System 47 (2014) 12e27

Contents lists available at ScienceDirect

System
journal homepage: www.elsevier.com/locate/system

Revisiting narrative frames: An instrument for investigating

language teaching and learning
Gary Barkhuizen*
School of Cultures, Languages and Linguistics, University of Auckland, Private Bag 92019, Auckland 1142, New Zealand

a r t i c l e i n f o a b s t r a c t

Article history: This article revisits the use of narrative frames as a research instrument first proposed in a
Received 20 May 2014 2008 System article. It starts by describing what narrative frames are and considering the
Received in revised form 20 July 2014 significance of narrative in research on language teaching and learning. It then re-
Accepted 8 September 2014
examines the strengths and limitations of narrative frames discussed in the original
Available online
article. Following a summary of recent narrative frames research, an example of a frame
used in an ongoing research project is presented and its use illustrated. Finally, to reinforce
Keywords:
the basic narrative frames tenets, ten essential features of narrative frames to be conscious
Narrative frames
Research methodology
of when designing and using frames are noted. Finally, suggestions for ways in which
Data collection narrative frames can be used in future research are made.
Language teaching © 2014 Elsevier Ltd. All rights reserved.
Language learning

1. Introduction

A colleague and I introduced narrative frames as a research instrument in a 2008 System article (Barkhuizen & Wette,
2008). In the article we described what narrative frames are, explained their purpose, discussed their strengths and limi-
tations as a research tool, and demonstrated their use with reference to data collected during a series of professional
development workshops for college English teachers in China. Using the same data set generated by the frames we later
analyzed the instructional practices of the teachers (Wette & Barkhuizen, 2009) and I investigated further their experiences of
research engagement in their working contexts (Barkhuizen, 2009). Since then I have continued to use narrative frames in my
research (e.g., Barkhuizen, 2011), and they have also been used by other researchers in a variety of ways to explore a range of
different topics. This is encouraging, since we ended our 2008 article by saying: “We hope that this article will engender
further exploration of the construction and use of narrative frames” (p. 384).
Without exception, these researchers claim in their research reports that narrative frames have been useful, serving their
research purposes by gathering easily analyzable data and generating relevant insights into the topic of their investigation.
Furthermore, some of the researchers have made advances in both the design and use of frames; combining them effectively
with other methods, for example. However, in reading the reports it appears to me that not all have been faithful to the spirit
(i.e., reflecting the concept of narrative) and practical utilization (i.e., design and application) of narrative frames. Based on
these publications and on my own further narrative frames research experience, in this article I revisit some of the original
tenets of narrative frames put forth in the 2008 article. What I have to say is not meant as a criticism of recent narrative frames
research, but rather a getting back on track of some of the conceptualizations and practices associated with narrative frames. I

* Tel.: þ64 9 9238197.

E-mail address: g.barkhuizen@auckland.ac.nz.

http://dx.doi.org/10.1016/j.system.2014.09.014
0346-251X/© 2014 Elsevier Ltd. All rights reserved.
 G. Barkhuizen / System 47 (2014) 12e27 13

believe doing so is timely, especially if narrative frames are going to be employed even more widely in research on language
teaching and learning, as it appears they might very well do.
I start by describing what narrative frames are and considering the significance of narrative in research on language
teaching and learning. I then re-examine the strengths and limitations of narrative frames discussed in the original article.
Following a summary of recent narrative frames research, I present an example of a narrative frame I have used in an ongoing
research project and illustrate how data generated by the frame might be analyzed. Finally, to reinforce the basic narrative
frames tenets, I note ten essential features of narrative frames to be conscious of when designing and using frames. I also
suggest some ways in which narrative frames can be used in future research.

2. What are narrative frames?

A narrative frame is “a written story template consisting of a series of incomplete sentences and blank spaces of varying
lengths. It is structured as a story in skeletal form. The aim is for participants to produce a coherent story by filling in the
spaces according to their own experiences and their reflections on these” (Barkhuizen, 2011, p. 402). Stories represent the
meanings storytellers make of their experiences. In telling stories people make sense of the events in the lives they have lived
or they imagine living. In the re-telling they shape and re-shape those experiences. Narrative inquiry as an approach to
research aims to understand these experiences from the perspective of those who experience them. Barkhuizen and Wette's
(2008) article provides a brief overview of the rationale for narrative inquiry in the field of teacher education, and more recent
work has located narrative more widely in both language teaching and language learning research (e.g., Barkhuizen, 2011,
2013; Barkhuizen, Benson, & Chik, 2014).
What we see from this literature is that narrative research varies considerably in its methodological approaches to investi-
gating language teaching and learning; from broad thematic analyses of extensive written data (e.g. teacher journals) to detailed
conversation analyses of short excerpts of interview data (e.g. small story analysis; see Norton & Early, 2011). Approaches to
analysis also vary according to the attention paid to the form of narratives, their content, and both the local and wider socio-political
context of their construction. Narrative frames elicit content that relates to the researcher's research question. The responses from
one respondent may not generate a lot of content, but a well-designed frame will ensure that the content is at least relevant, and
when combined with the responses of other respondents will produce substantially more content. The structure of the frame
enables respondents to construct their telling in narrative form, with the meaning making that allows, and the outcome offers the
researcher a coherent snapshot of a written reflection of experience in narrative form. In sum, from the respondents' perspective,
the frames “provide guidance and support in terms of both the structure and content of what is to be written. From the researcher's
perspective the frames ensure that the content will be more or less what is expected (and required to address the research aims)
and that it will be delivered in narrative form” (Barkhuizen & Wette, 2008, p. 376). As far as possible, the design of the frame should
embed contextual cues to enrich the spatiotemporal dimensions of the narrative (see below), and in addition it is up to the
researcher to consider macro socio-political contexts during the process of analysis (Pavlenko, 2007).
The strengths and limitations of narrative frames were summarized in Barkhuizen and Wette (2008, pp. 381e383), but it is
worth repeating and elaborating on them here. Readers at this stage might like to refer to Appendix A to see an example of a
frame. Firstly, the strengths:

