Changing assessment: washback on teacher perceptions ... - CiteSeerX

Teaching and Teacher Education 15 (1999) 253—271

Changing assessment: washback on teacher perceptions and actions Liying Cheng* Educational Psychology, Faculty of Education, 6-110h Education Building North, University of Alberta, Edmonton, Canada T6G 2G5 Received 11 August 1997; received in revised form 30 June 1998; accepted 2 July 1998

Abstract Public examinations are often used as instruments and targets of control in the school system. A belief that assessment can leverage educational change has often led to top-down educational reform strategies. In 1993, the Hong Kong Examinations Authority introduced major changes into its existing Hong Kong Certificate of Education Examination in English with the intention of bringing about a positive washback effect on classroom teaching. As a result of this change in the examination syllabus, teachers in Hong Kong secondary schools are experiencing and/or implementing changes in their present teaching situations. This paper presents findings of teachers’ perspectives of the examination change in relation to their teaching situations and classroom practices. Classroom observations and interview methods were employed to investigate the relationship between testing and teaching. The paper also identifies areas in which the examination has impact on the teaching context. 1999 Elsevier Science Ltd. All rights reserved.

1. Introduction As far back as 1877, Latham (1877) characterised the examination system as an ‘encroaching power’ * Tel.: (403) 4921144; fax: (403) 4921318; e-mail: liying@gpu. srv.ualberta.ca. The dissertation study on which this article is based received the 1998—99 TOEFL Award for Outstanding Doctoral Dissertation Research. The Test of English as a Foreign Language (TOEFL) Policy Council sponsors this annual award for dissertation research that makes a significant and original contribution to knowledge about language tests and testing. The award committee described Cheng’s dissertation as ‘‘a carefully designed study at the intersection of language testing and teaching practice’’.

that was influencing education, blurring distinctions between liberal and technical education, and narrowing the range of learning through forcing students to prepare by studying with crammers and in cramming schools. It has been asserted that the ‘encroaching power’ also permitted external bodies — a university or a government agency — to exert control over the internal operations of educational systems that were becoming increasingly complex (Spolsky, 1994). Popham (1987) outlined the traditional notion of measurement-driven instruction to illustrate the relationship between instruction and assessment: assessment directs teachers’ attention to the content of test items, acting as powerful ‘curricular magnets’.

0742-051X/99/$ — see front matter 1999 Elsevier Science Ltd. All rights reserved. PII: S 0 7 4 2 - 0 5 1 X ( 9 8 ) 0 0 0 4 6 - 8

254

L. Cheng / Teaching and Teacher Education 15 (1999) 253–271

Whatever the teaching objectives cover, tests can be written to measure them, can be administered to all, and re-administered as necessary to those who fail (Noble & Smith, 1994b). Noble & Smith point out that the essence of current measurement-driven reform in assessment is to build a better test that will drive schools toward more ambitious goals and reform them toward a curriculum and pedagogy geared more to thinking and less to rote memory and isolated skills — the shift from behaviourism to cognitive-constructivism in teaching and learning beliefs and the assessment shift from large-scale standardised testing to performance assessment. Proponents of performance assessment believe that what is assessed is what gets taught. Following the belief (Resnick & Resnick, 1992): You get what you assess. You do not get what you do not assess. You should build assessments toward how you want educators to teach, ‘High-stakes testing’ could drive reform if it follows better psychology and pedagogy and employs more appropriate measurement forms, namely performance assessments. If tests affect curriculum and instruction, as the argument goes, performance assessment could serve as an impetus for a thinking-oriented curriculum geared toward developing higher-order abilities and problem-solving skills (Honig, 1987; Resnick & Resnick, 1992). Instruction directed toward preparation for a performance assessment promotes better instructional practice (Baker, et al., 1992). A better test will produce better results in learning. Teaching to the test, accepted by scholars as inevitable and by teachers as necessary, becomes a virtue, according to this line of thinking (Noble & Smith, 1994b). There are quite a number of studies showing that high-stakes testing influences teaching and learning

Performance assessment is defined as a ‘‘systematic attempt to measure a learner’s ability to use previously acquired knowledge in solving novel problems or completing specific tasks’’. (Gipps, 1994, p. 99). High-stakes tests are those offering future academic and employment opportunities upon results. They are usually public examinations or large-scale standardised tests such as the Hong Kong Certificate of Education Examination (HKCEE) under study.

in various ways. Smith et al. (1990) found that pressure to improve students’ test scores caused some teachers to neglect teaching materials that the external test did not include. Herman and Golan (1991) found that teachers adjusted the sequence of their curriculum based on what was included on the test. Mathison (1987) discovered that teachers altered their instructional materials to resemble the format of mandated tests. Stodolsky (1988) found that the existence of mandated testing led teachers to neglect team teaching approaches. Madaus (1988) noted that teachers taught to the test when they believed important decisions, such as student promotion, would be based on test scores. Romberg et al. (1989) found that, as a result of the emphasis on test results, teachers increased attention to paper-and-pencil computation, yet decreased attention to project work. They concluded ‘‘the greater the consequences attached to the test, the more likely it would be to have an impact on teaching’’ (see also Alderson & Wall 1993, Wall and Alderson 1993, and Romberg, Zarinnia & Williams, 1989). As has been true of previous reform efforts, assessment is central to the current educational reform debate for at least three reasons. First, assessment results are relied upon to document the need for change. Second, assessments are seen as critical agents of reform. Third, assessment results are used to demonstrate that change has or has not occurred. Linn (1992) commented that educational assessments are expected not only to serve as a monitor of educational achievement but also to be powerful tools of educational reform. Petrie (1987) pointed out that it would not be too much of an exaggeration to say that evaluation and testing have become the engine for implementing educational policy. An important assumption of the current beliefs about examination consequences is that deleterious effects on teaching and learning can be overcome by switching to performance-based assessments. Those deleterious effects are associated with the traditional assessment paradigm of test-based accountability. The new performance-based assessments are closely linked to curriculum frameworks. It is argued (Baker, Aschbacher, Niemi & Sato 1992; Honig, 1987; Linn, 1992; Popham, 1987;


Noble & Smith, 1994b) that new performancebased assessment can be designed to be so closely linked to the goals of instruction as to be almost indistinguishable from them. Rather than being a negative consequence, as it is now with some high-stakes uses of existing standardised tests, it is asserted that teaching to these proposed performance-based assessments would be considered a virtue. However, such a reform strategy was counterargued by Andrews (1994a, b) as a ‘blunt instrument’ for bringing about changes in teaching and learning. The actual teaching and learning situation is clearly far more complex than proponents of alternative assessment suggest. Each different educational context (school environment, messages from administration, expectations of other teachers, and students) plays a key role in facilitating or detracting from the possibility of change. The current study is situated within this line of debate. In line with the overall world-wide assessment paradigm shift in education, Hong Kong is no exception. For the past 15—20 years, the Hong Kong Examinations Authority has made consistent efforts to bring about positive washback on teaching and learning by means of major public examination changes. The focus of the current study was to find out how the effort to change this public examination toward a more constructive approach in assessment — a more appropriate assessment (Biggs 1995) — would actually change teaching toward a more constructive approach in Hong Kong secondary schools.

