Date of Award


Document Type


Degree Name

Doctor of Philosophy (PhD)



First Advisor

James H. Geer


The information processing perspective of cognition and emotion has been a fruitful area of research inquiry in recent years. As a result of this recent spate of interest, emotional processing biases in anxiety have been consistently demonstrated, indicating that anxious individuals possess a processing priority for threatening information. This study is an attempt to further examine the processing biases which characterize anxious and nonanxious individuals. Pilot work was conducted to investigate the utility of a novel approach, the affective categorization task, to examine the affective meaning of lexical stimuli. Pilot research using the affective categorization task with socially anxious individuals suggested that socially anxious participants exhibited a tendency to evaluate subliminally presented threatening information more accurately than nonanxious participants. Nonanxious individuals were more accurate in detecting the affective content of neutral and positive information. Furthermore, socially anxious participants took longer to make affective decisions to emotional information, regardless of valence, than did nonanxious individuals. Based on that preliminary data, a second and more complete experiment was conducted to replicate and extend those findings with generally anxious individuals. A secondary aim of the full study was to explore the relationship between affective categorizations and underlying associative network representations. Overall, results from the affective categorization task were quite similar to those obtained in pilot work. Anxious individuals evidenced an enhanced ability to correctly classify subliminally presented threatening information, whereas, nonanxious participants demonstrated an enhanced ability to correctly classify subliminally presented information that was neutral or positive in affective tone. Signal detection analyses, however, indicated that these results were primarily due to a response bias, or tendency for anxious participants to categorize all subliminally presented information as threatening. Such a bias in responding was not observed in nonanxious participants. There were no differences in decision time to emotionally valent information between anxious and nonanxious individuals. Additionally, contrary to expectations, no group differences were found in network representations using the Pathfinder (Schvaneveldt, 1990) methodology. Implications will be discussed in terms of information processing theories of emotion and cognition.