Content uploaded by Dwayne Ball
Author content
All content in this area was uploaded by Dwayne Ball on May 02, 2022
Content may be subject to copyright.
Society for Consumer Psychology
The Role and Measurement of Attachment in Consumer Behavior
Author(s): A. Dwayne Ball and Lori H. Tasaki
Source:
Journal of Consumer Psychology,
Vol. 1, No. 2 (1992), pp. 155-172
Published by: Society for Consumer Psychology
Stable URL: http://www.jstor.org/stable/1480363
Accessed: 30/04/2010 10:05
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=socconpsych.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
Society for Consumer Psychology is collaborating with JSTOR to digitize, preserve and extend access to
Journal of Consumer Psychology.
http://www.jstor.org
JOURNAL OF CONSUMER PSYCHOLOGY, 1(2), 155-172
Copyright ? 1992, Lawrence Erlbaum Associates, Inc.
The Role and Measurement of
Attachment in Consumer Behavior
A. Dwayne Ball and Lori H. Tasaki
Department of Marketing
University of Nebraska-Lincoln
Of recent interest in consumer behavior research is the consumer's use
of owned
possessions
to develop
and maintain
self-concept.
This
study
presents
a measure
of a central
concept
in this area-attachment.
A conceptual
definition of the
construct is proposed
and
is related to social-cognitive
theories
of the self.
The
role of attachment
in the relationship
between
people
and possessions
is dis-
cussed. Evidence is presented
for
the
reliability
and
predictive
validity
of a simple
measure of attachment and
for the discriminant
validity
of the construct. Rela-
tionships
between attachment and other important
consumer behavior
con-
structs
are explored.
Over the past 10 years, consumer behavior theory has begun to concern itself
with postpurchase psychological processes other than primarily satisfaction
and complaining behaviors. Some examples are the experiential
dimensions of
product ownership (Hirschman & Holbrook, 1981; Holbrook & Hirschman,
1982;
Marks, Higgins, & Kamins, 1988), emotional responses
to products after
purchase (Westbrook, 1987), ritualized behaviors toward possessions
(McCracken, 1986, 1988; Rook, 1985; Tetrault & Kleine, 1990), and changes
in involvement over time (Richins & Bloch, 1986).
In particular, consumer researchers have suggested that possessions play a
role in maintaining and supporting the consumer's self-concept and sense of
identity (Belk, 1987, 1988, 1990; Csikszentmihalyi
& Rochberg-Halton, 1981;
Myers, 1985; Schultz, Kleine, & Kerman, 1989). Simultaneously, there has
been a revival of interest among psychologists in the concepts of self and
identity, which usually expand on the issue of identity development most
closely associated in the past with Erickson's (1959) theories.
Some psychological theorists adhere to a social-cognitive tradition (e.g.,
Requests for reprints should be sent to A. Dwayne Ball, Department of Marketing, College
of Business Administration, University of Nebraska-Lincoln, Lincoln, NE 68588-0492.
156 BALL AND TASAKI
Gergen, 1971; Greenwald, 1988; Greenwald & Pratkanis, 1984; Kihlstrom &
Cantor, 1984; Markus, 1977), and some perceive the self as an agent which
intergrates
the individual's experience (e.g., Blasi, 1988; Deci & Ryan, 1991;
Loevinger, 1976).
A central issue from a consumer
behavior
perspective
is the extent to which
an owned object serves the functions of defining and maintaining the self-
concept or identity of a consumer. One would expect different affects and
behaviors toward an object that serve these functions than toward an object
that does not. One protects and cultivates one's self. For example, if a home,
piece of furniture,
article of clothing, and so on constitute part of a consumer's
identity, we might expect more protective behaviors, greater effort spent on
maintaining
the object, and greater
emotional difficulty
in accepting deteriora-
tion or loss of the object than if the object is not so much a part of identity.
In this subfield of postpurchase
consumer
behavior, most interesting phenom-
ena are related to, or extensively modified by, the extent to which the object
is part of the consumer's self.
The study presented here attempts to provide a measurement of the con-
struct of attachment-the extent to which an individual uses an object to
develop and maintain a cognitive structure of self. This construct must first
be discussed in light of theories of the self before its measurement may be
considered. Social-cognitive theories, although not prominent in the posses-
sion-and-self literature in consumer behavior, offer a useful basis for this
discussion.
THE SELF AS COGNITIVE STRUCTURE
Greenwald (1988) presented a theory of self-development
in which the self is
an organization
of knowledge. For example, some statements
pertaining
to this
organization may be: "I am a kind person," or "I will refuse to engage in
unethical
business practices
no matter how much my company will reward it."
The kinds of elements present
in the knowledge structure can be characterized
by five stages of mental representations
of the self: features,
objects, categories,
propositions,
and schemata. The more advanced stages of complexity in repre-
senting the self as a knowledge structure (propositions and schemata) are
available by late adolescence, although their content certainly continues to
change through adulthood.