2. The research context The role of school as a selection device has been particularly evident in the Hong Kong educational system from early days (Biggs, 1993). Assessment in Hong Kong is still geared towards providing data to select students: not only in the HKCE (Hong Kong Certificate of Education) at the end of Secondary 5 and HKAL (Hong Kong Advanced Level) at the end of Secondary 7 for tertiary selection, but also in primary schools, to provide the data for

255

banding in the Secondary School Places Allocation exercise. The Hong Kong educational system is characterised as an examination-led system where what goes on in the classroom is largely dictated by what happens in the public examination halls (Fullilove, 1992). Given the traditional importance placed on public examination qualifications in the Hong Kong community, the Hong Kong Examinations Authority inevitably exerts great influence on what happens in the senior classes of the secondary schools. The Authority staff are very conscious of the fact that ‘‘if it is not examined, it won’t be taught’’ (Hong Kong Examinations Authority, 1994b, p. 80) and much thought is given to how the examination process can be used to bring about positive and constructive change to the system — commonly referred to as positive washback. In Hong Kong, certain washback effects are expected whenever language examinations are introduced. Some examples of the positive washback effect of public examinations in Hong Kong are reported in the literature (See Andrews & Fullilove, 1994; Andrews, 1993; Fullilove, 1992; Johnson & Wong, 1981). Morris (1990) emphasises that any change in the Hong Kong educational system must first involve a change in the examination system. Many of the major innovations of recent years have been designed with the expectation that the examination changes will help classroom teachers to keep a better balance between teaching and skillbuilding, on the one hand, and examination preparation on the other. However, what is still not clear is the nature and the scope of washback effects of public examinations in Hong Kong. A search of the literature indicates that washback effects are more perceived or assumed than supported by empirical data (Alderson & Wall, 1993; Andrews, 1994b).

Banding is decided upon the SSPA (The Secondary School Places Allocation) in Hong Kong. The SSPA is the process by which primary students are allocated to secondary schools. The SSPA covers the whole of the primary curriculum. Band 1 schools have the best intake of students with Band 5 the lowest. However, banding is not an absolute scale. The level of students is also related to the geographical location of the schools.

256


In 1993, the Hong Kong Examinations Authority (HKEA) introduced major changes to its existing examination syllabus in English at Secondary 5 — known as the Hong Kong Certificate Examinations in English (HKCEE) in accord with the Target Oriented Curriculum (TOC) initiative in Hong Kong. The immediate changes in the syllabus were reflected in an integrated Listening, Reading and Writing exam paper requiring students to perform simulated ‘real life’ tasks, and an increase in the weighting of the oral paper. The intention was to improve the current English language teaching and learning standard in Hong Kong secondary schools. The intended washback effect of changing its public examination was to positively influence the teaching of English toward the new philosophy in teaching and learning such as a change from non-interactive teacher-dominating talk to more real-life like and task-based teaching approaches. The above intention was to bring positive washback on teaching by means of its major public examination change into a more task-based and integrated approach. This was particularly evident

The HKCEE is a public examination taken by the majority of secondary students at the end of the fifth year of their secondary school. Students either proceed to further studies at the Sixth form level, or leave school and seek employment. A Target Oriented Curriculum (TOC) is a curriculum with clear targets and stimulating approaches for lively and effective teaching, learning and assessment. Better opportunities will be provided for children to learn and to develop their ability to think creatively. Such a curriculum initiative is viewed as the only significant landmark in the post second world war history of curriculum reform within schools in Hong Kong (Morris et al., 1996). A task refers to the idea of some kind of activity designed to engage the learner in using the language communicatively or reflectively in order to arrive at an outcome other than that of learning a specified feature of the L2 (Ellis, 1994). It can be a real-world activity or a contrived, pedagogical activity (Nunan, 1989b). In this study, a task-based approach in language teaching and learning involves learners in using the language communicatively and reflectively to carry out real-life like tasks in the classroom. A integrated approach in the context of this study refers to the approach in teaching and learning which involves the integration of the major four language skills in carrying out certain real-life tasks. Such approaches are specified as the major changes made in the new HKCEE.

at the policy-making level in Hong Kong. However, the focus of the current study was to discern whether or not the effort to change this public examination toward a more appropriate assessment (Biggs, 1995) would actually change teaching toward a more integrated and task-based approach in the Hong Kong secondary schools. There are two separate syllabuses, namely the examination syllabus by HKEA (Hong Kong Examinations Authority) and the teaching syllabus by the CDC (Curriculum Development Council), coexisting in Hong Kong secondary schools. However, the HKEA examination syllabus is the syllabus that receives teachers’ attention in schools. Morris (1990, p. 19) points out that ‘‘the curriculum in practice is defined by publications of the HKEA, specially public examination scripts and their accompanying marking schemes.’’ The textbooks in use locally, too, generally reflect the examination syllabus in their format, emphasis and approach (Fullilove, 1992). The new 1996 HKCEE was used for the first time in classroom teaching in September 1994 in Hong Kong secondary schools. The first cohort of students sat for the revised examination in May 1996. The new HKCEE consists of four papers only, with major changes in paper III and Paper IV. Paper III: Integrated ¸istening, Reading and ¼riting consists of Part A: Short Tasks and Part B: Extended Tasks. In Part A, candidates are required to select from and to make use of the information they hear and/or read in order to carry out a variety of short tasks. Part B requires students to process information by selecting and combining data from both spoken and written sources in order to complete various writing tasks (Hong Kong Examinations Authority, 1994a). Paper IV: Oral English, has changed greatly from Reading Aloud and Guided Conversation in the old examination paper to taskbased Role Play and Group Discussion in the new. Both papers of the new 1996 HKCEE require students to take an active role, participate fully in language interaction, and to carry out tasks by using different integrated language skills. Interviews were conducted with the Deputy Secretary and three senior English subject officers from the HKEA (See also Cheng, 1997).


