Given the large number of hypotheses stated above, a number of potential statistical analyses of the collected empirical data are possible. The performed analyses were geared towards explaining previous lexical decision task findings and finding attributes of information processing biases in depressed people which were particularly valuable to model.
The first set of analyses address a replication of previous lexical decision tasks, using the short stimulus durations, and will be compared to results on the valence identification task. Thus, each subject's data is averaged across the word categories rather than accounting for differential responses to some words of a given valence.
The mean of each subject's reaction times in each valence and SD, on each task was calculated. Reactions to stimuli below 150ms were discarded for both tasks. Reaction times over 5 seconds were also discarded. Additionally, reaction times to nonwords on the lexical decision task, to stimuli which were incorrectly identified on the lexical decision task, and to stimuli whose valence was identified as incongruent with the normed valence on the valence identification task were removed from the computation of means.