Propositions relevant to the self are generally of the form: "I am (quality
or category membership)," "I have (object, belief, etc.)," or "I (will behave,
am behaving, or shall behave in some fashion)." The emergent property of
schemata
is consistency, such that a field of propositions
begins to be processed
and integrated into a self-consistent whole. The propositions to be integrated
ROLE AND MEASUREMENT OF ATTACHMENT 157
into a schema are evaluated on the basis of "narrative coherence, analogy,
logical proof, cognitive consonance ... cognitive balance ... self-consistency
... legality ... morality ... and empirical validity .. ." (Greenwald, 1988,
p. 32). Individuals develop schemata that explain their personal histories and
current behaviors in light of causes attributed to themselves or other agents
(narrative, causal, and inferential schemata) and form a framework within
which to evaluate their own behaviors and fix self-worth (evaluative self-
schemata).
Propositions and schemata relevant to the self may include objects of con-
sumption. If an object is included in an evaluative self-schema, that object is
tied to the individual's self-worth. A very simple schema along such lines
would be a set of three propositions: (a) to own a large, impressive house is
evidence of some good personal quality; (b) I own a large, impressive house;
and (c) therefore I have this good quality. Although an individual may never
say this explicitly, even to himself or herself, the ownership of the object
nonetheless supports self-worth.
Greenwald
(1988) discussed four facets of the self: the diffuse
self, the public
self, the private self, and the collective self. These facets are all present to
varying
extents in normal adults. Each facet reflects a different
ego task (Breck-
ler & Greenwald, 1986) and basis for self-evaluation with some relevant audi-
ence of the self or others. The ego task in the diffuse self (which is the unformed
proto-self present
in a young child) is merely hedonic satisfaction,
and the basis
for self-worth is finding it.
In other conceptions of self, there have also been facets or divisions. The
private self of the social-cognitive approach has rough correspondences to
Erikson's (1959) ego-identity, Csikszentmihalyi and Rochberg-Halton's
(1981) "idiotic self," Lee's (1990) psychological dimension of the self, and
Schultz et al.'s (1989) individuation dimension. The public and collective
selves of the social-cognitive approach have similarities to Erikson's self-
identity, Lee's social dimension of the self, and Schultz et al.'s integration
dimension.
In Greenwald's (1988) social-cognitive approach, the ego task for the
public self is social recognition and status, and the basis for self worth is the
approval of others. The private self seeks individual achievement up to some
personal standard, and the collective self seeks to attain the goals of a refer-
ence group, which have been internalized by the individual as his or her own
goals.
The public and private facets are of interest in understanding the role of
possessions in the definition and maintenance of the adult self. The collective
self is also of interest, but, as explained later, is beyond the scope of this article.
The self plays to several audiences in the support of self-worth:
a private, inner
audience (private self) and audiences of others (public self) and internalized
158 BALL AND TASAKI
others (collective self). A self-directed individual is playing primarily to the
private audience, an other-directed individual to the public audience, and a
socially responsible individual primarily to the internalized collective audi-
ence.
The focus of this article is the extent to which a possession comes to support
the self, rather than the particular
audience from which a sense of self-worth
is obtained. However, note that few individuals are likely to be playing entirely
to one audience. Therefore, measurement of attachment must take both the
inner and external audiences into account.
ATTACHMENT
The number
and kinds of self-propositions
that could include an owned object
are certainly very large. When schemata (i.e., combinations of propositions)
are considered, it is apparent that there are an infinite number of ways that
an individual can cognitively "use" a possession, a desired possession, or even
previously discarded possession as support for self-worth. To use all such
propositions or schemata in a measure of attachment is not feasible. In any
case, the attachment construct (defined next) concerns the extent to which the
object is used for self-concept support, rather than the particular propositions
and schemata used. Therefore, those measuring attachment should concen-
trate on detecting the existence and importance to self-worth of all schemata
that utilize a particular possession.
The formal definition of attachment proposed here is: the extent to which
an object which is owned, expected to be owned, or previously owned by an
individual, is used by that individual to maintain his or her self-concept. The
domain of this construct includes evidence that the internal consistency of
self-schemata is dependent on ownership (planned, current, or past) of the
object. When the private facet of the self is considered, there should be: (a)
internal rehearsal of the schemata that connect the possession to the self
(accompanied by positive feelings of self-worth) and (b) feelings of self-loss
when an object is lost. When the public facet of the self is in use, external
rehearsal
of possession-linked
self-schemata
and positive and negative feelings
arising because of the reaction of others to the possession are observed.
Although the collective self is undoubtedly an important facet, it is most
evident in behaviors and activities directed toward a reference
group. Attach-
ment in this case tends to involve collectively owned property,
such as a church
building, or the product or service produced by one's company. There are
definite feelings of ownership in these matters, as well as legal ownership in
some cases, and there may be very strong feelings associated with the objects
in any case, but such objects and relationships fall outside the scope of this
article. Questions of the definitions of possession and ownership become con-
ROLE AND MEASUREMENT OF ATTACHMENT 159
fused with the issue of physical control of the fate of the object when it is
collectively owned. We focus on attachment manifested in the public and
private facets of the self.
INVOLVEMENT
AND ATTACHMENT
At this point, it is worth mentioning a construct in the consumer behavior
literature that appears
very similar to attachment,
but conceptually is not. This
is involvement (e.g., Costley, 1988). Although involvement has been defined
in
various ways as importance, it is generally conceived as a property of the
relationship between a person and a product category, rather than a specific
possession. The category of products does not acquire the meaning and sig-
nificance of the particular possession. The behaviors and feelings appropriate
to an owned object are of a different class of phenomena than those for the
product category as a whole.
