Bifactor modelling
Bifactor analysis provides a principled method for achieving this goal. It has been used extensively to study the structure of internalizing symptomology (Clark et al., 1994; Steer et al., 1995; Zinbarg and Barlow, 1996; Steer et al., 1999; Simms et al., 2008; Steer et al., 2008; Brodbeck et al., 2011) and has consistently revealed a substantial amount of shared variance, often termed ‘general distress’ or ‘negative affect’ (Clark and Watson, 1991; Clark et al., 1994). In addition, separate specific factors for depression and anxiety are consistently observed, with the depression-specific factor tapping symptoms of anhedonia (Clark et al., 1994; Steer et al., 1999; Steer et al., 2008) and anxiety-specific factors tapping symptoms of anxious arousal (Clark et al., 1994; Steer et al., 1999; Steer et al., 2008) and worry (Brodbeck et al., 2011). Although bifactor modeling of internalizing symptoms is well established, it has not, to date, been used to inform studies of anxiety- and depression-related deficits in decision-making. Using bifactor analysis to estimate scores for each participant on latent dimensions of internalizing symptoms, we can investigate whether impoverished adjustment of learning rate to volatility is primarily linked to general factor scores (i.e. to symptom variance common to both anxiety and depression) or to anxiety-specific or depression-specific factor scores.