attachment Conditions

 

 

Condition Mean SD
Attachment-Death Story 9.98 2.86
Attachment-Embedded Story 10.84 2.74

 

52

Replication of Patterns of Word Recall

For Hypothesis 7, it was predicted that subjects in the two attachment-
related experimental conditions would recall significantly more attachment words
than non-attachment words. Subjects in these two experimental conditions were
combined into one group, and a t—test was used to compare Total Attachment
Word Recall scores versus Spider Word Recall scores. Results of the t-test were
significant [t(162)=11.76, p<.001], but did not support the hypothesis. Contrary to
the hypothesis, subjects in the two attachment relevant conditions remembered
significantly more spider-related words than attachment-related words (see Table
15). Indeed, a paired-samples t-test follow-up analysis to this showed that all
subjects recalled more spider words than attachment words [t(322)=14.47,
p<.001]. See Table 16.

Table 15
Mean Word Recall for Combined Attachment Conditions

 

 

 

Word group Mean SD
Total Attachment words 4.09 1.64
Spider words 6.31 2.06
Table 16

Mean Word Recall for All Subjects

 

 

Word group Mean SD
Total Attachment words 4.15 1 .69
Spider words 6.21 2.06

 

53

Hypothesis 8 predicted that the Attachment—Embedded Story and
Attachment-Death Story groups were expected to replicate two patterns of
response in recall found previously in the literature: (1) High Anxious Attachment
subjects in the attachment-relevant experimental conditions were expected to
show a significantly higher recall of negative attachment words than all other
subjects, including those in the Neutral Story and Spider Story conditions, and (2)
Low Anxious Attachment subjects in the attachment-relevant experimental
conditions were expected to have a significantly higher recall of positive
attachment words than all other subjects, including those in the Neutral Story and
Spider Story conditions. These predictions were tested by conducting two
analyses, using a 2 (High/Low Anxious Attachment group) by 4 (Experimental
Condition) ANCOVA for each test. The first ANCOVA was run using positive
attachment words as the dependent variable, and the second with negative
attachment words as the dependent variable. The results of the analysis did not
support the hypotheses. The interaction between High/Low Anxious Attachment
and Experimental Condition was not significant for Positive Attachment words [
F(3, 312)=.37, p>.05], nor for Negative Attachment words [F(3, 312)=1.16,
p>.05]. Main effects were not found for condition or level of anxious attachment
in either of the analyses (see Tables 17, 18, 19, and 20).

Supplemental Analyses

Given the nonsignificant results using the Anxious Attachment subscale of

the RAQ, the nine individual scales were examined in relation to the dependent

variables. Significant results were found for only one of the scales, which are

54

Table 17

Summary of Two-wa y ANCO VA Testing Hypothesis 8: Positive Attachment

 

 

Words
Factors F Values df p 772 Power
Experimental Condition (EC) .64 3 .59 <01 .18
Anxious Attachment (AA) 1.53 1 .22 <01 .23
EC*AA .37 3 .77 <01 .12
Trails A (C) 1.56 1 .21 <01 .24
Depression (C) .41 1 .52 <01 .10
Spider Fear (C) 3.85 1 .05 .01 .50
Corrected Model .48

 

Note. C = Covariate

 

 

 

Table 18
Estimated Mgginal Means and Standard Errors Tested in Hypothesis 8
Experimental .
Condition Vanable
Anxious . .
Attachment Posntlve Attachment Words
. High 2.60(.19)
Spider Story Low 2-21(-19)
Attachment-Death High 2.23(.19)
Story Low 2.22(.19)
Attachment-Embedded High 2.26(.19)
Story Low 2.08(.19)
High 2.24(.19)
Neutral Story Low 2.13(.19)

 

55

Table 19

Summary of Two-wa y ANCO VA Testing Hypothesis 8: Negative Attachment

 

 

Words
Factors F Values df p 772 Power
Experimental Condition (EC) 2.53 3 .06 .02 .62
Anxious Attachment (AA) <.01 1 .96 <01 .05
EC*AA 1.16 3 .33 .01 .31
Trails A (C) .28 1 .60 <01 .08
Depression (C) .28 1 .60 <01 .08
Spider Fear (C) .96 1 .33 <01 .17
Corrected Model .66

 

Note. C = Covariate

Table 20

Estimated Marginal Means and Standard Errors Tested in Hypothesis 8

 

Experimental

 

 

Condition Variable
Anxious .
Attachment Negative Attachment Words

. High 1.78(.19)
Sp'de' Stay Low 1.97(.19)
Attachment-Death High 1.52(.19)
Story Low 1 .73(.19)
Attachment-Embedded High 2.14(.19)
Story Low 2.11(.19)
High 2.19(.19)
Neutral Story Low 1.78(.19)

 

56

presented here. These further analyses were conducted substituting the
Compulsive Care-Seeking scale for the Anxious Attachment variable. As the
assignment to experimental conditions was originally dependent on use of the
High/Low Anxious Attachment variable, these post-hoc analyses resulted in
uneven group numbers (see Table 21).

The analyses related to these hypotheses were re-run, using the
substitute attachment variable: Hypotheses 3, 4, and 8. The analysis for
Hypothesis 8 had promising results with stronger power, but the other two sets

Table 21
Sample Size per Cell Using High/Low Compulsive Care-Seeking

 

 

 

Experimental Condition Compulsive Care-Seeking
High Low
Spider Story 47 33
Attachment-Death Story 40 43
Attachment-Embedded Story 43 37
Neutral Story 43 37
Total 173 1 50

 

of analyses were not significant. Again, Hypothesis 8 predicted these two
patterns: (1) High (Attachment dimension) subjects in the attachment-relevant
experimental conditions were expected to show a significantly higher recall of
negative attachment words than all subjects in the Neutral Story and Spider Story

conditions, and (2) Low (Attachment dimension) subjects in the attachment-
S7

relevant experimental conditions were expected to have a significantly higher
recall of positive attachment words than all subjects in the Neutral Story and
Spider Story conditions. The 2 (High/Low Compulsive Care Seeking) X 4
(Experimental Condition) ANCOVA using Negative Attachment Words as the
dependent variable was not significant for the interaction [F(3,312)=.51, p>.05)].
However, when using Positive Attachment Words as the dependent variable, the
interaction between Compulsive Care Seeking and Experimental Condition was
significant [F(3,312)=4.50, p<.01)]. Main effects were not found for either
condition [F(3,312)=.31, p>.05)] or High/Low Compulsive Care-Seeking
[F(1,312)=.04, p>.05)]. The assumption regarding the equality of error variances
was not violated (Levene’s F(7,315)=1.01, p>.05)]. The effect size of the
interaction was eta-squared=.04. Cohen’s guidelines for interpreting eta-squared
effect sizes are as follows: small=.01, medium=.06, and large=.14. See Tables
22 and 23.

This outcome only partially supported the hypothesis. Subjects in the Low
Compulsive Care-Seeking group/Attachment-relevant experimental conditions
had mean scores on Positive Attachment Words second only to High Compulsive
Care-Seeking subjects in the Spider Condition (see Table 23), but post-hoc t-test
analyses revealed that they did not differ significantly from other subject groups.
There were only significant group differences involving the High Compulsive
Care-Seeking/Spider Story group, who scored higher than four other groups (See
Table 23). These post-hoc analyses need to be interpreted with caution,
however, as the chance of a Type I error increases when multiple tests are

conducted separately and not controlled for using ANOVA or other similar
58

statistical analyses. When the two attachment-relevant conditions were
combined, significant within-group differences were found between Low and High
Compulsive Care-Seeking groups for recall of positive attachment words
[t(161)=2.15, p<.05)]. The difference was found in the direction expected, with
the group low in Compulsive Care-Seeking having a higher recall score.

Table 22

Summary of Two-wa y ANCO VA Testing Hypothesis 8 using Compulsive Care-
Seeking: Positive Attachment Words

 

 

Factors F Values df p 772 Power
Experimental Condition (EC) .31 3 .82 <01 .11
gapulsive Care-Seeking .04 1 .84 (.01 .06
EC*CC 4.50 3 <.01 .04 .88
Trails A (C) 1.21 1 .27 <01 .20
Depression (C) .52 1 .47 <01 .11
Spider Fear (C) 4.77 1 .03 .02 .59
Corrected Model .88

 

Note. C = Covariate

59

Table 23

Estimated Marginal Means and Standard Errors Tested in Hypothesis 8

 

 

 

Experimental '
Condition Vanable
Compulsive . -
Care-Seeking Posrtive Attachment Words
. Higha 2.73(.18)
Spider Story LOWb,c 1_94(,21)
Attachment Death Story LOWap 2.37(-18)
Attachment-Embedded Highb.c 1-93(-13)
Story Lowab 2.43(.20)
Highab 2.24(.18)
Neutral Story Lowb 2,12(,2o)

 

Note. Different subcript letters indicate significant mean differences: b: p<.05, c: p<.01

60

DISCUSSION

The goal of the current study was to refine the predictive power of
attachment theory regarding cognition. The route used to do so was
investigating mechanisms underlying the activation of attachment in cognitive
research. Four main areas of examination regarding concepts and paradigms
were: the differentiation of attachment from other anxiety-related constructs,
whether an anxiety effect might exist, differentiating between methods employed
to activate attachment, and the replication and extension of patterns of word
recall found previously in the literature. In order to achieve this goal, a block
random experiment was conducted with 323 female subjects with two
independent factors: level of anxious attachment (high or low) and memory task
outcomes after one of four conditions (Spider Story, Attachment-Death Story,
Attachment-Embedded Story, and Neutral Story).

Discussion of Results

Differentiating Attachment System Activation

It was hypothesized that activating the attachment system accesses a
cognitive associative network which is specific to attachment. Prior research in
this area has proceeded with this concept as an implicit basis, but this has not
previously been explicitly tested. The first specific experimental prediction was
that subjects who were primed by reading one of the attachment-relevant stories
(a story about death or a story with attachment words embedded in the narrative)
would remember more attachment words from the memory recall list than

subjects in the other two conditions (Spider Story or Neutral Story). This was

61

employed to test whether activation of an associative network dedicated to the
attachment system would occur.

In order to test the discrimination of network activation, a comparative
condition was applied. A narrative created to stimulate the associative network
related to spiders was used. “Priming” of spiders was used, in part, because
attachment activation also induces attachment-related anxiety; the spider task
provided a comparable condition thought to activate both an associative network
and related anxiety which would not overlap with relationship or attachment
themes. This was employed to test whether priming of attachment activated
attachment only and could be differentiated from the activation of other systems.

