Positive-negative asymmetry of priming effects on impression formation

Download Positive-negative asymmetry of priming effects on impression formation

Post on 11-Jun-2016

218 views

Category:

Documents

4 download

Embed Size (px)

TRANSCRIPT

<ul><li><p>European Journal of Social Psychology, Vol. 23,l-I6 (1993) </p><p>Positive-negative asymmetry of priming effects on impression formation </p><p>TOMOKO IKEGAMI Department of Psychology, Aichi University of Education, 1 Hirosawa, Iga ya-cho, Kariya Aichi, 448, Japan </p><p>Abstract </p><p>The present study investigates the effects of evaluative processing of hostile and friendly words on impression formation. Two experiments were conducted. Subjects were first asked to complete sentences using hostile (Experiment I ) or friendly (Experiment 2) words. Then they participated in an ostensibly unrelated impression task, where they rated a stimulusperson on a series of trait scales based on aparagraph of behavioural descriptions that was ambiguous regarding hostility (Experiment 1) or friendliness (Experiment 2). It was found that the hostile sentence completion tasks increased the likelihood that subsequent behavioural information would be assimilated into the primed hostile constructs, whereas the friendly sentence completion tasks increased the likelihood that behavioural information would be contrasted to the primed friendly constructs. The underlying mechanism of the positive-negative asymmetry of priming effects was discussed in terms of the socio-cultural expectancy for normality of friendli- ness being evoked by priming manipulations. In addition, a supplementary experiment was conducted to assess the feeling states following the hostile or friendly priming tasks. It was found that the hostile priming elicited negative affects, whereas friendly priming had no eflects on the feeling states of subjects. This asymmetric effect of priming on the feeling states was also considered as a result of the influences of the socio-cultural expectancy. </p><p>INTRODUCTION </p><p>People may form a personality impression of a particular person from hidher utter- ances and behaviours in a variety of situations. However, it is often the case that behavioural information about others has multiple meanings and hence is interpre- table in various ways. The processes whereby inferences and interpretations are drawn </p><p>This paper was originally submitted for the Special Issue, Positive-negative asymmetry in affect and evaluations. </p><p>00462772/93/0 1000 1-1 6$13.00 0 1993 by John Wiley &amp; Sons, Ltd. </p><p>Received 29 January 1991 Accepted 15 January 1992 </p></li><li><p>2 T. Ikegami </p><p>from the given behavioural information are determined to a certain extent by a perceivers cognitive structure. In recent years, social cognitive psychologists, taking an information processing approach, have been exploring why the same individual will interpret the same information differently in different occasions or contexts. This temporary contextual effect on impression formation has been demonstrated by numerous experimental studies using priming technique, and has been conceptua- lized as category or construct accessibility. The notion of category accessibility was in fact first proposed by Bruner (1957) and still has been a key concept for understanding the contextual effects on person impressions in recent social cognitive studies. Most of the previous studies have demonstrated that subjects tend to assimi- late their impressions and evaluations of a target person to the constructs activated in the priming tasks. </p><p>A provocative demonstration was reported by Higgins, Rholes and Jones (1977). Subjects were first exposed to 10 words during an experiment presented to the subjects as a study on perception. In one condition, four trait words related with braveness were included in the list, whereas in another condition four trait words related with recklessness were included. After being exposed to one of these sets of trait words as part of the perceptual study, subjects participated in an ostensibly unrelated task of impression formation. Here, they were presented with a paragraph of behavioural descriptions about a stimulus person, Donald, whose behaviour could be interpreted either as recklessness or as braveness. Subjects were asked to characterize Donald in their own words and to evaluate him. As a result, it was found that in characteriza- tion of Donald subjects typically used trait words and phrases that were semantically or descriptively similar to the priming words. Besides, the overall evaluation of Donald increased with the favourableness of the implications of the prime words. The findings reported by Higgins et al. (1977) suggest that the effect of information on ones impression may be influenced substantially by prior, objectively irrelevant experiences that increase the accessibility of a set of constructs for use in interpreting the information. </p><p>Srull and Wyer (1979, 1980) carried out a series of studies to explore in more detail the underlying mechanisms of the effects of category accessibility and