1. The narrative frames enable respondents to write narratively by scaffolding them through the specially designed
narrative structure. For people who are not familiar or comfortable with writing about their personal experiences
(what form? what style? what level of intimacy?) the frame ‘walks them through’ the self-reflective writing. Frames
require reflection on experience, and their design should enable this.
2. The design of the frames ensures (or it should) that researchers obtain the information they want in order to achieve
the aims of their study. However, there is still some flexibility in what the respondents actually write. The flexibility is
guaranteed by the blank spaces, but it is the researcher who is responsible for making sure that the frame's design
generates content (experience and reflection on that experience) that is directly relevant to the topic of the research.
3. The spatial and structural restrictions do allow the narrators to tell their stories of experience; not long, detailed stories,
but nevertheless storied snapshots of the respondents' life which, together with stories of other participants in similar
contexts, is informative and relevant for the researcher.
4. The structured nature of the data makes for easier analysis. Typically, a thematic analysis (Barkhuizen et al., 2014) is carried
out on the frames, and much of the categorizing of the data into themes is already achieved because of the structure of the
frame; i.e., certain themes will be guaranteed because the prompt in the frame will elicit them as responses.
5. The frames limit the quantity of data that is collected; the sample of respondents can be relatively large, therefore. The
researcher can determine upfront how much data will be collected through the way the frame is designed (e.g., the
number of responses required and their length) and also by the number of participants it is distributed to.
6. Frames serve an exploratory purpose. They are useful for entry into a new or unfamiliar research context (regional,
social, cultural, educational, etc.) where they can provide an introduction to that context. As a preliminary instrument for
data collection they become especially useful when combined (possibly later) with other methods, such as interviews
and field observations. This is perhaps the main strength of narrative frames. We never claimed in our original article
(Barkhuizen & Wette, 2008) that narrative frames should be the only data collection instrument used for a project, or
14 G. Barkhuizen / System 47 (2014) 12e27

that they alone would generate data which would conclusively answer any particular research question. Narrative
frames work well when they are exploratory, preliminary or integrated with other compatible data-collection methods.

Narrative frames also have their limitations, some of which include the following:

1. Martin Bygate, in his commentary on Nguyen Gia Viet's narrative frames case study (Nguyen & Bygate, 2012), asks: “What
question or puzzle is the procedure intended to illuminate?” (p. 59). This is an important question. If neither the
researcher nor the respondent is absolutely clear about what is being elicited, the frame will in all likelihood produce data
that are inappropriate for answering the research question or puzzle. A potential limitation exists therefore when the
design of the frame does not explicitly make clear what written response the frame is requiring from the respondents.
2. The process of writing the frames may frustrate some respondents who may need or desire more space to compose
their responses. They want to say more, and have more to say, but there is no space to do so. Some narrative frames
respondents have actually told me that they have experienced this, and others have tried to squeeze in more text by
writing in the margins or on the back of the page.
3. Researchers may also feel that at times they would like more content than that captured in the spaces available between
prompts. They would, of course, be aware beforehand that they were going to receive limited data because of the con-
straints of the frame structure, but they may nevertheless feel short-changed if it is evident that there is potential for more.
4. The content structure of the frames (i.e. the narrative sequence of prompts) may not be compatible with the way a
particular writer would like or be able to structure his or her own story. There may also be incompatibilities between
the way the frame is structured and the cultural writing conventions of respondents.
5. There is no opportunity for respondents to write about other (non-framed) topics, unless such opportunities are built
into the design of the frame (see Appendix A, to be discussed later, for example).
6. There may be prompts which are interpreted in a way not intended by researchers. This is typically a design flaw, and
should be prevented during a frame trialling process.
7. The potential exists for respondents to write a list of unconnected ideas in response to the prompts if the frames are
poorly designed or the instructions are not clear. Again, piloting the frame with a different group of respondents or a
trial run with the target respondents using a similarly constructed practice frame should prevent this.
8. Depending on how the frames are used, they have, ironically, the potential to de-personalize the respondents' stories of
experience, especially when combined and analyzed together with many other completed frames. The individuals and
their stories get lost in the amalgamation of data and their reduction into themes and categories, and possibly even
numbers (e.g., frequency counts and statistics, if a quantitative analysis is done). Using the frame in conjunction with
one or a series of personal interviews, for example, would help to avoid this limitation. The respondent's responses
could be explored further during the interview(s).

3. Summary of narratives frames research

In this section I present in the form of a table (see Table 1) a summary of recent research that has employed narrative
frames, either as the sole data-collection instrument or in combination with other methods. I briefly describe the purpose of
the research and of the frame. I also comment on the design of the frame and offer a brief critique where relevant. In sum,
some of the design weaknesses include: too many topics in any one frame, meaning that it is somewhat unfocused; the topic
of the frame not being clear; the frame not producing a narrative (no spatiotemporal dimensions) but rather a list of facts or
opinions; the format being a list of prompts one under the other rather than in prose paragraph form; and, invitations for
deletions or insertions potentially disrupting the narrative coherence. On the other hand, some positive developments
included: the use of frames in conjunction with other methods such as interviews and field observations; frames used as
pedagogical tools; frames used contingently so that subsequent frames are adjusted to meet specific unforeseen needs;
helpful cues embedded within frames without distracting from the prompts; the use of qualitative data analysis software;
and, the use of more than one language in the design of the frame as well as to complete the frame.

