Plowden (1967)

Notes on the text

Volume 2

Preliminary pages Foreword and Contents
Appendix 1 Teachers' questionnaire
Appendix 2 Health of school children

The 1964 National Survey:

Appendix 3 1964 National Survey
Appendix 4 Regression analyses
Appendix 5 Data from the schools
Appendix 6 Infant starters
Appendix 7 Standards of reading of 11 year olds
Annexes to the National Survey

Appendix 8 Social services and primary education
Appendix 9 The Manchester Survey
Appendix 10 National Child Development Study
Appendix 11 School organisation and effects of streaming
Appendix 12 Gypsies and education
Appendix 13 Management of primary schools
Appendix 14 Variation in LEA provision

Volume 1

Report (full text)

Articles

about Plowden

The Plowden Report (1967)
Children and their Primary Schools

A Report of the Central Advisory Council for Education (England)

London: Her Majesty's Stationery Office 1967
© Crown copyright material is reproduced with the permission of the Controller of HMSO and the Queen's Printer for Scotland.

Volume 2 Appendix 11
The organisation of junior schools and effects of streaming
[pages 544 - 594]

[Note In the printed version the two tables in Part I of this appendix were numbered I and II; the tables in Part II were numbered I - XIX. In this web version I have numbered them consecutively from 1 to 21 and amended references to them in the text accordingly.]

This Appendix is an abridged version of the two reports submitted to the Council by the National Foundation for Educational Research. A fuller version was published in November 1966 as part of a Supplement to Educational Research. The reports were written by Mrs Joan Barker Lunn, assisted by CJ Tuppen and Mrs J Bouri.

The research project on which the material is based was commissioned by the Department of Education and Science in the expectation that its preliminary findings would be of help to the Central Advisory Councils. We are grateful to the Department for this foresight as well as to the Foundation who have so helpfully provided preliminary reports in advance of their own publication. The present report relates to children in different years of junior courses. The Foundation are following up those children who were at the beginning of the junior course when the study was made and this will continue until they are in their final year of the junior school. The whole report will probably be published at the end of 1968.

The organisation of junior schools and effects of streaming

National Foundation for Educational Research: a preliminary report

Part I: Junior schools and their type of organisation

Introduction

The National Foundation for Educational Research was asked by the Department of Education and Science to investigate the effects of 'streaming' and 'non-streaming' in junior schools.

The inquiry involved three main parts. Firstly, a general survey of current practices was conducted; secondly, a study of 100 junior schools (50 streamed and 50 unstreamed) was undertaken, to enable some assessment to be made of the effects of 'streaming' and 'non-streaming' on the intellectual and social development of pupils; and, finally, a more intensive study was made of 10 schools to supplement the evidence yielded by the larger scale investigations.

The survey of general practices, which is discussed in the present report, had two aims. The first was to discover the incidence of streaming and the methods of organisation practised in junior schools; the second was to obtain the information required to select a suitable sample of schools for the second stage of the study.

A fact-finding questionnaire was constructed to obtain information about the ways in which classes are formed; the criteria used in deciding the correct stream for each child; and the opportunities for transfer between streams. Questions were also asked about the general circumstances of the school and its pupils, e.g. number of teachers and classrooms, parental occupations, etc. In the early summer of 1963, the questionnaires were sent to a stratified random sample of 2,290 primary schools in England and Wales, and the heads in each of these schools were asked to complete them.

The sample was selected from the population of all primary schools in England and Wales, with the exception of schools which had less than 26 pupils on the school roll. The latter were omitted because the survey was concerned mainly with larger schools in which each year group had to be divided into a number of classes. All local education authorities except one agreed to cooperate. The sample was stratified by type of school and by number of pupils on the school roll. There were six categories for type of school: junior urban; junior rural; junior and infant urban; junior and infant rural; all-age urban; all-age rural; and four categories for school roll number.

Different sampling fractions were used for the different types of school and size categories (see details of the sample in Annex 1), and the sample was drawn from the lists made available by the Department of Education and Science.

One thousand eight hundred and sixty-nine schools returned a completed questionnaire, an overall completion rate of 82 per cent.

Of the 18 per cent that failed to return the questionnaires some schools had closed or had become infants' schools. In others, staff illness, or a recent change in staff, or reconstruction work in progress made it impossible for the questionnaires to be completed. It would therefore seem justifiable to assume that the sample returns were adequately representative, and they have been weighted according to the total junior school population (see Annex 1 for details).

In addition to the stratified random sample discussed above, a special sample of non-streamed schools was obtained by asking all local authorities for lists of their non-streamed junior schools or departments. These were supplied by the majority, and a sample of 50 non-streamed schools was selected, in addition to those which had occurred by chance in the random sample. The 50 schools completed a questionnaire.

These additional schools were not included in the analysis of the random sample but were analysed separately in cases where it was felt that the results obtained from the rather small non-streamed random sample required confirmation. In these instances, reference has been made to the 'special LEA Sample'.

All-age schools were included in the random sample but since the number of these is steadily diminishing and the total number of returns was only 113, they were discarded.

The discussion of the results relates to the situation in junior and junior mixed with infants' schools in 1963. Throughout the Report, the percentages in the tables are based on the weighted sample returns and are representative of the junior school population. The actual number of schools supplying information is also shown in each table (for details of the sample see Annex 1).

1. TYPES OF JUNIOR SCHOOL ORGANISATION

1.0 Introduction

Many schools, both large and small, are faced with the problem of finding an appropriate method of grouping their pupils into classes. In large schools, the number of pupils in each age intake is high. The intake has to be divided into several classes, and a decision has to be made as to what should be the basis or criteria for their formation. Sometimes children of one year group are allocated to classes according to their performance in standardised or school-made tests and/or the teachers' judgements, the more able pupils to the upper streams and the less able to lower streams. Less frequently, the children of one year group are assigned at random to different classes. In small schools, the problem may be how best to divide the pupils into classes, when the number of available teachers or classrooms is fewer than the number of year groups. A common method of coping with this problem is to divide the children according to age, with, in certain instances, allowance for some subsequent demotion or promotion according to ability.

One of the aims of this study was to categorise the methods used for allocating junior pupils to classes. The method adopted by a school in this respect has been designated in this Report as the type of junior school organisation. The different types of organisation found in junior schools or departments present a pattern so varied and complex that no simple classification would be adequate; and indeed, only those schools which follow a consistent pattern throughout the year groups in the formation of their classes can be categorised at all satisfactorily.

The type of junior school organisation chosen would seem to depend upon two major factors. The first of these is the number of classes in the school, itself largely determined by the number of pupils and of available staff. The second factor is what might loosely be called the head's social and educational philosophy which is likely to express itself in his attitudes to homogeneous or heterogeneous grouping.

1.1 Number of classes and type of organisation

The forms that homogeneous or heterogeneous ability grouping can take depend upon the number of classes and the number of children on roll. For this reason the schools have been divided into three main sizes representing various degrees of limitation on the possible forms of organisation. This has been done in the following way:

1. Large schools Those with two or more classes within each year group of pupils. For example a junior school taking pupils over the four year span 7-11 is classified as large if it has eight or more classes. These schools are large enough to stream consistently if they wish to do so.

2. Middle-sized schools Those with more classes than year groups but fewer than twice as many. An example would be a four year junior school with five, six, or seven classes. Some streaming is possible in these circumstances but not throughout the school.

3. Small schools Those in which the number of classes is equal to fewer than the number of year groups in the school.

1.2 The social and educational views of the head

The educational or social philosophies of heads are not easy to determine or to classify. Nevertheless two contrasting tendencies can be distinguished. If a head thinks that he can best allow for individual differences among his pupils and facilitate teaching by grouping children of like ability or attainment together, he tends to form classes on the basis of the results of a general ability test or some measure of attainment in reading and/or arithmetic. If on the other hand he believes that such practices are educationally harmful, he will assign children to classes by some other, possibly random, procedure.

Some heads do not pursue a consistent policy in this respect. For example in some of the larger schools the youngest two or three year groups were unstreamed and the fourth year streamed. The reasons for adopting this mixed type of organisation are not explicitly stated. It may be that some heads believe that an arrangement that is suitable for one age group is not necessarily so for another. Alternatively, a head who favours non-streaming may feel constrained, because of the number of pupils involved or as a result of outside pressures or the demands of the eleven plus examination, to introduce streaming into the fourth year.

1.3 The main types of organisation

Five main types of organisation can be distinguished, and these are listed below. A sixth, miscellaneous, category has been added and in this group are included those schools which were inconsistent in their methods of grouping. The seventh category is the one-class junior school, which has, of course, no choice in this respect.

Type 1. Homogeneous streaming

This form of ability grouping is found only in large schools. The children are assigned to classes within each year group, on the basis of ability and/or attainment. Thus the most able pupils are placed in the top (usually the A) stream, the less able are assigned to other streams.

Type 2. Non-streaming

This is a form of heterogeneous ability grouping, occurring only in large schools and found in two forms:

(a) Parallel classes
Within the year group, the children are divided into classes each containing pupils from all ability levels. For example, pupils of similar ability are matched and then divided equally into parallel classes, thus giving within each class a complete cross-section of the ability range. Or, more rarely, children are allocated to classes within the year group according to the initial letter of the surname, leading to effective randomisation of ability within each class.

(b) According to age
The children are divided, within each year group, according to their age, thus producing classes having a narrow age range but a complete cross-section of ability. This form of grouping has been categorised as Type 4 in all schools except large schools, where it is, in effect, a type of 'non-streaming'.

