ArticlePDF Available

Re-thinking grammar: The impact of embedded grammar teaching on students' writing and students' metalinguistic understanding


Abstract and Figures

This paper reports on a national study, involving a mixed-method research design comprising a randomised controlled trial (RCT), text analysis, student and teacher interviews and lesson observations. It set out to investigate whether contextualised teaching of grammar, linked to the teaching of writing, would improve student outcomes in writing and in metalinguistic understanding. The RCT involved 744 students in 31 schools in the south-west and the Midlands of England, and was a blind randomisation study. Classes were randomly allocated to either a comparison or intervention group, after the sample had been matched for teacher linguistic subject knowledge (LSK). The statistical data were complemented by three interviews per teacher and three interviews with a focus student in each class, plus three lesson observations in each class, giving a data-set of 93 teacher interviews, 93 student interviews and 93 lesson observations. In addition, the final pieces of writing produced for each scheme of work were collected. The statistical results indicate a significant positive effect for the intervention, but they also indicate that this benefit was experienced more strongly by the more able writers in the sample. The regression modelling also indicates that teacher LSK was a significant mediating factor in the success of the intervention. The qualitative data provide further evidence of the impact of teacher knowledge on how the intervention was implemented and on students’ metalinguistic learning. It also reveals that teachers found the explicitness, the use of discussion and the emphasis on playful experimentation to be the most salient features of the intervention. The study is significant in providing robust evidence for the first time of a positive benefit derived from the teaching of grammar, and signals the potential of a pedagogy for a writing which includes a theorised role for grammar.
Content may be subject to copyright.
This article was downloaded by: [Mrs Helen Lines]
On: 27 November 2011, At: 08:00
Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered
office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
Research Papers in Education
Publication details, including instructions for authors and
subscription information:
Re-thinking grammar: the impact
of embedded grammar teaching
on students’ writing and students’
metalinguistic understanding
Debra A. Myhill
, Susan M. Jones
, Helen Lines
& Annabel
School of Education, University of Exeter, Exeter, UK
Available online: 16 Nov 2011
To cite this article: Debra A. Myhill, Susan M. Jones, Helen Lines & Annabel Watson (2011): Re-
thinking grammar: the impact of embedded grammar teaching on students’ writing and students’
metalinguistic understanding, Research Papers in Education, DOI:10.1080/02671522.2011.637640
To link to this article:
Full terms and conditions of use:
This article may be used for research, teaching, and private study purposes. Any
substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,
systematic supply, or distribution in any form to anyone is expressly forbidden.
The publisher does not give any warranty express or implied or make any representation
that the contents will be complete or accurate or up to date. The accuracy of any
instructions, formulae, and drug doses should be independently verified with primary
sources. The publisher shall not be liable for any loss, actions, claims, proceedings,
demand, or costs or damages whatsoever or howsoever caused arising directly or
indirectly in connection with or arising out of the use of this material.
Re-thinking grammar: the impact of embedded grammar teaching
on students writing and students metalinguistic understanding
Debra A. Myhill*, Susan M. Jones, Helen Lines and Annabel Watson
School of Education, University of Exeter, Exeter, UK
(Received 11 October 2011; nal version received 28 October 2011)
This paper reports on a national study, involving a mixed-method research design
comprising a randomised controlled trial (RCT), text analysis, student and tea-
cher interviews and lesson observations. It set out to investigate whether contex-
tualised teaching of grammar, linked to the teaching of writing, would improve
student outcomes in writing and in metalinguistic understanding. The RCT
involved 744 students in 31 schools in the south-west and the Midlands of Eng-
land, and was a blind randomisation study. Classes were randomly allocated to
either a comparison or intervention group, after the sample had been matched for
teacher linguistic subject knowledge (LSK). The statistical data were comple-
mented by three interviews per teacher and three interviews with a focus student
in each class, plus three lesson observations in each class, giving a data-set of 93
teacher interviews, 93 student interviews and 93 lesson observations. In addition,
the nal pieces of writing produced for each scheme of work were collected. The
statistical results indicate a signicant positive effect for the intervention, but
they also indicate that this benet was experienced more strongly by the more
able writers in the sample. The regression modelling also indicates that teacher
LSK was a signicant mediating factor in the success of the intervention. The
qualitative data provide further evidence of the impact of teacher knowledge on
how the intervention was implemented and on students metalinguistic learning.
It also reveals that teachers found the explicitness, the use of discussion and the
emphasis on playful experimentation to be the most salient features of the inter-
vention. The study is signi cant in providing robust evidence for the rst time of
a positive benet derived from the teaching of grammar, and signals the potential
of a pedagogy for a writing which includes a theorised role for grammar.
Keywords: grammar; writing; metalinguistic knowledge
1. Introduction
The debate about the place of grammar in the English curriculum has a long
history, with research reports and professional arguments on the topic spanning
over 50 years. Moreover, it is a debate which crosses national boundaries and is
common to most Anglophone countries. Detailed reviews of international evidence
for and against the benets of teaching grammar, stimulated by a renewed emphasis
on grammar in the National Curriculum for English and the National Literacy
Strategy, have been conducted by Hudson (2004), Wyse (2004) and most recently,
by the Evidence for Policy and Practice Information and Co-ordinating Centre
*Corresponding author. Email:
Research Papers in Education
2011, 128, iFirst Article
ISSN 0267-1522 print/ISSN 1470-1146 online
Ó 2011 Taylor & Francis
Downloaded by [Mrs Helen Lines] at 08:00 27 November 2011
(EPPI-Centre) Review Group for English (2004). It is a debate which has not only
been theoretical and pedagogical, but one in which the public have regularly and
enthusiastically participated (Gordon 2005). At the same time, there are interna-
tional concerns about childrens standards of writing. In Australia, following the
1996 National School English Literacy Survey, the Minister for Schools acknowl-
edged that too many children did not achieve a minimum acceptable standard in
literacy (Masters and Forster 1997). In the USA, the call for a writing revolution
(National Commission on Writing in Americas Schools and Colleges [NCW] 2003)
to address the problem of children who cannot write with the skill expected of
them today (NCW 2003) has been followed by major policy change in No Child
Left Behind (United States Department of Education 2002). This has reintroduced
grammar as part of the raising standards agenda.
2. Theoretical framework
In tandem with political and ideological constructions of the grammar debate is an
academic discussion between linguists and educationalists on the value of explicit
and systematic teaching of grammar. The discipline of linguistics is, of course, much
broader than the topic of grammar, including amongst other things, phonetics, prag-
matics, evolutionary linguistics, language acquisition and neurolinguistics. In gen-
eral, educationalists focus only on grammar, the structure of the language, and in
some cases, on language acquisition. Increasingly, when discussing grammar, lin-
guists draw on the principles of contemporary linguistic theories, which are descrip-
tive and sociocultural in emphasis, or as Carter (1990, 104) describes them,
functionally oriented, related to the study of texts and responsive to social pur-
poses. They contend that a better understanding of how language works in a variety
of contexts supports learning in literacy. Denham and Lobeck (2005) draw on empir-
ical studies to claim that, in the multicultural, linguistically diverse classrooms of the
USA, linguistic knowledge is a tool which can inform teachers approaches to lan-
guage study in the classroom. Hudson (2004) offers a theoretical argument that edu-
cation needs linguistics, noting the distinction between traditional prescriptive
grammar and the very different approaches of modern linguistics. But teachers and
educationalists remain sceptical. The minutes of a meeting of the Linguistic Society
of America with the National Council for the Teaching of English (NCTE), a power-
ful body representing English teachers, to discuss how to better integrate linguistics
into the English/Language Arts curriculum note that NCTE was not eager to step
in as partners in such a project (initiated by linguists) (LSA 2006, 1). Indeed, many
educationalists not only lack enthusiasm for grammar teaching but see it as poten-
tially detrimental to childrens language development. Braddock, Lloyd-Jones, and
Schoers review (1963) of research into composition concluded that teaching gram-
mar was harmful, and Elbow (1981, 169) argued that nothing helps [their] writing
so much as learning to ignore grammar. As is evident from the publication dates
already cited, this has been a long-running debate which remains unresolved.
2.1. The nature of the evidence base
There are, however, many conceptual and methodological aws in much of the
extant research base used to provide evidence for this debate. Research is repeatedly
used selectively to justify a pre-determined position or to support a particular
2 D.A. Myhill et al.
Downloaded by [Mrs Helen Lines] at 08:00 27 November 2011
stance. Both Hudson (2004) and Wyse (2004) use research evidence to support their
opposing standpoints, drawing both on differing evidence for their claims and offer-
ing different interpretations of the same study. Bateman and Zidonis (1966) note
that although most research in this area produces inconclusive results, these are then
almost always construed as negative results. Tomlinson (1994), critiquing the meth-
odological validity of the research of Robinson (1959) and Harris (1962), noted that
the conclusions from their studies were what many in the educational establishment
wanted to hear (Tomlinson 1994, 26). Indeed, Tomlinson claims that most research
into the effectiveness of grammar teaching does not stand up to critical examina-
tion and many articles are simply polemical (Tomlinson 1994, 20).
The EPPI systematic review of the impact of grammar teaching on writing
(EPPI 2004) highlights many of the aws in this eld of enquiry. The research
question which informs the study (What is the effect of grammar teaching on the
accuracy and quality of 516 year olds written composition?) takes an over-simpli-
ed view of causal relationships between grammar teaching and written composi-
tion. It ignores the multifaceted nature of learning and the complex social, linguistic
and cognitive relationships that shape learning about writing. In particular, it does
not engage with some of the key factors related to teaching and learning which
might have an effect on the results. Teacher beliefs about the value of grammar,
their level of linguistic and pedagogic subject knowledge and teacher effectiveness
in the classroom are important variables which are not considered. The background
to the study does note that research might be needed to consider the effect/impact
on students writing skills of teachers grammatical knowledge (EPPI 2004, 12)
but no subsequent account is taken of this. Likewise, there is no adequate conceptu-
alization of grammar teaching: although the background to the study begins to
explore the changes in linguistic theories and the international political and peda-
gogical trends in grammar teaching, these are not used to reformulate the research
question. Thus the review considers research on grammar teaching but ignores the
considerable and signicant differences between the teaching of grammar in the UK
in the early twenty-rst century, and the teaching of grammar in different countries,
in different decades, and in different contexts. The clear conclusion that there is
no high quality evidence to counter the prevailing belief that the teaching of the
principles underlying and informing word order and syntax has virtually no inu-
ence on the writing quality of 516 year olds (EPPI 2004, 4) is actually predicated
on just three studies rated of medium or high signicance, two of which are at least
25 years old (Bateman and Zidonis 1966; Elley, Barham, and Wylie 1975) and none
of which were conducted in the UK. One of the review team subsequently reected
that our published reviews begged a lot of questions (Locke 2005, 3).
2.2. Grammar and writing
Although there is a considerable number of international studies purportedly investi-
gating the impact of grammar teaching on writing, there is almost none in which
the grammar is taught in the context of writing lessons with a view to developing
childrens writing. In many of the studies (e.g. Bateman and Zidonis 1966; Elley
et al. 1975, 1979; Robinson 1959) isolated grammar lessons are taught and the writ-
ing used to determine impact is produced in a different context. Fogel and Ehris
(2000) study is perhaps unique in taking as its starting point an identi
ed writing
problem, the tendency of some ethnic minority children to use non-standard Black
Research Papers in Education 3
Downloaded by [Mrs Helen Lines] at 08:00 27 November 2011
English Vernacular (BEV) in their writing. The study sets out to examine how to
structure dialect instruction so that it is effective in teaching Standard English (SE)
forms to students who use BEV in their writing (Fogel and Ehri 2000, 215) and
found a signicant improvement in avoidance of BEV in the group who were given
both strategies and guided support. They argue that their results demonstrated that
the approach used had claried for students the link between features in their own
nonstandard writing and features in SE (2000, 231). The Fogel and Ehri study
moves the eld forward by beginning to look at the pedagogical conditions which
support or hinder the transfer of grammatical knowledge into written outputs. It sig-
nals the importance of taking greater account of the subject knowledge of the tea-
cher and undertaking a more ne-grained analysis of pupils linguistic learning.
