PROOF COVER SHEET Author(s):
Michael Grenfella* and, Vee Harrisb
Article title:
Making a difference in language learning: the role of sociocultural factors and of learner strategy instruction
Article no:
RCJO_A_744326
Enclosures:
1) Query sheet 2) Article proofs
Dear Author, 1. Please check these proofs carefully. It is the responsibility of the corresponding author to check these and approve or amend them. A second proof is not normally provided. Taylor & Francis cannot be held responsible for uncorrected errors, even if introduced during the production process. Once your corrections have been added to the article, it will be considered ready for publication. Please limit changes at this stage to the correction of errors. You should not make insignificant changes, improve prose style, add new material, or delete existing material at this stage. Making a large number of small, non-essential corrections can lead to errors being introduced. We therefore reserve the right not to make such corrections. For detailed guidance on how to check your proofs, please see http://journalauthors.tandf.co.uk/production/checkingproofs.asp.
2. Please review the table of contributors below and confirm that the first and last names are structured correctly and that the authors are listed in the correct order of contribution. This check is to ensure that your name will appear correctly online and when the article is indexed. Sequence
Prefix
Given name(s)
Surname
1
Michael
Grenfella* and
2
Vee
Harrisb
Suffix
Queries are marked in the margins of the proofs.
AUTHOR QUERIES General query: You have warranted that you have secured the necessary written permission from the appropriate copyright owner for the reproduction of any text, illustration, or other material in your article. (Please see http://journalauthors.tandf.co.uk/preparation/permission.asp.) Please check that any required acknowledgements have been included to reflect this.
AQ1: Please confirm the corresponding author’s email address is correct – it is not spelt the same way as your surname AQ2: Shepherd 2011 is not listed in the references – please supply full publication details for the reference list or remove citation AQ3: Baker 2006 is not listed in the references – please supply full publication details for the reference list or remove citation AQ4: DCSF 2007 is not listed in the references – should this be DfE 2007? If not, please supply full publication details for the reference list or remove citation AQ5: Please note: the numbering of tables has been altered according to those that appear in the main text and those which are appendices. Tables 1 and 2 are now Appendices 1 and 2, and Tables 3–6 have been re-named Tables 1–4. Please confirm all in-text citations to specific tables are still correct. AQ6: Cohen, Weaver and Li 1998 is not listed in the references – please supply full publication details for the reference list or remove citation AQ7: Please supply brief biographical notes on authors AQ8: Bell 2004: please supply a volume number for this journal issue (if possible), and page numbers for the article AQ9: Carr and Pauwels 2005: please supply location details (city, country/state) for publisher AQ10: Cassen and Kingdon 2007: please supply location details (city, country/state) for publisher AQ11: Cenoz and Valencia 1994 is not cited in the text – please indicate where it is referred to or remove from the reference list AQ12: Chamot et al. 1996: please supply page numbers for this chapter AQ13: Dreyer and Oxford 1996: please supply page numbers for this chapter AQ14: Gilborn and Mirza 2000: if you are citing an electronic source, please supply a URL that links directly to the document, rather than the organisation homepage. If you are citing a published document, please supply the name and location details of the publisher. AQ15: Goh 1998 is not cited in the text – please indicate where it is referred to or remove from the reference list AQ16: Grenfell and Harris 2007: please supply the date(s) of the meeting at which this paper was presented AQ17: Holliday 2003: please supply page numbers for this chapter AQ18: Ikeda 2006: please supply location details (city, country/state) for publisher AQ19: Macaro and Erler 2005: please state what sort of document this is – a dissertation, internal report, etc. and give the author’s affiliation/organisation by which it was produced AQ20: Macaro et al. 2007: please supply page numbers for this chapter AQ21: Naiman et al. 1978 is not cited in the text – please indicate where it is referred to or remove from the reference list AQ22: National Literacy Trust 2006: the URL supplied does not work, and this document cannot be found on the Literacy Trust’s website – please check your source details and re-supply a
working URL to the document you used as a reference AQ23: NferNelson 2003 is not cited in the text – please indicate where it is referred to or remove from the reference list AQ24: Oxford and Schramm 2007: please supply page numbers for this chapter AQ25: Rubin 1990: please supply page numbers for this chapter AQ26: Rubin et al. 2007: please supply page numbers for this chapter AQ27: Shapiro 2007: please supply page numbers for this chapter AQ28: Stern 1975 is not cited in the text – please indicate where it is referred to or remove from the reference list AQ29: Takeuchi et al. 2007: please supply page numbers for this chapter AQ30: Watkins 2007: please supply page numbers for this chapter AQ31: White et al. 2007 is not cited in the text – please indicate where it is referred to or remove from the reference list AQ32: White et al. 2007: please supply page numbers for this chapter AQ33: Wong Fillmore 1979: please supply page numbers for this chapter AQ34: Table 3: Cohen 1988 is not listed in the references – please supply full publication details for the reference list or remove citation AQ35: The resolution of Figure 5 is too low – the legend and other text within the figure will not be legible in print. Please re-supply a high-resolution image file (preferably either TIFF or EPS format). Also, the key for Figure 1 says ‘Monlingual’ – please provide corrected Figure if possible.
devia 15/11/12 12:43 2008 – (B5)
RCJO_A_744326
(XML)
RefSty-(F Chicago au-date)
The Curriculum Journal Vol. 00, No. 0, Month 2012, 1–32
Making a difference in language learning: the role of sociocultural factors and of learner strategy instruction
5
Michael Grenfella* and Vee Harrisb a
School of Education, University of Southampton, Southampton, UK; bDepartment of Education, Goldsmiths College, University of London, London, UK
There has been increasing concern over the poor performance and lack of interest in modern language learning among secondary-school students. Although there is some evidence as to the under-achievement of boys in modern languages (ML), there is less information as to the degree to which other factors such as social class, and bilingual or monolingual status play a role; the focus of much ML research being on individual rather than sociocultural differences in language learning. The present study took place in two London schools with 120 students aged 12–13 years learning French. Using multiple regression analysis, it brings together psychological and sociocultural perspectives to examine the role of these factors on performance and motivation. In addition, it explores if these same factors are also significant in terms of students’ responses to explicit instruction to teach students the strategies they need to operate autonomously. Findings suggest that at this early stage in their language learning career, there was no significant difference in terms of gender on performance or motivation but all bilingual students, regardless of whether they were in the control or experimental classes, outperformed their monolingual peers in listening comprehension; linguistic features of their home background appeared to provide some explanation. Furthermore, the strategy instruction was also a significant factor in determining progress. The 12–13 age range may be a critical time for students, when teaching them how to learn may reverse an otherwise downward spiral in achievement and motivation. Keywords: data analysis; learning strategies; pedagogy; modern foreign languages
Introduction The last two decades have seen something of a crisis in modern language (ML) learning in English secondary schools. It has been described AQ2 recently as: ‘close to extinction’ (Shepherd 2011), because the number of students taking French and German General Certificate in Secondary AQ1 *Corresponding author. Email:
[email protected] ISSN 0958-5176 print/ISSN 1469-3704 online Ó 2012 British Curriculum Foundation http://dx.doi.org/10.1080/09585176.2012.744326 http://www.tandfonline.com
10
15
20
25
30
35
40
2
45
50
55
60
65
70
75
80
85
M. Grenfell and V. Harris
Education (GCSE), taken at age 16, has more than halved in the last 16 years. The picture is almost identical for those studying French A-level (taken at age 18) since 1991 (Curtis and Shepherd 2009). While the new millennium has seen concern over the under-achievement of boys in ML performance (Department for Education and Skills [DfES] 2007; Sedghi and Evans 2011), social class has been largely ignored. In contrast, as Whitty (2012) points out, social class inequalities in education have been a constant feature of English education research since the early part of the twentieth century, broadening out in the 1980s to include other social differences such as gender, sexuality and race. However, within second language acquisition research the learning process has been perceived as an individual cognitive process located in the mind of the learner. The last 10 years have seen this view challenged by those for whom it is a social enterprise whereby meaning is co-constructured by interlocutors in a particular sociocultural context. Summarising the two positions, LarsenFreeman (2007, 784) comments: ‘Our field is beset by dialectics: learning versus use, psychological versus social, acquisition versus participation, and yet, it is focussing on the dynamic coupling of each pair that is likely to be the most productive’. One initiative to reverse the downward trend in the study of ML has been explicitly to teach students language-learning strategies (LLS) (Macaro and Mutton 2009; Cohen and Macaro 2007; Harris 2007; Grenfell and Harris 1999). The LLS research field itself has also witnessed a shift from psychological to sociocultural perspectives. With some notable exceptions (Wong Fillimore 1979), early studies into LLS could be characterised as being more from a psychological perspective, stemming from Rubin’s (1975) seminal study ‘What the ‘‘Good Language Learner’’ can teach us’, which sought to uncover what it is that these individuals do that separates them from their less proficient peers. O’Malley and Chamot’s (1990) seminal taxonomy identified cognitive strategies, such as inferencing, and metacognitive strategies such as planning, monitoring and evaluation. It also included a limited number of social strategies defined as those involved in interacting with others. It is only within the last decade that LLS studies have explored how what might be termed ‘the wider sociocultural context’ impacts on strategy use. In line with Gipps’ (1999) study of the sociocultural aspects of assessment, the definition of ‘sociocultural’ in the present article includes economic and political contexts, as well as social and cultural ones, thus connecting to Whitty’s (2012) discussion of social inequalities. Norton (2000), for example, explains how a seemingly shy, unforthcoming immigrant can be dismissed as a ‘poor language learner’, rather than acknowledging the power relations that distance her from unsympathetic, impatient native speakers. This restricts her opportunities to engage in
The Curriculum Journal
3
