Notes on:  Murdock-Perrieral, L & Sedlacek, Q (2018)  Questioning Pygmalion in the twenty-first century: the formation, transmission, and attributional influence of teacher expectancies.  Soc Psychol Educ.  21: 691–707.

[Very good review of the erly work on teacher expectations and the problems arising in showing stable effects. Interesting asides on 'naive realism' and suggestions for future research on empathy to encourage 'situational' understandings. Not specifically focused on race,but mentions it]

There seem to be two issues — experimental and intervention, with emphasis on recursive processes of interaction. There may also be 'possible mediators including perceptual biases, confirmation biases, stereotyping, and attributional biases'. Empathy could also be important. More research is needed. (691)

Teachers clearly form and act upon expectations which may be positive or differentially positive. They do not always translate into different forms of education or interaction patterns. The same issues affect other professionals, for example in the amount of time they devote to clients. Teachers may attempt to achieve fairness, although that does not always carry the same meaning. Teacher expectations may be influential contributors to student outcomes, and should be empirically investigated.  This is particularly so when more and more information about students is collected in the form of individual plans, report cards, stories about students. Informal channels may be even more important. Information may be gathered long before actually meeting students, and might be 'an invaluable resource', but there may also be Pygmalion effects or self-fulfilling prophecies, well established in the classic social psychological literature going back to the late 1960s [Rist 1970 is particularly crucial]

It is clear that lots of teacher interaction is crucial for children's education, including practices of individualisation, but at the same time negative treatment of low expectancy students is inappropriate. Teachers might not be very good at concealing negative expectations, however, and instead, it might be better to 'leverage social psychological processes that interrupt potentially detrimental cycles' (693), focusing on the link between expectations and 'attributional processes' — 'attribution for, and responses to, student behaviour and academic performance'.

Expectancy effects have been taken for granted, and much research just debates issues such as how accurate they are, but the whole literature needs re-examining in order to generate '"wise interventions"' in the recursive processes and repercussions which interactions can generate. It is the longitudinal transmission of teacher expectancies that need to be studied. Empathy might actually be important here.

They want to use the term 'expectancy effects' to describe phenomena where teacher beliefs about students influence student outcomes, emphasising expectancies rather than specific contents of thinking, self-fulfilling prophecies rather than outcomes that actually affect behaviour in real-world interactions. Experimental evidence seems to suggest that expectancies can change interactions, especially if there is 'an explicit power differential'. The classic here is Rosenthal and Jacobson with their erroneous student test scores, and the controversy and failures to replicate the study generated. It seems that where strong beliefs that informed judgments already conflicted with incompatible expectancies supplied by the experimenters, the latter were rejected, showing that test scores alone were probably not sufficient to generate expectancy effects. Other work suggested that effects were already occurring in classroom settings — that older siblings had an effect on expectations, for example. Many other studies of teacher expectancies took place, and a meta-analysis in 1984 suggested that expectancy effects 'tended only to appear if induced before a student and teacher had known each other for more than two weeks' (695). Other research suggested modest self-fulfilling prophecy, and other work attributed self-fulfilling prophecies to the '"accuracy" of teacher expectations.

This [accuracy claim] in turn was criticised as being based largely on research showing correlation between teacher expectations and prior achievement. If teachers received extensive information about reading and writing abilities as test results, their expectancies were only slightly better aligned with performance than if teachers were provided with less achievement data — more information did not lead to more accurate expectancies. There were methodological critiques especially with long-term studies which underestimated '"dynamic social interactive processes that create disparities"' in interactions between teachers and students  ( 696). Even Rist (1970) might be brought to agree because his long-term ethnographic work showed that past expectancies can influence prior achievement data itself. However, 'regardless of whether teacher expectations are "accurate" or "inaccurate", researchers agree that strong expectancy effects may occur under certain circumstances', and the point is to identify factors that might produce them.

One issue is whether expectation effects are compounded over time, whether they grow more extreme or dissipate [neither, it seems, but they might just endure or persist, maybe therefore remaining damaging] There is some evidence of this persistence, but in some subjects only — mathematics but not reading, for example. Other researchers have disagreed. National contexts also seem different. Overall, 'teacher expectancies merit further study'.

We now turn to the issue of how expectancies are enacted. One early theory suggests four factors:

1.Positive or negative classroom climate  [eye contact and other means generated by the teacher] or lack of feedback provided to students
3.time in curriculum quality afforded to each student
4.performance and response opportunities afforded to each student [opportunities to answer questions and patients while waiting for answers]

[Reminds me of all the problems of research with studies of open questions and all that in  ORACLE]. A subsequent study  suggested that climate creation and input patterns seemed the most consequential, and further work [we are up to 2010 now] suggested that low expectancy students got fewer or less valuable learning experiences.

Classroom climate might operate through 'subtle, unconscious behaviour' — 'smiling, nodding and eye contact' which seems to happen more often with students for whom teachers have higher expectations, while those with low expectations appear to receive more emotional support, despite reports from students indicating 'the opposite' [studies referenced throughout]. Teachers believed that students would not be aware of any special emotional bonds to favourites, but empirical findings suggested they did — less than 20% of students accurately predicted who students would identify as the teacher's pet, a large measure of unawareness that students perceived differential treatment: teachers pets were negative but pervasive. The whole  area is  as a 'manifestation of naïve realism… A social psychological phenomenon by which individuals perceive themselves as unbiased observers of objective social realities' (697) [an awful lot of this about with pupil perceptions about racial prejudice as well]. Another study has seen that when students are struggling to answer questions, teachers give more clues to those they think of as bright.