This is a form of homogeneous ability grouping across year groups, found mainly in middle-sized schools. The most common method is to have a class for the bright children of one year group, and another for the duller children of two year groups. Alternatively all classes may be drawn from two year groups; for example, the brightest nine and ten year olds may be allocated to one class, and the remainder to another. Other combinations are also possible. This type of organisation results in classes with a wide age range but which are more or less homogeneous in terms of ability.

Type 4. According to age

This type of organisation, found in middle-sized and small schools, is a form of heterogeneous ability grouping. It involves pupils being assigned to classes on the basis of age, without reference to their attainment or ability.

Type 5. 'Traditional standard'

This form of homogeneous ability grouping is found mostly in small schools and seems to be a partial survival of the old system of 'standards'. It usually involves the retention or accelerated promotion of a number of pupils each year and its effect is to produce a wide age range but a more or less homogeneous level of attainment within classes.

Type 6. Other methods (miscellaneous)

These, found in large, middle-sized and small schools, are for the most part combinations and variations of the methods already described, although some schools use other criteria (e.g. sex, Welsh language) for grouping pupils. This category includes both homogeneous and heterogeneous ability grouping.

Type 7. One class junior schools

The relationship between the type of junior school organisation chosen and the number of classes in the school is illustrated in Table 1.

Table 1 Relationship between type of school organisation and number of junior classes

1.4 The number of junior pupils involved in and the proportion of schools using each type of organisation

Table 2 below shows the approximate number of junior pupils affected by each type of organisation and the percentage of schools using each. From these figures it will be seen that a junior school pupil in 1963 was much more likely to be in a school using a form of homogeneous rather than heterogeneous ability grouping - in fact at least 56 per cent (1) of all pupils were in schools using homogeneous ability grouping.

Table 2 The approximate number of junior pupils involved in and the proportion of schools using each type of organisation

Summary of main conclusions

The following provides a summary of the main findings of each section of Part I; for a full discussion see NFER publication New Research in Education, November 1966.

2. LARGE SCHOOLS AND THEIR TYPE OF ORGANISATION

1. In large schools (i.e. eight or more classes) the predominant type of organisation was homogeneous streaming - used by 65 per cent of large schools. Six per cent of large schools only were entirely unstreamed, and five per cent were unstreamed except for one year group. Four per cent kept the two younger year groups unstreamed but used homogeneous streaming in the two older year groups. Twenty per cent of large schools used other mixed methods.

2. The average size of classes in all large schools was equal, namely 36 pupils. However, class sizes in schools using homogeneous streaming were more varied than those in non-streamed schools. Higher ability streams tended to have more pupils than lower ability streams, whereas all classes in non-streamed schools had approximately the same number of pupils.

3. There were proportionally more girls than boys in the higher ability streams and conversely more boys than girls in the lower ability streams.

4. There was a difference in the average ages of A, B and C streams in all year groups. The A streams had the highest average age and the lowest ability streams were the youngest.

The most important findings are points three and four above. These indicate that the practice of homogeneous streaming may penalise boys and younger children of a year group.

Some of the disadvantages of being a 'younger' child in a school year group might be removed if all children were allowed to start school at the same point in the year, thus removing the advantage of longer schooling, which the 'older' child now has. It seems logical for children to start school together at the beginning of the school year, since at every stage of the education system children born within a school year are treated as a group (e.g. they move up the school together, take 11 plus at the same time, etc.).

3. METHODS OF ALLOCATION TO 'STREAMS' AND CLASSES IN LARGE SCHOOLS AND THE PUPIL'S CHANCES OF TRANSFER

1. Schools using homogeneous ability grouping commonly assigned children to classes in their first year at junior school on the basis of their infant school record. After the first year, an internal examination became the most common criterion.

2. Schools using non-streaming paid more attention to age and relatively little to infant school record as a criterion for assigning children to classes.

3. After the initial assignment to a class or stream in the first year of the junior school, the chances of a pupil being transferred are very slight.

Taking the school year 1961-62 as a typical year, it appears that movement between streams was relatively small. Once children had been assigned to their streams at seven plus, most of them would remain in the same stream throughout the four years of the junior school. Against this we should put Vernon's estimate, based on the known correlations of the measures used, that about 10 per cent of all children should be upgraded or downgraded each year if relative homogeneity is to be preserved.

Thus reliability and validity of the methods used for grading children to A, B or C streams when they first enter the junior school appear to be of the utmost importance; it is surprising to find that so many junior schools graded their pupils without the help of standardised tests and without making allowance for age.

Since the same largely unscientific criteria as are used to grade the children initially are also used to assess whether second, third or fourth year children should be regraded, it is hardly likely that grading errors will be recognised and rectified at these later stages. Lack of transfers could, of course, be due to an accurate assessment at seven plus, or could be a result of a self-fulfilling prophecy in that membership of a stream tends to condition learning.

4. MIDDLE-SIZED SCHOOLS AND THEIR TYPE OF ORGANISATION

1. The major type of organisation in middle-sized schools (five, six or seven junior classes) was vertical streaming (28 per cent). Thirty-six per cent used other methods, 19 per cent according to age, and 17 per cent traditional standard.

2. Thirty-eight per cent of the classes in schools using vertical streaming were composed of children from two or more year groups. These classes tended to consist of average and slow pupils, while brighter pupils were more often in classes of one year group only. Vertical streaming was found mainly in six and seven class schools.

3. In schools grouping according to age, 70 per cent of classes were composed of one year group only, while 30 per cent had two or more year groups. According to age was found mainly in five class schools, and nearly half of these had a remedial class, thus leaving one class each for the four year groups.

4. Traditional standard method was found mainly in five and six class schools.

5. The average number of pupils in the class in middle-sized schools was 34.5 although the actual numbers in any class varied more than in large schools; there was little difference in class size between the various types of organisation.

5. SMALL SCHOOLS AND THEIR TYPES OF ORGANISATIONS

1. It was found that the major types of organisation in small schools (four classes or less) were traditional standard method (used by 35 per cent) and grouping according to age (also used by 35 per cent).

2. In schools using the traditional standard method, the most frequent type of class was that composed mainly of average children but with a few bright younger pupils who had been promoted. Next most frequent were classes in which a few slow pupils had been 'kept down'.

3. Very often small schools were forced to form classes of more than one year group either because of unequal numbers in the different year groups or through having to allocate four year groups into three classes. The head teacher may, in such circumstances, form classes on the basis of the traditional standard method, or he may group according to age.

4. Seventy-six per cent of classes in small schools were composed of two or more year groups.

5. Some head teachers used the traditional standard method - i.e. kept slow pupils down and promoted bright pupils - when they were not forced by circumstances to mix year groups.

6. The average number of pupils in the class was 28 in schools using traditional standard and 29 in those grouping according to age.

6. INTRA AND INTER-CLASS GROUPING

A Intra-class grouping

1. Class teaching (30 per cent of all classes) was used less frequently than group teaching for the 3Rs. The least common intra-class practice was for children to remain in the same group for reading, mathematics and English (nine per cent of all classes).

2. There was a tendency for more attainment grouping in schools using heterogeneous ability grouping (i.e. schools using non-streaming or according to age).

3. Some schools appeared to teach the class as a whole and there was no grouping for the 3Rs even though the class was heterogeneous with respect to ability (28 per cent using non-streaming, 22 per cent in middle-sized schools grouping according to age). This may include schools using individual teaching methods.

4. Some classes were divided into sub-classes on the basis of attainment/ability and pupils were taught the 3Rs in these groups.

5. The smaller the school, the more use made of group rather than class teaching.

6. Grouping was used equally frequently for reading and mathematics, but less frequently for English.

B Inter-class grouping

1. Setting between classes was extremely rare. Four per cent of all schools 'set' for English and six per cent for both mathematics and reading.

2. Twenty-one per cent of schools using non-streaming, 47 per cent of middle-sized schools using according to age and 27 per cent using traditional standard had a remedial class.

3. The most common form of remedial group in large schools was the one which met irregularly or for part only of the school day.

4. Remedial help of any kind was much rarer in smaller than large schools.

5. A high percentage of large (81 per cent) and middle-sized schools (53 per cent) segregated their slower pupils in a special class. One in five of schools using non-streaming and a half of those grouping according to age (middle-sized schools) formed remedial classes, even though they did not separate their other ability levels.

6. Slower pupils in small schools were not segregated, although some were 'kept down' with a class of younger pupils (35 per cent of small schools used this method).

7. NEW TRENDS IN JUNIOR SCHOOL ORGANISATION

1. The major types of organisation being introduced in junior schools were non-streaming (19 per cent), grouping by age (12 per cent), group teaching (15 per cent), and individual teaching (10 per cent).

2. Nineteen per cent of schools at present using homogeneous streaming intended to introduce non-streaming or to extend it beyond the first year.

3. Thirty-six per cent of schools using non-streaming 'streamed' their fourth year pupils, but 24 per cent intended to change this and unstream them.

4. The main advantages claimed for non-streaming were: pupils more socially adjusted (51 per cent); pupils not labelled A, B, C (27 per cent); wide ability range (22 per cent); contented staff/easier for staff (23 per cent).

5. The main advantages claimed for homogeneous streaming were: easier for the staff/contented staff (23 per cent); smaller ability range in the class (19 per cent); more attention to the backward (22 per cent); meets the needs of the child (19 per cent); and higher standard of attainment (18 per cent).

6. Advantages claimed by schools using vertical streaming were similar to those of schools using homogeneous streaming.

7. Advantages claimed for grouping according to age in middle-sized schools were similar to those claimed for non-streaming (see above).

8. Advantages given by/for small schools represent advantages resulting from being in a small school, rather than from particular types of organisation. The main advantages claimed were: more individual attention; child's needs met; happy family atmosphere.

