Comparing Scales to Measure Compulsive Buying: an Exploration of Their Dimensionality

Leslie Cole, Louisiana State University
Dan Sherrell, Louisiana State University
[ to cite ]:
Leslie Cole and Dan Sherrell (1995) ,"Comparing Scales to Measure Compulsive Buying: an Exploration of Their Dimensionality", in NA - Advances in Consumer Research Volume 22, eds. Frank R. Kardes and Mita Sujan, Provo, UT : Association for Consumer Research, Pages: 419-427.

Advances in Consumer Research Volume 22, 1995      Pages 419-427


Leslie Cole, Louisiana State University

Dan Sherrell, Louisiana State University


Researchers have recently shown increased interest in understanding negative consumption behaviors, such as drug addiction (Hirschman 1992) and compulsive buying (Faber, O'Guinn and Krych 1987; Scherhorn, Raab and Reisch 1990). In particular, compulsive buying behavior has been examined from a phenomenological (O'Guinn and Faber 1989) as well as a conceptual (Valence, d'Astous and Fortier 1988) perspective.

At present, two research teams have focused on the development of instruments which appropriately tap the consumer's propensity to engage in compulsive buying behaviors (Valence, d'Astous and Fortier 1988; Faber and O'Guinn 1992). However, each group of researchers has taken a different approach to the task of scale development.

Valence et al. (1988) identified four conceptual dimensions associated with compulsive buying: a) tendency to spend; b) reactive aspect (i.e., presence of irresistible urge to buy); c) post-purchase guilt; and d) family environment. During refinement efforts, the fourth dimension of family environment was dropped from the scale development and subsequent analyses showed the other three dimensions to load together on one factor.

Faber and O'Guinn (1992) employed a phenomenological approach to the development of a scale designed to identify compulsive buyers in the general population. By examining in-depth interviews with self-reported compulsive buyers, Faber and O'Guinn (1992) constructed a screening scale to identify compulsive buyers. They suggest that various constructs such as self-esteem, materialism and credit usage are associated with compulsive buying, although their compulsive buying clinical screening scale is apparently based on the unidimensional compulsive buying construct.

The purpose of this paper is twofold: a) to empirically identify the various conceptual dimensions captured by each scale; and b) to compare the nomological and predictive validity exhibited by each scale. The evidence presented in pursuit of these two objectives should assist researchers interested in the topic of compulsive buying in deciding how best to measure the construct.


Defining Compulsive Buying

Negative consumption behaviors must be studied in order to more fully understand the effects of the consumption process on society and the well-being of others (Hirschman 1992; Wells 1993). One such negative consumption behavior is compulsive buying. A closely related consumption behavior without such negative consequences is impulsive buying (Rook and Hoch 1984). The presence of lack of volitional control and prepurchase planning present in impulsive purchasing is similar to that of compulsive buying. However, given the lack of negative consequences from impulsive buying in general, the focus of the present study is strictly on compulsive buying.

O'Guinn and Faber (1989) view compulsive buying as an addictive behavior with the following definition:

a response to an uncontrollable drive or desire to obtain, use or experience a feeling, substance, or activity that leads an individual to repetitively engage in a behavior that will ultimately cause harm to the individual and/or others (p. 148).

This definition is similar to that suggested by Valence et al. (1988), who identified three constructs associated with compulsive buying behavior: 1) a strong emotional activation (increase in psychological tension); 2) a high cognitive control (an acknowledgement that buying will reduce the tension); and 3) a high reactivity (looking for tension reduction, rather than ownership). It also should be noted that the O'Guinn and Faber (1989) definition is sufficiently general to allow for non-purchase consumption (e.g., anorexia/bulimia, or gambling). However, the objective of the study is to examine compulsive behavior as it relates to uncontrollable purchasing activity.

Conceptual Development of Valence et al. (1988) Scale

The first scale developed to tap the constructs underlying compulsive buying behavior was developed by Valence, d'Astous and Fortier (1988) and stems from early conceptual work done by Faber, O'Guinn, and Krych (1988). During the early stages of this scale's development there were four dimensions involved with this measure. The first dimension was identified as "tendency to spend", wherein a compulsive buyer should exhibit a higher propensity to spend than a noncompulsive buyer.

