Priming Of Social Attitudes

Abstract
Previous research has indicated that social behaviour can be automatically activated when primed by traits (Higgins, Rholes & Jones, 1977). The present study investigated whether participants are more like to interrupt an experimenter and unseen confederate when primed by words semantically associated with rudeness, compared with positive and neutral words. It represented a replication of the Bargh, Chen and Burrows experiment (1996) with one alteration; the total number of interruptions rather than time taken to interrupt was measured. 54 undergraduate students aged between 19 and 25 participated in a between-subjects experiment, and were randomly assigned to one of three experimental levels. Participants were asked to complete a Scrambled Sentence Task containing either rude, polite or neutral words. In order to indicate their completion, participants had to interrupt the conversation of the experimenter. As hypothesised, particpants allocated to the rude condition were significantly more likely to interrupt the experimenter than those in the neutral or polite conditions.
Introduction

The extent to which one has intentional control over their own thoughts and behaviours has formed the basis of much research in the field of social psychology. During the 1970’s the distinction between automatic and conscious thought processes emerged, and has became the focus of much attention (Bargh, 1989). Research concentrating on social cognition and attitude formation has documented that many phenomena are unintentional or automatic in nature. Stereotypes, for example, can be automatically elicited merely by the presense of physical features commonly associated with the stereotyped group. Similarly, the presense of an ‘attitude object’ can automatically elicit an attitude, which in turn exerts influence on behaviour (Bargh, Chen & Burrows, 1996). Both are examples of priming, which refers to the effect observed when exposure to a certain stimulus influences responses to a second stimulus. In social psychology, priming can be understood in terms of the tendency for recent information to influence subsequent thoughts. An early example of this came from Higgins, Rholes and Jones (1977). In this research, particpants read a passage involving a man attempting certain ambitious physical feats. Prior to this, particpants were told they were particpating in a memory task, and were given a list of attributes to read. Half of the particpants were ‘positvely primed’ and given words such as ‘adventurous’ and ‘brave’. The other half were ‘negatively primed’ and given words such as ‘foolish’ and ‘reckless’. After reading the passage, particpants were asked to give their impressions of the man in the story. Those who had been positively primed formed more positive impressions of the man in the passage than those who had been negatively primed.
Forgas and Bower (1987) looked at the effect of priming on how people judge information about others. Participants were assigned to one of two conditions. In the first condition, participants were given information that primed a happy mood. In the second, particpants were given information to prime a sad mood. Participants in both conditions were then given identical person descriptions to read. They found that those who had been primed to experience a happy mood formed more positive impressions of the people in the person descriptions than those primed to experience a sad mood.
An experiment by Bargh, Chen and Burrows (1996; exp2) demonstrated that priming influences behaviour, investigating the behavioural consequence of automatic stereotype activation based on the premise that a typical stereotypical view of the elderly concerns slowness. Participants were presented with scrambled sentences containing words that related to elderly stereotypes, or sentences containing neutral words. Importantly, the authors ensured that none of the words in the ‘elderly’ condition were directly related to slowness. This ensured that any observed behaviour change was attributable to the stereotype of elderly being activated (and the associated assumptions of slowness) rather than focusing attention on the single trait of being slow. Following this, particpants were asked to leave the room, and were timed walking down a hallway to return to the waiting area, to test whether priming participants with the concept of elderly would automatically and unconsciously change their behaviour to emulate the elderly. As predicted, participants who were primed with the stereotypical information took longer to walk down the hallway than those who received the neutral information.
Carver, Ganellen, Froming and Chambers (1983) demonstrated the priming effect of aggression on particpant’s subsequent behaviour. Participants were divided into two conditions and given scrambled sentences containing either aggressive or neutral concepts, diguised as part of a seemingly unrelated study. They were then asked to participate in an experiment of human learning where particpants were able to punish another participant (actually a confederate) by administering electric shocks for incorrect responses. Those in the aggressive condition administered stronger shocks than those in the control condition.
The studies discussed thus far have indicated that priming can influence both perceptions and behaviour. From this premise, Bargh et al. (1996, exp 1) investigated whether this effect overrode the typical processes one uses in everyday life, such as social judgement. In this experiment, 34 participants were informed that they were taking part in a test of language ability, and presented with a scrambled sentence test. They were randomly assigned to one of three conditions priming conditions, rude, polite and neutral. Participants were asked to complete the task individually, then notify the experimenter in another room. When the participant entered the second room, they found the experimenter in conversation with a second participant (a confederate). The critical outcome measure of the study was the length of time the participant took to interupt the conversation between the experimenter and confederate. They found that those in the rude condition did interupt significantly faster than those in the polite and neutral conditions. However, the results suffered from significant ceiling effects; 21 of the 34 participants did not interupt at all.
The present study was a replication of Bargh et al.’s(1996) experiment and borrows heavily from their methodology, but with one critical difference. In an attempt to address the methodological issues caused by the strong ceiling effects observed by Bargh et al., the current experiment measured the total number of people to interupt in each condition, rather than the time taken to interupt.
Based on the findings of Bargh et al., the experimental hypothesis predicted that significantly more particpants in the ‘rude’ condition would interupt the experimenter than those in the ‘polite’ and neutral conditions. Furthermore, it was predicted that there would be no significant difference in the number of participants in the polite and neutral conditons who interupted the experimenter.
Method
Participants
A total of 54 undergraduate psychology students, 34 females and 19 males, aged between 19-25 (mean age 20.3), volunteered to participate in the experiment.
Materials
Each of the participants was presented with “Scrambled Sentence Test” which was presented as a test of language ability. Comprising 30 items, participants had to use listed words to form a grammatically correct four-word sentence as quickly as possible. There were three versions of the scrambled-sentence test: for the rude and polite conditions, 15 of the 30 items contained words that were associated with the trait in question. In the neutral condition, these 15 were replaced with neutral words. The remaining 15 items were idetical across the three conditions.