OTHER CONSTRUCTS AND ATTACHMENT
Emotional Significance
A number of constructs should have simple and direct relationships
to attach-
ment. As attachment and the time of ownership increase, so should the emo-
tional significance of the object. The emotional significance of a possession is
the total strength of associations with significant events or people in the
person's life, both good and bad. A possession with low attachment will
probably (but not always) have little emotional significance,
whereas a posses-
sion with high attachment may require some time to acquire emotional sig-
nificance.
Kind of Object
Attachment should vary across the population with respect to the kind of
object: One tends to use a house or car more for the purposes of self-concept
maintenance than a pair of shoes or a television, for example. Although
Americans are clearly attached (on the average) to their automobiles, for
example, it is difficult to say what the attachment level should be for many
kinds of objects. There should be wide individual variation in attachment to
any given object, because there is wide individual variation in how Americans
define themselves. Objects that are socially visible; expensive; reflective of the
individual's roles, relationships, accomplishments, and experiences; and usu-
ally "personalized" by the efforts of their owners are clearly more likely to
reflect self.
160 BALL AND TASAKI
The "Stage" of Ownership
The extent to which an object is used to maintain identity should vary with
the passage
of milestones in the relationship
between the person
and the object.
One such milestone is purchase
or acquisition itself. One would expect attach-
ment to rise after acquisition, up to the point at which the person's identity
changes or the possession (through deterioration, perhaps) is no longer able
to support the self. After the individual begins to consider getting rid of the
object, a decline in attachment is expected, and a further decline is expected
after disposal.
For the purposes of this study, the stages of ownership are defined as: (a)
preacquisition,
(b) early ownership (owned less than the median time for such
objects), (c) mature ownership (owned more than the median time for such
objects), (d) predisposal (thinking about getting rid of the object but have not
yet), and (e) postdisposal.
Social Desirability
But other constructs, such as social desirability (Crowne & Marlowe, 1964),
should not be strongly related to attachment. Social desirability
should reflect
the extent to which an individual seeks to present himself or herself in a
socially acceptable manner. Although the original interest in researching
this
tendency was substantive,
measures
of social desirability
have come to be used
more often as a indication of threat to the validity of a self-report
measure of
attitude or behavior. Thus, if social desirability
correlates highly with a mea-
sure of attachment, it is reasonable
to suspect the measure of contamination
by subjects seeking to present themselves as more or less connected to their
possessions than they really are.
Trait Materialism
Materialism is the "importance a person attaches to wordly possessions"
(Belk, 1985, p. 266) and seems to be related to attachment at first glance.
However, materialism is defined as a psychological trait, unconnected to any
possession in particular.
In addition, neither the definition
of materialism nor
its subscales (Possessiveness,
Envy, and Nongenerosity) imply that materialis-
tic people should use possessions for the purpose of maintaining
a concept of
self and nonmaterialistic
people should not. Over a broad range
of possessions,
little relationship should be expected between materialism and attachment.
To recapitulate,
the purpose of this work was to develop an instrument to
measure attachment and to find evidence of its validity. Reliability and unidi-
mensionality of the measure of attachment was assessed, and evidence of
content, predictive,
and discriminant
validity was found through relationships
with other constructs just presented.
ROLE AND MEASUREMENT OF ATTACHMENT 161
METHOD
Sample and Administration
Our sample consisted of 188 students from a large midwestern university and
143 adults from the surrounding community. Participation was requested
from students enrolled in a large lecture section of an introductory market-
ing class. Administration of the questionnaire
was conducted in groups of 30
to 40 students. Adults were members of various organizations listed with the
local Chamber of Commerce, and administration was performed during a
regular meeting of the group. In compensation for their participation, a do-
nation of $3.00 per person was offered to the organization or to the individ-
ual.
Administration of the questionnaire
involved a multistage process. In the
initial stage, each subject was given four lists of the same 10 objects. The
objects listed were a car, family home, living room decoration other than
furniture,
pair of everyday
shoes, watch, television, wallet, nice piece of jewelry
other than a watch, souvenir of a trip, and something collected as a hobby.
Instructions for the first list asked the subjects to circle all items they were
planning to acquire. They were to identify the object to be purchased,
but had
not yet purchased it (for mass produced items, the brand and type needed to
be identified). Subjects
were instructed to circle all items currently owned, all
items that they were currently
thinking seriously of getting rid of, and all items
they had discarded in the past year in the second, third, and fourth lists,
respectively.
Based on responses to the object list, each subject was assigned two objects
from among those circled. Assignment was carried out to obtain as equal a
distribution as possible across the 10 objects and the five stages of ownership.
At this point in the administration, it was not possible to separate the early
and mature ownership stages, because this separation was to be based on the
median time that each type of object was owned. Therefore, the balancing
pattern given to the questionnaire administrators
merely requested twice as
many subjects be assigned to objects circled on the second list.
