International Journal of Mental Health and Addiction

, Volume 10, Issue 6, pp 927–935

Difficulty Identifying Feelings, Distress Tolerance and Compulsive Buying: Analyzing the Associations to Inform Therapeutic Strategies

Authors

    • Department of PsychologySouthern Illinois University Edwardsville
  • Daniel J. Segrist
    • Department of PsychologySouthern Illinois University Edwardsville
Article

DOI: 10.1007/s11469-012-9389-y

Cite this article as:
Rose, P. & Segrist, D.J. Int J Ment Health Addiction (2012) 10: 927. doi:10.1007/s11469-012-9389-y

Abstract

Difficulty identifying feelings (a component of alexithymia) and distress tolerance both appear to play a role in impulse-control problems. The goal of the present study was to build upon past research by developing a model of the relations between these constructs and compulsive buying. Participants from the United States and Canada completed a survey containing well-established measures of demographic variables, difficulty identifying feelings, distress tolerance and compulsive buying. In support of a hypothesized model, the three constructs were significantly related in predicted directions and distress tolerance fully mediated the relationship between difficulty identifying feelings and compulsive buying. These results confirm the relationship between alexithymic tendencies and distress tolerance and extend previous findings concerning the problematic behaviors (e.g., substance abuse, pathological gambling) of people who have difficulty identifying their feelings. They also highlight attributes and skills (e.g., tolerating distress, identifying feelings) which clinicians might beneficially target while working with clients who buy compulsively.

Keywords

Compulsive buyingAlexithymiaDistress toleranceDifficulty identifying feelingsSelf-regulation

For people who buy compulsively, buying can become such a strong preoccupation (Ridgway et al. 2011) and feel so uncontrolled that it causes significant personal and interpersonal distress (Kellett and Bolton 2009; Davenport et al. 2011; Dittmar 2004; O’Guinn and Faber 1989). As the negative consequences of such buying increase, compulsive buyers might use additional buying binges to relieve distress or boost mood (Billieux et al. 2008; Kellett and Bolton 2009; Kukar-Kinney et al. 2009; Miltenberger et al. 2003; Thornhill et al. 2012). These buying patterns and their consequences are probably exacerbated by the variety of emotional difficulties compulsive buyers experience (Mowen and Spears 1999; Mueller et al. 2011), including, for example, mood and anxiety disorders (Black 2001; de Zwaan 2011). Several studies have identified comorbidity between compulsive buying and other disorders (e.g., McElroy et al. 1994; Christenson et al. 1994) and in one study, 9.3 % of a voluntarily admitted inpatient sample had significant problems with compulsive buying (which was the most commonly reported impulse-control problem; Grant et al. 2005).

Compulsive buying can be distinguished from similar phenomena (such as buying that is merely immoderate or low financial-management skill) by the emotional difficulties and impairment associated with it (Hollander and Allen 2006; Kellett and Bolton 2009; Koran et al. 2006). Furthermore, compulsive buying is more than just a component of a more fundamental trait. Depression (Mueller et al. 2011; de Zwaan 2011), narcissism (Rose 2007) and neuroticism (Mowen and Spears 1999) predict compulsive buying, but these associations tend to be statistically weak or moderate. Nor is compulsive buying just a symptom of obsessive-compulsive disorder (OCD): Christenson et al. (1994) observed with their sample that the majority of people with compulsive buying problems did not exhibit any other symptoms of OCD. Ridgway et al. (2008) have argued that compulsive buying is best conceptualized as having features of both OCD and impulse-control disorders.

Compulsive buying has drawn substantial research and clinical attention only relatively recently (Kellett and Bolton 2009) and some authors have noted that there is a relative paucity of treatment literature on this problem (Benson and Gengler 2004; Steffen and Mitchell 2011; Steketee and Frost 2003). Nevertheless, research has already made clear that poor emotional self-regulation is central to compulsive buying and should probably be targeted during therapy. Compared to normal buyers, compulsive buyers are more likely to (a) feel badly prior to shopping (Dittmar 2004; McElroy et al. 1994; Miltenberger et al. 2003) and (b) experience a decline in positive feelings (Dittmar 2004) and an increase in negative feelings such as regret and shame (Dittmar 2004; Edwards 1993) after the “high” of purchasing is over. Given that, for at least some compulsive buyers (cf. Thornhill et al. 2012), negative mood appears to serve as a buying trigger and buying episodes may ultimately contribute to this trigger, it seems appropriate to consider how skilled compulsive buyers are at understanding the feelings that they experience. After all, choosing healthy rather than unhealthy strategies to cope with aversive feelings may require a nuanced understanding of one’s emotional experience (e.g., “I’m feeling more ashamed than angry”).

