Chapter Title: Understanding the Problem: Student attrition and retention in

university Language & Culture programs in Australia

Book Title: The Doubters' Dilemma

Book Subtitle: Exploring student attrition and retention in university language and
culture programs
Book Author(s): Mario Daniel Martín, Louise Jansen and Elizabeth A. Beckmann
Published by: ANU Press

Stable URL: https://www.jstor.org/stable/j.ctt1rqc98q.6

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms
This content is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives
4.0 International License (CC BY-NC-ND 4.0). To view a copy of this license, visit
https://creativecommons.org/licenses/by-nc-nd/4.0/.

ANU Press is collaborating with JSTOR to digitize, preserve and extend access to The Doubters'
Dilemma

This content downloaded from

217.64.17.122 on Wed, 26 May 2021 07:41:05 UTC
All use subject to https://about.jstor.org/terms
 1
Understanding the Problem:
Student attrition and retention in
university Language & Culture
programs in Australia

The more students learn, the more value they find in their learning,
the more likely they are to stay and graduate … the purpose of higher
education is not merely that students are retained, but that they
are educated. In the final analysis, student learning drives student
retention. (Tinto, 2002, 4)

1.1. Why is it important to understand

student attrition and retention?
What makes students decide to study a language at university?
What makes those same students decide to continue or stop studying
a language? As university academics and administrators supporting
language and culture (L&C)1 studies, how do we quantify these decisions

1 In this book, the term ‘language and culture’ or ‘L&C’ is used to encompass all higher
education courses/units/programs in ‘languages other than English’ (LOTE) taught at universities,
and explicitly includes the teaching of concepts and materials related to the cultures entwined
with those languages. This terminology has been embraced by the Languages and Cultures
Network for Australian Universities (www.lcnau.org) because it evidences the fundamental
tenet that a language cannot be taught, or learned, effectively without reference to cultural
contexts and competencies. For stylistic reasons, ‘languages’ is occasionally used in Tables and
Figures, but always implies ‘L&C’.

This content downloaded from

217.64.17.122 on Wed, 26 May 2021 07:41:05 UTC
All use subject to https://about.jstor.org/terms
 The Doubters' Dilemma

of students in ways that provide effective input into resourcing? How

do we know if there are more students giving up on their L&C studies
than on their studies in other disciplines, and whether this indicates a
problem with L&C teaching? Is there anything that can, or should, be
done to help students stay in L&C programs?2
These are questions that occupy all language teachers, policy makers
and administrators at tertiary level, whatever the languages taught
and whatever the institution. At the very least, universities should
be able to measure accurately the rates at which students leave or
stay in L&C majors. Ideally, universities should also understand
exactly which factors—such as teaching style, workloads, or student
characteristics—affect attrition and retention, and which are most
influential.
Attrition is usually defined as the number of non-completing
students (i.e. students who have not yet finished their program of
study) who are enrolled in a specific university, school, discipline or
program in a given year, but not enrolled in that same program the
following year (Gabb, Milne and Cao, 2006, 3). Research into attrition
(and its corollary,  retention)—that is, examining the numbers and
characteristics of students who withdraw from, or stay in, university
study, and the reasons why they do so—has a long tradition in some
countries such as the United States of America (USA, e.g. Pascarella
and Chapman, 1983; Pascarella and Terenzini, 1991, 2005; Tinto,
1975, 1987, 1993, 2006; Wesely, 2010) but has only relatively recently
become an area of interest in the United Kingdom (e.g. Jones, 2008;
Ozga and Sukhnandan, 1997; Yorke et al., 1997), New Zealand
(Zepke and Leach, 2006) and Australia (e.g. Foster, 2010; James, Krause
and Jennings, 2010; Krause, 2005; McInnis, 2001; McInnis, Hartley,
Polesel and Teese, 2000; Pitkethly and Prosser, 2001; Taylor and
Bedford, 2004). The numbers involved are not insignificant: Pitkethly
and Prosser (2001) calculated that about a third of all Australian
students entering university at that time did not graduate, and
that half of those who withdrew did so in their first year. The two

2 Throughout this book the word ‘program’ refers to a large group of courses or multiple majors
that constitute a pathway to a certified award, such as a degree. ‘Course’ is used in the sense of a
defined unit of university study, usually equivalent to a semester of face-to-face or online teaching.
‘Major’ refers to a cluster of courses that together indicate a defined level of expected learner
competence.‘Honours’ refers to a one-year pre-doctoral research-orientated pathway.