The second specific prediction was that subjects exposed to the Spider
Story would recall more spider words from the memory recall task than the
subjects in the three other conditions. Unexpectedly, there were no significant
group differences shown in the tests of either of the specific hypotheses. The
failure of the attachment-exposed subjects to show a higher recall for attachment
words diverges from prior studies using a neutral comparison condition of initial
priming (Mikulincer et al., 2000, Mikulincer et al., 2002, and Pereg & Mikulincer,
2004). These studies concluded that subjects primed with attachment cues
showed quicker reaction times to subsequent attachment stimuli. However, one
relevant difference may be in the target words used: Mikulincer and colleagues
(2002) employed names of persons known or not known to the subjects, as well
as non-words, and asked their subjects to identify whether the items were words
or non-words. The authors report that high anxious attachment subjects had fast

processing times for names of attachment figures, in both the neutral and
62

attachment priming conditions. This suggests that high anxious attachment
persons are more readily able to access and process references to their
attachment figures. It is important here to point out a potential bias in choice of
target words. Target words belonging to groups such as people we know or
which link to personal memories are categorized differently in terms of memory
processing than words which fall within our knowledge base of general concepts;
the former is considered episodic memory, the latter, semantic memory. The two
categories of memory have been found to be processed in different areas of the
brain (Solso, 1998). It is also known from past studies in cognitive psychology
that specific kinds of information result in better memory recall. Two such kinds
of information which are relevant to both “names of attachment figures” and
“names of close persons” are (1) deeply processed information and (2)
information which refers to ourselves (Solso, 1998). Thus, names of attachment
figures and other close persons may be processed differently anatomically, more
deeply (set into a previously created, deep associative network, as opposed to a
neutral word like “hat"), and may access self-referential cognition. A high level of
anxious attachment is also related to preoccupation with attachment themes or
events. It could be that Mikulincer and colleagues” (2002) results are a reflection
of a combination of high initial preoccupation with attachment and target words
that are susceptible to better recall (compared to general concepts of
attachment), as opposed to having activated the attachment system and
impacting word recall. Considering further that the kind of prime was irrelevant to

the outcome provides support for this idea.

63

Mikulincer and colleagues (2002) also present a second (similar) study in
the same paper, indicating the same quick reaction to attachment figure names
will occur under two events measured by the study: high anxious attachment, or
the presence of threat cues. Thus, only secure subjects were shown to have
responded differentially to the threat. This may further suggest that increased
vigilance may be induced in secure persons, but can tell us no further about
activation of attachment. Maier and colleagues (2004) also reported on a study
featuring a neutral versus attachment prime. They only report significant
differences for secure subjects, who reacted more quickly to two kinds of positive
statements about the self, when exposed to the attachment prime. Although this
is promising, it should be noted that these subjects did not show significant
differences for attachment statements, and thus, the results only provide
information about a tendency of secure persons to process positive information
more quickly when an attachment cue is used. Again, evidence targeting
whether attachment systems are stimulated may only be inferred and not firmly
concluded from this study. Also, the current study differs in that our measure
focused on high or low anxious attachment, and low levels of anxious attachment
cannot be equated with having a secure attachment classification. Mikulincer
and colleagues (2000) reported on a study whereby “stress word’ priming did
result in fast processing of attachment words (semantic category); however, the
same issue of problematic differentiation of what is being primed existed in this
study as well, with high anxious attachment subjects providing quick reaction
times to attachment words, regardless of prime (and secure subjects only

responding thus to the stress word prime).
64

Therefore, while our finding that attachment priming did not result in
significant differences in recall of attachment words is one that diverges from
these studies, it also starts to provide an answer to a question the other three
studies were not designed to address. First of all, it was 391 shown that initiating
attachment activation (by two methods) clearly related to processing attachment
information more quickly. Given the additional information that one prior study
showed a processing preference for positive information beyond attachment, and
that another showed preferential processing for attachment words when primed
by a “stress word,’ a tentative explanation may be that a cognitive network
associated with attachment is not activated in these studies, but rather a broader
concept, such as negative or positive affect.

It was further found that activating a spider associative network resulted in
m significant differences for those subjects and recall of spider words. This, too,
differs from past research which found that spider priming is related to selective
attention to spider-related information across subjects with varying levels of
spider phobia (Kindt & Brosschot, 1998). It may be that combining the
attachment and spider sets of words together resulted in processing interference.
Subjects in the current study listened to the list being read, which taps into
auditory short-tenn memory. However, the spider words in the list were more
concrete (e.g., spider, web, hairy) and thus easier to form a visual mental image
of than the more abstract attachment words (e.g., caring, supportive, rejecting).
One hypothesis is that easily formed mental images from spoken words may
utilize visual-spatial short-term memory, as well as auditory short-term memory.

Auditory working memory is processed along the phonological loop, and visual-
65

spatial information is processed in short-tenn memory along a different cortical
circuit, the visual spatial sketchpad (Kalat, 1998; Kolb & Whishaw, 2000). These
have been shown to be different anatomical circuits in the brain, and that
activating both at the same time results in more cereme blood flow in the cortex
than activating either circuit alone (Kalat, 1998). Therefore, one alternative
explanation for our results is that having one set of words (spider) which may use
both circuits, and thus accessing more mental resources, would result in higher
recall than a set which relies primarily on just one circuit. Indeed, additional
analyses showed that a_ll subjects recalled significantly more spider words than
attachment words.
Anxiety Effect

The second set of hypotheses predicted that exposing persons with high
vigilance to a specific domain to an intense activation of that domain would result
in anxiety strong enough to disrupt memory processing. During the course of
activation it is believed that worries, fears, or anxiety related to attachment are
engaged, along with an associative network. Spider-fear is an established area
of research, and the exposure to spiders was ascertained to be less likely to
stimulate relationship-based worries or anxiety than many other constructs which
induce anxiety. It was thus chosen to elicit both a different kind of associative
network, as well as anxiety thought to be unrelated to attachment. Both domains
(attachment and spider) were stimulated in comparable conditions. The memory
performance was examined of persons in each condition who had a high initial
vulnerability to activation of that domain. Surprisingly, there were no significant

results for either domain tested (attachment or spiders). For persons high in
66

either anxious attachment or spider fear, there was no relationship between
intensely activating the domain and ability to recall words or letters and numbers.
There are a few different considerations when examining this finding, and
attempting to answer the question of whether an anxiety effect exists.
First of all, one of the current study's premises was that the specific domains
would actually be activated. As drawn from the discussion for the first set of
hypotheses, there was a lack of evidence to support that presupposition. A
secondary question would be whether the domains could be judged to be
intensely activated—again, it is not clear whether specific attachment or spider
domains were activated or whether the methods stimulated a broader negative
affect domain. However, it should be noted that differences in level of anxiety
were seen, with persons exposed to the two “intense” conditions (reading about
spider nightmares or a family death) reporting higher anxiety after the experiment
than those subjects who read similar stories about a computer shopping trip (with
one having embedded attachment words in the narrative). It is also interesting to
note that only the women who read the stories about a mother's death
experienced an increase in state anxiety from previously measured levels, and
those who read either of the computer stories had a significant decrease in
anxiety. Differences across level of vulnerability to vigilance also were found to
impact post-experimental anxiety: high anxious attachment persons endorsed
greater anxiety than those with low anxious attachment, when placed in the
death story condition. A significant difference was not found for high / low levels
of spider fear for women who read the spider story. Although these findings are

suggestive of domain activation, an alternative explanation could be that the two
67

stories employed were different in intensity and quality of stimulation. It may be
that reading a story about the death of a young women’s mother, then asked to
briefly reflect upon what that would be like for oneself, is more intense and more
apt to emotionally effect someone than reading about nightmares regarding
spiders.

Other important aspects to examine concern whether increases in anxiety
occurred, and whether levels were broached that disrupted cognitive processing.
Again, the women reading the death vignette were the only group to show an
increase between baseline levels of state anxiety and anxiety after reading the
story. However, this increase was not significant. Neither levels of post-
experimental state anxiety nor amount of change (increase or decrease)
impacted memory performance. Examining scores normed on a general
population for age provides further insight here. The group of women with the
highest scores (high levels of anxious attachment who read the story about
death) did not have an abnormally high level of post-experimental anxiety,
scoring “within normal limits” at the 59th percentile.

The current findings do not actually diverge from previous studies, as the
sample used was nonclinical. Past research in the area has found that clinical
levels of anxiety impact cognitive functioning when the population studied has
been diagnosed with either an anxiety disorder alone (Dibartolo, Brown, &
Barlow, 1997; Livingston, Haak, & Jennings, 1996), or comorbid with depression
(Livingston, Haak, 8 Jennings, 1996; Waldenstein et al., 1997). However, the
same relationship has not been found for nonclinical samples (Hoffman & al'Absi,

2003; Palav et al., 2000; Waldensten et al., 1997), and in one study, not even for
68

clinically anxious patients (Zalewski, Thompson, & Gottesman, 1994). One point
of support for the prediction is that previous research has found a relation
between high levels of anxious attachment and clinical diagnoses of anxiety
disorders (Fonagy et al., 1996; Rosenstein & Horowitz, 1996; Warren, Huston,
Egeland, & Sroufe, 1997). However, the current study was based on levels of
anxious attachment in a general undergraduate population, as opposed to
groups categorized into attachment classifications, which the studies cited relied
upon. It would appear that a faulty assumption when studying a nonclinical
population is that high levels of domain vigilance coupled with intense activation
of that domain will result in clinically significant levels of anxiety.

What can be concluded from these findings is that women who are initially
more vigilant to attachment issues report an increase in anxiety, and highest
overall levels of anxiety, after being exposed to a story that induces negative
affect. Vigilance to spider themes does not appear to result in a similar increase
for women who read a story about spiders. The current study’s manipulations
did not result in abnormally high levels of state anxiety, which may be necessary
in order to impact cognitive functioning. It is possible that different, more intense
domain exposures would elicit very high levels of anxiety that would impact
working memory. However, the two factors of interest (domain exposure and
vulnerability) appear to influence levels of anxiety separately. Other researchers
have failed to find a link between anxiety and bias for negative information in a
subclinical population (Yovel & Mineka, 2004). These results, along with prior
research, appear to weigh against finding such an effect when studying a

nonclinical population.
69

Differentiating Methods of Activating Attachment

Two different methods of activating attachment were employed, one overt
(the story about death), the other covert (the story with attachment words
embedded). These different methods were meant to reflect differences in
methods currently being employed across studies in this area. It was predicted
that the overt, intense activation of attachment would decrease memory
performance for women when compared to those given the covert, embedded
attachment words vignette to read. A trend was noted in support of the
hypothesis, but the result was not significant.

