Underachievement and Teacher Expectations


After all the classic stuff trying to find the causes of underachievement in homes, the emphasis in sociology of education shifts (around 1970) to looking at factors inside schools.  Perhaps something about schools helps turn students from working class backgrounds into failures/underachievers. 

The new emphasis was expressed in MFD Young's question - what sort of education is it that working class students fail at? (Young 1971).  Schools apparently exist to cater for the mass of the population - yet 3/4 of population do not do very well in them.  The old approach looked to the culture/language/environment of homes as explanations: the "new sociology of education" looked at what was actually going on in schools. 

A significant break was introduced by this approach.  The role of schools in creating inequality had been largely ignored before. In studies like Plowden, the (best) schools were OK, the problem was to make more parents aware of their part in educating their offspring. At most, the only real problem with schools was financial - the processes of education, the techniques and principles were seen as basically sound, at least following the examples of best practice.  Teachers too were basically good professional people, best left to get on with the task of teaching each child according to their merits, needing at most a bit of encouragement to use the new "progressive methods" Plowden admired so much (God, how sad it all is to look back at this era from the 1990s, after a long period of haranging and blaming teachers!!)  All this was to be challenged by the new sociology of education, which was very critical of education processes, and of teachers and their professional skills (and was very unpopular as a result) (and still is -- teachers have had enough criticism from the other wing of the political spectrum)!

The "New Sociology"

The first stage of the argument was to attack the old sociology, its findings and the methods that produced them.  This old sociology was in a rut, it was argued, it tended to focus on the same old factors of "home circumstances".  We can see why if we look at research like Plowden:  her choice of factors was heavily determined by what teachers saw as the issues.  The researchers turned to teachers for estimates of ability of students, or of parental interest.  Researchers also incorporated a fair few assumptions of their own into their estimates of the "goodness" of homes, and  often "found" in working class homes what they "knew" all along - that these homes were "deficient" in certain ways. This could be put kindly - these homes were "culturally deprived".

The same thing can be found in the work on US black children, (where the concept of cultural deprivation came from). Lots of estimates  exist here about the organisation of the home, the marital stability of black families, the quality of mothering, care and speech.  Much rather clever number-munching was based on these initial qualitative judgements, just as in Britain, but when closely examined, these judgements often turn out to be very dubious indeed, likely to reflect the values, biases, and estimates of middle class white researchers rather than the reality of the actual culture of clients. 

What was needed was new, more sympathetic examination of the culture of apparently inadequate groups.  This in turn meant new methods of enquiry, using methods borrowed from anthropology.  These involve participating in cultures and observing them, trying to understand them through the eyes of the participants NOT imposing your own value judgements. This approach drew strength from arguments at a more abstract level found in "symbolic interactionism" (see Hammersley & Woods 1976).

The New Findings

When you do this with groups of underachievers - what you do find?  That their culture like all human culture is perfectly adequate for its own circumstances and NOT simply a poor version of "proper" culture.  People are not deprived of culture, but simply have a different one. 

Famous anthropological studies like the one of Cherokee Indian children (Dumont & Wax 1971), or of the US black residents of Trackton (Brice Heath 1986) showed that the social and cultural rules relating to speaking in groups, questioning, child -rearing, or addressing adults and strangers were different in these communitites,but that this was often misunderstood as a lack of competence.

Maybe the best example is the work on language.  Linguistic deprivation is perhaps the best argument to link social class and underachievement.  It is a very commonly found and popular explanation of underachievement, often rendered in a rather simplified way.  B Bernstein, despite his sophistication is usually seen as arguing that working class children only have restricted codes - they can only use simple sentences, with lots of repetition, simple pronouns, context-bound speech etc.  Whether Bernstein ever actually meant his work to be taken like this is in doubt (see file ). In some teachers' hands this turns into arguments like: Cornish kids speak only in grunts, because they spend most of their formative years with animals.

Now compare that kind of work with the work of Labov (in Keddie 1973), on the language of  US black children.  The familiar "deprivation" arguments were found all over the earlier work on US black language - it was restricted, simple, incapable of complex thoughts etc.  Labov, a linguist, argues that this just does not sound like human language at all.  All humans are and must be capable of complex speech, all can use it to escape context and express general concepts.  It is just that we express ourselves in different variants of language. 

Labov set out to re-examine "Negro" language and discovered it to be as complex, rich, diverse and capable of abstract thought as any language.  It was simply different from standard speech: Labov termed it Nonstandard Negro English (NNE).   Why had earlier researchers not noticed these complexities?  Apart from the issue of cultural prejudice or bias, the answer was really an obvious one - people only use complex speech when they feel at ease, when they feel the listener wants to know what they think.  The context of the research is vital. 