9. Special difficulties encountered by head teachers were: mainly inadequate number of staff; large classes; inadequate accommodation; too wide ability range and age range within the class.

Over 50 per cent of head teachers mentioned some special difficulty. Middle-sized schools mentioned the shortage of staff more often than other schools.

8. SOME GENERAL FINDINGS OF THE SURVEY

1. Qualified and unqualified staff

Seventy-one per cent of all schools were entirely staffed by qualified teachers. The larger schools were more likely to have an unqualified teacher. Schools using homogeneous streaming and non-streaming were equally well staffed.

2. Staff turnover

This was related to the size of the school but not to the type of school organisation.

3. Accommodation

Large schools had better facilities; they were more likely to have a room for the head, a staffroom and a school hall. Schools using vertical streaming were more likely to have a spare classroom than schools using any other type of organisation.

4. Schools of different socio-economic categories

Schools classified as middle class or upper working class were compared with those classified as middle or lower working class.

It was found that the latter group tended to be worse off in terms of number of unqualified staff and turnover of staff. Their heads spent more time teaching. Also they were less well provided with accommodation - however, these schools tended to be small in size, which would explain this finding. The socio-economic categories were based upon the head's subjective estimate and the above findings must be interpreted with caution.

Part II The effects of streaming and non-streaming in junior schools

Section 1 Introduction

1.1 The aims of the investigation

The inquiry was initiated at the request of the Department of Education and Science and was designed to study the effects of streaming and non-streaming in junior schools.

The aim of the project is to provide evidence concerning the ways in which children are assigned to different classes, and grouped within their classes, in both streamed and non-streamed schools; the attitudes of teachers towards these various forms of organisation, the methods of teaching they employ and the facilities with which they are provided; and, finally, the effects of different methods of grouping and treatment on the attainments, personalities and social adjustment of the pupils concerned.

To this end, the inquiry has been undertaken at three levels:

1. A broad survey of grouping practices in a nationally representative sample of junior schools.

2. A comparative study of 50 matched pairs of streamed and non-streamed junior schools.

3. An intensive study carried out in 10 junior schools.

A report of the survey of practices in junior schools was completed and submitted to the Plowden Committee earlier in the year. (2) The study of 100 schools (50 matched pairs of streamed and non-streamed junior schools) falls into two parts: a longitudinal study of pupils over four years and a cross-sectional study which enables certain preliminary comparisons to be made between the two types of organisation based upon samples of children, at present in different years of their course. It is this cross-sectional study with which this Report is concerned. The longitudinal inquiry, which should yield more reliable information concerning the effects of the contrasting forms of organisation, cannot, of course, be completed until the four years, planned for the follow-through, have elapsed.

In interpreting what follows, it should be remembered that all the findings are based upon cross-sectional data and that this kind of information has certain disadvantages. The most notable of these is that inferences about the growth of children based upon two groups, tested at the same time, but differing in age, are less reliable than inferences about the same group of children tested or examined at intervals. One cannot, in fact, legitimately generalise from one year group to another and assume that, because the ten year olds of the present sample behaved in certain ways, the seven year olds will resemble them closely three years from now. Nor, in cross-sectional samples can conclusions be drawn about the direct effects of a change in organisation or in placement, particularly as this may affect children of initially different levels of ability. It is hoped that the follow-through studies will enable firmer conclusions to be drawn in these respects.

A further limitation should be noted. Not all the data from the cross-sectional study could be analysed in time and a number of important questions - the answers to which might well qualify some of the statements made - have had to be left aside for the present. Where this is so, and where it is reasonable to suppose that subsequent analysis may throw a new or different light upon the interpretation of the factual data given, this is pointed out in the text. Finally, it should be emphasised that the question - to stream or not to stream? - reveals itself, as the research continues - to require a far more complex and nuanced answer than the propagandists on both sides would have one believe.

1.3 The sample of 100 schools

As a result of the major survey, details of a large number of streamed schools and of 43 non-streamed schools were available. To increase the size of the non-streamed sample of schools, the Foundation wrote to all local education authorities in England and Wales, asking for addresses of all non-streamed schools. These schools were invited to complete a questionnaire and on the basis of the information thus provided, a further 55 non-streamed schools were selected.

Each of the non-streamed schools was then matched with the most similar streamed school which could be found, using the following criteria for matching:

(a) Type of school: junior or junior-with-infants; urban or rural.

(b) Number of classes in the school.

(c) Average number of pupils per class.

(d) Geographical region.

(e) Percentage of children in LEA attending non-selective schools.

(f) Predominant socio-economic class of parents.

(g) Approximate matching of percentage of parents in professional, clerical and skilled occupations.

1.4 Exploratory research

One aim of the inquiry is to discover whether children tend to reach higher levels of achievement in the basic school subjects under one system than under the other. For this purpose, tests of attainment are, of course, appropriate. Another stated aim is to examine the effects of streaming and non-streaming on children's personality characteristics and social adjustment. In these areas, however, the choice of suitable measuring instruments depends on the identification of modes of behaviour that are both relevant and susceptible to reasonably objective appraisal. For example, there are references in the literature to concepts such as 'the atmosphere of the school'. Clearly such a concept needs to be defined in operational terms before any meaningful assessment could be undertaken. Furthermore, it is reasonable to suppose that some of the outcomes of streaming or non-streaming depend not so much on the forms of organisation per se as on the attitudes and practices of the teachers who operate within them.

For these reasons it was felt essential to carry out exploratory research, in order to discover which aspects of the school, teacher, and pupil could and should be given particular attention in the 100 schools study.

The exploratory research involved six streamed and six non-streamed schools (not included in the 100 schools sample) and consisted of interviews with the heads and staff, and visits to classes. The staff who were interviewed were selected so that the sample included both men and women, with varying amounts of experience and teachers of different year-groups and different streams. Thirty-one interviews were carried out - each lasting about one hour. These interviews were unstructured, so that the teachers could express their views freely, as well as explain their methods of teaching. Each interview was taped and later transcribed.

1.5 The significance of 'traditional' and 'progressive' teaching methods and attitudes

The interviews and the visits to schools suggested that teachers in streamed schools differed from those in non-streamed schools. both in their teaching methods and in their attitudes.

It also became clear that these differences between teachers in their methods and attitudes might possibly outweigh the effects of streaming or non-streaming per se. A good deal of attention was therefore given to the construction of questionnaires to assess the method of teaching and the attitudes of the teachers. These questionnaires were based upon the activities described and the opinions offered in the interviews. (See Section 2 for further details of Questionnaires S.1 and S.3.)

1.6 Hypotheses

On the basis of the interviews and visits, a number of hypotheses were formulated, some for testing in the cross-sectional study and others in the longitudinal study. Outlined below, are those hypotheses tested in the cross-sectional study and discussed in this Report. Indicated beside each hypothesis is the number of the section in which it is discussed.

(a) Streaming and non-streaming may have different effects on children's attainments. (Section 3.)

(b) Pupils' attainments may be affected by the methods of teaching used, irrespective of the type of organisation chosen by the school. (Section 3.)

(c) Pupils' attainments and social behaviour may be influenced by teachers' attitudes. (Section 4.)

Although the project started with 100 schools, for various reasons certain schools had to be excluded, so that data from only 84 schools, 42 streamed, 42 non-streamed, were analysed in the cross-sectional study. Details of the 42 matched pairs of schools can be found in Annex 1.

Eight hundred and seventy-one teachers and approximately 30,000 junior pupils were involved in the cross-sectional study.

Section 2 The characteristics and attitudes of teachers in streamed and non-streamed schools

2.1 Introduction

It will be recalled that the two samples of 50 schools were matched as far as possible for their principal characteristics - size, socio-economic status, geographical location, and so on - and were differentiated by their choice of internal organisation, into streamed and non-streamed schools. The interviews and visits carried out, as exploratory research, suggested that this difference in organisation was reflected in, or arose from, very different views held by the teaching staff which in turn affected the choice of teaching methods.

From the material recorded at the interviews therefore, three questionnaires were devised to enable this hypothesis to be tested in some depth.

2.2 Information supplied by class teachers

The first of these questionnaires was concerned with biographical information, sex, age, experience and training.

The second was concerned with more or less objective information about the methods of teaching used and the ways in which children were organised for learning. It fell into three sections:

(a) A section dealing with the classroom facilities as perceived by the teachers. This yielded a composite 'facilities' score.

(b) A section in which teachers were asked to indicate how often they used various types of lesson, for example 'formal sums'. From the responses made, two scores were derived: a 'traditional lessons' score, indicating how often the teacher used lessons of the more 'traditional' type and a complementary 'progressive lessons' score. Each of these scores was studied by the method of item analysis, in order to eliminate items (i.e. types of lesson) which were not correlated with the total score. The final versions contained the following items:

'Traditional' - writing class-prepared compositions; learning lists of spellings; formal grammar -understanding parts of speech; saying and learning tables by rote.

'Progressive' - projects - in which the child does his own 'research'; pupils working or helping each other in groups; practical arithmetic, e.g. measuring, apparatus work; free activities.

(c) A section on the use of class teaching, mixed ability group teaching and ability grouping within the class.

2.3 A comparison of streamed and non-streamed schools

Although, of course, few teachers or schools conformed to a clear-cut average, the analysis of the data revealed a considerable degree of polarity between the two types in terms of the age and experience of the teachers, the methods they chose to use and their perception of the facilities available to them. Table 3 brings this out clearly, indicating that, with one exception, all the differences found are statistically significant and (where this is appropriate) point in the same direction. The climate in the unstreamed school - if we are to judge by what its teachers say about themselves, their methods and their attitudes - is more permissive and tolerant, less structured and places less emphasis on the more traditional methods of class teaching than its streamed counterpart. Since, moreover, all the traits studied proved to be intercorrelated positively, this suggests that they are something more than an arbitrary collection of more or less discreet [discrete?] and separate characteristics; that they do represent a coherent pair of opposing syndromes, likely, if they are at all intense in anyone school, profoundly to affect the pupils exposed to them.

