Cognitive Factors in Evaluative Conditioning: The Role of Attention, Working Memory, and Contingency Appraisal

In evaluative conditioning, the liking of a stimulus (the conditioned stimulus; CS) is being changed by pairing it with another either liked or disliked stimulus (the unconditioned stimulus; US). In terms of classical Pavlovian conditioning, the shift in the evaluative response can be referred to as a conditioned response. In the last two decades, there has been a great deal of research on whether this type of Pavlovian conditioning depends on controlled (conscious) processes, and particularly the participants' explicit knowledge about the stimulus pairings (contingency awareness). The empirical evidence on the (causal) role of contingency awareness in evaluative conditioning, however, is rather inconsistent. Whereas some studies found conditioning effects in the absence of contingency knowledge, others reported the effect to be restricted to contingency-aware participants. Most researchers refer to contingency knowledge as the recognition of CS-US pairings in a test after conditioning. The contingency recognition performance is thus expected to depend on several cognitive resources like attention and working memory. Only a few studies separately addressed the role of these cognitive factors in the acquisition awareness and evaluative conditioning. Furthermore, the propositional knowledge about the CS-US contingency may be more or less sophisticated. In addition to merely memorizing the stimulus pairings (contiguities), a participant might, for example, learn something about the statistical contingency between stimuli, as well. The effects of these cognitive factors on evaluative learning was investigated in the experiments of this dissertation. A model describing the relation between different propositional-learning processes that may influence the acquisition of contingency knowledge and evaluative conditioning is presented in the synopsis. The thesis comprises three manuscripts that are reporting five experiments in total on the impact of attention, working memory and propositional knowledge about the CS-US contingencies (contingency judgments) on the occurrence and the magnitude of evaluative conditioning. In manuscript A, an experiment is reported (N=109) that investigated the kind of attention required in evaluative learning. Particularly, it was tested whether attention, rather than contingency awareness, is sufficient for evaluative conditioning to occur. In principle, attention can be focused either on the stimuli (the CS and the US) or on the contingency between the stimuli. Since the acquisition of contingency awareness might require a focus on the CS-US contingency, it would be interesting to see whether evaluative conditioning occurs if attention is paid to the stimuli but not to the contingencies. Therefore, three secondary tasks were implemented during conditioning in which the focus of attention was either directed to the stimulus contingencies, diverted completely from the stimuli, or diverted selectively only from the contingencies while maintaining a stimulus focus. Evaluative conditioning effects occurred only in those participants who attended the CS-US contingencies, but not when attention was diverted from the stimuli or from the contingencies. These results show that mere attention is not sufficient for evaluative conditioning to occur, if attention is diverted selectively from the CS-US contingencies. Since a contingency focus is assumed to be essential for the acquisition of contingency awareness, the results are in line with the assumption that evaluative conditioning relies on the participants' CS-US awareness during the acquisition. In addition to requiring a specific form of attention, the acquisition of contingency knowledge is assumed to depend on working memory resources being available during the exposure to the CS-US pairings. In a series of three experiments (N=109) reported in manuscript B, working memory capacity was manipulated during evaluative conditioning by means of phonological distraction. Particularly, in a verbal conditioning paradigm, the encoding of the CS-US contingencies (pairs of words) was disrupted by the playback of irrelevant speech which is assumed to gain automatic access to the same working memory module. Both, the processing and the production of irrelevant speech were shown to reduce both contingency memory and the magnitude of evaluative conditioning. These results imply that evaluative learning requires working memory resources, and they challenge the assumption that evaluative conditioning relies on automatic processes. However, the conditioning effect was not restricted to participants who were able to recall the CS-US contingencies in these experiments. Thus, though working memory may be important during conditioning in order to encode the CS-US contingencies, evaluative learning does not necessarily depend on the acquisition of long-term knowledge about the contingencies. Manuscript C is concerned with the content of the contingency knowledge. According to the propositional account, evaluative conditioning is assumed to rely on the formation and evaluation of propositional knowledge about the predictive relation between CS and US. Thus, stronger evaluative conditioning effects may be expected if the participants perceive a strong statistical CS-US contingency than if they perceive a weak contingency. In the experiment reported in manuscript C (N=31), the evaluative conditioning procedure was combined with a contingency learning paradigm requiring the participants to simultaneously judge the contingencies of four CS-US pairings. The subjective appraisal of contingency was manipulated by varying (a) the objective CS-US contingency and (b) the density of the US which is known to bias contingency judgments (the outcome density effect). This method allows to modulate propositional knowledge about the contingency independently of objective values of contingency during the acquisition of evaluative conditioning. Indeed, contingencies were judged to be stronger in case of low US density irrespective of the level of the objective contingency. More importantly, comparable effects of evaluative conditioning occurred with low and with high objective contingency, but the magnitude of conditioning increased with subjective contingency judgments. These results indicate that evaluative conditioning is sensitive to propositional knowledge about the CS-US relations (i.e. contingency judgments). Evaluative conditioning appears to (a) depend on contingency awareness and (b) vary with subjective beliefs about the CS-US contingency. This result is in line with the propositional account of evaluative conditioning. Taken together, the results show (a) that evaluative conditioning depends on those cognitive resources (i.e. attention and working memory) which enable the acquisition of contingency knowledge and (b) that propositional knowledge modulates the magnitude of the conditioning effect. It is difficult to explain these findings by referring to evaluative conditioning as an automatic process of association formation. Rather, the results imply that propositional learning about the CS-US contingencies is involved in the acquisition of evaluative responses. This form of affective-evaluative learning was shown to require a specific focus of attention (a contingency focus) as well as working memory resources to be available during acquisition. Furthermore, the liking of a CS seems change as a function of the subjective appraisal of the statistical CS-US contingency