Lots of conventional research had involved black children being brought into laboratories in universities, given toys by big, posh,  white male reseachers and asked to describe what they saw - in those circumstances, we get halting, incoherent, unfluent speech.  Labov did it differently, using local black researchers, talking to several kids at a time, putting them at their ease e.g. by using taboo words -  his results showed fluent, complex, articulated speech.  It was NOT "proper" English but it was just as good as proper English, or maybe even better: comparing an NNE speaker and an educated black speaker in discoursing about God, Labov found the educated one so concerned with qualifications, reservations, careful moderation etc., that he actually said very little. As a result, Labov is not terribly impressed with "educated speech" of the kind teachers took such efforts to impress upon their charges - an elaborated code for him often means simply an over-elaborate style - using 10 bland words where 2 will really do.

What of British working class speech?  Bernstein once found it restricted, monotonous, repetitive, context-bound. Could that be because of context?  Rosen thinks it quite possible (1972).  Maybe if Bernstein had employed local people, in context and put respondents at ease he would have found fluent, elaborate working class speech too?  Maybe if Bernstein had been more of a linguist (when he wrote the legendary paper in 1958 that teacher education courses since have enshrined), he would not have referred to middle class speech as reflecting middle class codes, but middle class style. In any event, we need many more sympathetic studies of real working class speech, and maybe some more less sympathetic and critical studies of actual school speech (see Barnes et al 1969) before we can cheerfully write off children's speech as "deprived".

Much the same point is argued in Keddie about all the famous aspects of "deprived" cultures - are homes disorganised and unstimulating OR differently organised and differently stimulating?  How much of the accusation of "cultural deprivation" reflects the values and assumptions of researchers who cannot understand different arrangements - and tend to see these as simply poor versions of "proper" arrangements (their own preferred styles.

Exercise 1

1. Record some examples of speech from speakers belonging to different social classes, and see if you can detect any differences in terms of the qualities of the speech (eg duplicate Labov or Bernstein)

2. Experiment with the influence of context on your own speech - when are you fluent and abstract, and when are you terse, or monosyllabic? Are there any occasions where speech is superfluous?

3. Record some "middle class speech" - eg school lessons, college lectures, speeches at ceremonies, debates (eg in Parliament), intellectual discussions on TV etc. How much of the impressive linguistic behaviour you are witnessing is "style", and how much is effective communication? What makes speech like this acceptable in its context?

Teachers' Judgements

If sociological researchers are guilty of bringing their own value judgements in - maybe teachers in schools do this too?  Maybe teachers have low opinions of working class or black students and see their differences as deficits.  It seems that this is exactly what happens.  Teachers expect working class/black students to be poor readers, less able, "subnormal" and "troublemakers" (See some of the classic studies like Becker 1971, Goodacre 1971, or Sharp & Green 1975 ). These expectations may be realistic ones, or they may be unfair or premature, and cloud teachers' judgements of students' real abilities. 

In practice, teachers are not very good at distinguishing the ability of students from much more dubious judgements about their social worth, Keddie argued in a famous piece (1971) - children that are culturally different, that use different forms of English, dress differently, behave differently can easily be seen as dim, thick, or as troublemakers.  All this is often misunderstood because teachers think what they are doing is simply educating students in some value-neutral way - but education is NOT value-neutral, but bound up in a fundamental way with middle class style.  Students have to conform NOT to "education" but to a style of education. Bourdieu has done a great deal to spell this out in his own distinctive way --  see file

What an odd style it often is too - styles of dress that no-one outside schools upholds any more, styles of careful polite speech, styles of behaviour that teachers themselves often find it difficult to adhere to:  not even really middle class styles, but those of 100 years ago - schools are "museums of values". 

This sort of approach is heavily critical - and asks questions like: is it necessary to speak "proper" English, is it essential to wear ties, always keep to the left in corridors, keep your hair at an approved length, avoid plastic jewelry?  Schools and teachers spend so much time enforcing and policing these styles - no wonder culturally different people find it so off-putting (and no wonder they find lots of chances to be deviant and break rules)(on this, see Reynolds 1976).

Exercise 2

We have been rather critical of the values and assumptions made by teachers, reflecting a common view in the 70s and 80s - but how fair is it? Much of the evidence is a little shaky in practice -- much of it assumes that the answes teachers give to interviewers are 'what they really believe', for example, but all the teachers I know have a scathing view of their charges and their job, really as a kind of coping strategy. Below is an exercise I once gave to trainee teachers...

1. Ask a teacher to describe a typical working class child and listen for the value-judgments if any

2. Ask a teacher to describe a "problem" child they have encountered and spot the values - are any of them related to social class? how reliable are they? what clues or cues about children are involved in these judgements?

