A good place to start with any book is at the beginning. In this chapter, I trace the history of the teacher expectation field. I present the findings from the first study in the teacher expectation field and examine how that ground-breaking work influenced the future directions that research took. Also, I present a teacher expectation model that encapsulates the steps in the expectation process, from the teacher forming her/his expectations, through to the student interpreting these and acting accordingly.
In 1948, Merton, a sociologist, proposed the self-fulfilling prophecy. The underlying premise was that, when we believe something to be true, we act in particular ways that can cause our beliefs to become true. He provided the example of a bank, the Last National Bank, which was doing well financially â until Black Wednesday. On that day, when the manager of the bank arrived at work, he noticed that business was unusually brisk and mused that he hoped folks at the local steel plant and mattress factory had not been laid off. He later found out that someone had begun a rumour that the bank was in financial trouble and about to collapse. The rumour spread. Anxious depositors rushed to the bank to withdraw all their savings. Because the bank could not sustain the large volume of withdrawals, it became insolvent. Hence, âthe self-fulfilling prophecy is, in the beginning, a false definition of the situation, evoking a new behaviour which makes the originally false conception become trueâ (Merton 1948: 195). Merton went on to describe how much of the prejudice then evident in American society, and the outcomes of that prejudice, could be explained by the self-fulfilling prophecy effect. For example, he explained how, in Mississippi at the time, the state spent five times as much on educating white students as on African American students, because African Americans were considered inferior intellectually. Of course, this led to African Americans achieving less in school â the self-fulfilling prophecy effect. Interestingly, expectation research, or research into the self-fulfilling prophecy effect, since Mertonâs original proposal, has mostly been carried out within the disciplines of social psychology and educational psychology, rather than sociology, the discipline within which it was originally conceived.
Early beginnings of expectancy research
During the 1950s, Robert Rosenthal was a young psychologist who became interested in how experimenters could unconsciously influence their subjects during experiments to obtain the results they were expecting. In some very early experiments (Rosenthal 1963), thirty experimenters were told to ask almost 400 participants to look at photos of a set of people and rate them, from extreme failure to extreme success, on a 10-point scale. Half the experimenters were told their participants would be likely to rate the people in the photos positively, so above five on a 10-point scale, and the other experimenters were told that their participants were likely to rate those in their photos below five. The experimenters read exactly the same instructions to all participants. Rosenthal found that, indeed, the mean was above five for those who expected their subjects to rate the photos positively and below five for the other group of experimenters, and the differences in the ratings of the two groups were statistically significant.
Rosenthal conducted similar experiments in laboratories in which experimenters were training rats to go through mazes (Rosenthal and Fode 1963). He noticed that, if laboratory assistants were told that the rats were smart, the rats learned to go through the mazes more quickly than if he told the laboratory assistants that the rats were dull. Actually, there was no difference in the rats; they had been randomly assigned to the laboratory assistants. Rosenthal found that he got the same results, even when the rats were put into Skinner boxes, where the rats needed to learn to flick a switch if they were to obtain food. In the first set of experiments, it was thought that perhaps the laboratory assistants had handled the rats differently, depending on whether they believed their rat was smart or dull, but, in the Skinner box experiments, the laboratory assistants did not touch the rats, and yet the same results were obtained. Rosenthal proposed that the laboratory assistants must have interacted with the rats differentially but subconsciously, depending on whether they believed that the rats were smart or dull, and something about the way in which the laboratory assistants had interacted with the rats had caused the rats to learn more quickly or slowly. Thus, the laboratory assistantsâ expectations had been fulfilled. Rosenthal concluded, therefore, that an experimenterâs expectations could determine, to a large and significant degree, the types of result that could be obtained in experimental studies. He suggested that this led to researchers having their hypotheses confirmed.
Rosenthal (1963) published his results in an American science journal. In pondering the implications of his studies, Rosenthal queried whether, if a senior teacher told a beginning teacher that a particular student was a slow learner, this prompt would subsequently lead to that expectation being fulfilled. It happened that Lenore Jacobson, a school principal, read the article and offered her school as a potential site for the first ever teacher expectation experiment.
The first teacher expectation experiment
The primary aim of the first experiment was to see whether teachersâ expectations could be manipulated when teachers were given false information about their students. In their book, Rosenthal and Jacobson (1968) began by outlining how the expectations of teachers could lead to potentially differential teaching quality for African American and white students, and for students from poor socioeconomic backgrounds, as opposed to those from middle-class families, even when the students had similar initial achievement. They proposed that teacher expectations could be one explanation for the widening achievement gap between African American and white students as they progressed through school, because these different students were exposed to contrasting learning experiences and teacher input. This discussion provided the background for the Pygmalion experiment, as it became known.
Oak School, where the experiment took place, was in a low socioeconomic area, and the student achievement was generally low. Before the experiment began, Rosenthal and Jacobson used a little-known, mostly non-verbal, IQ test (Test of General Ability, or TOGA) with all the students in the school. They dressed it up as being from Harvard, where Rosenthal was based, and called it the Harvard Test of Inflected Acquisition. The teachers were told that it was a new test that was able to predict which students would suddenly bloom during that year. The test was administered to the students four times: first, as a pretest, to provide baseline data; second, eight months after the experiment began; third, one year after the first administration; and fourth, two years after the initial baseline data were collected.