2.3. Teachers linguistic subject knowledge
Concerns that teachers linguistic knowledge is insufcient are neither new nor
restricted to the UK. Gurrey (1962, 14) observed that teachers lacked a thorough
training in pure grammar and more recent concerns about the level of linguistic
knowledge of English teachers have been expressed by Hudson (2004) in the UK,
and by Koln and Hancock (2005) in the USA and Gordon (2005) notes teachers in
New Zealand recognised their own, inadequate linguistic knowledge (Gordon
2005, 50). These latter observations are, however, made by linguists, not teachers.
A Qualications and Curriculum Authority (QCA) (1998) survey of teachers in the
period immediately following the introduction of the National Literacy Strategy
indicated considerable lack of condence in linguistic knowledge, particularly with
sentence grammar, and uncertainty about implicit and explicit knowledge. The
report concluded that there was a signicant gap ... in teachers knowledge and
condence in sentence grammar and this has implications for ... the teaching of
language and style in texts and pupils own writing (QCA 1998, 35). From a peda-
gogical perspective, linguistic subject knowledge (LSK) is more than the ability to
use appropriate terminology, as it also involves the ability to explain grammatical
concepts clearly and know when to draw attention to them. Andrews suggests that
it is likely to be the case that a teacher with a rich knowledge of grammatical con-
structions and a more general awareness of the forms and varieties of the language
will be in a better position to help young writers (Andrews 2005, 75), and Gordon
(2005) found that teachers who developed more secure linguistic knowledge were
able to see beyond supercial error in childrens writing as evidence of growing
syntactic maturity. Previously, for these teachers the writing virtues of their
pupils often went unseen and unacknowledged because of their own lack of knowl-
edge about language (Gordon 2005, 63). In contrast, weak linguistic knowledge
can lead to an over-emphasis upon identication of grammar structures without
fully acknowledging the conceptual or cognitive implications (Myhill 2003) of that
teaching. Equally, it can lead to sterile teaching, divorced from the realities of lan-
guage in use: Applebee (2000), for example, notes two studies in the USA which
showed that topic sentences and paragraph patterns taught in school bear little
resemblance to those found in real
2.4. The role of metalanguage
Central to the issue of LSK and the debate about the role of grammar in developing
writing is the question of the value of grammatical terminology and access to this
4 D.A. Myhill et al.
Downloaded by [Mrs Helen Lines] at 08:00 27 November 2011
metalanguage. Cognitive psychologists have repeatedly signalled the importance of
metacognition (Buttereld, Hacker, and Albertson 1996; Hayes and Flower 1980;
Kellogg 1994; Martlew 1983; Wallace and Hayes 1992) in the writing process,
because writing is a process which demands self-monitoring (Kellogg 1994, 17).
Metacognition, put simply, is thinking about ones thinking: it is a conscious pro-
cess, described by Flavell (1976, 232) as
ones knowledge concerning ones own cognitive processes or anything related to
them, e.g., the learning-relevant properties of information or data. For example, I am
engaging in metacognition if I notice that I am having more trouble learning A than
B; if it strikes me that I should double check C before accepting it as fact.
Metacognitive knowledge plays a role in every stage of the writing process: in
moving planning from an over-emphasis on content to greater consideration of the
strategic goals of the task (Hayes and Flower 1980); in supporting the development
of a model of their audience, for reecting on rhetorical and content probabilities
(Kellogg 1994, 213); in the process of revision (Alamargot and Chanquoy 2001)
and in developing self-regulation (Englert, Raphael, and Anderson 1992). Bereiter
and Scardamalia (1982) argue that the benet of metacognition is that it renders
normally covert processes overt and provides labels to make tacit knowledge
more accessible (1982, 57) and in summarising the ndings of their intervention
study, Englert, Raphael, and Anderson (1992, 441) are insistent that the importance
of the students increased mastery over the language of the writing process cannot
be over-emphasised: both are thus signalling the importance of a metalanguage,
though not necessarily grammatical language. Metalinguistic knowledge is a subset
of metacognitive knowledge, though there is surprisingly little empirical research in
this aspect of writing. The most comprehensive consideration of metalinguistic
knowledge is Gomberts (1992) model of metalinguistic awareness, designed to
inform an understanding of oral development and how children learn to read. He
proposes two levels of cognitive control of linguistic knowledge: epilinguistic,
where linguistic processing is controlled automatically by linguistic organisations in
the memory, and metalinguistic, when, the individual is in conscious control of lin-
guistic decision-making. Gombert argues that there is a developmental hierarchy
between epilinguistic control and metalinguistic awareness (2003). Van Lier
(1998), however, contests the hierarchical assumptions of Gomberts proposition,
questioning whether epilinguistic awareness is necessarily a precursor of metalin-
guistic awareness. Yet this principle of a cognitive shift from implicit to explicit
knowledge is a prevalent one, including at policy level. QCA (1998) describe the
learning trajectory as moving from implicit knowledge, derived from experience, to
analysis, based on grammatical terminology, developing into understanding of func-
tion and effect, leading nally to explicit knowledge. Explicit knowledge is dened
as knowledge that can identify and account for connections and distinctions
between different examples of usage, enhance reading and improve writing (QCA
1998, 20). Van Lier questions the value of metalinguistic knowledge measured in
solitary demonstrations of knowledge and argues that being able to articulate meta-
linguistic knowledge is less important than being able to demonstrate it: the ability
to control and manipulate the material at hand is more signicant than the ability
to describe a linguistic feature using grammatical terminology (1998, 136). Van
Liers concern that metalinguistic knowledge is not transferred into linguistic
Research Papers in Education 5
Downloaded by [Mrs Helen Lines] at 08:00 27 November 2011
performance is central to the issue of whether grammar supports writing develop-
ment. Myhill and Jones (2007) found that secondary-age writers were often able to
articulate explicit choices made during text production, but were not always able to
describe these in metalinguistic terms; equally, they found that some procient writ-
ers had automated linguistic decision-making and no longer thought explicitly about
metalinguistic choices. On the other hand, Carter (1990) maintains that the demise
of formal grammar teaching and with it the absence of a metalanguage in the class-
room has been disempowering, preventing learners from exercising the kind of
conscious control and conscious choice over language which enables both to see
through language in a systematic way and to use language more discriminatingly
(Carter 1990, 119).
This review of research into the relationship between grammar and writing
indicates that as Andrews (2005, 74) observed, there is still a dearth of evidence
for the effective use of grammar teaching of any kind in the development of writ-
ing. It is evident that there remains a pressing need for robust large-scale research
which seeks to establish valid causal relationships, but which also seeks to go
beyond simple causeeffect paradigms to understand the complexity of the issue. In
particular, writing research needs to adopt an inter-disciplinary framework, which is
cognisant of linguistic, cognitive and sociocultural perspectives, in order to reect
with validity the complexity of classrooms as teaching and learning contexts. It is
important to be mindful of the cognitive demands of writing production, and the
challenge all writers face of keeping in mind the conceptual message together with
their rhetorical objectives and at the same time appeal to linguistic knowledge to
express their ideas correctly and appropriately (Van Gelderen and Oostdam 2005,
215). However, it is equally important to foreground the linguists perspective that
terminology and rules are pointless if your mind hasnt grasped the concepts behind
the terminology (Keith 1997, 12). In addition, the cognitive and linguistic chal-
lenges of writing need to be bounded by an acknowledgement that writing is mate-
rial social practice in which meaning is actively made, rather than passively relayed
or effortlessly produced (Micciche 2004, 719). The study reported here sought to
operate within such an inter-disciplinary framework and to answer the research
question: What impact does contextualised grammar teaching have upon pupils
writing and pupils metalinguistic understanding?
3. Methodology
Teaching is a complex, multifaceted and situated endeavour which resists simplistic
causal explanations between pedagogical activity and learning outcome; equally,
writing is perhaps the most complex activity learners undertake, drawing on
cognitive, social and linguistic resources. Accordingly, this study adopted a mixed-
method approach located within an inter-disciplinary conceptual framework, com-
bining a cluster randomised controlled trial (RCT) with multiple regression analysis
and a complementary qualitative study.
The EPPI review of the effect of grammar teaching (EPPI 2004) concluded by
calling for a conclusive, large scale and well-designed randomised controlled trial
(EPPI 2004, 49) into the impact of grammar teaching on writing. We would argue
that no such RCT could be conclusive because of the complexity of both the
empirical question and the educational context. Instead, this study adopted a mixed-
method approach to investigate specically the complex causal relationships
6 D.A. Myhill et al.
Downloaded by [Mrs Helen Lines] at 08:00 27 November 2011
between pedagogical support for teaching grammar, teacher subject knowledge and
improvement in writing. One intervention can be realised in multiple outcomes for
learners and one intervention strategy can be multiply interpreted and mediated in
the classroom by different teachers. Indeed, as Shadish, Cook, and Campbell (2002,
5) argue, causal relationships are rarely deterministic: to different degrees, all cau-
sal relationships are context dependent. Therefore, to complement the statistical
data derived from the experimental study, and to provide in-depth understanding of
the theoretical, pedagogical and contextual implications of the statistical data, the
experimental component was embedded within a qualitative design. This mixed-
method approach is important for RCTs conducted in educational contexts: indeed,
Moore, Graham, and Diamond (2003) argue that
to undertake a trial of an educational or social intervention without an embedded qual-
itative process evaluation would be to treat the intervention as a black box, with no
information on how it worked, how it could be improved, or what the crucial compo-
nents of the intervention were.
Likewise, Shadish, Cook, and Campbell (2002, 71) recommend the addition of
qualitative methodologies to experiments to provide better interpretation and avoid
errors in applying research outcomes to practice.
The collection of data fell broadly into three categories: the baseline data
required before the intervention began; the impact data which measured the effect
of the intervention and the qualitative data which provided interpretative and con-
textual information. The impact data comprised the pre- and post-test writing scores,
including both the total scores, and the scores on the sub-components of the writing
test. The qualitative data comprised lesson observations of the sample classes, tea-
cher interviews, writing conversations with the sub-sample of pupils and the writing
produced in the teaching episodes.
3.1. The RCT
In educational settings, conventional RCTs following the medical model are rarely
appropriate as frequently interventions are at the level of the school or the class,
rather than that of the individual. Instead, the cluster trial where randomisation is
conducted at the level of the group or cluster (Moore, Graham, and Diamond
2003, 680) is increasingly viewed as a more valid and robust design. This study uti-
lised a cluster RCT where the independent variable was the pedagogical support
materials and the dependent variable was the quality of writing. The intervention
group received detailed pedagogical support materials and were trained in their use,
whilst the comparison group received only an outline scheme of work with no ped-
agogical support.
3.1.1. The sample
In this study, as a cluster trial, randomisation occurred at the level of the group
rather than that of the individual (Donner and Klar 2000; Murray 1998). Only one
class was used from each school to avoid any cross-over effects from one group to
another. Neither the teacher nor the class knew there was a distinction being made
between an intervention and a comparison group. The teachers and students knew
Research Papers in Education 7
Downloaded by [Mrs Helen Lines] at 08:00 27 November 2011
that the focus of the research was writing, but they were unaware of the specic
grammar focus. The student sample was 32 Year eight mixed ability classes in com-
prehensive schools with between 24 and 30 students in each class.
The process of recruiting the schools made use of Local Authority school directo-
ries in the south-west and the west Midlands. In order to create as representative a
sample as possible, an initial process de-selected the small number of schools which
were single-sex, selective, or atypical in age range (e.g. 1418 age range). This was
because these schools could distort the sample in terms of gender, attainment or
socio-economic status. This left a sample which comprised mixed, comprehensive
schools with an age range of 1118. Using a random number generator, each school
was given a number, creating a numbered list of the schools. Schools were then con-
tacted in numerical order to invite them to participate in the project, and if there was
a non-response or if they declined to participate then the next school on the list was
approached. Random sampling continued until at least 16 schools from each area
had been recruited. Nineteen schools in the south-west and 28 schools in the west
Midlands were approached to secure the sample of 16 in each area.