For each object assigned, subjects were asked to respond to a questionnaire
concerning self-object identity and emotional significance. For example, be-
cause a particular
subject might have been planning to buy a pair of everyday
shoes, have currently owned one, been thinking of discarding one, and have
recently discarded
one, the instructions on his or her questionnaire
made clear
which of these pair of shoes was under consideration. If he or she currently
owned the object and was not planning to discard it, the number of years and
months owned was requested. As a final task, subjects were asked to fill out
a questionnaire containing social desirability and materialism scales and to
supply demographic information.
162 BALL AND TASAKI
Measurement of Attachment
Attachment was measured using a 9-item Likert scale (Table 1). Ten items
were originally generated
to reflect the two facets of the self in the attachment
domain-the public and private selves. One of the 10 items was dropped
because it correlated poorly with the other items. A balance was maintained
among the items so that they somewhat equally reflected the two public
self-aspects and two private self-aspects of the domain of attachment already
mentioned.
Items 2 and 6 reflect the private self in internal rehearsal of the schemata
that connect the self to the possession. Items 5 and 9 are concerned with the
loss of self upon loss of the object, also a concern of the private self. Items 3
and 8 reflect the external rehearsal of schemata that connect the possession
with the public self, and Items 1, 4, and 7 concern the public selfs attention
to the reaction of others to the possession. Thus, the areas of the domain of
attachment previously mentioned were covered by the measure.
The verb tenses of the items varied slightly depending on whether or not
TABLE 1
Attachment Scale for the Early Ownership, Mature Ownership, and
Predisposition Stages
Respondents were instructed to fill in the blanks mentally with the object being rated (e.g., a
car). A 6-point Likert scale ranging from agree (6) to disagree (1) was used.
1. Imagine for a moment someone making fun of your car. How much would you agree
with the statement, "If someone ridiculed my car, I would feel irritated."
2. How much do you agree with the statement, "My car reminds me of who I am."
3. Picture yourself encountering someone who would like to get to know you. How much
do you think you would agree with the statement, "If I were describing myself, my car
would likely be something I would mention."
4. Suppose someone managed to destroy your car. Think about how you would feel. How
much do you agree with the statement, "If someone destroyed my car, I would feel a
little bit personally attacked."
5. Imagine for a moment that you lost your car. Think of your feelings after such an
event. How much do you agree with the statement, "If I lost my car, I would feel like I
had lost a little bit of myself."
6. How much do you agree with the statement, "I don't really have too many feelings
about my car."a
7. Imagine for a moment someone admiring your car. How much would you agree with
the statement, "If someone praised my car, I would feel somewhat praised myself."
8. Think for a moment about whether or not people who know you might think of your
car when they think of you. How much do you agree with this statement, "Probably,
people who know me might sometimes think of my car when they think of me."
9. Imagine for a moment that you have lost your car. Think about going through your
daily activities knowing that it is gone. How much do you agree with the statement, "If
I didn't have my car, I would feel a little bit less like myself."
aReverse scored.
ROLE AND MEASUREMENT OF ATTACHMENT 163
the object was currently owned, was a probable future acquisition, or was a
former possession. Items were usually prefaced by a sentence or two. State-
ments like "imagine for a moment... ." were used to place participants
in an
appropriate mind set and to encourage greater consideration of items.
A Cronbach's coefficient alpha of .93 over all objects and stages was cal-
culated for attachment. Factor analysis confirmed the existence of a single
factor accounting for 87% of the common variance.
Previous measurements
of attachment have used either a statement that the
object was a favorite possession, implying high attachment (e.g., Belk, 1985;
Myers, 1985; Wallendorf & Arnould, 1988), or a statement that the object
would be very hard or easy to part with (Schultz et al., 1989). Instead, we
measured attachment using a more traditional method derived from a domain-
sampling model (Nunnally, 1978). This Likert-type approach is both more
precise and provides a means of assessing attachment across the vast number
of objects which are neither favorites nor irrelevant to the self-concept, but
somewhere inbetween.
Measurement of Emotional Significance
Table 2 shows the three-item scale of Emotional Significance. These were
selected from a list of five items designed to tap the associations of the object
with significant people and events in a person's life. The three items chosen
had the greatest internal consistency. Like the Attachment scale, the verb
tenses of the Emotional Significance
items varied slightly depending on stage
of ownership. Otherwise, the items were the same from stage to stage. Cron-
bach's coefficient alpha for the three-item scale was .76 when the sample was
aggregated across all objects and stages.
Measurement of Social Desirability
Nineteen of the items from Crowne and Marlowe's (1964) 33-item scale were
used. The items in this scale are rated true or false, with 1 point allocated if
an item is answered in the "socially acceptable"
direction, and 0 points if not.
TABLE 2
Emotional Significance Scale for the Early Ownership, Mature
Ownership, and Predisposition Stages
Respondents were instructed to fill in the blanks mentally with the object being rated (e.g., a
car). A 6-point Likert scale ranging from agree (6) to disagree (1) was used.
1. My car reminds me of important people in my life.
2. My car reminds me of important things I've done or places I've been.
3. If I lost my car, another one like it wouldn't be as meaningful to me.
164 BALL AND TASAKI
The 19 items chosen from the scale were those used by Carlson and Grossbart
(1988), who found a Cronbach's
alpha of .75. In our study, the 19-item scale
had a Cronbach's alpha of.69.