Interestingly enough, a difference between the two groups was shown
regarding changes in anxiety from pre- to post-experiment. Only women who
read the death story reported increases (although not significantly different from
prior levels); women who read the computer story with embedded attachment
words showed a significant decrease in anxiety. One explanation for this,
however, is that the embedded words made no impact regarding attachment
activation, as the control group who read the same story without attachment
words also showed a significant decrease in anxiety. Furthermore, neither level
of post-experimental anxiety nor amount of change in anxiety was related to the
total recall of words.

No prior studies have directly compared the ability to recall number of
words based on prime comparison. Although several studies have used
attachment or threat type primes with neutral prime groups for comparison
(Mikulincer, et al., 2000; Mikulincer, et al., 2002; Pereg & Mikulincer, 2004), and

others have investigated similar recall tasks of headlines (Pereg & Mikulincer,
70

2004), attachment sentences (Poole, 1995), or word recall (Rowe & Camelley,
2003), reports of overall performance (how many items recalled) were not
reported.

Drawing on a preceding discussion, it is uncertain whether the specific
domains were activated or if a broader domain of negative affect was induced
instead. The present study offered some evidence that the two priming methods
were different regarding the change of anxiety that occurred. However, one
method was not differentiated from the control condition regarding this change in
anxiety, and thus may not have activated the attachment domain. Different
priming methods did not result in significantly different memory performance;
however, a trend in the direction predicted suggests that further examination of
the impact of differing priming methods may enhance research in this area.
Replication of Patterns of Word Recall and Additional Analyses

It was hypothesized that the patterns of recall previously found in the
literature would be replicated or extended to memory performance (versus
reaction time outcomes). First, it was predicted that women who were exposed
to attachment cues (via either the story about death or the story with embedded
attachment words) would remember more attachment words from the word list
than non-attachment words. Unexpectedly, women primed with attachment cues
remembered significantly more non-attachment words (spider words) than
attachment words. In fact, all subjects, regardless of condition, had higher recall
of spider words than attachment words.

Second, it was predicted that patterns specific to attachment status would

be replicated and extended in this study. Specifically, it was expected that
71

women high in anxious attachment, when given attachment cues, would recall
significantly more negative attachment words and fewer positive attachment
words than all other subjects. Conversely, it was predicted that women low in
attachment anxiety, when given attachment cues, would recall significantly more
positive attachment words and fewer negative attachment words than all other
subjects. There were no significant differences between groups for either of the
analyses. These results are contrary to previous findings using attachment or
negative primes and memory recall (Beinstein-Miller, 1999; Hahn, 1995; Pereg &
Mikulincer, 2004; Rowe & Camelley, 2003), as well as similar studies using
reaction time outcomes (Baldwin & Kay, 2003; Baldwin & Meunier, 1999; Maier
et al., 2004; Mikulincer et al., 2000; Mikulincer et al., 2002).

A probable explanation for a lack of results here is a combination of
arguments already addressed: adding concrete, easily visualized words such as
the spider words probably created a processing interference that was
unexpected, the current study employed a measure that only targeted high or low
anxious attachment (with low attachment not being entirely comparable to high
levels of secure attachment), and that it may be these priming methods and
variables employed tap into broader domains than expected.

Another explanation for a lack of results may be found through examining
the additional analyses substituting a different attachment scale from the RAQ,
Compulsive Care-Seeking, for the Anxious Attachment scale. One of the group
differences found was unexpected (women with high levels of Compulsive Care-
Seeking who read the spider story had the highest recall of positive words

compared to other groups). The expected pattern for recall of positive
72

attachment words was found for women exposed to attachment cues (those low
in Compulsive Care-Seeking recalled higher levels of positive attachment words,
and those high in the dimension recalled fewer positive attachment words). An
explanation for these findings may be that reading the attachment-relevant
stories stimulated a preoccupation with negative attachment themes for the
women high in Compulsive Care-Seeking, as expected, but that reading the
stories about spiders activated a negative domain which did not tap into negative
preoccupation with relationship themes, but rather fears, worries or anxiety about
insects that this group was motivated to avoid by focusing on the positive words.
It does not explain, however, why similar results were not found using the
Anxious Attachment scale. This calls into question whether the anxious
attachment measure used captures the characteristics of the construct in a way
similar to other, more commonly used measures. The additional analyses did not
reveal significant results for negative attachment words. Significant outcomes for
securely attached subjects in this area of research have mainly been linked to
positive stimuli, and has been attributed to how secure persons are thought to
respond to attachment themes with positive expectations and memories (Maier et
al., 2004; Pereg & Mikulincer, 2004). This suggests that Compulsive Care-
Seeking may target aspects of secure attachment underlying the results of
previous research. Related questions to ask are: to what extent do different
measures either overlap with each other, solidly defining the dimensions of
anxious and secure attachment, and also, to what degree do these measures
capture characteristics outside the generally agreed-upon limits of attachment

features?
73

It has been questioned in the past whether certain measures or constructs
attributed to attachment are better attributed to personality traits or cognitive
style. Crowell, Fraley, and Shaver (1999) reported on efforts to discern whether
certain attachment dimensions were “redundant” with traits in the five-factor
model of personality. Results showed moderate overlap of the anxious
attachment dimension with neuroticism, and the secure dimension with
agreeableness and extraversion. Pereg and Mikulincer (2004), also reported
interesting findings on attachment and cognitive style. They measured
attachment and asked subjects to rate the causes of hypothetical relationship
scenarios based on perceptions of how intemal/extemal, stable/unstable, and
global/specific they were. When considering the negative relationship scenario,
secure persons tended to rate them as having unstable and specific causes,
whereas the anxiously attached subjects perceived these situations as having
stable, global causes. These attributional patterns are well-established in the
social psychology literature as perceptions of control in negative situations, with
stable, global, and internal attributions linked to vulnerability to depression, and
unstable, specific, and external attributions linked to optimism (Brehm & Kassin,
1996). It should be noted that this was one of a series of attachment studies
conducted, one of which included priming and memory tasks. Using the same
attachment scale (Mikulincer, Florian, & Tolmacz, 1990), these researchers
reported that compared to being given a neutral prime, secure persons given the
negative one had a higher level of positive recall, and that anxiously attached
persons given the negative prime had a lower level of positive recall. A question

to answer would be whether recalling positive information was truly due to the
74

influence of attachment, or whether it could be in part due to the presence or
absence of depression or optimism. This information about the relation of
attachment and personality or cognitive traits suggests the need to discern to
what degree these other traits may be influencing attachment research. The
differing results between the current study and previous research may be an
artifact of using the RAQ Anxious Attachment scale as Opposed to more
commonly employed scales, and there being (unknown) intervening variables
present which are responsible for the results (in the case of past research) or the
lack thereof (in the case of the current research).
Clinical Implications

Insecure attachment has long been established in connection to
psychopathology, including such clinical diagnoses as depression, anxiety
disorders, and borderline personality disorder (Dozier, Stovall, & Albus, 1999).
Understanding the cognitive component of attachment is a valuable aspect of
considering therapeutic interventions in these disorders. In order to better know
how to intervene with the highly anxiously attached individual, it is helpful to
understand the mechanisms leading to (for example) flooded emotional reactions
leading to obsessive ruminations or preoccupation with attachment relationships.
However, the findings in this investigation appear to negate the specificity of
attachment-relevant networks as of import. Rather, clinicians may infer from this
a general vulnerability towards negative affect, and bias to negative cognitions,
for those persons high in attachment anxiety. Additionally, our outcomes suggest
that a focus on traits such as optimism or level of depression may make more

sense than restricting treatments based on attachment cognitive networks. Such
75

traits may drive attachment relationship tendencies; thus, it appears that key
areas for interventions would appear to be broader negative affect and
cognitions. Rowe and Camelley’s work (2003) suggests that priming secure
attachment schemas induces a bias towards positive information, which may in
turn influence or tap into ones’ overall optimistic stance. Future research is
needed, however, to delineate whether this would be a useful intervention with
clinical patients, or whether broader-based interventions (priming positive affect)
would work as well or better.
Strengths and Limitations

The present study has both strengths and limitations which are relevant to
a discussion of the outcomes. First, a significant strength of the study is the
employment of an experimental design with random assignment to conditions
and manipulation of conditions of interest. Utilizing such a design allows for a
closer control of variables and the opportunity to measure causal mechanisms.
Second, the study is the first to empirically test methods of activating attachment
systems using comparison groups. The methodology of the study was also
unique to research in this area, with both independent and dependent variables
specifically targeting attachment themes (as opposed to threat, stress, or
negative affect). This allowed for examination of whether attachment was being
activated or not. The study was also first to use two different kinds of techniques
to activate attachment (as reflected by the variety of techniques used in the
literature) in order to test a comparison of the two methods, and determine
whether they differentially affected memory outcomes. Another strength of the

study is that it was limited to female subjects, as verbal recall and levels of
76

anxiety have been found to differ significantly by gender. This controlled for a
potential confound.

A number of limitations also existed in the study. First, there were
limitations related to measures and procedures. One limitation that has already
been discussed is the addition of spider words to the word list that may have
been much easier to recall, resulting in a possible interference of processing that
could not be statistically controlled for. Another is the use of a word list that was
not previously validated (i.e., did the attachment words really link to an
attachment associative network?) The measurement of anxiety in the study was
also potentially problematic. First, pre- and post-experimental anxiety were
measured in different venues, the pre- version being given online and the post-
version given in-person. There is potential for a social desirability effect here for
only one of the data points, which would impact the measure of how much
subjects’ level of state anxiety had changed after the experimental manipulation.
The second issue is the use of a self report questionnaire for the anxiety
measure; a study employing physiological measurements would be greatly
strengthened, as the conceptual questions are better answered by the subject’s
actual state of anxiety, as opposed to their perception of it. Another limitation
was the use of an attachment measure that hasn't been previously used in this
particular area of research. Although West and colleagues (1998) cross-
validated the measure against Shaver and Hazan’s self-report scale (1997) , the
measures, of course, did not overlap completely. The particular appeal of the
RAQ is that it targets different dimensions of attachment that are, for example,

combined in order to reflect such characteristics as anxious attachment. It is
77

difficult to determine whether the RAQ’s Anxious Attachment scale can be
equated to other measures which have targeted anxious attachment, and have
been used by researchers in this area. However, in examining those other
measures, they vary as to whether they are dimensional or categorical, or if they
even measure the same categories of attachment (Bartholomew & Horowitz,
1991 ). The study and subsequent conclusions would have benefited by the use
of one of the two most commonly applied measures, The Experiences in Close
Relationships Scale (Brennan, Clark, & Shaver, 1998), or the Mikulincer, Florian,
and-Tolmacz (1990) adaptation of Hazan and Shaver's prototype attachment
scale (1987).