Because the study set out to investigate the impact of teachers LSK upon the
intervention, the sample was stratied rst by teacher LSK in order to avoid acci-
dental bias in either the intervention or comparison group. The LSK of the teachers
was established through a specically designed test of grammar knowledge in a
questionnaire establishing baseline information about the teachers professional
background and subject knowledge. It included questions regarding professional
and academic background; questions focusing on teachers perceptions of their own
knowledge of English literature and their beliefs about the use of literature to sup-
port the teaching of writing; questions focusing on teachers grammar knowledge
and their beliefs about the use of grammar to support the teaching of writing. The
score for each teacher on the LSK test was added to the database. The teacher
scores were ranked and then in turn allocated to either group A or group B, thus
ensuring that each group was broadly matched for LSK. Finally, the two groups
were randomly allocated to the comparison or intervention group.
Having established the participating schools, teachers and classes, a comprehen-
sive set of baseline data was collated (see Table 1). This drew heavily on national
test data, collected in all schools, and comparable across the sample. At student
level, data on attainment in English were gathered. Students in England sit for
national externally marked tests at age 11 (Key Stage 2 tests) and are required to
report a National Curriculum level for English at age 14 (Key Stage 3). At Key
Stage 2, the test score is available as an overall National Curriculum level for Eng-
lish, plus a separate level and raw score for writing. The student data also recorded
whether a child was in receipt of Free School Meals or whether English was an
additional language for them. At teacher level, the baseline data was collected
through the questionnaire described above. At school level, the data were drawn
from the most recent ofce for standards in education (OFSTED) inspection report,
which provides comparable contextual information about each school inspected.
3.1.2. The intervention
The intervention comprised detailed teaching schemes of work in which grammar
was embedded where a meaningful connection could be made between the grammar
point and writing. Both the comparison and intervention groups taught the same
8 D.A. Myhill et al.
Downloaded by [Mrs Helen Lines] at 08:00 27 November 2011
writing genre over a three-week period once a term, and addressed the same writing
learning objectives from the Framework for English, part of the English govern-
ments National Strategies for raising educational attainment (see Table 2). Both
groups were given the same written outcomes for each genre studied: the opening
of a story; a written speech and a portfolio of three specied types of poem. A
medium term plan was provided for each group, which outlined the time frame,
learning objectives, assessed outcomes, accompanied by a range of relevant stimu-
lus resources. Thus, for both the intervention and comparison groups, the learning
focus, the period of study, the learning objectives and the assessed written outcomes
were the same.
Table 1. A summary of the baseline data.
Data collected
Student Key Stage 2 English level
Key Stage 2 Writing level
Key stage 2 Writing raw score
Predicted Key Stage 3 English level
Free School Meals or not
English as an additional language or not
Teacher Years of service
Initial undergraduate degree subject
LSK score
School Free School Meals (average; above average; below average)
Ethnic diversity (average; above average; below average)
% of students achieving ve GCSEs or more, including Maths and English
% of students with special educational needs
Contextual value-added measure
Most recent OFSTED inspection grade
OFSTED Section 10 English result
Table 2. Learning objectives addressed in the schemes of work.
Learning objectives for writing from the framework for English
Year 8
Developing viewpoint, voice and ideas
Varying sentences and punctuation for clarity and effect
Improving vocabulary for precision and effect
Developing varied linguistic and literary techniques
Using grammar accurately and appropriately
Developing viewpoint, voice and ideas
Varying sentences and punctuation for clarity and effect
Improving vocabulary for precision and effect
Developing varied linguistic and literary devices
Structuring, organising and presenting texts in a variety of forms
on paper and on screen
Using grammar accurately and appropriately
Generating ideas, planning and drafting
Varying sentences and punctuation for clarity and effect
Improving vocabulary for precision and impact
Developing varied linguistics and literary techniques
Research Papers in Education 9
Downloaded by [Mrs Helen Lines] at 08:00 27 November 2011
The detailed teaching schemes for the intervention group were designed by the
project team, and explicitly sought to introduce grammatical constructions and termi-
nology at a point in the teaching sequence which was relevant to the genre being stud-
ied; for example, exploring how the use of rst or third person can position the
narrator differently or looking at how expanded noun phrases can build description in
poetry (further information about the teaching strategies can be found in Myhill,
Lines, and Watson 2011). The teaching focus was on effects and constructing mean-
ings, not on the grammatical terminology: building on the concept of writing as
design (Myhill 2010; Sharples 1999), the goal was to open up what we have called a
repertoire of possibilities, rather than to suggest correct or formulaic ways of writing.
A set of pedagogical principles informed the design of the teaching schemes:
The grammatical metalanguage is used but it is always explained through
examples and patterns.
Links are always made between the feature introduced and how it might
enhance the writing being tackled.
The use of imitation: offering model patterns for students to play with and
then use in their own writing.
The inclusion of activities which encourage talking about language and effects.
The use of authentic examples from authentic texts.
The use of activities which support students in making choices and being
designers of writing.
The encouragement of language play, experimentation and games.
3.1.3. Pre- and post-test writing tasks
The impact of the teaching on student writing was determined by a pre- and
post-test sample of writing. Both the pre- and post-test writing samples were a rst
person narrative, drawing on personal experience and written under controlled con-
ditions. The test design and marking were led by Cambridge Assessment, who were
responsible for setting and marking the national Writing Test at Key Stage 3 until
2006. In order to avoid any possible bias created by the precise choice of writing
task, the topic was selected to avoid known gender preferences in writing and to
avoid any need for having had a particular experience. To ensure that there was no
task bias, a cross-over design was adopted where half the sample completed task 1
as the pre-test and task 2 as the post-test, while the other half of the sample
reversed the order in which these tests were taken. Both sample sets were indepen-
dently marked by Cambridge Assessment. For each set of scripts, Cambridge
Assessment provided a rst markers set of marks, a second markers set and a res-
olution mark, adjudicated by a third senior marker if the rst two marks were very
different. They indicated which set of marks should be used for the purposes of the
study, but allowed us to see any variability in the marking behind those chosen g-
ures. The marking was based on the national Key Stage 3 mark scheme format, the
nal mark being made up of three components: sentence structure and punctuation;
text structure and organisation and composition and effect. Cambridge Assessment
devised the training materials for marking; undertook the administration to select
and train a marking team; delivered a training day for each marking round and
ensured the usual standardisation checks during the marking. The markers did not
know from which treatment group the writing had derived.
10 D.A. Myhill et al.
Downloaded by [Mrs Helen Lines] at 08:00 27 November 2011
3.1.4. Attrition and delity
Fidelity is a problematic concept in a naturalistic educational setting such as this, as
identical implementation of the intervention teaching materials is neither possible
nor desirable. Teachers were not asked to follow the lesson plans rigidly; they were
allowed to adapt materials to suit the needs of their students, but were also asked to
remain as close as possible to the materials. All 32 participating schools remained
in the project throughout the year-long period of the study, but it was decided to
exclude one schools data from the nal analysis because of low delity to the
study. For example, the teacher taught lessons which were not focused upon writ-
ing, and also regularly had other teachers teaching her lessons. Therefore, the nal
sample used for analysis comprised 31 teachers in 31 schools.
The original student sample was n = 900, but after removing one class on the
grounds of low delity, and removing those students who were not present for both
the pre- and post-tests, the nal sample was n = 744, representing an attrition rate of
3.2. The qualitative study
The qualitative data are comprised of four data-sets: classroom observations and fol-
low-up interviews once per teacher per scheme of work; writing conversations once
per scheme of work with a focus child from each class; and the written outcomes
from the teaching of each genre. The full qualitative data-set includes 93 lesson
observations; 93 teacher interviews; 93 writing conversations, plus the writing sam-
ples of narrative ction, argument and poetry from each class.
For the lesson observations, a schedule was designed to capture a record of how
the teachers taught the three writing genres and how students responded. The sche-
dule recorded the sequence of activities and the grammatical, literary and linguistic
terminology used by the teacher, and prompted for comments on teacher interaction,
student responses and observers reections on the use of contextualised grammar
The teacher interviews were conducted using a semi-structured interview sche-
dule which explored their pedagogical decision-making in the lesson observed, and
their reections on the lesson and students learning. For the intervention group only,
this included their pedagogic evaluations of the schemes of work. The specic ques-
tions about the lesson observed were not pre-planned but drew on the lesson obser-
vation, for example, probing why the teacher gave a particular example, or the
teachersreection on a particular students response. In addition, each interview
sought to generate teachers beliefs about writing, through prompts designed to stim-
ulate open-ended discussion. The nal teacher interview, following the poetry
scheme of work, was an extended interview which directly explored teachers beliefs
about the value of grammar teaching: this was not directly addressed in the earlier
interviews to avoid highlighting grammar as the focus of the study, although many
of the teachers did refer to grammar without prompting in the earlier interviews.
The writing conversations with a student were conducted following the lesson
observation. In each of the sample classes, one student was selected as a focus
student for these conversations; the overall sample of focus students was stratied
by gender in order to avoid any data distortions which might be inuenced by
gendered attitudes towards writing. The writing conversations were shaped by a
Research Papers in Education 11
Downloaded by [Mrs Helen Lines] at 08:00 27 November 2011
semi-structured interview schedule and by stimulus prompts. The rst section of the
interview was principally to facilitate engagement by asking broad questions about
writing and their own perceptions of learning about writing, but it did include a
question which probed their perspectives as learners on the lesson observed. As
with the teacher interviews, the precise questions asked at this point were framed
by what had been observed in the lesson. The second section of the interview
explored students metalinguistic understanding of their own and others writing.
This section of the interview used stimulus prompts to lead the discussion. For each
genre, a sample of writing in that genre written by students of the same age was
selected, not as a model of excellence but as a starting point for discussion. The
interview also used the focus students own writing produced during the teaching as
a stimulus for discussion.
The qualitative data enabled the study to be sensitive to the complexities of
classroom learning, and to provide a more nuanced picture of the way the interven-
tion was realised in practice. It will both inform future development of the inter-
vention, and also to contribute to theory and understanding of the relation between
context, mechanism and outcome (Moore, Graham, and Diamond 2003).
3.3. Ethical considerations
Ethical considerations have been informed by the institutional Research Ethics pol-
icy, the British Educational Research Association (BERA) Revised Ethical Guide-
lines for Educational Research (2004) and the ESRC Research Ethics Framework.
The original proposal underwent an institutional ethical review and was awarded a
Certicate of Approval.
However, the blind randomisation design created a particular ethical problem as
it was not possible to tell participants which treatment group they were in, or the pre-
cise focus of the study. All participants were informed that the study was researching
writing, but not that it was investigating the impact of grammar teaching on writing.
Thus the informed consent was partially compromised. In order to address this, all
participants were informed at the outset that all research results and a full outline of
the conduct of the research would be communicated to them at the end of the study.
A further ethical problem specic to this study was the need to minimise the
effects of designs that advantage ... one group of participants over another (BERA
2004, 8): half the teachers received pedagogical support which could give their clas-
ses an educational advantage/disadvantage. This was addressed in two ways. Firstly,
the period of trialling the support materials attempted to remove any threat of nega-
tive impacts upon student learning during the main study, and both groups were
given matching learning objectives, stimulus resources and written outcomes.
Secondly, at the end of the project, the outcomes of the study were disseminated
directly to participating schools through a teacher conference so that any benecial
impacts could be adopted more widely.
4. The outcomes of the study
4.1. The impact of grammar teaching on students writing
The outcome variable considered in the analysis was the difference between the
post- and the pre-test percentage marks for each of the 744 participants in the study.
The results indicate that both intervention and comparison groups improved over
12 D.A. Myhill et al.
Downloaded by [Mrs Helen Lines] at 08:00 27 November 2011
the time period of the study, with the overall mean value of the improvement
between post- and pre-test scores being 9.24%. However, for the intervention group
(n = 412), the mean outcome was 11.52%, which contrasts with a mean outcome of
6.41% for the comparison group (n = 332). So overall, and ignoring the effect of
any other covariates, a simple two-sample t-test suggests a highly signicant
(p < 0.001) positive difference of 5.11% marks for the intervention in terms of
improvement in writing attainment. In laypersons terms, those students in the inter-
vention group improved their attainment in the post-test writing test signicantly
more than the comparison group. This represents the rst robust statistical evidence
for a benecial impact of the teaching of grammar in students writing attainment.