Individual differences in difficulty identifying feelings are a well-researched component of alexithymia, a broad syndrome that includes not only difficulty identifying feelings, but also difficulty describing feelings, restricted fantasizing ability and a tendency to focus attention on external events (Taylor et al. 1997). Alexithymia has been associated with a variety of problems that, like compulsive buying (cf. Ridgway et al. 2011), are characterized by poor impulse control. Eating disorders (e.g., Loas et al. 2001; Schmidt et al. 1993), substance-related addictions (e.g., Handelsman et al. 2000; Loas et al. 2001) and process addictions such as pathological gambling (e.g., Parker et al. 2005) and internet addiction (e.g., DeBerardis et al. 2009) have all been linked to alexithymia. Numerous addictive behaviors also correlate (negatively) with emotional intelligence (Kun and Demetrovics 2010), a construct strongly related to alexithymia (Parker et al. 2001).

However, the components of alexithymia may not be equally strong predictors of impulse control problems (or other psychopathological symptomology; see Grabe et al. 2004; Mason et al. 2005). In a study of women, van Strien and Ouwens (2007) found that difficulty identifying feelings (DIF) was a much stronger correlate of domain-general impulsivity than difficulty describing feelings. Mitrovic and Brown (2009) found that, among all alexithymia components, DIF had the strongest correlation with problem gambling. Loas et al. (2001) observed that DIF scores showed the largest difference (among all alexithymia components) between a group of clients diagnosed with a substance use disorder or an eating disorder and a non-clinical group. Reid et al. (2008) found that DIF was the alexithymia component that best predicted hypersexual behavior.

One reason DIF may be associated with impulse-control problems is that people prone to DIF may have difficulty tolerating distress. When people feel badly but are unsure what their aversive feelings are, they may feel incapable of managing their mysterious, aversive feelings and choosing healthy coping strategies (Anestis et al. 2011). Unhealthy coping strategies, such as those that provide a “high” but involve future negative consequences, may then be more tempting. Thus, poor distress tolerance (i.e., an inability to tolerate aversive states; see Simons and Gaher 2005) may play a role in sustaining the association between DIF and impulse-control problems. With respect to addictive behaviors, distress tolerance “has been implicated as a key mechanism across the stages of addiction, from substance use initiation among adolescents to relapse following cessation among dependent individuals” (Richards et al. 2011, p. 191). Recent evidence confirms that distress tolerance is a predictor of nicotine dependence (Leyro et al. 2011), alcohol problems among men (Simons and Gaher 2005) and abstinence duration among pathological gamblers (Daughters et al. 2005).

Kellett and Bolton (2009) have noted that, like other impulse-control problems, compulsive buying is not explained by any one cause; it appears to be multi-determined (see also Davenport et al. 2011). With the intent to examine likely but as yet unidentified predictors of compulsive buying, we hypothesized that DIF and distress tolerance (operationalized as intolerance of distress) would both correlate positively with compulsive buying. We further hypothesized, based on the previously mentioned idea that impulsive, unhealthy behaviors might be more likely to occur when people feel unable to tolerate aversive but difficult-to-identity feelings, that distress tolerance would mediate the relationship between DIF and compulsive buying. We tested our hypotheses in a sample of internet workers spread across the United States and Canada. By doing so, we avoided the generalizability problem that is common to the many compulsive buying studies that use undergraduate samples (Claes and Müller 2011).

Method

Participants

Two-hundred thirteen adults (Mage = 31.88, SDage = 10.29) completed an online survey. The sample was 47 % male and 53 % female. Seventy-seven percent of the participants were European American, 5 % were African American, 3 % were Hispanic American, 11 % were Asian American, 0.50 % were Native American and the remainder described their ethnicity in a way that did not fit any specified category. Participants’ mean household income was in the $25,001–$50,000 (U.S.) range. Participants’ mean education level was between the associate’s (two-year) degree and bachelor’s (four-year) degree response categories. Because both income and education were measured on ordinal scales, these variables were dichotomized (through median splits; 0=low, 1=high) in preparation for the analyses reported below.

Procedure

Participants were recruited through http://www.mturk.com/, an online marketplace where workers willing to do simple computer tasks are matched with requesters who need computer tasks done. The task was described on the web site as an online survey about personality, habits and buying experiences and workers who agreed to complete the task were paid for their participation. Buhrmester et al. (2011) have documented that samples obtained through http://www.mturk.com/ are more demographically diverse than convenience samples, including typical samples of online participants. Workers registered with http://www.mturk.com were eligible for our survey only if their work approval rate was greater than 93 % (suggesting their prior computer work done through the web site was of high quality) and they had a U.S. or Canadian address at the time they registered as a worker. This study was approved by a university IRB.

