Investigating trial types putatively evidencing semantic conflict in the Stroop task.

Abstract

Interference in the Stroop task is thought to arise from various stages of processing, including the semantic and response stages. Different experimental methods have been used in an attempt to dissociate the cognitive processes involved in these stages. The work presented in this thesis evaluates two such methods that have been popular, namely the use of a two-to-one response mapping variant of the task and using colour-word distractors that are not valid response options (non-response set trials). The results from a series of experiments which utilised behavioural and eye-tracking measures, provided (Bayesian) evidence that two-to-one mapping trials do not involve additional interference compared to non-word neutral trials. Studies that have utilised this method are likely to have been measuring facilitation instead of the intended semantic-based interference, which has obvious ramifications to the conclusions of those studies. The experiments that evaluated non-response set trials indicated them as a better alternative, although during the course of the investigation, it was found that the make-up of Stroop interference is affected by experimental design. This is problematic to extant models of selective attention, as they cannot account for such findings. This led to further investigations of the cognitive mechanisms involved in processing relevant and irrelevant information during the Stroop task. The findings revealed that bottom-up implicit learning processes have a greater role in the allocation of attention and establishing task relevant stimuli, than previously thought. These concepts have generally not been given much consideration in theoretical accounts and the results from these experiments highlight their importance. The methodological and theoretical implications of the findings in this thesis are discussed in the context of theories of selective attention in the Stroop task and automaticity.