3. Methodology The purpose of the current study was to investigate how a change in a high-stakes public examination might influence classroom teaching, in relation to the reactions of the local education context and also with regard to aspects of teachers’ attitudes and behaviours. It stresses ‘‘the importance of context, setting and subjects’’ frame of reference’ (Marshall & Rossman, 1989, p. 46). The methodology aimed to capture the reality, variation and complexity of changes in day-to-day classroom practice as well as within the whole local educational context. Given the nature of this public examination and its impact on the education system as a whole, this study investigated the relevant education context before investigating particular areas of the teaching and learning in depth. I identified three levels of washback effect of the 1996 HKCEE in terms of curriculum change. Agencies of the three levels are (1) decision-making agencies, (2) the intervening agencies, and (3) the implementing agencies (see Fig. 1). The HKEA made the decisions with its subject committee, which consists of persons nominated by the Director of Education, English subject examination officers, language experts from tertiary institutions and school teachers. The HKEA piloted

257

the revised syllabus and went to schools to get opinions from teachers. It then recommended that the textbook publishers revise textbook materials, and also informed tertiary institutions about further teacher education. Then it was up to the schools and teachers to decide how they were going to carry out their teaching according to the syllabus. Such a process usually signifies a cycle of a curriculum change. However, Smith et al. (1994) pointed out that definitions of the situation — the images of the problems a given policy should solve, as well as the characteristics of pupils, teachers, curriculum, assessment, and educational change, etc. — held by policy makers are shaped and translated imperfectly by practitioners. Teachers and principals redefine and reinterpret the messages about policy that they receive. They then act — adapt, teach, learn, and evaluate — according to their own definitions of the situation (Blumer, 1986). Therefore, the identification of the gaps noted in Fig. 1 would greatly improve our knowledge and understanding of how and in what areas a public examination change can actually influence the Hong Kong school curriculum. Research methods were designed to investigate the washback phenomenon. Triangulation was used to tackle problems at different levels. Survey studies were carried out into

Fig. 1. Washback explanatory model.

258


teachers and students’, attitudes, and were partially reported in separate papers (Cheng, 1997, 1998). Key informant interviews were carried out at different levels, and classroom observations were undertaken at the third level, and are reported below.

4. Classroom observation procedures and instruments The process of exploring the influence of this public examination change took about three years. Based on the nature of the data collection, I visited nine Hong Kong schools (23 lessons observed) as part of the baseline classroom investigation, which was carried out between November 1994 and February 1995. The purpose of the investigation was to observe how different schools reacted in face of the major public examination change, and what aspects of teaching and learning within the schools were the most sensitive to the change, and might be the areas of washback intensity. The main reasons behind the initial choice of different types of schools was that it was not clear at that time how washback would occur in schools and whether teachers in different types of schools (schools consisting of students of higher or lower levels of proficiency) would react to the examination change differently. Such studies in the above areas have been well documented (Fullan, 1993; Fullan with Stiegelbauer, 1991; Fish, 1988; McDonnell et al., 1990; Herman & Golan, 1991; Dorr-Bremme & Herman, 1983, among others). However, it was found, through the baseline study, that the differences among schools in Hong Kong are not as obvious as the differences among teachers themselves. The findings of the baseline study were identical with those of the research studies that concentrate on teachers (Bailey, 1984; Katz, 1996; Harrison, 1996; Freeman, 1996). The main study was carried out in two rounds between November 1994 and February 1995 (the

Fig. 2. Stages of classroom observations.

1st round — 8 lessons), and between November 1995 and February 1996 (the 2nd round — 14 lessons). Only three out of the nine teachers in the baseline study were eligible for the main study due to practical research constraints, some teachers from the baseline study left the teaching profession in the 2nd year, or moved to another country. Others were assigned to teach totally different levels of students. As required by this study’s research methodology, only teachers teaching Secondary Five students (also called Form 5, the grade HKCEE is conducted) in both years under the old and new syllabus were observed. One teacher simply said that she didn’t feel comfortable to be videoed again in the second round. The classroom observation procedure is summarised in Fig. 2. This study was carried out within the first two years of the period when the new 1996 HKCEE was introduced into schools. Teachers who taught Form 5 classes (F5), both years were observed for the major study: F5 classes under the old HKCEE syllabus and F5 classes under the new syllabus. Classroom observations were carried out to observe the same teachers in the same schools teaching the same forms of students.

4.1. Choice of teachers under major study

Washback intensity refers to the degree of washback effect in an area or a number of areas of teaching and learning affected by an examination (Cheng, 1997).

Based on the methodological considerations of this study, criteria for identifying potential teachers who were teaching F5 classes both in 1994 and 1995 were identified (Fig. 3):


259

Fig. 3. Comparative nature of the major classroom observations. Table 1 Numbers of F5 lessons observed in both 1994 and 1995

Ada Betty Cathy Total

1994 (old syllabus) F5 lessons

1995 (new syllabus) F5 lessons

Total Total lessons observed

4 2 2 8

6 4 4 14

10 6 6 22

E teachers who were willing to accept observations of their teaching, E teachers who agreed to have their lessons videorecorded, E teachers who were willing to encourage their students to ignore the video camera, E school administrators (principal and panel chair) who had a positive attitude toward observations. Initially nine teachers agreed to participate in the first year observation. There were eight females and one male, all qualified teachers. This group of teachers was not meant to be representative of teachers of English in Hong Kong secondary schools. The teachers were selected using purposive sampling (Patton, 1987), and the main purpose was to select teachers based on whether they could provide a rich variety of information about classroom teaching and learning activities in schools in relation to the new 1996 HKCEE. As the sample is

very small, it will be appropriate to look at each teacher as an individual case (Hargreaves & Fullan, 1992). Pseudonyms — Ada, Betty and Cathy — are used in this study to refer to the three teachers who participated in the main study. 4.2. Observation procedures After the initial school visits and baseline classroom observations of 23 lessons, detailed procedures for major classroom observation and the observation instrument were decided upon. The main study involved eight lessons from F5 lessons under the old syllabus (the 1st round) and fourteen lessons from F5 lessons under the new syllabus (the 2nd around) which were observed (see Table 1). Most observations took place in the winter term when a long and constant stretch of teaching was carried out. During those observations, a video camera was set up before each lesson at one corner of the classroom facing the students. I sat at the back of the classroom and made notes on the interaction as the lesson went along.

A panel chair is the subject head in Hong Kong secondary schools. Efforts have been made to have both male and female teachers for the study, yet none of the male teachers except one approached agreed to be observed. They gave various reasons about why they did not wish to be observed.

The HKCEE takes place in May every year. Intensive exam practice usually starts as early as February each year. All observations were carried out before the start of such exam practices.

260


The following predictions were made about the washback effect of the new syllabus on classroom teaching. Compared to teaching under the old syllabus: E The teacher will assign more practice opportunities to students. E The teacher will assign more class time to students’ activities such as role-play and group discussion. E The teacher will talk less. The students will talk more. This part of the observation will show how much students contribute to the classroom interaction and how much the teacher does. E The teacher will use more authentic materials from real-life sources. In this study, ‘authentic materials’ refers to materials taken from real-life sources, rather than textbooks or exam practice books. 4.3. The instrument The observation scheme was designed based on the analysis of the data from the baseline study mentioned above and Part A of COLT (Communicative Orientation of Language Teaching) category definitions (Frohlich et al., 1985, pp. 53—56). The following five categories were designed to describe classroom activities in order to investigate such aspects as whether the lesson was student-centred or teacher-centred, how many learning opportunities were provided and what pedagogical materials teachers used in teaching, e.g. real-life materials, main textbooks or practice exam papers. (I) Time: How is time segmented within the lesson as a percentage of class time? (II) Participant organisation: Who is holding the floor/talking during the segments of the lesson as a percentage of class time? (III) Activity type: What teaching and learning is realised through various activities as a percentage of class time? (IV) Content: What are the teacher and the students talking, reading, or writing about or what they are listening to? (V) Material used: What types and purposes of teaching materials were involved?