However, the simple overall comparison above needs to be rened to allow for
signicant effects of school, teacher and student explanatory covariates, so as to
obtain a more nuanced understanding of the effect of the intervention. To achieve
this, multiple regression analysis was applied to the individual student level data
(n = 744). The difference in percentage marks between the post- and pre-writing test
scores for each individual was again used as the outcome variable. The explanatory
covariates were rstly, whether the individual was in the intervention or comparison
group, and then the baseline measures (see Table 1) at student, teacher and school
level, where school and teacher covariate values were attributed to each student
according to his or her school or class membership. Interaction terms between these
covariates were also considered in the modelling. As the covariates included both
categorical measures (e.g. Free School Meals) and continuous measures (e.g. LSK),
the multiple regression modelling could be described more specically as a multi-
way analysis of covariance.
A range of model selection procedures were then used to determine which vari-
ables and interactions were statistically signicant in impacting the response from
amongst the intervention indicator and the other student-, teacher- and school-level
covariates. The model selection was by guided both by theoretical considerations
and size of estimated coefcients, as well as purely statistical procedures such as
stepwise regression techniques and formal statistical tests of differences in explana-
tory power between competing models. Final model selection was then checked
using full residual analyses and standard diagnostic procedures. The results from the
model selection indicated that the most appropriate reduced model involved the
explanatory variables/factors or interactions listed in Table 1. All terms in this model
are very highly statistically signicant (at a .1% level) with the exception of the tea-
cher LSK score which is signicant only at the 5% level. None of the factors, inter-
actions or variables not reported in Table 1 were statistically signicant at the 5%
level and, more importantly, a formal F-test shows no signicant difference in model
t between the reduced model in Table 1 and the full model containing all of the
original variables and interactions. Overall, the reduced model explains only 9% of
the raw variation in the pre- and post-test differences (R
= .09$, adjusted R
= .08$),
but low overall explanatory value is not atypical in educational studies and taken
overall this model is very highly signicant (F
statistic = 10.42, p-value < .00001).
4.1.1. Relationships between the intervention and student writing attainment
Table 3 shows that the intervention had a more marked positive effect on able
writers. It also shows that, over the period of the intervention, able writers in the
comparison group made less progress in writing than less able writers, whereas able
Research Papers in Education 13
Downloaded by [Mrs Helen Lines] at 08:00 27 November 2011
writers in the intervention group made signicant progress. The implications of
these results are rstly that able writers receiving conventional teaching of writing
may be stalling in their progress at age 14, and secondly that the attention to the
relationship between grammar and writing may be particularly appropriate to their
learning needs. The intervention may have been pitched too much towards able
writers: it drew on understanding of students linguistic development from a previ-
ous study (Myhill 2008, 2009). Further research would be necessary to establish
whether using the same pedagogic strategy of embedded grammar teaching but
addressing different aspects of writing more relevant to lower attaining writers
needs would be more successful. It is also possible that able writers had higher lev-
els of cognitive understanding of the concepts being taught, and that they were
more effective in transferring the learning into their own writing.
4.1.2. Relationships between the intervention and teachers LSK and experience
The research design set out to investigate not only whether contextualised grammar
teaching might support learning about writing and improve writing attainment but
also whether teachers condence with grammatical knowledge, or their length of
teaching experience might inuence the intervention in any way. The statistical data
indicate that both these factors were signicant mediating factors. Students in inter-
vention classes with the most and the least experienced teachers beneted less than
those teachers with 510 years experience. There is no supporting evidence or illu-
mination in the qualitative data to explain this result, but it is possible that teachers
with the most experience were more resistant to altering rmly embedded practices,
and that more recently qualied teachers followed the intervention materials in a
more technicist manner, focusing on the activities rather than on the learning.
Further research would be required to draw more robust conclusions.
In terms of LSK, those students in intervention classes with teachers with higher
subject knowledge benetted more than those with teachers who had lower subject
knowledge. This is a less surprising nding, because of the very obvious
pedagogical relationship between LSK and the nature of the intervention. The
teaching materials provided some support for linguistic knowledge through the
Table 3. Coefcient estimates (standard errors, t-values and p -values) for student, teacher
and school level covariates in nal selected model.
Model coefcient Estimate
t-value p-value
Intercept 6.1727 (3.5285) (1.739) (0.723)
Student level
Above average writing level 7.4006 (1.9037) (3.887) ( < 0.001)
Intervention group and above average writing
8.1246 (1.3863) (5.860) ( < 0.001)
Teacher level
Teacher LSK score 0.5817 (0.2760) (2.108) (0.031)
Teacher has 510 years experience 5.9832 (1.3091) (4.570) ( < 0.001)
School level
Satisfactory result in last school inspection 7.0165 (1.9258) (3.643) ( < 0.001)
Good or excellent result in last school
8.3394 (2.0151) (4.138) ( < 0.001)
% SEN in school 0.6411 (0.1610) (3.983) ( < 0.001)
14 D.A. Myhill et al.
Downloaded by [Mrs Helen Lines] at 08:00 27 November 2011
accompanying resources, but nevertheless required condent mastery of grammar,
particularly in making meaningful connections for writers between a linguistic con-
struction and a piece of writing, and in being able to cope with childrens questions.
The qualitative data provide considerable further evidence of this relationship
between the intervention and the teachers linguistic knowledge, and will be dis-
cussed in more detail later in this paper.
Overall, then, the statistical analysis indicates a positive impact of the use of
contextualised grammar teaching on student writing. To date, this is the rst large-
scale study which has found this result. However, the statistical data also indicate
differential effects which are important both for theory, policy and practice. Firstly,
the intervention was more benecial to able writers than weaker writers; secondly,
the LSK of the teacher was a signicant inuencing factor; and thirdly, the length
of teaching experience inuenced learner outcomes. This data allow us to construct
scenarios which describe the different factors which shape the way the intervention
plays out in practice. So, for example, a student who is an able writer in a class
with a teacher of 510 years experience and good LSK is likely to benet more
from this pedagogical approach than a student who is a weaker writer in a class
with a recently qualied teacher whose LSK is limited. These differences are impor-
tant when considering the implications of the study for policy or practice as they
underline the need to take more account of contextual factors beyond an apparent
causal relationship between the pedagogical approach and writing improvement.
4.2. The impact of pedagogical support materials on the teaching of grammar
Given the overall positive results of the intervention, but also the evidence that
other factors, such as LSK, were inuential, it is important to understand how the
teachers used the materials and their views of them. In the teacher interviews, teach-
ers reected on how they had used the intervention materials; in addition, the inter-
viewer used the lesson observation notes to invite discussion of particular aspects
of the lesson relevant to the focus on grammar and writing. Nvivo was used to ana-
lyse the codes inductively: the coding process derived two over-arching categories
from the data, each with its own themes. These were rstly, comments which
addressed teachers views and use of the materials; and secondly, comments which
reected teachers perspectives of student response to the materials (see appendix).
The themes will be italicised in the text below for ease of identication.
4.2.1. Teacher reections on their own practice using the materials
Two themes, Resources/Activities and Timing, were more general comments on the
materials. Many teachers had found the pace of the lessons too fast, and noted tim-
ing issues. That notwithstanding, the comments about the resources and activities
were highly positive: teachers liked the schemes of work and even found them
brilliant to share with the other teachers. These comments tell us little specically
about the intervention, other than conrming professional approval. However, this
approval appeared to assist Fidelity to the teaching intervention. More than half the
teachers maintained they had followed the materials closely because the scheme of
work is really, really good or because students were succeeding in the lessons and
really learning. The theme, Adaptation, gave an insight into the ways adjustments
were made; although we wanted teachers to follow the schemes closely, it was
Research Papers in Education 15
Downloaded by [Mrs Helen Lines] at 08:00 27 November 2011
recognised that teachers had to be free to adapt to meet the needs of their students.
Many adaptations were pragmatic, often to timing, or exchanging one resource for
another similar one. But it is relevant to the research focus that some of the adapta-
tions were to avoid or circumvent the grammar focus because the teacher was not
condent with the material or because the teacher was concerned it would have
just completely have gone over their head.
This concern about the grammar element of the intervention materials was also
reected in the theme, LSK, in which approximately half of the teachers expressed
their anxiety about how well they could handle grammar explanations or student
questions. One teacher reected that
I didnt nd it easy because Im struggling to get my head round understanding some
of these things myself and I think sometimes it shows and I think sometimes the kids
know, and sometimes they throw out answers that Im not quite sure about and I think
my unsureness comes across.
At the same time, teachers felt the materials had given them Condence in various
aspects of writing pedagogy such as the way that I teach imaginative writing, but
signicantly, no teacher directly argued that the teaching materials had developed
their condence with grammar.
One of the pedagogical principles behind the design of the teaching materials
was being explicit in highlighting grammar features where they related to the writ-
ing being taught. Teachers were not informed about the pedagogical principles until
the end of the project so it is noteworthy that, after comments on the practical
aspects of the materials, the theme of Explicitness was the most frequently articu-
lated. Teachers spoke about how the materials had been more explicit about linguis-
tic features than their own normal practice, using words such as explicitly, focused
and specic repeatedly. Whilst teachers felt they frequently taught students about
vocabulary and word choice, the emphasis on syntactic features in the materials
was noted as less usual. Explicit use of grammatical terminology was also noted,
with several teachers reecting that they had never taught determiners and modal
verbs. For one teacher, there was a recognition that the use of terminology sup-
ported students articulation of their understanding about writing, when previously
they hadnt perhaps been given the terminology to be able to express it very
clearly. Many of the comments in this theme identied this explicitness as a distin-
guishing feature of the materials, different from standard practice, and in some cases
participation in the project had changed their viewpoints: Ive changed my mind
about this one, it is crucial to teach children explicitly how to write well.
4.2.2. Teacher reections on student responses to the materials
A substantial set of responses referred to high levels of Student Engagement with
the materials, and there is probably a correspondence between this and the teachers
approval of the materials. There may well, therefore, be halo effects here which
would not be sustained in the longer term. There appeared to be an element of
Surprise at students responses; as one teacher admitted, students have engaged at
a much higher level than I was expecting rst off. Some of this surprise related to
students being able to cope with activities their teachers had not expected them to
manage. This links to another strong theme in the teacher reections, that some
16 D.A. Myhill et al.
Downloaded by [Mrs Helen Lines] at 08:00 27 November 2011
things were Too Difcult for the students. Aspects perceived as difcult always
related to the grammar focus: one teachers students struggled with the noun
phrases, whilst another teachers students found the idea of viewpoint quite dif-
cult to grasp and they kept saying Is it rst person or third person?”’ Conversely,
a high number of comments reected the development of student Understanding,
and the development of explicit knowledge about grammar and writing: they obvi-
ously got the whole thing about connectives as to which ones work best when and
why and what have you. Signicantly, this understanding was purposeful knowl-
edge applied to writing; as one teacher noted, Theyve far more understanding of
what it is theyre doing it and why theyre doing it.
Four of the themes have direct links with the pedagogical principles underpin-
ning the intervention. Teachers enjoyed the Discussion opportunities provided in the
schemes and reported brilliant discussions where students were willing to risk
opinions about language more. They recognised that the discussion helped to create
transferable knowledge, Making Connections for learners between the grammar and
the writing, and generating greater learner independence where students were mak-
ing a link without me. This chimes with the aims of the teaching materials which
sought to avoid generating formulaic knowledge about writing and which endeav-
oured to open up students to a repertoire of possibilities to draw on in writing. One
strategy to support this was the use of activities which encouraged language play.
Teachers felt that this Experimentation had created space for valuable play, and par-
ticularly required that they handed over control a little bit and allow them to just
experiment. Teachers also believed that students were making informed decisions
about their writing, and that they had understood the importance of making choices
and making decisions about your reading and your writing.
