EDUCATION BUREAU
Study on Small Class Teaching in Primary Schools in Hong Kong
Final Report
December 2009
Maurice Galton and Tony Pell
List of Contents
SECTION PAGE
Executive Summary i-xi
Introduction 1
Background to the Study 2
Research measures used in the Study 7
The Consultant’s visits to the schools 9 Teacher professional development 11 The replication of teachers’ views on small classes 12 Further consideration of the observation data 18
Impact of small classes on the Territory-wide System
Assessment (TSA) 26
Factors contributing to school success 30
Relative performance of schools with high proportions
of disadvantaged pupils 35
Performance in small classes compared to Reference
Schools 41
Reviewing the results of the SCT Study (2004-08) 52
Some general conclusions 60
Recommendations 64
References 70
Appendixes 74
Executive Summary
I. The final year of the small class teaching study (SCT) has seen the completion of the fieldwork. During the year the P3 small classes from Cohort 1 (the 2004/05 P1 small class sample) returned to normal size classes in P4 and Cohort 2, having been in small P1 classes during 2005/06 and P2 small classes in 2006/07 returned to P3 normal size classes. A third cohort, who had started in P1 normal size classes in 2006/07 moved on to ‘normal’ classes in P2. It has thus been possible to compare small classes with normal size ones in the same schools over the first three years of the pupils’ primary schooling and to replicate these comparisons in P1 and P2.
II. The study sought to address two broad research questions:
What are the benefits of SCT in the local context?
What teaching strategies, professional support and resources are necessary in order to maximise the benefits of SCT in Hong Kong primary schools?
III. A matched sample of 15 schools (known as the reference schools) was added to the study in 2006/07. P1, P2 and P3 classes in these schools during 2006/07 moved up in 2007/08 to P2, P3 and P4 respectively, thus affording a further set of comparisons. Attainment tests of Chinese, English and mathematics and student questionnaires were administered. The Teacher Questionnaire, which solicited practitioners’ views on the merits of SCT during the second year of the study (2005/06), was repeated. In addition, an analysis of the Territory-wide System Assessment (TSA) scores was undertaken using the 2004 results to draw a matched sample of 37 schools to compare with the ‘experimental’
schools taking part in the SCT study. Comparisons were then made using the 2005, 2006 and 2007 results.
IV. In most years some 700 classes have been tested. This has resulted in approximately 20,000 pupils taking the end of P1 tests in each of the subjects, all of whom were followed for a second year. Nearly 23,500 took the end of the P2 tests in each subject, the majority having entered the study at the beginning of P1. The figure for the number of pupils taking the end of P3 test was 20,500 while just over 11,000 of these pupils continued to the end of P4. Of the total sample who started in P1 classes approximately 53.9% started in small classes, 27.2% came from the control classes in the same schools with the remaining 18.9% being drawn from the reference schools.
V. The consultant completed his visits to the participating schools and made a third visits to the 8 case study schools (the repeat sample). There continued to be considerable variation in the quality of instruction between schools and between teachers. About 25% of observed lessons showed real improvement.
The main weakness in the others concerned the inappropriate use of group work.
In too many group activities only one pupil at a time was actively engaged while other members watched. This sometimes led to poor behaviour because pupils became distracted.
VI. The Teachers Questionnaire was again administered during 2007/08. If anything, opinions on the perceived benefits of SCT were stronger than on the earlier occasion when the survey was distributed. It was thought impossible to give individuals extra help, cater for slow and gifted learners, match tasks to individual needs, mark pupils’ work during lessons and make use of across-the- age-group peer tutoring in normal classes. Holding extended class discussions, differentiating work by task and doing group work was thought to be difficult but possible when the number of pupils in the class was greater than 25, while setting practical tasks, giving oral feedback, doing pair work, involving pupils in assessing their work and using same-age-group peer tutoring was deemed equally feasible in both small and normal classes. These responses suggest many teachers now accept the project’s stance that most of the common teaching approaches can be used in both small and normal classes, although with more difficulty in the latter situation. As such it marks a change in opinion from the beginning of the study when teachers demanded that they should be informed about ‘SCT methods.’
VII. Teachers were also asked to what help they received in implementing SCT and which of these supports they most valued. Generally, as in the earlier survey, teachers most valued practical help with technology and from outside experts such as educational psychologists and social workers. Next important was the support of the Principal and the curriculum leader. Sharing good practice with colleagues was less valued because any useful knowledge gained could not be put to immediate good use on account of existing workloads.
VIII. Teachers ranked catering for diversity as of greatest concern. Providing extra tuition outside normal lessons and the use of same-age-group peer tutoring continues to be strongly favoured strategies for coping with this problem, but more use is now being made of setting differentiated tasks. As a result grouping practices have become more flexible with pupils placed in either mixed or ability groups according to the task demand.
IX. A cluster analysis of the teacher observation data yielded four distinct teaching approaches. The first of these, named sustained individual and pair enquiry consisted of teachers who asked the highest number of open, challenging questions made more statements of ideas and offered greater amounts of informing feedback. Their interactions with pupils took place either individually or in pairs and were more often sustained into the next 30 second time unit. The second approach, labeled group task monitoring, consisted of teachers who spent much time listening or watching pupils. These teachers issued the highest number of task directions which often involved setting up and organizing work in groups. The third approach, that of whole class instruction, represented a more traditional form of teaching, and consisted of teachers who spent nearly two thirds of the time talking to the whole class without any particular child in focus. A large proportion of this time was spent making statements of fact, giving directions and listening to pupils read in unison. The final approach also employed a high proportion of whole class teaching but the focus was on questioning of all kinds rather than statements.
Teachers in this cluster were labeled whole class questioners.
X. Across the three subjects the use of the different approaches tended to be more evenly distributed in small classes suggesting that many teachers were still experimenting with ways of maximizing the advantages of having to cope with fewer pupils. Normal classes were dominated by whole class instruction.
English, teachers too favoured this more traditional approach in small classes, while for mathematics in small classes pair work predominated. As pupils moved from P2 to P3 more group work was attempted. When classes were ranked in order of attainment and attitudes at P1 and P2 few difference emerged between the four teaching approaches. In P1 sustained individual and pair enquiry improved learning disposition in Chinese while in mathematics whole class questioning had a similar effect. In P2 pupils taught by whole class questioners in English made most progress. A similar breakdown could not be accomplished with P3 since only Cohort 1 classes were small ones and numbers did not allow an extended analysis of this kind. When the data was aggregated across the three subjects no differences in attainment were found but in P2 classes the use of whole class instruction produced the lowest levels of learning disposition while whole class questioning produced the highest.