Table 3 Differences between streamed and non-streamed schools

It must be emphasised, however, that these (Table 3) are central tendencies and some elucidatory comment is perhaps required.

2.4 Age and experience of teachers in the two types of school

In general, a higher proportion of the teachers in non-streamed schools tend to be younger and to have less experience than those in streamed schools.

Table 4 Age distribution of teachers in streamed and non-streamed schools

Table 4 shows that a quarter of the teachers in non-streamed schools are aged 26 or below, whilst a quarter of those in streamed schools are verging on their fifties or are older. However, the overlap in age and experience is in fact greater than the difference.

As a group, younger teachers tended to have more 'progressive' opinions, particularly in their rejection of eleven plus selection, their 'permissiveness' and their 'tolerance of noise'. Those with two years' experience or less were the most hostile to 11 plus.

2.5 Class facilities

According to the teachers, the facilities (good lighting, new desks, little outside distraction, plenty of space, etc.) in the unstreamed schools were better. It should be emphasised that this reflects what the teachers thought about what they had, and is not derived from a strictly objective comparison; but there is no reason to suppose that the non-streamed schools were not in fact superior in this respect.

2.6 Teaching methods

Teachers in streamed schools, on average, tended to make more frequent use of 'traditional lessons' and less frequent use of 'progressive lessons' than teachers in non-streamed schools. But the greater frequency of 'traditional lessons' (streamed schools) was in fact more true for teachers of seven plus children than for teachers of the ten year olds, not surprisingly when one considers the meaning of 'traditional lessons' at the different ages. 'Traditional lessons' could be defined as learning the basic skills in a formal way - an activity considered important for seven year olds even by 'permissive' teachers (there was no correlation between use of 'traditional lessons' and attitudes for the teachers of seven year olds). But by the time a child reaches the fourth year of the junior school, fewer teachers believe in the necessity for daily recitation of tables or practice in formal sums. The teacher who did use 'traditional lessons' at ten plus, tended also to be 'obsessional' in outlook (4) (the correlation between the use of 'traditional lessons' and the 'permissive'/'obsessional' scale was 0.327).

The greater frequency of 'progressive lessons' (5) in non-streamed schools was in fact more striking in the case of teachers of ten year old children than in that of teachers of seven year olds. The use of 'progressive lessons' also was found to be correlated with the teacher's 'permissiveness' in fourth year classes, but less so in the earlier years. One possible explanation for the low correlation between teacher attitude and use of 'progressive lessons' at seven plus is that use of apparatus (e.g. Dienes, Cuisenaire) and consequent group work is becoming increasingly common with these year groups and is introduced by the school rather than being chosen by the individual class teacher. Lessons in formal sums (i.e. mechanical computation) and arithmetic tests and other tests were more frequent in streamed schools whereas practical arithmetic, in general, and the use of apparatus, were more common in non-streamed schools.

In the streamed schools, perhaps in accordance with their greater emphasis on structuring experiences, on teaching children in specific ways and upon the more traditionally systematic and formal approach, significantly more use was made of tests of various kinds to check progress and diagnose difficulties. The only aspect of method studied which did not yield a statistically significant difference was that concerning the use of sums in problem arithmetic. It is possible that differently organised schools are in fact similar in this respect; alternatively the test itself may not have been sufficiently discriminative to detect any difference which might exist.

Table 3 has shown the difference between schools over all year groups; while Table 5, below, shows the frequency of different types of arithmetic lessons at seven plus and at ten plus. Note that there are no significant differences in the frequency of problem sums at seven plus or formal sums at ten plus.

Table 5 Frequency of tests and of different types of arithmetic lesson in streamed and non-streamed schools

The general differences between the two types of school are summed up in the composite score on the two complementary 'traditional' and 'progressive' scales. How these are distributed is shown in Table 6 which sets out the medians and quartiles of the two groups.

Table 6 Traditional and progressive scores of teachers in streamed and non-streamed schools

It should be noted that, on both scales, scores range from four to 24 and that the medians are considerably nearer the 'traditional' and 'non-progressive' ends of the scale. In terms, therefore, of the views expressed, teachers in both types of school are fairly 'traditional'; the difficulties between them would seem to be nuances rather than marked divergences of opinion.

2.7 Attitudes of teachers in streamed and non-streamed schools

The exploratory research indicated a number of fundamental attitude areas that are relevant in this inquiry. Each area could be represented as a dimension along which teachers could be placed according to the degree to which they manifest the attitude. Six attitude areas were investigated, namely:

(a) 'Permissive'/'Obsessional'
(b) Attitude towards physical punishment
(c) Attitude towards noise in the classroom
(d) Attitude towards A stream children
(e) Attitude towards 11 plus selection
(f) Attitude towards streaming.

(a) 'Permissive' versus 'obsessional'

This scale was intended to rank teachers in terms of the 'permissiveness' of their attitudes towards junior school children. The labels 'permissive' and 'obsessional' must not, of course, be interpreted literally. It is convenient to use some title to describe those teachers who were most prone to object to children fidgeting, to demand clean hands and good manners; the teachers who rated the three Rs more highly than 'self-expression' and vice-versa.

There was a tendency (highly significant statistically) for teachers in non-streamed schools to be more 'permissive' in this sense than those in streamed schools. For the content of this scale, see Annex 3, Items 6, 14, 17, 20, 23.

Graph No. 1 Permissiveness

It will be seen from the graphs that the distribution of the scores for the two groups of teachers is very similar, and in fact, the modal statement was the same. The difference arises from a group of teachers in non-streamed schools who hold firm views towards the extreme of the scale. The statement which most clearly differentiates the streamers from the non-streamers is quoted above the graph.

(b) Attitudes towards physical punishment

There was, too, a similar tendency (p<0.001) for teachers in non-streamed schools to show disapproval of physical punishment and in streamed schools to show approval. The items forming this scale can be found in Annex 3 (items: 4, 11, 13, 22, 25, 30).

Graph No. 2 Attitude to physical punishment

It will be seen from the graph that, as with 'permissiveness', the difference arises principally from a small group in the non-streamed schools who have extreme scores on the scale. The statement most clearly differentiating the two groups is that concerning physical punishment in general.

(c) Attitudes towards noise in the classroom

The differences between teachers in streamed and non-streamed schools were highly significant (p<0.001). Teachers in non-streamed schools were tolerant of noise (Annex 3 Items: 2, 7, 18, 27, 31) and in streamed schools less tolerant.

Graph No. 3 Attitude to noise

Graph No. 3 shows a clear displacement of scores of the teachers in non-streamed schools towards one end of the scale with differences on most of the statements.

The three more general attitude scales also differentiate between teachers in the two types of school, and if indeed they reflect strongly held views, may deeply influence pupil-teacher relationships.

(d) Attitude towards A stream children

On this scale individuals rank themselves in terms of attitudes towards children in A streams. It contained items suggesting that A stream children worry about marks, tend to become conceited, etc. (see Annex 3, Items: 3, 5, 8, 10, 26). Teachers in streamed schools had a more favourable attitude to A stream children - presumably believing in the value of A streams for bright children - than teachers in non-streamed schools. The difference was significant (p<0.001).

Graph No. 4 Attitudes to A stream children

Here the graph shows a rather different distribution with a clearer division of opinion - certainly among substantial groups of teachers in the two types of school. The most discriminating statement is, it will be noted, concerned with the attitude attributed to children.

(e) Attitudes towards 11 plus selection

Teachers in non-streamed schools tended to disapprove of selection, while teachers in streamed schools tended to approve. The difference was highly significant statistically (p<0.001). (See Annex 3, Items: 9, 12, 16, 21, 29).

Graph No. 5 Attitude to the 11 plus

There seems to be a considerable spread of opinion in both groups about this with, however, the bulk of teachers in non-streamed schools further towards one extreme than their colleagues. Two points are worthy of note: that the most discriminating statement concerns the justice of the 11 plus; and that substantial proportions of teachers in both groups are opposed to the system.

(f) Attitudes towards streaming

In many ways this attitude scale is the most important for this research project. If a class teacher does not share the beliefs of his head teacher about the advantages of streaming or non-streaming, it is possible that these advantages will not be realised. This scale can be used as a means of identifying teachers who are committed to one point of view or the other. It has been found that in streamed schools there are some teachers who favour non-streaming, and in non-streamed schools there are some teachers who favour streaming, and in both types of school there are teachers who are uncommitted either way. However, the average differences between teachers in streamed and non-streamed schools were highly significant (p<0.001).

The attitudes of teachers towards streaming were studied in greater detail in order to discover the extent of teachers' agreement with their school's policy. As a first step the attitude scores were used to categorise the teachers into three groups ('pro-streaming', 'neutral' and 'anti-streaming') according to Guttman's method of intensity analysis. (6)

Table 7 shows the proportion of teachers in each attitude category.

Table 7 Attitudes of teachers according to type of school organisation

The first point to be noted is that substantial proportions of teachers are in favour of streaming. Since of all junior schools in the country which are large enough to do so, at least 65 per cent stream and only 11 per cent clearly do not, (7) it seems reasonable to suppose that the majority of junior school teachers in England and Wales are in favour of the practice, that a substantial proportion are undecided, and that only relatively few are firmly committed to the opposing view. This finding must be stressed at a time when some writers suggest that the death knell of streaming has already sounded. Coupled with the finding concerning parents' views, reported later in this report (Section 6), it suggests that any universal change recommended may meet with considerable opposition - particularly since the attitudes for or against streaming seem to form part of a whole syndrome of views, practices and beliefs.