The second dimension, "reactive aspect", dealt with the individual's response to strong urges to purchase. Thus, an individual exhibiting compulsive buying behavior might feel that the motivations or urges to purchase are irresistible or beyond their control, while noncompulsive buyers would not view such motivations to purchase as uncontrollable.

The third dimension associated with compulsive buying by Valence et al. (1988) was post-purchase guilt. Researchers have reported evidence that individuals who engage in compulsive buying often felt remorse over their behaviors (Faber, O'Guinn and Krych 1987; O'Guinn and Faber 1989).

Finally, a fourth dimension was determined to be family environment, that is, the environment (which would include relationships among the family members) in which one grew up should suggest a predisposition to engage in negative consumption behaviors, such as compulsive buying (d'Astous, Maltais and Roberge 1990). However, in the final version of the Valence et al. (1988) scale, the dimension of family environment was dropped because of poor internal consistency.

The Valence et al. (1988) scale has been used in several studies. Scherhorn, Raab and Reisch (1990) employed the scale to study self-reported compulsive and "normal" German consumers. The scale exhibited a Cronbach's alpha of .92 across both samples, but produced different factor structures in the compulsive and normal consumer samples. The non-compulsive subjects' scale responses produced two factors, an irresistible urge to purchase and a certain amount of postpurchase guilt. The compulsive buyer group's scale answers resulted in three factors: a external urge to purchase and an internal urge to purchase, along with a third factor, noted to be postpurchase guilt, similar to the noncompulsive group. These results suggest that the Valence et al. (1988) scale strongly taps into a motivational construct related to compulsive buying.

D'Astous, Maltais and Roberge (1990) used the Valence et al. (1988) scale to study compulsive buying in adolescents. The scale displayed a Cronbach's alpha of .78 for the sample respondents. D'Astous, Maltais and Roberge (1990) found adolescents to exhibit a generalized urge to buy, influenced by their peers.

Faber and O'Guinn's (1992) Compulsive Buying Screening Scale

Building upon their earlier work, Faber and O'Guinn (1992) developed a scale intended to screen compulsive buyers out of the general population. Previously, the use of self-reported compulsive buyer samples had limited researchers in examining the phenomenon of compulsive buying behavior. In addition, some individuals may have difficulty in distinguishing compulsive buying from impulsive buying behavior. Nataraajan and Goff (1992) point out that distinct differences may exist between compulsive shoppers and compulsive buyers, much in the same way that shopping may be conducted for reasons that may differ from the motivation for purchasing. Consequently, a screening instrument for compulsive buying is badly needed.

Faber and O'Guinn (1992) conducted in-depth interviews of identified compulsive buyers to gather descriptions of the types of behaviors and feelings these individuals reported during compulsive buying activities. Additionally, items were included based upon prior research efforts and theoretical concerns and tended to focus on the behaviors associated with compulsive buying.

Currently, there are no published studies which use the Faber and O'Guinn clinical screening scale. However, work done by these authors has been crucial in calling attention to the problems associated with finding such a small segment within a given population.

Study Objectives

The purpose of this study is to empirically compare the Valence et al. (1988) and Faber and O'Guinn (1992) compulsive buying measures to assess their: a) dimensionality (both within and across scales) and b) relative performance in identifying compulsive buyers. Based on the discussion presented above, it is apparent that the two instruments were developed for slightly different purposes using different methodologies. While each scale has been developed from a different point, it is expected that each scale should tap into similar conceptual dimensions. The Valence et al. (1988) instrument seems more oriented toward measuring the degree of irresistible urge to purchase (or shop), while the Faber and O'Guinn (1992) measure is designed to capture behavioral and financial indicators of compulsive buying behavior.

The Faber and O'Guinn (1992) measure has norms associated with its use to identify compulsive buyers while the Valence et al. (1988) scale does not. Consequently, direct comparisons of performance are problematic. Comparisons of the groups of consumers identified by each scale across a common set of constructs suggested by the literature as associated with compulsive buying should enable some preliminary conclusions to be drawn.


A study was conducted to identify the factor structures underlying the two compulsive buying scales, assess the level of nomological validity exhibited by each scale, and examine the predictive validity of the scales. Five separate constructs were identified from the compulsive buying literature as being closely associated with compulsive buying tendencies. These items were used to compare the nomological validity of the Valence et al. (1988) and Faber and O'Guinn (1992) scales.