The above observation scheme combines the methods of real time notes and transcription of videotaped episodes of classroom activities at the level of verbal interaction between teachers and students. This scheme led the researcher to observe and describe the interactional events that took place in the classrooms in order to understand how learning opportunities were created in the context of the new 1996 HKCEE. The analytical method was a post-hoc analysis of videotaped recordings. In addition to the above observation scheme, real time notes and transcriptions of lesson episodes were also used in order to obtain both a general and a specific picture of the lessons observed, as they were essential to help to explain in detail the classroom interactions and discourse.

5. Classroom interaction outcomes Data analysis was carried out according to the five categories of the instrument and detailed recordings of episodes during the lessons were made. The analysis of the five categories enables us to see if there were any differences among the classes at the level of practice opportunity, activity types, content/language focus, modality, and the use of materials by each teacher over the two-year period. The analysis involved a calculation of the amount of time in average percentages (see Fig. 4 and Table 2) that teachers and students spent on the five categories of the observation instrument. A note of caution needs to be mentioned here that the percentages are descriptive data bearing no statistical tests of comparison. This part of the data analysis was followed by the analysis of instructional patterns and in-depth interviews, which allows us to explore in more detail the interaction between teachers and students. 5.1. Participant organisation Three basic patterns were observed: (1) Is the teacher working with the whole class or not? (2) Are students divided into groups or are they engaged in individual work? (3) If they are engaged in group work, how is it organised? They are represented in the table as:


261

Table 2 Segmented classroom activity of all three teachers’ lessons by percentage of class time under the old and the new HKCEE Activity type

Ada old n"4

Ada new n"6

Betty old n"2

Betty new n"4

Cathy old n"2

Cathy new n"4

Teacher: Pre-lesson activities Teacher: Lecturing, describing, explaining, narrating, directing

1 59

3 35

1 40

3 53

0 14

2 62

0

20

0

0

72

4

0

0

21

5

0

0

8 22 10 0

1 26 12 3

15 15 8 0

8 18 8 5

9 5 0 0

12 7 0 13

Teacher & Student: Checking answers for exercises together Teacher & Student: Reading aloud Student: Individual work Student: Group work Student: Oral Student: Listening Total in percentage"100%

Fig. 4. Participant organisation of the lessons carried out by the three teachers.

E ¹'S/C (teacher to students or class as a whole, e.g. in the case of lecturing) E S'S/C (students to students or class as a whole, e.g. in the case of oral presentation)

E Group work (students are working on a certain task in groups of two or more students) E Individual (individual work, e.g. on exercises or listening)

262


E ‘Choral’ in the figure refers to the whole class repeating a model provided by the textbook or the teacher. Differences between the two years can be seen from the figure. Looking at the participant organisation of the interactive patterns of the classroom activities carried out by the three teachers in Fig. 4, a small increase in group work was seen in 1995 compared with 1994 (see Fig. 4), though the increase varied from teacher to teacher. This suggested a possible washback effect of the new HKCEE, as group work is required for its oral exam. Paper IV: Oral consists of Role-play and Group discussion whereas the old exam format consists of Reading and Dialogue and Conversation. From this point of view, changing the format of the exam may have had some impact at this level on classroom teaching. 5.2. Activity types and content as a percentage of class time The purpose of looking at activity type in classroom teaching is to explore what kinds of teaching and learning are realised through various activities as a percentage of class time. By investigating the content of those activities carried out in the classroom, we can explore the subject matter of the activities — What are the teachers and the students talking, reading, or writing about or what they are listening to? After each lesson was segmented and participant organisation of classroom interaction patterns were analysed, it was then possible to look closer at the activity types carried out according to the segments analysed. Activity types were grouped into (1) teacher activities, (2) teacher and student activities carried out together and (3) student activities. Each activity was classified, such as discussion, drill or singing. Frequently, activities consisted of two or more episodes as mentioned earlier. Findings related to the content will be reported in this section only as a percentage of class time. Whether the content of a particular activity was management (classroom procedures or disciplinary routines) or language input (form, function or discourse) was analysed when detailed analysis of excerpts of

transcriptions of lesson episodes was carried out, which will be reported in a separate paper. Table 2 shows: (1) what types of activities were carried out in the lessons and how lessons were segmented according to the percentage of time (duration of time) devoted to them by the three teachers; and (2) who was holding the floor during the lessons, teacher or student, and in what ways. The analysis of the 22 lessons conducted by the three teachers shows the following patterns in Table 2. The percentages given in Table 2 are mean percentages for the numbers of lessons observed in each of the two years for each teacher. In this way, differences of Ada, Betty and Cathy’s own lessons over the two years can be observed. It can be seen that teacher talk as a percentage of class time increased under the new 1996 HKCEE for both Betty and Cathy. One of the research foci was to explore whether, as a result of the new 1996 HKCEE, teachers could provide students with more practice opportunities, which would suggest less teacher talk. When Betty and Cathy were asked the reason for having more teacher talk, they both explained that it was because of the new examination. As their students would take the new HKCEE, they felt that they had to explain more to their students (talk more) and provide explanations in detail in order to make sure that students could meet the new examination requirements. Another activity type specified above was teacher and student joint activities. It was found that Ada and Cathy used checking answer activities. When such lesson episodes were analysed, there were frequent and short turns between the teachers and students. These are regarded as ideal for providing practice opportunities (see Allwright, 1980; Tsui 1994). However, it was found that these turns were used for simple questions and answers for them. This type of activity was usually related to a grammar-related exercise or a comprehension exercise with multiple choice responses. Therefore, the interaction between the teachers and students would be based on responses to choose the correct answer, A, B, C or D. The students simply answered A, B, C or D, according to whichever was the answer they had arrived at. Then the turn went to the teacher for the next exercise item. This type of activity seemed to be interactive, but little was


achieved in terms of actual learning opportunities for the students, who were observed to be bored and reluctant to participate.