4. An illustrative case: the home tutor project

In this section I present an example of a narrative frame (see Appendix A) that was used as part of a larger study which
investigated the problems and successes of adult migrant and refugee English learners in New Zealand. All the learners were
enrolled in services provided by English Language Partners New Zealand (http://englishlanguage.org.nz/) an organization
which operates in 23 independent centers across the country. Its mission is to provide an accessible, nationwide, community-
based service that supports settlement through English for adult refugees and migrants. In addition to social and literacy
classes, it offers one-on-one tutoring for one hour per week in the homes of refugee and migrant English learners. Forty-one
learners from 20 different countries completed the narrative frame. Participants included 30 females and 11 males, who lived
in 13 different regions in New Zealand. Learners starting the home tutoring program vary in their levels of English proficiency,
from complete beginner to quite advanced; the latter engaging in the program because, for instance, they desire to become
more familiar with the Kiwi accent or to gain confidence engaging with New Zealand society outside the home environment.
All participants were adults, ranging in age from the mid-twenties to over seventy years old.
Table 1
Summary of studies using narratives frames.

Study Research aim Topic of frame(s) Design of frame(s) Design critique Other methods Method of analysis Author evaluation
used
Barkhuizen (2009) To investigate college Identifying a classroom Seven sentence starters, Frame is perhaps too None. A thematic analysis of the “The frame provides
English teacher research problem and a related in English, followed by short to cover so many written responses and a teachers with a coherent
engagement and practice research project to blank spaces of varying different, but related, quantitative frequency structure within which
in China. address the problem. lengths formatted as a aspects of the topic. calculation of the themes. they can concentrate on
prose paragraph. communicating their
experiences and
reflections” (p. 116).
Wette and To investigate college Three frames focused on: Each with seven or eight Instructions to such a None. Responses coded for They “serve two very
Barkhuizen teaching of English for second language teaching sentence starters, in large group of themes using N-Vivo important purposes: (a)
(2009) academic purposes in background, language English, followed by respondents might not (qualitative data analysis they scaffold writers who
China (see Barkhuizen, curriculum and materials blank spaces of varying have always been software), first by one may be unfamiliar with
2009) development, and lengths formatted as a effective e they were researcher and then writing reflective,
assessment in the prose paragraph. mainly delivered second- checked by the other, especially narrative,
language curriculum. hand, not by the followed by a accounts of their
researchers. quantitative calculation experiences, and (b) they
of the themes. make it possible to collect

G. Barkhuizen / System 47 (2014) 12e27

data from a large number
of participants” (p. 201).
Barnard and To explore what high Three frames explored First two very short; four Perhaps not completely Completion of Frames subjected to “The information thus
Nguyen (2010) school teachers workshop participants' and two sentence starters evident what the frames frames followed by constant comparison collected served as a
participating in personal approaches to respectively. Third frame were requiring from audio-recorded method typical of useful platform to
professional language teaching and with seven starters listed respondents. Not group lesson thematic analysis triangulate data gathered
development workshops their attitudes to TBLT. one under each other. All structured as narrative planning, including coding and [from other methods.]
in Vietnam know and starters in Vietnamese. (e.g. no spatiotemporal observation of categorizing of themes. The implication is that
believe about task-based dimension). Not planned lessons with small numbers of
language teaching (TBLT). formatted as prose. and stimulated participants, narrative
recall/group frames can more usefully
discussion sessions. complement other data
than serve as the main, or
only, source of
information” (p. 83).
Swenson and To understand the A broad range of topics, Free-writing introductory Inserted cues very useful. None. Through a thematic “Narrative frames
Visgatis (2011) experiences of four temporally sequenced, section with prompts, Not always clear, analysis, 12 themes provide more in-depth
students from Japan ranging from pre- followed by a lengthy however, what some of identified and coded. information than surveys,
participating in an departure expectations, frames template in the other starters Frequency counts carried focus attention on the
overseas study abroad experiences during the English with sentence required of respondents. out and reported in points of concern to
internship so as to work internship and starters formatted as a Overall, frame had a good tables. researchers … are
determine what changes homestay, and reflections prose paragraph, with temporal flow, with relatively easy to
could be made to the pre- on return to Japan. cues in parentheses after opportunities for complete, and give a
departure program. some starters. evaluation. fuller picture of the
experience” (p. 448)
Macalister (2012) To explore the use of The story of learning and Nine sentence starters, A temporal dimension, None, as purpose Qualitative content “Narratives frames were
narrative frames as a using English in different each on a separate line but story ends abruptly. was to examine the (thematic) analysis, firstly found to be a valuable
needs analysis tool in the contexts reporting both followed by several blank Clearly and appropriately effectiveness of by two independent tool in the needs analysis
design of an EFL challenges and successes. lines of equal length elicits very personal using the frame as a readers and then by both for the new EFL
curriculum for trainee spread over two pages, in responses. needs analysis collaboratively. curriculum and are a tool
seamen in Kiribati. English. instrument. that could be used more
(continued on next page)