Finally, at least half the teachers believed the teaching materials had supported
Writing Improvement. It is important to be cautious in claims made on the basis of
this as the research project was obviously addressing writing improvement and
again there may be halo effects here. For some teachers there was a perception of a
direct relationship between the intervention and the improvement:
I do feel that all this work weve been doing on working with words, working with
sentence structures, working with how you work with these different word classes to
enhance your writing, Im sure thats fed into it because it was beautiful work.
In conclusion, then, the evidence from the teacher interviews suggests two levels of
impact of the intervention teaching materials on the teaching of grammar. Firstly,
the teachers found the materials high-quality, stimulating strong student engage-
ment. These are qualities which could apply to any teaching materials and are not
specic to the grammar intervention. The second level of impact is speci cto
grammar teaching. The teachers comments suggest that some of the pedagogical
principles underpinning the planning had signicant impact on their teaching. The
explicitness of the attention to linguistic features built into the teaching materials
and the creation of multiple opportunities for discussion are the most prominent of
these, but the emphasis on playfulness and experimentation, and on developing rep-
ertoires of possibility through making connections and encouraging informed deci-
sion-making are also important. Conversely, in addition to expressing concern about
their linguistic knowledge, some teachers demonstrated their anxiety in practice by
avoiding some of the grammar points in the lessons, and their strong sense that the
Research Papers in Education 17
Downloaded by [Mrs Helen Lines] at 08:00 27 November 2011
materials might be too difcult for students may reect their own difculty as much
as student difculty.
4.3. The impact of the intervention on students metalinguistic understanding
The student interviews included the writing conversations, where the student was
invited to comment on a prompt piece of writing of the same genre under study,
and to reect on their own writing so far. The coding of students metalinguistic
understanding drew on Gomberts taxonomy (1992) of metalinguistic understanding
as the basic framework for analysis. Gombert identied six categories of metalin-
guistic understanding (metaphonological, metalexical, metasemantic, metasyntactic,
metapragmatic and metatextual), which will be summarised here as word-, sentence-
and text-level understanding. For a fuller account of this aspect of the study, see
Myhill (2011a). Students in both the intervention and comparison groups were
given the same prompts and the same opportunities to reveal metalinguistic under-
standing: however, students in the intervention group made more comments dened
as evidence of metalinguistic understanding, and made more elaborated comments.
This implies that one element of the success of the intervention was in developing
metalinguistic understanding about writing. But their condence in articulating
metalinguistic understanding was not equal across word-, sentence- and text-level
domains, as will be outlined below.
4.3.1. Word-level metalinguistic understanding
Without doubt, the greatest condence in discussing linguistic choices in writing
was at word level. The students were very aware of the impact and signicance of
word choices, understanding that words helped create a picture in your mind or
contributed to the creation of a particular viewpoint, showing what the person in
the story is actually seeing. They understood the potential emotiveness of word
choices, of words that are going to sink into the heart and how the choice of
appropriate vocabulary makes you feel something that the writer wants you to
feel. When talking about his own piece of argument writing, one writer observed
that he had chosen the words downhearted, inhumane, defenceless, neglected to
describe a character in order to make people feel sorry for him.
The prompt text for argument included a mix of formal and informal vocabulary:
for example, the writer used elderly rather than old, which is a more formal
option, and conversely used the verb carted, rather than an alternative such as
taken. Many students recognised that word choices were important in increasing
the formality of text, and several commented on the choice of the word older gener-
ation rather than older people
, a choice which one students described as not posh
but pretty civilised. The word disgraceful stimulated many comments, which often
combined a recognition that the word was emotive and a more formal choice. As
one student observed: Disgraceful is a really strong word to use ... instead of say-
ing Oh it s horrible, because that probably wouldnt grab your attention if you just
said its horrible because its quite a common word to use, lots of people use it.
4.3.2. Sentence-level metalinguistic understanding
Gomberts (1992) denition of metasyntactic understanding is that it represents
conscious reasoning about syntax and deliberate control over accurate use of
18 D.A. Myhill et al.
Downloaded by [Mrs Helen Lines] at 08:00 27 November 2011
grammar. One difculty with Gomberts taxonomy is that it relates to oral develop-
ment and is particularly focused, therefore, on young learners and how they develop
grammatical accuracy in speech. In our study, only the weakest writers presented
grammatical inaccuracies in their writing, and then only rarely. More relevant to this
study was these students ability to understand and discuss how syntactic choices
helped writers to meet their rhetorical goals. The teaching schemes drew explicit
attention to aspects of syntactic variety as part of a repertoire for shaping text: the
possibilities of short sentences for impact, varying sentence lengths to create textual
rhythm, and about altering the syntactic structure of sentences to shift focus. Over-
all, however, it is this aspect of metalinguistic understanding which seemed to be
least secure in this sample.
Greatest condence was shown in discussing the use of short sentences for par-
ticular effects and in recognising the value of variety in sentence length within a
text. The way a short sentence can draw attention to itself as a kind of prosodic
exclamation seemed to be understood: students talked of short sentences which
grab attention and make[s] you think about it. However, whilst the students have
grasped that variety might be a good writers choice, they are less assured in articu-
lating any reasons linked with prosody or textual rhythm. Indeed, in some cases, it
was thought to be more to do with supporting reader understanding than creating
an effect: If theres a really long sentence youre probably thinking, oh well I cant
remember what the beginning of the sentence was.
Understanding the possibilities of syntactic variety in altering how a sentence is
read was much less well understood, despite this being one focus of the teaching.
Some writers were able to make fairly precise comments about syntactic choices, as
this writer evidences: Ive put the connective at the start like I said before and I
wouldnt normally do that, normally it would be The pig will have nowhere to
sleep or something like that, but Ive put Despite the fact that the pig will have
nowhere to sleep. However, this writer was unable to explain why this choice had
been made and for what purpose. It is possible that this writer, and indeed many of
the other students in the study, do not yet have sufcient explicit syntactic
knowledge to articulate their emergent metasyntactic understandings.
4.3.3. Text-level metalinguistic understanding
Gombert (1992) argued that metatextual understanding, a more global awareness of
coherence and cohesion, develops later than metalexical and metasyntactic under-
standing. There were fewer comments coded as metatextual in this study, but the
teaching focus was directed more to word- and sentence-level aspects of writing,
which may account for this. In the poetry unit, we had anticipated there might be
more opportunities for metatextual discussion, because the poem as a text is very
visible as a whole and often deliberately organised to exploit textual possibilities.
However, students seemed to struggle to talk about a poem as a text, focusing more
on layout than structure, for example, the writer who observed that the poem
couldnt be a story because its too short and it couldnt just be a passage of writ-
ing, because each sentence starts the same way.
There was more evidence in the narrative and argument units of students having
awareness of the text as a whole. In the narrative scheme, one writer articulated a
clear text-level decision that Its going to be told in rst person by one of the
tramps called Toby and another explained her use of tense to manage different time
Research Papers in Education 19
Downloaded by [Mrs Helen Lines] at 08:00 27 November 2011
points: it will be past at the beginning from when hes telling the story of his past
but towards the main bit it will be present, so it will go from past to present. Stu-
dents were also condent explaining plot development and narrative structure,
including understanding that narrative can be non-chronological and you dont nec-
essarily have to put the opening at the beginning, you can put it in the middle.In
the argument scheme, students were developing awareness of how the argument
was structured across the text:
Theyve put the important points at the top and important points at the bottom and in
the middle theyve kind of tried to persuade you its bad. Theyve put a summary of
the whole kind of speech in a couple of sentences at the end which is good, kind of
refreshes your memory.
In summary, the data not only show that the intervention had a positive impact
upon developing metalinguistic understanding but also signals that metalinguistic
learning about writing is socially constructed, and in this case heavily shaped by
what teachers value in writing. Many of the comments, both those demonstrating
strong metalinguistic understanding and those showing more limited understanding,
frequently mirrored closely what teachers had said in the lessons. In particular,
because some of the teachers did not have sufcient LSK to handle metalinguistic
discussion condently, students understanding was correspondingly limited. Many
of the teachers were more comfortable teaching about word choices than syntactic
variety, which, as will be outlined below, reects relatively less assurance with syn-
tax than with word grammar.
4.4. The impact of teacher LSK on the teaching of grammar
The statistical data, as reported earlier, indicated that teacher LSK was a mediating
factor in inuencing student outcomes. The teachers scores on the LSK test were
very evenly spread across the cohort (see Table 4) with a mean result of 60% and a
fairly high standard deviation of 15.8%. Only one of the teachers in this sample
had a degree which included a linguistics element, and this teacher, not surprisingly,
scored 86%. Given that older teachers may have been taught grammar as part of
their own education, the results were analysed to see if there was any relationship
Table 4. Teachers LSK scores.
20 D.A. Myhill et al.
Downloaded by [Mrs Helen Lines] at 08:00 27 November 2011
between years of teaching experience and linguistic knowledge, but there is no
strong evidence that this is so. The top four results do include three teachers with
more than 28 years experience, but equally the second lowest result is from a tea-
cher with 23 years experience.
The lesson observations in particular, but also the teacher and student inter-
views, provided explanatory evidence of how LSK impacted upon the teaching.
Overall, there were three ways in which a lack of condence in grammar was rea-
lised in the classroom or in reections on practice:
making meaningless comments about grammar;
the use of semantic denitions and
syntactic confusions.
4.4.1. Meaningless grammar
Teacher comments to students during observed lessons sometimes included advice
to writers which either made no sense at all, or was insufciently elaborated or
explained to be meaningful for students. One teacher told her students that if you
use verbs, adverbs or nouns, you will be able to write a very powerful description,
which is not helpful it would be hard to write at all without using verbs, adverbs
or nouns and, moreover, it is perfectly possible to write weak and ineffective
descriptions using verbs, adverbs and nouns. Another set of less helpful comments
related to the idea of sentence variety, which was a teaching focus of the schemes
of work. Teachers regularly advocated the use of variety: variety is important;
make sure you have sentence variety. However, there was rarely any explanation
of why this variety was benecial, and implying that variety, of whatever quality,
was a good thing. In contrast, one teacher with good LSK, gave a more precise and
meaningful reason for using varied sentence lengths which made a link between the
linguistic feature and how it might impact upon the writing. In the argument
scheme of work, she was discussing how students could use contrast in sentence
length in different ways: in a long sentence you can detail the cruelty and a short
sentence you can refer to sudden death for impact.
Another tendency was to promote the use of a particular linguistic feature for
think about where you put your punctuation for effect;
use sentences for effect;
vary vocabulary for effect and
short sentence used for effect.
To an extent, these reect the teaching materials which repeatedly encouraged
discussion about the effects of grammar features, but many teachers lacked the
applied linguistic knowledge which allowed them to move beyond the phrase for
effect to a more text- or context-specic discussion of the possible effects created.
4.4.2. The use of semantic denitions
Another pattern of response which links to lack of condence in handling linguistic
terminology was the strong tendency to give students semantic denitions for word
Research Papers in Education 21
Downloaded by [Mrs Helen Lines] at 08:00 27 November 2011
classes, rather than linguistically precise descriptions. So verbs were regularly dened
as doing words, thus leading to student difculties when they encountered verbs
which do not appear to involve any action (e.g. are, will, wonder, consider), but espe-
cially when they encountered words which were not verbs but which implied an
action (as in I am deeply opposed to hunting where many children identify hunting
as the doing word because of its implied action). Frequently, the semantic denitions
offered were partial, such as adverbials are size words, or a noun is the name of an
object. Adjectives were regularly dened as describing words, without acknowl-
edgement of the descriptive power of lexical verbs, nouns and adverbs. The tendency
to use semantic denitions is inuenced by common practice over many years in both
primary and secondary classrooms, practice which is often endorsed in commercial
materials. Occasionally, however, the semantic denition was completely unique as
was the case with the teacher who described an adverb as an action plus word.