5.3 ¹eaching materials used It was found that the three teachers used the same kinds of teaching materials. They were commercially written materials, occasionally accompanied by audio-visual materials (cassette tapes for listening). The written materials were exam practice books. The books were structured in the same way as the HKCEE formats for both years under the old and the new examinations. The ultimate aim of both F5 teaching materials was to prepare students for the HKCEE. This was the same for both years. However, there were occasions — two lessons out of the 14 lessons observed in 1995 under the new HKCEE — when Ada and Cathy used newspapers during their teaching. They used the news from the newspapers as a topic for oral work and discussion. However, no main textbooks were used for F5 classes in 1994 or 1995. In Hong Kong secondary schools, students at the F4 level tended to use both main textbooks as well as practice examination books. However, for F5 only practice books (examination papers) were used. F5 was highly examination oriented in teaching. In addition, many past examination papers were used with F5 students in 1994. In 1995, past examination papers were not available at the time of the research, as the 1995 cohort of F5 students was the first group to take the new HKCEE examination. Therefore, practice books were used instead, which were designed according to the formats of the examination. For example, since the 1994 HKCEE had five papers, the practice books were divided into five parts. As the new HKCEE has four papers, the practice books automatically consisted of four books. Each book focused on one examination paper. Although examination practice books were used in both years, the formats of the practice books were quite different, especially for Paper III: Integrated ¸istening, Reading and ¼riting and Paper IV: Oral. Therefore, the difference in the HKCEE exam papers led to differences in classroom activities. For example, the old Oral paper consisted of Reading Aloud and Picture

263

Conversation. The new Oral is in the format of RolePlay and Group Discussion. Therefore, classroom activities in 1995 were based on role-play and group discussion to reflect the new examination format. To reiterate, the types of teaching materials for both F5s were not much different. They were both practice books. However, the contents were different. The new 1996 HKCEE led to changes in such activities as are required by the new HKCEE exam format.

6. A close look at three teachers’ oral lessons The following is a closer look at the oral lessons that the three teachers carried out under the old and new syllabuses, as oral was an intended area of washback from the HKEA due to the change in exam weighting and formats (HKEA, 1993). Two oral lessons from the 1st and 2nd rounds of observations conducted by each teacher were analysed here. These oral lessons were chosen as a comparative analysis since they were observed within the same cycle of teaching of the three teachers. A cycle consisting of six school days is a teaching arrangement of the school subject contents. It is important to note here that although the three teachers scheduled their lessons as oral, it was observed that there were short listening, reading and writing tasks conducted during the lessons. It is also important to note that findings are usually derived from limited numbers of observations and with certain emphasis. There is simply so much happening in the classrooms that needs to be observed. And the actual teaching is far more complicated than that of most of our research instruments. 6.1. Ada Two oral lessons taught by Ada were observed and analysed. In the 1994 lesson, of the 70 minutes of the lesson (68 min recorded and analysed), A stopwatch was used to record the time for classroom activities. A unit of one minute was used. Usually most of the teachers cannot start the lesson at the exact time the lesson is scheduled to begin. If there is no recess (continued overleaf ).

264


Ada talked for 56% of the lesson time. The teacher was in control 69% of the lesson time. 6% of the lesson was devoted to teacher-led activities. There was a total of 22% lesson time used for group or pair work on dialogue reading and picture description, one student being the examiner and the other being the examinee or vice versa. 9% of the lesson time was used for students’ oral presentations in front of the class. Most of the teacher-talk time was spent on working with the vocabulary required to give picture descriptions, and commenting on the language points and sentence structures before or after students’ oral presentation. In 1995, during the 70 minute lesson (69 minutes recorded), the teacher talked for 36% of the lesson time, which was a decrease from 1994. 4% of the lesson was spent on leading classroom activities. Students’ pair and group activities occupied 43% of the lesson time and 13% was spent on student oral presentations. Students’ practice opportunities increased from 31% of the lesson time in 1994 under the old HKCEE to 56% of the lesson in 1995 under the new. This finding suggests that the new HKCEE Oral paper may have had some influence on teaching. There is an increase in the time spent on students’ activities similar to the new HKCEE oral format — role play and group discussion. However, the main pattern of the lesson was the same in both years. It follows the pattern: teacher explanation — student activities — teacher comment and further explanation. But, the atmosphere in 1995 was lively as students spent most of the time working and talking in groups. Their attention was focused on the tasks they were doing. The teacher went around the classroom to give individual comments, and then commented to the whole class once in a while. From the overall analysis of the two oral lessons, it can be seen from Table 3 that teacher-dominating

Table 3 Comparison of two oral lessons carried out by Ada in 1994 and 1995 Activities

Teacher 1994 Teacher 1995

Pre-lesson activities Lecturing and explaining Leading classroom activities

7% 56% 6%

Activities

Students 1994 Students 1995

Pair work or group work Oral presentation

22% 9%

4% 36% 4%

43% 13%

talk and activities declined, and students’ activities were on the increase. These findings indicate that the format differences between the two syllabuses might have led to certain changes in the general pattern of the lessons taught by Ada. The new HKCEE seemed to have led to an increase in the percentage of lesson time being spent on the activities required in the new exam. Furthermore, this teacher did not assign silent individual work among her students. Instead the major student activities were either talking in pairs or within groups, or giving oral presentations. Students received more opportunities to talk in a more creative manner in 1995 than they did in 1994. This change was due, to a large extent, to the basic requirements of the examination. Ada commented later in the interview about why certain changes happened in her teaching due to the exam. Definitely yes, there are certain changes. I think teachers are under a lot of pressure. Teachers who are in the exam levels are under a lot pressure to deliver, you know, good grades. At least, in my schools, they won’t acknowledge that, but it is there. Teachers have to work for the exam. Teachers, according to Ada, make certain changes in their teaching for the new exam as their

before the lesson, there is no time in between the last lesson and the following lesson. Even if the teacher in the previous lesson leaves the classroom exactly on time, students still need some time to relax and change textbooks, etc. Therefore, teachers usually spend the first few minutes talking to students about homework or daily news. Teachers might use this period of time to fix the tape recorder for listening, etc.

All interviews were carried out in English. Interviews with the three teachers during the second year observation were audiotaped. Each of the interviews was about two hours. They were all transcribed.


students’ exam scores are related to their reputation in the school. However, Ada commented further on the possibility that the change of the exam (the new 1996 HKCEE) would eventually change teaching. Definitely the exam will change teaching, but I don’t think it will idealise it. I don’t think it will make it into an ideal situation. You know, what I would want to do is to allow more language in the classrooms, more spoken language. In that, the students can feel more. They allow the language to become a part of themselves. 6.2. Betty Two oral lessons were observed in 1994 and 1995. In the 1994 lesson, during the 70 minutes of the lesson (69 min recorded), the teacher talked for about 39% of the lesson time. Twenty-three percent of the time was spent on dialogue reading, the teacher being the examiner and students being the examinees, or vice versa. Twenty percent of the time was spent on individual student work, and 13% on pair work, with one student as the examiner and the other as the examinee. The major interaction between the teacher and students was choral work — reading a dialogue aloud. There was no student initiation of questions. In a 70-minute 1995 lesson, the teacher talked for 55% of the lesson time, allocating 11% for listening activities, and 18% for oral group activities. Twelve percent of the lesson time was spent on organisation of activities and activities themselves. The main pattern of the two lessons was IRF: Initiation — Response — Feedback (Sinclair & Coulthard, 1975; Mehan, 1979; Cazden, 1985), but there were many interactions in which the teacher initiated questions and students answered the questions. The teaching material used was confined to the practice workbook (Paper IV for the lesson). It can be seen from Table 4 that there was more pair work and group work (18%) carried out in the 1995 class than in 1994. The increase is from 13% in 1994 to 18% in 1995. Only half as much time was devoted to individual work in 1995 (11%), com-