communication with peers or colleagues and thus to develop her language. Although less common, the sociocultural approach has also been extended to strategy instruction (SI), where learners are explicitly encouraged to develop the tools they need to succeed. The majority of such studies focus on the cognitive processes within the individual, examining changes in performance or strategy deployment as a result of the SI. However, in outlining his social autonomy model, Holliday (2003) takes a broader view, criticising teachers who provide ‘imperialistic’ SI based on their own alien cultural values rather than encouraging students to choose strategies according to their own personal goals. Although there is increasing evidence for the effectiveness of SI (Hassan et al. 2005), many of the studies focus on adults and university students, often learning English, rather than on young school students learning a ML such as French. Furthermore, little is known about the ‘weight’ of the SI relative to other potentially powerful factors, such as gender and background that may impact on a learner’s progress. Although socio-economic background has been shown to play an important role in determining educational achievement generally (Cassen and Kingdon 2007), to what extent does this apply specifically to ML? And do such factors also come into play in determining students’ response to the SI itself? The present article addresses these questions in the context of a programme of listening and reading SI undertaken over the course of nine months with near-beginner learners of French in two London secondary schools, each with a different socio-economic intake. A combination of qualitative and quantitative measures was used. Student progress was measured using tests. However, recognising the importance of assessing not just their performance but the students’ shifting use of strategies as they tackle a task (Manchon 2008), 27 casestudy students undertook think-alouds in listening and reading comprehension. The students also engaged in semi-structured interviews regarding their existing strategy use and their reactions to the SI. Details can be found in Harris (2007, 2006). However, the present article discusses the quantitative findings, as the focus is the relative contribution to student progress of a range of sociocultural and individual factors. Although limited, some reference will be made to the qualitative findings where they can shed light on particular issues raised. Background Although the ‘good language learner’ (GLL) was a key concept in establishing the study of learner strategies as a legitimate field of research, the approach has been criticised subsequently for two main
90
95
100
105
110
115
120
125
130
4
135
140
145
150
155
160
165
170
175
M. Grenfell and V. Harris
reasons; the first broadly psychological, the second sociocultural. What characterises the successful learner is not the number of strategies used but how they are used and orchestrated in combination with other strategies and vary according to task (Cohen and Macaro 2007). Using prior knowledge, for example, may be unreliable if the listener jumps to unwarranted conclusions, without monitoring if it makes sense. The second criticism of the GLL research was the emphasis on the individual, rather than on the group, on the cognitive rather than sociocultural. Subsequently, some studies have adopted a more sociocultural approach. Takeuchi, Griffiths, and Coyle (2007) summarise studies that discuss individual, situational and group differences, exploring the relationship between strategy use and gender, proficiency, and motivation. They point out that much of this research is inconclusive or contradictory, sometimes because the cultural contexts themselves differ. For example, Dreyer and Oxford (1996), in their study of South African students, found that females use strategies more often than males, particularly social and metacognitive strategies. In contrast, in his investigation of students in Singapore, Wharton (2000) reported that men use more strategies than women and neither Griffiths (2003) in New Zealand nor Nisbet, Tindall, and Arroyo (2005) in China found any difference. Social class, however, is not included in Takeuchi, Griffiths, and Coyle’s (2007) review. They themselves acknowledge the relationship between a wide range of both individual and sociocultural factors in their plea that: . . . future research, rather than trying to isolate these many variables from each other as much previous research has done, needs to develop methods for including consideration of the multiple factors which affect the ways that individual learners learn in specific situations. (Takeuchi, Griffiths, and Coyle 2007, 92)
Such critiques of the GLL perspective have implications for the other major area related to LLS – namely explicit strategy instruction. A key claim is that ‘intervening in learners’ strategic behaviour can improve learning processes and ultimate attainment’ (Cohen and Macaro 2007, 4). There is increasing consensus over the broad sequence of steps to be followed in such SI programmes, from ‘awareness raising’ through modelling and practising new strategies to evaluating strategy use (Rubin et al. 2007; O’Malley and Chamot 1990). Although initially the learners’ use of strategies is heavily scaffolded, the support is gradually removed to ensure that students can operationalise them independently in a way that is appropriate to the task in hand (Macaro 2001; Grenfell and Harris 1999).
The Curriculum Journal
5
Recognising that strategies are not deployed in a decontextualised vacuum, one aim of the systematic Evidence for Policy and Practice Information and Co-ordinating Centre (EPPI-Centre)1 review of SI (Hassan et al. 2005, 2) was to: . . . uncover differential effectiveness for different languages, different learners (school, university, adult), different stages of learning (beginner, intermediate, advanced) and different language skills (reading, writing, listening, speaking, overall ability etc). In doing so, we hoped to explore why different types of strategy training might or might not work.
While there is some evidence for the effectiveness of SI with adult learners of English, Hassan et al.’s findings highlight the need for more interventions with school-level learners learning ML and for more longitudinal studies. The large-scale ‘Learning how to learn’ (LHTL) study by James et al. (2007) provides invaluable insights into staff responses to the initiative across subjects and across primary and secondary schools. Implications of SI, specifically for ML and English collaboration at secondary-school level, are explored by Harris (2008). However, a systematic study of the impact of SI on ML students in terms of both individual and sociocultural differences has not as yet been undertaken. Oxford and Schramm (2007) argue that the psychological and the sociocultural should be complementary, explaining that a critical, sociocultural framework explores how issues of power, oppression, imperialism, and resistance come into play in the teaching and learning of second language strategies. Although there are very few of such studies, the definition connects with Whitty’s (2012) reflections on developments in the field of the sociology of education. It is with these considerations in mind that the present study was undertaken. The first step in the investigation was to examine the extent to which research into the key role of factors such as social class and gender in determining educational success across the curriculum apply specifically to secondary-school students learning ML. The study moved on to investigate the impact of SI relative to these other factors. Finally, it explored whether any potential benefits of the SI extend to all students or if they are limited to particular groups according to, for example, their gender, background or bilingual status. The next section begins by summarising research into individual and sociocultural differences in language learning success, reflecting on their possible weight compared with SI. For example, gender may be a more significant factor than the SI for the students in the study, given that more boys than girls opt out of continuing their studies of ML to GCSE level (DfES 2007; Coleman, Galaczi, and Astruc 2007; Sedghi and Evans 2011).
180
185
190
195
200
205
210
215
220
6
225
230
235
240
245
250
255
260
M. Grenfell and V. Harris
Although the evidence is grouped under key headings (‘gender’ or ‘attainment’ for example), inevitably there is some overlap between the factors. The headings also include whether each factor will or will or will not have a role to play in determining students’ response to the SI itself. To date little is known about whether there are differences in response to SI, even within the adult learning context. The majority of SI research has focused on the impact of the intervention in terms of test results or shifts in patterns of strategy use. Because factors such as gender influence progress in language learning, it seems likely that they will play a role in determining the students’ response to the SI. It may or may not succeed, for example, in reversing boys’ under-achievement. Since few studies of SI have attempted to bridge the gap between the psycholinguistic and the sociocultural, the literature review will raise as many questions as it answers, particularly in relation to students’ response to SI. Furthermore, breaking down boundaries between research fields risks failing to treat each in sufficient depth. The role of sociocultural and individual differences in language learning progress Social class The present study was based in two London schools: Moreton and West (pseudonyms). West School is a large, multi-ethnic, mixed comprehensive school in a working-class area of southeast London. Moreton School is a small, mixed, voluntary-aided catholic school, serving a predominantly middle-class population in a London suburb. The schemes of work, along with teachers’ records, indicate that generally the national curriculum levels at Moreton School are higher across the curriculum than at West School. The issue in the present study was the extent to which social class impacts specifically on the rate of progress in ML. In spite of the government’s drive to raise standards in England, social class is still the strongest indicator of educational success (Gilborn and Mirza 2000; Bynner and Joshi 2002). Eligibility for free school meals, for example, is strongly associated with low educational achievement generally (Cassen and Kingdon 2007). Social class may also be associated with the level of challenging behaviour in the classroom and high or low parental expectations. Reay (2006) vividly describes how working-class pupils may be marginalised by teachers’ behaviour towards them. Little is known about the links between social class and ML performance. However, it seems likely that the students’ attitudes towards language learning, whether positive or negative, will play a role in their progress, especially in the light of Bartram’s (2006) study of the importance of parental influence on attitudes to language learning. Moreton School students may have the advantage of holidaying in France or Spain and
The Curriculum Journal
7
hence may be more motivated, since they see the relevance of learning a new language (Buttjes and Byram 1991). Motivation is a complex controversial phenomenon; for example, whether it is a psychological attribute or a sociological condition and how the unit of analysis, once defined, can be researched. In contrast to Gardner and Lambert’s (1972) comparison between instrumental and integrative motivation, Do¨rnyei (2001) emphasises its dynamic nature as it changes over time and also the key influence of the learning situation and classroom ethos. As early as 2002, Chitty (2002) highlighted the link between social class and motivation, pointing out that students opting out of language learning at age 14 are much more likely to come from inner-city schools. The trend towards languages as an elitist subject has continued (CILT 2006) and the Guardian warns that the study of ML is becoming a privilege largely restricted to private schools (Curtis and Shepherd 2009). It seems possible that the more positive attitudes of the middle-class students may also make them more willing to engage in LHTL. Qualitative data exploring motivation issues in more depth are discussed in Harris (2006) and Harris and Prescott (2005). The focus here is on the quantitative evidence to weigh up the potential impact of social class on these students’ progress and attitudes towards language learning, relative to the other factors.