Measurement of Materialism
The measure of materialism used in this study is heavily derived from Belk
(1985). Belk used 24 items and three subscale scores-Envy, Nongenerosity,
and Possessiveness-to obtain an overall measure of materialism. Many of
Belk's items were used in this study, but several were replaced by items that
seemed to be more generalizable.
Sixteen Likert-type items were borrowed from Belk's scale, and we devel-
oped 4 more items. An item analysis was performed, and 16 items had suf-
ficient item-total correlations to be retained in the final materialism scale
(Table 3). The factor structure of the scale was the same as Belk's original
scale, with the same interpretation
for the three oblique factors: Envy, Non-
generosity, and Possessiveness. Cronbach's alpha for the overall 15-item
scale was .76, roughly similar to the values of .66, .73, and .68 obtained for
Belk's original scale.
TABLE 3
Materialism Items and Subscales
Envy
(.78) When friends have things I cannot afford, it bothers me.
(.77) I don't like watching others enjoy things I cannot have.a
(.66) I am bothered when I see people who buy anything they want.
(.57) There are certain people I would like to trade places with.
(.54) I don't seem to get what is coming to me.
(.50) When friends do better than me in competition, it usually makes me happy for
them.b
Nongenerosity
(.60) I enjoy sharing what I have.b
(.52) I don't like to lend things, even to good friends.
(.48) I don't like watching others enjoy things I cannot have.a
(.46) I enjoy having guests stay in my home.b
(.43) I worry about people taking my possessions.a
(.39) I enjoy donating things to charities.b
Possessiveness
(.62) I get very upset if something is stolen from me, even if it has little monetary value.
(.53) I don't get particularly upset when I lose things.
(.39) I don't like to have anyone in my home when I'm not there.
(.29) I tend to hang on to things I should probably throw out.
Note. Factor loadings are in parentheses.
aAdded to Belk's (1985) scale. bltem reversed.
ROLE AND MEASUREMENT OF ATTACHMENT 165
RESULTS
Overview
Though each individual rated 2 of the 10 objects, each object was treated
separately as though it had been rated by a unique individual. This doubled
the sample size available for the analyses,
but raised an issue regarding
the lack
of independence of the observations. However, this form of independence is
not likely to reduce the validity of the analyses. Even though the assignment
of stages and objects to subjects was not completely random, it was nonsys-
tematic with a strong random component, because the questionnaire
adminis-
trators chose from among several objects at each stage for most subjects.
Dependence among the observations could only take the form of response
biases arising from using the same person at each of two stages or for two
particular
objects, linking the two objects or two stages. Because the choice of
the two objects
and two stages for the subjects
followed no pattern,
and because
response biases are not likely to be large, the dependence of the observations
is not likely to have affected the results.
To separate
the early ownership
stage from the mature ownership stage, the
median time owned was calculated for each of the 10 objects owned, but not
at the predisposal
stage. If a subject had owned an object less than the median
time, his or her object was assigned to the early ownership stage; if an object
was owned for more than the median time, the object was assigned to more
the mature ownership stage.
Because the number of objects of each type at each stage was not quite
equal, each observation was weighted to give equal influence to each of the 10
objects at each stage. The weighted mean of attachment for the mature owner-
ship stage, for example, represents equal influence for homes, cars, wallets, and
so on. Unless otherwise specified, all analyses that aggregate across objects
used these weights.
To observe the variation of attachment over the range of the 10 objects and
over the five ownership stages, means of the 50 resulting cells were computed
and a two-way analysis of variance (ANOVA) performed. To quantify the
relationship
between attachment and emotional significance,
materialism,
and
social desirability, product-moment correlations, combining data across all
object types and ownership stages, were computed.
Attachment, Objects, and Ownership Stages
As shown in Table 4 the objects that one might expect to be most closely
associated with identity typically have higher mean attachment levels. Homes,
hobby items, cars, and personal jewelry might be expected to be items reflective
166 BALL AND TASAKI
TABLE 4
Attachment Means by Object and Ownership Stage
Ownership
Stage
All Early Mature
Object Stages Preacquisition Ownership Ownership Predisposal Postdisposal
Family home 4.18 4.46 4.63 4.63 4.43 2.74
Hobby item 3.42 3.53 3.17 3.67 3.38 3.33
Car 3.35 3.85 3.47 3.54 2.75 3.14
Nice jewelry 3.33 3.56 3.71 3.73 2.22 3.43
Living room 3.11 3.64 3.32 3.54 2.15 2.88
decoration
Shoes 2.97 3.11 2.84 3.26 2.97 2.66
Souvenir 2.95 3.34 3.16 3.39 2.40 2.44
Watch 2.93 3.62 3.07 3.25 2.72 1.98
Television 2.69 3.12 2.67 2.55 2.71 2.42
Wallet 2.69 3.11 2.67 2.94 2.49 2.22
Totals 3.16 3.53 3.28 3.45 2.82 2.72
Note. Observations weighted to give equal influence to each object and stage.
of the self, and shoes, watches, wallets, televisions are less reflective
of the self.