The self-esteem construct has been central to much of the development of compulsive buying measures (Faber and O'Guinn 1989). Self-esteem was measured by a scale determined to tap an individual's state self-esteem (Heatherton and Polivy 1991). Low self-esteem has been shown to be related to compulsive buying in a number of previous studies (Faber and O'Guinn 1992, Scherhorn, Reisch and Raab 1990; d'Astous, Maltais and Roberge 1990).

Early work in this area suggested that compulsive individuals were strongly motivated to reduce tension and anxiety through the shopping process (Faber, O'Guinn and Krych 1988; O'Guinn and Faber 1989). Furthermore, it has been proposed that compulsive buyers have a greater need for the activity than do members of the general population (Faber, O'Guinn and Krych 1988). This need for shopping was operationalized as involvement with the shopping process and motivations to shop. The shopping involvement scale was a reduced form of the Zaichowsky (1986) involvement scale. Shopper motivations were measured with items which suggested shopping motivated by a need to escape (Attaway 1989).

In this study, the Richins and Dawson (1992) materialism scale was used. Developed as a measure of consumer values, rather than as a personality trait, this scale purports to measure three dimensions of materialism: success, happiness, and acquisition centrality. The first of these, materialism as a sign of success, suggests that individuals who wish to show their success materially are likely to engage in behavior which appears materialistic. Materialism as happiness is operationalized as acquisition which is important to overall happiness and life satisfaction. Finally, Richins and Dawson (1992) operationalize acquisition centrality as that type of materialism which represents consumption excess. Previous studies have used the Belk (1985) materialism scale, where materialism was inferred from measures of personality traits (O'Guinn and Faber 1989; Scherhorn 1990; Faber and O'Guinn 1992). The scale developed by Richins and Dawson conceptualizes materialism as a consumer value, allowing for more direct measure of the construct.

Previous studies indicated that consumers who engage in compulsive buying often experience credit card abuse (Faber and O'Guinn 1988b; d'Astous 1990). In another Faber and O'Guinn study, it was shown that compulsive buyers had a greater proportion of their monthly income going to service debt than did general consumers (Faber and O'Guinn 1992). In this study, credit card usage was operationalized with a multi-item measure focusing on credit card behaviors and developed by d'Astous (1990).

A questionnaire containing measures for each of the constructs described above as well as the two compulsive buying scales was administered during class time to a convenience sample of 337 college students from a large southern university, with 319 usable questionnaires being returned.

While there is debate over the use of student samples (Wells 1993), the college student population has some relevance to shopping and purchasing issues. For many college students, their college experience is one of the first opportunities to make shopping and purchase decisions in a fairly autonomous manner. This expectation was investigated through focus groups held with students similar to those used in the main sample. Themes which emerged from the focus groups showed that college students did, indeed, show some autonomy over their finances. Many students described their parents giving them lump sums of money for semester expenses. Further, many related that their parents had given them credit cards to use while they were at school. While individuals may not have formed stable purchase patterns at this stage in their life, they should still have a significant store of general purchase knowledge and experience. In addition, for purposes of scale comparison, the use of a homogeneous population would seem to be suggested (Calder, Phillips and Tybout 1981). Therefore, the use of a convenience sample of college students was deemed appropriate.


The Valence et al. (1988) scale is a multi-item measure using a five point Likert scale, ranging from (1) "strongly disagree" to (5) "strongly agree". The Faber and O'Guinn (1992) scale is also a multi-item instrument measured on a five point scale where the respondents are instructed to answer how often they have behaved in a certain manner. The scale endpoints range from (1) "never" to (5) "very often".

Self-esteem was measured using a scale designed to tap an individual's state self-esteem (Heatherton and Polivy 1991). Respondents were instructed to respond according to their degree of agreement with scale items, using (1) "not at all" to (5) "extremely". The shopping involvement scale was a reduced form of the Zaichowsky (1986) involvement scale. Using a seven point semantic differential scale, respondents were instructed to indicate their feelings about the shopping process. Shopper motivations were measured using a reduced form of a shopping motivation scale developed by Attaway (1989) and based on the functional, symbolic and experiential needs expressed by Park, Jaworski and MacInnis (1986). Using a five point Likert scale, respondents were asked to express their level of agreement with the statements provided.

Materialism was measured using a scale developed by Richins and Dawson (1992). Items reflecting each of the three dimensions of success, happiness and acquisition centrality was measured using a five point Likert scale, with responses ranging from (1) "strongly disagree" to (5) "strongly agree".