270-
275
280
285 Gender The last 20 years have been marked by a preoccupation with the underachievement of boys across many areas of the curriculum (Department for Education [DfE] 2010, 2007). It has been evident in ML in England (Nuffield 2000; Ofsted 2004), Australia (Carr and Pauwels 2005) and Canada (Kissau 2006). The problem in England has been further exacerbated since students are allowed to opt out of studying a language after age 14. In 2004 the Chief Inspector for Schools had already expressed his concern that French and Spanish might become the preserve of middle-class girls (Bell 2004). The situation has continued to deteriorate (Rodeiro 2009). Figures regarding the gender imbalance in ML in England are often drawn from a comparison of GCSE results. However, it is not clear at what age such disenchantment on the part of boys occurs and whether the SI might contribute to reversing the process. Furthermore, the link between gender, motivation and attainment has to be taken into account. Given boys’ under-achievement, the female students may be more successful language learners and more motivated than their male peers, prior to the intervention. They may, therefore, be more receptive to the SI. On the other hand, Harris et al. (2001) point out that high attainers of either gender can become complacent, believing that their existing use of strategies already guarantees them success. The possibility that boys in particular might benefit from the SI is increased in
290
295
300
305
8
310
315
320
325
330
335
340
345
350
M. Grenfell and V. Harris
the light of the study by Jones and Jones (2001). This suggests that boys value explicit discussion of what and why they are learning and how to approach it. They also benefit from developing greater skills of selfregulation (Watkins 2007). Thus, the issue is whether these 12–13-year-old boys are already under-achieving and whether the SI can reverse the downward trend? This may apply particularly to reading, compared with listening, since Shapiro (2007) suggests that girls’ superior attainment is more pronounced for skills linked to literacy, even in their mother tongue (L1). Bilingualism Definitions of bilingualism are complex and include considerations of when the languages are learned (in parallel or consecutively), whether learners are first, second or third generation, different levels of proficiency within the skill areas and the support and prestige offered for their development in the school and in the community (Anderson 2008; Cummins 2000). For many years, bilingual students were considered problematic for the special needs they represented; competence in the home language was seen as interfering with the development of English. The view was discredited by Cummins’ ‘common underlying proficiency model’, which argues that although on the surface the two languages are different, underneath they are fused in a single, central, processing system AQ3 (Baker 2006). Nevertheless, some groups do under-perform and the start of the millennium saw a raft of initiatives to support them (DfES 2004, 2006). Uncovering the reasons for their under-performance is particularly complex since bilingualism is linked to class and race and there is considerable variation in educational achievement between ethnic groups (Gilborn and Mirza 2000). Within ML, there is a growing body of evidence that bilingualism facilitates the acquisition of a third language (see Jessner [2008] for an overview of current findings). However, in light of Norton’s (2000) and Cummins’ (2000, 2001) research, it seems likely that a complex range of factors may be involved in the progress of bilingual students learning French in these English secondary schools, including not only their general proficiency level, ethnic group, home environment and sense of identity but also the schools’ ethos and policies in relation to recognising and validating that identity. It is, therefore, hard to predict the likely impact of the SI weighted against the role of the students’ bilingual or monolingual status itself. Do bilingual students make more progress as they have already developed some of the strategies? Or for that very reason do they, therefore, perceive the SI as a ‘waste of time’? Do they indeed make the connections between the ‘natural’ home language learning environment and the school classroom? In summary then, the question is whether the students’ bilingual status facilitates their progress
The Curriculum Journal
9
in ML and if bilingual or monolingual status is a factor in determining their response to SI. 355 Attainment ‘Attainment’ is the term used by the English government to capture the learner’s level at a particular moment against set criteria regardless of age. Hence, for example, their concern to analyse over time student attainment AQ4 by gender (DfE 2010) and by minority ethnic group (DCSF 2007). Attainment thus expresses the outcome of the interaction of a number of complex factors, including those already discussed, such as motivation. While some of these factors might fluctuate over time, it seems likely that a student’s existing attainment level is a significant factor in determining their progress; successful high attainers making greater progress over the course of the year than low attainers. Furthermore, high attainers’ wider vocabulary may also enable them to benefit most from the SI. A common distinction in terms of reading and listening strategies is made between ‘bottom-up’ strategies such as word-for-word translation and more global ‘top-down’ strategies such as inferencing. Macaro, Graham, and Vanderplank (2007) argue that it may be that learners’ lexical knowledge needs to exceed a certain ‘threshold level’ if they are successfully to exploit top-down strategies such as prior knowledge. So for some low attaining students in the present study, the SI may simply be ‘beyond’ them. However, the evidence is inconclusive. In one of the few studies of individual differences in response to SI, Ikeda (2006) reports on its impact on 210 Japanese university English as a foreign language (EFL) students. She noted that for the lower proficiency group the intervention was not effective enough to make them change their strategy use and surmised that they did not have enough EFL reading ability to make use of the topdown strategies that were the main focus of the SI. In contrast, Kusiak’s (2001) study showed particularly positive results for the lower proficiency group and Van Der Grift’s (2007) weaker listeners in the experimental classes made no fewer gains in listening than those in the control class. The possible negative impact of a ‘threshold level’ has to be balanced against low attainers’ increased motivation, since there is some evidence that SI enhances motivation (Chamot et al. 1996). Rubin (1990, 282) highlighted its potential benefits for low attainers, linking it to motivation: ‘Often poor learners don’t have a clue as to how good learners arrive at their answers and feel they can never perform as good learners do. By revealing the process, this myth can be exposed’. Thus, the low attaining students’ limited lexical repertoire might be compensated for by their increased motivation. Baines, Blatchford and Chowne’s study (2007) may be of some interest here, even though it was not concerned with teaching LLS specifically. Rather, the focus was on
360
365
370
375
380
385
390
395
10
400
405
410
415
420
425
M. Grenfell and V. Harris
collaborative group work skills, the aim being to explore the impact of training primary-school students to develop them. They found that low, middle and high attainers made equal progress over the year; nor was there any difference according to gender. In summary, then, the issue here is to establish whether, in spite of their attainment level, low attainers can make as much progress as high attainers and whether they can benefit from the SI. The present study thus sought to explore the following research questions: (1) What are the potentially powerful factors such as socio-economic background, gender, bilingual status, and attainment, which make a significant difference to 12–13-year-old students’ performance and motivation in ML? (a) How strong a predictor of success is social class specifically in ML? (b) Do girls already out-perform boys at the age of 12–13? (c) Is students’ bilingual status a problem or an advantage? (d) To what extent does prior attainment predict future performance? (2) What is the impact of SI relative to these other factors? Does it make a significant difference, and if so, does it play more or less of a role than gender or background? (3) Within the group exposed to SI, what role do these other factors play, if any? Do some sub-groups benefit more from the SI than others? (a) Are middle-class students more motivated to ‘learn how to learn’? (b) Do boys respond positively to explicit discussion about ‘how to learn’? (c) Do bilingual students perceive SI as unnecessary since they are already familiar with LLS? (d) Do low attainers in particular value learning the tools they need to be more successful?