An examination of Table 7 also suggests that on the average, there appears to be one teacher in each school whose views are in conflict with the policy of the school. This, however, is rather misleading. A more detailed analysis revealed that in a few schools less than half the staff were in agreement with the head's policy, while in others all were in accord. Similarly, there were differences (discussed in Section 5) in the views of A, B, C and D stream teachers.

Graph No. 6 Attitudes to streaming

This graph reveals a pattern of response most clearly differentiating the two groups of teachers. Whereas in both groups (as Table 7 showed more clearly) there are those whose opinions are contrary to the practice of their school, in each the majority seem to adhere to the appropriate norm. The most discriminating statement, significantly, is the one which touches on the fate of the bright child in a non-streamed school. A clear majority of teachers in streamed schools agree that such a child will be neglected; a comparable majority in non-streamed schools oppose this view.

2.8 Streamed and non-streamed schools embody different philosophies

It seems apparent from the foregoing that schools using streaming or non-streaming do not merely differ in their organisation: and indeed that the strictly organisational aspects may be the least important in their bearing upon the development of the pupils. The streamed school seems to be more systematic in its approach, concentrates more on conventional lessons, gives more attention to the 3Rs and is likely to be more 'traditional' and at least overtly more authoritarian. Its staff is likely to be somewhat more experienced and older, to approve of bright children, of 11 plus selection and of streaming as a means of adapting to individual differences. The unstreamed school presents an apparent contrast. Its younger teachers hold more 'permissive' views on such things as noise, cleanliness and manners; they disapprove of physical punishment and of the differentiation implicit in streaming and in the 11 plus procedures. Their teaching tends to place more emphasis on self-expression, learning by discovery and practical experience. In short, the aims and practices of the two kinds of school seem to embody different views about children and different philosophies of education.

In practice, however, the contrast is probably not so marked as this dichotomy may seem to imply, although, of course, specific schools, where staff and head share very firm convictions, may represent quite thorough-going, even doctrinaire, embodiments of these contrasting approaches. In view, however, of the frequency of streaming in junior schools, it is clear that it is the non-streamed school which, in the statistical sense at least, is deviant. It would also appear that the staffs of non-streamed schools are slightly more in agreement with each other than with the rest of their colleagues. How far this is due to teachers choosing non-streamed schools on account of their views and how far it is due to the influence of the school and the head, it is difficult to say. It seems possible that some of the heads of non-streamed schools, because they are using a new and fashionable method, have been accorded special treatment by their LEA and allowed to select teachers in accordance with their own viewpoint.

Many of the questions raised by this part of the study can be answered only by further analysis and the collection of additional data. One thing is, however, clear and should be kept in mind with regard to Section 3. Any effect which may be shown to be associated with streaming or non-streaming is unlikely to be purely and simply due to the form of organisation used. Teaching method, the ideas which underlie disciplinary systems, the views about children held by their teachers - in short the whole climate of relationships built up by what teachers say and do and what they appear to their pupils to imply, may well be the critical factors. The comparisons which follow will show the broad differences associated with contrasting types of general attitude and consequent organisation. In many schools of both types, however, there are teachers whose views and at least some of whose practices, are in conflict with those of their colleagues. One might expect the influence of such teachers to attenuate any differences that might result from the differences in organisation. The extent and nature of this it will only be possible to elucidate in the later comparisons.

Section 3 Comparison of attainments in streamed and non-streamed schools

3.1 The difficulty of finding tests which favour neither type of organisation

Tests of reading, English, mechanical arithmetic and problem arithmetic were given to the pupils in all four junior years in the cross-sectional study. Each test was devised to be suitable for all ages from seven plus to ten plus, so that in the longitudinal study, measures of gain in achievement could be made. These tests were specially produced for this project by the test service of the NFER, and their content was of the kind usually demanded for juniors by teachers and education authorities. As was pointed out earlier, however (paragraph 2.7), the teaching methods and the aims of the streamed schools are typical of the majority, whereas non-streamed schools in which teachers held different views are in the minority: hence, the tests used in the cross-sectional study probably reflect the majority view of teachers. In this sense, they can legitimately be held to be biased against the non-streamed schools, in that they favour outcomes of methods of teaching to which less attention is paid in these schools, and do not include measures of some of the objectives which many non-streamed schools consider to be of great importance. On the other hand, it might reasonably be argued that the bias implicit in the tests is a correct one. On this we hold no views, the important point to make here being that the comparison, at this stage, is in terms of the norms and objectives of the majority of teachers. Comparisons have yet to be made on the basis of how far other aims and objectives proposed by the non-streamers are attained.

3.2 The tests

Of the four attainment tests used, the most impartial was probably the reading test. Both types of school share the aim of teaching children to read with understanding and, within its limitations, the test gives an accurate measure of how far this has been achieved. The test does not, of course, give a measure of the use and enjoyment of reading, nor can the inquiry furnish evidence of this at the present stage.

The test of English is rather more 'biased'. It consists principally of tests of comprehension, probably important in themselves for both groups, but on which children who have had some formal practice and teaching are likely to have at least a marginal advantage. Also included in the test are some items based on formal grammar (parts of speech, punctuation) which would favour those pupils whose teachers scored highly on the 'traditional lessons' scale and such teachers were more common in streamed schools.

The two tests of arithmetic are somewhat different from each other. One was concerned with problem arithmetic and since this was that aspect of teaching which did not differentiate between the two kinds of school and teacher, it is reasonable to assume that any bias was probably slight. The test of mechanical arithmetic is another matter. Table 3 indicates that, over all year groups, daily practice in formal sums was more common in the streamed schools, and one would expect that this would result in a greater facility in this test. Any differences found here then might perhaps be ascribed to the teaching method used rather than to the form of organisation as such. (8)

In addition to the attainment tests, two measures of ability were used: the NFER Primary Verbal Test I was given to children in the second, third and fourth years; and a Verbal/Non-verbal test, yielding two separate scores was given to fourth year children. (9) This latter test was the one used in 1957 in the Population Investigation Committee Follow-up Study (Douglas JWB The Home and the School.)

Finally, it should be stated that two parallel versions of the attainment tests were used. A random half of the schools worked the A version and the other half the B ones. This has the advantage that, in effect, two independent studies were carried out simultaneously and each acted as a check upon the other.

3.3 The results of the attainment testing

The overall results of the attainment testing suggest that, on the various outcomes measured, pupils in streamed schools performed better than their counterparts in non-streamed schools. Most of the differences found were statistically significant.

Few of the differences, however, were large; most amounted to less than one third of a standard deviation and only those derived from the mechanical arithmetic score approached or surpassed half a standard deviation. In effect, what these differences mean is that, on the average, the streamed group got two or three more answers right in tests having 30 or 40 items in them. Some would consider this a small price to pay for non-streaming, if other educational and personal advantages were obtained.

Attention should also be drawn to another feature of the score distributions. In attempting to interpret, in educational terms, statistically significant differences in mean scores, one must ask whether the higher or lower marks were spread similarly in the compared groups or whether the difference is due mainly to a larger proportion in one or other group, obtaining higher or lower marks. The standard deviation of the test scores showed that in streamed schools there was less homogeneity in performance and that more children in these schools were getting higher marks. There was little difference in the percentage of children scoring low marks in the two groups of schools. The most likely interpretation of this seems to be that the most able children in the streamed schools score more highly than do their counterparts in the unstreamed ones. This may or may not be true, however, the longitudinal study should provide more reliable evidence.

The table which follows below illustrates two other points. It shows the correlations, at the beginning and at the end of the junior school, between the type of organisation used by the schools and the results of the four tests of attainment. It will be noticed that, whereas in every case except one the correlations reflect the superiority of the streamed schools in the aspects of attainment measured, there are two gradients of what might be called the intensity of association.

Table 8 Correlation between type of school and test

The more the test reflects 'traditional' educational practices and emphases, differentiating the two kinds of school, the closer the association is found to be. It is at its highest for mechanical arithmetic and at its lowest for reading. English and problem arithmetic occupy intermediate positions. Secondly, it is apparent that the associations are strongest at seven plus and weakest at 10 plus. Other workers have found that children taught by 'progressive' or 'active' methods tend to make a slower start, but catch up with those taught by more 'traditional' methods by the end of the primary school. (10, 11)

3.4 Conclusion

In concluding this section, it is necessary to remind ourselves that this Report is concerned with cross-sectional data and that therefore comparisons between successive age groups are of dubious validity. It should be noted that although the two groups of schools were matched as far as possible, one slight difference did emerge. The streamed schools had slightly more pupils from somewhat higher social backgrounds (see Annex 1). Further, the test of verbal ability given to the second, third and fourth year pupils revealed a small superiority in the eight and nine year old children in the streamed schools, though neither this test, nor the verbal/non-verbal test showed any marked difference in the10 plus groups. The superiority of second and third year children on the verbal ability test in streamed schools could be a result of the school organisation, and is not necessarily a cause of the attainment differences found in the two types of school.

Finally, we would draw attention to something which is a feature of the streamed schools and which might have a bearing upon the test results - at least in a marginal way. Table 3 (Section 2) shows that children in streamed schools were more frequently tested, though not necessarily with standardised tests of the kind used in this study. It is thus possible that they were more 'test wise' or at least more habituated to test situations.