265

Table 4 Comparison of two oral lessons carried out by Betty in 1994 and 1995 Activities

Teacher 1994

Teacher 1995


3% 39% 23%

4% 55% 12%

Activities

Students 1994

Students 1995

Individual work Pair work or group work

22% 13%

11% 18%

pared with 1994 (22%). However, there was an increase in the time spent on teacher’s talk, which was different from the previous assumption — a decrease of teacher talk and an increase in students’ practice opportunities. The increase of teacher talk was from 39% in 1994 to 56% in 1995. Betty, the teacher during the interview explained the reason for the increase of teacher talking time. She said, in a verbatim reply: Yes. Probably it is because the old type of examination is very structured. The old type one has been there for quite some time. Being a very experienced teacher in the old Hong Kong exam format, we all know what is the requirement of the examination. So we are very structured. We get the students to read. We get the students to look at the pictures and learn vocabulary. But with the new one, we don’t really know what questions would come up. Because it emphasises on communication, so I have to create different situations for the students. I have to teach them the basic skills. She also mentioned that she certainly gave the new F5 students more opportunities to talk in class. When asked about whether the new HKCEE had brought some significant changes in her teaching methods. She said: Yes, it has. Especially, for example, I think the biggest impact is on Paper IV, the oral part. I have started a lot of role-plays. We have a lot of group discussions. That is not just in class. It is even the whole school thing. For

266


example, in the English week, we created situations for students so that they could have an environment where they could talk or speak to other students or things like that, whereas there was no need for us to do such things in the previous year. The biggest change observed between the two lessons taught by Betty was that there was much more interaction between the teacher and the students in 1995 than in 1994. Although the teacher did not talk a lot in the 1994 lesson, that lesson was carried out according to the old examination format — practising reading a dialogue aloud with interaction between the teacher to students, and students to students. Choral reading aloud dominated the whole lesson. There was no interaction, only simulations of dialogue practice between the teacher and the students and students among themselves. It is difficult to perceive changes in just a few lessons. However, interviewing the teacher provided insights into what had happened as a result of the examination change. In the interview, the teacher talked about the major differences in her oral lesson teaching in 1994 and teaching in 1995, and the new oral teaching methods she had adopted: With the old F5 class, it is more like learning more vocabulary. This is flower. This is a tree. You must know how to say these words. They also have to read. The oral skills for the previous F5s are reading skills, I think. [Reading aloud?] That is it. They see the words. And they must know how to pronounce it, and they have to remember intonation and stresses of certain words. But with the new CE syllabus, the emphasis is on communication. That is the problem. The students have to reveal how they would solve a particular problem. It is, I think in that way, students became more natural in their responses rather than reading aloud. Although the interaction pattern of her class has not changed much, the way that the lesson was organised changed to a great extent. The teacher’s perception of her role and her students’ role changed. The nature of the classroom activities

Table 5 Comparison of two oral lessons carried out by Cathy in 1994 and 1995 Activities

Teacher 1994

Teacher 1995


4% 81% 4%

4% 53% 11%

Activities

Students 1994

Students 1995

Individual work Pair work or group work

8% 3%

18% 14%

carried out was definitely different in 1994 than it was in 1995 especially in the oral activities. 6.3. Cathy Cathy’s talking time decreased from 81% of a lesson in 1994 to 53% of a lesson in 1995. In the 1995 lesson, there were increased practice opportunities for students’ individual work, pair work and oral presentation. The increase was from 8% in 1994 to 18% in 1995 for individual work, 3% in 1994 to 14% in 1995 for pair work and group work. (see Table 5) However, the interaction pattern of the lessons in both years was still the traditional IRF: Initiation — Response — Feedback. The focus of the lesson was still the teacher to the whole class. In lessons observed in the two years, the teacher followed a pattern starting with assigning a language task, lecturing, explaining language points, and then checking answers from the students. During most of these lessons, the teacher talked to the whole class. Students only provided very limited responses such as ‘yes’ or ‘no’ answers. There was no student initiation of questions. The interaction pattern did not show much change when she taught according to a different practice book or using either of the two syllabuses. The interaction pattern mentioned above, however, can be further explained from the interview. When I asked Cathy whether she thought teaching under the old exam syllabus and the new syllabus was different, she replied: Concerning the methods, no. But concerning the materials, yes. Well, I don’t know about


other teachers. But to me, I am still using the same methods. I am still asking the students to do the exercises. Well, take F5 for example, they are in Term Two now, they are doing, well, exam practices. I am still using the same method. I am still asking the students to do it in class. Do it as homework at home. Then they have to tell me the answers. They have to tell me where and how they get the answers, where they find the support, and where they find the proof of it, whatever. It is the same.

In her 1994 lessons, Cathy began by assigning a reading task. She took the students through the comprehension exercises of the pre-assigned reading passage one by one. The lesson continued for almost nine minutes before she started to ask the students a question. Only a few students answered the question. During the lesson (66 min video-recorded), the teacher talked for 81% of the class time. In the whole lesson, there was only 3% of the lesson time left for students’ pair work. In the 1995 lesson observed, it was observed that there was a decrease in teacher talk - 53% compared with 81% in 1994. Teacher-led activities took 11% of the lesson. Fourteen percent of the lesson was spent on pair work. During the remaining part of the lesson, students either worked on their exercise books or carried out listening activities. There was not much change in the way the teacher organised her lessons. From the later interview, it was clear that she believed that checking the answers and making sure students knew how to get the right answers was her major responsibility. During the interview, she also mentioned that students’ proficiency level was the key problem in implementing the new syllabus. To her, if a class of students was good, (i.e. they had a high level of proficiency), there was no need for the teacher to talk and explain much. The students would be able to carry out tasks on their own. However, if the level of a class of students was low, the teacher would have to spend a lot of lesson time to explain the theory part, then give them extra practice during lunch hour as was the case with her 1995 F5 class.

267

I would use the oral lessons teaching them the theory part first. If they want to ask questions properly, what should they do or whatever. They have to do the practice in class. In F5, I am focusing on the discussion, Part Two of the oral exam. I have to give them extra lessons. In oral class, I teach them the theory. And then every afternoon, they would have extra lessons. They come to my office, group by group. And they would have a discussion just like the examination. And they have practices at that time during the lunchtime. And I tried, before I gave them extra lessons, I try to let them practice in class. But I don’t think the results would be good. Because I can’t supervise all of them.