Other limitations include the inability to generalize the findings to a
broader population, and the sample size. The current study was limited in that
the sample consisted of young female adults who were mainly Caucasian.
Future studies would benefit from looking at male subjects, differing age groups,
and different racial/ethnic groups. Another issue concerned the number of
subjects in the study. Although an a priori power analysis was conducted, the
effect size used in the analysis was medium, an estimate based on a reaction
time study, as opposed to word recall. The largest effect size in the current study
fell between small and medium (found in the additional analysis with the
Compulsive Care-Seeking variable). It is possible that the effect of the
(manipulated conditions on the memory outcomes are smaller than those
previously found in reaction time studies. An increase in sample size would have
increased the observed power for the one— and two-way ANCOVAs used in the

analyses (power for corrected models ranged from .39 to .77, and averaged .57).
78

If, indeed, the power was not high enough to detect small or small-medium effect
sizes, an increase may have resulted in more significant outcomes for the study.
However, the means reported for groups were not in the direction(s) predicted.
Future Directions

Further research is needed in order to clarify several different issues in the
area of attachment and cognition research. First, it is not clear whether the
priming methods commonly employed in this area actually activate a domain
specifically devoted to attachment, or a broader domain of negative affect. The
current study did not provide evidence that attachment was clearly activated.
However, a methodological issue with target words called into question our
results concerning this. In order to answer this question, it is recommended that
future studies use an experimental design with a control condition and a
comparison condition, as the present study did. Using primes and target words
that are specific to attachment is necessary, as well. Suggestions for improving
upon the current studYs design, as well as that of prior research, are to carefully
consider the attachment and comparison target words used for outcome
measures, and to choose ones that are comparable in how easily they are
processed and retrieved. Levels of ease of processing that might be considered
are whether words are semantic or episodic, deeply processed or shallowly,
concrete and easily visualized or abstract, and degree of self referential
processing. This will enhance studies which predict that certain events, such as
priming attachment, are the reason for particular patterns of recall. Another issue
which needs to be addressed to answer this question is differentiating

attachment activation from broader domains (such as negative affect) as well as
79

more specific domains (such as spider fear). Using comparison condition primes
of negative affect and examining whether cognitive outcomes vary significantly is
one way to begin examining this area.

Second, it would be useful to examine the extent to which personality
traits or cognitive styles may be responsible for cognitive outcomes, versus
attachment styles. One aspect of this kind of exploration is to differentiate
between theoretical issues and measurement problems. Not all self reports of
attachment will measure dimensions and styles comparably, which could impact
cognitive outcome research. Regarding theoretical implications, cognition
studies are needed which delineate where the influence of attachment ends and
personality or cognitive styles begin, or indeed, if there is significant overlap
between them. Initial research should target the traits of neuroticism,
agreeableness, optimism, extraversion. and depressive cognitive styles, as prior
research has found significant correspondence between them and attachment
dimensions.

Third, it appears that differences among priming methods need to be
further examined. This is, of course, secondary to confirming that priming
actually does activate attachment systems. The present study indicated that
different techniques resulted in differing change in state anxiety from baseline.
One particular focus here might be to examine different cognitive performance
tasks, such as reaction time tests and Stroop interference tasks, which may be
impacted by anxiety differently.

Finally, it is important to determine whether anxiety effects on cognition

occur in child samples. A conclusion from the current study is that such an effect
80

is unlikely to be found in an adult nonclinical population. It is important,
however, to answer questions arising from research on attachment groups and
intelligence. Some researchers make a strong case that testing situations may
induce very high levels of anxiety in anxiously attached children; furthermore, it is
thought that attachment is more of a primary system in childhood than for adults.
It would be helpful clinically as well as theoretically to understand whether the
lower cognitive scores found for anxiously attached children in some previous
studies is due to a temporary state of high anxiety or whether these scores are
truly reflective of the usual functioning of these children. Suggestions to
approach this question are using attachment classifications (categorical rather
than dimensional, as much of the childhood research has relied upon the group
classifications), diagnostic interviews, both nonclinical and clinical child samples,

and physiological measures of anxiety rather than self report scales.

81

REFERENCES

Abraham, 8., van ljzendoom, M.H., Scharf, M., Koren-Karie, N., Joels, T.,
8 Mayseless, O. (1995). Stability and discriminant validity of the Adult
Attachment Interview: A psychometric study in young Israeli adults.

Developmental Psychology, 30, 771-777.

Ainsworth, M.D.S., Blehar, M.C., Waters, E. 8 Wall, S. ( 1978). Patterns of
attachment: A psychological study of the Strange Situation. Hillsdale: Lawrence
Erlbaum Associates.

Ainsworth, MOS (1991). Attachment and other affectional bonds across
the life cycle. In C.M. Parkes, J. Stevenson-Hinde, 8 P. Marris (Eds.),
Attachment across the life cvcle (pp. 33-51). New York: Routledge.

Alexander, K.W., Quas, J.A., 8 Goodman, GS. (2002). Theoretical
advances in understanding children’s memory for distressing events: The role of
attachment. _Develobmental Review. 2_2, 490-519.

Bakerrnans-Kranenberg, M.J., 8 van ljzendoom, M.H. (1993). A
psychometric study of the Adult Attachment Interview: Reliability and
discriminant validity. Developmental Psychology, 29, 870-879.

Baldwin, M.W., 8 Kay, A.C. (2003). Adult attachment and the inhibition of
rejection. Journal of Social and Clinical Psychology, 22, 275-293.

Baldwin, M.W., 8 Meunier, J. (1999). The cued activation of attachment
relational schemas. Social Cognition, 17, 209-238.

Barlow, DH. (2000). Unraveling the mysteries of anxiety and its disorders
from the perspective of emotion theory. American Psychologist, 55, 1247-1263.

Bartholomew, K., 8 Horowitz, L.M. (1991). Attachment styles among
young adults: A test of a four-category model. Journal of Personaliy and Social
Psychology, 61, 226-244.

Beck, AT (1987). Beck Depression Inventory. San Antonio: The
Psychological Corporation.

Belsky, J., Spritz, B., 8 Cmic, K. (1996). Infant attachment security and
affective-cognitive information processing at age 3. Psychological Science, 7,
1 1 1-1 14.

Borenstein, M., Rothstein, H., Cohen, J., Schoenfeld, D., 8 Berlin, J.
(2000). Power and Precision 2._0. Englewood: Biostat, Inc.

82

Bowlby, J. (1969). Attaclymentand IOfi 1 : Attachment. New York: Basic
Books.

Bowlby, J. (1973). Attachment and loss 2: Separation: Anxiety and anger.
London: The Hogarth Press.

Bowlby, J. (1980). Attachment and loss 3: Sadness and depression. New
York: Basic Books.

Brehm, SS, 8 Kassin, SM. (1996). Social Psychology. Boston:
Houghton Mifflin Company.

Brennan, K.A., Clark, C.L., 8 Shaver, PR. (1998). Self-report
measurement of adult attachment: An integrative overview. In J.A. Simpson 8

W.S. Rholes (Eds), Attachment theopy and close relationships (pp. 46-76). New
York: Guilford.

Bretherton, l., 8 Munholland, K.A. (1999). lntemal working models in
attachment relationships: A construct revisited. In Cassidy, J., 8 Shaver, P.R.

(Eds.), Handpook of attachment: Theopy, research, and clinical applications (pp.
89-111). New York: The Guilford Press.

 

Buckelew, SR, 8 Hannay, J.H. (1986). Relationships among anxiety,
defensiveness, sex, task difficulty, and performance on various
neuropsychological tasks. Perceptual and Motor Skills, 63, 711-718.

Chronister, R.B., 8 Hardy, S.G.P. (2002). The limbic system. In Haines,
D.E. (Ed.), Fundamental Neuroscience (pp. 493-504). Philadelphia: Churchill
Livingstone.

Crowell, J.A., Fraley, RC, 8 Shaver, PR. (1999). Measurement of
individual differences in adolescent and adult attachment. In Cassidy, J., 8

Shaver, P.R. (Eds), Handbook of attachment: The0[y, research, and clinical
applications (pp. 434-465). New York: The Guilford Press.

Dibartolo, P.M., Brown, TA, 8 Barlow, DH. (1997). Effects of anxiety on
attentional allocation and task performance: An information processing analysis.
Behavio_ur Research and Therapy, 35, 1 101-1 1 1 1 .

Dikmen, S.S., Heaton, R.K., Grant, I., 8 Temkin, NR. (1999). Test-retest
reliability and practice effects of Expanded Halstead-Reitan Neuropsychological

Test Battery. Mmal of the lntemational Neuropsychological Socieg, 5, 346—
356.

Dozier, M., Stovall, K.C., 8 Albus, K.E. (1999). Attachment and
psychopathology in adulthood. In Cassidy, J., 8 Shaver, P.R. (Eds.), Handbook

83

of attachment: TheoLy, research, and clinical applications (pp. 497-519). New
York: The Guilford Press.

Eysenck, MW. (1982). Attention and Arousal: Cognition and
Performance. Berlin: Springer.

Eysenck, MW. (1985). Anxiety and cognitive-task performance.
Personality and Individual Differences, 6, 579-586.

Eysenck, M.W., 8 Calvo, MG. (1992). Anxiety and performance: The
processing efficiency theory. Cognition and Emotion, 6, 409-434.

Feeney, J.A. (1995). Adult attachment, coping style and health locus of
control as predictors of health behavior. Australian Journal of Psychology, 47,
1 71 -1 77.

Finney, JG. (1985). Anxiety: lts measurement by objective personality
tests and self-report. In A.H. Tuma 8 J. Maser (Eds), Anxieg and the anxieg
disorders (pp. 645-679). Hillsdale: Lawrence Erlbaum Publishers.

Fox, N.A., 8 Card, J.A. (1999). Psychophysiological measures in the
study of attachment. In Cassidy, J., 8 Shaver, P.R. (Eds), Handbook of

attachment: Theory, research, and clinical applications (pp. 226-245). New York:
The Guilford Press.

George, C., 8 Solomon, J. (1999). Attachment and caregiving: The
caregiving behavioral system. In Cassidy, J., 8 Shaver, P.R. (Eds), Handbook
of attachment: Theory. regarchLand clinical applications (pp. 649-670). New
York: The Guilford Press.

George, C., West, M., 8 Pettem, O. (1999). The Adult Attachment
Projective: Disorganization of adult attachment at the level of representation. .
In J. Solomon and C. George (Eds), Attachment Disorganization (pp. 318-346).
New York: Guilford Press.

Goldberg, S. (1995). Introduction. In S. Goldberg, R. Muir, 8 J. Kerr

(Eds) Attachment theogy: Social, developmental, and clinical persmctives (pp.
1-15). Hillsdale: The Analytic Press.

