Processing Asymmetries of Emotionally Valenced Stimuli

Abstract: The central phenomenon investigated concerns the valence-based process asymmetry found in several earlier studies (e.g. Pratto & John, 1991; Taylor, 1991), where negative stimuli seem to initiate more thorough processing than positive stimuli. This finding was consistent in the three empirical studies forming this dissertation. In Study 1 (three experiments) emotionally valenced words were presented at the centre of pictures of emotional faces (angry, sad, disgusted, happy). The results showed a general positive valence advantage (PVA) in reaction times for positive words relative to negative words. Furthermore, except for the words imposed on pictures of angry faces in the first experiment, the PVA increased when words were imposed on pictures of happy faces. In Study 2, the priming effects of emotional words were tested. In both experiments, two conditions were used. When valence had been activated by a valence categorisation task, negatively primed words resulted in prolonged RTs in a subsequent word recognition test. In the other condition, where the participants just read the word, no valence dependent latency differences were found. The results provide support for a non-automatic appraisal of the valence when using words as stimuli.In Study 3 (three experiments) neutral words were imposed on emotional pictures (negative, neutral and positive) in the encoding phase. In the following test phase the phenomenological quality of the memory was measured using the Remember-Know paradigm (Tulving, 1985). The results showed a decrease in the frequency of “remember” responses, suggesting less episodic detail in the retention experience of the to-be-remembered item, when the words had been presented on negative pictures relative to positive pictures. The third experiment tested whether the neutral words were primed by the valence of the encoding context (picture) and whether such affective information could be used as a cue in a subsequent source monitoring task (Johnson, Hashtroudi, Lindsay, 1993). However, no support for the valence-as-a-cue hypothesis was found. The results of the three studies are discussed within the framework of appraisal theories (Lazarus, 1991; Scherer, 2001) and Taylor’s mobilization-minimization hypothesis, and an evolutionary explanation for the processing asymmetry is considered.