430
435
440
Research methods Project participants and school context Data were not collected about each individual student’s socio-economic background but Ofsted2 reports indicate that West School has a higher number of students entitled to free school meals than the national average and a higher number of students with special educational needs (SEN). Standards on entry are below the national average. In contrast, Moreton School has fewer students with SEN than the national average and students’ attainment level on entry is above the national average. The
The Curriculum Journal
11
national curriculum levels in Moreton School are higher than for West School, suggesting, in relation to the earlier discussion of a ‘threshold’ level, that the learners may have a larger lexical repertoire. The investigation was ‘quasi-experimental’ in design, using intact class groups: in each school, two parallel classes of 30 students aged 12–13 years learning French. The classes had been designated by the schools as ‘top sets’. As is usual in experimental research such as clinical trials, there was a control and an experimental class. Both classes followed the ML scheme of work in each school but the experimental classes were also exposed to an explicit SI programme (see Harris [2006, 2007] for a detailed description of the intervention). Thus, it was considered ethically acceptable, as it was a question of possible ‘added value’ for the experimental classes and the control classes would not necessarily be disadvantaged. Parental permission was sought and granted. The students had begun to learn the language at the start of the previous year, with approximately two hours a week of lessons; thus approximately 80 hours of language learning prior to the intervention. Within the 120 students across the control and experimental classes, 51% of the students were female in each school; 33% of the students at West School and 3% at Moreton School (i.e. 18% across both schools) were designated as ‘bilingual’. For the study’s purpose, the designation of the students as ‘bilingual’ was mainly determined by the school’s ‘English as an additional language’ list. Students between stages 3 and 5 in English on the list had been educated in the UK from birth or an early age and a question in a pre- and post-intervention attitude questionnaire established if they were regularly exposed to another language within the extended family and community networks. They were thus operating in two languages from an early age. They came from a wide range of cultural backgrounds, typical of London, including Vietnamese, Somali, Turkish, Yoruba and Tamil (National Literacy Trust 2006). There was, therefore, not a particular dominant group in terms of the link to social class. While there was a risk of both teachers and pupils in the experimental classes sharing experiences with those in the control classes, it was decided to locate the classes in the same schools, rather than different schools, in order to maximise parity of socio-economic background. The experimental classes were taught by the two teachers working on the project and the control classes by another experienced teacher in each ML department. The SI lessons were developed through close collaboration between one of the authors and the teachers (Harris 2008; Harris and Prescott 2005) and was based on the model of four steps described earlier to withdraw the support gradually until the students could use the strategies independently. Given the need for extensive practice in the strategies involved in each skill area, it was decided there was insufficient time to deal with all four
445
450
455
460
465
470
475
480
12
M. Grenfell and V. Harris
language skills in sufficient depth over the course of the project. A pilot study conducted with the teachers finalised the choice of reading and listening as the skill areas (Harris 2007) since the students reported finding listening comprehension one of the hardest skills and reading was felt to be a useful, related ‘way in’. Both are receptive skills but reading strategies might be more accessible because the learner has greater control 490of the process, as they have the time to read and re-read. Furthermore, SI in these skill areas has been more extensively researched than speaking AQ5 (Cohen and Macaro 2007). Appendices 1 and 2 list the strategies in both skill areas. In relation to each specific skill, Oxford et al.’s (2004) revised version of Ikeda and Takeuchi’s reading questionnaire was drawn on to 495 identify the reading strategies. Their focus on task-based strategy assessment overcomes some of the vagueness of generic strategy taxonomies (‘advance organisation’ or ‘inferencing’ for example) and provides clear descriptions specific to reading (‘I use the title to help predict the contents’, ‘If I don’t understand something, I guess its 500 meaning using clues from the text’); a particularly important factor given the students were only 12 years old. Similar reasons guided the selection of listening strategies from Van Der Grift (1997, 2003), since while the terms remain generic, definitions and representative examples are given. 505 AQ6 Bearing in mind Cohen, Weaver, and Li’s (1998) observation of the difficulties learners experience with unclear strategy definitions, the wording was adapted and some strategies ‘unpacked’ into sub strategies to make them more accessible to the pupils. In addition, where possible, the wording was similar across listening and reading strategies to encourage transfer between the two skills. In the overall selection of the 510 strategies from the aforementioned taxonomies, care was taken to ensure that they included both metacognitive and cognitive strategies since, as Graham, Macaro, and Vanderplank (2007, 182) suggest, metacognitive strategy development appears to be a common feature of the more successful SI listening programmes. Oxford et al. (2004) highlight 515 metacognitive planning, monitoring and evaluation by listing strategies under ‘before/while/after reading’ and this was adopted for the layout of the strategy checklists given to the students. The selection included both ‘top-down’ and ‘bottom-up’ listening and reading strategies to reflect the 520 complex parallel interaction between world knowledge (topic, genre, etc. to build a conceptual framework for comprehension) and linguistic knowledge or compensatory strategies to cope with lack of such knowledge. The SI initially focussed on reading strategies rather than listening 525 strategies, as they are easier to model and arguably easier to use, since the learner has time to reflect on the areas of difficulty and return to them (Grenfell and Harris 1999). Over a nine-month period, the experimental classes were taught 25 lessons or parts of lessons incorporating SI; the
The Curriculum Journal
13
control class was merely exposed to the same reading and listening texts.3 The importance of a gradual approach, in which clusters of strategies are first presented in familiar contexts, and the need for even more practice than had been anticipated are discussed in Harris and Prescott (2005), and an analysis of the qualitative data on the students’ opinions of the SI in Harris (2007).
530
535 Research measures In order to examine whether SI results in improved performance, a reading and listening test in French was administered pre- and postintervention. Since there are no national standardised tests at this level (Macaro and Erler 2005), the tests first had to be piloted. Space does not permit a discussion of the difficulties raised in devising the tests but some of the issues are discussed in the EPPI-Centre review (Hassan et al. 2005). For example, it was important to ensure that the nature of the tests allowed all the strategies to be used. Whereas the reliability of the listening test, estimated by Cronbach’s alpha, was high (0.839), it was slightly lower for the reading test (0.703). The tests were marked by the research officer, according to agreed criteria, and a sample marked by one of the authors. Inter-rater reliability was high: 0.995 for the reading test and 0.986 for the listening test. Drawing on Chambers (1999), who explored motivation with cohorts of ML learners of a similar age in England, a general attitude score was constructed by taking the mean of responses to Question 1 (‘How do you feel about learning a new language like French?’) and Question 2 (‘How hard do you think you try to learn French?’). While somewhat crude as a measure, this score was used to give a broad indication of level of motivation. Semi-structured interviews and think-aloud protocols were carried out with 27 case-study students of different attainment levels and of both monolingual and bilingual status. A detailed description of the findings is not included in the present article but they are discussed by Harris (2007, 2006). Research methodology Separate independent samples t-tests and between-groups analyses of variance (ANOVAs) could have been used to assess the impact of SI on the performance of the experimental classes compared with the control classes. However, the study sought to examine the role that factors other than the SI played in students’ progress, and their relative influence. Multiple regression analysis (MRA) allows comparison of an intervention’s impact with that of other factors determining students’ progress
540
545
550
555
560
565
570
590
595
600
605
610
615
(þ) 0.05* Experimental group School R
(þ) 0.16*** 0.01
0.01 (þ) 0.05* 0.01 (þ) 0.06* (þ) 0.06* Female
0.02
Bilingual
(þ) 0.03{
0.01
(–) 0.04*
Bilingual
(þ) 0.27*** (þ) 0.03{ (þ) 0.39*** (þ) 0.06* 0.01 Higher score
Prior listening
(þ) 0.05* (þ) 0.24*** (þ) 0.25*** 0.03 Higher score
Prior reading
585
(þ) 0.03{
(–) 0.03{
(–) 0.17*** (þ) 0.40*** (þ) 0.13***
Gender
(þ) 0.07** (þ) 0.06* (þ) 0.47*** Higher score
Prior attitude
Summer listening
(þ) 0.14*** (þ) 0.05* Higher score
580 (þ) 0.06* Higher score
Summer reading
Notes: (þ) and (–) indicate direction of effect (i.e. sign of correlation before it was squared); empty squares indicate r2 5 0.010; {p 5 0.100; *p 5 0.050; **p 5 0.010; ***p 5 0.001.
School Gender Bilingual Prior listening Prior reading Prior attitude Summer listening Summer reading Summer attitude Higher is:
School
575
Experimental group
Table 1. Squared correlations (r2) between predictors and each other, and with summer scores.
14 M. Grenfell and V. Harris
The Curriculum Journal
15
(Brantmeier 2004; Hatch and Lazarton 1991). Since the aim of this aspect of the present study was to explore the relative ‘weight’ of the SI, the effect of being in the experimental group was one of the factors included in the regression. The other factors included for comparison were: students’ prior attainment (their autumn score on the same test); school attended (as an indication of socio-economic background); gender; bilingual status; and the autumn combined attitude score (as an indication of prior motivation). To explore whether the SI was especially valuable depending on any of these variables (for example, whether it was particularly useful for students at one school), the interaction of each of these factors with the SI was also included in the regressions. The outcome measures were the students’ gain scores (from autumn to summer) in the reading and listening test and students’ gain scores in the attitude questionnaire over the same period; the focus being on progress made rather than on final attainment level. Regression and correlation analyses were run in Statistical Package for the Social Sciences (SPSS). The regressions were hierarchical, showing separately how much of the total variance was explained by the intervention and its interactions; the default ‘Enter’ method was used at each of the two hierarchical stages. In line with Hatch and Lazaraton’s (1991) guidelines, the linearity of relationships between predictors and dependent variables was examined, co-linearity diagnostics were satisfactory and there were no strong indications of heteroscedacity in the residuals plots. Results AQ5 Selected descriptive statistics are shown in Tables 1 and 2. Numbers of cases included in each analysis are shown in Table 3.4 Table 4 (column A)
Table 2.