Following the pre-test, the teachers were presented with class lists in which from one to nine students had been nominated as students whose trajectory of academic gains would suddenly steepen that year â the âbloomersâ. Over the entire school, 20 per cent of students had, in fact, been randomly selected to be the bloomers. One year later, when the students were tested, overall there were substantial gains in intellectual growth for the students who had been identified as bloomers when compared with their peers. Similar results were found for reading, where, overall, the identified students made more progress than the control-group students. Greater intellectual and reading gains were noted for the students in Grades 1 and 2, however, than for the students in Grades 3â6, and girls in the experimental group gained more in intellectual growth than boys. Students in the school were in ability tracks, and the bloomers in the middle track made much greater reading progress over the year of the experiment than did those in the high track.
By way of explanation for the overall results from the first year of greater gains for the younger students, Rosenthal and Jacobson suggested that younger students might be more susceptible to the effects of teachersâ expectations than older students. They also argued that teachers might not have been so readily influenced by the false information in relation to the older students, whom they might have already known, or at least that teachers might have been aware of their reputations. A further possibility was that the younger students might have been more responsive to the ways in which teachers communicated their expectations to students. Although they did not measure how the teachers interacted with the students, Rosenthal and Jacobson proposed that the teachers must have interacted differently with those for whom their expectations had been raised, causing improved intellectual performance in the students â the self-fulfilling prophecy.
Towards the end of the first year of the experiment, the teachers were asked to rate all their studentsâ classroom behaviour. The ratings for the control and experimental students were then compared. The behaviours fell into three clusters. The first cluster, labelled intellectual curiosity, comprised teacher ratings for studentsâ intellectual curiosity, the probability that the students would be successful in the future, and how interesting the students were. The second cluster, adjustment, was made up of teachersâ ratings for the adjustment of students, and for how happy, appealing, affectionate, and non-hostile the students were. The final group of variables, need for approval, was made up of only one item, which did not relate to the other variables. Overall, the experimental group was rated by their teachers as showing more intellectual curiosity than their control-group peers. The effect was particularly evident among the younger students, but there was also a statistically significant difference in the teachersâ ratings of intellectual curiosity between the two groups at Grade 6. In terms of ratings for adjustment, there was a trend for those labelled the bloomers to be considered happier than the control group, but there were no statistically significant differences found for any of the other variables. With regard to need for approval, the Grade 1 students in the experimental group were viewed by their teachers as being more self-directed and less anxious to obtain the approval of others than were those in the control group, but this was the only statistically significant difference found between the bloomers and non-bloomers. Hence, there was no overall difference in need for approval between the experimental and control groups.
As mentioned above, the children were tested eight months into the experiment as a measure of how quickly the experimental effect came into being. It was found that, although there was some evidence of a gap in intellectual performance appearing at that time, the differences between the experimental and control groups were not statistically significant, and so the expectancy advantages that were becoming evident were described as a trend at that point. However, in both reading and social studies, experimental students were graded higher by their teachers than their control-group peers, after just one semester. Further, the bloomers in fifth and sixth grade (the only students to sit the achievement tests) gained approximately 5 percentile points more than the control group in the vocabulary test and more than 10 percentile points for work study skills. Moreover, the advantage that these children received on the achievement tests was still evident one year later.
The final round of testing occurred two years after the initial pre-test. This final testing was designed to determine whether any of the advantages experienced by the bloomers as a result of the experiment still existed two years later. The new teachers were not told which students were part of the experimental group, because the experimenters wanted to see if any increases in intellectual growth from the first year were dependent on the experimental children being with a teacher whose expectations had been positively influenced. The follow-up test showed that the expectancy advantage for the original Grades 1 and 2 bloomers had disappeared after two years, although the experimental students in Grade 5 profited such that their gain in relation to the control group was statistically significant. Rosenthal and Jacobson posited that the younger children, who appeared to be more easily influenced by their teachers, might have needed the continual influence of high expectations for their advantage to be maintained. The middle-track students also benefitted from enhanced teacher expectations, and the experimental girls continued to show an advantage in reasoning two years later when compared with the control group, whereas, for the boys, the advantage for the experimental group was in their verbal score.
Overall in the Pygmalion study, Rosenthal and Jacobson suggested that a difference in the interaction quality of teachers with students between the experimental and control groups might have led to the gains of the bloomers. Rosenthal noted that, in his earlier work, experimenters who had been primed to believe that their rats were smart were observed to engage more pleasantly and were more enthusiastic and friendly towards them than were those who believed that their rats were dull. Hence, he proposed that it was not outrageous to consider that teachers might have been more friendly, interested, warm, expressive, and encouraging towards the bloomers rather than the non-bloomers. Rosenthal also suggested that teachers might have provided the bloomers with more, and more positive, feedback than they did with the control group, and that they may have communicated their expectations unwittingly, perhaps through non-verbal channels. However, Rosenthal did caution that, because teacher interactions, whether verbal or non-verbal, had not been measured in the experiment, any suggestions in relation to how high and low expectations might be communicated to students were purely speculative.
Reactions to the Pygmalion study
The Pygmalion experiment (Rosenthal and Jacobson 1968) resulted in headlines in some prominent United States newspapers when it was first published, and it is still favourably included in many education textbooks. Moreover, the study was cited in several American court cases, resulting in the elimination of tracking in one state, a ban on the use of intelligence tests to identify students for special education classes in another state, and the initiation of desegregation in one southern city (Spitz 1999).
Many academics were enthusiastic about the study causally implicating teacher expectations in the racial, social-class and gender injustices and inequalities of society (Jussim et al. 1996). Others advocated utilizing teacher expectancy to raise intelligence and tackle poor educational performance (Spitz 1999). Some in the general public surmised that much of the poor achievement of some students could be attributed to low teacher expectations! However, many of these claimants misinterpreted or exaggerated the effects that Rosenthal and Jacobson had found. For...