XI. A similar analysis was carried out on the pupil observation data. Again 4 types of pupil were identified bearing close similarity to the patterns of behaviour exhibited by pupils in earlier studies of primary classrooms in England. Type 1 pupils were labeled solitary workers. They were on task for almost 95% of the lesson but rarely interacted either with the teacher or their peers. Type 2 were intermittent workers. These pupils also mainly worked on their own but tended to become distracted when the teacher was engaged elsewhere in the classroom.
Type 3 were also very industrious workers but were more often part of a group and so were labeled active collaborators. These pupils also concentrated on the task for 90% of the lesson. The final group of pupils with the lowest levels of on-task behaviour were called attention grabbers. They either sought or received more of the teachers’ undivided attention than any other pupil in the class, were more often moving around the classroom and received more praise than other pupils.
XII. When the distribution of pupil types was examined there were few differences between the size of the class or the year. Small classes and P3 classes had more active collaborators. P3 classes also had fewer attention grabbers. No differences emerged in attainment, attitudes or subjects. When the relationship between teaching approach and pupil types was examined whole class instructors had a significant higher proportion of solitary workers. These pupils while on-task were mostly listening to the teacher talking. The lack of any overall significant relationship between teaching and the patterns of pupil behaviour suggests that the latter may be a function of personality and temperament so that some attention grabbers may be shy, anxious introverts who seek constant reassurance from the teacher.
XIII. The impact of being in small classes on the TSA was also investigated. The 2004 results were used to select 37 other schools which were paired with those in the SCT study while also taking account of the social and economic characteristics of the school’s intake. Comparisons were then made between this matched sample, the experimental schools and the reference group. In the
three samples the proportion of pupils achieving basic competence in 2005 was below the average for all Hong Kong schools in all three subjects in P3. By 2006 this situation had been remedied except for English where only the reference schools achieved comparable results with the territory-wide average.
In 2007, the first year when the pupils in the experimental schools were in small P3 classes, there was little change compared to 2006 in any subject. Using the 2005 results to predict the 2007 competencies gave only one (mathematics) non-significant positive residual gain in favour of the experimental schools.
XIV. When the scores on the various sub-tests were analysed, no significant differences in favour of the experimental schools occurred in English or mathematics. In 2007 the experimental schools scored significantly below both the matched sample and the reference schools at individual listening in Chinese but did better on the second audio-visual information test. For these sub-tests a
‘light sampling’ procedure was adopted and this adds to the uncertainties in interpreting results. It would seem that main reason for improved scores was the teachers’ greater familiarity with the structure of the assessment since the greatest gains were made in 2006, the third year of the TSA operation.
XV. Some of the factors contributing to school success were examined by comparing the characteristics of the six most successful experimental schools (in terms of combined attitude and attainment profiles) with the four least successful. In the most successful schools Principals were more experienced, held firmer beliefs in the value of SCT for improving pupil attainment and took a more active role in curriculum and teacher learning development. They also placed greater emphasis on parental support and participation which was significantly higher in the more successful schools. There was a tendency for the more successful schools to have a greater proportion of individual and pair sustained enquirers and the less successful ones to have more group task monitors although these differences did not quite reach a statistically significant level.
XVI. The performance of the 5 schools in the experimental sample with relatively high proportions of disadvantaged pupils was examined. There was little difference in the attitudes and motivation of pupils irrespective of whether they were in schools with a high population of disadvantaged pupils or attended schools with a standard mix of pupils. In the first Cohort pupils from the disadvantaged schools held their own in both Chinese and mathematics attainment during the P1 and P2 years but lost ground in P3. In Cohort 2 pupils from the disadvantage schools did slightly better in both these two subjects in P1 and maintained this advantage in Chinese while achieving parity in mathematics during the P2 year. In P3 when Cohort 2 pupils moved back to normal classes those from disadvantaged backgrounds lost ground. Most of the deterioration in performance was due to girls. In English, where the pupils from disadvantaged backgrounds start from a low base, the position worsened year by year. The conclusion reached is that providing pupils from disadvantaged backgrounds are not too far behind on entry to primary school being in a small class is a positive advantage such that they maintain parity with pupils in schools with standard populations during the P1 year, although the magnitude of this effect decreases as pupils move up the school.
XVII. Further evidence in support of this conclusion comes when the progress of pupils in normal classes in the disadvantaged schools is compared with normal classes in the rest of the control schools. In both Chinese and mathematics pupils from the schools with a high proportion of disadvantaged pupils did worse in both P1 and P2 compared to the schools with a standard mix of pupils.
In P3, however, the pupils from disadvantaged backgrounds caught up which reversed the trend in the small class sample. In English scores continued to deteriorate year by year. Being in a normal class and attending a school with a high proportion of disadvantaged pupils had little effect on either motivation or attitudes when compared to normal classes in the remaining experimental schools.
XVIII. A series of comparisons was carried out between classes from the reference group of schools and those in the SCT study. For attitudes and motivation there was little difference between the various samples in P1 and P2. English is disliked most, mathematics least. In every case attitudes decline year by year.
The boys’ decline is sharpest in languages, the girls in mathematics. In the move to P3 the pattern is similar but in mathematics, boys’ attitudes in Cohort 1, which remained in small classes, showed the least decline. However this ground was lost on returning to normal classes in P4.
XIX. Comparing the attainment of pupils as they move through the P1 and P2 years produced similar patterns in both the reference and experimental schools. Any differences were mainly due to the differences at the start of P1 rather than being in a small class. In both school samples and in both languages boys make less progress relative to girls but in mathematics, although both girls and boys had approximately equal scores at the start of P1, they showed greater improvement by the end of P2. The superiority of the reference group at the start of primary school was mainly due to the lower percentage of pupils born in Mainland China. This has consequences in terms of parental support.
XX. In examining the changes from the beginning of P2 to the end of P3 the reference schools enjoyed an initial advantage because the pre-test was taken in September and not June. The reference group then maintained this advantage on the tests at the end of P2. The gender pattern in both experimental schools and reference schools was similar to that in P1 where girls outperformed boys in languages but boys were ahead in mathematics by the end of P3. When Cohorts 1 and 2 were compared the latter has an advantage in P2 and maintained this in P3 although the pupils had returned to normal classes. The final comparison compared Cohort 1’s move through P3 to the end of P4. The gender patterns partially replicated the earlier analyses with girls doing better in languages but girls maintaining parity with boys in mathematics. In Chinese, Cohort 1 pupils closed the gap on the reference schools (both genders) by the end of P3 but fell behind again at the end of P4 when they returned to normal classes. In English pupils in Cohort 1 were behind in every test but in mathematics they caught up the reference group at the end of P3 only to fall behind again at the end of P4. The inference from these results is that the attainment of pupils at the beginning of each school year was the prime determinant of pupils’ progress as they move from P1 to P4. In so far that there are fluctuations in the relative progress in Cohorts 1 and 2, compared to
the reference schools, these are likely to be random and mainly due to the expertise of teachers taking a particular class in any given year.