These semantic denitions led to some very confused discussions with students
because they held on very tightly to the semantic denition and applied it with
absolute logic. In one lesson observed, the teacher had earlier suggested that a noun
was something you can touch, which resulted in the following exchange:
Teacher: What are the rules for whether it is a noun or a verb or something else?
Student: Its if you can touch it.
Student: Can you touch it?
Student: Can you go to it?
Teacher: Can you touch hockey? But hockey is a noun.
Student: You play hockey so it must be a doing word.
Teacher: Is her a noun?
Student: Yes, you can touch her.
Teacher: Can safe be a noun?
Student: Safe is a feeling not a thing.
Sometimes the students recognised the aw in these semantic denitions. In one
student interview, one girl suggested she could improve her writing by adding more
adjectives to strengthen the description, but then added If you think about it, all
words are adjectives because theyre all describing things. A noun is describing.
4.4.3. Syntactic confusion
The aspect of the LSK test in which teachers scored least well was the set of ques-
tions on clauses and syntax, suggesting this is an area of particular challenge. In the
interviews, many teachers articulated a specic anxiety about clauses. In the lessons
observed, teachers often chose to focus on sentence variety in terms of sentence
length, as this is very easy to handle, and requires no grammatical explanations.
Many ungrammatical variations on sentence types were developed, with the gram-
matical distinctions of simple, compound and complex being extended with con-
cepts such as more simple, very complex , a normal sentence and a
more than
average sentence. Linked to the issue of using semantic denitions for word clas-
ses, described above, at a syntactic level, the concept of grammatical simplicity was
confused with semantic simplicity, so simple sentences were short sentences and
complex sentences were long sentences. This led to one student offering this
explanation in an interview:
22 D.A. Myhill et al.
Downloaded by [Mrs Helen Lines] at 08:00 27 November 2011
Student: There was a cat is a simple sentence; a complex sentence is like, There
was a slim, something, something ginger cat.
In contrast, some teachers demonstrated more condent management of discussion
of syntactic features which linked them very explicitly to the way they were work-
ing in the specic piece of writing under focus. In the narrative ction scheme of
work, one teacher responded to a students draft with the feedback that I like the
way youve kept some short sentences in to build the tension, and another drew
attention to the way adverbials can create a sense of place and setting: Weve got a
real sense of the environment with adverbials in there. On other occasions, teachers
deftly drew attention to the subtleties of making changes to the standard subject
verb order of a sentence:
Look at this and the way its been changed. Sometimes you can change the
structure of a sentence to make it more interesting.
Look whats happened by changing the word order. As a writer you can with-
hold information and build a sense of expectation.
These data reveal how basic problems with declarative knowledge of linguistic
metalanguage, particularly syntactic knowledge, and knowledge about the mobility
of word classes in English generate very real problems in working constructively
with grammar in the context of writing. Where teachers own subject knowledge
was limited, there were frequently corresponding applied pedagogical problems in
providing adequate denitions or explanations of linguistic terms. The reliance on
semantic explanations, rather than functional explanations, often confused students,
and teachers sometimes found it hard to handle student questions or conceptual
problems with condence. Signicantly for the focus of this study, limitations in
LSK meant that some teachers struggled to make meaningful links for students
between a linguistic feature and its effect or purpose in a specic text. Conversely,
where teachers had greater command of the LSK, they were better able to make
purposeful connections between grammar and writing, and were more condent
managing discussion about effects and possibilities. Teachers with condence in
LSK helped writers shape text creatively; teachers who lacked condence provided
formulaic recipes for success.
5. Implications for theory, policy and practice
The study represents the rst large-scale study in any country of the benets or
otherwise of teaching grammar within a purposeful context in writing. It stands in
contrast to previous studies which were either small-scale (Bateman and Zidonis
1966; Fogel and Ehri 2000) or which investigated whether discrete grammar
instruction improved writing outcomes (Elley et al. 1975, 1979), and is the only
study of its kind conducted in England. It is also signicant in combining comple-
mentary qualitative and quantitative methods, which have provided not only statisti-
cal evidence of the impact of the intervention on writing outcomes, but also a
richly nuanced understanding of how the intervention worked. The intervention
materials were theoretically informed by the principles of writing as design (Cope
and Kalantzis 2000; Kress 1994; Kress and van Leeuwen 1996; Sharples 1999), in
which creators of text make design choices from an available repertoire of
Research Papers in Education 23
Downloaded by [Mrs Helen Lines] at 08:00 27 November 2011
possibilities. The use of grammar as a design tool (Myhill 2010, 2011b) elaborates
thinking about writing as design beyond the visual and the multimodal to include
specic linguistic decision-making. Alongside this, therefore, the intervention also
sought to develop students metalinguistic understanding through explicit instruction
and through opportunities for discussion. The strong positive effect of the interven-
tion signals for the rst time the potentiality of grammar as an enabling element in
writing development and evidences a clearly theorised role for grammar in writing
pedagogy. However, the result that the intervention beneted able writers most is
also important. Further studies should usefully investigate whether teaching materi-
als designed around the identied linguistic learning needs of less able writers
would have a benecial effect; and whether the use of materials which incorporate
explicitness and discussion of linguistic focuses but without the use of the metalan-
guage might be more supportive to less able writers.
The study also highlights the importance of teachers LSK in mediating metalin-
guistic knowledge in the writing classroom. The teacher interviews and lesson
observations indicate that teachers need to possess condence in declarative knowl-
edge of grammar, particularly syntactic knowledge, if they are to be able to handle
students questions and misunderstandings effectively. This would include an ability
to dene and explain metalinguistic terminology appropriately. This is an aspect of
grammar teaching which has been systematically overlooked at both policy level
and in research; there is neither understanding nor agreement about how best to
explain grammatical terminology. At the same time, however, the study makes it
clear that declarative knowledge alone is insufcient. As Parr (2009) argues, one
facet of pedagogical subject knowledge for the teaching of writing is the ability to
articulate and make accessible to developing writers that which is implicit and often
at a level below conscious (2009, 147); on the one hand this implicit knowledge
may be about writers and the process of writing, but on the other hand, it is about
texts and how texts work. Teachers need to be able to apply their LSK to published
texts and to childrens own writing, identifying signicant linguistic features and
being able to make connections for writers between a feature and its impact on a
text or reader.
At policy level, the study suggests judicious caution about too simplistic an
advocacy of or legislation for a specic pedagogical practice. Whilst there is now
robust evidence from the data in favour of the use of grammar in an embedded way
within the teaching of writing, the study certainly does not suggest that this would
be of universal benet. Rather it emphasises the complex inter-relatedness of many
factors in the realisation of educational benet; particularly in terms of learners
needs, teachers attitudes and experience and teachers subject knowledge. Policy
development needs to take these interacting factors into account and, in particular,
consider how to develop professional pedagogical ownership of policy in ways
which foster principled adaptation to meet learners needs and interests.
Finally, we think it is important to acknowledge the limitations of the study, and
especially the RCT design. Whilst there are benets in the large-scale quantitative
data, the RCT may be too focused on causal relationships between an intervention
and student outcomes, paying insufcient attention to other factors. In seeking to
generalise, it always has the potential to miss the particular. The emphasis on prin-
ciples such as intervention delity, blind randomisation and bias can serve to
exclude the very variables which are most signicant should the intervention be
generalised into professional practice. Further studies which build on this study
24 D.A. Myhill et al.
Downloaded by [Mrs Helen Lines] at 08:00 27 November 2011
might consider the use of a series of smaller-scale, but nonetheless statistically
robust, interventions which combine the testing of the impact of specic interven-
tions with full involvement of the teachers in understanding the pedagogical princi-
ples underlying the intervention. In particular, we would recommend that teachers
design and develop the teaching materials for any intervention themselves, with
guidance from the research team, thus taking ownership of the pedagogical princi-
ples which inform the study.
Notes on contributors
Debra A. Myhill is a professor of education in the Graduate School of Education at the
University of Exeter and professorial fellow at the University of Wollongong, Australia. She
researches in language education, principally in the teaching of writing and the relationship
between grammar and writing. She has recently co-authored the book Writing voices:
Creating communities of writers (London: Routledge) with Teresa Cremin.
Susan M. Jones is a senior lecturer in the Graduate School of Education at the University of
Exeter. Her research interests are in language education, including talk for learning, writing
and writing processes, and gender.
Helen Lines is a research fellow in the Graduate School of Education at the University of
Exeter and was lead research fellow on the ESRC project reported here. Her own research
interests are assessment in writing, and how teachers conceptualise quality in writing.
Annabel Watson is a PhD student in the Graduate School of Education at the University of
Exeter. She won an ESRC-funded studentship linked to the project reported here and was
part of the research team. Her doctoral research is investigating in teachers beliefs about the
role of grammar in writing pedagogy.
Alamargot, D., and L. Chanquoy. 2001. Through the models of writing. Dordrecht: Kluwer
Andrews, R. 2005. Knowledge about the teaching of sentence grammar: The state of play.
English Teaching: Practice and Critique 4, no. 3: 6976.
Applebee, A. 2000. Alternative models of writing development. In Perspectives on writing:
Research, theory and practice, ed. R. Indrisano and J. Squire, 90111. Newark, NJ:
Bateman, D.R., and F.J. Zidonis. 1966. The effect of a study of transformational grammar
on the writing of ninth and tenth graders. Champagne, IL: National Council of Teachers
of English.
Bereiter, C., and M. Scardamalia. 1982. From conversation to composition: The role of
instruction in a developmental process. In Advances in instructional psychology, Vol. 2,
ed. R. Glaser, 164. Hillsdale, NJ: Lawrence Erlbaum.
Braddock, R., R. Lloyd-Jones, and L. Schoer. 1963. Research in written composition.
Urbana, IL: National Council of Teachers of English.
British Educational Research Association (BERA). 2004. Revised ethical guidelines for edu-
cational research: 2004. Nottingham: BERA.
Buttereld, E.C., D.J. Hacker, and L.R. Albertson. 1996. Environmental, cognitive and meta-
cognitive inuences on text revision. Educational Psychology Review 8, no. 3: 23997.
Carter, R., ed. 1990. Knowledge about language. London: Hodder and Stoughton.
Cope, B., and M. Kalantzis. 2000. Multiliteracies: Literacy learning and the design of social
futures. London: Routledge.
Denham, K., and A. Lobeck. 2005. Language in the schools: Integrating linguistics into
K12 teaching. Hillsdale, NJ: Lawrence Erlbaum.
Research Papers in Education 25
Downloaded by [Mrs Helen Lines] at 08:00 27 November 2011
Donner, A., and N. Klar. 2000. Design and analysis of cluster randomization trials in health
research. London: Arnold.
Elbow, P. 1981. Writing with power: Techniques for mastering the writing process. New
York, NY: Oxford.
Elley, W.B., I.H. Barham, H. Lamb, and M. Wylie. 1975. The role of grammar in a
secondary school curriculum. New Zealand Council for Educational Studies 10: 2641.
Elley, W.B., I.H. Barham, H. Lamb, and M. Wylie. 1979. The role of grammar in a second-
ary school curriculum: Educational research series no. 60. Wellington: New Zealand
Council for Educational Research.
Englert, C.S., T.E. Raphael, and L.M. Anderson. 1992. Socially mediated instruction:
Improving students knowledge and talk about writing. The Elementary School Journal
29, no. 4: 41149.
Evidence for Policy and Practice Information and Co-ordinating Centre (EPPI-Centre) Review
Group for English. 2004. The effect of grammar teaching (syntax) in English on 5 to 16
year olds accuracy and quality in written composition. London: EPPI-Centre.
Flavell, J. 1976. Metacognitive aspects of problem-solving. In The nature of intelligence
hillsdale, ed. L.B. Resnick, 2316. Hillsdale, NJ: Erlbaum.
Fogel, H., and L.C. Ehri. 2000. Teaching elementary students who speak black English
vernacular to write in standard English: Effects of dialect transformation practice.
Contemporary Educational Psychology 25: 21235.
Gombert, J.E. 1992. Metalinguistic development. Hemel Hempstead: Harvester-Wheatsheaf.
