Readers of this blog will be familiar with research by Bargh et al that showed that priming with stereotypes led to assimilation of those traits by the primed subjects. To recall, when asked to wait, those primed with rudeness interrupted the experimenter more than those primed with patience/nicety primes. Also those primed with elderly stereotype walked more slowly, away from the experimenter after the end of the alleged experiment, than those in neutral condition. These are all very well known findings and more so as they have been included in numerous pop-sci books.
Sceepar et al provide a very sober qualification for the above findings, contrasting how stereotype traits priming (abstract concept priming) do indeed lead to assimilation of stereotype in behavior/self-schema; but exposing to concrete, extreme exemplars belonging to stereotyped category has an opposite and paradoxical effect of leading to social comparison with the exemplar and thus leading to contrast behavior or behavior where the subject tries to distance oneself from the stereotype. Thus, if primed with the stereotype of professors, and thus the trait intelligence, and then subjected to an intelligence/knowledge test, then the subjects would perform better than baseline condition; but if primed with Einstein (a concrete exemplar of professor/intelligence), then one may lead to compare with Einstein, deduce that one is not so great, but indeed stupid or a bozo in comparison, and thus perform badly in the subsequent test based on this self-comparison. This is what they theorized and this is what they found.
In the second study they reused the Bargh paradigm of priming with elderly stereotype and replicated the results; the twist they added was adding a condition in which after priming with the elderly stereotype, an elderly exemplar was presented; this condition led to comparison and thus to contrast behavior whereby the subjects walked faster after the experimental manipulation in this exemplar condition.
Their third study was essentially a study to nail down the mechanism (social comparison) behind the contrast behavior observed. After priming with Einstein and professor in two separate conditions, the subjects were exposed to a lexical task, that was designed to discover if concepts like intelligence , stupidity had been primed and if so , was any of them also bound to the self schema. They found that indeed in the Einstein condition the concept of stupidity was bound to the concept of self, thus only exemplars, like Claudia Schiffer or Einstein led to social comparison, but not abstract notions like supermodels or professors. They end with an advice to Mick Jagger which has to be read in original to be savored.
We close with perhaps one of the more trivial of these implications. This concerns advice we might offer celebrities such as Mick Jagger and other stars known for their predilection for supermodels. If these people share the popular stereotype of supermodels found among our participants, they would be wise to restrict themselves to a single such partner (i.e., an exemplar) on intellectual as well as moral grounds.
I now also post some snippets form the excellent article, freely available on the web, which you should read in its entirety.
Individuals are exemplars in the sense that they are concrete instantiations of a given category (Smith & Zfirate, 1990, 1992). We focus on familiar exemplars that genuinely exemplify their category stereotype rather than deviate from it and thus where impression formation on the level of stereotypes has already occurred (Fiske & Neuberg, 1990). Our argument is thus not that such exemplars suppress stereotypic traits but rather that they make the instantiation of the relevant trait distinct (Stapel & Spears, 1996) by means of an actor-trait link (Stapel, Koomen, & van der Pligt, 1996). This distinction between stereotypes and the traits they activate on the one hand and exemplars that may exemplify these traits on the other is crucial for social perception and resulting automatic behavior. We argue that the activation of an exemplar (such as Einstein) may well lead to different behavioral output than activation of traits (intelligent) primed either directly or by means of a stereotype (professor). As already noted above, priming traits and stereotypes can lead to corresponding behavior in the perceiver. In other words, such priming can result in behavioral assimilation, thereby paralleling findings from the social judgment domain (e.g., Higgins, Rholes, & Jones, 1977; Srull & Wyer, 1979, 1980). Activated traits produce assimilation because they function as an interpretation frame and cause perceptual input to be interpreted in line with this trait construct (Higgins, 1996; Schwarz & Bless, 1992; Stapel et al., 1996). These findings are comparable with those on the effects of perception on behavior, where trait (or stereotype) priming seems to function as a general framework for action. In this case, instead of guiding the interpretation of perceptual input, the trait guides behavioral output. In both cases, highly accessible abstract constructs direct ongoing processes in an assimilative fashion.
A considerable number of social judgment studies demonstrate a different effect, however: Under appropriate conditions, the priming of exemplars has been shown to lead to judgmental contrast (see Herr, 1986; Schwarz & Bless, 1992; Stapel, Koomen, & van der Pligt, 1997). Priming the trait hostility causes a person who is subsequently judged to be seen as more hostile, but priming the exemplar Adolf Hitler causes this person to be seen as less hostile (Herr, 1986). Exemplars exert different effects because they are used as comparison standards (Herr, 1986; Sherif & Hovland, 1961; Stapel et al., 1996, 1997; Stapel & Spears, 1996). Priming Hitler still activates the concept of hostility, thereby eliciting a potential assimilation effect, but the engagement in a comparison induces a contrast effect that overrides the assimilative effect of the activated construct. In the example given above, a person called Donald is compared to Hitler. It is this comparison that evokes the contrast response ( “Well, Donald could not be that hostile” ), whereas the assimilative effects of the activated concept of hostility do not become apparent. However, the exemplar must be sufficiently extreme to override any assimilation evoked by the activated trait. Research by Herr, Manis, Schwarz, Stapel, and their colleagues has repeatedly shown that extreme and specific exemplars exert judgmental contrast effects (see Herr, 1986; Herr, Sherman, & Fazio, 1983; Manis, Nelson, & Shedler, 1988; Schwarz & Bless, 1992; Stapel et al., 1996, 1997). In general, it seems that a comparison (and a resulting contrast effeCt) ~ is rendered more likely if the primed construct constitutes a relevant comparison standard and is sufficiently extreme and concrete (e.g., Hitler). Given that exemplars produce contrasts in judgment by evoking a comparative judgment between exemplar and the target of judgment, the next question is whether such exemplars may evoke in perceivers social comparison with themselves. After all, if primed with Hitler, people may not only make a comparison with Donald–it is also likely that, in a relevant judgment situation, they will compare Hitler with themselves (hopefully with positive results). Recent research indicates that such comparisons with the self are very probable, if not inescapable, especially if they evoke a relevant dimension of comparison. Gilbert, Giesler, and Morris ( 1995 ) showed that social comparison is a spontaneous process: Individuals do not intentionally compare themselves with others but simply do it (see also Festinger, 1954; Wood, 1989), especially when these others are distinctive and extreme (Parducci, 1992; Parducci & Wedell, 1990). Following this analysis, we expect extreme exemplars to evoke a spontaneous comparison with the self.