620
625
630
635
640
645
Descriptive statistics. Experimental group
Control group
Listening test autumn Listening test summer Listening test change Reading test autumn Reading test summer Reading test change Attitude autumn Attitude summer Attitude change
650
Overall
n
Mean
SD
n
Mean
SD
n
Mean
SD
48 48 48 47 47 47 53 46 46
50 56 6 54 72 18 3.62 3.20 –0.41
17 16 12 20 18 17 0.66 0.84 0.61
52 52 52 60 60 60 61 53 53
48 62 14 48 74 26 3.69 3.57 –0.16
15 14 14 18 15 18 0.75 0.74 0.58
100 100 100 107 107 107 114 99 99
49 59 10 50 73 22 3.66 3.39 –0.28
16 15 14 19 16 18 0.71 0.81 0.60
655
660
695
700
99 107 99
Listening change score Reading change score Attitude change score
29%*** 46%*** 6%***
0.41 0.85 0.07
Cohen’s f2 10%** 5%* 4%*
18%** 6%{ 12%{
Variance explained (R2)
0.33 0.13 0.15
Cohen’s f2
47%*** 52%*** 18%***
Variance explained (R2)
0.87 1.09 0.23
Cohen’s f2
Final model
Notes: f2 effect sizes of 0.02, 0.15 and 0.35 are considered small, medium and large, respectively (Cohen 1988); {p 5 0.100; *p 5 0.050; **p 5 0.010; ***p 5 0.001.
n
Dependent variable
685 Variance explained (R2)
Effect of adding group and interactions
675
Variance explained (R2)
680
Main effect of group
670
Model before including group
665
AQ34
690
Table 3. Overall effect sizes of regression models.
16 M. Grenfell and V. Harris
22%*** 20%** –0.64 ^Lower achievers 36%*** 20%** –0.55 ^Lower achievers
10%** 7%** 0.27 ^Experimental group 5%* 2%{ 0.13 ^Experimental group
Reading gain Variance Zero-order (n ¼ 107) explained: Semi-partial Beta
Gender 1% 0% –0.01
1% 2%{ 0.15 ^Girls 0% 1% 0.12
School 8%** 2%{ 0.22 Zero-order: ^West school Semi-partial: ^Moreton 22%*** 4%** –0.25 ^West School 1% 0% 0.07
0% 0% 0.05
0% 3%* –0.21 ^Low autumn scores
1% 4%** 0.24 ^Higher attitude 0% 0% 0.04
5%* 6%* –0.27 Less decline for lower prior attitude
3%{ 1% 0.11 ^Bilingual 0% 0% –0.05
Prior attainment (autumn score)
5%* 4%* 0.22 ^Bilingual
Bilingual
Prior attitude
1% 0% 0.04
1% 0% 0.08
0% 1% 0.10
0% 0% 0.03
0% 0% 0.05
1% 0% –0.03
0% 0% 0.02
2% 4%* 0.22 ^Moreton School
3%{ 1% –0.13
1% 1% 0.12
0% 0% 0.03
Prior Gender Bilingual attitude
4%* 5%** –0.33 ^West School
School
D. Possible factors predicting progress within the experimental classes; interaction with experimental group
Notes: Variances explained are based on unadjusted R2; ^indicates which group or subgroup made more progress; {p 5 0.100; *p 5 0.050; **p 5 0.010; ***p 5 0.001.
Attitude/ Variance Zero-order 4%* motivation explained: Semi-partial 5%* change Beta 0.24 (n ¼ 99) Less decline in experimental group
Listening gain (n ¼ 99)
Variance Zero-order explained: Semi-partial Beta
B. Effect size
Prior attainment (autumn score)
Experimental group
A. Outcome measure and sample size
C. Main effects. Possible factors predicting progress across control and experimental classes
Table 4. Effect of individual predictors in the regression models.
The Curriculum Journal 17
710-
715
720
725
730
735
740
745
18
750
755
M. Grenfell and V. Harris
lists the measured outcomes in terms of progress in listening and reading comprehension and in positive attitude. Column C indicates the relative weight of each of the factors to contributing to the students’ progress. Thus, across the control and experimental classes, it shows the contribution of students’ gender to their progress, for example, as well as of the SI itself. In contrast, Column D, ‘Possible factors predicting progress within experimental classes’, does not apply to the control classes and only indicates any differences in response to the SI according to prior attainment, gender or bilingual status. Significant results are indicated by asterisks, non-significant trends by a plus sign.5
760
765
770
775
780
785
790
Progress in listening The MRA model explains 47% of variance in the listening gain score. In other words, all the factors included in the analysis, such as the SI, gender and prior attainment, account for 47% of the differences between the students’ gain scores. In terms of the first research questions as to the relative weight of the factors and their relationship to the SI, Column C in Table 4 shows that the following are significant predictors of the gain in listening scores in order of effect size, as measured by the zero-order correlation: (1) Prior attainment (as indicated by the autumn term scores) (22%); low attainers making the most progress over the course of the year (2) Membership of the experimental group (10%) (3) School attended (8%); West School students gaining more (4) Bilingual status (5%)6 Thus, across control and experimental groups, and regardless of the SI, low attainers made the greatest progress. The higher the score in the autumn term, the less the gain by the summer term; a histogram revealed that this was not because of a ceiling effect. Socio-economic background, in terms of school attended was also a significant predictor. Even though the autumn term scores of pupils in West School were below those of Moreton school, and were still below them in the summer term, they made significantly more progress. Of further interest is that bilingual students (again across both the experimental and control classes) achieved greater gain scores than their monolingual peers (Figure 1). Finally, the data in column C also suggest that gender does not play a significant role in contributing to the gain scores in listening at this stage in the students’ language learning career. It is encouraging to note, however, that the intervention was a significant factor in contributing to progress in listening, with students in the SI group gaining an average of 14% compared with 6% in the control group.
The Curriculum Journal
19
Turning to the third research question, within the sub-groups of the experimental classes (Column D), it seems that the listening strategy instruction benefitted all students regardless of their prior attainment or prior attitude, their gender or bilingual status. Only the school attended played a role. Figure 2 shows that students in the experimental class in West School made more progress than both those in Moreton School and
795
800
805
810
815
Figure 1. Listening score by bilingual/monolingual status.
820
825
830
Figure 2. Listening test scores by membership of experimental group and school attended.
835
20
840
845
850
855
860
865
870
875
880
M. Grenfell and V. Harris
in the control classes. It might be assumed that this was because there was a larger proportion of bilingual students in West School. However, there were similar numbers of bilingual students in both the control and experimental classes in West School and the MRA indicates that there was no significant interaction between group and bilingualism. Progress in reading The MRA model explains 52% of variance in the reading gain score. The following were significant predictors in order of effect size, indicating their relative weight in contributing to the progress of these 12–13-year-old students: (1) Prior attainment, as indicated by the autumn term scores (36%); lower attainers making more progress (2) School attended (22%); West School pupils gaining more (3) Membership of the experimental group (5%) Prior attainment and school attended appeared to play a more important role in reading gains than in listening gains, and the intervention slightly less of a role. However, a histogram suggested that a ceiling effect may be involved, so some caution must be exercised in interpreting the results. Although data in column C indicate that neither gender nor prior attitude were significant factors in ML reading progress across the control and experimental classes, there was a non-significant trend of 3% in favour of bilingual students. Unlike listening comprehension, these students did not start out at a disadvantage in their reading, but they improved more than their monolingual peers (Figure 3). Finally, the semi-partial correlations suggest a non-significant trend of 2% þ in favour of girls, suggesting that even at this early stage in their learning career, they were making more progress, albeit limited, than the boys in reading comprehension. Within the experimental classes (column D), the reading strategy instruction benefitted all students regardless of their socio-economic background, gender, bilingual status or prior attitude and attainment. It is only the semi-partial correlation that suggests that students with low autumn term scores were more likely to benefit from the reading instruction than high attainers. Progress in positive attitude While still significant, only 18% of the variance in the attitude gain score was explained by the MRA model (compared with 47% for listening and 52% for reading); factors other than social class, gender, bilingual status or the SI accounted for the remaining 82%. This may reflect the brevity of
The Curriculum Journal
21
the questions, providing only a crude measure of students’ attitudes, and the same reason may account for its lack of significance as an independent variable. Figure 4 shows that the attitudes of all students towards their language learning was already declining at the end of their second year of language learning, although at this stage neither the students’ socioeconomic background, nor their gender, nor bilingual status played a role in this decline. It is encouraging to note that the attitudes of students in
885
890
895
900
905 Figure 3. Reading gain score by monolingual and bilingual status.
910
915
920
Figure 4. Attitude scores by group.