15
Table 1 (continued )

16
Study Research aim Topic of frame(s) Design of frame(s) Design critique Other methods Method of analysis Author evaluation
used
widely in curriculum
design” (p. 120).
Shelley, Murphy & To examine the Covers four topics, Introduction with Instructions invite Follow-up Frames data analyzed “Participants reported
White (2013) development of language including experience as a instructions. Four omission of irrelevant interviews with six using NVivo that they found the open
teacher cognitions as they language teacher, sections of between starters, which might of the ten independently and then structure of the narrative
transition from classroom transition into blended seven and eight sentence disrupt coherence of participants. collaboratively, and later frame … accessible,
to distance, online and learning, problem solving starters, each with usefulstory, as might invitation verified with participants thought-provoking and
blended teaching. and a specific teaching heading. to insert other ‘issues’. during interviews. professionally
activity. Not formatted as prose. rewarding” (p. 571).
Not all starters flow into
the next logically.
Hiratsuka (2014) To investigate how Frames related to three First frame included Frames certainly produce After completion of Thematic analysis to “The narrative frame
narrative frames affected video-clips of three cycles sixteen numbered a coherent piece of the frames, search for commonalities technique proved to be a
the learning of high of teaching respectively, sentence starters, reflective writing, but interviews with and differences. Themes beneficial tool for
school students in team- were observed by formatted in paragraph questionably narrative e selected students organized into two main improving learner
teaching contexts in students, who format. Cycle 2 and 3 lacks temporal and the teachers. categories. autonomy. It facilitated
Japan, particularly their commented on the frames shortened to dimension. Published the students becoming

G. Barkhuizen / System 47 (2014) 12e27

development of learner observed class, their thirteen numbered frames not in prose responsible for their
autonomy. teachers, themselves, starters. All frames format, but originals were learning” (p. 169).
reasons for comments, completed in both English (personal
and suggestions for and Japanese. communication, April,
future classes. 2014).
Xu (2014) To investigate Chinese Five broad topics One frame divided into Useful cues given for 104 frames Qualitative content “Narrative frames and
University English covered: demographic three sections, including some prompts. Coherent received. Follow-up analysis conducted on the individual in-depth
teachers' research information; institutional sentence starters, mid- throughout, though more interviews with frames. interviews were used in
practice and the policies; research sentence chunks and in the form of personal four teachers, in this study. They were not
construction of their engagement and activity; ends of sentences. statements and opinions which they were used as separate parts but
identities as researchers. research motivation; and Formatted as three than a narrative. asked to comment as complements to one
challenges and solutions. paragraphs. Originals in on their frame another” (p. 245).
Chinese; published form responses plus
in English. further questions.
 G. Barkhuizen / System 47 (2014) 12e27 17

For illustrative purposes in this article I focus on the short section of the frame that elicits responses related to the partic-
ipants' imagined future successes in learning and using English. Refugees and migrants who come to New Zealand imagine the
lives they will live when they get there, and continue to imagine the future after they have arrived. They will imagine problems
and, especially for those who undertake to learn English, they will imagine successes. Although there is no space to do so in this
article, such ideas are relevant to the work on psychological possible selves (Markus & Nurius, 1986), Ideal L2 selves (Do €rnyei,
2009) and on learners who live their lives in socially-oriented imagined communities. Kanno and Norton (2003) point out that
“the notion of imagined communities provides a theoretical framework for the exploration of creativity, hope, and desire in
identity construction” (p. 248). In other words, imagining themselves as ‘ideal’, legitimate members of future-oriented com-
munities has an impact on learners' participation in current identity construction and engagement in language learning.
Some narrative frames were completed by the learners themselves, while others were actually written by the tutors, who
worked closely with their learners to elicit their story and to fill in the narrative frame. This was usually incorporated into the
weekly one-hour lesson as a learning activity. Following all requirements approved by my university's ethics committee, the
learners were approached by their own tutors and asked if they were willing to take part in the study.
The frame (see Appendix A) was structured as follows:

1. Part 1 asked, in the form of optional question prompts, for some free writing to provide background information about
the learners; e.g. where they live in New Zealand, which country they come from, what language(s) they speak, and
who the members of their families are. One rationale for this free-writing section was to allow respondents some
writing (and reflection) space outside the confines of the frame itself. It also encouraged some ‘brainstorming’ before
starting to write the frame (see Swenson & Visgatis, 2011).
2. Part 2 was divided into three parts, all of which consisted of between 4 and 6 sentence-starter prompts followed by a
space for written responses to complete the sentence: (a) the first part asked learners to identify problems they
experienced with English in the past and some early successes; (b) the second part prompted responses to do with
progress they were currently making in the present; and (c) the third part asked respondents to imagine success with
using and learning English in the future. The sentence starters and their following blank writing spaces were not or-
dered in list form, one after the other, like one might find on a questionnaire. Instead, they were formatted very much
like a paragraph.
3. Part 3 offered the learners' the chance (if they wished) to add anything else, in free writing, about their successes in
learning and using English in New Zealand. Again, this space gave respondents the opportunity to add any further ex-
periences that might have been prompted by the frame but were unable to within the frame itself because of lack of space.