XXI. This assertion was confirmed by a series of regression analyses in which the start of P1, P2, and P3 scores were used to predict the end of P2, P3 and P4 scores. The start of P1 score predicted 42.4% of the explained variance up to the end of P2, the start of P2 explained 64.3% up to the end of P3, the start of P3 accounted for 75% up to the end of P4 while the end of P3 scores accounted for 79% of the explained variation. In none of these analyses did being a member of a small class result in a statistically significant regression coefficient.
As pupils move from year to year other factors such as place of birth and parental support assume less significance. The contribution to a pupil’s predicted score resulting in attendance at a particular school also decreased, for when a multilevel regression analysis was performed, with pupil variables and school as the two levels, schools accounted for some 8% of the variance by the end of P2 and P3 but only 4% by the end of P4. Moving from a small to a normal class in the P4 year causes the aggregated scores of pupils to fall by 1.9%.
XXII. The final analysis compared the 23 data sets collected for each subject over the lifetime of the project. Pupils in Cohort 2 who made more progress during the P1 year than the pupils in normal classes in Cohort 3 regress in P2 and do less well than either these control classes or the pupils in the reference schools. In P3, however, the year of the TSA examination and the year Cohort 2 pupils return to normal classes these pupils, who have experienced small classes for two years, regain lost ground. Cohort 1 pupils outperform pupils in the control classes in P3 but not in P1 or P2 and the return to large classes in the P4 year results in a dip in performance in comparison with both the control and reference school samples.
XXIII. Regression analysis confirmed these trends. The strongest predictor of the end of P2 attainment was the end of P1 score accounting for 67.4% of the explained variation. The start of P1 score contributed a further 3.7%. Using the end of P2 score to predict the P4 score accounts for 70.9% and using the end of P3 score just over 80%. When attainment is omitted from the analysis then only 14.2%
of the variance is explained and being in a small class does not contribute to this percentage. Using the multilevel regression model some 5% of the predicted end of P4 performance can be attributed to school differences. Such differences can be attributed in various degrees to factors such as the Principal’s active leadership, the proportion of staff with qualifications in the subject they are teaching, levels of parental support and the socio-economic characteristics of the surrounding neighbourhood from which a school draws its pupils.
XXIV. To see if the advantages of attending a particular school were greater in small classes compared to the controls the analysis was conducted separately for each condition. The P3 score accounted for 81.8% of the explained variance in the predicted P4 scores in the small classes and 77.9% in the controls. Girls obtained a bigger advantage in small classes and the active leadership of the Principal also made a significant contribution. Examining the school contribution, the same schools rarely contribute to both the small and the
control class analysis. One school made a positive contribution in small classes but a negative one in the controls. These results support the view that a major determinant of pupil progress is the expertise of the teacher who takes a particular class at a particular point in time. When all measures of attainment are excluded from the regression analysis 14.2% of the variation in the predicted P4 scores is explained in the control classes and about 8% can be attributed to school differences of the kind set out in the previous paragraph.
The corresponding figure for small classes is 27.1%. Thus while there is variation between schools in their effective use of SCT there is an even larger variation within schools. The most likely explanation of these within school differences is the individual teacher’s classroom expertise.
XXV. The above results allow a number of research questions which were posed at the start of the study to receive qualified answers. In all of the analyses undertaken, effect sizes were always small to very small. The central research question asked whether pupils in small classes make better progress than those in normal ones. The answer was a partial yes, in that pupils in Cohort 2 outperformed Cohort 3 in the P1 year. However this advantage was lost in P2 although these pupils regained lost ground during the P3 year by which time they had returned to normal classes. Being in Cohort 1 had no positive effect in either P1 or P2 but these pupils did better in P3 although this advantage was lost when they returned to normal classes in P4. In none of the comparisons did being in a small class bring about improvements in motivation, self esteem and attitudes towards Chinese, English or mathematics. The results are therefore inconclusive and in line with previous studies where class numbers are in the mid twenties.
XXVI. In so far there were gains in any year these seem to fluctuate randomly between the experimental classes, the controls and the reference group. This suggests that a major determinant of pupils’ progress is the expertise of the particular teacher who takes the class in any one year. This view is supported by the regression analyses which show that the end of year score in the previous year increasingly accounted for most of the predicted variance as pupils moved from P1 to P4. Another factor was the TSA test which pupils took in P3. In this year, pupils in both Cohort 1 and 2 made up previously lost ground on the reference classes although those in Cohort 2 were no longer in small classes. The inference here is that there is a certain amount of ‘teaching to the test’ which makes the P3 year results difficult to interpret.
XXVII. The second research question asked whether it is important to have a small class only in P1 or whether the effect of being in a small class for more than a year was additive. Being in a small class in P1 was a clear advantage for Cohort 2 but not Cohort 1, probably because teachers in the first year of the study were adapting to the new conditions. The answer to the second question appears to be a negative one since Cohort 2 did not maintain their advantage in terms of attainment when pupils moved to P2 and although Cohort 1 made attainment gains during the P3 year this is probably had more to do with preparation for the TSA tests. The benefits of being in a small class prior to returning to a normal one were not conclusive. Cohort 1 showed no benefit when moving to P4 but Cohort 2 improved on moving to P3. However, the effect of the TSA
again probably influenced this result. There is no consistent evidence therefore that being in the small class for three years was better than being in one for two.
XXVIII. The third question concerned the relationship between teaching and outcomes in the small classes. This has become particularly relevant given the decision to extend SCT to all public sector primary schools at the start of the 2009/10 academic year. Although four distinct teaching approaches were identified none appeared to offer an outright advantage in terms of attainment although there was some indication that a mix of individual and pair work that encouraged enquiry, combined with the use of whole class questioning improved pupils’
orientation to learning. Working in groups resulted in considerable off task behaviour which suggests that many teachers are still struggling to find ways of implementing this strategy in a satisfactory manner. Normal classes more often adopted a whole class approach where for over two-thirds of the lesson pupils either watched or listened to the teacher. Teachers taking part in the SCT study still appeared to be experimenting with different teaching approaches which, in itself, may account for the limited and inconsistent results in small classes.
XXIX. The fourth question concerned the effect of small classes on pupils from disadvantaged backgrounds. Here, there was evidence to suggest that being in a small class helps disadvantaged boys in particular to catch up in Chinese and maintain parity up to the end of P2. In mathematics the boys gradually outperform the girls and also maintain parity with the rest of the sample until the end of P2 but as with Chinese, any benefit gained from being in a small class is gradually eroded during subsequent P3 year. In a reverse of the general pattern it is the boys rather than the girls who benefit most from reductions in class size. However a condition of making progress is that the pupils must not be too far behind on entry to primary school. Thus in English, where the entry scores of both boys and girls in the schools with a high concentration of disadvantaged pupils were considerably below those of the remaining experimental sample, the gap widened year on year. When attitudes in the core subjects (Chinese, English and mathematics) were examined, both girls and boys scores in the disadvantaged schools showed few differences compared to those from the remaining schools in the experimental sample but girls’ attitudes in all three subjects declined at a faster rate.