Design
?he design of the experiment was between subjects, and had three experimental conditions. Participants were randomly assigned to each condition. The independent variable was the condition that the participant was assigned to and had three levels; rude, polite and neutral. The dependent variable was the number of participants who interupted the experimenter.
Procedure
Participants took part in the experiment one at a time. They were informed that they were to participate in a language ability study, and their consent to participate was obtained. Each participant received an envelope that contained 30 scrambled sentences, and were told the task was concerned with grammatical construction. They were asked to form a grammatically correct four-word sentence from a list of five-word scrambled sentences. Particpants were then given one of three versions of the test (rude, polite or neutral) of the scrambled-sentence test, and asked to complete it as quickly as possible. Upon completion, participants were asked to find the experimenter in a second room and notify them of their completion, in order to move on to the next experimental task. Participants was randomly assigned to each condition, to which the experimenter was blind. When the participant entered the second room, the experimenter appeared to be engaged with another unseen particpant (actually a confederate). The experimenter and confederate continued their discussion until interupted by the participant. The confederate noted which of the participants elected to interrupt.
Results
Table 1: Total number of participants who did/did not interrupt in each condition (polite, neutral, rude).
Word Type
Polite (n)Neutral (n)Rude (n)
Did not interrupt15108
Interrupted3612
The dependent variable of the experiment was the total number of participants that interrupted the experimenter in each word type condition, and as can be seen from Table 1, participants in the rude condition interrupted most frequently (n = 12). Participants in the polite condition interrupted the least (n = 3). The propensity to interrupt increased across the levels of the independent variable; a higher number of participants in the neutral condition (n = 6) interrupted than in the polite group. A higher number again interrupted in the rude group.
The results of this experiment were statistically analysed used a Chi-square goodness-of-fit test. This test was selected as the data was categorical, numerical and discrete. It produced a goodness-of-fit between the observed and expected values. If priming had no effect on interrupting behaviour, distribution across the conditions would be equal. The Chi-square test demonstrated whether the observed frequencies differed significantly from the expected frequencies. The result of the Chi-square was significant; participants in the rude condition were significantly more likely to interrupt the experimenter than participants in the neutral or positive condition, x2 (2, N = 54) = 7.50, p< 0.5.
Discussion
This study examined whether priming traits (rudeness and politeness) could have an impact on social behaviour. It was hypothesized that participants primed with rude traits would be more likely to interrupt an experimenter ostensibly engaged in a conversation with an unseen ‘participant’ than those who were primed with positive or neutral traits, and this hypothesis was fully supported by the results of the experiment.
The study confirmed that people, when influenced by a rude prime condition, would demonstrate a negative impact on social behaviour, whilst the social behaviour of those who were influenced by a polite or neutral prime condition were unaffected.
This finding supported the results of Bargh et al. (1996) despite the change to the dependent variable. This provides compelling evidence for their hypothesis regarding the automatic activation of social behaviours in the face of certain environmental features (in both cases, the rude primes). Not only does the replication confirm the original study findings, but the change in the dependent variable in the study under discussion improves on the methodological design flaw of the original. Interestingly, the present study also supported the original study finding of no significant difference in interrupting behaviour between the polite and neutral groups. This indicates that it is rude behaviour that is most likely to be automatically activated. Priming a participant with polite trait information does not appear to increase an individual’s politeness, or a significant difference between the neutral and polite levels would be observed. This is somewhat at odds with previous research which indicated a bias towards increased postivity via priming (Higgins et al., 1977; Forgas & Bower, 1987). However, these studies looked at the effect of priming on judgement, and not on behaviour. If the result of the present study is compared with Carver et al. (1983) study of negative priming and electric shock administration, both show an increased bias towards the propensity to prime negative behaviour.
Although the change in the dependent variable from time to interrupt to presence of interrupting behaviour resolved the issue of ceiling effects described by Bargh et al. (1996), whilst producing a similar result, some detail was lost. Moving from a numerical scale (time) to a categorical measurement meant that the choice of statistical analysis available was limited. A future replication might involve measuring the total number of participants that interrupted, and the time taken among those that did.
A second factor that should be considered is the possibility that those in the rude condition were simply a ruder group of participants and were more likely to interrupt despite the priming effect. To control for this possibility, a future replication could incorporate a personality questionnaire to control for propensity towards rude behaviour.
The current study was a replication of Baugh et al.’s (1996) experiment 2. To resolve the original study’s problem of ceiling effects, a change was made to the dependent variable; instead of measuring the time taken to interrupt, the total number of those interrupting in each condition was measured. The results supported the finding of the original study, presentation of the rudeness trait primes subsequent rude behaviour; therefore social behaviour can be automatically triggered without conscious thought.
References
Bargh, J. A. (1989). Conditional automaticity: Varieties of automatic influence in social perception and cognition. In J. S. Uleman, & J. A. Bargh (Eds.), Unintended Thought (pp. 3-51). New York: Guilford Press.
Bargh, J.A., 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, 230-244.
Carver, C., Ganellen, R., Froming, W., & Chambers, W. (1983). Modelling: an analysis in terms of c ategory accessibility. Journal of Experimental Social Psychology, 19, 403–421.
Forgas, J. P. & Moylan, S. J. (1987). After the movies: The effects of transient mood states on social judgments. Personality and Social Psychology Bulletin, 13, 478-489.
Higgins, E. T., Rholes, W. S., & Jones, C. R. (1977). Category accessibility and impression formation. Journal of Experimental Social Psychology,13, 141–154.