IMPLICIT AND EXPLICIT ATTITUDE CHANGE positive or 2004).Participants were asked to open virtual doors to a phote d attitude Pet et al found that di the door (e ino the de esulted in either (a bout the target and implic prescntationof negative picture,which wa b)an im mtation of a po tive ni de long pre hoth e uniquely influenced by the imm rem ADE lar t s hy positive when opening these doors res ulted in an im Cftconpnopositionalreasoaiewihanaditionalinirctctcg For this interpretation,it is important to note that the new info of a positive d pro on of a negative pict reFrom the perspectiveof th nformation was not I in the ded on proc of EC and thu Hence was cesses of propositional which suggestsan bout th of ther al reae ing.Even thoush this in on is ob ent The sixth case direc propo tive evaluations (see Figure 8).This attern implies ng tha g to the Al n (a) given lac with the omentarily sidered set of subjectively erge wher iative stru re and (b)the es a basis for pev on pos ognitive dis. is directly opposite to ce en as compared with reiected alternatives (Brehn when they recognize either (a)that does not have or (h)that the chosen alternative has Influence 个↓ fcatrcshiacnotpcscntinthcrcictedahcrmaicd M. often emphasize(Brehm,1956)or delibera 1986)positive of th tive inform ation search n.in tum,leads to Figure 7.Case 5:Direct influence on pr sitional reas g and diree s in explicit and implicit ws reflect a Drawing on Gawronski and Strack's(2004)res of in proposition nature of cogmtive nce.one co argue tha attitude toward an unknown target by repeatedly pairing this target with either positive or negative images. Shortly afterward, partic￾ipants received positive or negative verbal information about the target that either confirmed or disconfirmed the initially condi￾tioned attitude. Petty et al. found that disconfirming information about the target influenced both explicit and implicit attitudes. However, whereas explicit attitudes were completely reversed by the counterattitudinal information, implicit attitudes showed some￾what weaker effects, such that implicit attitudes were only neu￾tralized rather than reversed. When the subsequently presented information confirmed the initially conditioned attitude, both ex￾plicit and implicit attitudes remained stable. From the perspective of the APE model, the prior induction of explicit and implicit attitudes by means of EC represents an example of Case 1. More important, however, the subsequently induced change in attitudes can be interpreted as an example of Case 4, involving a direct effect on propositional reasoning with an additional indirect effect on associative evaluations mediated by propositional reasoning. For this interpretation, it is important to note that the new infor￾mation presented after the EC manipulation represented proposi￾tional statements about the target. Moreover, the particular valence of this information was not directly implied in the statements. Instead, the valence was implied only indirectly, such that it had to be inferred. Hence, extracting the valence of this information required processes of propositional reasoning, which suggests an indirect influence on associative evaluations that was mediated by propositional reasoning. Even though this interpretation is obvi￾ously post hoc, it is consistent with the somewhat weaker effects on implicit attitudes, such that implicit attitudes were only neu￾tralized, whereas explicit attitudes were completely reversed. Case 5. The fifth case involves a direct influence on associa￾tive evaluations and a direct influence on propositional reasoning, with the two influences having distinct effects, such that evaluative judgments are not based on the propositional implications of associative evaluations and propositional processes leave associa￾tive evaluations unaffected (see Figure 7). This case implies dif￾ferent patterns of change for explicit and implicit attitudes, with the two being generally uncorrelated (or negatively correlated). According to the APE model, such patterns should emerge when (a) a given factor leads to changes in pattern activation or asso￾ciative structure and (b) the same factor provides a basis for new propositions that imply an evaluation that is directly opposite to the one implied by associative evaluations. An interesting example for such a pattern can be found in a study by Deutsch and Strack (2002; reported in Strack & Deutsch, 2004). Participants were asked to open virtual doors to a photo gallery on a computer screen. Depending on the particular color of the door (i.e., blue vs. red), opening the door resulted in either (a) an immediate, brief presentation of a negative picture, which was followed by a delayed, long presentation of a positive picture, or (b) an immediate, brief presentation of a positive picture, which was followed by a delayed, long presentation of a negative picture. Results indicate that subsequent associative evaluations of the two types of doors were uniquely influenced by the immediate, short￾term consequences associated with a particular kind of door. That is, associative evaluations of a particular type of door were more positive when opening these doors resulted in an immediate, brief presentation of a positive picture than when it resulted in an immediate, brief presentation of a negative picture. In contrast, evaluative judgments of the two types of doors were more positive when opening these doors resulted in a delayed, long presentation of a positive picture than when it resulted in a delayed, long presentation of a negative picture. From the perspective of the APE model, these results suggest that participants’ associative evalua￾tions depended on processes of EC and thus on the immediate, short-term consequences of their actions. Evaluative judgments, in contrast, were influenced by participants’ propositional beliefs about the long-term consequences of their actions, thus leading to antagonistic effects for associative evaluations and evaluative judgments. Case 6. The sixth case involves a direct influence on propo￾sitional reasoning and an additional indirect influence mediated by associative evaluations (see Figure 8). This pattern implies corre￾sponding changes in explicit and implicit attitudes, with changes in explicit attitudes being partially mediated by changes in implicit attitudes. According to the APE model, such patterns should emerge when (a) a given factor leads to changes in pattern acti￾vation or associative structure, (b) associative evaluations are consistent with the momentarily considered set of subjectively valid propositions, and (c) the same factor provides a basis for new propositions that indirectly imply an evaluation of the same valence. An illustrative example for such a pattern comes from research on postdecisional attitude change. Research in the cognitive dis￾sonance tradition has consistently shown that decisions between two equally attractive alternatives lead to more favorable evalua￾tions of chosen as compared with rejected alternatives (Brehm, 1956). A common explanation for this spreading of alternatives effect is that people experience an uncomfortable feeling of post￾decisional dissonance when they recognize either (a) that the rejected alternative has positive features that the chosen alternative does not have or (b) that the chosen alternative has negative features that are not present in the rejected alternative (J. M. Olson & Stone, 2005). Hence, to reduce this uncomfortable feeling, people often emphasize (Brehm, 1956) or deliberately search for (Frey, 1986) positive characteristics of the chosen alternative and negative characteristics of the rejected alternative. This kind of selective information search, in turn, leads to more favorable evaluations of the chosen alternative but to less favorable evalu￾ations of the rejected alternative. Drawing on Gawronski and Strack’s (2004) research on the propositional nature of cognitive dissonance, one could argue that Figure 7. Case 5: Direct influence on propositional reasoning and direct (antagonistic) influence on associative evaluations, leading to noncorre￾sponding changes in explicit and implicit attitudes. Open arrows reflect a lack of influence. IMPLICIT AND EXPLICIT ATTITUDE CHANGE 707