7. Discussion and conclusion To summarise the observations of the three teachers above, it can be seen that very similar classroom activities were observed in the oral lessons of the three teachers, particularly in 1994. The oral activities then emphasised oral reading skills; vocabulary and sentence structures required by Dialogue Reading and Picture Conversation. In 1995, however, the emphasis had shifted to Role Play and Group Discussion. The skills emphasised in the new HKCEE would focus on the expression of meaning, and were less likely to be on the linguistic accuracy of utterances. Dialogue Reading, on the other hand, tended to limit student talk to the production of isolated sentences which are assessed for their grammatical accuracy. The interaction pattern of each teacher’s teaching in the classroom, however, did not reveal much change. The lessons were carried out in similar ways in both years. These lessons were highly teacher-controlled. In both years, teacher talk time especially for Betty and Cathy dominated the majority of the class time (see Table 6). Teachers tend to follow whatever they feel comfortable with in their teaching. But there were two levels of changes that could be observed. First, there were differences (changes) within teachers in the allocation of time in teacher and student talk vs. teacher and student activities in both years. In Ada’s 1995 lesson,

268


Table 6 Summarising the observations of the three teachers Ada

Teacher talk Student talk

1994 69 31

Betty 1995 44 56

student talk (activity time) overtook teacher talk (activity time). For Cathy, teacher talk (activity time) decreased by 21% in 1995. Second, there were differences between teachers in both years. Cathy’s lessons were much more teacher-controlled compared with the other two in general. Table 6 summarises the oral lessons from the three teachers according to the percentage of class time spent on activities. The observations revealed that there were new activities in the 1995 lessons. Teachers tried to use various teaching activities according to their understanding of the integrated approach and task-based approach as encouraged by the new HKCEE. One of the reasons Betty gave was that they did not really know what the new HKCEE format would be, and they did not know exactly how they should teach according to the new examination syllabus. Nor could they rely on any past exam papers. Thereafter, they just tried their own ideas. This washback effect was seen to occur with some of the teachers who were positive about their teaching like Betty. The change to the new exam provided an opportunity for them to try out new ideas and activities in schools. The students would not complain, as they did not know much about the new HKCEE either. Other teachers tended to rely on the practice books for the new 1996 HKCEE. In that case, they would use activities suggested by the textbook writers. Furthermore, it can be seen that there were increasing opportunities for students’ activities in 1995. Teachers assigned much more time for activities such as role play and group work. Note also that some out-of-class time was used for these types of activities (e.g., Betty’s comments about English Week activities; Cathy’s comments about practice discussion at lunchtime). As a result, class time

1994 65 35

Cathy 1995 71 29

1994 89 11

1995 68 32

spent on student individual work decreased. However, there was no evidence in the lessons observed that teachers employed more authentic materials. Teachers relied on the revised practice examination books most of the time. In the 1994 lessons, the teachers tended to use past examination papers. In 1995, however, they relied on the revised practice workbooks designed especially for the new HKCEE papers. One teacher did mention the use of English newspapers in her lessons and her requests to students to buy the newspaper. She mentioned that she needed the newspapers to help students cope with Paper III: Integrated Listening, Reading and Writing of the new examination. However, she did not use newspapers in the old syllabus. When comparing the three teachers’ teaching over the two-year period, it can be seen that the general pattern of teaching approaches had not changed much. In 1995, the teachers still taught in a manner similar to 1994. There were differences between teachers, not within the teachers themselves. It can be seen that although the change of the HKCEE could push them to change their classroom activities given the importance of this public examination, it has not changed them in their fundamental beliefs and attitudes about teaching and learning, the roles of teachers and students, and how teaching and learning should be carried out. The following verbatim interview further illustrates this point. When I asked the teachers individually whether they thought that the new HKCEE actually brought about any changes in teaching methodology, the answer was a definite ‘yes’. Betty’s comment exemplified this point. Yes, after all especially for teachers who are teaching graduate classes, F5s, F7s even, of


course the main guidelines are usually from the Examinations Authority. All right, no matter what we are going to say about the curriculum, we have to follow the curriculum. We have to follow the syllabus, but truly where we go, at the front line, the first thing the teachers and students are very concerned about is ‘Are the students able to perform well in the exam? OK.’ If the Examinations Authority is going to change something or anything, I think all teachers will be concerned about this item. I think the teachers will just do their best in order to make their students take the examinations well. I think most of the teachers believe this. A key assumption underlying the change of public examination to change teaching is that the outcome is something that is required of all, regardless of their differing capacities and interpretations. I found substantial variation in teachers’ beliefs about how to teach and how students learn. Teachers also varied in their classroom practices. The change of the public examination could, to a large extent, change the content of teaching, even the way activities were carried out, but very little change in the interaction pattern between teachers and students could be found. McDonnell and Elmore (1987) point out that reform instruments tend to depend on coercion to create uniformity. If one expects practitioners to change themselves and their students, an environment conducive to such change must be fostered. The teaching context, school environment, messages from the administration, and expectations of other teachers facilitate or detract from the possibility of change. Fullilove (1992, p.131) also commented that ‘‘the nature and strength of this washback effect and the benefits or disadvantages of the washback depend in large measure upon the educational system as a whole, upon the nature of the other participants involved in writing or establishing the competing curricular, and, of course, upon the types of examinations in question’’. Noble and Smith (1994b) concluded that to be consistent with cognitive-constructivist beliefs about learning and teaching, educational reform

269

efforts directed toward instructional improvement should acknowledge the challenges presented by such conceptual changes. Changes of this type are not simply brought about by the imposition of new teaching materials. Conceptual changes are seldom achieved without attending to the beliefs of those who are the targets of change — teachers — and the conditions of the environments in which they function — schools and student levels of proficiency.

References Alderson, J.C., & Hamp-Lyons, L. (1996). TOEFL preparation courses: A study of washback. ¸anguage ¹esting, 13(3), 280—297. Alderson, J.C., & Wall, D. (1993). Does washback exist? Applied ¸inguistics, 14(2), 115—129. Allwright, D., & Bailey, K. (1991). Focus on the language classroom: An introduction to classroom research for language teachers. Cambridge: Cambridge University Press. Allwright, R.L. (1984). Why don’t learners learn what teachers teach? — The interaction hypothesis. In D.M. Singleton, and D.G. Little (Eds.), ¸anguage learning in formal and informal contexts, (pp. 3—18). Dublin: IRAAL. Andrews, S., & Fullilove, J. (1993). Backwash and the use of English oral speculations on the impact of a new examination upon sixth form English language testing in Hong Kong. New Horizons, 34,46—52. Andrews, S. (1994). The washback effect of examinations: Its impact upon curriculum innovation in English language teaching. Curriculum Forum, 1(4), 44—58. Baker, E., Aschbacher, P., Niemi, D., & Sato, E. (1992). CRESS¹ Performance assessment models: Assessing content area explanations. Los Angeles: University of California, Centre for Research on Evaluation, Standards, and Student Testing. Biggs, J.B. (1995). Assumptions underlying new approaches to educational assessment. Curriculum Forum, 4(2), 1—22. Biggs, J.B. (1993). The quality of teaching and learning in an expanding tertiary system. New Horizons, 34, 102—104. Blumer, H. (1986). Symbolic interactionism: Perspective and Method. Berkeley: University of California Press. Cazden, C.B. (1985). Classroom discourse. In M.C, Wittrock, (Ed.), Handbook on ¹eaching (pp. 432—463). New York: Macmillan. Cheng, L. (1997). How does washback influence teaching? Implications for Hong Kong. ¸anguage and Education, 11(1), 38—54. Cheng, L. (1998) Impact of public English examination change on students’ perceptions and attitudes toward their English learning. Studies in Educational Evaluation. 24(3), 279—301.