Labelling Theory Reconsidered

It seemed that work like that cited above had offered a new and appealingly simple account of the underattainment of working class children and those from other groups. To be precise, low expectations led to low achievement, it was thought, especially after the publication of the highly influential study by Rosenthal and Jacobson (1968) (which randomly assigned completely phoney test scores to children, told teachers to expect spurts in attainment from the "high scorers", and then found real gains in those children). All that was necessary to eradicate underachievement, it followed, was to prevent teachers from having low expectations of these unpopular children (eg by telling them about the strengths of Cherokee culture, or whatever). However, there are problems with the whole approach:

1. Labels do not always "stick" - eg they may be contradicted by other labels, fought off by children or denied by their parents etc. Hargreaves et al (1975) suggest that labelling theory only works in specific circumstances (basically, when labels are reinforced,  consistent, and applied by an authoritative "significant other"). In another file, I have likened this to the discovery of the 'active audience' in media studies -- people do not always accept the views of the world offered by films or TV either
2. Expectations can lead to "negative self-fulfilling prophecies", where labellees are so outraged by low expectations that they set out to prove the labeller wrong - or the reverse for high expectations. This possibility was outlined systematically in one of the earliest contributions (Merton 1956),  but is sometimes forgotten. In fact it is not a bad description of my own response to a very discouraging school report just before I took my A-levels -- they predicted bare passes and said it was all my fault for being the wrong sort of chap. I thought to myself -- I'll show the buggers: I got A, A, and B!
3. It is hard to prove that labelling is the major factor in attainment. Rogers argues this well (1986), pointing to problems in isolating the mechanisms whereby labels are transmitted to children, for example (researchers' attention shifted from streaming and grouping practices to  ever more subtle processes of interaction, such as asking diffferent sorts of questions, smiling, interacting more frequently - but as these get more subtle, they get less easy to identify satisfactorily in research and become more ambiguous and negotiable for the recipients). 
Similarly, eliminating other factors in attainment is difficult (such as maturation - especially in longitudinal studies). Rogers also points out, incidentally, that replication of the Rosenthal & Jacobson findings is rare, and studies that show contrary findings are much less well-known. (See in the same collection Rist - a famous longitudinal study showing that early expectations by American teachers had apparently hardened into decisions to allot groupings reflecting ability at the end of the first year, and Brophy's & Good's careful and systematic study which showed that teacher expectations did seem connected to their behaviour towards children in their class, but then found no connection in a follow-up!).

There are other problems too, developed well in critical  arguments about labelling theory in another field -- delinquency. For a general discussion, click here


It seems easier to show that teachers have low expectations of some groups of children than to show that these expectations are unfair or that they cause underachievement. Certainly, there are several other factors at work in schools that might produce underattainment, and it is hard to isolate teacher expectations alone. Despite its popularity, then, especially with young and idealistic trainee teachers, the labelling theory/self-fulfilling prophecy approach is still not entirely accepted. Nevertheless, teacher values and expectations, their 'aesthetics' as would call them, are still well worth studying in their own right!


Since revisiting this file, quite a while after it was first written, I have become especially interested in the issue of 'non-verbal behaviour', or 'body language', and I have written a little additional rant about it. Try it here


Barnes, D et al  Language, the Learner and the School, Harmondsworth, Penguin, 1969
Becker, H     "Social class variations in the teacher-pupil relationship" in Cosin B et al (eds) School and Society, London, Routledge and Kegan Paul 1971
Brice Heath, S    "Questioning at home and at school..." in Hammersley M (ed) Case Studies in Classroom Research, Milton   Keynes, Open University Press, 1986
Brophy, J & Good, T "Naturalistic studies of teacher expectation effects" in Hammersley M (ed) ibid.
Dumont, R & Wax, M "Cherokee school society and the intercultural classroom" in Cosin B et al(eds) op.cit.
Goodacre, E    "Teachers and their pupils' home background" in Cosin B et al (eds) op,cit.
Hammersley M & Woods P (eds) The Process of Schooling, London, Routledge and Kegan Paul, 1976 (see  especially Introduction)
Hargreaves D et al  Deviance in Classrooms, London, Routledge, 1975
Keddie, N     "Classroom Knowledge" in Young M (ed)  Knowledge and Control, London, Collier-Macmillan, 1971
Keddie, N (ed)  Tinker, Tailor..., Harmondsworth, Penguin, 1973
Merton, R  Social Theory and Social Structure,  New York, Free Press, 1956
Reynolds, D   "The Delinquent School" in Hammersley  M & Woods P (eds) op.cit.
Rogers, C     "Research into teachers' expectations and their effects" in Hammersley, M (ed) op.cit.
Rosen, H   Language and Class..., London, Falling Wall Press, 1972
Rosenthal, R & Jacobsen, L Pygmalion in the Classroom, New York, Rinehart & Wilson, 1968
Sharp, R & Green T  Education and Social Control, London, Routledge and Kegan Paul, 1975

 back to start page