Credit card usage was measured with a scale developed by d'Astous (1990). The scale was a multi-item five point measure which asked how often the respondent engaged in the behavior, with responses ranging from (1) "never" to (5) "very often".

Scale Evaluation

Several steps were taken to compare the two compulsive buying scales. First, the Valence et al. (1988) (VDF) scale and the Faber and O'Guinn (1992) (FOG) scale were subjected to confirmatory factor analysis using the dimensions identified in previous studies. The VDF scale was hypothesized to based on three separate dimensions of tendency to spend, reactive aspect, and post-purchase guilt. The FOG scale has been presented as a unidimensional scale. Then, the two scales were combined in a confirmatory factor analysis to see if they revealed common dimensions in respondents' answers.

The nomological validity of the scales was investigated by correlating the summed scale responses with related constructs identified in the literature (i.e., self-esteem, shopping involvement, shopping motivation, materialism, and credit usage). Finally, the predictive validity of the two scales was examined in two stages. Each set of scale items was used in a cluster analysis to generate distinct groups of respondents. These respondent groups were then profiled on the set of five related constructs described above and the cluster solutions compared across the two compulsive scale groups. The second stage of predictive validity investigation involved building groups of compulsive and non-compulsive buyers based on the VDF scale and using the FOG scale items as predictive elements in a discriminant analysis. This analysis allowed the examination of the ability of the FOG scale to identify compulsive buyers in a specific population.


Scale Dimensionality

Valence et al. scale. The three dimensions identified by Valence et al. for their compulsive buying scale were used to develop a confirmatory factor model for analyzing the data from the respondents. Table 1 presents the factor loadings and fit statistics generated via LISREL VII (J÷reskog and S÷rbom 1989). The goodness-of-fit (GFI) and the adjusted goodness-of-fit (AGFI) indices exhibit respectable levels of fit (values in the mid .80s and higher have been suggested as evidence of acceptable fit (Bagozzi and Yi 1988)).

The internal consistency estimates of composite reliability for the three dimensions were .69, .86 and .61, respectively. The variance extracted estimates were moderate at best. Of the three dimensions, only one (reactive aspect) had a variance extracted estimate above .50.

Finally, the majority of the standardized residuals were below + 2.00, which is indicative of a reasonable model fit (Anderson and Gerbing 1988). The modification indices suggested that two of the scale items (one item from the tendency to spend dimension and one item from the postpurchase guilt dimension) should be correlated with each other across constructs. Therefore, the dimensions proposed by Valence et al. seem to be supported reasonably well by our data, although the amount of variance explained is not high. However, since the scale attempts to measure a type of consumption behavior that has been estimated to be present for only 10-15 percent of the population (Faber and O'Guinn 1992), the scale's performance is a reasonable one.

Faber and O'Guinn scale. Faber and O'Guinn posit a unidimensional scale with the behavioral construct of compulsive buying as the latent variable behind the scale items. A single factor confirmatory model was estimated using the sample data. The results of that analysis are reported in Table 2. The proposed single factor model does a reasonably good job of reproducing the observed correlation matrix, as evidenced by the fit statistics of GFI=.96 and AGFI=.91. The composite reliability is .76, while the average variance extracted is low (.33), but similar to the results for the Valence et al. scale. As was the case with the Valence et al. scale, there are two items with low factor loadings.

An additional model was run with the all the items from both scales. The confirmatory model was specified with the three factors from the Valence et al. scale, along with the single factor from Faber and O'Guinn scale. The results of this model test are presented in Table 3. As can be seen, the combined model did not represent the data as well as the separate models did. The goodness-of-fit indices were lower and the Chi-square statistic was noticeably higher. The combined model results suggested that the same two items from the Valence et al. scale wanted to load across the Valence constructs, while there were two Faber and O'Guinn scale items with low loadings. Interestingly, however, none of the separate scale items from the Valence et al. or the Faber and O'Guinn scale showed tendencies to load with the opposite scale. These results suggest that the two scales are measuring separate sets of constructs (or simply different dimensions of compulsive buying behavior).

Nomological Validity Comparisons

A correlation analysis was run between the summed scale scores for the Valence et al. scale, the Faber and O'Guinn scale, and the constructs of self-esteem, shopping involvement, shopping motivation, materialism, and credit card usage. These constructs had been identified earlier from the literature as conceptually linked to compulsive buying behavior. Table 4 shows the results of these comparisons.