270


Eckstein, M.A., & Noah, H.J. (Eds.) (1993). Examinations: Comparative and international studies. Pergamon Comparative & International Education Series. (Vol 12) Oxford: Pergamon Press. Education Department, Hong Kong. (1989). Report of the working group set up to review language improvement measures. Education Department, Hong Kong. Erickson, F.E. (1986). Qualitative methods in research on teaching. In M. Wittrock (Ed.), Handbook of research on teaching. (3rd ed., pp. 119—161). New York: Macmillan. Frohlich, M., Spada, N., & Allen, P. (1985). Differences in the communicative orientation of L2 classrooms. ¹ESO¸ Quarterly, 19:27—56. Fuhrman, S.H. (Ed.) (1993). Designing coherent education policy: Improving the system. San Francisco: Jossey-Bass. Fullilove, J. (1992). The tail that wags. Institute of ¸anguage in Education Journal, 9, 131—147. Gipps, C.V. (1994). Beyond ¹esting: ¹oward a ¹heory of Educational Assessment. London: The Falmer Press. Herman, J.L., & Golan, S. (1991). Effects of standardised testing on teachers and learning — Another look. CSE ¹echnical Report 334. Los Angeles: University of California. Centre for Research on Evaluation, Standards and Student Testing. Hong Kong Examinations Authority. (1993). Hong Kong certificate of education examination 1996 — Proposed English language syllabus. Hong Kong: Hong Kong Examinations Authority. Hong Kong Examinations Authority. (1994a). Hong Kong Certificate of Education examination 1996 — English ¸anguage p. 187. Hong Kong: Hong Kong Examinations Authority. Hong Kong Examinations Authority. (1994b) ¹he work of the Hong Kong Examinations Authority — 1977—93. Hong Kong: Hong Kong Examinations Authority. Honig, B. (1987). How assessment can best serve teaching and learning. Assessment in the service of learning: Proceedings of the 1987 E¹S Invitational Conference. Princeton, NJ: Educational Testing Service. Johnson, F., & Wong, C.L.K.L. (1981). The interdependence of teaching, testing and instructional materials. In J.A.S. Read (Ed.), Directions in language testing (pp. 277—302). Singapore: Regional Language Centre. Latham, H. (1877). On the action of examinations considered as a means of selection. Cambridge, Deighton: Bell and Company. Linn, R.L. (1992). Educational assessment: Expanded expectations and challenges. CSE ¹echnical Report 351, Centre for Research on Evaluation, Standards, and Student Testing. University of Colorado at Boulder. Madau, G. (1985). Test scores as administrative mechanisms in educational policy. Phi Delta Kappa, 66 (9), 611—617. Madau, G. (1988). The influence of testing on the curriculum. In L. Tanner (Ed.), ¹he politics of reforming school administration (pp. 83—121). London: The Falmer Press. Marshall and Rossman. (1989). Designing qualitative research. California: Sage.

Mathison, S.M. (1987). ¹he perceived effects of standardised testing on teaching and curriculum. Unpublished dissertation, University of Illinois at Urbana-Champaign, Urbana. McDonnell, L.M., & Elmore, R.F. (1987). Getting the job done: Alternative policy instruments. Educational Evaluation and Policy Analysis, 9(2), 133—152. Mehan, H. (1979). ¸earning lessons: social organisation in the classroom. Cambridge, MA.: Harvard University Press. Miles, M.B., & Huberman, A.M. (1994). Qualitative data analysis: An expanded sourcebook. New York. Sage. Morris, P. (1990). Curriculum development in Hong Kong. Education Paper 7. Hong Kong: Faculty of Education, The University of Hong Kong. Morris, P. et al (1996). ¹arget oriented curriculum evaluation project — Interim report. INSTEP, Faculty of Education, The University of Hong Kong, Hong Kong. Moskowitz, G. (1971). Interaction analysis — A new modern language for supervisors. Foreign ¸anguage Annals, 5, 221—221. Noble, A.J., & Smith, M.L. (1994a). Measurement-driven reform: Research on policy, practice, and repercussion. CSE ¹echnical Report 381, Centre for Research on Evaluation, Standards, and Student Testing. Arizona State University. Noble, A.J., & Smith, M.L. (1994b). Old and New beliefs about measurement-driven reform: The more things change, the more they stay the same. CSE ¹echnical Report 373. Centre for Research on Evaluation, Standards, and Student Testing. Arizona State University. Petrie, H.G. (1987). Introduction to ‘‘evaluation and testing’’. Educational Policy, 1, 175—180. Popham, W.J. (1987). The merits of measurement-driven instruction. Phi Delta Kappa, 68, 679—682. Resnick, L.B., & Resnick, D.P. (1992). Assessing the thinking curriculum: New tools for educational reform. In B.R. Gifford, & M.C. O’Connor, (Eds.), Changing assessments: Alternative views of aptitude, achievement and instruction. London Kluwer. Romberg, T.A., Zarinnia, A., & Williams, S.R. (1989). ¹he Influence of Mandated testing on mathematics instruction: Grade 8 teachers’ perceptions (monograph). Madison: University of Wisconsin, National Centre for Research in Mathematical Science Education. Smith, M.L., Edelsky, C., Draper, K., Rottenberg, C., & Cherland, M. (1990). The role of testing in elementary schools. CSE ¹echnical Report 321. CRESST. University of California. Smith, M.L., Noble, A.J., Cabay, M., Heinecke, W., Junker, M.S., & Saffron, Y. (1994). What happens when the test mandate change? Results of a multiple case study. CRESS¹ Report 380. Centre for Research on Evaluation, Standards, and Student Testing. Arizona State University. Spada, N. (1989). A look at the research process in classroom observation: A case study. In C. Brumfit and R. Mitchell. (Eds.), Research in the ¸anguage Classroom (pp. 81—93). Modern English Publications in association with the British Council.

L. Cheng / Teaching and Teacher Education 15 (1999) 253–271 Spolsky, B. (1994). The examination — classroom backwash cycle: Some historical cases: Proceedings of the International ¸anguage in Education Conference 1994. Hong Kong: The University of Hong Kong.

271

Stodolsky, H. (1988). ¹he subject matters: Classroom activity in math and social studies. Chicago: University of Chicago Press. Wall, D., & Alderson, J.C. (1993). Examining washback: The Sri Lankan impact study. ¸anguage ¹esting, 10(1), 41—69.