Gombert, J.E. 2003. Implicit and explicit learning to read: Implication as for subtypes of
Dyslexia. Current Psychology Letters: Behaviour, Brain and Cognition 10, no. 1. Special
Issue on Language Disorders and Reading Acquisition.
ment202.html (accessed March 2006).
Gordon, E. 2005. Grammar in New Zealand schools: Two case studies. English Teaching:
Practice and Critique 4, no. 3: 4868.
Gurrey, P. 1962. Teaching English grammar. London: Longmans.
Harris, R.J. 1962. An experimental enquiry into the functions of and value of formal
grammar in the teaching of English. Unpublished PhD thesis, London University.
Hayes, J.R., and L.S. Flower. 1980. Identifying the organisation of writing processes. In
Cognitive processes in writing, ed. L.W. Gregg and E.R. Steinberg, 330. Hillsdale, NJ:
Lawrence Erlbaum.
Hudson, R. 2004. Why education needs linguistics. Journal of Linguistics 40, no. 1:
Keith, G. 1997. Teach yourself English grammar. The English and Media Magazine, No 36.
Kellogg, R.T. 1994. The psychology of writing. Oxford: Oxford University Press.
Koln, M., and C. Hancock. 2005. The story of English grammar in US schools. English
Teaching: Practice and Critique 4, no. 3: 1131.
Kress, G. 1994. Learning to write. London: Routledge.
Kress, G., and T. van Leeuwen. 1996. Reading images: The grammar of visual design .
London: Routledge.
Locke, T. 2005. Grammar wars beyond a truce. English Teaching: Practice and Critique
4, no. 3: 1
LSA. 2006. Minutes of the meeting of the committee on language in the school curriculum.
LSA: Albuquerque. (accessed
July 2006).
Martlew, M. 1983. Problems and difculties: Cognitive and communicative aspects of writ-
ing. In The psychology of written language: Developmental, educational perspectives,
ed. M. Martlew, 295333. Chichester: Wiley.
Masters, G., and M. Forster. 1997. Literacy standards in Australia. Melbourne: Australian
Council for Educational Research.
Micciche, L. 2004. Making a case for rhetorical grammar. College Composition and
Communication 55, no. 4: 71637.
Moore, L., A. Graham, and I. Diamond. 2003. On the feasibility of conducting randomized
controlled trials in education. British Education Research Journal 29, no. 5: 67389.
Murray, D. 1998. Design and analysis of group-randomized trials. Oxford: Oxford
University Press.
26 D.A. Myhill et al.
Downloaded by [Mrs Helen Lines] at 08:00 27 November 2011
Myhill, D.A. 2003. Principled understanding? Teaching the active and passive voice.
Language and Education 17, no. 5: 35570.
Myhill, D.A. 2008. Towards a linguistic model of sentence development in writing. Lan-
guage and Education 22, no. 5: 27188.
Myhill, D.A. 2009. Becoming a designer: Trajectories of linguistic development. In The sage
handbook of writing development, ed. R. Beard, D. Myhill, J. Riley, and M. Nystrand,
40214. London: SAGE.
Myhill, D.A. 2010. Ways of knowing: Grammar as a tool for developing writing. In Beyond
the grammar wars: A resource for teachers, students on developing language knowledge
in the English/literacy classroom, ed. T. Locke, 12948. London: Routledge.
Myhill, D.A. 2011a. The Ordeal of Deliberate Choice: Metalinguistic development in sec-
ondary writers. In Past, present, future contributions of cognitive writing research to
cognitive psychology, ed. V. Berninger, 24774. New York, NY: Psychology Press/Taylor
Francis Group.
Myhill, D. A. 2011b. Grammar for designers: How grammar supports the development of
writing. In Insight and impact: Applied linguistics and the primary school, ed. S. Ellis,
E. McCartney, and J. Bourne, 8192. Cambridge: Cambridge University Press.
Myhill, D.A., and S. Jones. 2007. More than just error correction: Children sreections on
their revision processes. Written Communication 24, no. 4: 32343.
Myhill, D.A., H.E. Lines, and A. Watson. 2011. Making meaning with grammar: A reper-
toire of possibilities. Metaphor 2: 110.
National Commission on Writing in Americas Schools and Colleges. 2003. The neglected
R: The need for a writing revolution. New York, NY: College Board.
Parr, J. 2009. Building professional knowledge to teach writing: Teacher pedagogical content
knowledge and student achievement in writing. Presented at the 45th UKLA annual
conference, Literacy Today and Tomorrow, July 1012, 2009, in University of Greenwich.
Qualications and Curriculum Authority (QCA). 1998. The grammar papers: Perspectives
on the teaching of grammar in the National Curriculum. London: QCA.
Robinson, N. 1959. The relation between knowledge of English grammar and ability in
English compositions. Unpublished MEd thesis, University of Manchester.
Shadish, W.R., T. Cook, and D. Campbell. 2002. Experimental and quasi-experimental
designs for generalized causal inference. Boston, MA: Houghton Mifin.
Sharples, M. 1999. How we write: Writing as creative design. London: Routledge.
Tomlinson, D. 1994. Errors in the research into the effectiveness of grammar teaching.
English in Education 28, no. 1: 206.
United States Department of Education. 2002. No Child Left Behind (NCLB) Act of 2001.
Washington, DC: US GPO.
Van Gelderen, A., and R. Oostdam. 2005. Effects of
uency training on the application of
linguistic operations in writing. L1 Studies in Language and Literature 5, no. 2: 21540.
Van Lier, L. 1998. The relationship between consciousness, interaction and language
learning. Language Awareness 7, no. 2&3: 12845.
Wallace, D.L., and J.R. Hayes. 1992. Redening revision for freshmen. In Reading empirical
research studies: The rhetoric of research, ed. J.R. Hayes and R.E. Young, 35269.
Hillsdale, NJ: Lawrence Erlbaum.
Wyse, D. 2004. Grammar. For writing? A critical review of empirical evidence. British
Journal of Educational Studies 49, no. 4: 41127.
Research Papers in Education 27
Downloaded by [Mrs Helen Lines] at 08:00 27 November 2011
Appendix: Summary of coding themes
Number of
Teacher-focused themes
Explicitness Comments claiming the schemes of work were more
explicit about grammar or teaching writing than in their
normal practice
Comments about the quality of the resources and activities 59
Fidelity Comments which seem to suggest the teacher has stuck to
the schemes of work
Adaptation Comments which refer to things the teacher changed or
dropped from the schemes of work
Comments which refer to teacher uncertainty about LSK 21
Condence Comments which refer to how the schemes of work gave
them condence
Timing Comments which refer to problems with timing or amount
of material to be covered
Student-focused themes
Surprise Comments in which teachers express surprise at some
aspect of students work
Comments which refer to how the schemes of work or
students are making connections between grammar and
Making informed
Comments which refer to students making conscious
choices in writing
Experimentation Comments which claim the schemes of work encouraged
students to experiment or play with language
Discussion Comments which refer to the schemes of work provoking
student discussion
Comments about ways in which the writing had improved
or not
Comments which refer to activities or resources which
engaged the students
Understanding Comments which refer to students understanding 45
Too difcult Comments which refer to things not taught because of a
perception they were too difcult for the students
28 D.A. Myhill et al.
Downloaded by [Mrs Helen Lines] at 08:00 27 November 2011
... Several of the included articles indicate conceptual confusion about grammar among teachers and an uncertainty concerning their subject knowledge and the application of grammatical knowledge (Camps & Fontich, 2019;Myhill, Jones et al., 2012;Watson, 2012Watson, , 2015. In addition, one major assertion within the data is that teachers' grammatical and linguistic subject knowledge is crucial to their ability to teach grammar based on a descriptive model (Cushing, 2018;Myhill, 2000;Myhill, Jones et al., 2012;Smagorinsky et al., 2011;Van Rijt et al., 2019;Watson, 2012Watson, , 2015. ...
... Several of the included articles indicate conceptual confusion about grammar among teachers and an uncertainty concerning their subject knowledge and the application of grammatical knowledge (Camps & Fontich, 2019;Myhill, Jones et al., 2012;Watson, 2012Watson, , 2015. In addition, one major assertion within the data is that teachers' grammatical and linguistic subject knowledge is crucial to their ability to teach grammar based on a descriptive model (Cushing, 2018;Myhill, 2000;Myhill, Jones et al., 2012;Smagorinsky et al., 2011;Van Rijt et al., 2019;Watson, 2012Watson, , 2015. A relationship between teachers' low degree of grammatical subject knowledge, their negative associations and reported prescriptive practices of grammar teaching is visible Watson, 2012;Wilson & Myhill, 2012). ...
... The underscoring of the importance of converting theoretical knowledge into practice is discernible in the material. Several studies emphasize that the challenges of grammar teaching cannot only be ascribed to the lack of language courses in teacher education, since teachers also need educational tools to put the knowledge and findings from research into practice (Camps & Fontich, 2019;Clark, 2010;Myhill, Jones et al., 2012;. Teacher education plays an important role in teachers' engagement with research. ...
Conference Paper
Full-text available
Denna artikel behandlar finskans och somaliskans roll i den göteborgska språk- hierarkin. Analysen bygger främst på intervjuer med talare av de två språken, kompletterat med statistik och en fotodokumentation av språkens synlighet i fyra Göteborgsområden. Finskan var länge det största minoritetsspråket i Göteborg, men det har liten synlighet och hade länge relativt låg status. Somaliskan som är ett nyare minoritetsspråk i Göteborg tillskrivs låg social status men präglas av en större medial och visuell synlighet. Idag har finskan högre mobilitet i Göteborg än för tjugo år sedan, medan somaliskan präglas av lägre lokal mobilitet, även om språket är betydligt mer spritt globalt. Vi menar att skillnaderna mellan de två språkens status och synlighet i Göteborg med fördel kan förklaras genom en intersektionell analys som tar hänsyn till platsens flerskiktade historia och de historiska och nutida talares kroppar som bebor denna plats.
... Several of the included articles indicate conceptual confusion about grammar among teachers and an uncertainty concerning their subject knowledge and the application of grammatical knowledge (Camps & Fontich, 2019;Myhill, Jones et al., 2012;Watson, 2012Watson, , 2015. In addition, one major assertion within the data is that teachers' grammatical and linguistic subject knowledge is crucial to their ability to teach grammar based on a descriptive model (Cushing, 2018;Myhill, 2000;Myhill, Jones et al., 2012;Smagorinsky et al., 2011;Van Rijt et al., 2019;Watson, 2012Watson, , 2015. ...
... Several of the included articles indicate conceptual confusion about grammar among teachers and an uncertainty concerning their subject knowledge and the application of grammatical knowledge (Camps & Fontich, 2019;Myhill, Jones et al., 2012;Watson, 2012Watson, , 2015. In addition, one major assertion within the data is that teachers' grammatical and linguistic subject knowledge is crucial to their ability to teach grammar based on a descriptive model (Cushing, 2018;Myhill, 2000;Myhill, Jones et al., 2012;Smagorinsky et al., 2011;Van Rijt et al., 2019;Watson, 2012Watson, , 2015. A relationship between teachers' low degree of grammatical subject knowledge, their negative associations and reported prescriptive practices of grammar teaching is visible Watson, 2012;Wilson & Myhill, 2012). ...
... The underscoring of the importance of converting theoretical knowledge into practice is discernible in the material. Several studies emphasize that the challenges of grammar teaching cannot only be ascribed to the lack of language courses in teacher education, since teachers also need educational tools to put the knowledge and findings from research into practice (Camps & Fontich, 2019;Clark, 2010;Myhill, Jones et al., 2012;. Teacher education plays an important role in teachers' engagement with research. ...
Full-text available
This article reports a literature review of peer-reviewed studies of L1 grammar teaching, published in international English-language journals from 1999 to 2019. The review generates interpretative themes of representations of L1 grammar teaching and thus offers guidance for teachers and future research on grammar education. The articles (n=31) were analyzed using reflexive thematic analysis to generate interpretative themes concerning representations of grammar teaching within the field. The results show that grammar teaching is described as challenging , mainly due to differing perceptions of grammar between linguists and teachers. In addition, the second theme captures grammar teaching as something that aims to foster metalinguistic understanding. Finally, contextualized grammar teaching is discernible which is carried out through reflective interaction and experimentation. This study suggests that future research should reconceptualize grammar teaching building on these findings using practice-based research approaches where teachers are offered agency in the research process.