These cautions are reiterated to guard against too much emphasis being placed on the results of this part of the inquiry. As they stand they lend small support to the controversialists on either side. They indicate that organisational factors may well mask more important and pervasive influences and that these are perhaps less obscurely to be discerned in the facts reported in Section 2. They also indicate that the effects of teachers' beliefs and attitudes are to be looked for not so much perhaps in differences in formal attainments as in children's beliefs and feelings about themselves and their underlying attitudes to school and what it stands for. (12) It is not in the nature of cross-sectional evidence to throw much light upon questions such as this, though we may hope that something will emerge from the follow-through study. One point does, however, seem to be clear. A mere change in organisation - the abandonment of streaming, for example - unaccompanied by a serious attempt to change teachers' attitudes, beliefs and methods of teaching, is unlikely to make much difference either to attainments or - though this is less certainly based on the present evidence - to the quality of teacher-pupil relationships.

Section 4 The influence of teachers' methods and attitudes

4.1 Introduction

It has often been suggested that a teacher's effectiveness depends upon his attitudes more than upon his qualifications or even his length of experience. Thus, the individual teacher's attitudes towards streaming may be predicted to have some effect upon his pupils, modifying to some extent the effect of the school's policy of streaming or non-streaming.

This, it would appear from earlier sections, is probably one of the most crucial problems of the whole streaming research. Unfortunately, in the time at our disposal, it has not been possible to go into it as thoroughly as even this cross-sectional data warrants. What follows in this section is a partial examination of the problem. We have been able to examine certain sub-groups of teachers in non-streamed schools and to show some relationship between the teachers' attitudes and their pupils' performances. This section must be regarded as exploratory and is only a 'first look' at the data.

Two probes have been made, both based upon contrasting groups of teachers in non-streamed schools. The unstreamed schools were chosen simply because any effects which the views and practices of teachers might have upon the performance and personality variables of their pupils, would not be complicated or concealed by the differences in status of A, B and C streams. From among the 226 teachers of nine and 10 year old children, in the non-streamed schools, the 28 most convinced supporters of streaming and the 24 most convinced supporters of non-streaming were chosen. Certain of the comparisons which follow (Table 9) are based upon these two groups. It was also possible, since the 'obsessive'/'permissive' scale tended to correlate with the streaming/anti-streaming scale, to distinguish two sub-groups within the supporters and opponents of streaming. Eight teachers who strongly favoured streaming were also at the 'obsessional' extreme on the appropriate scale, while 13 of the teachers who favoured non-streaming were at the extreme 'permissive' end. The comparisons which could be made between the pupils taught by these teachers are shown in Table 9.

Time has not permitted, even with these limited samples, the examination of all the possibly worthwhile questions and what follows is a somewhat arbitrary selection from among those variables which showed statistically significant correlations with teacher attitudes.

It should also be remembered that any relationships found would in any case be small and not very clear-cut, simply because each class had been with its teacher for less than one year. This is not very long for that teacher's influence to modify the effects of home background and previous teaching.

4.2 Some of the effects of teachers' attitudes to streaming

Table 9, below, is based upon the two contrasting groups of teachers of third and fourth year children and upon the results of the attainment tests, a scale of 'test anxiety', a sociometric scale dealing with the number of friendless or isolated children in the class and one of the ratings made by the teacher concerned of the number of children considered a pleasure to have in the class. The data dealing with attainments are for the third year only (nine year olds) since two analyses could not be done in the time and it would have been meaningless to study nine and ten year olds together on this particular variable. All the other analyses are based upon the nine and ten year olds considered as one group.

So far as attainments are concerned, the results sustain and even reinforce the findings of the broad survey. The scores achieved by the pupils whose teachers are in favour of streaming on mechanical arithmetic and on English are higher than those of the anti-streamers - this is strikingly so in mechanical arithmetic. The pro-streamers in the non-streamed schools made frequent use of 'traditional lessons' in contrast to the convinced non-streamers, which suggests that attitudes to streaming are closely related to teachers' preferred types of lesson, rather perhaps than to the particular ethos of the school in which they find themselves. Their use of more formal methods resulted in higher mechanical arithmetic scores although, in fact, their pupils tended to come, on the average, from homes slightly lower on the socio-economic scale, than those in the classes of the non-streamers. The differences in the scores on the English test - which is somewhat less open to the influence of traditional types of teaching - are much smaller and not statistically significant.

In the rest of the table there is the suggestion that the views and methods of the teacher have - in an educational system where differences between the 'progressives' and the 'traditionalists' may be ideologically considerable without resulting in more than marginal modifications in actual practice - a more marked effect upon the pupils' attitudes to themselves and to their work. The additional analyses of anxiety scores seem to point to this. It was found that the two samples did not match perfectly for socio-economic status and to check whether this might have influenced the results, the scores of girls from socio-economic group 3 (parents in the skilled category) were analysed separately. This confirmed the tendency for girls in classes under a pro-streaming teacher to display more anxiety about being tested than similar girls under an anti-streaming teacher. Girls of below average ability seem to have reacted in a similar fashion under either kind of teacher whereas their more able sisters showed considerably more anxiety under pro-streamers.

We might tentatively sum up this part of the table by saying that it confirms the common view that girls are more anxious than boys (which may account for their higher scores on mechanical arithmetic) and more responsive to nuances in the views and practices of their teachers. The less able girls probably find any kind of test a source of anxiety, but the brighter ones find them more anxiety-provoking with a teacher who believes in competition as a spur. What as yet we do not know from this data is how higher or lower scores on the test anxiety scale are related to higher or lower performances on the various kinds of test - that is, whether the anxiety (rather crudely displayed and measured) acts as a spur or an inhibitor.

The last four lines of the table show some apparently slight differences. There seem to be a few more children perceived as 'isolates' in the classes of the pro-streamers, a difference of about four per cent in the aggregate. This, however, only seems small until we consider the possibly serious distress which might be felt by the friendless, isolated child. The final line probably reflects a combination of two factors: the rather sterner view of discipline taken by the pro-streamers and perhaps the greater tolerance of those who generally dislike drawing too many distinctions.

Table 9 The effects of teachers' attitudes to streaming - a study of teachers of third and fourth year juniors in non-streamed schools

4.3 'Permissive' and' obsessional' teachers

On the basis of extreme scores on the 'permissive'/'obsessional' scale, two very contrasting groups of teachers were selected from the pro- and anti-streaming groups. Thirteen highly 'permissive' anti-streamers were found and eight markedly 'obsessional' pro-streamers: between them they taught classes containing 326 children of nine and ten. The numbers are small and any differences have to be considerable to reach statistical significance. Table 10 summarises the results of the analyses which were made.

Table 10 Teachers pro-streaming and also holding 'obsessional' attitudes, compared with teachers anti-streaming and 'permissive'

From this table, the generally greater test anxiety shown by girls again emerges and this time is even more clearly shown to have a relationship to the teacher's views and methods. More striking is the fact that the 'obsessionals' disapprove of 'bad manners' - aspects of behaviour more common among boys - which is reflected in their tendency to rate proportionally more boys as difficult than do their 'permissive' colleagues. This in turn may go at least part way to explain why there are significantly more isolates among boys in their classes. It seems not unreasonable to suggest that the teachers probably manifest disapproval of certain boys and that this induces the class to reject them also.

One final, highly tentative hypothesis might be put forward. The 'permissive' teachers' classes did not achieve lower anxiety scores than those achieved in the anti-streaming teachers' classes generally - in fact the mean is slightly higher. This suggests that there may be a non-linear relationship between 'permissiveness' and anxiety, and that beyond a certain point of 'permissiveness', anxiety may in fact mount.

4.4 Conclusion

It must again be emphasised that in the nature of the data available, no very striking distinctions could be expected from the analyses of this section. It is the more interesting, therefore, that - particularly in the number of 'isolates' and in the test anxiety scores of girls - some fairly reliable differences have been found. The nature of these underlines the complexities of the problem under review and suggests that the more important and far reaching effects of streaming or non-streaming lie much less in formal attainments than in the human and personal aims of education. The emphasis, too, is shifted from the forms of organisation - the effects of which may be counteracted consciously or unconsciously by the teacher - to the teacher himself, his personality, views and attitudes, and consequent choice of methods. Organisation may perhaps facilitate or hinder the attainment of particular ends; the really important factor is undoubtedly the teacher who passes the whole day with an impressionable group of young children.

The following tables show the allocation to each stream of children having had six, seven, eight and nine terms of infant schooling.

Table 11 The length of infant schooling of children in different streams

There is a consistent trend for the A streams to have more children with nine terms infant schooling than would be expected by chance, and relatively fewer of those children who had spent a short time in the infants' school. For the lower streams, the trend is reversed. This tendency was found to persist even in the final school year: this accords with the low rate of transfer between streams that was noted in the survey report. The length of infants' schooling would appear to have had a lasting effect on the child's educational career, at least to the age of ten plus.

The differences between streams with regard to the length of infants' schooling are statistically highly significant (p<0.01) in both sizes of schools. It will be seen that the trend was more marked in the three and four stream schools. The remedial classes in the 100 school study, defined as 'the lowest stream combining two or more year groups and the E stream of the largest schools', showed the same tendency to a startling degree. Almost two fifths of the children in these classes had had only six terms infants' schooling. On the other hand, only 12 per cent with nine terms in the infants' school were found in a remedial class.

Table 12 The length of infant schooling of children in remedial classes

These results are in line with the findings of previous similar investigations. (13, 14, 15)

Some of the disadvantages of being a 'younger' child in a school year group might be removed if all children were allowed to start school at the same point in the year, thus removing the advantage of longer schooling which the 'older' child now has. It would be more logical for children to start school together at the beginning of the school year, since at every stage of the education system children born within the same year are treated as a unit (i.e. they move up the school together, take 11 plus at the same time, etc.). The fact that a quarter of the children in the sample had missed a full year's schooling as compared with their older classmates, does not seem compatible with the policy of treating them as one age group. A review of current practice may well be thought necessary.