For purposes of illustration I include here one complete narrative frame (see Extract 1) e the story of Irena (a pseudonym)
from Bulgaria. The text in italics is the sentence starter prompts and the underlined text is the content which is the focus of
this discussion.
Extract 1: Irena, from Bulgaria
I come from Bulgaria, I speak Bulgarian and some Russian. I came to NZ to marry my husband, Jack. He is also Bulgarian
but he has lived in NZ 15 years. He speaks very good English. In Bulgaria, I was a chemical engineer. I can't work in NZ
until I learn more English. I've lived in NZ for one year and I live in Mount Maunganui. We lived in Auckland for six
months then moved here for lifestyle and for Pete's work. My adult daughter and family still live in Bulgaria and I miss
them very much. We talk on the telephone.
When I first came to NZ my English was only hello, goodbye and thank you. Soon after I arrived, a problem I experienced
with English was I felt very stressed because I didn't understand people. I couldn't read signs and shopping was hard. I
felt helpless. Another problem I had with my English was I felt very lonely. I could only talk to my husband. I felt like a
child, I got stressed by mail and phone calls. After a while, however, I started to learn some English and I met my very
nice neighbour. Now I know a little when people talk to me. I don't like it when people are impatient and say “it doesn't
matter”. I realized things were getting better when I was able to go shopping and talk to people. I worked [volunteer] at
the store, I could hear people talking and sometimes I knew some words. I could remember phrases (e.g. how's it
going?). In those early days, another example of success with my English was I answered the phone when my home tutor
called me. I can ask for things in a shop. I understood my teacher. I could tell her what I wanted to learn.
Now, my English is growing fast. I can talk to people and tell them what I want. I can ask people to talk slower or say
something again. I find I can more easily talk to people at my work. I can chat with people I know well. What I can really
do well when I use English is talk to my neighbour. I can ask for help and help people. Also, I remember once I was able to
bargain at the flea market. However, when I use English I'm still having problems with using the phone because I can't see
the person's face and body language. I would like to be able to solve this problem by practicing with my home tutor and
husband. We use a list of telephone expressions. I keep this by my phone. I practice calling businesses to ask questions.
In the future, I would like to use English to WORK! I have worked for 21 years and I miss work. I feel helpless because I
have no money. I want to help other people learn English. I imagine that I will also be able to learn to drive. I will know
that I have learned enough English when I understand people talking and I can read books. I miss reading books. Learning
18 G. Barkhuizen / System 47 (2014) 12e27

English is important for me because all people in NZ speak English. I must learn English because I live here. I want to work
and have friends I want to know what people say. That's the end of my story.
The responses were analyzed in the following way: Each learner's written responses to the three relevant starters (imagined
successes in using and learning English) were extracted from the full story. There were 41 separate extracts, each of which was
entered into a row on a spreadsheet, with the sentence starter responses in sequence in separate columns. Commonalities among
learners were then sought in the data. To do this, responses to each sentence starter from each of the individual stories were in turn
examined (i.e. all 41 responses to the first starter, In the future I would like to use English to …). The data were analyzed by following
the procedures typical of qualitative thematic analysis (Barkhuizen et al., 2014; Hayes, 2013; Riessman, 2008): that is, themes were
identified and coded, and interpretations of patterns among themes were made. As with any qualitative coding procedures, codes
were consistently applied to themes; e.g. any mention of family intervention in learning and use were coded as Family involvement.
Definitions of themes (see below) guided this coding. Other researchers, again typical of thematic analyses, may have identified
different themes and defined these in their own way. The analysis was made easier because all responses to the same starter
appeared in the same spreadsheet column. Frequencies of the various themes were also calculated (see Table 2). The frequencies
represent the number of times a particular theme was mentioned by the learners. Sometimes more than one theme was evident in
any single response to a particular starter. Numbers do not, therefore, add up to 41 (i.e. one per learner).

4.1. Qualitative findings: descriptions of successes

The analysis of the responses to the success starters revealed 15 main themes. These themes are organized further into the
following categories: (1) linguistic, (2) instrumental, (3) interactional, (4) support, and (5) affect. The themes, with illustrative
narrative frame responses, are as follows:

4.1.1. Linguistic
Speaking:

a. … talk to everyone more easily.

b. … speak English like Kiwis.
c. … speak to everyone as it is an international language.

Responses here have to do with speaking fluently or easily, or achieving successful oral communication with Kiwis. The
ideal self of the learner who supplied response (c) above would be someone who would be able to speak English to everyone
in his/her imagined New Zealand community.
Understanding e language related:

a. … understand people talking.

b. … to understand more English words.

Table 2
Ranked imagined successes. Note: Superscripts indicate the top five ranked successes themes in the totals column.

Theme In the future I would I imagine that I will I will know that I have learned Total
like to use English to … also be able to … enough English when I …
Linguistic 20 16 22 58
Speaking 12 11 14 37(21%)1
Reading and writing 4 2 3 9
Understanding e generally 2 2 3 7
Understanding e language related 2 1 2 5

Interactional 22 12 12 46
Daily living 13 7 6 26(15%)3
Understanding e daily living 1 1 1 3
Social interaction 6 4 5 15(9%)4
Using the phone 2 2

Instrumental 21 17 6 44
Employment 15 10 4 29(17%)2
Further education 3 4 2 9
Teaching English 3 3 6

Support 5 4 7 16
Family involvement 5 2 3 10
English lessons 2 1 3
No need for support 3 3

Affect 1 2 8 11(6%)5

TOTAL 69 51 55 175
 G. Barkhuizen / System 47 (2014) 12e27 19

These responses indicate that learners imagine themselves understanding their interlocutor's turn at talk, or under-
standing written text, with specific reference to a language system or skill (e.g. spoken or written language, vocabulary,
accent).
Understanding e generally:

a. … understand more easily.

b. … understand a lot of English.
c. … to understand strangers.

These responses did not specify any particular feature of their understanding, merely indicating that they imagine un-
derstanding an English-speaking interlocutor's turn at talk or some written text.
Reading and writing:

a. … read newspaper, use English internet more.

b. … writing letters.
c. … I can read books. I miss reading books (novels).

Ideal learner selves are able to read texts such as library books, newspapers and bus timetables, and to write in English.

4.1.2. Instrumental
Employment:

a. … work in the hospitality industry.

b. … get a better job.
c. … WORK!

The employment theme refers to learners securing a job in their imagined community, or changing jobs, or being fulfilled
in the workplace. Their ideal L2 selves would thus be productive, contributing members of the community.
Further education:

a. … advance my studies.
b. … join classes for painting.
c. … achieve success in academic study.