Key Measures

To minimize respondent error, participants used 1 (strongly disagree/never) to 7 (strongly agree/very often) scales with all but the demographic measures. Thus, higher scores on each of the key variables represent higher standing on the construct (or poorer psychological functioning). To measure DIF, we used the seven-item DIF subscale of the Toronto Alexithymia Scale (TAS-20), a subscale which represents the first and most internally consistent of three factors that emerge in analyses of the TAS-20 (Bagby et al. 1994; Parker et al. 2003). We measured distress intolerance with Simons and Gaher’s (2005) Distress Tolerance Scale. (We refer to our operationalized variable as distress intolerance because we reversed the response scale used by Simons and Gaher 2005, to ensure that higher scores reflect low ability to handle distress.) Compulsive buying was measured with the six-item scale developed by Ridgway et al. (2008), a scale that overcomes several disadvantages of older measures of compulsive buying (see Ridgway et al. 2008). Each scale was internally consistent (DIF: α = 0.91; distress intolerance: α = 0.93; compulsive buying: α = 0.78).

Although for some purposes compulsive buying can be treated as a categorical variable (e.g., Christenson et al. 1994; Miltenberger et al. 2003; Thornhill et al. 2012), there are several good reasons to treat compulsive buying as a continuous or dimensional variable in some circumstances. Researchers have repeatedly noted that compulsive buying tendencies vary along a spectrum (e.g., Sood and Vaswani 2009) or continuum (e.g., d’Astous 1990; Dittmar 2004; Edwards 1993) and many authors have noted the problems with assuming psychological problems exist only in categories (e.g., Edens et al. 2006; Krueger et al. 2005; Widiger and Trull 2007). Although criteria have been proposed for a compulsive buying disorder (McElroy et al. 1994), there is no consensus on whether compulsive buying warrants designation as a disorder or on how it should be diagnosed (see Black 2007; Hollander and Allen 2006) or classified (Faber 2011). Moreover, recent meta-analyses have demonstrated that continuous measures of psychological problems tend to be more reliable and valid and offer several theoretical advantages over categorical measures (Markon et al. 2011). Therefore, for both theoretical and measurement reasons, we followed the norm for compulsive buying research that is published in psychological (and consumer behavior) journals and operationalized compulsive buying as a continuous variable.

Results

As shown in Table 1, compulsive buying was significantly related to DIF and distress intolerance, as we predicted. (It was also negatively related to age, as past research has shown [e.g., Koran et al. 2006].) DIF was positively related to distress intolerance (as predicted) and negatively related to age and highest level of education achieved.
Table 1

Intercorrelations among and descriptive statistics for key study variables

 

M (SD)

Sex

Age

Income

Educ.

DIF

Dist. Intol.

Comp. Buy.

Sex

0.53 (0.50)

 

0.22*

−0.01

−0.02

−0.06

0.06

0.06

Age

31.88 (10.29)

  

0.13

0.12

−0.18*

0.01

−0.18*

Income

0.70 (0.46)

   

0.18*

−0.11

−0.07

−0.04

Education

0.58 (0.50)

    

−0.20*

−0.04

−0.06

DIF

18.98 (10.10)

     

0.50*

0.21*

Dist. Intol.

3.75 (1.19)

      

0.26*

Comp. Buy.

15.03 (5.91)

       

N’s range from 208 to 213 due to occasional missing data. For sex, 0=male, 1=female. For income and education, 0=low, 1=high. DIF difficulty identifying feelings. Dist. Intol. distress intolerance. Comp. Buy. compulsive buying

* p < .05

Multiple regression analyses provided preliminary support for the hypothesis that distress intolerance mediates (or accounts for the shared variance between) the association between DIF and compulsive buying. As shown in the first column of Table 2, DIF significantly predicted compulsive buying (β = 0.18, p = .01) even when sex, age, income and education were controlled. DIF also predicted distress intolerance (β = 0.54, p < .001) when the same demographic variables were controlled. When compulsive buying was simultaneously regressed on DIF and distress intolerance (as well as the control variables—see the third column of Table 2), the effect of distress intolerance was significant (β = 0.23, p = .01) whereas the effect of DIF substantially shrunk (β = 0.05, p = .52). These results suggest that distress intolerance fully mediates the DIF-compulsive buying relation.
Table 2

Multiple regression analyses

Predictors

Model 1 (DV: Compulsive Buying) β’s

Model 2 (DV: Dist. Intol.) β’s

Model 3 (DV: Compulsive Buying) β’s

Sex

0.11

0.05

0.10

Age

−0.17*

0.09

−0.19*

Income

0.01

−0.04

0.02

Education

−0.02

0.06

−0.03

DIF

0.18*

0.54*

0.05

Dist. Intol.