5.5 Experience and age of the teacher

The teachers of the A streams tended to have more experience - and consequently were older - than the teachers of other streams.

Table 13 Experience and age of the teacher

This tendency is in line with the observations made by Jackson (16) in the 10 streamed schools that he studied. There is also a tendency throughout the schools for younger children to be taught by younger and less experienced teachers.

5.6 Facilities in the classroom

The data available to us were analysed to see if there were any tendencies for anyone stream to be given classrooms with better facilities than the others (desks, lighting, noise, space). For the most part, any differences were not significant statistically. There were, however, two exceptions. In the three-stream schools, the lower streams were significantly favoured with regard to floor space (p<0.01). This accords with the observed tendency for lower stream classes to be smaller. It was also found that, in the three-streamed schools, the lower streams were placed in significantly less well-lit classrooms (p<0.05). There seems to be no more plausible explanation for this difference than that, where differential conditions exist, these tend to be allocated to the detriment of the lower streams. Most schools would have fairly uniform facilities, but enough schools would have relatively ill-lit classrooms facing north for their allocation to constitute a measurable tendency.

5.7 Changes of teachers

Taking all the year groups together, there was no significant tendency for some streams to have suffered more than others from changes of teacher during the year. When the total variables affecting the fourth year, however, were subjected to a complete analysis, a small but significant correlation (r = 0.247) was found between type of class and the number of changes of teacher. There was in fact a slight but positive association between low ability stream and change of teachers. The correlation for the first year children was not significant. More detailed work must be done before this point is completely clear.

5.8 Teachers' attitudes

In two-streamed schools, some tendencies were found which might indicate differences of attitude among teachers of different streams. Greater differences were found in schools with three or more streams, where there is probably a greater awareness of the fact that the school is streamed.

(1) In two-streamed schools, there were relatively more highly 'permissive' teachers in B streams than in A streams. No such differences were found in the larger streamed schools.

(2) In the larger streamed schools, there was a consistent trend for teachers of higher streams to favour 11 plus selection more than teachers of lower streams. In the two-stream school there was a less pronounced tendency for relatively more B stream than A stream teachers to disapprove of selection.

(3) Teachers of A streams in large schools were the group who disapproved most strongly of noise; there were no consistent differences between teachers of middle and lower streams.

(4) In two-streamed schools, a slightly higher proportion of B stream teachers favoured streaming than A stream teachers (45 per cent and 40 per cent). In larger schools, more A stream teachers were favourable (58 per cent) than middle stream teachers (42 per cent) or lowest stream teachers (32 per cent). This may suggest that the status-hierarchy of children and teachers is much more apparent in large streamed schools than in two-streamed schools.

(5) In all streamed schools there was a definite tendency for A stream teachers to be the most favourable towards A streams. Teachers of the lowest streams were the least favourable (but not nearly as unfavourable as the teachers in non-streamed schools).

Results (4) and (5) both show the same trends among teachers of different ability streams. Teachers of A streams tend to favour streaming and to perceive favourable qualities in A stream classes. Teachers of the lowest streams tend to be neutral in attitudes to streaming and less favourable towards A stream children.

Section 6 Attitudes of parents

(In cooperation with the Government Social Survey)

6.1 Introduction

In the course of a more extensive survey of parental opinion 224 parents of children in the schools included in the present study were interviewed in autumn 1964. They were the parents of selected children in their second year at four large junior schools, situated in Chester, Leicester, Wellingborough and Manchester. The children had been selected from the sample in such a way that a range of social classes and all levels of ability were represented. Two socio-economic categories were distinguished-UPPER (non-manual and skilled manual occupations) and LOWER (semi-skilled and non-skilled manual occupations). Five categories of pupils' ability were distinguished, as measured by the NFER Primary Verbal Test.

Two of these schools were streamed and two were entirely non-streamed.

The parents were interviewed in their homes by staff of the COI, using an extensive pre-coded interview schedule. The COI had included a number of questions which were relevant to the streaming project. These have been analysed at the NFER, making use of our special information on the pupils, i.e. their streams, measured ability and parental occupation.

The findings which follow are tentative, since they are based on small samples, particularly when broken down into smaller sub-groups. Statistical tests have been applied where possible. These results should be interpreted with caution and regarded as providing suggestive rather than conclusive evidence.

6.2 Parents' knowledge of school organisation

Parents were asked: 'Do you know how the children of ...'s age are put into classes at (present school)? Are they put into classes by age, or do they put the quicker ones into one class and the slower ones into another?'

It was found that 85 per cent of parents of children in streamed schools could correctly identify the school organisation (i.e. streamed or non-streamed). In non-streamed schools, only 69 per cent of parents knew that streaming was not used; 17 per cent imagined that streaming by ability was used.

Table 14 Parents' knowledge of school organisation

6.3 Parents' attitudes to streaming and non-streaming

Parents were then asked: 'For children of ...'s age, could you tell me which of these you think is better? For the quicker and slower children to be mixed together in one class, or for the quicker children to be put in one class and the slower in another?' The first alternative, of course, refers to non-streaming and the second to streaming.

Streaming was the more commonly chosen alternative. In the streamed schools it was chosen by 79 per cent of parents; in the non-streamed schools by 65 per cent. (Streamed and non-streamed schools differed significantly, x² = 6.70, p<.01). There were six per cent of 'Don't knows' in each case. Less than one third of the parents of children in non-streamed schools approved of non-streaming.

Table 15 Percentages of parents preferring streaming and non-streaming

It seemed possible that parents in different social categories would differ in their attitudes. It was found, however, that this was not the case. The percentages preferring each method were almost identical.

Parents of B stream children appeared to be less favourable to streaming than either A stream or C stream parents; although this difference was not statistically significant (x² = 5.67 p<.06). Table 16 shows that 28 per cent of B stream parents preferred non-streaming, as compared with an overall 15 per cent. The sample is rather small, however.

Table 16 Percentages of parents preferring streaming and non-streaming (streamed schools only)

Next, a study was made of the attitudes of parents with children above average in ability, compared with those whose children were average and below. The cut-off point was taken as a standardised score of 105 on the primary verbal test. This made possible a comparison of parents' attitudes in streamed and non-streamed schools, regardless of the stream to which pupils had been allocated.

There appeared to be an association between the child's ability and his parents' attitudes to streaming. Parents of children with above average ability definitely tended to favour streaming; parents with less gifted children were much more likely to favour non-streaming. But even in the latter group, 'streamers' outnumbered 'non-streamers' by almost two to one.

Table 17 Percentages of parents preferring streaming and non-streaming - according to ability of child

The 49 parents who preferred non-streaming were asked: 'What do you think are the advantages of having the quicker and slower children mixed in one class?' The advantage most frequently mentioned was that it gives a sense of competition or spurs on the backward (this was mentioned by nine out of ten B stream parents who preferred non-streaming). Other advantages mentioned were that brighter children can help slower ones and that the feeling of differences between streams is avoided.

A similar question on advantages of streaming was put to the parents who preferred streaming. The percentages giving various responses are shown in Table 18, with separate figures for A, B, C streams and for two levels of ability in non-streamed schools.

Table 18 Perceived advantages of streaming by parents preferring streaming

As Table 18 is based on the responses to an open-ended question, it is difficult to interpret the data. The numbers are small, but some trends deserve comment. In non-streamed schools, parents of bright children tended to mention mainly the advantage to the bright child, and parents of less bright children were more likely to mention advantages which would accrue to the slower child in a streamed class.

6.4 Parents' preferences for secondary schools

'If you had the choice, what type of secondary school would you like ... to go to when he/she is 11?' In reply to this question 40 per cent of all parents chose the grammar school. Twenty-four per cent either had not thought about it yet or did not mind (the children were eight years old). Sixteen per cent chose technical schools, 13 per cent secondary modern schools, four per cent comprehensive schools, and three per cent other schools. None of the four areas where this research was carried out had a comprehensive school.

When asked what they thought would be the advantages of the chosen type of schools for their child, those who chose grammar schools tended to mention that it would lead to a good or better job, had good future prospects, and that there was a better standard of education and teaching.

Technical schools were thought to lead to a good job and to have a varied, interesting or wide curriculum. Those who chose secondary modern schools tended to mention mainly the varied, interesting or wide curriculum.

Parents were also asked if there was any type of secondary school which they would particularly dislike their child to go to. Seventy-seven per cent had no dislikes. The numbers mentioning a dislike are very small; suffice it to say that coeducational schools were first on the list, and secondary modern schools were second. The commonest reasons given for disliking these schools were that they contained rough, bad mannered or illiterate children and manifested bad discipline or behaviour or lax morals.

Let us now see how parents' choices were related to their children's abilities and to their socio-economic status.

In the sample, 35 per cent of children scored at or above 105 on the primary verbal test. A relatively high proportion of children from families in the upper social group obtained scores above 105, as compared with those in the lower social group. (See Table 19 below.)

Table 19 Percentages of pupils in each social group scoring above and below 105 on the primary verbal test

The preferences of each of the four sub-groups were studied separately (see Table 20). Seventy per cent of the upper social group with bright children hoped for a grammar school, but so did 40 per cent of the upper group and 26 per cent of the lower group with less gifted children.

Evidence that social class influences to some extent the choice of school is seen in Table 20. Sixteen per cent of parents of the lower social group with gifted children preferred the technical schools, which, while not chosen by parents in the upper group with bright children, were a popular choice among other parents.

The low proportion of parents actually choosing a secondary modern school bears no relation to the proportion whose children must eventually attend these schools. Again, choice of technical school was widely unrealistic when compared with availability.

