Understanding some characteristics of interpretation bias in general distress

Abstract

Interpretation bias, referred to as the tendency to interpret ambiguous information negatively, is believed to be related to the maintenance or even development of emotional disorders. Prior research has documented associations between interpretation bias and various disorder symptoms, and even demonstrated that manipulating biases could potentially alleviate such symptoms. Although the field has rapidly progressed in a promising direction, research gaps still remain regarding the fundamental conceptualizations of interpretation bias. For instance, it is still unclear whether bias should be differentiated into automatic versus strategic processing components, or whether biases for different types of content matter – implying differential specificity towards disorders – are empirically distinct. A better understanding of the overlapping and distinct features of interpretation bias in relation to psychopathology is paramount, especially given contemporary developments in psychopathology promoting a greater focus on such processes that transcend traditional across disorder diagnoses.

In this thesis, four studies were conducted to explore the conceptualizations of interpretation bias under the context of overlapping psychopathological dimensions. Study One investigated the latter notion by examining the structure of a widely-used questionnaire measuring depression and anxiety. Study Two examined the empirical convergence between a measure of automatic versus strategic interpretation bias – as well as their associations with anxiety, depression, and general distress – while also exploring the potential influence of language. Study Three presented a meta-analysis of the links between interpretation bias and general anxiety/social anxiety, synthesizing previous findings and utilizing these to further understand some characteristics of bias. Finally, Study Four investigated the conceptualization of bias for different content matter using factor analyses, further examining its potential specificity and longitudinal associations with different psychopathological variables.

The results of Study One demonstrated that anxiety and depression were subsumed by a hierarchical unidimensional construct possibly reflecting general distress, congruent with contemporary notions. Study Two found that there was no convergence between the automatic and strategic measures of interpretation bias, and that their associations with psychopathology depended on participants’ first language. The potential distinction between automatic and strategic bias was reinforced in Study Three, though meta-analytic findings implied that automatic bias measures may not show a reliable link to anxiety. Said analyses also highlighted unexpected patterns in the specificity of bias content, which were followed up on in Study Four. Study Four found that there seemed to be a unidimensional underlying conceptualization of bias regardless of ambiguous content, whilst the differentiation of bias into distinct content types was not apparent. This unidimensional construct had comparable links to different psychopathological variables. From the longitudinal analyses it further seemed that psychopathological symptoms were driving subsequent changes in interpretation bias, but not vice versa.

The current series of studies improves our understanding of some fundamental characteristics of interpretation bias. However, it also raises various unanswered questions regarding its conceptualization, particularly highlighting that future research should adopt more methodologically-focused approaches to investigate whether different types of bias content can indeed be distinguished, and whether these might differentially relate to psychopathology. Such knowledge would help to refine interventional approaches targeting interpretation bias moving forward.