XXX. The final question investigated the impact of the various contextual variables such as parental support and the leadership characteristics of the school Principal on the performance of pupils in the small classes. Teaching approaches did not appear to influence attainment except in mathematics where there was a very small effect size in favour of sustained enquiry and where pupils either interacted with the teacher individually or in pairs. Pupils who were taught by teachers using this sustained enquiry approach or where whole class teaching emphasized questioning rather than instruction had marginally better attitudes but again the effect sizes were very small. The effect of these contextual variables diminished as pupils moved up the primary school. Thus the advantage of attending a particular school was greatest in P1, as were the degree of parental support and the active leadership of the Principal. Comparing the most and least successful schools in the SCT study provided a further
distinguishing characteristic. Principals of successful schools supported their teachers by providing non contact time for professional development. In none of the various analyses did the teacher’s training, their experience, level of qualification or their experience of teaching small classes make a significant difference.
XXXI. As teachers’ involvement in the study increased over time different forms of professional support were required. Initially teachers required mentoring to help replicate the existing successful practice of other colleagues. Visits from members of the school support team acting in this capacity were therefore helpful. At the next stage teachers became more interested in curriculum tasks so that workshops by subject experts and visits to other schools to observe other teachers’ classrooms were most valuable. At the point where teachers began to take ownership for changing their classroom practice, rather than seeking prescriptions from outsiders, learning circles became the most important means of professional development.
XXXII. Teachers were generally satisfied with the technical support they received.
Their greatest need was more non contact time to engage in professional development activities and to make the necessary resources for group activities.
It was often the case that these resources could be obtained from the surrounding environment rather than having to be manufactured. Thus in mathematics pupils could identify and measure various shapes within the classroom rather than have the teacher produces several sets of cardboard rectangles, circles and triangles. More successful schools found ways of providing additional non contact time for professional development.
XXXIII. In summary, although the results in terms of outcomes, both attitudes and attainment were not conclusive, there were many positive findings. First, and most importantly, there was strong evidence from the cluster analysis that teachers were beginning to experiment in the use of different teaching approaches. While teachers in regular classes mostly used whole class instruction, teachers in small classes were asking more challenging questions, making use of pair and group work and engaging individual pupils in more sustained interactions. This has come about because over the period of the study teachers have taken ownership of these changes, having accepted that there is no specific repertoire of prescribed strategies that constitute ‘small class teaching’. Learning circles have played a key part in bringing about this transformation.
XXXIV. Second, as a result of these improvements teachers in small classes have been able to improve the performance of pupils more evenly across the whole ability range whereas in regular classes improvements tend to be concentrated within the more able group. Third, smaller classes appeared to have particular benefits for children in schools with a high proportion of disadvantaged pupils during the initial year of primary schooling. Fourth, teachers have felt more relaxed and enthusiastic when teaching a smaller class and fifth, SCT was shown to work best when Principals took an active part in curriculum development and in the teachers’ professional development. These results suggest that over time, extending SCT to all primary schools will result in substantial benefits providing these encouraging features continue to be developed.
XXXV. There are therefore a number of recommendations that follow on from these findings:
Principals need to be encouraged to play a more active part in the implementation of SCT so that they can frame the professional development of their staff around the six principles which underpinned the approach used in this study. Distributed leadership to school curriculum leaders should not therefore be conceived simply as delegation but as a collegial sharing of responsibilities.
Both inter-school sharing across subjects and intra-school sharing around pedagogic issues should be encouraged using the ‘learning circles’
approach which has been employed successfully in the SCT study.
Learning circles are important because they focus at any one time on a specific pedagogy, allow teachers to observe and evaluate each others classroom practice and thereby enhance the participants’ sense of professionality. This allows teachers to move from a position where they looked to others to tell them what they should do to become an effective small class practitioner to a point where they are prepared to take responsibility for developing appropriate pedagogies. Support for this initiative should be sought from appropriate members of staff in the University Departments of Education and should be coordinated by a core team. The aim should be to promote teaching for understanding by increased pupil talk and participation through extended whole class discussion, together with the increased use of pair and group work, thus easing the transition from the current emphasis on whole class instruction where pupils spend most of their time watching or listening to the teacher.
These initiatives should concentrate initially on improving the quality of questioning during whole class discussion and with the effective use of peer tutoring during pair work where more knowledgeable pupils are able to help slower learners, thereby allowing teachers to provide more individual attention. The more difficult task of implementing collaborative group work should be left till there is evidence that the use of more effective whole class teaching strategies have begun to take root.
In coping with diversity, particularly in schools with a high concentration of disadvantaged pupils where teachers claim that there is a wide spread of ability in many classes, Principals should ensure that those aspects of classroom organisation and pedagogy which international research has shown can bring about significant improvement in pupils’ attainment are in place. These include the development of flexible grouping strategies, so that teachers can sometimes concentrate on slower learners while the more advanced work independently, the use of peer tutoring and the creation of resources which relate as far as possible to the pupils’ daily lives. The biggest improvement would come about, however, if teachers could raise the expectations of these disadvantaged pupils and convince them that it is often lack of effort rather than ability which stops them from making progress. Coping with diversity in schools with high concentration of disadvantaged pupils is particularly important in the years after P1 where the initial gains are eroded. The situation with regard to English, particularly in these disadvantaged schools, needs to be
reviewed in order to overcome the deficit that many of these pupils bring with them on entry and from which they never recover.
Some thought should be given to timetabling. Lessons which encourage greater pupil participation are difficult to manage in a 35 minute session.
Many schools have introduced at least one double session per week while others now work to a 45- or 60-minute timetable. If pupils are to be encouraged to participate more actively and to be given more thinking time during discussion then 35 minutes may be too short a time in which to implement such practice effectively.
Principals should be encouraged to allocate teachers to a limited number of year groups so that they have opportunities to teach more than one class in any particular year group. In the SCT study many practitioners only taught a single small class in P1, P2 or P3. They therefore had no opportunity to revise a particular lesson until a year later so that progress was often slow. If ideas from the learning circles are to develop at a reasonable pace then teachers need opportunities to test them out in one lesson, reflect on the outcomes and then make appropriate changes with a parallel class. Having a smaller range of year groups to teach would also cut down on preparation and hence offer possibilities of reducing the existing workload.