22
930-
935
940
M. Grenfell and V. Harris
the experimental classes are less likely to become negative than those in the control classes. A histogram suggested that this was not because of a floor effect. Turning to the responses of the experimental classes to the SI, Figure 5 indicates that there was a non-significant trend of 3% for the interaction between experimental group and prior attitude. There was thus a tendency for students in the experimental classes with very low autumn attitude scores to increase their scores slightly by the summer. In the control group there was no such tendency. The semi-partial correlation of 4% in Table 4 column D suggests that students in Moreton School were less likely to develop negative attitudes than those in West School. Discussion The limitations of the present study are apparent. The sample sizes do not quite reach the general guidelines and devising appropriate tests proved challenging. The ‘Hawthorne effect’ may have had a role to play; involvement in the intervention rather than its particular characteristics
945
950
955
960
965
AQ35 Figure 5. Attitude change by group and prior attitude.
The Curriculum Journal
23
possibly determining the greater progress of the experimental classes. Nevertheless, the findings suggest some useful avenues for further research with larger cohorts of students, of varying ages and backgrounds and learning languages other than French. Although the focus of this article is on the psychological and sociocultural aspects of the quantitative data, some reference will be made to the qualitative data, where appropriate. The results appear promising, at least for this group of students. The first research question set out to explore the potentially powerful factors that might make a difference to the level of progress and motivation in 12–13-year-old students across the control and experimental classes. The issue in relation to socio-economic background was its impact specifically on ML. In spite of the fact that student behaviour in West School is more challenging than in Moreton School, and that one might surmise that their working-class parents might have lower expectations and be less likely to take holidays abroad, it was the West School students across the control and experimental classes who made greater gain scores in listening (the possible ceiling effect making it difficult to draw a similar conclusion in relation to reading). The reasons are not apparent and further research is needed in a range of schools serving different populations. Furthermore, it was low attainers across both schools who made the most progress, suggesting perhaps a need for differentiated approaches to stretch high attainers. Finally, West School students appeared to benefit more from the listening SI. A possible explanation is offered when comparing the success of the reading and listening SI later in this section. Like socio-economic background, assumptions about the role of gender are also challenged, since across control and experimental groups these 12– 13-year-old girls did not make more progress than boys in listening and there was only a non-significant trend in the semi-partial correlation for reading. Further studies are needed to establish whether, at this relatively early stage in their language learning, factors such as social class and gender have not yet significantly impacted on students’ progress and are more likely to come into play when students reach the 13–14 age range. Turning to the bilingual students, the study confirms other investigations indicating that knowing several languages can be an asset not a problem. It could be argued that students should have been utilising their heightened awareness to achieve more than their monolingual peers already in Year 7, particularly in terms of their listening skills. However, the limited level of ML classroom language in that initial year may be too simple for them to make the connections to rich exposure in the home environment. Over the course of the following year, even without deliberately realising it, they may have started to transfer their listening strategies to the classroom. The non-significant trend in relation to their
970
975
980
985
990
995
1000
1005
1010
24
1015
1020
1025
1030
1035
1040
1045
1050
1055
M. Grenfell and V. Harris
reading progress (3%), compared with the significant correlation in listening (5%) would appear to support the connections with the aural skills developed in the home environment, since, as Anderson (2008, 80) points out: ‘ . . . although comprehension skills may be well developed in certain contexts, this does not necessarily imply confidence in speaking or indeed literacy skills’. The interviews with the bilingual case-study students highlight the impact of the home environment (for a detailed description see Grenfell and Harris [2007]). For example, Martelle talked about how she infers meaning: I use the words that I do know, and the words that I don’t, I put the ‘something’ there instead of it. Like if they were saying ‘I work in’, I would think of what sort of person they are, what their job could be, or I close my eyes, and sort of bring my spirit out, and get myself into that word, what it can mean.
Asked how she learned to use this strategy, she explains how her mother told her what to do if she did not understand what she was saying in Patois: ‘Well I just picked it up, ’cause my mum always said to me, if you just think of all the different possibilities it could be, jumble them up, and that’s what I done’. The home environment then means that bilingual students may be faced with more exposure to working out meanings than their monolingual peers (Grenfell and Harris 2007). This may be one reason underlying the significant difference between monolinguals and bilinguals in the present MRA results. With regard to the second and third research questions regarding the relative weight of the SI, and whether it benefitted all students, it is encouraging that it had a significant impact on both reading and listening performance, for boys as well as girls, and for bilingual as well monolingual students. Qualitative data shed some light when comparing the impact of the SI across the two skill areas. While both of the variances explained are slight, the relative weight of the intervention differs; listening (10%) being slightly greater than reading (5%). The reasons are unclear, especially since a questionnaire for the experimental classes (Harris 2007) indicates that 30 pupils (55%) reported that the reading SI had helped them most, compared with 10 (19%) reporting that the listening SI had helped most, and 14 (26%) indicating they had helped ‘the same’. However, the listening SI may have made more of an impact as both teachers and students were more attuned to it, having already engaged in the reading SI prior to the listening SI. A further reason may be that students found listening more difficult than reading and hence having explicit instruction on how to tackle listening texts was more
The Curriculum Journal
25
useful, even if many of the students were not themselves aware of it. As one student commented: ‘listening strategies helped me not to panic’. This may explain why West School students appeared to have gained more in terms of listening than Moreton School. Possibly, students from the middle-class background are more exposed to complex language in their L1, developing valuable inferencing strategies which they transfer to their second language. One student, for example, referred to being obliged to sit and watch the news with his parents and listening to their subsequent discussion over dinner: ‘There are words and stuff I don’t understand but I kinda figure it out from the pictures on TV and the way my Dad is always on about the same things’. It is interesting to note that there was no significant difference in students’ responses to the SI according to gender. The girls’ higher level of motivation in language learning may have been counterbalanced by the boys’ appreciation of the explicit nature of the instruction, like those in the research of Jones and Jones (2001). However, it is also possible that the general decline in motivation becomes more marked for boys than for girls from the age of 13 onwards. Given that the SI appears to reduce this decline, integrating it into the curriculum throughout the entire span of a student’s education might contribute to reversing the achievement gap between boys and girls. Such a move may also be of benefit to low attainers, since they, like high attainers, were able to access the SI. Indeed, the significant semi-partial correlation for reading would suggest that low attainers in particular benefited from the SI, and it appeared to have had a particularly positive impact on their attitudes. Understanding exactly how to go about their language learning may have, as Rubin (1990) suggested, served to demystify the process.
1060
1065
1070
1075
1080
1085 Conclusion In bringing together both psychological and sociocultural perspectives, the present study contributes not only to an understanding of the impact of SI but also of other factors influencing the progress of 12–13-year-old students learning French. At this early stage in their learning, gender appears to be less influential than expected. In contrast, being bilingual may be an advantage that has not yet been fully recognised. One strand within more recent LLS research notes that being a ‘good language learner’ is not sufficient; there must be opportunities to deploy the strategies (Norton and Toohey 2001). If bilinguals’ ‘good’ language learning skills are not valued in school, then they may not fulfil their true potential. It is promising that SI appears to play a significant role in determining the progress of young school students aged 12–13 learning a ML. Given the high rate at which students in England are opting out of ML after age 14, it appears that there is all to play for in the preceding two
1090
1095
1100
26
M. Grenfell and V. Harris
years. Learning how to learn languages may be one pedagogical approach that is effective with all students and may contribute to encouraging them to continue their ML studies. 1105 Acknowledgements We are very grateful for the grant from the Society of Educational Studies which made the study possible. We are also indebted to Mike Griffiths, Goldsmiths College for his considerable expertise in the field of statistics and his patient but rigorous support throughout.
