This study investigated theoretical claims that different emotional disorders are associated with different patterns of cognitive bias, both in terms of the cognitive processes involved and the stimulus content that is preferentially processed. These claims were tested by comparing clinically anxious (generalized anxiety disorder [GAD], posttraumatic stress disorder [PTSD]) and clinically depressed children and adolescents on a range of cognitive tasks measuring attention, memory, and prospective cognition, with both threat-related and depressogenic stimulus materials. The results did reveal some relative specificity of processing in that the anxious participants exhibited a greater selective attentional bias for threat relative to depressogenic material with no such difference being apparent in the depressed sample. However, this bias was only clear-cut on a dot-probe measure of attentional processing and not on a modified Stroop measure, and indeed threat-related bias on the 2 tasks was uncorrelated. On the prospective cognition task, anxious participants exhibited an other-referent bias in their risk estimations regarding future negative events that was absent in the depressed sample. No specificity effects were evident on the memory task. The results are discussed in terms of the strengths and weaknesses of carrying out direct comparisons across groups and tasks versus drawing conclusions from overall patterns across multiple studies.