Besides learning English, learners desire to continue to higher education, or train for a specific job-related qualification.
Often further study is unavailable to learners because they lack the required English proficiency.
Teaching English:

a. … help beginners of English.

b. … train as a home tutor some time in the future.
c. … teach others in my Sri Lankan community to speak English.

Some learners see themselves doing so well with English learning that they plan in their imagined future life to teach
English within their community or in their home countries, if they return.

4.1.3. Interactional
Daily living:

a. … go to the supermarket on my own.

b. … carry out my everyday living.

In their imagined world learners would carry out their regular daily activities successfully in English.
Understanding e daily living:

a. … to understand the name, the price, and which size, which country the product come from.
b. … to understand the boys' school teachers.

This theme refers to understanding an interlocutor's turn at talk or written text but it includes a reference to accom-
plishing some aspect of the learners' everyday living activities.
Social interaction:

a. … make friends with local Kiwi.

b. … can fit in NZ life.
20 G. Barkhuizen / System 47 (2014) 12e27

Interacting successfully with English speakers and Kiwis, and making new friends, is a clear sign of successful community
membership. In this theme, the learners emphasize the social dimension of their imagined ideal selves.
Using the phone:

a. … being able to answer the telephone.

b. … make myself understand on the phone.

This theme concerns success at answering the phone or engaging in a conversation on the phone. Telephonic communication is
obviously stressful for learners, and being able to conquer their fear in an imagined world appears to be particularly desirable.

4.1.4. Support
Family involvement:

a. … to be able to talk with and understand the boys' school teachers.

b. … help my son with his school work.
c. … speak to my grandchildren who live in Germany.

Family involvement means interacting with family members in English (e.g. children who are learning English at school or
relatives living overseas), or helping family members with their English. In their imagined communities, English will be a part
of family life.
English lessons:

a. … study English in university.

b. … I don't think I'll ever stop learning.

Some learners expressed the desire to continue learning English, indicating that some kind of formal English learning will
be taken up, such as community English classes and literacy workshops.
No need for support:

a. … can do everything by myself.

b. … don't need my English tutor anymore.
c. … have no problems.

A measure of success is when learners feel they know enough English to no longer need further formal teaching or support
from proficient English speakers. The English learning mission is well-advanced; the learners have become successful,
legitimate members of their imagined English-speaking community.

4.1.5. Affect
a. … stop worrying about my English and feel confident in most situations.
b. … feel confident using it every day everywhere.
c. … no longer feel frightened to speak to New Zealand people.

Emotional responses accompanying the reported imagined successes of the learners are evident throughout the narrative
frame data. Particularly desired are ideal selves who are confident about using English.

Quantitative findings: Ranking of successes

Table 2 shows the results of the frequency count of responses to the three featured starters. It is not surprising that
Speaking is ranked the highest (21%). It is through casual English conversations outside the home that migrants and refugees
make most contact with other New Zealanders, in shops, on the street, at schools and in the bank, thus providing them with
the opportunity to interact and integrate with those outside their own (ethnic, national, linguistic) communities. Re-
spondents thus imagine their future ideal selves in a world where they are able to converse easily in English with members of
the broader English-speaking society.
Employment is ranked second highest (17%). This too is not surprising. It was mentioned 29 times. It is often a lack of
adequate English proficiency which prevents migrants and refugees from finding a job, especially in an occupation in which
they may already have considerable experience in their home countries. Imagining a future where they will speak English at a
level that will enable them to find employment in a job they desire is quite understandably a priority for the learners. The
third ranked response, Daily living, is also highly ranked (3rd, at 15%). As the learners expressed throughout the frames data,
living a life in a community by simply getting on with everyday activities and not feeling anxious about one's English pro-
ficiency in order to do so must be a wonderful image - clearly a trait of an ideal L2 self. When Daily living is combined with
Understanding e daily living (3 mentions) and Using the phone (2 mentions, and certainly a daily living activity) the total for
 G. Barkhuizen / System 47 (2014) 12e27 21

this theme is 31, which would make it ranked second, indicating the high priority given to speaking in the process of fulfilling
daily living tasks in the participants' imagined future worlds. A high level of proficiency would obviously mean much less
communicative effort on the part of the learners, as well as a better chance of getting tasks done successfully. Imagining
oneself in such a situation is attractive to English learners. Social interaction ranks 4th with 15 mentions (9%). Legitimate
community membership means interacting appropriately and effectively with other members of that community. In this case,
these members would be other English speakers living in New Zealand. A successful self would interact with them using
English efficiently and resourcefully, and would not feel afraid or anxious when doing so (Affect, ranked 5th, at 6%).
In this section I have merely shown, for the purposes of illustration, how an analysis of the frame could be conducted. A full
report of the study would, of course, include a more in-depth interpretation and discussion of the findings, particularly in
relation to relevant literature. In the larger study of which this frame was a part, narrative interviews were conducted with the
tutors of the learners. The focus of these interviews was twofold: to further explore the learners' learning and use of English,
from the tutors' perspective, and to investigate the tutors' instructional practices and their own imagined successes as tutors.
The frames and interviews combined provided a rich data source for understanding the experiences of the tutors and the
learners and their relationship (see Barkhuizen, 2011).

5. Ten essential features of narrative frames

Based on my experience of using narrative frames since publication of the 2008 article and on my review of other research
which has utilized frames, I now present ten features which I believe are essential to the effective design and administration
of narrative frames. I have divided the features into seven design features and three administration features. To reinforce each
of these, I draw on the home tutor study described above.

