To date, assessment screens for shopping disorders have primarily been rooted within the impulse-control or obsessive-compulsive disorder paradigms. Consequently, a new screening tool for assessing shopping addiction was developed. Initially, 28 items, four for each of seven addiction criteria salience, mood modification, conflict, tolerance, withdrawal, relapse, and problems , were constructed.
These items and validated scales i. The highest loading item from each set of four pooled items reflecting the seven addiction criteria were retained in the final scale, The Bergen Shopping Addiction Scale BSAS. The scores of the BSAS were positively associated with anxiety, depression, and low self-esteem and inversely related to age. Females scored higher than males on the BSAS.
The BSAS is the first scale to fully embed shopping addiction within an addiction paradigm. A recommended cutoff score for the new scale and future research directions are discussed. Together, these changes represent an increasing recognition of non-chemical addictions. However, at present, most non-chemical addictions are not yet embedded in the psychiatric nosology.
This includes shopping addiction, despite this disorder having been recognized in the psychiatric literature for over a century Kraepelin, Andreassen , p. For the most part, this empirical research also suggests that the typical shopping addict is young, female, and of lower educational background Black, ; Davenport et al. Research has also linked those with problematic shopping to individual characteristics typical for other addictive behaviors Aboujaoude, Some of this research has involved the five-factor model of personality Costa and McCrae, ; Wiggins, Extroversion has been positively associated with shopping addiction Balabanis, ; Mikolajczak-Degrauwe et al.
Neuroticism has also been consistently been related to shopping addiction Wang and Yang, ; Mikolajczak-Degrauwe et al. Neurotic individuals, typically being anxious, depressive, and self-conscious may use shopping as means of reducing their negative emotional feelings. Conscientiousness, on the other hand, appears to be a protective factor Mowen and Spears, ; Wang and Yang, ; Andreassen et al. People with low conscientiousness scores appear to shop due to low ability to be structured and responsible Andreassen et al.
Also, the relationship between agreeableness and shopping addiction appears to be more ambivalent. Some studies have reported a positive relationship Mowen and Spears, ; Mikolajczak-Degrauwe et al. High degrees of agreeableness may represent a protective factor for developing shopping addiction or addiction of any kind , as such individuals typically avoid conflicts and disharmony. Since addictive behaviors often create conflicts with others, it seems reasonable that shopping addiction would be negatively related to agreeableness.
At the same time, agreeable people may be more prone to fall for exploitative marketing techniques since they easily trust others. Finally, the openness to experience trait has typically been unrelated to shopping addiction Mowen and Spears, ; Wang and Yang, ; Andreassen et al. However, at least two studies have reported a negative relationship Balabanis, ; Mikolajczak-Degrauwe et al. Addictive behaviors may also be related to individual differences in self-esteem and psychological distress.
Empirical research has consistently reported significantly lower levels of self-esteem among shopping addicts Davenport et al.
However, this may be related to depression, which has been shown to be highly comorbid with problematic shopping Aboujaoude, It has also been suggested that self-critical people shop in order to escape, or cope with, negative feelings Rick et al. As existing research is primarily based on cross-sectional studies, we know little about the directionality of these relationships.
In this regard, it should be noted that shopping addiction has been explained as a way of regulating neurochemical e. To that effect, an fMRI-study reported a significant difference in activation of reward and pain circuit systems between shopping addicts and non-shopping addicts during purchasing decisions Raab et al. Therefore, an argument could also be made for a biological basis for this condition.
In addition, several items of existing scales appear outdated with regards to modern consumer patterns. Another limitation is that the existing scales are relatively lengthy. Although two new scales have been developed more recently Christo et al. Existing problematic shopping scales typically involve one or several of these symptoms, but fail to fully incorporate them all.
In addition, since new Internet-related technologies can greatly facilitate the emergence of problematic shopping behavior because of factors such as accessibility, affordability, anonymity, convenience, and disinhibition Wang and Yang, ; Widyanto and Griffiths, ; Aboujaoude, , there is a need for a psychometrically robust instrument that assesses problematic shopping across all platforms.
Given this background, a shopping addiction scale [the Bergen Shopping Addiction Scale BSAS ] was developed, containing a small number of items that reflect the seven aforementioned elements of addiction, thus ensuring its content validity in an addiction framework.
It was hypothesized that the new shopping addiction scale would correlate highly with measures of similar constructs convergent validity and less with measures of more divergent or unrelated constructs discriminant validity; Nunnally and Bernstein, Accordingly, the following hypotheses were investigated: Materials and Methods Sample The sample comprised 23, participants 15, females and 8, males.