University Departments of Education need to consider ways in which the findings from this study could be used to reinforce existing and future initial teacher education programmes on matters such as teaching for understanding, catering for diversity and helping pupils to develop as independent thinkers. In this way these novice teachers, with the support of their University tutors, can act as agents of change. At the level of initial teacher education, courses should perhaps concentrate, initially, on making students more effective ‘whole class’ practitioners leaving more sophisticated teaching approaches such as group work to a later stage of their professional development.
1. Introduction
1.1 This final year of the Small Class Teaching (SCT) study has seen the completion of the fieldwork. As this therefore constitutes the final report in the series, it may be helpful to readers if a brief outline of the main features of the research design together with a short summary of the findings of the previous three interim reports is included. The latter summaries will be placed in the appropriate sections of this report. During the 2007/08 school year the P3 small classes from Cohort 1 (the 2004/05 P1 small class sample) in the 37 experimental schools returned to normal size classes in P4. It is now possible to compare these classes with normal size classes from the same schools who were in P4 during the previous year (2006/07). These latter pupils have been followed since they were in P2 ‘normal’ classes at the start of the study.
1.2 Cohort 2, having been in small P1 classes during 2005/06 and P2 small classes in 2006/07 returned to P3 normal size classes in the 2007/08 school year. In addition, a third cohort, who had started in P1 normal size classes in 2006/07 moved on to ‘normal’ classes in P2. It has thus been possible to compare small classes with normal size ones in the same schools over the first three years of the pupils primary schooling and to replicate these comparisons in P1 and P2 using the second cohort. It has also been possible to measure the effect of three years’ experience of small classes on P4 comparisons (Cohort 1) and two years’ experience of small classes on P3 comparisons (Cohort 2).
1.3 A further matched sample of 15 schools (known as the reference schools) was added to the study in 2006/07. P1, P2 and P3 classes in these schools during 2006/07 moved up in the 2007/08 school year to P2, P3 and P4 respectively, thus affording a further set of comparisons. The complete administrative plan for affording the various comparisons outlined in the previous paragraphs is included at Appendix I. No further observations were carried out during this final year of the study, but the Teacher Questionnaire, which solicited practitioners’ views on the merits of SCT during the second year of the study (2005/06), was repeated. In addition, an analysis of the Territory-wide System Assessment (TSA) scores was undertaken using the 2004 results to draw a matched sample of 37 schools to compare with the ‘experimental’ schools taking part in the SCT study. Comparisons were then made using the 2005, 2006 and 2007 results.
1.4 To summarise this report attempts to answer the following questions:
• Do pupils in small classes make better progress in attainment and improved attitudes and motivation compared to pupils in normal size classes?
• Is it more important to have a small class in P1? Are any attainment and attitude changes cumulative so that pupils with three years experience of small classes do better than those with just two years? Are there any effects of Cohort 1 pupils moving back to regular classes in P4?
• Are any changes in attainment and attitudes associated with particular kinds of teacher and pupil behaviours?
• Do small classes confer additional benefits on schools with an above average number of pupils from disadvantaged backgrounds?
What other possible factors such as parental support, school leadership, teachers’ background and opinions influence classroom behaviour and performance in the ‘experimental’ schools participating in the SCT study?
2. Background to the Study
2.1 The SCT study began at a point in time when there was considerable controversy surrounding the issue of reducing class size. Demographic changes in some areas of Hong Kong SAR led to a fall in the school population with the possibility of teacher redundancies. There was suggestion that rather than lose experienced teachers the opportunity should be taken to reduce the number of pupils in ‘normal’ classes to further enhance the quality of teaching and learning. In response to this appeal the then Education and Manpower Bureau (EMB) commissioned a SCT study. The broad aims of this programme of research were twofold; first to ascertain the benefits of SCT in the local context and second, to determine what teaching strategies, professional support and resources were necessary to maximise the effectiveness of SCT in Hong Kong’s primary schools.
2.2 Not surprisingly, the decision to set up the SCT study was seen by some as a delaying tactic on the part of the EMB. The research of Blatchford et al (2002, 2003a) in England and that of the Tennessee ‘STAR’ (Student Teacher Achievement Ratios) project (Finn& Achilles 1999) was frequently cited in support of the claim that pupils did better in smaller classes. However, a more complete reading of this and other research suggested that the above conclusion was an over-simplification and that there were valid reasons why further research in the local context was required. A review by Hattie (2005) has summarised the results of some 164 studies and found that the average effect size of small class was 0.12 which is relatively small and according to Hattie represented about 9% improvement on the average pupil’s performance in a large class. In another American study, the Wisconsin Student Achievement Guarantee in Education (SAGE) the estimated effect size due to smaller classes was 0.20 against an overall annual improvement of 1.20 (Molnar et al 1999).
2.3 Hattie goes on to discuss a number of critical comments concerning research into class size. Hanushek (1999) for example, points out that there was a high attrition rate in the STAR classes so that less than 50% of the students remained in the study by the time they entered the fourth year of schooling, although an analysis which controlled for this factor found it resulted in slightly bigger effect sizes (Nye et al 1999).
2.4 More relevant to the present study was the fact that in both the STAR and SAGE studies a small class consisted of 15 pupils, whereas in Hong Kong the numbers range from 20 to 25. Blatchford et al (2002) found that significant gains in test scores only occurred when the class size was less than 18 although they argue (Blatchford et al 2003) that classes of less than 25 pupils can still be important, particularly for those pupils from poorer socio-
economic background who generally enter school with few academic accomplishments and therefore need to catch up.
2.5 Most studies conclude that the benefits accruing from smaller classes follow the ‘law of diminishing returns,’ in that the effects are greatest with the youngest classes (Tomlinson 1990; Blatchford et al 2002). Nevertheless, Finn et al (2005) claim there are lasting benefits resulting from early exposure to small classes including better behaviour, increased engagement and higher levels of graduation once the pupils return to regular classes.
2.6 A further argument for setting up the SCT study arises from the fact that most of the studies reviewed by Hattie (2005) took place in Western countries where regular classes are in the range 25-30. Fuller (1987) found 9 studies which had been conducted in developing countries where the typical regular class consisted of 44 pupils and found no differences were attributable to class size. More recently, an analysis of the TIMSS (Third International Mathematics and Science Study) found only four out of 40 countries where class size made a difference (cited in Hattie 2005). These researchers conclude that effective teachers in these developing countries have created a range of teaching strategies that ensure learning takes place in larger classes and that these techniques may not be so appropriate when the numbers of pupils are reduced somewhat.
2.7 Moreover, the general explanation for the gains in small classes in Western countries is that there is less disruption and more pupil engagement in learning.