All the validity constructs were significantly associated with both compulsive buying scales. Self-esteem was correlated negatively with both the Valence et al. scale (r=-.32) and the Faber and O'Guinn scale (r=-.28). Materialism was correlated more highly with the Valence et al. scale than the Faber and O'Guinn scale, although the difference was not large.







There were three validity constructs that showed large differences between scales were shopping involvement (VDF r=.38; FOG r=.23); shopping motivation (VDF r=.56; FOG r=.31); and credit card usage (VDF r=.36; FOG r=.71). Whereas the two items related to shopping (involvement and motivation) were correlated higher with the Valence et al. scale than the Faber and O'Guinn scale, the credit card usage item showed a much stronger relationship with the Faber and O'Guinn scale. This finding suggests that maybe the Valence et al. scale taps into a behavior more closely associated with shopping, while the Faber and O'Guinn scale measures compulsive buying behavior more closely.

Predictive Validity Comparisons

Cluster analysis. The first stage of the predictive analysis was an attempt to group respondents using both compulsive buying scales and to compare the identified groups. The results of these analyses are presented in Tables 5 and 6.

In determining the number of clusters to use as a solution, the percentage change criteria was used with the agglomeration schedule. In doing so, it appeared that for both scale analyses, there appeared to be a large jump in going from three to two clusters. There was some concern that using a two group solution would result in people who showed extreme levels of compulsive buying being grouped with individuals who reported some compulsive tendencies, but were not extreme across the whole range of scale items. For this reason, a three group solution was used for each scale analysis.







Following the decision to use a three group cluster solution for both scales, the means of the three groups for each scale were entered into separate, nonhierarchical cluster routines to generate respondent groups and profiles for the scale items.

Valence et al. Scale Cluster Profiles. In examining the profiles of the groups of respondents identified by the cluster analysis, it is apparent that the individuals in cluster three are more extreme on all but three of the scale items. Respondents in clusters two and three are similar on items relating to: a) "I feel guilty after buying a product"; (b) "..fear being perceived as irrational in my buying behavior"; and (c) "I often respond to direct mail offers". The groups identified by the cluster solution for the Valence et al. scale can be characterized as: cluster 1 - Normal Buyers; cluster 2 - Impulsive/Guilty Buyers; and cluster 3 - Compulsive Buyers. While the choice of labels is subjective, the scale clearly identifies respondents who differ in terms of being able to resist uncontrollable urges to shop or buy; feel guilty about unnecessary purchases; or view shopping/buying as a means of reducing stress. This is consistent with the conceptual taxonomy of buyers developed by Valence et al. 1

Faber and O'Guinn Scale Cluster Profiles. The profiles of the respondents grouped according to the Faber and O'Guinn scale also show clear demarcations between relatively normal buying behavior and compulsive buying patterns. Cluster 3 respondents again showed higher scores on all scale items except one: making the minimum payment on their credit cards. Respondents in cluster 2 were significantly different from the more "normal" subjects in cluster 1, in terms of being more motivated to spend money, buying things they couldn't afford, or worrying about what other people would think if they knew how much that person spent. However, respondents in cluster 2 showed some similarities to cluster 1 respondents in viewing buying as a way to feel better or being anxious/nervous if they didn't go shopping frequently. Cluster 1 could be labeled Normal Buyers; cluster 2 respondents could be called Impulsive Buyers; and cluster 3 subjects were clearly Compulsive Buyers. As with the Valence et al. scale results, the Faber and O'Guinn group profiles are distinct with respect to buying behaviors and motivations.

Validity Construct Comparisons. Tables 5 and 6 also display the results of comparing the identified cluster groups across the five validity constructs. For both the Valence et al. scale and the Faber and O'Guinn scale, the groups labeled as compulsive score significantly lower on self-esteem matching reported evidence by Faber and O'Guinn 1992; and significantly higher on the remaining measures of shopping involvement, shopping motivation, materialism, and credit card usage.