The main effects for type of object and for stage of ownership were significant
at p < .05, but the interaction between them was not.
With the possible exception of the preacquisition stage, variations across
ownership stages are predictable.
Attachment rises slightly from early to ma-
ture ownership, then declines as the consumer begins to think about disposing
the object and finally does. The relatively
large decline from mature ownership
to predisposal probably reflects the adjustment of evaluative schemata to
achieve cognitive consistency. If one's identity is changing or the object no
longer supports identity, the act of considering disposal should be concurrent
with distancing oneself from the object and adjusting or eliminating proposi-
tions that tie the object to one's self-esteem.
The fact that attachment, for most objects, is higher in the preacquisition
stage than in early ownership is a bit surprising. A plausible explanation is
provided, however, by examining this relationship with respect to age (Figure
1). The drop in attachment is a function of the younger age group (ages 15 to
24), which comprised
slightly over half the sample. Younger consumers proba-
bly have much more changeable
identities and are less sophisticated, especially
with respect to infrequently
purchased
items. They are more likely to imagine
themselves as different people, build propositions relating themselves to the
object and its personal
and social consequences
which support
a new and better
self, and then discover that those consequences
do not materialize
after
acquisi-
tion. Hence, younger people are likely to experience a drop in attachment for
many objects from preacquisition to early ownership.
ROLE AND MEASUREMENT OF ATTACHMENT 167
Attachment 0 ages 15 to 24
El ages 25 to 44
4 [7 ages over 44
3
Pre-acquisition Early
ownership
NOTE: Values
weighted
to give
equal
influence to each
object.
FIGURE 1 Interaction between age of subject and the stages of preacquisition
and early
ownership.
Attachment, Emotional Significance,
Social Desirability, and Materialism
Table 5 shows correlations between attachment and emotional significance,
social desirability, and materialism across all 10 objects and five ownership
stages. As expected, attachment is not contaminated by social desirability
and
does not correlate highly with trait materialism.
As expected, attachment covaries strongly with emotional significance, al-
though it is incorrect to conclude that attachment and emotional significance
are the same construct. Figure 2 demonstrates that, although attachment and
emotional significance generally covary across ownership stages, they diverge
significantly at one important point. Emotional significance begins at a low
level and rises through the mature ownership stage. This is to be expected:
Emotional significance reflects important memories, which cannot be present
before acquisition, and require time to accumulate. Attachment, on the other
hand, reflects the schemata used to connect the self to the object, and these
may be present prior to acquisition. The drop in both measures in the predis-
TABLE 5
Correlations Between Attachment and Emotional
Significance, Social Desirability, and Materialism
Attachment
Emotional significance .503*
Social desirability -.069**
Materialism .159*
*Statistically significant at p < .0005. **Not statistically
significant at p < .05.
168 BALL AND TASAKI
Attachment
4 -
o Average
attachment
2 - 0 Average
emotional
significance
- I I I I
Pre- Early Mature Pre- Post-
acquisition ownership ownership disposal disposal
NOTE:
Values
weighted
to give equal
influence
to each
object.
FIGURE 2 Pattern of attachment and emotional significance.
posal stage probably reflects similar cognitive consistency processes of elimi-
nating propositions that give the object meaning and value to the individual.
DISCUSSION
The measure of attachment proposed here varies as expected with other con-
structs on the basis of the conceptual definition. Evidence of reliability
is quite
strong, with Cronbach's
alpha at .93, and unidimensionality
seems present as
well.
Attachment should vary coherently with the type of object and with the
stage of ownership, as this measure does, reflecting predictive validity. Fur-
thermore,
attachment should be strongly related to emotional significance,
but
should be differentiated
from it when an object reflects the self; attachment is
something other than important memories. For the proposed measure, this is
the case, because the preacquisition stage of ownership
is often associated with
high attachment
but not with high emotional significance, especially for young
consumers.
This latter finding provides
some evidence toward the discriminant
validity of attachment.
Attachment should not be related to social desirability because such a
relationship would imply contamination, and the proposed measure is un-
related to social desirability. Finally, attachment should be distinct from trait
ROLE AND MEASUREMENT OF ATTACHMENT 169
materialism because materialistic individuals
should not use all of their posses-
sions to support the self, even though possessions may be generally important
to them. Nonmaterialistic consumers should nonetheless be attached to at least
a few of the things they own. Thus, only a small correlation between material-
ism and attachment was expected and was found. This lack of strong covari-
ance between attachment and social desirability and materialism is further
evidence of discriminant validity.
These findings, along with the face validity of the measure, support the
notion that what is being measured here is attachment, not some other con-
struct such as emotional significance or materialism. This variable is a criti-
cal one, especially in the study of postpurchase consumer behavior, and its
measurement in a precise and valid way is important to the field. Although
there are limitations to the methodology of the study, the measure of attach-
ment proposed here shows sufficient validity to make it useful in further
research.
Any future work on the measurement of attachment should address
conver-
gent validity, an aspect of construct validity. The multitrait-multimethod
matrix (Campbell & Fiske, 1959) is the basic tool for this analysis, often
augmented by an analytic measurement model such as that available in LIS-
REL (Bagozzi, 1980). Convergent validity is the tendency for multiple mea-
sures of a construct, using different
measurement
techniques,
to correlate more
highly with each other than with measures of other constructs. Because multi-
ple measures of attachment were not attempted, this could not be done.