--

--

0.23*

R2

0.07

0.28

0.11

F

3.02*

15.26*

4.05*

N’s range from 200 to 203 due to occasional missing data. DV dependent variable. DIF difficulty identifying feelings. Dist. Intol. distress intolerance

*p < .05

To fully test for mediation, we analyzed the indirect effect of DIF on compulsive buying through distress intolerance using a bootstrapping technique for which Preacher and Hayes (2008) have documented advantages over earlier methods (e.g., Sobel 1982; Baron and Kenny 1986) of analyzing indirect effects. Using this technique, which involved extracting 10,000 samples from the data and estimating the indirect effect from the resampled data set, we determined that the indirect effect of DIF on compulsive buying through distress intolerance was significant. The bootstrapped point estimate was 0.06 and the 95 % bias-corrected and accelerated confidence interval (see Efron 1987) for the indirect effect was 0.02 to 0.11 (i.e., it did not include zero). The results of this indirect effect test further support the hypothesis that distress intolerance mediates the DIF-compulsive buying association.

For comparison, we analyzed (with the same techniques) an alternative model in which DIF served as a mediator of the relation between distress intolerance and compulsive buying. In these analyses, DIF failed to account for a substantial portion of shared variance between distress intolerance and compulsive buying and the indirect effect was not significant. This alternative model, therefore, seems implausible on both logical and empirical grounds.

Discussion

Although this study is limited by its non-experimental design and self-report methods, the results enhance our understanding of compulsive buying and test ideas that were previously based only on clinical observation. Barth (2000) has observed that clients with buying (and eating) problems exhibit low tolerance for their feelings and often fail to fully process them. Krueger (2000) has emphasized that such clients need help identifying and verbalizing their emotions. Our results are consistent with these clinical observations and our mediational results in particular suggest that poor distress tolerance may sustain the link between DIF and compulsive buying. In light of suggestions that people who are relatively confused by their feelings may be less capable of constructively coping with distress (Anestis et al. 2011; Barth 2000), our results strengthen the case for helping clients with DIF improve their distress tolerance skills.

By focusing on feeling identification and distress tolerance, the present study builds on the emphasis Kellett and Bolton’s (2009) cognitive-behavioral model of compulsive buying places on internal states as buying triggers. Distress from a variety of sources may trigger buying episodes, including the feelings of regret, guilt and shame that compulsive buyers may experience after a buying binge (Kellett and Bolton 2009; see also Dittmar 2004). According to the cognitive-behavioral model of compulsive buying, buying can distract consumers from distress by directing attention elsewhere and creating a tangible goal that consumes attention (i.e., the goal to purchase). Krueger (2000) has described this process as part of an “action symptom” that distracts the compulsive buyer from distress that could otherwise be processed more deeply. Our results align with these perspectives by showing that consumers who are less able to identify and tolerate aversive feelings are more prone to maladaptive buying habits.

The sample of participants in our study, which consisted of North Americans registered with an online marketplace for computer work, somewhat limits the generalizability of the study results. However, the sample was much more diverse than the community and university samples used in most psychological research (cf. Buhrmester et al. 2011) and our intent was to test our hypotheses with data obtained from across the continua of each of our key variables. With the limitations of our study in mind, we suggest that one implication of our results is that helping clients who buy compulsively by teaching them to persevere through unpleasant feelings (e.g., helping them to build distress tolerance skills through Dialectical Behavior Therapy; Linehan 1993) may be beneficial. In addition, helping these clients implement alexithymia-reducing techniques such as those developed by Vanheule et al. (2011) and Kennedy and Franklin (2002) may also be of value. These techniques include practicing the verbalization of difficult situations and emotional responses to those situations, validating clients’ feelings and having clients write about emotional experiences (i.e., identifying and labeling feelings as well as associated physical reactions) in a journal. Related techniques that may be constructive include assisting clients who buy compulsively to conceptualize specific aversive feelings (and perceptions of being overwhelmed by those feelings) as triggers for problem buying (cf. Barth 2000; Billieux et al. 2008). These efforts may help clients understand the connections between their antecedent psychological states and their consequent buying behavior (cf. Ertelt et al. 2009).

The number of people needing treatment for buying problems may increase as societies adopt more consumerist values (Kellett and Bolton 2009). We should acknowledge, however, that the detection of compulsive buying problems can be a challenge for clinicians. Benson and Gengler (2004) have noted that because compulsive buying is thought of as “the ‘smiled upon addiction’ (Catalano and Sonenberg 1993), society conspires against our taking compulsive buying seriously” (p. 453). Moreover, compulsive buying is often not a client’s presenting problem (Benson 2000; Goldman 2000); it may be more likely to surface as therapy proceeds (Benson 2000). In light of these difficulties, our results suggest that clinicians working with clients exhibiting difficulty identifying feelings or poor distress tolerance should inquire about excessive buying and similar impulse-control problems.

Copyright information

© Springer Science+Business Media, LLC 2012