Dichotomous thinking in borderline personality disorder
This study examined dichotomous thinking in borderline personality disorder (BPD). In particular, it sought to determine whether dichotomous thinking is characteristic of individuals with BPD, and whether it is unidimensional—all-good or all-bad, and thus synonymous with splitting, or multidimensional—characterized by a mix of positive and negative attributes. The other objectives of the study were to determine whether dichotomous thinking in BPD is correlated with (1) self-injury, and (2) borderline severity. 16 individuals with BPD, 16 individuals with other personality disorders (OPD), and 16 normal controls (NPD) were presented with 6 brief film clips depicting interpersonal situations of varying emotional intensity. 2 of the clips depicted themes thought to be specific to borderline pathology: rejection, abuse, and a relationship crisis. 4 clips depicted themes that were not specific to BPD, and included emotionally positive and neutral themes. After viewing the clips, subjects were asked to rate the characters using visual analogue scales with bipolar trait descriptions. Borderline severity and personality diagnoses were assessed using the PDQ. Frequency of self-injury was assessed using the Self-Injury Interview (McKay, Greiner, Greisberg, Napolitano & D'Andrea, 2000).
The BPD group exhibited significantly more dichotomous thinking than the NPD group; differences between the BPD and OPD group were not significant. Dichotomous thinking in the BPD was predominantly multidimensional, rather than unidimensional. It was not limited to borderline-specific stimuli, but occurred in emotionally positive and neutral circumstances as well. Dichotomous thinking was correlated with the total number of borderline traits endorsed, but was not correlated with frequency of self-injury.
Analyzed at the aggregate level, none of the 3 subjects groups exhibited splitting or purely unidimensional evaluations. However, visual inspection of the data revealed limited instances of splitting within each group. Splitting by the BPD and OPD groups was more frequent than splitting within the NPD group. Additionally, splitting for the BPD and OPD groups, unlike the NPD group, occurred most frequently in response to positive stimuli. Notably, splitting accounted for a very small proportion of the total ratings made by individuals in the BPD group, and 37% of the BPD group did not split.