The mean age was Educational level ranged from primary school Procedure In the first stage of scale development, 28 potential items were included. The scale was constructed based on the seven basic components of addiction originally proposed by Brown and developed and modified by Griffiths Four items for each component were constructed. Since addiction was considered a main construct comprising seven different components, a second-order model was initially set up, with addiction i. For some items the wording was similar to that used in the diagnostic criteria for pathological gambling American Psychiatric Association, and the Game Addiction Scale Lemmens et al.
The questions were distributed via the online edition of five nationwide newspapers in Norway in March, April and May The survey was available online for 1 day up to 1 week on the various newspaper websites. Information about the study purpose was provided immediately after participants clicked the survey link. Consent to participate was deemed as given since participants completed the questionnaire.
Also, after survey completion, participants were provided interactive feedback on their shopping habits. No other incentives were offered in return for participation. All questions were collected anonymously and no interventions were made. Participant responses were stored by an Internet survey agency before being passed over to the research team.
Instruments Demographics Participants were asked for information about their age, gender, level of education, relationship, and occupational status. The Bergen Shopping Addiction Scale In order to develop a new, brief, updated shopping addiction scale, a pool of 28 items were first created. This pool was based on the seven addiction criteria outlined earlier in the paper. Four items for each addiction criterion were constructed based upon diagnostic criteria for pathological gambling American Psychiatric Association, , the Game Addiction Scale Lemmens et al.
The response options were completely disagree 0 , disagree 1 , neither disagree nor agree 2 , agree 3 , and completely agree 4 see Table 1. Higher scores indicate higher levels of shopping addiction. Initial pool items for the shopping addiction scale1. Compulsive Buying Measurement Scale This scale comprises 13 items for assessing compulsive buying tendencies Valence et al. Items are answered on a 5-point scale anchored from Strongly disagree 1 to Strongly agree 5 e.
Higher scores reflect more compulsive buying. The CBMS, initially a pool of 16 impulse-control based items, was the first quantitative measure of compulsive buying tendencies. It is brief, suitable for adolescents and adults, widely used, and exists in several languages. Although the CBMS has proven reliable and valid, it has also been criticized for being outdated Andreassen, Four items reflect each of the personality traits of the established Five-Factor Model of personality Costa and McCrae, ; Wiggins, All items are answered on a 5-point scale ranging from Very inaccurate 1 to Very accurate 5.
Studies have shown acceptable psychometric properties of this personality measure e. All items are answered along a 4-point frequency scale ranging from 0 to 3. However, the frequency categories are different for almost each question asked. Rosenberg Self-Esteem Scale This scale comprises 10 items for assessing levels of self-esteem Rosenberg, Items are answered on a 4-point scale using anchors of Strongly agree 0 and Strongly disagree 3 e. The higher the score, the higher the self-esteem.
A recent meta-analysis provided good support for its factor structure and psychometric qualities Huang and Dong, Statistics Descriptive statistics including frequencies, means, standard deviation and Cronbach alphas were calculated for all study variables see Table 2.
The second-order factor comprised shopping addiction whereas the seven first order factors Salience, Mood modification, Conflict, Tolerance, Relapse, Withdrawal and Problems were each reflected by four items. The fit of these models was investigated by confirmatory factor analyses using AMOS, version In the final model, correlations between error terms were allowed providing this had substantive meaning Byrne, The final scale was further investigated by Cronbach alpha and corrected item-total correlations.
In order to investigate the convergent validity of the new scale, the zero-order correlation with the CBMS Valence et al. Finally, for investigating the convergent as well as the discriminative validity of the new scale, a hierarchical regression analysis was conducted where the new scale BSAS comprised the dependent variable.
The predictors included in the first step were gender, age, and marital status. In the second and final step, symptoms of depression, anxiety, and self-esteem were included, as well as the measures of the five-factor model of personality. Results Scale Construction The second-order factor structure is shown in Figure 1. The standardized second-order factor loadings ranged from 0.
The highest first order loading for each of the seven factors ranged from 0. The standardized factor loadings ranged from 0. The error terms of the two first items were allowed to correlate as they both reflected inner thoughts or states.
The correlation coefficient was 0. The mean inter-item correlation coefficient was 0. Furthermore, both scales showed similar correlation patterns with the other study variables see Table 2.