This is a flaw in our study, which we do not take lightly. However, the
evidence of reliability, unidimensionality,
content validity, predictive validity,
and discriminant validity presented here, although not complete from a con-
struct validity point of view, is nonetheless substantial evidence of the validities
of the measure and construct.
Attachment should predict many postpurchase
behaviors, among them the
amount of time to disposal (and, hence, repurchase of a replacement object
within its product class, if needed), and behaviors surrounding protection and
maintenance of the object and the amount of effort and money devoted to
them. Such knowledge would be useful to many marketing practioners in
understanding purchase cycles and aftermarkets,
for example, as well as to
marketing theorists attempting to understand consumer psychology.
Furthermore,
understanding
attachment should help in understanding
the
emotional aspects of product ownership. Obvious psychological conflicts arise
when a product no longer performs satisfactorily and yet is still a part of the
consumer's identity, for example. What is the process by which an individual
becomes disattached, and how could this be affected by marketing actions?
Here are implications not only for practitioners of the traditional marketing
arts but also for those studying garbage disposal and recycling. For another
matter, what is the process of attachment? An intriguing result found here
170 BALL AND TASAKI
indicates that, for young consumers at least, attachment may sometimes pre-
cede ownership.
A possession has a life cycle over time in terms of its relationship with the
owner. We speculate that the use of owned objects to support identity will
systematically vary with critical events-not only the obvious ones, such as
purchase and disposal-but also with more difficult-to-capture
events such as
the critical comments of others, performance successes and failures of the
object, and the use of the object in self-esteem enhancing, goal-directed
activi-
ties.
More important, the constantly changing character of identity itself must
be recognized as a factor in attachment. The propositions and schemata con-
nected to the self are constantly changing as new propositions about the self
are introduced and processed. Cognitive consistency will result in change in
many schemata as events occur and as the maturing individual integrates his
or her personality and acts out of this integrity. People will psychologically
outgrow many possessions and acquire attachments to others that reflect the
new self. If there is, in fact, such a thing as a "core self' (Belk, 1988), it may
consist of a small number of simple propositions related to higher level quali-
ties such as intellectual honesty, spirituality,
and connectedness.
If the core self
tends to become more prominent and well-defined as the person matures
psychologically, we might expect lower attachments to possessions in general
(although not necessarily
fewer possessions) and high attachment to only a few
objects that reflect an integrated core.
An obvious benefit to developing a theory in this area is the study of the
life cycle of the person, particularly
the changing
nature
of identity. Also useful
would be the study of personal relationships:
Inasmuch as attachment reflects
a relationship
between a person and an object, theoretical constructs from the
person-person relationship literature will be valuable. Finally, cross-cultural
psychology literature should be tapped to explore not just cross-cultural dif-
ferences in what people become attached to, but to suggest whether there are
cross-cultural differences in the processes of attachment and disattachment.
Research methods in this area should tend to be more longitudinal than
consumer behavior studies typically are, requiring
quantitative measurement
of attachment and other constructs at multiple points in time. In-depth inter-
viewing and interpretivist approaches also have value, particularly when
focussed on extracting the complexes of schemata surrounding an object,
assessing the place of the object in the definition and evaluation of self, and
exploring the role of life events in altering attachment. The scale presented in
this article, although designed specifically for quantitative hypothesis testing,
could also be a useful prompting device in qualitative research methods.
Identity is the central concept in understanding
the person, and attachment
is the central construct in understanding
the role and meaning of possessions
to their owners. This is a new area of consumer research, with vastly more
ROLE AND MEASUREMENT OF ATTACHMENT 171
hypotheses than data points, and is potentially valuable to the broad under-
standing of consumption behavior.
ACKNOWLEDGMENTS
We thank those in the Marketing Department at the University of Nebraska
for their support and Professors James Gentry, Sanford
Grossbart,
and Robert
Mittelstaedt for their advice.
REFERENCES
Bagozzi, R. P. (1980). Causal models in marketing. New York: Wiley.
Belk, R. W. (1985). Materialism: Trait aspects of living in the material world. Journal
of Consumer
Research, 12, 265-280.
Belk, R. W. (1987). Identity and the relevance of market, personal, and community objects. In
J. Umiker-Sebeok (Ed.), Marketing and semiotics (pp. 151-164). Berlin: Morton de Gruyter.
Belk, R. W. (1988). Possessions and the extended self. Journal of Consumer
Research, 15, 139-
167.
Belk, R. W. (1990). The role of possessions in constructing and maintaining a sense of past. In
M. Goldberg, G. Gorn, & R. Pollay (Eds.), Advances in consumer research (Vol. 17, pp.
669-676). Provo, UT: Association for Consumer Research.
Blasi, A. (1988). Identity and the development of the self. In D. K. Lapsley & F. C. Power (Eds.),
Self, ego and identity:
Integrative approaches (pp. 226-242). New York: Springer-Verlag.