Finn et al. (2003) argue that it is mainly the differences in engaged time that differentiate between small and large classes. Averaging the results of various studies there appears to be an improvement of around 13.9% in the ‘on-task’
behaviour of pupils in small classes compared to those in larger ones where the average recorded was 56.6% (ibid: 327). In the UK levels of engaged time typically range from the mid sixties to the low seventies as a percentage of all observations. Clearly it would seem valid to suggest that if teachers spend around a third of their time not teaching, and this figure is reduced by around 10% in smaller classes, then echoing Harnischfeger and Wiley’s (1976) conclusion, time on task is likely to be a major determinant of pupils’
improving progress. In Hong Kong, however, three years of observation has regularly recorded values of engaged time of around 90%, irrespective of where the class is small or of regular size and this seems common to many countries around the Pacific Rim. Moreover, as Bray (2000) has shown, many children in these countries receive after-school coaching which adds to engagement and complicates the analysis further. If Hong Kong pupils can be shown to benefit academically from the experience of smaller classes it is unlikely therefore that Finn et al’s (2003) explanation of increased engagement will suffix in the local context.
2.8 Hattie (2005) considers two further possible explanations of why small classes might lead to improvements in academic performance (or rather why such classes don’t yield bigger effect sizes). The first of these examines the proposition often put forward by practitioners that in smaller classes more attention can be given to individual pupils. Few studies of class size have
included systematic observation of pupils and teachers, a notable exception being that of Blatchford et al (2005) who concluded that more one to one teaching took place in smaller classes with more pupil-teacher initiations. On the other hand Evertson and Folger (1989) found no statistically significant differences between small and large classes in the STAR project although pupils in the smaller classes tended to initiate exchanges with the teacher more often in seeking help. Other studies in Canada (Shapson et al. 1980) and in Australia (Bourke 1986) have also found little differences in teachers’ and pupils’ behaviour patterns across small and regular size classes. These results are not surprising when the mathematics of class size reduction on individualisation is considered. Buckingham (2003) has pointed out from a review of these observational studies that, typically, about 50% of each lesson consists of teachers directly addressing the class. In a typical 40 minute lesson in a class of 40 pupils then each individual would receive 30 seconds attention if the remaining half of the lesson was individualised and the teacher distributed her attention evenly around the class. Reducing the class to 25 pupils would increase the figure to 48 seconds. Thus under the most optimistic set of criteria a pupil would receive one and a half minute’s additional attention during the course of a school day.
2.9 This brings Hattie (2005) to the main reason why so many studies produce such small differences when small classes are compared to regular size ones.
He points to the mass of evidence to suggest that despite a world wide reform movement designed to change the way that ‘teachers teach,’ the dominant pedagogy still consists of mainly of teacher talk (most of it to the whole class) interspersed with some rapid questioning which according to Edwards and Mercer (1989) takes the form of ‘cued elicitations’ (where teachers offer further help in their reply to a pupil’s initial response as a way of guaranteeing a correct answer) followed by short bouts of group or pair work. Yair (2000) estimates that despite various reforms, some well intentioned, about 80% of American teachers continue with this didactic approach, 15% attempt to change but fail, leaving only 5% who are able to innovate successfully. These figures tend to support the historical analysis of teaching since the mid nineteenth century by Cuban and Tyack (1995) which argues that teaching has always been a conservative profession. What is true of teaching in general is therefore also likely to be true of the particular case of class size reduction.
2.10 However, the above conclusion does not necessarily vindicate the arguments of those who suggest that the money spent on reducing the size of classes could be spent to better advantage elsewhere in the educational system. An international study of teachers’ work lives (Galton & MacBeath 2008) has shown that most practitioners, worldwide, work long hours (including most evenings and weekends) are under considerable personal stress and feel that their family life has deteriorated over the past five years. Hong Kong teachers typically work some 10 hours more per week than do their colleagues in Western countries. Among the factors which could help alleviate this stress most teachers cite a reduction in the size of classes for the obvious reason that it would reduce marking and administration time, the main reason for having to take work home in the evenings and at weekends. However, now that beginning with the 2009/10 school year class sizes will be reduced in most
public sector primary schools (following the announcement by the Chief Executive in October 2007) it is surely important that teachers give the research results serious consideration in looking for ways of maximising the advantages of smaller classes.
2.11 Maximising the advantages of smaller classes so that effect sizes improve must involve changes in the way that teachers teach. This is certainly the premise that has guided the present SCT study. Faced with a lack of specific empirical evidence that distinguishes between less successful small classes and more effective ones it has been necessary to turn to the more general studies of classroom practice; to seek out those aspects of teaching that correlate most positively with pupil attainment. Hattie (2005) takes a similar view and has conducted an analysis of some 500 meta-analyses involving a total of around 300,000 individual research studies. Taking this analysis into account, the SCT study has been guided in its approach by the following six broad pedagogic principles aimed at developing pupils’ understanding:
• A clear statement of the learning objectives which are communicated to pupils as an ‘advanced organiser’ at the beginning of the lesson and reviewed at the end to assess if students have mastered what has been taught. The aim is to move away from stating objectives in terms of outcomes (e.g. to count to 100 in tens, or to learn the story of the ‘Greedy Cat’) and instead to focus on the purpose of the learning and the processes through which it is to be achieved, using key phrases such as, to understand, to compare, to identify etc. An analysis of this kind helps to focus the teacher’s attention on the most appropriate form of pedagogy for achieving such objectives.
• The use of extended questioning techniques which allows thoughtful discourse (Brophy 2004) to occur during dialogic whole class discussion (Alexander 2008a). This involves the use of more open questions, and allowing pupils more ‘thinking time’ between answers.
• Increasing pupil participation, particularly during class discussion.
Whenever possible therefore, investigation of pupils’ ideas precedes instruction. Thus in an English lesson on ‘shops’ the teacher might first ask the class about local shops, what they like to buy etc., before putting up flash cards with the sentences, “This is a toy (sweet, cake) shop.”
• Developing a spirit of cooperation between pupils through the use of group and pair work. For this to result in effective learning it is necessary for pupils to formulate, with the help of the teacher, their own rules for working in groups and to be trained in ways of maintaining group activity and of arriving at decisions through consensus rather than by majority vote (Kutnick et al 2005). An important element in the training is to provide opportunities for groups to evaluate how well they worked together and to think about how they might improve their collaboration on future occasions. This can be undertaken as part of a general class debriefing or by getting pupils to engage in some form of self-evaluation activity.
• Providing appropriate feedback which allows pupils to reflect on their work and if possible to discover errors for themselves and to correct them.