They conclude with :
The relation between perception and action is not as simple and straightforward as it originally appeared to be: Perception does affect our behavior but not necessarily always in an assimilative manner. Whereas abstract concepts have been shown to lead to behavioral assimilation (Bargh et al., 1996; Chen & Bargh, 1997; Dijksterhuis & van Knippenberg, 1998a, 1998b; Levy, 1996), in this contribution we have demonstrated that concrete exemplars can evoke behavioral contrast. Thinking about Einstein apparently lowers one’s intellectual capacity, and bearing the 89-year-old Dutch Queen Mother in mind puts an extra spring in one’s step. Support for behavioral contrast effects was obtained in two very different domains. In Study 1, intellectual performance was influenced by priming exemplars designating intelligence and nonintelligence. Intellectual performance may be seen as complex behavior involving highly differentiated stages and components.In Study 2, on the other hand, we demonstrated contrast effects on genuinely motor behavior by measuring walking speed. Hence, contrast effects were obtained in two completely different behavioral domains, from higher order mental processes to lower order locomotion. We also used different procedures to arrive at contrast effects. In Study 1, participants were primed with exemplars only (although we also assume that these exemplars evoked their associated stereotypes and traits). In Study 2, we activated stereotypes prior to activating the exemplars, simulating the notion that individual exemplars are perceived against the background of already-activated stereotypes.By demonstrating contrast in this experiment, we showed that behavioral contrast produced by exemplars can counteract previously activated stereotypes.
Study 3 was designed to shed light on the underlying process. On the basis of the literature on social comparison and social judgment, we formulated the hypothesis that concrete exemplars evoke comparisons (e.g., Stapel et al., 1996, 1997). Activating Albert Einstein, in other words, elicits the comparison of this exemplar with the self. This comparison leads to a conception of the self as more stupid or less intelligent. It is this conception, we argue, that leads to stupid behavior. Of course, these ideas rest on the assumption of the comparison being made and the resulting conclusion ( “Boy, am I dumb!” ) being drawn. Study 3 provided support for this assumption. Thinking about the exemplar Einstein indeed led to an association between the self-concept and stupidity as evidenced by facilitated response latencies, whereas thinking about the stereotype of professors did not. Indeed, as might be expected, priming professors activated the concept of intelligence. Priming Einstein also activated intelligence, but the fact that stupidity was also made accessible in combination with the priming of self-reference helps to explain why participants perform less intelligently (contrast) in Study 1.
Overall, these findings are important in being the first to demonstrate evidence of contrast outside the domain of judgment and in the realm of behavior. Although there is growing evidence for automatic behavior, previous research in this domain has always shown that behavior reflects the contents of primed constructs. The notion that automatic behavior can also go against the grain of the prevailing context suggests that automatic behavior may be more varied, and its implications more wide ranging, than previously supposed. Although an exemplar is a potential object of comparison, this does not necessarily mean that people exposed to exemplars always engage in a comparison. So when do people compare? In the typical social judgment paradigm, an exemplar is compared with another social target (e.g., Donald), whereas here we are interested in comparisons between an exemplar and the self. The research of Gilbert et al. (1995) shows that a comparison is easy to make and may even be spontaneous. They proposed that the chances of a comparison increase when the object of comparison is (a) recently encountered, (b) explicitly judged, and (c) extreme (see also Parducci, 1992) and argued that comparisons are hard not to make under these circumstances. The conditions under which exemplars were used in our study seem to conform to these criteria, but we should not assume that just any exemplar will evoke behavioral contrast.
To conclude, the present research integrates theorizing on contrast effects in social judgment with that on automatic social behavior to provide the first direct demonstration of contrast effects in automatic behavior. One of the more reassuring messages to emerge is therefore that automatic behavior is not a one-way street ending up in assimilation. The fact that contrast as well as assimilation can occur may help to explain the diversity of social behavior as much as do the strategies of interpersonal and intergroup differentiation that people consciously employ. The consequences of both assimilative and contrastive automatic behavior are far reaching and should be the subject of future research.
Final departing thoughts: I have always believed that the right brain works using exemplars, while the left brain is abstract; I further believe that Men use abstract reasoning, while females rely more on exemplar based memory: should stereotype research paradigms be defined separately for split-brain and male and female participants; can we observe the hypothesized effects in experiments in the hypothesized directions?
Dijksterhuis, A., Spears, R., Postmes, T., Stapel, D., Koomen, W., Knippenberg, A., & Scheepers, D. (1998). Seeing one thing and doing another: Contrast effects in automatic behavior. Journal of Personality and Social Psychology, 75 (4), 862-871 DOI: 10.1037/0022-35220.127.116.112
Bargh, J., Chen, M., & Burrows, L. (1996). Automaticity of social behavior: Direct effects of trait construct and stereotype activation on action. Journal of Personality and Social Psychology, 71 (2), 230-244 DOI: 10.1037/0022-3518.104.22.168