Steele and Aronson first detected the phenomenon of stereotype threat: Black and white Stanford University undergraduates, matched for SAT scores, were given part of the Graduate Record Exam (GRE) under two different conditions. One group was told that the test would be a reliable measure of their intelligence, whereas it was presented simply as a set of problems recently devised by the experimenters, who were interested in the participants’ opinion on them, to the other one. There was no difference in the obtained scores of the white students in the two different conditions, but the black ones obtained significantly worse scores when the test was presented as a measure of their intelligence. Steele and Aronson explained this finding by suggesting that black students are always aware of the stereotype that they are less intelligent than white ones, and can thus not concentrate well on the test. This hypothesis of a general underestimate of black students’ true ability due to the stereotype threat has been confirmed in a number of other studies as well, e.g. by Brown and Day, who administered Raven’s Matrices Tasks. In this experiment we are testing whether Cambridge undergraduates might suffer from a similar stereotype threat of exceptional intelligence, whereby they are expected to perform highly irrespective of personal circumstances and situational variables such as anxiety. This was investigated by dividing the students into two groups - unbeknownst to them - setting them two different kinds of Raven’s items - one easy and one very hard one - to complete immediately prior to a speeded test (the Wechsler Digit Symbol) and a non-speeded test (vocabulary). As other studies have shown, anxiety might have a particular impact on speeded tests.1 In general, it is also possible that ` students profit from the phenomenon of stereotype boost, whereby the increased arousal induced by the difficult Raven’s items among high IQ individual may actually enhance performance. This is also known as Pygmalion effect.
The subjects were 60 Cambridge University undergrad- uates, of which 29 were assigned easy Raven’s matrices and 31 had to complete hard Raven’s items.
Two kinds of Raven’s test sheets were used (either both of them containing easy or hard Raven’s items) and the Wechsler Digit Symbol as well as a non-speeded vocabulary test printed on two separate sheets.
The experimenter gave out easy Raven’s items to people in the test room that were sitting in easily visible positions and hard ones to the ones who could see them, in order to ensure maximum possible anxiety conditions in this experimental design. The experi- ment consisted of four sessions: Raven’s test sheets (5min), Wechsler Digit Symbol test (90s), Raven’s test sheets (5min), vocabulary test (untimed).
We have tested two H 0 hypotheses, the first one of which is H 0 ,a that there is no statistically significant difference in the performance on the Wechsler Digit Symbol task between the group that was administered the easy Raven’s items and the one that was assigned the hard items. H 0 ,b states that there is no statistically significant difference between the performance of the former and latter group on the vocabulary test. Contrary to our expectations regarding the Cambridge University student stereotype threat or boost, we retained both H 0 ,a as well as H 0 ,b by conducting a two-sided t-test in both cases. In the comparison of the performances of the two different groups on the Wechsler Digit Symbol task, we found t (58)0 . 05 = 0 . 24 and thus we retain H 0 ,a.2 Similarly, comparing the performances on the vocabulary tasks led to retention of H 0 ,b.3
The fact that the expected results were not obtained does not prove that there is no such thing as the Cambridge Student stereotype threat or boost, since there are some weaknesses in the experimental design. First of all, one of the major issues is the assumption that students who were set the easy Raven’s items would gaze away from their test sheets and wait in silence for the remaining minutes of assessment, and thus cause state anxiety in other people with harder Raven’s items. Many highly-achieving people use a different strategy, e.g. they use the remaining time in order to go through all the questions they had already solved trying to correct for possible mistakes. Anpther reason might be that they have a drive for constant intellectual stimulation and consider the prospect of waiting without doing anything productive as boring. In order to avoid this behavioural phenomenon, we could tell participants that there is a Prize Draw for the first 10 people that finished with their test sheet and answered every question correctly. Secondly, we
4 LAURA IMPERATORI (MUR)
cannot expect all students completing a test under time pressure to be influenced by their neighbours behaviour in such a distinct way. Most people will have learnt to focus on their exam sheet by the point in their educational career when we are assessing them and will not get as distracted by their neighbours’ behaviour as we are assuming here. Thirdly, Cambridge Students might suffer from the stereotype threat in a more subtle way, by having greater state anxiety because they know they are competing against the world’s best students, so the actual behaviour of others does not influence this effect of intrinsic knowledge.
In conclusion, this experiment has demonstrated that there is no statistically significant difference in the performance on the Wechsler Digit Symbol task as well as on the vocabulary test between the group that was administered the easy Raven’s items and the one that was assigned the hard items. There are three possible conclusions we could draw from this:
(1) There is no such thing as the Cambridge University student stereotype threat or boost.
(2) Cambridge Students do not necessarily finish with their test sheets straight away and might spend remaining time of any kind of assessment to check for possible errors.
(3) Cambridge Students might not get distracted as much by their neighbours’ behaviour, but they might have a greater state anxiety for other reasons. We would need to conduct further experiments to explore each one of them in more detail.
1. Ryan P Brown and Eric Anthony Day, The difference isn ’ t black and white: Stereotype threat and the race gap on raven ’ s advanced progressive matrices.de, Journal of Applied Psychology 91 (2006), no. 4, 979.
2. Robert Rosenthal and Lenore Jacobson, Pygmalion in the classroom: Teacher expectation and pupils ’ intellectual development., Holt, Rinehart & Winston, 1968.
3. Magaretj Shih, Todd L Pittinsky, and Geoffrey C Ho, 9 stereotype boost, Stereo- type Threat: Theory, Process, and Application (2012), 141.
4. Clau M Steele and Joshua Aronson, Stereotype threat and the intellectual test performance of african americans., Journal of personality and social psychology 69 (1995), no. 5, 797.
1 Dr Kate Plaisted-Grant was asked for research papers undermining this hypothesis.
2 Since the F-test led to F (30 , 28) = 2 . 14, a two-sample, unpaired t test, for unequal sample variances, was performed.
3 We were able to use the more powerful two-sample, unpaired t test, for equal sample variances, since the F-test led to F (30 , 28) = 1 . 30, which is below the tabulated value of 2 . 11. We found t (58)0 . 05 = 0 . 5, well below the tabulated value of 2 . 0 and thus we retain H 0 ,b as well.