Table 20 Percentages of parents choosing various types of secondary school

Summary of results

1. Streamed and non-streamed schools embody different philosophies (paragraph 2.8).

The streamed school seems to be more systematic in its approach, concentrates more on conventional lessons, gives more attention to the 3Rs and is likely to be more 'traditional'. Its staff is likely to be somewhat older and more experienced, to approve of A stream children, of 11 plus selection and of streaming. The non-streamed schools present an apparent contrast. Its younger teachers hold more 'permissive' views on such things as manners, noise and cleanliness; they disapprove of streaming and A streams, and 11 plus procedures. Their teaching tends to place more emphasis on self-expression, learning by discovery and practical experience.

Although different patterns emerge for streamed and non-streamed schools, not all teachers in each type conform to the predominant pattern. A minority in each type of school holds the views and opinions usually expressed by teachers in the other type of school.

Any effect which may be shown to be associated with streaming or non-streaming is unlikely to be purely and simply due to the form of organisation used. Attitudes of teachers and their teaching methods may well be the critical factors.

2. A straight comparison between streamed and non-streamed schools showed that pupils in the streamed schools had slightly higher mean scores on the attainment tests (NB But see conclusion, paragraph 3.4). The differences were greater the more the test reflected 'traditional' educational practices; they were largest for mechanical arithmetic and smallest for reading. It may happen when the effects of teachers' attitudes and beliefs are examined, that differences between the two types of school will disappear altogether. Or that they may become larger. The differences may also disappear when the two types of school are controlled for social class.

3. A study of two contrasting groups of teachers, supporters of streaming and supporters of non-streaming, was made in non-streamed schools. What can be discerned at this stage is that the teachers' attitudes are just as important or perhaps even more important than the type of school organisation. The teachers in non-streamed schools, believing in streaming, obtained a higher level of arithmetic computation from their pupils, while those believing in non-streaming had fewer social isolates in the class, caused less anxiety about tests and found more of their pupils a pleasure to have in the class. (Paragraph 4.4.)

4. A study was made of the characteristics of pupils allocated to, and teachers taking, the different ability streams in streamed schools. Month of birth and length of infants' schooling had a lasting effect on the educational career of children in streamed schools. (This effect is being examined in non-streamed schools.)

A steady decline in average age was seen as one proceeded from the A streams to the lower streams. Also, children in lower streams, especially in remedial classes, tended to have had a shorter period in the infants' school. (Paragraph 5.4.)

A streams tended to have the more experienced and, consequently, older teachers than other streams (paragraph 5.5). A stream teachers (in three or four-stream schools) were more in favour of streaming than teachers of other streams. (Paragraph 5.8.)

5. From a study of a small sample of parents of pupils in the two types of school, it was found that the majority of parents preferred streaming to non-streaming. (Paragraph 6.3.)

Of the parents who had a preference for a secondary school, the majority preferred their child to go to a grammar school. (Paragraph 6.4.)

Annex 1

Details of matched pairs of schools

Initially, 50 matched pairs were selected. By the time that all the tests and questionnaires had been completed, a number of schools had changed their organisation in some way which made them unrepresentative of streaming or non-streaming. They were therefore excluded from the analysis and it became necessary to rematch the remaining schools. This operation resulted in 42 matched pairs, giving a total of 84 schools for the cross-sectional study. It will be appreciated that when a number of pairs had lost one member, rematching would not produce pairs which were as good as the original ones. An effort has been made to obtain the best possible pairs, either by rematching or by elimination of schools which could not be rematched.

Criteria for matching schools

(a) Type of school: Junior or junior-with-infants; urban or rural. (All the schools were non-denominational and all were situated in England.) The results of matching on this criterion are shown in the table below.

(b) Number of classes in the school. All the schools had at least eight junior classes. The two members of a matched pair could differ in size by a maximum of one class.

The 42 streamed schools had a total of 437 classes;
The 42 non-streamed had a total of 434 classes.

The average number of pupils per class were:

In streamed schools: 35.90 (standard deviation 6.27)
In non-streamed schools: 35.25 (standard deviation 4.89)

(d) Geographical region: North, Midlands, South. The two members of a matched pair were both located in the same region. The numbers of schools in each region were:

(e) Percentage of children in the local authority who attended non-selective schools. It was felt that the forces tending to make a school concentrate on 'preparing for the 11 plus' might be related to the proportion of children admitted to selective schools. We therefore matched the schools on this criterion, allowing a margin of +/- 5 per cent.

(f) Predominant socio-economic status of parents. The head teachers were asked to estimate the percentage of parents in each of five categories: professional, clerical, skilled, semi-skilled, unskilled. This information was used to form four main categories of schools. All but three pairs of matched schools were both drawn from the same category, the number of schools in each being:

(g) Approximate matching of percentage of parents in professional, clerical and skilled occupations. As a further precaution, matching of schools was effected within each of these socio-economic categories. The members of a pair resembled one another, both having either more or less than 10 per cent of professional parents, 20 per cent of parents in the clerical category, and 20 per cent in the category of skilled workers. This was based on the head teacher's estimate. At a later date, information, given by class teachers, for each individual child was obtained. The table below is based on the latter, more accurate information. The streamed schools in the sample were from a slightly higher socio-economic background.

Table 21 Percentage of children in each socio-economic category, according to type of school organisation

Annex 2 Test reliabilities

Reliabilities of tests, calculated by the Kuder-Richardson Formula 20

Annex 3 Teachers' attitudes

Reproducibilities of Guttman Scales for Teachers' Attitudes (Questionnaire S.3)

These were calculated according to the formula proposed by Guttman. (See SA Stouffer et al. Measurement and Prediction, Princeton, 1950. Page 117.) Values obtained from two separate random samples of 100 teachers are given; also the minimal marginal reproducibilities (MMR), the number of items in each scale, and the number of trichotomous items.

Teachers' attitudes - Questionnaire S.3

1. Streaming makes slow children feel inferior.
2. There is too much emphasis on cutting down noise in schools.
3. A stream children have wider interests than other children, both inside and outside school.
4. I think a good slap in the right place at the right time does an awful lot of good.
5. Children in A streams tend to become conceited about their abilities.
6. Children must be taught to have decent manners.
7. I don't mind a reasonably high working noise in my class.
8. Bright children should not be streamed off from the rest of their age group.
9. The 11 plus exam is an entirely fair method of assessing a child's abilities.
10. Children in A streams worry too much about marks.
11. Physical punishment does no good at all to any child.
12. Bright children deserve a special academic course in a separate school when they are 11 years old.
13. An occasional hard slap does children no harm.
14. I cannot stand children fidgeting in class.
15. It is socially wrong to segregate children into streams.
16. The 11 plus exam can prevent slackness in junior schools and this is a good thing.
17. Teachers should demand clean hands in school.
18. I would not allow talking in a class of 35 or more children.
19. The bright children will be neglected in non-streamed classes.
20. Naturalness is more important than good manners in juniors.
21. An 11 plus exam is more fair than relying on record cards and teachers' assessments.
22. If children in my class are insolent, they have to be slapped.
23. Opportunities for self-expression through movement, painting and writing poetry, are more important than concentrating on the 'three Rs'.
24. In a streamed school, one gets far more done for the slow learner.
25. I'm quite prepared to spank bottoms for disobeying rules.
26. The atmosphere in A streams is too competitive.
27. Nothing worthwhile will be achieved by a class that talks while it works.
28. Non-streaming would be impossible with large classes of 40 or more.
29. Any exam that segregates children into separate schools at 11 plus is undesirable.
30. Physical punishment is out of the question and completely unnecessary.
31. A quiet atmosphere is the one best suited for all school work.
32. Without streaming, neither the bright nor the dull get the best from what the school could offer.
33. With non-streaming, I would find it impossible to keep duller children occupied while bright children received attention.

Footnotes

(1) The figure is probably nearer 70 per cent since many children in schools classified under Type 6 were in fact in homogeneous groups.

(2) Part I of this Appendix

(3) For further details, see STOUFFER, SA, GUTTMAN, L, et al. Measurement and Prediction, Princeton, NJ, 1950.

(4) For the meaning of this, see Section 2.7 and Annex 3.

(5) For the 'progressive lessons' scale see paragraph 2.2

(6) GUTTMAN, L. and SUCHMAN, EA, Intensity and a zero point for attitude analysis, American Social Review, 12, 57. 1947.

(7) See Part I of this Appendix.

(8) The cross-sectional study had to be mounted rapidly and there was not time to devise measures of the less conventional and traditional objectives of primary education. For the longitudinal study, the tests described above are being supplemented by others devised to be more in accord with the aims and philosophy expressed by the non-streamed schools. In particular, a new arithmetic test has been devised to assess children's understanding of mathematical principles and another concerned with creativity, is being constructed to redress somewhat the balance in the measures of attainment in English.

(9) Budgetary considerations forbade the use of this test for the younger age groups.

(10) MORRIS, J, Standards and Progress in Reading, NFER publication (to be published in 1966).

(11) GARDNER, DEM, Testing Results in the Infant School; Long term Results of the Infant School.

(12) Attitudes of children will be measured in the longitudinal study.

(13) JACKSON, B (1964) Streaming: an Education System in Miniature. London: Routledge and Kegan Paul.

(14) JINKS, PC (1964) An Investigation into the Effect of Date of Birth on Subsequent School Performance. Educational Research, Vol. VI, No.3.

(15) FREYMAN, R (1965) Further Evidence on the Effect of Date of Birth on Subsequent School Performance. Educational Research, Vol. VlII, No. 1.

(16) JACKSON, B (1964) Streaming: an Education System in Miniature. London: Routledge and Kegan Paul.