... JoTTER Vol. 12 (2021) ã Georgina Thorpe, 2021 208 Myhill et al. (2012) demonstrate that these findings can be replicated in a larger study when the focus remains on contextualised grammar teaching for writing. Their sample consisted of 744 students with 31 teachers from randomly selected schools, with a comparison group and an intervention group (who received explicit grammar instruction) randomly assigned. ...
... Their sample consisted of 744 students with 31 teachers from randomly selected schools, with a comparison group and an intervention group (who received explicit grammar instruction) randomly assigned. The researchers focused on a grammar of choice intended to be used in writing, stating they aimed "to help writers to recognise how making grammatical choices could shape their texts for communicative purposes" (Myhill et al., 2012(Myhill et al., , p.1248, echoing the aims of the National Curriculum. The researchers note the "naturalistic context" of their study (ibid., p.1248), which means that "although each teacher in the intervention had the same training and the same set of materials, it was neither possible nor ethical to attempt to achieve identical implementation" (ibid., p.1250). ...
... The study found that "embedded teaching of grammar relevant to the writing being studied had an overall beneficial effect on students' achievement in writing" (Myhill et al., 2012(Myhill et al., , p.1252, with the most significant improvement "in composition and effect" (ibid., p.1253). There are, however, a number of caveats to this finding. ...
This paper explores an investigation, based on action research, into the impact of contextualised grammar teaching on a high-attaining Year 9 class’s understanding of grammar. Prompted by previous findings that have shown traditional teaching of grammar as a set of rules to have little benefit for student writing, these lessons sought to present grammar as a tool to create effect in written work. The findings suggest that students are able to make use of grammatical concepts learned through contextualised grammar teaching, and that such teaching does not have any detrimental effect on enjoyment or confidence. This paper argues for the potential of contexualised grammar teaching as an answer to grammar’s contested position in the National Curriculum for some classes, and as a way to engage students more meaningfully in the study of grammar.
... Results suggest moderate to large effect sizes compared with control groups (ESs 0.61 to 0.94). Recently, Wyse and Torgerson (2017) revisited the evidence for effectiveness of grammar teaching, concluding that, at primary level, it is ineffective while at secondary level they discuss in detail how two robust experimental studies (Elley et al., 1976 andMyhill et al., 2012) There is disagreement about the degree or extent of direct instruction that is effective in teaching. However, almost all international researchers of writing whose mixed methods studies we analysed have concluded that some direct instruction is required if teachers are to ensure positive learner outcomes in writing. ...
... Incorporating collaborative discussion arguably interfaces in some instances with studies discussed in the later section on Integration. For example, an innovative means of teaching grammar in context (Myhill et al., 2012) involves a multi-componential approach, linking reading (of model texts), small group discussion (of the effects of grammatical features) and writing (using the discussed feature). Whereas studying grammar out of context seemed purposeless, the use of small group discussion to investigate the effects of writers' choices is emerging as a fruitful endeavour supporting adolescent writing composition. ...
Technical Report
Full-text available
A review for the New Zealand Ministry of Education, as background to the 2022 Literacy and Communications & Mathematics strategy.
... Ze is er tevens op gericht om leerlingen te laten nadenken over de effecten van hun formuleringen. Onderzoeken rapporteren dat grammar for writing positief effect heeft op schrijfvaardigheid van leerlingen (Myhill, Jones, Lines & Watson, 2012). ...
Full-text available
Leerlingen krijgen doorgaans vanaf klas 3 of 4 havo/ vwo jaarlijks modules correct formuleren aangeboden. Ze leren daarbij formuleerfouten (zoals foutieve samentrekkingen en contaminaties) herkennen en verbeteren. In hoeverre sluiten deze modules aan bij de formuleerproblemen van de leerlingen? Tot voor kort kon deze vraag niet goed worden beantwoord, omdat er onvoldoende empirische gegevens waren over formuleerfouten in leerlingenteksten. Twee recente studies hebben nieuwe data aangeleverd. Het eerste deel van dit artikel bevat een overzicht van wat er inmiddels bekend is over formuleerfouten in leerlingenteksten. Het tweede deel van dit artikel presenteert ontwerpprincipes voor nieuw formuleeronderwijs. Nieuwe modules (correct) formuleren zouden zich niet moeten richten op vermeende formuleerproblemen maar op reële formuleerproblemen van leerlingen. Bovendien kunnen ze bijdragen aan de ontwikkeling van bewuste taalvaardigheid.
... al. (2011) found that teachers who emphasised the importance of meaning-making, purpose, the use of linguistic devices, and their effect on the reader, elicited effective writing from students (Fisher et al., 2011). In addition, educators who have applied a Hallidayan perspective (Halliday & Matthiessen, 2014;Halliday, 1978), focusing on the social functions of different text types, as well as their linguistic and textual structures, have emphasised the importance of metalinguistic dialogue between teachers and students during the compositional process (Healy & Gardner, 2021;Myhill, Newman & Watson, 2020;Myhill, 2019;Myhill, 2018;Myhill, Jones & Wilson, 2016;Myhill & Newman, 2016;Myhill, Jones, Lines & Watson, 2012). A solid grounding in teaching grammar in the context of student's own writing might obviate increasing incursions into the classroom of formulaic writing programs, by making compositional processes meaningful events in which the teacher and students co-construct texts around an appropriate selection of linguistic devices, designed to create particular effects for different audiences. ...
Full-text available
For over a decade, there has been growing concern about declining ‘standards’ of school students’ writing (Gardner, 2018; Wyatt-Smith & Jackson, 2020). Teachers’ own writing experiences and writer identities are important considerations in developing teacher preparedness and skill in the teaching of writing. It cannot be assumed that pre-service teachers entering university have the pre-requisite skills and experience to effectively teach writing. This study investigated the pre-entry writing practices of first year Primary and Early Childhood Education (ECE) BEd students at one Australian university. Findings show they most frequently wrote informal, digital texts. It is suggested a lack of experience of writing extended texts, required in the primary English curriculum, may contribute to the decline in school students’ achievement in writing, when assessed against national standard benchmarks (Gardner, 2018; Wyatt-Smith & Jackson, 2020). International studies suggest the teaching of writing has not been addressed well in initial teacher education (ITE) courses (Brindle et al., 2016; Rietdijk, Janssen, van Weijen, van den Bergh & Rijlaarsdam, 2017; Wyatt-Smith & Jackson, 2020). This study supports Wyatt-Smith and Jackson’s (2020) view that greater attention should be given to the teaching of writing in teacher education courses in order to produce the next generation of ‘teacher-writers’ capable of improving the quality of writing in primary schools.
This article is concerned with how developing writers use grammar knowledge to make choices in writing and the extent to which this knowledge is conscious and the choices deliberate. Drawing on case study data with 24 school-aged writers involved in a three-year longitudinal study, the article reports the pattern of conscious and unconscious choices being made. It will reveal how curriculum emphasis has tended to prioritise grammatical use over rhetorical purpose, thus raising grammatical use to conscious awareness. The data also shows how linguistic understanding appears first in text before being articulated as a conscious choice; thus raising a question as to the value of conscious and explicit grammatical understanding. The article will argue for a pedagogy that creates a space to talk about choices in order to lift unconscious writing decisions into conscious and useable awareness.
Full-text available
The article sums up the situation of teaching and using local official languages in society and education.
The abstract nature of grammar makes metalinguistic thinking a challenge for both teachers and students. However, it is suggested writing conferences in which students are encouraged to reflect on grammatical choices and their impact on meaning may be an effective means to develop metalinguistic awareness. This paper draws on cognitive linguistics and mentor texts to investigate what impact a concept-led, dialogic approach to grammar teaching has in the context of student-teacher discussions. By means of writing conferences between a teacher and three Year Five students, the paper explores how students made effective grammatical choices, as a result of metalinguistic dialogue with their teacher. Six concepts, scope, action chains, deixis, attentional windowing, fictive motion, and figure and ground, provided the explicit foci for imagining narrative scenes and appropriate grammatical choices. The findings suggest these concepts may have an enduring effect on students’ ability to make independent and creative choices in their writing.
Full-text available
In group-randomized trials, a frequent practical limitation to adopting rigorous research designs is that only a small number of groups may be available, and therefore, simple randomization cannot be relied upon to balance key group-level prognostic factors across the comparison arms. Constrained randomization is an allocation technique proposed for ensuring balance and can be used together with a permutation test for randomization-based inference. However, several statistical issues have not been thoroughly studied when constrained randomization is considered. Therefore, we used simulations to evaluate key issues including the following: the impact of the choice of the candidate set size and the balance metric used to guide randomization; the choice of adjusted versus unadjusted analysis; and the use of model-based versus randomization-based tests. We conducted a simulation study to compare the type I error and power of the F-test and the permutation test in the presence of group-level potential confounders. Our results indicate that the adjusted F-test and the permutation test perform similarly and slightly better for constrained randomization relative to simple randomization in terms of power, and the candidate set size does not substantially affect their power. Under constrained randomization, however, the unadjusted F-test is conservative, while the unadjusted permutation test carries the desired type I error rate as long as the candidate set size is not too small; the unadjusted permutation test is consistently more powerful than the unadjusted F-test and gains power as candidate set size changes. Finally, we caution against the inappropriate specification of permutation distribution under constrained randomization. An ongoing group-randomized trial is used as an illustrative example for the constrained randomization design. Copyright © 2015 John Wiley & Sons, Ltd.
Full-text available
Alamargot, D. & Chanquoy, L., (2001). Through the models of writing. Dordrecht-Boston-London : Kluwer Academic Publishers. Denis Alamargot and Lucile Chanquoy’s book offers a vivid and original presenta- tion of main trends in the research field devoted to writing. First, it provides both young and senior scientists with a comparative view of current theoretical models of composition, with different levels of reading made available: each element of these models is clearly situated in its historical context, and scrutinized in its further evo- lution. Second, this well documented theoretical analysis of writing mechanisms is checked against empirical data extracted from a lot of updated experimental studies; and lack of necessary data is thought to be underlined and defined when noted. Following the usual description of writing phases initially proposed by Hayes and Flowers, the first part of this book presents planning, translating and revision processes and compares them to other researchers’ conceptions (from Bereiter and Scardamalia, to Kellogg or Galbraith). Such presentations of isolated models do ex- ist in literature; but the present work really gives a good comparative analysis of components inside each of models, in a clear and cumulative way; a fine-grained ob- servation of differences between similarly-looking models is also performed.
In this chapter we provide a thumbnail sketch of the key English resources relevant to an understanding of academic discourse. In Chapter 2 we introduced the basic SFL concept of stratification and the way in which SFL sees context as a more abstract level of meaning realised through language. This conception is outlined in very general terms in Figure 3.1, which positions genre as made of meanings and thus ‘realised’ through language.
The mechanic should sit down among levers, screws, wedges, wheels etc. like a poet among the letters of the alphabet, considering them as the exhibition of his thoughts, in which a new arrangement transmits a new Idea to the world. (Robert Fulton - nineteenth-century engineer). A synergy between the work of a mechanic and the work of a writer is not the most obvious one, perhaps, yet Robert Fulton’s analogy (Barlex and Givens 1995: 48) between the mechanic and the poet is an apt one. Both have to create products from the materials available, be that physical materials or linguistic resources; both have to test things out to see how they work; both have to make choices and decisions about the purpose of their work; and both have to evaluate their work critically before presenting it to the world as a new creation. At the heart of this creative activity is the concept of design. This chapter sets out to illustrate how, within a theoretical framework that conceives of writers as designers, linguistic resources and linguistic understanding play a crucial role in supporting the development of design capability in writing. Writers as designers. I should like to lay the foundation for this chapter by elaborating the framework of writers as designers before narrowing the focus more specifically to the role of grammar and linguistics in a pedagogy of writing.