Breckler, S. G., & Greenwald, A. G. (1986). Motivational facets of the self. In E. T. Higgins &
R. Sorrentino (Eds.), Handbook of motivation and cognition (pp. 145-164). New York: Guil-
ford.
Campbell, D. T., & Fiske, D. W. (1959). Convergent
and discriminant
validation
by the multi-trait
multi-method matrix. Psychological
Bulletin, 56, 81-105.
Carlson, L., & Grossbart,
S. (1988). Parental
style and consumer socialization of children. Journal
of Consumer
Research, 15, 77-94.
Costley, C. L. (1988). Meta-analysis of involvement research. In M. J. Houston (Ed.), Advances
in consumer research (Vol. 15, pp. 554-562). Provo, UT: Association for Consumer Research.
Crowne, D. P., & Marlowe, D. (1964). The approval
motive. New York: Wiley.
Csikszentmihalyi, M., & Rochberg-Halton, E. (1981). The meaning of things:
Domestic symbols
and the self. New York: Cambridge University Press.
Deci, E. L., & Ryan, R. M. (1991). A motivational approach
to the self: Integration
in personality.
In R. Dienstbier (Ed.), Nebraska symposium
on motivation: Vol. 38: Perspectives
on motivation
(pp. 237-288). Lincoln: University of Nebraska Press.
Erikson, E. H. (1959). Identity and the life-cycle. New York: Norton.
Gergen, K. J. (1971). The concept of self New York: Holt.
Greenwald, A. G. (1988). A social-cognitive account of the selfs development. In D. K. Lapsley
& F. C. Power (Eds.), Self, ego and identity:
Integrative
approaches (pp. 30-42). New York:
Springer-Verlag.
Greenwald, A. G., & Pratkanis, A. R. (1984). The self. In R. S. Wyer & T. K. Srull (Eds.),
Handbook of social cognition (Vol. 3, pp. 129-178). Hillsdale, NJ: Lawrence Erlbaum Associ-
ates, Inc.
172 BALL AND TASAKI
Hirschman, E. C., & Holbrook, M. B. (Eds.). (1981). Symbolic consumer behavior.
Ann Arbor,
MI: Association for Consumer Research.
Holbrook, M. B., & Hirschman, E. C. (1982). The experiential aspects of consumption:
Consumer
fantasies, feelings, and fun. Journal of Consumer
Research, 9, 132-140.
Khilstrom, J.F., & Cantor, N. (1984). Mental representations
of the self. In L. Berkowitz (Ed.),
Advances in experimental
social psychology (Vol. 17, pp. 2-47). New York: Academic.
Lee, D. H. (1990). Symbolic interactionism: Some implications for consumer self-concept and
product symbolism research, In M. Goldberg, G. Gorn, & R. Pollay (Eds.), Advances in
consumer research (Vol. 17, pp. 386-393). Provo, UT: Association for Consumer Research.
Loevinger, J. (1976). Ego development. San Francisco: Jossey-Bass.
Marks, L. J., Higgins, S., & Kamins, M. A. (1988). Investigating the experiential dimensions of
product evaluation. In M. J. Houston (Ed.), Advances in consumer research (Vol. 15, pp.
114-121). Provo, UT: Association for Consumer Research.
Markus, H. (1977). Self-schemata
and processing
information
about the self. Journal
of Personality
and Social Psychology,
35, 63-78.
McCracken, G. (1986). Culture and consumption: A theoretical account of the structure and
movement of the cultural meaning of consumer goods. Journal of Consumer Research, 13,
71-84.
McCracken, G. (1988). Culture and consumption. Bloomington: Indiana University Press.
Myers, E. (1985). A phenomenolgical analysis of the importance of special possessions: An
exploratory study. In E. Hirschman & M. Holbrook (Eds.), Advances in consumer research
(Vol. 12, pp. 560-565). Provo, UT: Association for Consumer Research.
Nunnally, J. C. (1978). Psychometric theory. New York: McGraw-Hill.
Richins, M. L., & Bloch, P. H. (1986). After the new wears off: The temporal context of product
involvement. Journal of Consumer
Research, 13, 280-285.
Rook, D. W. (1985). The ritual dimension of consumer behavior. Journal of Consumer
Research,
12, 251-264.
Schultz, S. E., Kleine, R. E., & Kernan,
J. B. (1989). These are a few of my favorite
things:
Toward
an explication of attachment as a consumer behavior construct. In T. K. Srull (Ed.), Advances
in consumer research
(Vol. 16, pp. 359-366). Provo, UT: Association for Consumer Research.
Tetreault, M. A., & Kleine, R. E. (1990). Ritual, ritualized
behavior, and habit: Refinements
and
extensions of the consumption ritual construct. In M. Goldberg, G. Gorn, & R. Pollay (Eds.),
Advances in consumer research (Vol. 17, pp. 31-38). Provo, UT: Association for Consumer
Research.
Wallendorf, M., & Arnould, E. J. (1988). My favorite things: A cross-cultural
inquiry into object
attachment, possessiveness, and social linkage. Journal of Consumer
Research, 14, 531-547.
Westbrook, R. A. (1987). Product and consumption-based
affective responses and post-purchase
processes. Journal of Marketing Research, 24, 258-270.