1110 Notes
1115
1120
1125
1130
1135
1. The EPPI-Centre was established to develop a systematic approach to the organisation and review of evidence-based work. Its work and publications engage health and education policymakers and practitioners in discussions about how researchers can make their work more relevant and how to use research findings. 2. Ofsted is the main inspection agency in the UK and is charged with raising education standards through its activities: http://www.ofsted.gov.uk/ 3. Explicit reference to strategies is unusual in textbooks at this level in England. Rather, both textbooks and teachers see reading and listening activities essentially as testing rather than teaching opportunities 4. It should be noted that the sample sizes do not reach the general guidelines recommended in Tabachnick and Fidell (2007), where m is the number of predictors: 138 (50 þ 8m) for the overall model and 115 (104 þ m) for individual predictors. This may limit the power of the analysis to find statistically significant effects. On the other hand, no corrections have been made for multiple comparisons, which may inflate significance levels. Table 3 also shows the variance explained by each of the models, and how much of this is explained by being in the experimental group and/or by its interactions. 5. Multiple regression comparison can be misleading if factors correlate with each other (Tabachnick and Fidell 2007). It can be useful if studies indicate both the squared simple correlation (otherwise known as the zero-order correlation [representing each factor’s total contribution to the outcome]) as well as the squared semi-partial correlation (representing its unique contribution), as in Table 4. An unusual complication that occurs in some studies is that the semi-partial correlation can be higher than the zero-order correlation, or even pull in opposite directions, if other variables in the analysis are acting as ‘suppressor’ variables; that is, if they remove variance from that predictor which is unrelated to the outcome (Tabachnick and Fidell 2007). 6. With the exception of ‘school’, representing socio-economic background, the semipartial correlations indicate a similar relative weighting of the factors. The other exception is that only the semi-partial correlations suggest that students with a more positive attitude from the outset made greater progress.
AQ7 Notes on contributors References 1140
Anderson, J. 2008. Towards an integrated second-language pedagogy for foreign and community/heritage languages in multilingual Britain. Language Learning Journal 36, no. 1: 79–89. Baines, E., P. Blatchford, and A. Chowne. 2007. Improving the effectiveness of collaborative group work in primary schools: Effects on science attainment. British Educational Research Journal 33, no. 5: 663–80.
The Curriculum Journal
AQ8
AQ9 AQ10 AQ11
AQ12
AQ13
AQ14
27
Bartram, B. 2006. An examination of perceptions of parental influence on attitudes to language learning. Educational Research 48, no. 2: 211–21. Bell, D. 2004. A new paradigm for modern foreign languages? Language World Winter . Brantmeier, C. 2004. Statistical procedures for research on L2 reading comprehension: An examination of ANOVA and regression models. Reading in a Foreign Language 16, no. 2: 51–69. Buttjes, D., and M. Byram. 1991. Mediating languages and cultures. Clevedon, UK: Multilingual Matters. Bynner, J., and H. Joshi. 2002. Equality and opportunity in education: Evidence from the 1958 and 1970 birth cohort studies. Oxford Review of Education 28, no. 4: 405–25. Carr, J., and A. Pauwels. 2005. Boys and foreign language learning: Real boys don’t do languages. : Palgrave Macmillan. Cassen, R., and G. Kingdon. 2007. Tackling low educational achievement. : Joseph Rowntree Foundation. Cenoz, J., and J.F. Valencia. 1994. Additive trilingualism: Evidence from the Basque country. Applied Psycholinguistics 15, no. 2: 195–207. Chambers, G. 1999. Motivating language learners. Clevedon, UK: Multilingual Matters. Chamot, A.U., S. Barnhardt, P. El Dinary, and J. Robbins. 1996. Methods for teaching learning strategies in the foreign language classroom. In Language learning strategies around the world: Cross-cultural perspectives, ed. R.L. Oxford, . Manoa, HI: University of Hawai’i Press. Chitty, C. 2002. The inclusive curriculum: An education for the benefit of all young people? Forum 44, no. 3: 99–102. CILT. 2006. Language trends 2006. London: CILT, the National Centre for Languages. http://www.cilt.org.uk/research/languagetrends. Cohen, A.D., and E. Macaro. 2007. Language learner strategies: 30 years of research and practice. Oxford, UK: Oxford University Press. Coleman, J.A., A. Galaczi, and L. Astruc. 2007. Motivation of UK school pupils towards foreign languages: A large scale survey at Key Stage 3. Language Learning Journal 35, no. 2: 245–80. Cummins, J. 2000. Language power and pedagogy: Bilingual children in the crossfire. Clevedon, UK: Multilingual Matters. Cummins. 2001. Negotiating identities: Education for empowerment in a diverse society. 2nd ed. Ontario, CA: California Association for Bilingual Education. Curtis, P., and J. Shepherd. 2009. Modern languages remains barrier for state pupils. The Guardian, January 15. http://www.guardian.co.uk/education. Department for Education. 2007. Minority ethnic pupils in the longitudinal study of young people in England. DfE. http://www.education.gov.uk/publications. Department for Education. 2010. Neighbourhood statistics: Small area pupil attainment by pupil characteristics in England. DfE. http://www.education.gov.uk/publications. Department for Education and Skills. 2004. Aiming high: Supporting effective use of EMAG. Nottingham, UK: DfES Publications. Department for Education and Skills. 2006. Raising expectations for ethnic minority students. DfES. www://findoutmore.dfes.gov.uk/2006/07/ethnic_minority.html. Department for Education and Skills. 2007. Gender in education: The evidence on pupils in England. London: HMSO. Do¨rnyei, Z. 2001. Motivational strategies in the language classroom. Cambridge, UK: Cambridge University Press. Dreyer, C., and R.L. Oxford. 1996. Learning strategies and other predictors of ESL proficiency among Africaans speakers in South Africa. In Language learning strategies across the world: Cross-cultural perspectives, ed. R.L. Oxford, . Manoa, HI: University of Hawai’i Press. Gardner, R.C., and W.E. Lambert. 1972. Attitudes and motivation in second language learning. Rowley, MA: Newbury House. Gilborn, D., and H.S. Mirza. 2000. Educational inequality: Mapping race, class and gender. Ofsted. www.ofsted.gov.uk
28
1190 AQ15
1195 AQ16
1200
1205
1210 AQ17 AQ18
1215
1220
1225 AQ19
AQ20
1230
M. Grenfell and V. Harris
Gipps, C. 1999. Socio-cultural aspects of assessment. Review of Research in Education 24: 355–92. Goh, C. 1998. How ESL learners with different listening abilities use comprehension strategies and tactics. Language Teaching Research 2, no. 2: 124–47. Grenfell, M., and V. Harris. 1999. Modern languages and learning strategies: In theory and practice. London: Routledge. Grenfell, M., and V. Harris. 2007. The strategy use of learners of a third language. Paper presented at the Fifth International Conference on Third Language Acquisition and Multilingualism, September , in Stirling, UK. Griffiths, C. 2003. Patterns of language learning strategy use. System 31, no. 3: 367–83. Harris, V. 2006. Language learning strategies across the curriculum. Report by the Society for Educational Studies. Harris, V. 2007. Exploring progression: Reading and listening strategy instruction with near-beginner learners of French. Language Learning Journal 35, no. 2: 189–204. Harris, V. 2008. A cross-curricular approach to ‘learning to learn’ languages: Government policy and school practice. Curriculum Journal 19, no. 4: 255–68. Harris, V., and J. Prescott. 2005. Learning strategy instruction: In theory and in practice. Scottish Languages Review 11. http://www.strath.ac.uk/media/faculties/hass/scilt/slr/ issues/11/1_Harris_Prescott_images.pdf Harris, V., H. Ingvadottir, B. Jones, R. Neuburg, I. Palos, and I. Schindler. 2001. Helping learners learn: Exploring strategy instruction in languages classrooms across Europe. Graz, Austria: European Council for Modern Languages. Hassan X., E. Macaro, D. Mason, G. Nye, P. Smith, and R. Vanderplank. 2005. Strategy training in language learning: A systematic review of available research. EPPI-Centre. http://eppi.ioe.ac.uk/cms/Default.aspx?tabid¼296 Hatch, E., and A. Lazarton. 1991. The research manual: Design and statistics for applied linguistics. Boston, MA: Heinle and Heinle. Holliday, A. 2003. Social autonomy: Addressing the dangers of culturalism in TESOL. In Learner autonomy across cultures, ed. D. Palfreyman and R. Smith, . London: Palgrave Macmillan. Ikeda, M. 2006. EFL Reading strategies: Empirical studies and an instructional model. : Shohakusha. James, M., R. McCormick, P. Black, P. Carmichael, M.J. Drummond, A. Fox, J. MacBeath, et al. 2007. Improving learning how to learn: Classrooms, schools and networks. London: RoutledgeFalmer. Jessner, U. 2008. Teaching third languages: Findings, trends and challenges. Language Teaching 41, no. 1: 15–56. Jones, J., and G. Jones. 2001. Boys’ performance in modern foreign languages: Listening to learners. London: CILT. Larsen-Freeman, D. 2007. Reflecting on the cognitive-social debate in second language acquisition. Modern Language Journal 91: 773–87. Kissau, S. 2006. Gender differences in motivation to learn French. Canadian Modern Languages Review 62, no. 3: 401–22. Kusiak, M. 2001. The effect of metacognitive strategy training on reading comprehension and metacognitive knowledge. EUROSLA yearbook 1: 255–74. Macaro, E. 2001. Learning strategies in foreign and second language classrooms. London: Continuum. Macaro, E., and L. Erler. 2005. Raising achievement of young beginner readers of French through strategy instruction. Manuscript. Macaro, E., S. Graham, and R. Vanderplank. 2007. A review of listening strategies: Focus on sources of knowledge and success. In Language learner strategies: 30 years of research and practice, ed. A.D. Cohen and E. Macaro, Oxford, UK: Oxford University Press. Macaro, E., and T. Mutton. 2009. Developing reading achievement in primary learners of French: Inferencing strategies versus exposure to ‘graded readers’. Language Learning Journal 37, no. 2: 165–82.