This kind of informing feedback is designed to help pupils develop strategies for problem solving and for developing understanding so that eventually they are able to self-regulate their own learning and become what Galton (2007) has termed metacognitively wise. It contrasts with corrective forms of feedback where the teacher identifies the error for the pupil and then demonstrates the correct solution. There is evidence that although corrective feedback is useful when the lesson is focused on what Biggs (1994) and Biggs and Collis (1982) termed surface rather than deep learning, the feedback activities associated with self-regulation (providing cues, reinforcing pupils sense of self-efficacy etc) result in massive effect sizes of the order of 0.90 compared to corrective feedback with more modest values of 0.37 (Hattie and Timperley 2007). In particular, the use of praise for getting a correct answer only results in an effect size of 0.14 so that it is important when praising students to relate it specifically to some particular quality of the pupil’s response and to remember also to praise pupils whose answer may be inappropriate but who nevertheless have made a determined effort to arrive at an acceptable solution. Praise for effort can have a powerful motivational effect.
• Creating a framework for assessment for learning which follows the ideas developed by Black and Wiliam (1998). In this approach tests and worksheets are not the sole means by which teachers determine whether a pupil has mastered a topic. When moving from teaching for transmission to teaching for understanding (Good and Brophy 2002) it is equally important to probe the thought processes of pupils as they carry out the set task. Thus debriefing students, listening to their explanations, getting students to identify when they need help etc. are all important assessment tools and link closely with the kinds of informative feedback discussed previously (Black et al. 2003).
2.12 It is important to stress that these are but guiding principles. How the teacher puts these principles into practice will vary according to the specific contexts operating in the particular classroom such as the physical circumstances, the qualities of the pupils etc. This is in accordance with Gage’s (1978) definition of pedagogy as the science of the art of teaching. The principles of effective teaching derive from our knowledge of how human beings learn but the art of teaching, putting these principles into practice is built on teachers’ experience;
what is usually termed craft knowledge. Thus Hong Kong teachers in this study have developed several ways of increasing pupil participation, quite unlike those seen in Western classrooms including a practice known as ‘daily talk’ where the teacher and pupils greet each other at the start of the lesson using the target language that was taught on the previous occasion.
2.13 In concluding this section therefore it should by now be clear that the initial request by teachers that the then EMB provide a prescriptive list of approved SCT methods was a distraction from the task that lay ahead of all involved in the SCT study. The principles of effective teaching are the same in classes of all sizes. The advantage of a smaller class is that it is easier to put these
principles into practice and to monitor their effect on pupils. The task of the then EMB school support team has been to work alongside teachers in developing ways that enable these principles of teaching to operate within the local culture.
3. Research measures used in the Study
3.1 The design of the study was such that it involved three levels of analysis; that of the school, the class and the pupil. An overview of the various research instruments used is shown in Table 3.1.
Table 3.1 Overview of the measures used in the SCT Study
School Level Class level Pupil Level
Population characteristics
• Parents’ survey
• Head’s survey
• Year Group
• Subject
Teacher characteristics
• Gender
• Experience
• Qualifications
• Training
• Subject specialism
• Survey of opinions Class size
Observation
• Questions asked
• Statements made
• Feedback given
• Whole class
• Groups/pairs
• Individual
• Sustained
Pupil characteristics
• Gender
• Age
• Place of birth
• S.E.S.
Outcome measures
• Attainment
• Attitudes
• Motivation
• Self-esteem Observations
• Time on task
• Pupil-pupil talk
• Pupil-teacher talk
• Target’s setting
3.2 At school level a questionnaire was administered to the parents of both P1 cohorts and to parents from the reference schools. It asked for information about such matters as place of birth of their children, level of education, type of employment, salary, accommodation and also explored the ways that they supported their children’s learning outside school (extra tuition, helping with homework, making educational visits etc.). In each administration response rates were high (around 90%) and, not surprisingly the degree of support given to a child outside school was highly correlated with family background and income. A factor analysis provided an overall measure of parental support with an internal consistency of 0.70. The scale had 11 items with loadings greater than 0.65 principally concerning the educational level of parents, their joint income, and the home resources available to support the child’s learning, the frequency of educational visits and the parents’ involvement in school
activities. Headteachers were also given a questionnaire, similar to the ones used to investigate teachers’ opinions about class size but which also explored the strategy for implementing SCT in the particular school.
3.3 At class level, background data on the teachers’ qualifications and training was collected from official records. Teachers’ opinions about SCT were solicited using a questionnaire based on earlier studies by Bennett (1996) and Blatchford (2003). This was first administered in 2005/06 and repeated in the 2007/08 school year. For the second administration a number of items were added that attempted to explore how far teachers felt that some of the practices developed for use in smaller classes were also applicable in normal size ones.
In each of the first 3 years of the study a sample of teachers (around 100 in any year) were observed using a modified version of the two original ORACLE (Observational Research and Classroom Evaluation) Project systematic observation instruments (Galton et al 1980). This codes teachers’
and pupils’ behaviour at 30 second intervals during the lesson. Members of the then EMB research team were trained to use the instrument by the consultant.
A detailed version of the observation schedule is contained in the Appendix II.
3.4 At pupil level details of the pupils’ background were again obtained from official records. Attainment tests were specially constructed by the then EMB.
In Chinese and English there were sections on listening, reading and writing.
Items were mainly of a multiple choice variety. The pupils’ attitudes to Chinese, English and mathematics and the motivation and self-esteem measures were based on instruments used by Galton et al (2003) in studies of pupils transferring from primary to secondary school in England. Despite initial fears of using ‘Likert’ type 5 point scale questionnaires with younger pupils (For the start and end of P1 a 3-point scale was adopted and the number of questions reduced) the scales had high internal consistency (α> 0.90 in most cases) and showed remarkable stability from year to year. An exception in P1 was the motivation scale (Cronbach α =0.65) so that motivation and self esteem were merged into a single scale and this practice was retained in subsequent year groups for the sake of continuity. Observations on pupils were also carried out at the same time that the data on teachers was collected. In each class a stratified sample (3 boys and 3 girls) of six ‘target’ pupils was randomly selected for observation.
3.5 The quantity of data available for analysis was considerable. In most years some 700 classes were tested. This resulted in approximately 20,000 pupils taking the end of P1 tests in each of the subjects, all of whom were followed for a second year. Nearly 23,500 took the end of the P2 tests in each subject, the majority having entered the study at the beginning of P1. The corresponding figure for the number of pupils taking the end of P3 test was 20,500 while just over 11,000 of these pupils continued to the end of P4. Of the total sample who started in P1 classes, approximately 53.9% started in small classes, 27.2% came from the control classes in the same schools with the remaining 18.9% being drawn from the reference schools. Full details of the numbers of pupils involved in the various tests can be found later in a later section of this report (Table 12.1).
4. The Consultant’s visit to the schools
4.1 By the end of 2006/07 academic year the consultant had visited all but one school in the SCT study and observed approximately equal numbers of Chinese, mathematics and English classes. This was in addition to the systematic observations carried out by members of the then EMB research team. In the first year of the study it was decided that the consultant should also make an annual visit to 8 of the 37 experimental schools in attempt to monitor changes over the lifetime of the study. These eight so called ‘sample schools’ were selected mainly on the basis of their relative academic performance during the first year of the study supplemented by advice from the school support team who had detailed knowledge of the schools in question. The final visit to the sample schools took place in November 2007 when the remaining unseen school was also visited.