5.1. Design features

1. Purpose: The researcher must be clear about the purpose of the narrative frame. What is it trying to achieve? What data
does the researcher need for the project and how will these data integrate with any other data collected by other means?
What research puzzle is the frame trying to solve (see Martin Bygate's comment in the limitations section above)? Does
it have value beyond data collection (e.g. some pedagogical value)? It also helps if the respondents know the purpose of
the frame. Not only will they feel more at ease when completing it but they will be able to focus their responses more
directly on what the researcher is looking for. In the home tutor project, the tutors informed their learners about the
project. They therefore had the responsibility of explaining the purpose of the research to their learners. I had invited
them to do so during my interview with them, which took place prior to their working on the frame with their learner.
2. Topic: The tutor also explained the specific topic of the frame to their learner. The topic of a narrative frame needs to be well-
defined. This should be evident from a title or clear written instructions. Again, designing the frame around a topic (or
possibly two or more but related topics in a longer frame) is important for both researchers and respondents. If the topic is
vague or there are too many (unrelated) topics, the frame will potentially generate data that is not appropriate for the
research project. It will also be more difficult for the respondent to produce a coherent narrative and consequently for the
researcher to analyze it.
3. Experience: The topic of a frame is always about the experiences of the respondents; either those lived in the past or
those imagined in the future. It is about their lives e the stories of their lives that relate to the research puzzle and more
specifically to the topic of the frame. In other words, the frames capture details of how respondents experiences their
lives, when, where, how and why. For the home tutor learners, their experiences were their problems and successes of
learning and using English in their lives as migrants and refugees in New Zealand. The frame asked them to give actual
examples of problems and successes at different times of their immigrant lives, including their imagined futures.
4. Reflection: And it is also about how they make sense of these experiences. The frames should provide opportunities for
respondents to reflect on the experiences described in the frames, what Labov (1997) calls evaluation. These evaluations
indicate why something is important to the respondent and usually take the form of emotional responses to or beliefs
about what happens. In short, the frames produce more than just a list of facts about what happened or will happen,
they also require respondents to comment on those experiences. The home tutor frame (Appendix A), for example, asks
respondents to comment on “What I really do well …” and why “Learning English is important for me …”.
5. Spatiotemporal: To ground these experiences in particular contexts narrative frames need to have both a spatial and a
temporal dimension. That is, the design of the frames should generate stories which are located in particular places (e.g.
classrooms, schools, teacher education courses) and at particular times. The home tutor frame is designed chrono-
logically in three broad sections, In the past, Now and In the future, and then more specifically by means of phrases such
as “Soon after I arrived …”, “In those early days …” and “In the future …”. In terms of space the frame clearly locates the
respondents in New Zealand, firstly, by contextualizing the narrative regionally in Part 1 which requests some free-
writing and then in the very first sentence starter, “When I first came to New Zealand …”.
6. Coherence: Once the frame has been completed, the narrative should read like a coherent story. It is a story about
something, not a list of facts or opinions about something (which researchers could elicit using a questionnaire). From
the home tutor project see Irena's story (see Extract 1 above). Unfortunately, even with very well-designed frames
some respondents may still produce what we may consider to be an incoherent story; e.g. some spaces may be left
22 G. Barkhuizen / System 47 (2014) 12e27

blank and the story will thus have gaps. Researchers would need to make a judgement about whether or not a
particular frame is still useful. In some cases, frames may have to be omitted from the data set.
7. Formatting: As such, I strongly recommend that the blank frame be formatted like an incomplete paragraph (see
Appendix A), not as a numbered list of prompts one under each other starting against the margin. Doing so at least
creates the impression that a narrative in prose form is being written rather than responses to a list of survey questions.

5.2. Administration features

8. Instructions: Instructions to respondents need to be clear and detailed, in writing (on the frame itself) and also orally, if
possible. The instructions should inform respondents of the purpose of the research and more specifically the frame; the
topic of the frame, how it should be completed, and that a coherent story based on their experience should be the outcome
(i.e. not a series of disconnected responses). The written instructions on the home tutor frame were directed to the tutor.
These were merely a reminder, since they had been fully informed of the relevant instruction during their interview. The
learners in this case thus received the instructions second-hand. (The reason for this is that it was anticipated that some of
the learners might not have been able to understand a full list of written English instructions by themselves.)
9. Trialling: It therefore makes sense that a draft of the frame should be trialled with a similar group of respondents, or a
similar frame with the target group of respondents. Not only will this iron out any weaknesses in the design of the
frame it will also give respondents, in the latter case, the opportunity to practice writing narrative frames. Problems
that could be identified include poor wording of the prompts, insufficient writing space for some responses, and
inadequate instructions. Three home tutors volunteered to pilot the frame with their learners.
10. Language: There are many factors to consider when choosing the language of the frame. If it is practicable to design the
frame using the language of the respondents and the researcher too knows their language, then it makes sense to use
this common language for the frame. But things get more complicated if the researcher does not know the respondents'
language, or a group of respondents is multilingual, or a particular language is chosen so that the frames serve a
pedagogical purpose (e.g. writing the frames are used for target language writing practice in a classroom context). In
these cases, researchers will need to make decisions based on what is desired and practicable in their particular
research circumstances. In the home tutor study, the frame was normally used as part of the scheduled one-hour
English lesson. In fact, some tutors reported that they took multiple lessons to complete the frame with their
learner, even building other learning activities into the process. In other cases, more advanced learners simply
completed the frame independently in their own time.