Within scale comparisons showed that the impulsive and normal groups were not significantly different (p<.05) on self-esteem for each scale. For the Valence et al. scale, impulsive respondents were significantly different (p<.05) from compulsive subjects on shopping involvement and shopping motivation, but not significantly different from compulsives on materialism and credit card usage. For the Faber and O'Guinn groups, impulsive buyers were significantly different (p<.05) from compulsives on self-esteem, shopping motivation, materialism, and credit usage. Interestingly, compulsive buyers and impulsive buyers did not display significantly different ratings for shopping involvement.

Discriminant Analysis. It is difficult to directly compare the ability of the two compulsive buying scales to correctly identify compulsive buyers. The Faber and O'Guinn scale incorporates a scoring norm in its application, which typically results in around 10 percent of a group of individuals being identified as compulsive buyers. The Valence et al. scale contains no such norms or scoring to help group people. The end result of using the Valence et al. scale is simply the ability to array a set of subjects on a compulsive buying scale without knowing what cutoff points should be used to identify individuals as compulsive.

Given the characteristics of the two scales, it was decided to use subjects' scores on the Valence et al. scale as a means of grouping respondents and then using the Faber and O'Guinn scores as predictors. The samples' mean score on the Valence et al. scale was used as the cutoff point. Subjects falling two or more standard deviations away from the group mean were grouped into a compulsive and noncompulsive category, respectively. The resulting sample was then randomly split into halves to form a analysis and holdout sample. The results of the discriminant analysis are reported in Table 7.



Three items on the Faber and O'Guinn scale were significant predictors of compulsive buying group membership: a) "..have to spend any money left"; b) "..bought to make myself feel better" and c) ".. felt anxious when I didn't shop". These items correctly predicted group membership an average of 93.44 percent in the analysis sample and 89.66 percent in the holdout sample. The items in the Faber and O'Guinn scale do a good job of identifying individuals with compulsive buying tendencies, as measured by the Valence et al. scale.


The results of the analysis comparing the Valence et al. and Faber and O'Guinn scales suggest that both scales perform reasonably well and exhibit the dimensionality claimed for them by their authors. The Valence et al. scale taps compulsive buying tendencies, while the Faber and O'Guinn scale is able to identify the more extreme cases of compulsive buying behavior. The Valence et al. scale exhibited multiple dimensions in its confirmatory factor structure, while the Faber and O'Guinn scale showed only one. Additionally, attempts to combine the two sets of scale items proved ineffective, so there is the suggestion that each scale is measuring a different, but related underlying construct.

The cluster profiles exhibited by each scale were similar, but the Faber and O'Guinn respondent profiles were more differentiated than the Valence et al. profiles. Also, the credit usage of the compulsive buyer group from the Faber and O'Guinn scale analysis was more extreme than that of the compulsive buyer group from the Valence et al. scale analysis. Thus, the pattern of results suggest that the Faber and O'Guinn scale results in the identification of more extreme compulsive buyers, while the Valence et al. scale results in the measurement of a group of respondents' compulsive tendencies.

There is also a small amount of evidence suggesting that the Valence et al. scale leans more toward identifying compulsive shopping tendencies as opposed to compulsive buying tendencies. More of the Valence et al. scale items deal with the idea of shopping (six of 12 items mention or refer to shopping versus three of 13 items for the Faber and O'Guinn scale) than do the Faber and O'Guinn scale. Shopping involvement and motivation were correlated more highly with the Valence et al. scale than with the Faber and O'Guinn scale. In sum, it would appear that the two scales are compatible, but distinct. Thus, researchers working in this area might consider using both scales to capture as much of the variance in the construct as possible.

Future research in this area should concentrate on exploring the extent and validity of our suggestion about the compulsive shopping orientation of the Valence et al. scale. The use of different subject groups, different types of shopping behaviors, and additional antecedent variables may help pinpoint the distinct differences between the two existing scales to measure compulsive consumption behavior. Another area of needed work concerns the development of norms for the Valence et al. scale, paralleling the norms for the Faber and O'Guinn scale. Such benchmarking activity would make the two scales more directly comparable.


Anderson, James C. and David W. Gerbing (1988), "Structural Equation Modeling in Practice: A Review and Recommended Two-Step Approach," Psychological Bulletin, 103, 411-423.

Attaway, Jill (1989), Influence of an Expanded Framework of Shopping Motivations and Inclusion of Non-Store Retailers on the Choice Set Formation Process, unpublished dissertation, Baton Rouge, LA: Louisiana State University.

Bagozzi, Richard and Youjae Yi (1988), "On the Evaluation of Structural Equation Models," Journal of the Academy of Marketing Science, 16 (Spring) 74-94.

