Cooperation, ethnic salience and generalization of interethnic attitudes

Publication Year


Journal Article

This study compared the effectiveness of three theoretically-based conditions of intergroup cooperation in bringing about generalization of ethnic outgroup attitudes from a cooperation partner to the outgroup as a whole. Twenty-seven pairs of Dutch secondary school pupils were assigned at random to work together in triads to solve two word puzzles. The triads consisted of one Turkish pupil, always a confederate, and two Dutch pupils. The three conditions varied according to whether reference was made to the ethnic background of the confederate in both an introductory conversation and in the conversation-break between puzzles (High–High salience); only in the later break (Low–High); or not at all (Low–Low). Results show no differences between conditions in attitudes towards the partner, which were quite positive. However, attitude change only generalized in the two conditions in which ethnic membership was made salient (Low–High and High–High, which did not differ). These findings are discussed in terms of different models of intergroup contact, and how contact may actually work.

European Journal of Social Psychology
Type of Article
Journal Article
Full text

The following is an excerpt of the intervention methodology. For more information, please see the full text of the article on the publisher's website or through your institution's library.


Subjects. Fifty-four secondary school pupils [...] participated in the study. [...]

Design. Subjects were randomly assigned to one of the three conditions of a one-factor between-subjects design, created by manipulating whether reference was made to the ethnic background of the confederate in both an introductory conversation and in the conversation-break between puzzles [...]; only in the later break [...]; or not at all [...].


Partner Attitudes towards the Turkish and Dutch cooperation partner were measured via five 7-point scales on which subjects had to indicate to what extent the following words applied to their partner: [Text Stimulus A]. [...] In addition, subjects indicated on a 5-point scale their willingness to participate in a possible follow-up study with the two cooperation partners [...].

Attitude towards Turks in general was measured via a semantic differential, consisting of 14 items : [Text Stimulus B]. It was important to prevent subjects from making superficial associations between these measures and their recent experience with the Turkish confederate. We therefore used a visibly different scale from the partner ratings and we linked this measure to an ostensibly different study [...].


Subjects came to the laboratory individually. They were not known to each other before the experiment. There they worked in triads that always consisted of two Dutch students [...] and the same-age, male Turkish confederate. For nearly 2 hours they worked together on a task which consisted of two puzzle tasks. 

The ethnicity of the confederate was either made salient or kept non-salient during the introductory conversation and the conversation-break between the first and second puzzles. Both conversations lasted approximately 10 minutes. Where the Turkish background of the confederate was made an explicit topic, this was done by having the experimenter refer, in a standardized way, to topics such as: country of birth [...]. In the Low-High salience condition no attention to the ethnic background of the confederate was made until the break; then exactly the same topics were referred to as in the High-High condition. No justification was given for these references to the confederate’s Turkish background [...] nor were comparable references made to either of the Dutch subjects’ ethnic background. The topics were raised by a standard series of questions [...]. 

The cooperation task consisted of solving two crossword puzzles. Mutual dependence was created by giving each member of the triad only a third of the information that was needed to solve the puzzle. It was not therefore possible to solve the puzzle without consulting the other two group members. Essentially, this method is an adaption for research purposes of Aronson’s ‘Jigsaw’, a cooperative learning method [...]. Success was guaranteed by choosing a difficulty level whereby all triads could nearly complete the task. The success experience was further enhanced by standardized positive feedback. [...]. 

Six Turkish confederates participated in the study. Their age varied from 14 to 17 years. In six meetings, each lasting 90 minutes, they were trained to react in a standardized way during the task and conversations. They participated equally frequently in the three conditions. 

Manipulation Checks 

[...] Whether subjects perceived the salience of ethnic categories differently from one condition to the other was measured by having subjects rank the following eight situation descriptions according to how striking or salient they seemed: [Text Stimulus C]. [...]


Type of Prejudice/Bias