Hahn, MB. (1995). Adult attachment style, coping style, defensive
processes, and memory in college women. (Doctoral dissertation, Temple
University, 1995). Dissertation Abstracts lntemational 56, 2370.

 

Hazan, C., 8 Shaver, PR. (1987). Romantic love conceptualized as an

attachment process. Journal of Personalig and Social Psychology,52, 511-524.

84

Hazan, C., 8 Zeifman, D. (1999). Pair bonds as attachments. In Cassidy,

J., 8 Shaver, P.R. (Eds), Handbook of attachment: Theopy, research, and clinical
applications (pp. 336-354). New York: The Guilford Press.

Hofer, MA. (1995). Hidden regulators: Implications for a new
understanding of attachment, separation, and loss. In S. Goldberg, R. Muir, 8 J.

Kerr (Eds), flachment thepry: Social, developmental, and clinical perspectives
(pp. 203-230). Hillsdale: The Analytic Press.

Hoffman, R., 8 al'Absi, M. (2004). The effect of acute stress on
subsequent neuropsychological test performance. Archiveg. of Clinical

Neuropsychology, 19, 497-506.

Kalat, J.W. (1998). Biological Psychology. Pacific Grove: Brooks/Cole
Publishing Co.

Kindt, M., 8 Brosschot, J.F. (1998). Cognitive inhibition in phobia. British
Joumal of Clinical Psychology,37, 103-106.

Kirsch, SJ, 8 Cassidy, J. (1997). Preschooler’s attention to and memory
for attachment-relevant information. Child Development, 68, 1143-1153.

Kizilbash, A.H., Vanderploeg, RD, 8 Curtiss, G. (2002). The effects of
depression and anxiety on memory performance. Archives of Clinical

Neuropsychology, 17, 57-67.

Kolb, B., 8 Whishaw, LG. (2000). Fundamentals of Human
Neuropsychology. New York: W.H. Freeman and Company.

Lezak, MD. (1995). Neuropsychological Assessment. New York: Oxford
University Press.

Livingston, R.B., Stark, K.D., Haak, R.A., 8 Jennings, E. (1996).
Neuropsychological profiles of children with depressive and anxiety disorders.

Child Neuropsychology, 2, 48-62.

Maier, M.A., Bemier, A., Pekrun, R., Zimmermann, P., 8 Grossmann, K.E.
(2004). Attachment working models as unconscious structures: An experimental
test. lntemational Journal of _Behavioral Development, 28, 180-189.

Martin, N.J., 8 Franzen, MD. (1989). The effect of anxiety on
neuropsychological function. lntemational Journal of Clinical Neuropsychology,
_1_1, 1-8.

Mikulincer, M., Bimbaum, G., Woddis, D., 8 Nachmias, O. (2000). Stress
and accessibility of proximity-related thoughts: Exploring the normative and

85

intraindividual components of attachment theory. Journal of Personality and
Social Psychology, 78, 509-523.

Mikulincer, M., Florian, V., 8 Tolmacz, R. (1990). Attachment styles and
fear of personal death: A case study of affect regulation. Journal of Personam

and Social Psychology, 58, 273-280.

Mikulincer, M., Gillath, O., Halevy, V., Avihou, N., Avidan, S., 8 Eshkoli, N.
(2001). Attachment theory and reactions to others’ needs: Evidence that
activation of the sense of attachment security promotes empathic responses.
Journal of Personality and Social Psychology, 81, 1205-1224.

Mikulincer, M., Gillath, O., 8 Shaver, RR. (2002). Activation of the
attachment system in adulthood: Threat-related primes increase the accessibility
of mental representations of attachment figures. Jourflal of Personally and

Social Psychology, 83, 881-895.

Mikulincer, M., Hirschberger, G., Nachmias, O., 8 Gillath, O. (2001). The
affective component of the secure base schema: Affective priming with
representations of attachment security. Journal of Personalig and Social

Psychology, 81, 305-321.

Mikulincer, M., Shaver, PR, 8 Pereg, D. (2003). Attachment theory and
affect regulation: The dynamics, development, and cognitive consequences of
attachment-related strategies. Motivation and Emotion 27, 77-102.

 

Miller, J.B. (1999). Attachment style and memory for attachment-related
events. qumal of Social 3am Personal Relationships, 16, 773-801.

Orsini, A., Alfieri, P., Grossi, D., Raiano, S., Matarese, V., 8 Chiacchio, L.
(1982). Sex differences in adults’ spatial and verbal serial learning tasks. Acta

Neurologica, 4, 44-51.

Palav, A., O’Bryant, S., Ortega, A., 8 McCaffrey, RI. (2000). Effects of
depression and anxiety on CVLT performance in brain-injured persons. Archives

of Clinical Neuropsychology, 15, 813.

Pereg, D., 8 Mikulincer, M. (2004). Attachment style and the regulation of
negative affect: Exploring individual differences in mood congruency effects on

memory and judgment. Personally and Social Psychology Bulletin, 30, 67-80.

Poole, J.A. (1998). The influence of attachment representations on
emotions encoding and memory. (Doctoral dissertation, Simon Fraser
University, 1998). Dissertation Abstracts lntemational 59, 0438.

 

86

Reitan, R. M., andWolfson, D. (1985). The Halitead-Reitan Neurological

Test Batten: Theom and Clinical Inteppretation. Neuropsychology Press, Tucson,
AZ.

Rowe, A., 8 Camelley, KB. (2003). Attachment style differences in the
processing of attachment-relevant information: Primed-style effects on recall,
interpersonal expectations and affect. Personal Relationships, 10, 59-75.

Sobel, ME. (1982). Asymptotic confidence intervals for indirect effects in
structural equations models. In S. Leinhart (Ed.), Sociological Methodology 1982
(pp. 290-312). San Francisco: Jossey-Bass.

Solso, R.L. (1998). C_ognitive Psychology. Boston: Allyn 8 Bacon.

Spence, J.T., 8 Spence, K.W. (1966). The motivational components of
manifest anxiety—Drive and drive stimuli. In C.D. Spielberger (Ed.), Anxieg and
Behavior. London: Academic Press.

Spielberger, C., Gorsuch, R., 8 Lushene, R. (1970). Manual for the State-
Metv Inventory. Palo Alto: Consulting Psychologists Press.

 

Szymanski, J., 8 O'Donohue, W. (1995). Fear of spiders questionnaire.
Journal of Behavior Therapy and Experiental Psychiatm, 26, 31 -34.

van Emmichoven, l.A.Z., van lJzendoom, M.H., De Ruiter, C., 8
Brosschot, J.F. (2003). Selective processing of threatening information: Effects
of attachment representation and anxiety disorder on attention and memory.

Development and Psychopathology, 15, 219-237.

van lJzendoom, M.H., Dijkstra, J., 8 Bus, A.G. (1995). Attachment,
intelligence, and language: A meta-analysis. Social Development, 4, 115-128.

Vemon, LL, 8 Berenbaum, H. (2002). Disgust and fear in response to
spiders. Cognition and Emotion, 16, 809-830.

Waldstein, S.R., Ryan, C.M., Jennings, J.R., Muldoon, M.F., 8 Manuck,
SB. (1997). Self-reported levels of anxiety do not predict neuropsycholgocial
performance in healthy men. Archives of Clinical Neuropsychology, 12, 567-574.

Wechsler, D. (1997). Wechsler Memopy Scale-Third Edition. Psychological
Corporation, New York.

West, M., Rose, S.M., Spreng, S., Verhoef, M., 8 Bergman, J. (1999).
Anxious attachment and severity of depressive symptomatology in women.
Women 8 Health 29 47—56.‘

 

87

West, M., Sheldon, A., 8 Reiffer, L. (1987). An approach to the
delineation of adult attachment: Scale development and reliability. Journal of
Nervous and Mental Disease 175 738-741.

 

West, M., Spreng, S., Casares—Knight, P., Rose, SM, 8 Leiper, R. (1998).
A comparison of two measures to assess adult attachment. Journal of Genetic

Psychology, 159, 345-352.

Yovel, l., 8 Mineka, S. (2004). Hierarchical models of emotional disorders
and emotion-congruent cognitive biases. Personally and Individual Differences,
36, 679-694.

Zalewski, C., Thompson, W., 8 Gottesman, l. (1994). Comparison of
neuropsychological test performance in PTSD, generalized anxiety disorder, and
control Vietnam veterans. Assessment 1, 133-142.

 

88

APPENDIX A

Social Style and Memory Study
Online Consent Form

You are invited to participate in this study on social style and memory.
Your participation is completely voluntary. You will be asked to fill out some
questionnaires as well as participate in tasks measuring attention and memory.
The entire interview will take about 1 hour to complete in two sessions. The first
session is online and lasts up to 1/2 hour. The second session is in-person and
lasts up to ‘/2 hour. You will receive a total of 2 research credits for your
participation in this study when both sessions are completed. All of your answers
and responses will be kept completely confidential. Your privacy will be
protected to the maximum extent allowable bylaw.

If you have any questions or concerns about this research project, you
may contact Peggy O'Toole at (517) 355-9564, or Professor Anne Bogat at (517)
353-0812. If at any time you decide that you cannot or do not want to continue,
you may withdraw from the study. You may also decline to complete any item or
portion of the interview. Further questions about your rights as a research
participant can be answered by David Wright of the University Committee for
Research Involving Human Subjects at Michigan State University at (517) 355-
21 80.

Your submission of this form indicates your voluntary agreement to
participate in this study.

Thank you very much for your help.

89

APPENDIX 3

Demographics questionnaire

Please answer the following questions about yourself:

.5
>

go: Date of birth:

 

Education - indicate highest level attained: (Circle one)
vocational training/trade school/nursing

college - undergraduate (no degree)

college — undergraduate degree BA/BS

other (please indicate type):

 

w eeva

Ethnicity:
African American
Asian American
Caucasian

Hispanic

Native American

Other (please indicate):

 

P

Marital status:

A. Married

B. Partnered (unmarried, living together)
C. Single

D. Divorced

E. Widowed

5. What is your family’s total yearly income reported for the last year
for tax purposes? (Circle one)
under 310,000
310,000-319,999
320,000-329,999
330,000-339,999
340,000-349,999
350,000-359,999
360,000-369,999
370,000-379,999
$80,000-389,999
390,000-399,999
$100,000 and over

FF’FPF‘FPPP?

7. A. Are you currently employed? Yes/No
B. If so, do you work:
a) 40 hours/week or more
b) between 30-40 hours/week
c) between 20-30 hours/week
d) between 10-20 hours/week
e) less than 10 hours/week

8. Have you experienced any of the following in the past 3 years? (Circle
all that apply)

a. Death of a spouse/partner
b. Death of a close family member
c. Death of a good friend

9. Identify your attachment figure. This is the person you have been most
likely to turn to or degnd on for comfort or help when facing stress.