4.2 In the event it was only possible to follow the progress of relatively few teachers over the three year period. This was because some Principals adopted a policy of giving as many of the staff as possible experience of small classes so that a teacher observed in one year would have moved back to teaching normal classes in subsequent years. In other cases moves to different year groups also reduced the opportunities of seeing the same teachers on more than one occasion. For these reasons it was difficult to form a considered judgement about the changes in classroom practice that may have taken place over the lifetime of the project.
4.3 The general impression gained, which was confirmed by the more detailed systematic observation during the course of the study, was that there was considerable variation across schools and between teachers within the same school. In one mathematics lesson, for example, P2 groups of pupils in a small class were asked to classify different objects according to their shape (circle, square, rectangle etc). The groups worked badly because only one pupil explored the characteristics of the object at a time (rolling or sliding) while others had to watch. This led to excessive noise and a tendency to grab at the objects in order to get a turn. In reporting back it became clear that pupils had not understood the purpose of the lesson so the teacher then demonstrated what they should have done.
4.4 In the previous year the same teacher had been observed teaching the concept of ‘reversibility’ to another small P2 class. Although this was the first time, according to the teacher, that they had encountered this idea no attempt was made to explore with the use of counters or other objects, the difference between ten lots of 5 and 5 lots of ten etc. Instead, pupils worked in pairs on a worksheet consisting of a series of simple calculations with one pupil working out one arrangement (e.g. 5 x 2) and writing the answer on the worksheet while the other watched. The second pupil then did the alternative calculation (e.g. 2 x 5). Again the fact that pupils had to wait for a turn led to a certain degree of disruption. When reporting back pupils again showed they had little understanding of what they had been asked to do or why they had been required to do it.
4.5 This teacher seemed to have benefited little from the advice given by the school support team. She had not established rules for working collaboratively, did not debrief the class about how well (or badly) they had worked together.
Although it was established early on in the study that writing and similar tasks were best done individually with results then shared, thus avoiding some pupils having little to do in the groups, this teacher persisted with the opposite approach. At interview she appeared to have very low expectations of these pupils claiming as a reason for not instituting peer tutoring that all but two pupils in the class were still at beginning P1 level.
4.6 In contrast in a P1 English class with 27 pupils the teachers used a pencil box to teach the target language, “This is my ruler,” and “These are my pencils”
etc. This was done through class discussion. Each pupil then prepared worksheet on which they had to choose a name for the child portrayed in the pictures. They then played a game, in pairs, using cards with pictures of pens, pencils, ruler etc. One pupil had to select a card and asks, “Is this your/Are these your ruler/pencils etc?” as appropriate and the other then replied using the correct construction which the first pupil checked from the card. Then each pupil consolidated this learning by drawing a circle around the correct construction on their worksheet. The pair then swapped roles. At the end of the lesson pairs of pupils came out to the front of the class and acted out the sequence.
4.7 This lesson was in sharp contrast to the one seen in the previous year where a small P3 class spent most of their time constructing a birthday card in groups mainly by cutting out pictures from magazines and sticking them onto the card.
Since only one of the groups was actively mounting a picture at any one point in time most of the class had found themselves with little to do but sit and watch for long periods. In the second year of observation, however, this teacher was much more concerned to utilise time profitably. Thus it took time for the teacher to distribute the packs of cards to the pairs of pupils but while she was doing this the class sang a song in unison about what was in their pencil cases, thus again reinforcing the initial learning. Overall, many of the suggestions made by the school support team had been taken up. The rules for group work were clearly displayed, the target language was situated within a familiar context (i.e. the pupils’ pencil cases) so that the spoken words were more meaningful and during pair work both pupils were fully involved, while writing tasks were undertaken individually.
4.8 A rough estimate would suggest that only about 25% of the lessons that were seen achieved this standard. As on previous visits lessons were rated on a five point scale in terms of the six principles. Last year’s overall rating in the 8 repeat sample schools averaged between 2.0 and 2.5. This year the corresponding figure was 2.75. The main weaknesses in the other lessons, as it was in 2006/07, mainly concerned inappropriate use of groups, particularly in cases where only one pupil at a time had an active role so that the remaining members sat and watched. This situation usually led to inattention and sometimes misbehaviour on the part of the passive pupils. Part of the problem of group work was that in a 35 minute lesson there was little time for debriefing pupils and evaluating how well they had worked. In some schools
teachers normally used groups in ‘double’ lessons but felt it necessary to attempt a group task because the consultant was present. Thus what was seen may on occasions have represented atypical practice.
4.9 There were however other factors which distinguished between poor and satisfactory lessons. In the latter case nearly all the teachers had attended the learning circles and said at interview that they had benefited considerably as a result of the experience, because as one teacher remarked working with colleagues from other schools had, “made me rethink my practice.” One factor associated with the less satisfactory lessons was the low expectations of some teachers had of the pupils, as exemplified in the earlier example of the P2 mathematics lesson on reversibility. This negative view of the pupils’
ability was frequently used to justify rigid adherence to the textbook so that often the work set was not sufficiently challenging with a consequence that pupils lost interest, thus appearing to confirm the teacher’s assessment that
“these children were unable to concentrate” for any length of time.
5. Teacher professional development
5.1 In the 2007/08 school year most of the efforts by the school support team have been devoted to the schools joining the scheme for schools with a high proportion of disadvantaged pupils. Teachers from the study schools, along with those from the scheme, did attend a half-day seminar in December in 2007 when the consultant reported on the results from the third year of fieldwork and gave a presentation on maximising the benefits of SCT to cater for pupils’ learning diversity.
5.2 Throughout the lifetime of the study the school support team have provided substantial professional support to the participating schools. In the 2004/05 school year this mainly consisted of workshops on joint lesson planning and on the principles and practice of cooperative learning and individual school visits to clarify issues and to identify the various kinds of support required. In the spring of 2005 a group of teachers from participating schools visited small classes in Shanghai. In 2005/06 there was a shift of strategy away from formal seminars towards the development of learning circles in which teachers played a more active part, although a number of on-site workshops were delivered at the request of schools and the study tour to Shanghai schools was repeated in April 2006. During the year 8 learning circles were established in which 26 schools participated. In 2006/07 this programme was extended and a variation of the learning circles format, collaborative lesson planning circles introduced where teachers planned lessons jointly, observed each other’s teaching and then evaluated the results. By the end of the year there were 15 learning circles and 15 lesson planning circles were in existence. Outside expert facilitators were brought in to support the work of these circles on topics chosen by the teachers. In addition to these activities members of the school support team continued individual school visits to discuss progress and assess needs.
