The majority of research in the compulsive buying literature has used the Compulsive Buying Scale, a self-report questionnaire developed by Faber and O’Guinn (Journal of Consumer Research,19(3), 459–469, 1992). Although this scale has previously shown good psychometric properties, it is now over 30 years old and researchers have critiqued whether some items in the scale are still relevant to the current clinical picture of compulsive buying. The aim of this study was to re-investigate its psychometric properties with an item-level psychometric analysis, while also using differential item functioning to test whether items were biased toward demographic groups (i.e., gender, age). We conducted item response theory analyses on the Compulsive Buying Scale in a Canadian sample recruited through Prolific (N = 927) and tested whether there were differences in responding among men versus women, and younger versus older people. We found evidence for differential item functioning for items 2, 3, 5, and 7. The most pronounced effect was in item 5 (bought something to feel better), in which women endorsed this item more overall and had to score higher than men to indicate the same severity of compulsive buying. This suggests that the Compulsive Buying Scale may be biased in its measurement. We conclude with recommendations for developing new measures to accurately assess compulsive buying problems.