This person may be a parent, a spouse/partner, a friend, a family member or
someone else. You may have several people in your life whom you are close
to in different ways, or it may be difficult to think of one person who means
that much to you. Think of the arson you feel closest to right now.
Please report the relationship you have to this person

(circle one):

Parent—mother
Parent—father .
Spouse/romantic partner
Ffiend

Other family member (specify):
Other (specify):

 

P’S’H‘P’N.‘

 

91

APPENDIX C
Reciprocal Attachment Questionnaire

In this questionnaire, you will find questions about your relationship to your
attachment figure. Remember, your attachment figure is:

o The person you have been most likely to turn to or depend
on for comfort or help when facing stress.

Please circle a number to indicate how you feel each statement

applies to you.

1 2 3 4 5

Strongly Disagree Somewhat agree Agree

1. I turn to my attachment figure for many things, including comfort
and reassurance. 1 2 3 4 5
2. I wish there was less anger in my relationship with my attachment
figure. 1 2 3 4 5
3. I put my attachment figure’s needs before my own.
1 2 3 4 5
4. I get frustrated when my attachment figure is not around as much
as
I would like. 1 2 3 4 5 ‘
5. I feel it is best not to depend on my attachment figure.
1 2 3 4 5
6. I want to get close to my attachment figure but I keep pulling back.
1 2 3 4 5
6. I often feel too dependent on my attachment figure.
1 2 3 4 5
8. I can’t get on with my work if my attachment figure has a problem.
1 2 3 4 5
9. I enjoy taking care of my attachment figure.
1 2 3 4 5
10. I don’t object when my attachment figure goes away for a few
days.
1 2 3 4 5
11. I’m confident that my attachment figure will try to understand my
feelings. 1 2 3 4 5
12. I wish that I could be a child again and be taken care of by my

attachment figure. 1 2 3 4 5

92

1 2 3 4 5

Strongly Disagree Somewhat agree Agree
13. I worry that my attachment figure will let me down
1 2 3 4 5
14. I wouldn’t want my attachment figure relying on me.
1 2 3 4 5
15. I resent it when my attachment figure spends time away from me.
1 2 3 4 5
16. I have to have my attachment figure with me when I’m upset.
1 2 3 4 5
17. I rely on myself and not my attachment figure to solve my
problems. 1 2 3 4 5
18. When I’m upset, I am confident my attachment figure will be there
to listen to me. 1 2 3 4 5
19. I usually discuss my problems and concerns with my attachment
figure.
1 2 3 4 5
20. I feel abandoned when my attachment figure is away for a few
days.
1 2 3 4 5
21. l have a terrible fear that my relationship with my attachment
figure
will end. 1 2 3 4 5
22. I do not need my attachment figure to take care of me.
1 2 3 4 5
23. My attachment figure only seems to notice me when I am angry.
1 2 3 4 5
24. I talk things over with my attachment figure.
1 2 3 4 5
25. It’s easy for me to be affectionate with my attachment figure.
1 2 3 4 5
26. I expect my attachment figure to take care of his/her own
problems.
1 2 3 4 5
27. I’m afraid that I will lose my attachment figure’s love.
1 2 3 4 5
28. I feel lost if I’m upset and my attachment figure is not around.

1 2 3 4 5
93

1 2 3 4 5

Strongly Disagree Somewhat agree Agree
29. I’m furious that I don’t get any comfort from my attachment figure.
1 2 3 4 5
30. I’m so used to doing thing on my own that I don’t ask my
attachment
figure for help. 1 2 3 4 5

31. I’m confident that my attachment figure will always love me.
1 2 3 4 5
32. I’m never certain about what I should do until I talk to my
attachment figure. 1 2 3 4 5
33. I would be helpless without my attachment figure.
1 2 3 4 5
34. Things have to be really bad for me to ask my attachment figure
for help.
1 2 3 4 5
35. I get really angry at my attachment figure because I think he/she
could
make more time for me. 1 2 3 4 5
36. I often feel angry with my attachment figure without knowing why.
1 2 3 4 5
37. I feel that the hardest thing to do is to stand on my own.
1 2 3 4 5
38. I feel that there is something wrong with me because I’m remote
from
my attachment figure. 1 2 3 4 5
39. I don’t make a fuss over my attachment figure.
1 2 3 4 5
40. I don’t sacrifice my own needs for the benefit of my attachment
figure. 1 2 3 4 5
41. My attachment figure is always disappointing me.
1 2 3 4 5
42. When I am anxious I desperately need to be close to my
attachment
figure. 1 2 3 4 5
43. It makes me feel important to be able to do things for my
attachment

figure. 1 2 3 4 5
94

APPENDIX D

BDI

In answering these questions, think about each item carefully and circle the
answer out of the group of 4 items that best reflects how you have been feeling
during the past week.

1. [1] ldo not feel sad.
[2] I feel sad.
[3] I am sad all the time and can't snap out of it.
[4] I am so sad or unhappy that I can't stand it.

2. [1] I am not particularly discouraged about the future.
[2] I feel discouraged about the future.
[3] I feel I have nothing to look forward to.
[4] | feel that the future is hopeless and things cannot improve.

3. [1] I do not feel like a failure.
[2] I feel I have failed more than the average person.
[3] As I look back on my life, all I can see is a lot of failures.
[4] I feel I am a complete failure as a person.

4. [1] I get as much satisfaction out of things as I used to.
[2] I don't enjoy things the way I used to.
[3] I don‘t get real satisfaction out of anything anymore.
[4] I am dissatisfied or bored with everything.

5. [1] I don't feel particularly guilty.
[2] I feel guilty a good part of the time.
[3] I feel quite guilty most of the time.
[4] I feel guilty all of the time.

6. [1] I don't feel I am being punished.
[2] I feel I may be punished.
[3] I expect to be punished.
[4] I feel I am being punished.

95

7.

10.

11.

12.

13.

14.

[1] Idon’t feel disappointed in myself.
[2] I am disappointed in myself.

[3] I am disgusted with myself.

[4] I hate myself.

[1] I don't feel I am worse than anybody else.

[2] I am critical of myself for my weaknesses or mistakes.
[3] I blame myself all the time for my faults.

[4] I blame myself for everything bad that happens.

[1] I don't have any thoughts of killing myself.

[1] I have thoughts of killing myself, but I would not carry them out.
[2] I would like to kill myself.

[4] I would kill myself if I had the chance.

[1] I don‘t cry any more than usual.
[2] I cry more now than I used to.
[3] I cry all the time now.

[4] I used to be able to cry, but now I can’t cry even though I want to.

[1] I am no more irritated by things than I ever am.

[2] I am slightly more irritated now than usual.

[3] I am quite annoyed or irritated a good deal of the time.
[4] I feel irritated all the time now.

[1] l have not lost interest in other people.

[2] I am less interested in other people than I used to be.
[3] I have lost most of my interest in other people.

[4] I have lost all of my interest in other people.

[1] I make decisions about as well as I ever could.

[2] I put off making decisions more than I used to.

[3] I have greater difficulty in making decisions than I used to.
[4] I can’t make decisions at all anymore.

[1]l don't feel that I look any worse than I used to.

[2] I am worried that I am looking old or unattractive.

[3] I feel that there are permanent changes in my appearance that
make me look unattractive.

[4] I believe that I look ugly.

96

15.

16.

17.

18.

19.

20.

21.

[1] I can work about as well as before.

[2] It takes an extra effort to get started at doing something.
[3] I have to push myself very hard to do anything.

[4] I can't do any work at all.

[1] I can sleep as well as usual.

[2] I don't sleep as well as | used to.

[3] I wake up 1-2 hours earlier than usual and find it hard to get back
to sleep.

[4] I wake up several hours earlier than I used to and cannot get back
to sleep.

[1] I don't get tired more than usual.

[2] I get tired more easily than I used to.
[3] I get tired from doing almost anything.
[4] I am too tired to do anything.

[1] My appetite is no worse than usual.

[2] My appetite is not as good as it used to be.
[3] My appetite is much worse now.

[4] l have no appetite at all anymore.

[1] I haven't lost much weight, if any, lately.
[2] I have lost more than five pounds.

[3] I have lost more than ten pounds.

[4] I have lost more than fifteen pounds.

[1] lam no more worried about my health than usual.

[2] I am worried about physical problems such as aches or pains, or
upset stomach, or constipation.

[3] I am very worried about physical problems and it’s hard to think
of much else.

[4] I am so worried about my physical problems that I cannot think
about anything else.

[1] l have not noticed any recent change in my interest in sex.
[2] I am less interested in sex than I used to be.

[3] I am much less interested in sex now.

[4] l have lost interest in sex completely.

97

STAI-Y1

Directions: A number of statements which people have used to describe
themselves are given below. Read each statement and then circle the

APPENDIX E

appropriate number to the right of the statement to indicate how you feel right
now, that is, at this moment. There are no right or wrong answers. Do not
spend too much time on any one statement but give the answer which seems to
describe your present feelings best.

 

Not at
all

Some-
what

Moder-
ately so

Very
much
so

 

I feel calm

[1]

[2]

[3]

[4]

 

I feel secure

[11

[21

I31

[4]

 

I am tense

[1]

[2]

[3]

[4]

 

I feel strained

[1]

[2]

[31

[4]

 

| feel at ease

I1]

[2]

[3]

[4]

 

I feel upset

[1]

[21

[3]

[4]

 

VP’S’IPP’N.‘

I am presently worrying over

possible misfortunes

[1]

[2]

[3]

[4]

 

8.

I feel satisfied

[1]

[2]

[3]

[4]

 

9.

I feel frightened

[1]

[2]

[3]

I4]

 

10.

I feel c5mfortable

[1]

[2]

[3]

[4]

 

11.

I feel self-confident

[1]

[2]

[31

[4]

 

12.

I feel nervous

[11

[2]

[3]

[4]

 

13.

I am jittery

[1]

[2]

[3]

[4]

 

14.

| feel indecisive

[1]

[2]

[3]

[4]

 

15.

I am relaxed

L1]

[2]

[3]

[4]

 

16.

I feel content

[11

[2]

[31

[41

 

17.

I am worried

[1]

[2]

[3]

[4]

 

18.

I feel confused

[1]

[2]

13]

[4]

 

19.

I feel steady f

[1]

[2]

[31

[4]

 

 

20.

| feel pleasant

 

[1]

 

[2]

 

[3]

 

[4]

 

98

 

STAI-Y2

DIRECTIONS: A number of statements which people have used to describe

themselves are given below. Read each statement and then circle the

appropriate number to the right of the statement to indicate how you generally
feel. There are no right or wrong answers. Do not spend too much time on any
one statement but give the answer which seems to describe how you generally

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

feel.