the conditions in which these effects would occur. They reasoned that trait schemata should contain prototypic behaviours that exemplify the trait and that trait categories or schemata could be made accessible by priming behavioural instances of these categories or schemata as well as trait names. Subjects were asked to perform a sentence construction task in which they were exposed to instances of hostile or kind behaviour. After completing the sentence construction task, subjects partici- pated in an impression formation task, where subjects read a paragraph describing the events in one afternoon of a target persons life and rated the target person on a series of trait scales. In one condition, the paragraph described behaviours which were ambiguous with respect to hostility, and in another condition, ambiguous with respect to kindness. It was revealed that exposure to hostile or kind behaviours activated and increased accessibility of the trait schema in which the behaviours were contained, with the result that more ambiguous behaviours with respect to hostility or kindness were led to be interpreted in terms of that trait schema. This happened even if the second set of behaviours were presented in a completely unrelated context after the first set were presented in the preceding tasks. </p></li><li><p>Priming efects and impression formation 3 </p><p>One of the most important things which these studies suggest and should be noted here, is that the processes whereby prior processing of stimuli activates a particular trait schema and then assimilates subsequent behavioural information into that schema, may occur automatically without subjects conscious awareness. In a str- ingent test of this issue, Bargh and Pietromonaco (1982) demonstrated that the prim- ing effects on the impression ratings would appear even if the prime words were presented subliminally in a prior vigilance task. </p><p>Another important and interesting point is that the magnitude of priming effects are not symmetric for positive versus negative prime words. For example, the findings reported by Srull and Wyer (1979) showed that the priming effects were more pro- nounced and durable when the prime words were negative than when they were positive. In other words, more priming was required to activate favourable trait schemata than unfavourable ones (Wyer and Srull, 1981). This suggests that assimilat- ive processes should be more likely to occur in cases of negative priming than in cases of positive priming. </p><p>Besides, several other studies suggest that the level of conscious awareness or the recallability of the priming events may influence the typical pattern of assimilation effects on the impression ratings (Lombardi, Higgins and Bargh, 1987; Martin, 1985, 1986; Ikegami and Kawaguchi, 1989). It seems that the more con- sciously are priming stimuli processed, the less likely are assimilative effects to occur. </p><p>A study by Ikegami and Kawaguchi (1989) was an attempt to relate the issue of consciousness with the positive-negative asymmetry of priming effects. They con- ducted a series of experiments, using a modification of Bargh and Pietromonaco (1982)s paradigm. Subjects first performed a perceptual task in which they inciden- tally processed hostility- or friendliness-related words, consciously or subcons- ciously. To be more precise, in their experiments hostile or friendly jukugo words (Chinese compound words) were presented on a CRT screen either on the right or the left side of a fixation point with a dummy stimulus of # # being simulta- neously presented on the other side. In the conscious priming condition, stimulus words were presented for long enough to identify what they were, whereas in the subconscious priming condition words were presented so briefly that subjects were unable even to identify them. Subjects were instructed to press quickly and accu- rately the left or the right button on a response box corresponding to the side of the fixation point on which a jukugo word had appeared. In a second ostensibly unrelated task, subjects rated a stimulus person on 12 trait scales based on a paragraph of behavioural descriptions that was ambiguous, regarding hostility or friendliness. As a result, under the conscious priming condition, asymmetric effects were observed for positive versus negative priming words. Specifically, the assimilat- ive effects on the impression ratings were obtained when the prime words were related with hostility, whereas the effects were not obtained when they were related with friendliness. On the other hand, under the subconscious priming condition, the symmetric assimilation effects were observed for positive versus negative words. In sum, the conscious processing was influenced by the positivity and the negativity of conceptual stimuli, whereas the subconscious processing was not. In other words, it was demonstrated that the level of conscious awareness of priming events should be one of the critical factors for the positive-negative asymmetry of effects to appear. </p></li><li><p>4 T. Ikegami </p><p>In Ikegami and Kawaguchi (1989), it was presumed