There is some evidence of racial biases among New Zealand teachers and Maori students. An American study assigned randomly selected students as gifted and 'significantly different patterns of praise, criticism and attention emerged'. A meta-analysis 'found widespread evidence of racially biased expectations in real-world classrooms'. Even sceptics 'acknowledge that teacher expectancies may have a stronger impact on ethnic minority students', and there are findings about particular impacts on low income students. However a study of German teachers expectations for ethnically Turkish students 'were relatively accurate', but they over-expected for ethnically German and high income students .

Turning to curriculum quality and response opportunities, Rist looked at teacher judgements made in the first eight days of kindergarten, where there was no objective diagnostic information, and noted that some students were already being classified as 'having "no idea what was going on", largely on the basis of dialect, socio-economic status, family structure, and even hygiene and clothing', and those who are seated at the back of the classroom and receive less time and attention. A paper trail of low achievement was created for them, producing a '"caste like"' system. Another study looked at how teachers assign students to different sorts of reading groups and showed differences in preference to interact with students in the higher level ones: enjoyment was also higher as was time spent, resulting in 'expectancy reinforcement'. In another approach students were randomly reassigned from lower ability groups into higher ability ones and they achieved higher scores on standardised tests and better recommendations from teachers — 'these findings suggest the impact expectancies may have when reified via ability groupings' (699). Assessment and grading practices are clearly another mechanism, and in one experiment, essays were attributed to excellent or weak students and that 'significantly impacted the grades, although it might not follow that lower levels of learning like this might happen in the long term. It may not be the case with psychological reports for kindergarten children, another study shows, but seems to be more likely with assessments 'with more subjective criteria (e.g. essays)'.

Of course, 'the evidence that expectancy  effects can exist in some classrooms and contexts does not imply that they do exist in all contexts, nor that such effects are uniform in their strength or in the mechanisms of their enactment', and it may be the case that "such effects are minimal for most teachers because their expectations are generally accurate and opened to corrective feedback"'. Of course, expectations will also vary from teacher to teacher and context to context. Some teachers will be better at controlling their behaviour to minimise the influence of their expectations — one study suggests that personality variables are important including '"dogmatism, authoritarianism, field dependency, bias and cognitive rigidity"' (700)

The processes underlying the formation of expectancies is still little researched, possibly because they seem to be simple hypotheses based on perceptual bias. However, it seems unlikely that perceptual bias alone could account for these expectancy effects, and confirmation bias might also be involved, 'belief confirming information', where teachers might expect help to reward good students, but be a waste of time for poor students. This might manifest itself as 'differences in teachers' questioning techniques' or clue-giving patterns, perhaps 'selective noticing and remembering', including forgetting the failures of good students and remembering those of poorer ones. There might also be stereotyping, characterised by 'selective inferring' (701) — student actions might be interpreted and responded to differently according to expectations held. The example turns on why students fall asleep in the class -- eg  whether it's because they been working hard or because they don't care about school. This might explain the role of racial biases, where African-American students' activity leads to the inference of patterns of misbehaviour more often 'in an experimental context'.

Teachers also 'sometimes experience naïve realism… A tendency to think that one's own views and actions are objective and unbiased, whereas others actions are "biased by ideology, self-interest and irrationality"'. This can include 'the fundamental attribution error'— over estimating dispositions and underestimating situational factors when attributing factors to other people's behaviour. The result is to see some behaviour as situational and malleable, and others as 'dispositional and fixed'. With limited information, teachers might be more likely to attribute student behaviour to their dispositions, and when they draw inferences this might augment their expectancies [clearly at work with racial expectations like some of those reported in the YMCA study]

They want to study that possibility 'by manipulating expectancy formation and measuring its influence on attribution'. The hypothesis is that a strong positive expectancy for a student could lead to dispositional attributions for positive behaviour and situational attributions for negative behaviour, and the reverse. One study found evidence for this in a vignette study of graduate students, but they want to do more expansive research.

They also want to examine the effects of teacher empathy. Several studies already suggest it is significant: those who experience empathy 'are likely to make more situational and fewer dispositional attributions' [according to one rather flimsy experiment, where subjects were asked to place themselves in the shoes of one of the subjects in the experiment (702)].

To examine what happens long-term, we need to see how information that produces expectancies is actually transmitted. This might involve communications with other teachers, parents or administrators, cumulative record folders, or pupil characteristics [the ones listed here include 'sex, physical attractiveness, level of motivation, socio-economic status'], and another study adds ethnicity. Others have added intelligence and race. Family members might be important, reinforcing the study about student siblings. Unusual names might be 'correlated with negative teacher attitudes'. Formal and informal conversations might transmit expectancies. Student test scores seem less effective when presented alone.

There is a vast literature, but overall, 'we acknowledge that even the most severe teacher expectancies explain only a fraction of variation in students' educational outcomes and that significant expectancies may affect only a fraction of all teachers' (703). Transmitting expectancies is important, 'yet it is possible that such transmission is merely a rare and unusual phenomenon… [There is]… Considerable ambiguity about whether expectancy transmission from one year to the next is prevalent or rare'. They only have hypothetical influences at the moment and there could be quite different mechanisms, and more recent research has shown greater levels of complexity, for example in identifying 'potential new intervention points in students educational trajectories' (704) [and, of course they have not particularly focused on race].

Self-fulfilling prophecies are not the same as expectancy effects. The former are 'mediated by students' own behaviours and beliefs'. It is important to look at expectancies in real-world classroom contexts, and they are currently conducting pilot research. Nevertheless there are still important research questions remaining about how expectancies are transmitted between teachers and over time, how they actually influence attributions and the role empathy might play.