Belk, Russell, (1985), "Materialism Trait Aspects of Living in the Material World," Journal of Consumer Research, 12 (December), 265-280.

Calder, Phillips and Tybout (1981), "Designing Research for Application," Journal of Consumer Research, 8 (September)197-207.

d'Astous, Alain (1990), "An Inquiry into the Compulsive Side of "Normal" Consumers," Journal of Consumer Policy, 13 (March), 15-31.

d'Astous, Alain, Julie Maltais, and Caroline Roberge (1990), "Compulsive Buying Tendencies of Adolescent Consumers," in Advances in Consumer Research, Vol. 17, eds. Marvin Goldberg, Gerald Gorn, and Richard Pollay, Provo, UT: Association for Consumer Research, 306-312.

Faber, Ronald J. and Thomas C. O'Guinn (1992), "A Clinical Screener for Compulsive Buying," Journal of Consumer Research, 19 (December), 459-469.

Faber, Ronald J. and Thomas C. O'Guinn (1988b), "Compulsive Consumption and Credit Abuse," Journal of Consumer Policy, 11, 97-109.

Faber, Ronald J. and Thomas C. O'Guinn (1988a), "Dysfunctional Consumer Socialization: A Search for the Roots of Compulsive Buying," presented at the 13th International Association for Research in Economic Psychology Colloquium, Leuven, Belgium, September.

Faber, Ronald J. and Thomas C. O'Guinn (1989), "Classifying Compulsive Consumers: Advances in the Development of a Diagnostic Tool," in Advances in Consumer Research, Vol. 16, ed. Thomas K. Srull, Provo, UT: Association for Consumer Research, 745-748.

Faber, Ronald J., Thomas C. O'Guinn and Raymond Krych (1988), "Compulsive Consumption," in Advances in Consumer Research, Vol. 14, eds. Melanie Wallendorf and Paul Anderson, Provo, UT: Association for Consumer Research, 132-135.

Heatherton, Todd F. and Janet Polivy (1991), "Development and Validation of a Scale for Measuring State Self-Esteem", Journal of Personality and Social Psychology, 60 (6), 895-910.

Hirschman, Elizabeth C. (1992), "The Consciousness of Addiction: Toward a General Theory of Compulsive Consumption," Journal of Consumer Research, 19 (September), 155-179.

J÷reskog, Karl G. and Dag S÷rbom (1989), LISREL 7: A Guide to the Program and Applications, 2nd edition, Chicago: SPSS.

Nataraajan, Rajan and Brent G. Goff (1992), "Manifestations of Compulsiveness in the Consumer-Marketplace Domain," Psychology and Marketing, 9 (January), 31-44.

O'Guinn, Thomas C. and Ronald J. Faber (1989), "Compulsive Buying: A Phenomenological Exploration," Journal of Consumer Research, 16 (September), 147-157.

Park, C. Whan, Bernard J. Jaworski and Deborah J. MacInnis (1986), "Strategic Brand Concept-Image Management," Journal of Marketing, 50 (October), 135-145.

Richins, Marsha L. and Scott Dawson (1992), "A Consumer Values Orientation for Materialism and Its Measurement: Scale Development and Validation", Journal of Consumer Research, (December).

Rook, Dennis W. and Stephen J. Hoch (1984), "Consuming Impulses", Advances in Consumer Research, in Advances in Consumer Research, Vol. 12 eds. Elizabeth C. Hirschman and Morris B. Holbrook, Provo, UT: Association for Consumer Research, 36-40.

Scherhorn, Gerhard (1990), "The Addictive Trait in Buying Behaviour," Journal of Consumer Policy, 13 (March), 33-51.

Scherhorn, Gerhard, Lucia A. Reisch, and Gerhard Raab (1990), "Addictive Buying in West Germany, " Journal of Consumer Policy, 13 (December), 355-387.

Valence, Gilles, Alain d'Astous, and Louis Fortier (1988), "Compulsive Buying: Concept and Measurement," Journal of Consumer Policy, 11, 419-433.

Wells, William D. (1993), "Discovery-Oriented Consumer Research," Journal of Consumer Research, 19 (March), 489-504.

Zaichowsky, Judith (1986), "Measuring the Involvement Construct," Journal of Consumer Research, 12 (December), 341-352.