that the conscious processing of friendly or hostile words might evoke the chronic expectancy that people would emit positive traits and behaviours in usual cases, which is well known as Polyanna hypothesis (see Boucher and Osgood, 1969; Fiske, 1980). Consequently, the impact of positive priming words congruent with the expectancy was mitigated, but the impact of unexpected negative priming words was not. The subconscious processing of prime words hardly evoked such expectancy so that the impact of either positive or negative prime words was not affected. They reasoned further that under the subconscious priming condition, evaluative components of the words could not be processed sufficiently to evoke such expectancy. That is, purely semantic components of the words were processed in this case. Under the conscious priming condition, evaluative implications of the words could be processed well enough to evoke the socio-cultural expectancy on the subjects part. </p><p>Generally speaking, a trait word carries with it some evaluative components as well as descriptive components, and it is very probable that the processing of such components of the word should evoke some kind of evaluative expectancy on the subjects part. According to the reasoning of Ikegami and Kawaguchi (1989) the evoked evaluative expectancy may circumvent or suppress the purely cognitive activa- tion processes which are reflected in the symmetric assimilation effects in the subcons- cious priming condition. It is inferred that such evaluative information processing may occur when the prime words are processed consciously enough at least to identify what they are. </p><p>However, it is still questionable whether the subjects actually processed evaluative implications of the words under the conscious priming condition in Ikegami and Kawaguchi (1989). What the subjects had to do was just to decide whether a stimulus being presented was a jukugo word or not. Such kinds of tasks may not necessarily require subjects to process the evaluative implications of the words. Most of the other previous studies have investigated the effects of prior tasks that were character- ized as perceptual and nonevaluative processing tasks. Further investigations are needed to confirm that the positive-negative asymmetry of priming effects will appear under the condition where subjects apparently enter the evaluative implications of the words and hence they are reminded of the socio-cultural expectancy as described in Ikegami and Kawaguchi (1989). Thus the purpose of the present study is to examine whether the positive-negative asymmetry of priming effects obtained thus far will be observed when prime words receive such evaluative processing as mentioned above. </p><p>OVERVIEW OF THE STUDY </p><p>Two experiments were conducted to investigate the effects of prior tasks upon the impression formation of others. In the impression formation tasks subjects rated a target person on several trait scales based on a paragraph of behavioural descrip- tions that was ambiguous with regard to hostility (Experiment 1) or friendliness </p></li><li><p>Priming efects and impression formation 5 </p><p>(Experiment 2). In the prior priming tasks subjects were asked to complete sentences using hostile (Experiment 1) or friendly words (Experiment 2), in combination with 1. It was assumed that such sentence completion tasks should require of subjects evaluative processing of the conceptual words. Subjects had to comprehend evaluative as well as descriptive aspects of the words in order to convey their own personal beliefs. Subjects were exposed to six hostile (Experiment 1) or six friendly (Experiment 2) words in the priming tasks. </p><p>The hostile and the friendly words were selected among the stimulus words used in the priming tasks in Ikegami and Kawaguchi (1989). The selection was based on the hostility or friendliness ratings of each stimulus word on a 7-point scale which had been gathered in the preparatory study of Ikegami and Kawaguchi (1989). The averaged hostility rating values of the six hostile words used in Experiment 1 was 6.26, while the averaged friendliness rating values of the six friendly words used in Experiment 2 was 6.24. The distance from the midpoint of 4 was 2.26 and 2.24, respectively. Thus it was assured that the extremity (the distance from the midpoint) in hostility or friendliness implied by the prime words was equalized between Experiment 1 and Experiment 2. </p><p>Also, the same paragraphs of behavioural descriptions that had been constructed by Ikegami and Kawaguchi (1989) were used in the subsequent impression tasks. The paragraphs were Japanese versions of Srull and Wyers (1979) material with some changes. Since the symmetric effects had been obtained under the subconscious priming condition in Ikegami and Kawaguchi (1989), it had been proved that these hostile and friendly descriptive paragraphs in themselves were equally susceptible to the priming. The normative rating data, which had been collected in the prepara- tory study of Ikegami and Kawaguchi (1989), revealed that the hostile pa...</p></li></ul>