Category: Personality Disorders
Background: The body of research on distress intolerance (DI) is developing rapidly. Studies in community and clinical populations have linked DI cross-sectionally to anxiety, depression, substance misuse, and self-harm. Moreover, DI is a popular treatment target (e.g., in dialectical behavior therapy).
Researchers alternately use DI to refer to the perceived capacity to effectively manage aversive internal states (e.g., intrusive thoughts, painful emotions) and the behavioral act of withstanding uncomfortable scenarios. Moreover, some investigators have used autonomic nervous system markers as physiological indices of DI.
A potentially problematic feature of this field is that we do not yet know the extent to which these various indices of DI agree with one another. We argue that the concordance of these measures needs to be quantified so that investigators can understand whether they are capturing the same underlying construct with these diverse assessment methods. Stated differently, we need to find out whether these different modalities (e.g., self-report, behavior, physiology) are equally useful windows into a latent DI trait.
Method: We recruited a number of adult samples to evaluate the alignment of various DI assessment techniques. In our first study (N = 117), we administered self-report and behavioral tests of DI. The behavioral test was a computer task in which participants were instructed to trace the outline of a star shape. However, the cursor on screen moved in the opposite direction of the mouse, making it difficult to trace. Further, whenever participants traced outside the perimeter of the star or paused for more than 1 sec, a loud buzzer noise would sound binaurally through headphones. Participants were told they could discontinue the task at any time, but that they would receive a reward commensurate with their performance. We also monitored heart rate and electrodermal activity during this task.
In a set of two studies in new samples, we administered a variety of behavioral, self-report, and physiological measures of DI. For instance, in one study (N = 100) we assessed performance on a cold pressor task, in which a participant is invited to keep his or her hand in 4 degrees Celsius water for as long as possible, and the same tracing task as above. In another study (N = 170 and counting) we administer the tracing task, a speeded mental arithmetic task, and a breath holding task.
Results: Analyses from our first study revealed that self-reported DI was virtually uncorrelated with persistence on the tracing task (r = -.01) or physiological reactivity (rs ~= -.10). Correlational analyses in the additional samples produced the same pattern of results. In no case did correlations across modalities exceed |.10|.
Conclusion: Diverse measures of DI were not as concordant as expected across a series of studies. While this problem is not unique to DI (i.e., we see it to some extent with other phenotypes like fear), it does threaten to limit the validity of DI assessment in etiological and treatment research, especially studies that rely solely on self-report inventories. We advise more attention to—and creativity with—measurement issues in DI.