Almost Some- Almost

never times Often always
21. I feel pleasant [1] [2] [3] [4]
22. I feel nervous and restless [1] [2] [3] [4]
23. I feel satisfied with myself [1] [2] [3] [4]
24. I wish I could be as ha as
others seem to be ppy [1] [2] [3] [4]
25. I feel like a failure [1] [2] [3] [4]
26. I feel rested [1] [2] [3] [4]
27.l am “calm, cool, and collected” [1] [2] [3] [4]
28. I feel that difficulties are ilin u
so that I cannot overcome thzm g p [1] [2] [3] [4]
29. I wo too much over some-

, thing thairréally doesn’t matter [1] [2] [3] [4]
30. lam happy [1] [2] [3] [4]
31. l have disturbing thoughts [1] [2] [3] [4]
32. I lack self-confidence [1] [2] [3] [4]
33. I feel secure [1] [2L [3] [4]
34. I make decisions easily [1] [2] [3] [4]
35. I feel inadequate [1] [2] [3] [4]
36. lam content [1] [2] [3] [4]

 

37. Some unimportant thought runs
. through my mind and bothers me

[1]

[2]

[3]

[4]

 

38. I take disappointments so keenly that I
can't put them out of my mind

[1]

[2]

[3]

[4]

 

39. I am a steady person

[1]

[21

[3]

[4]

 

40. I get in a state of tension or
turmoil as I think over my recent
concerns and interests

 

 

[1]

 

[2]

 

[3]

 

[4]

 

99

 

APPENDIX F

Trail Making Test, Part A instructions

Sample item instructions

“On this page are some numbers. Begin at number 1 and draw a line from 1 to
2, 2 to 3, 3 to 4 and so on, in order, until you reach the end. Draw the lines as
fast as you can. Ready—begin.”

Test item instructions

“On this page are numbers from 1 to 25. Do this the same way. Begin at
number 1 and draw a line from 1 to 2, 2 to 3, 3 to 4 and so on, in order, until you
reach the end. Remember to work as fast as you can. Ready—begin.”

100

Appendix G
Spider-Story

This is a story about a 20 year old college student named Shayla who has
nightmares.

Shayla came home from her new job in the research lab feeling very
unsettled. They would be running an experiment requiring the handling of a live
tarantula. Shayla hadn’t expected how squeamish she would feel upon actually
seeing it. The professor had brought one into the meeting and several research
assistants had held it. Shayla had almost backed out of the room, and felt very
anxious. She knew she would have to have to get used to picking it up, but she
couldn’t even imagine holding one of the huge things in her hand.

That night, Shayla had terrible nightmares. The first thing she dreamed of
was entering the lab and putting on one of the testing smocks. She tied a strong
knot in th strings and began to look over her job checklist. Slowly, she became
aware of a tingling sensation on her abdomen. She scratched it several times
before noticing that she began to feel it on her chest and arms as well. Alarmed,
she peered down the front of the smock and saw dozens of newly hatched
tarantulas, climbing inside her clothing, and touching her skin everywhere. She
panicked, trying to rip the smock off, but only tightening the knot further. Shayla
screamed, but she couldn’t keep the creatures from climbing all over her, even
getting into her hair and on her face. The sensation of them was horrible, and
Shayla woke up with a start, shaking.

Later that night, Shayla had more nightmares. The second dream was in
the lab, where she was told to dissect a tarantula. Shayla looked at the one she
had in a jar and was afraid to open the lid, as it seemed very agitated. She
carefully soaked her cloth with formaldehyde and hesitated before swiftly opening
the jar. She tried to clamp down the cloth over the mouth of the jar, but her
tarantula scrambled up and attached itself to the cloth, puncturing it through with
its teeth. Fearfully, Shayla dropped the cloth, and the huge beast swiftly moved
across the table towards her. Shayla shrieked and retrieved her cloth, dropping it
on top of the writhing tarantula. Her breathing returned to normal as she
watched it die slowly. Shayla waited a while before lifting up a corner of the cloth
to look at it. Suddenly, the tarantula came back to life, clicking its mandibles and
running towards Shayla. It leaped off the table and landed on her. She
screamed and tried to shake it off, but it kept climbing and climbing.

Shayla had a third nightmare: she was getting up and making her
breakfast, when she noticed something tickling the back of her neck and her arm.
Annoyed, she twice tried to brush it off of her. Finally, she checked her arm. A
gigantic tarantula at least 8 inches long was climbing up the back of her arm.
Shayla tried to scream, but she felt paralyzed and couldn’t move. The horrible
thing crept slowly up her arm, the beady little wet eyes staring at her. Suddenly,

101

it reared back and then fell forward swiftly, sinking its huge, shiny mandibles into
Shayla’s arm oVer and over again. She tried to knock it off, but it wouldn’t let go.
Shayla woke up screaming.

Word count: 542

Spider-anxiety condition vignette questions

Imagine yourself in Shayla’s situation, going through the same experiences that
she did as you briefly answer the following questions. Refer to the story as
needed.

1.

P’P‘PP’N

What was the reaction that Shayla had when she first saw the tarantula?
Would you have had the same reaction?

Give one word to describe Shayla’s dreams:

Which of the nightmares was the worst one for you to read? Imagining
yourself in the situation, tell us why.

Have you ever had to work in a lab with tarantulas? What do you imagine it
would be like?

What happened in Shayla’s 1St nightmare? What part of Shayla’s first
nightmare did you dislike the most?

In Shayla’s second dream, she has to dissect a tarantula. Reread that
section—what part of that bothered you the most and why?

What happened in Shayla’s third nightmare? Have you ever had a dream
that was similar in any way?

Would you take a position in a research lab working with tarantulas? Why or
why not?

102

APPENDIX H

Attachment-Death Story
Shayla is a 20 year old college student whose mother has just died.

Shayla was trying to finish the reading for her government class, when the
doorbell rang. She looked out the window and saw her older sister on the front
steps. Shayla was surprised to see Maria at her new apartment, but delighted
that she had dropped by. As soon as Shayla opened the door, though, she could
tell that something was wrong. Maria’s face was blotched and it looked as
though she’d been crying. It took a moment to get the words out: “Mom’s been
killed in a car accident.” Shayla could not believe what she was hearing. She
had just seen her mother the day before when they had gone shopping for a new
computer together. Shayla felt as though she’d been struck, and could not
speak. Maria put her arms around her and held her while they both cried.

Their father had been gone on business that day. Maria had left a
message on his cell phone, but he didn’t call back for a few hours. Maria had
laid down on Shayla’s bed to try to relax, and Shayla answered the phone when
he called. It was horrible to tell her dad that her mom had died. She couldn’t
stop crying and there was a shocked silence on her father’s end of the phone.
Finally, he asked where Maria was and if she knew. He told Shayla in a shaking
voice that he loved them both and that he would be there as soon as possible.
Shayla spent the rest of the day thinking of her mom and waiting for her dad to
get there. She thought about the last morning she had seen her mother alive,
trying to remember the last words they had said to each other. She and her mom
had been so close. Shayla couldn’t imagine a world without her in it. She had
always been there for Shayla, when things were tough, to hold her when she
cried, or just to listen. Shayla just couldn’t believe that she would never see her
again.

Shayla’s father arrived that evening and stayed over at Shayla’s
apartment. The next day, the three of them drove home, and Shayla’s dad called
people to let them know and made funeral arrangements. Aunt Barbara came
over to help. Shayla usually lit up when she saw her favorite aunt, but there was
no joy in this occasion. She was exhausted with grief. Over the next few days,
Shayla never thought she could feel pain so intense; she missed her mom so
much it was like a physical aching inside. Her father and sister were in so much
pain, too. No one had expected this.

Maria walked around like she was in a daze and cried a lot. Shayla’s
father mostly kept to himself when he wasn’t on the phone, taking care of
arrangements. Shayla wondered why her mom had to die. She missed her so
much already, she didn’t know if she could get through this. Shayla hadn’t even
had a chance to kiss her mother goodbye, give her one last hug, or anything; she

103

wished with all her heart that her mother was back again, even just to say
goodbye to her and tell her that she loved her, one last time.

WORD COUNT: 542

Attachment Activation condition vignette questions

1.

:PP’N

9191

If you were Shayla, how would you react to the loss of your mother? Would it
be similar or different?

Imagine what it would be like to tell a family member that your mother had
died. How would you do this?

Give two words to describe how Shayla’s father reacts to the death:

Describe the emotion in this story. Is this how you or someone you know has
experienced a death?

How do you think the loss of her mother has impacted Shayla?

How do Shayla and Maria respond in different ways to the loss of their
mother? How are their reactions similar?

What part of this story did you like the least? Why?

If you knew it was going to be the last time you would see your mom, what
would you say to her?

104

APPENDIX l

Attachment-Embedded Words Story
Shayla is a 20 year old college student who has decided to buy a new computer.

Shayla needed a new computer. The first step she took was to decide
what type of computer she would prefer, a laptop or a desktop. Shayla
considered her needs: a desktop would have more power and she could run
more programs on it. However, a laptop was portable and she could take it to
the library and other places to work on papers. Shayla decided that she liked the
idea of being able to bring her computer along with her, so she decided on a
laptop. Next, she thought about how much money she could reasonably spend
on the purchase. Unfortunately, she didn’t have very much in her bank account.
However, her grandmother had called and left a message to expect a check in
the mail both to help out with the computer and because Shayla’s birthday was
coming up soon. Shayla decided she could spend around $1300.

Shayla then thought about computer memory, the processor speed and the
hard drive she wanted and whether or not she wanted a DVD burner. She
decided that she wanted to get at least 256 MB of memory and a 40 GB hard
drive. She also wanted to get a Pentium 4 processor or a Pentium M processor.
Shayla briefly considered getting the HP T-340 with a DVD burner, but when she
saw how much it cost, she quickly abandoned that idea. She then compared
computer pricing for different brands, given what she wanted. She ended up with
three computers that suited her needs: the Compaq V-5000, the Dell X-500, or
an IBM G-550. She liked the Compaq because it had a wide screen; however,
both the Dell and the IBM came with a graphics card that she liked.

The next day, Shayla was doing her reading for her government class, when
she heard the mail carrier dropping off the mail. She rushed outside and saw an
envelope from her grandmother. Inside was a birthday card along with a very
generous check—this was the death of her money troubles! Shayla thought it
would be a great day to go out and get her computer. Besides, her reading for
her government class was far too boring to finish: it was on the separation
between church and state. She decided she would rather not wait and went over
to a local store that was advertising discounts on computers, Anderson’s
Department Store. In the electronics department, a salesperson named Jared
approached and asked if he could assist her. Shayla replied that she was
interested in either a Compaq V-5000, a Dell X-500, or an IBM Thinkpod G-550.
Jared quickly went over some pricing for those units, and noted that they had run
out of Dell X-500s. Shayla heard the prices Jared quoted and thought he must be
completely divorced from reality. She turned to leave and saw an X-500 sitting
alone on the shelf directly behind her. It was available at $300 off the regular
price, so Shayla bought it immediately. She was very happy at the deal she
made. On her way out of the store, Shayla saw a piece of furniture that she could

105

not resist; it was a distressed bedside table which would be perfect for her room.
She bought that as well.