6. Conclusion: the future of narrative frames

In this article, I have revisited the narrative frame research instrument by defining what it is, addressing some of the
strengths and limitations associated with its use and summarizing recent research which has utilized frames as (part of) its
methodology. I contextualized this discussion by presenting an illustrative example of a frame used to explore the problems
and successes experienced by migrant and refugee English learners in New Zealand. I ended by listing ten features which I
believe are essential to bear in mind when designing and administering narrative frames. This overview has uncovered some
advances since our original 2008 article (Barkhuizen & Wette, 2008) which introduced the concept of narrative frames, some
of which have been highlighted in the summary of subsequent research above.
But what are the possibilities for the future use of frames? I see at least three areas which could be developed further, and I
end this article with these:

1. Using alternative formats to sentence starter prompts. This format is perhaps the easiest to design and to ensure narrative
flow, but they may also ‘leave the door open’ to unwanted responses. Within sentence prompts might tighten responses, but
they might also restrict them, of course. Creative thinking is needed to generate productive and effective formats. Embedding
cues in parentheses after some starters is a useful innovation (see, e.g., Hiratsuka, 2014; Swenson & Visgatis, 2011).
2. Integrating narrative frames with other research methods. This is a suggestion we made in our original article, and my
summary of recent research above shows that advances have been made in this area. Mixed methods research has two
main aims (Sandelowski, 2003); to reach a fuller understanding of what is being investigated, and to verify one set of
findings against another. To achieve this, the methods need to be properly integrated, and this is the challenge for those
who use narrative frames.
3. The use of Web.2 technologies. Most of the narrative frames used in the studies described above were of the pencil and
paper variety. Shelley et al. (2013) sent the frames to participants by email, as did Xu (2014). These frames were in
electronic form, but they were not Web 2, which means they were not on the Web and they were not interactive in any
way. The potential for Web-designed narrative frames is enormous: the frames could be multi-layered, offering
alternative routes along the way and conversations with the researcher before, after and possibly even during their
completion! Attached to the web-based frames could be interview schedules, journals, Facebook and Twitter group
links and other research instruments, as well as pedagogical materials. The possibilities are almost endless.
 G. Barkhuizen / System 47 (2014) 12e27 23

Appendix A
24 G. Barkhuizen / System 47 (2014) 12e27
G. Barkhuizen / System 47 (2014) 12e27 25
26 G. Barkhuizen / System 47 (2014) 12e27

References

Barkhuizen, G. (2009). Topics, aims, and constraints in English teacher research: A Chinese case study. TESOL Quarterly, 43(1), 113e125.
Barkhuizen, G. (2011). Home tutor cognitions and the nature of tutor-learner relationships. In P. Benson, & H. Reinders (Eds.), Beyond the language classroom
(pp. 161e174). London: Palgrave Macmillan.
Barkhuizen, G. (Ed.). (2013). Narrative research in applied linguistics. Cambridge: Cambridge University Press.
Barkhuizen, G., & Wette, R. (2008). Narrative frames for investigating the experiences of language teachers. System, 36(3), 372e387.
Barkhuizen, G., Benson, P., & Chik, A. (2014). Narrative inquiry in language teaching and learning research. New York: Routledge.
Barnard, R., & Nguyen, G. V. (2010). Task-based language teaching (TBLT): a Vietnamese case study using narrative frames to elicit teacher's beliefs. Language
Education in Asia, 1, 77e86.
€rnyei, Z. (2009). The L2 motivational self system. In Z. Do
Do €rnyei, & E. Ushioda (Eds.), Motivation, language identity and the L2 self (pp. 9e42). Bristol:
Multilingual Matters.
 G. Barkhuizen / System 47 (2014) 12e27 27

Hayes, D. (2013). Narratives of experience: teaching English in Sri Lanka and Thailand. In G. Barkhuizen (Ed.), Narrative research in applied linguistics (pp.
62e82). Cambridge: Cambridge University Press.
Hiratsuka, T. (2014). A study into how high school students learn using narrative frames. ELT Journal, 68(2), 169e178.
Kanno, Y., & Norton, B. (2003). Imagined communities and educational possibilities: introduction. Journal of Language, Identity, and Education, 2(4), 241e249.
Labov, W. (1997). Some further steps in narrative analysis. Journal of Narrative and Life History, 7, 395e415.
Macalister, J. (2012). Narrative frames and needs analysis. System, 40(1), 120e128.
Markus, H. R., & Nurius, P. (1986). Possible selves. American Psychologist, 41, 954e969.
Nguyen, G. V., & Bygate, M. (2012). Narrative frames. In R. Barnard, & A. Burns (Eds.), Researching language teacher cognition and practice (pp. 48e67). Bristol:
Multilingual Matters.
Norton, B., & Early, M. (2011). Researcher identity, narrative inquiry, and language teaching research. TESOL Quarterly, 45(3), 415e439.
Pavlenko, A. (2007). Autobiographic narratives as data in applied linguistics. Applied Linguistics, 28(2), 163e188.
Riessman, C. K. (2008). Narrative methods for the human sciences. Los Angeles: Sage.
Sandelowski, M. (2003). Tables or tableaux? The challenges of writing and reading mixed methods studies. In A. Tashakkori, & C. Teddlie (Eds.), Handbook of
mixed methods in social and behavioral research (pp. 321e350). Thousand Oaks, CA: Sage.
Shelley, M., Murphy, L., & White, C. (2013). Language teacher development in a narrative frame: the transition from classroom to distance and blended
settings. System, 41, 560e574.
Swenson, T., & Visgatis, B. (2011). Narrative frames to assess overseas experiences. In A. Stewart (Ed.), JALT2010 Conference Proceedings (pp. 441e452). Tokyo:
JALT.
Wette, R., & Barkhuizen, G. (2009). Teaching the book and educating the person: challenges for university English language teachers in China. Asia Pacific
Journal of Education, 29(2), 195e212.
Xu, Y. (2014). Becoming researchers: a narrative study of Chinese university EFL teachers' research practice and their professional identity construction.
Language Teaching Research, 18(2), 242e259.