The Curriculum Journal
AQ21 AQ22 AQ23
AQ24
AQ25 AQ26
AQ27 AQ28
AQ29
29
Manchon, R. 2008. Taking strategies to the foreign language classroom: Where are we now in theory and research? International Review of Applied Linguistics in Language Teaching (IRAL) 46, no. 3: 221–43. Naiman, N., M. Fro¨hlich, H.H. Stern, and A. Todesco. 1978. The good language learner. Clevedon, UK: Multilingual Matters. National Literacy Trust. 2006. What languages are spoken in the UK? National Literacy Trust. www.literacytrust.org.uk/Research/lostops3.htm NferNelson. 2003. Understanding potential: Cognitive abilities test. London: NferNelson Nisbet, D.L., E.R. Tindall, and A.A. Arroyo. 2005. Language learning strategies and English proficiency of Chinese university students. Foreign Language Annals 38, no. 1: 100–7. Norton, B. 2000. Identity and language learning: Gender, ethnicity and educational change. London: Longman/Pearson Education. Norton, B., and K. Toohey. 2001. Changing perspectives on good language learners. TESOL Quarterly 35, no. 2: 307–22. Nuffield. 2000. The Nuffield languages inquiry. Languages: The next generation. London: The Nuffield Foundation. O’Malley, J.M., and A.U. Chamot. 1990. Learning strategies in second language acquisition. Cambridge, UK: Cambridge University Press. Ofsted. 2004. Modern foreign languages at a glance 2002/03. London: HMSO. Oxford, R. 1999. Relationships between second language learning strategies and language proficiency in the context of learner autonomy and self-regulation. Revista Canaria de Estudios Ingleses 38: 108–26. Oxford, R., Y. Cho, S. Leung, and H. Kim. 2004. Effect of the presence and difficulty of task on strategy use: An exploratory study. International Review of Applied Linguistics in Language Teaching (IRAL) 42, no. 1: 1–47. Oxford, R., and K. Schramm. 2007. Bridging the gap between psychological and sociocultural perspectives on L2 learner strategies. In Language learner strategies: 30 years of research and practice, ed. A.D. Cohen and E. Macaro, . Oxford, UK: Oxford University Press. Reay, D. 2006. The zombie stalking English schools: Social class and educational inequality. British Journal of Educational Studies 54, no. 3: 288–307. Rodeiro, C.L.V. 2009. Some issues on the uptake of modern foreign languages at GCSE. Statistics report series 10. Cambridge, UK: Cambridge Assessment. Rubin, J. 1975. What the ‘good language learner’ can teach us. TESOL Quarterly 9, no. 1: 41–51. Rubin, J. 1990. How learner strategies can inform language teaching. In Language use, language teaching and the curriculum, ed. V. Bickley, . Hong Kong: Institute of Language in Education. Rubin, R., A.U. Chamot, V. Harris, and N.J. Anderson. 2007. Intervening in the use of strategies. In Language learner strategies: 30 years of research and practice, ed. A.D. Cohen and E. Macaro, . Oxford, UK: Oxford University Press. Sedghi, A., and L. Evans. 2011. GCSE results 2011: Exam breakdown by subject, school and gender. The Guardian, August 25. http://www.guardian.co.uk/news/datablog/ 2011/aug/25/gcse-results-2011-exam-breakdown Shapiro, J. 2007. English and media, In Genderwatch: Still watching, ed. K. Myers and H. Taylor, . Stoke on Trent, UK: Trentham Books. Stern, H.H. 1975. What can we learn from the Good Language Learner? Canadian Modern Language Review 31, no. 4: 304–18. Tabachnick, B.G., and L.S. Fidell. 2007. Using multivariate statistics. 5th ed. Boston, MA: Pearson. Takeuchi, O., C. Griffiths, and D. Coyle. 2007. Applying strategies to contexts: The role of individual, situational and group differences. In Language learner strategies: 30 years of research and practice, ed. A.D. Cohen and E. Macaro, . Oxford, UK: Oxford University Press Van Der Grift, L. 1997. The comprehension strategies of second language (French) listeners: A descriptive study. Foreign Language Annals 30, no. 3: 387–409.
1235
1240
1245
1250
1255
1260
1265
1270
1275
30
M. Grenfell and V. Harris
Van Der Grift, L. 2003. Orchestrating strategy use: Toward a model of the skilled second language listener. Language Learning 53, no. 3: 463–95. Van Der Grift, L. 2007. Recent developments in second and foreign language listening comprehension research. Language Teaching 40, no. 3: 191–210. Watkins, C. 2007. Learning and teaching. In Genderwatch: Still watching, ed. K. Myers 1280 AQ30 and H. Taylor, Stoke on Trent, UK: Trentham Books. Wharton, G. 2000. Language learning strategy use of bilingual foreign language learners in Singapore. Language Learning 50, no. 2: 203–43. White, C., K. Schramm, and A.U. Chamot. 2007. Research methods in strategy research: Re-examining the toolbox. In Language learner strategies: 30 years of research and practice, ed. A.D. Cohen and E. Macaro, Oxford, UK: Oxford University Press. AQ31 AQ32 1285 Whitty, G. 2012. A life with the sociology of education. British Journal of Educational Studies 60, no. 1: 65–75. Wong Fillmore, L. 1979. Individual differences in second language acquisition. In Individual differences in language ability and behaviour, ed. C.J. Fillmore, W.S.Y. AQ33 Wang and D. Kempler, . New York: Academic Press.
1290
1295
1300
1305
1310
1315
1320
The Curriculum Journal
31
Appendix 1. Reading strategy checklist Before reading 1. I work out from the layout what I am reading; e.g. if it’s instructions, or a short paragraph or a letter, or even an advert or a brochure 2. I try to get clues from any pictures and the title to help me guess what it will be about 3. I try to predict all the words and information that I might find in the text While reading 1. I don’t panic and switch off but I just tell myself it’s OK and keep reading even if it is hard 2. I just try to get the main ideas first and then read it again for the details 3. I skip over words that I do not understand 4. I try to spot familiar words that I do understand from when we learned them in class 5. I look out for cognates 6. I look out for the names of people or places and for punctuation clues 7. I think about all the possible things it could mean 8. If I don’t understand, I use my common sense to guess the meaning from the rest of the words in the sentence and what I have worked out so far 9. I say the difficult bits out loud or in my own head 10. I say in English what I have worked out so far in the sentence and substitute ‘something’ for the words I don’t know 11. I break the word or sentence up into bits that I may recognise 12. If I don’t understand one bit, I go back to it and read it over several times slowly 13. I try to use grammar clues to spot what kind of a word it is – a noun, a verb etc. After reading 1. I try to remember everything that I have read and then fit it altogether so that it makes sense 2. I check back to see if my first guesses still make sense
1325
1330
1335
1340
1345
1350
1355
1360
32
M. Grenfell and V. Harris
Appendix 2. Listening strategy checklist When I am listening to the CD (or to the teacher talking in French):
1370-
1375
1380
1385
1390
1395
1400
1405
Before listening 1. I work out from the layout (pictures and instructions) what I will hear; a conversation, directions round town, a railway announcement 2. I try to predict all the words and information that I might hear 3. I decide on the key words to listen out for While listening 1. I don’t panic and switch off but I just tell myself it’s OK and keep listening even if they do talk fast 2. I just try to get the main ideas, when the tape is first played, and then I listen again for details 3. I try to get clues from the tone of voice (questions? feelings?) and from gestures or background sounds 4. I skip over words that I do not understand so that I don’t miss what is said next 5. I don’t try to write and listen at the same time 6. I make pictures in my head of what is said 7. I try to spot familiar words that I do understand from when the teacher said them in class 8. I listen out for possible cognates and think about how they may sound different in French 9. I listen out for the names of people or places and think about how they may sound different in French 10. I break the sounds down into possible words and try writing them down to see if I can recognise them 11. I think about all the possible words it could be and things it could mean 12. I double-check words because a word may sound like English but not mean the same thing at all 13. If I don’t understand, I use my common sense to guess the meaning from the rest of the words in the sentence and what I have worked out so far 14. I say in English what I have worked out so far and substitute ‘something’ for the words I don’t know 15. If I don’t understand one bit, I listen out for it when the tape is played again 16. I try to use grammatical clues to spot what kind of a word it is – a noun, verb etc. After listening 1. I try to remember everything I have heard and then fit it all together so that it makes sense 2. I check back to see if my first guesses still make sense