Word count: 541

Attachment Priming condition vignette questions

Imagine yourself in Shayla’s situation, going through the same experiences
above as you briefly answer the following questions. Refer to the story as
needed.

:9 P’Nr‘

9°.“.°’.°'

Would you have decided on a laptop or a desktop?

Give one word to describe Shayla’s shopping trip:

Shayla turns around and sees something while shopping that surprised her—
what did she see? (alone)

What was the reading that Shayla was doing for class when the mail arrived?
Would you enjoy the content of such an assignment (why or why not)?
(separation)

If you had just bought a computer at Anderson’s, would you have considered
the second purchase as well? What was the second purchase? (distressed)
How did Shayla react to Jared’s quoting of the high prices? Would you have
reacted to him differently or in a similar way? (divorced)

What did Shayla do about the HP T-340? Would you have done the same?
(abandoned)

If, like Shayla, you received a check in the mail, would you have come to the
same conclusion that she did? Why or why not? (death)

106

APPENDIX J
Neutral Story

Shayla is a 20 year old college student who has decided to buy a new computer.

Shayla needed a new computer. The first step she took was to decide what
type of computer she would prefer, a laptop or a desktop. Shayla considered her
needs: a desktop would have more power and she could run more programs on
it. However, a laptop was portable and she could take it to the library and other
places to work on papers. Shayla decided that she liked the idea of being able to
bring her computer along with her, so she decided on a laptop. Next, she
thought about how much money she could reasonably spend on the purchase.
Unfortunately, she didn’t have very much in her bank account. However, her
grandmother had called and left a message to expect a check in the mail both to
help out with the computer and because Shayla’s birthday was coming up soon.
Shayla decided she could spend around $1300.

Shayla then thought about computer memory, the processor speed and
the hard drive she wanted and whether or not she wanted a DVD burner. She
decided that she wanted to get at least 256 MB of memory and a 40 GB hard
drive. She also wanted to get a Pentium 4 processor or a Pentium M processor.
Shayla briefly considered getting the HP T-340 with a DVD burner, but when she
saw how much it cost, she quickly relinquished that idea. She then compared
computer pricing for different brands, given what she wanted. She ended up with
three computers that suited her needs: the Compaq V-5000, the Dell X-500, or
an IBM G-550. She liked the Compaq because it had a wide screen; however,
both the Dell and the IBM came with a graphics card that she liked.

The next day, Shayla was doing her reading for her government class,
when she heard the mail carrier dropping off the mail. She rushed outside and
saw an envelope from her grandmother. Inside was a birthday card along with a
very generous check—this was the end of her money troubles! Shayla thought it
would be a great day to go out and get her computer. Besides, her reading for
her government class was far too boring to finish: it was on the division between
church and state. She decided she would rather not wait and went over to a
local store that was advertising discounts on computers, Anderson’s Department
Store. In the electronics department, a salesperson named Jared approached
and asked if he could assist her. Shayla replied that she was interested in either
a Compaq V-5000, a Dell X-500, or an IBM Thinkpod G-550. Jared quickly went
over some pricing for those units, and noted that they had run out of Dell X-500s.
Shayla heard the prices Jared quoted and thought he must be out of contact with
reality. She turned to leave and saw an X-500 sitting there on the shelf directly
behind her. It was available at $300 off the regular price, so Shayla bought it
immediately. She was very happy with the deal she made.

107

On her way out of the store, Shayla saw a piece of furniture that she could

not resist; it was a patterned bedside table which would be perfect for her room.
She bought that as well.

Word count: 542

Control condition vignette questions

Imagine yourself in Shayla’s situation, going through the same experiences
above as you briefly answer the following questions. Refer to the story as
needed.

5‘ P’Nr‘

Would you have decided on a laptop or a desktop?

Give one word to describe Shayla’s shopping trip:

Shayla turns around and sees something while shopping that surprised her—
what did she see?

What was the reading that Shayla was doing for class when the mail arrived?
Would you enjoy the content of such an assignment (why or why not)?

If you had just bought a computer at Anderson’s, would you have considered
the second purchase as well? What was the second purchase?

How did Shayla react to Jared’s quoting of the high prices? Would you have
reacted to him differently or in a similar way?

What did Shayla do about the HP T-340? Would you have done the same?
If, like Shayla, you received a check in the mail, would you have come to the
same conclusion that she did? Why or why not?

108

APPENDIX K

Letter-Number Sequencing Instructions

“I am going to say a group of numbers and letters. After I say them, I want you to
tell me the numbers first, in order, starting with the lowest number. Then tell me
the letters in alphabetical order. For example, if I say, B — 7, your answer should
be 7 - B. The number goes first, then the letter. If I say 9 - C - 3, then your
answer should be 3 — 9 - C, the numbers in order first, then the letters in

alphabetical order. Let’s practice.”

Practice items:

6-F
G- 4
3- W— 5
T- 7— L
1- J- A
Testitems:
1. 1
2
3
2. 1
2
3
3. 1
2
3
4. 1
2
3
5. 1
2
3
6. 1
2
3
7. 1
2
3

(6-F)
(4-G)
(3—5-W)
(7—L-T)
u—A—n

mxmfl<dzmflmar

I l
AGO-ah
I I
L
I

rum
3,:
(I.
S

I
VNUZ
l
a:
I
N G)
l

1
— s (2—7—C—K—S)

9 (3-5-9-P-Y)

Q 2 (2-4-7-E-M-Q)
F - 3 (3-5-8-F-H-W)
2 s

z—

E:
I.
lzl'l'l
| I
0|
I

(2-6-9-A-G -S)
1— c (1-3-43 C-R-Z)
— J - 2 — x -7 (2-5-7-9-J-T-X)
(1-4-8-D-E- --H -R)
A (2-5—6-9-A-H-N-S)
3 (1-3-4-9-B-D-K-R)
z (1-2-6-7-F-M-T-Z)

Nommmxa:
I
ng-‘me
I

—D
--6
-K—
--—1

109

APPENDIX L

Word list task

Instructions to subject:

“I am going to read a list of words to you. Listen carefully and try to remember as
many as you can, because when I am done, I will ask you to tell me all the words
that you recall. You can say them in any order, just say as many of them as you
can.”

Read word list to subject at a rate of 1 second per word; examiner writes
down all words that subject generates

Second trial of word list:

“I am going to read you the same list again. Like before, tell me as many of the
words as you can, in any order. Be sure to say words from the list that you told
me the first time.”

Read word list to subject at a rate of 1 second per word; examiner writes
down all words that subject generates

jealous
spider
protected
web
unloved
creeping
mothering
cobweb

. rejecting

1 0. itching

1 1 . nurtured

1 2. hairy

1 3. insecure
14. rapid

1 5. comforted
1 6.jaws

1 7. clinging

1 8. poisonous
19.trustworthy
20.fast

21 . longing
22. spindly
23. soothing
24. venom
25. yearn

26. crawling
27. caring

28. dangling
29. pining

OPHPPPPN.‘

110

30. bite

31 . attentive
32. moving
33. heartbreak
34. legs

35. supportive
36. scurry
37.farewel|
38. sting

39. reassured
40. bugs

111

APPENDIX M

Fear of Spiders Questionnaire
For this questionnaire, circle one number for each item, rating how well the
statement describes you and how you feel.

1=not at all like you and 7=definitely like you

 

1)lf I came across a spider now, I would 1 2 3 4 5 6
_get help from someone else to remove it. -

 

2)Currently, I am sometimes on the look _

outfor spiders. 1 2 3 4 5 6

 

3)lfl saw a spider now, I would think it will *
harm me.

 

4)lnowthinkalotaboutspiders. 1 2 3 4 5 6

 

5)l would be somewhat afraid to enter a
room now, where I have seen a spider 1 2 3 4 5 6
before.

 

6)l now would do anything to try to avoid a
spider.

 

7)Currently, I sometimes think about
getting bit by a spider.

 

8)lf I encountered a spider now, lwouldn’t
be able to deal effectively with it.

 

9)lf I encountered a spider now, it would ‘
. . . 1 2 3 4 5 6
take a long time to get It out of my mind.

 

 

10)lf I came across a spider now, I would 1 2 3 4 5 6
leave the room.

 

11) If I saw a spider now, I would think it 1 2 3 4 5 6
Will try to jump on me.

 

 

12)lf I saw a spider now, I would ask 1 2 3 4 5 6
someone else to kill It. _
13)lf I encountered a spider now, I would 1 2 3 4 5 6

haveimags of it trying to get me.

 

 

(1):)” I saw a spider now I would be afraid 1 2 3 4 5 6

 

15)Iflsawaspider now,lwould feel very 1 2 3 4 5 6
panicky.

 

 

 

 

 

 

 

 

 

 

 

16)Spiders are one of my worst fears 7 f 1 2 3 4 5 6
17.)| would feel very nervous if I saw a 1 2 3 4 5 6
spider now.

18)lf I saw a spider now I would probably

break out in a sweat and my heart would 1 2 3 4 5 6
beat faster.

 

 

112

APPENDIX N

Debrieflng Information

Thank you for participating in our study. I will now explain the purpose of
the experiment that you just completed. First, I would like to remind you that it is
very important that you do not discuss this information with any of your
classmates who might participate in the experiment. This is to avoid invalidating
the results of the research. All of the data you have provided will be kept
confidential and any identifying information you provided will be used only to
match data from your pre-interview file to today's data, and to provide you with
the appropriate amount of research credits.

The purpose of this experiment is to examine the effects that stable
relationship styles have on the ability to process information. We are examining
the extent to which predisposition to specific relationship styles leads to anxiety,
and how that anxiety may interfere with a person’s ability to remember and
repeat words or numbers and letters. We are also comparing conditions which
impact anxiety and seeing how those conditions differ in their outcomes. If you
have questions about this experiment or would like to obtain information
regarding the results, please contact the student investigator, Peggy O'Toole, at
355-9564.

Thank you again for your interest and participation.

Graduate student investigator: Peggy O’Toole, M.A.

otoolee1@msu.edu
51 7-355-9564

Faculty supervising investigator: G. Anne Bogat, Ph.D.
Department of Psychology

bogat@msu.edu
51 7-353-0812

113

   

iiiiiiijiiiijlijiillilijii