This content downloaded from

217.64.17.122 on Wed, 26 May 2021 07:41:05 UTC
All use subject to https://about.jstor.org/terms
 1. Understanding the Problem

fundamental questions asked by those who research student attrition

at university are therefore ‘How many students discontinue their
studies before completing a degree?’ and ‘Why do students do this?’
Higher education administrators and budget planners have commonly
used the very unsophisticated tool of raw enrolment numbers both
as a measure of attrition and a surrogate indicator of a program’s
success. For example, the Australian Department of Education,
Science and Training provided the first notable data set (1994–2004)
for Australian higher education using ‘simple measures of attrition at
an institution level [whereby] the attrition rate plus the retention rate
plus the completion rate for a given student population in a  given
year will equal 100 percent’ (Lukic, Broadbent and Maclachlan, 2004,
2). The authors noted that, for methodological reasons, the rates
included ‘students who leave a course at one university and enrol
the next year at another university … [and] those students who leave
university without completing their course, but who return later
to the same university’ (Lukic et al., 2004, 2). Using this measure,
the Commonwealth Government reported an average attrition rate
of 18 per cent for all students in Australian universities during the
period 1994–2002, although considerable variation was noted across
institutions and different student populations (Lukic et al., 2004).
At  The Australian National University, for example, attrition rates
for first year students in that period were somewhat higher than the
national average, and increased from 22 per cent to 24 per cent over
the eight-year period.
Shaw (2008) has subsequently argued that, because this method
of calculating attrition rates fails to allow for students who leave one
university but enrol at another university the following year, the real
national average attrition rate in 2002 was more likely to have been
around 10 per cent. The rudimentary nature of such a measure of
attrition is even more concerning when we remember that, as a critical
outcome of the Nelson reforms in the early 2000s, attrition rates
calculated in this way were used as performance indicators for the
allocation of Learning and Teaching Performance Funds to universities
(Gabb et al., 2006), despite the resultant data being relatively
untrustworthy. A subsequent review of base funding did show that
the government attrition rates used to measure and allocate university
performance funding had indeed been misleading: in reality, about

This content downloaded from

217.64.17.122 on Wed, 26 May 2021 07:41:05 UTC
All use subject to https://about.jstor.org/terms
 The Doubters' Dilemma

10 per cent of students who had been counted as ‘discontinuing’

their higher education had actually transferred to another university
(Lomax-Smith, Watson and Webster, 2010).
With this economic incentive, the paired parameters of attrition and
retention became popular focal points for researchers in Australasian
higher education in the early 2000s (e.g. McInnis and James, 2004;
Taylor and Bedford, 2004; Zepke and Leach, 2006), especially with
regard to the ‘first year experience’ (e.g. James et al., 2010; Krause,
2005; McInnis, James and Hartley, 2000; Nelson, Duncan and Clarke,
2009; Pitkethly and Prosser, 2001). These studies reported a clear
need, from both economic and pedagogical perspectives, to identify
students who are ‘at risk’ of withdrawing from individual courses, or
from university study as a whole, especially in their first year of study:
one wonders whether, if the institutions to which these potential
dropouts belonged had known what they were thinking and feeling
and why, things might have been done any differently to support
them (Krause, 2005, 58).

Such identification requires an understanding of relevant student

motivations. In a seminal work on retention, Tinto (1975) identified
four factors of key importance: instruction, academic success, anxiety
and motivation. Wesely (2010) explored the literature on these four
factors in the specific context of foreign language teaching (from a US
perspective). Overall, however, motivation remains little understood:
while students provide many reasons for leaving university before
graduating (Figure 1.1), Pitkethly and Prosser (2001, 186) argued
that the factors most likely to affect students’ failures or course
withdrawals seemed related more often to the students’ adjustment to
the university context rather than to their difficulties with intellectual
understanding of the relevant content, and hence have a local/national
context that must be considered. Whether the motivations to give up
study are the same for students in L&C programs is not clear: despite
the new interest in university student attrition as an area worthy of
empirical research, by the mid-2000s there was still little data available
with respect to attrition of students in L&C programs, even in the USA
(Wesely, 2010), and it is not clear how generalisable many research
findings are to the special circumstances of L&C teaching.

This content downloaded from

217.64.17.122 on Wed, 26 May 2021 07:41:05 UTC
All use subject to https://about.jstor.org/terms
 1. Understanding the Problem

• Wrong choice of program

• Poor quality of the student experience
• Inability to cope with the demands of the program
• Unhappiness with the social environment
• Matters related to financial need
• Dissatisfaction with aspects of institutional provision
• Problems with relationships and finance
• Pressure of work (academic and employment)
• Learning efficiency (students’ general cognitive skills)
• Self efficacy (self reliance, locus of control, self directedness)
• Quality of instruction (perceptions of the quality of teaching)
• Course difficulty (in relation to available academic support and counselling)
• Interaction with academic staff
• Goal commitment (planning skills, motivation)
• Time for learning (planning and organising study programs)

Figure 1.1. Factors known to have a negative effect on student completion

Source: After Longden, 2006; Taylor and Bedford, 2004, 376; Tinto, 1975; Weston, 1998;
Yorke, 1999.

1.2. Attrition as a concern for university

Language & Culture programs in Australia
One could argue that the most striking characteristic of L&C programs
in Australian universities is the relative scarcity of students. Although
it is a complex statistic to calculate in any national or pan-national
context, a snapshot of language teaching in Australian higher education
in the early 1990s noted that just 2 per cent of higher education
students were studying languages, with the highest proportion being
in the Australian Capital Territory (Leal, Bettoni and Malcolm, 1991).
Subsequent estimates suggested a relative increase in interest—Hajek
(2001) reported 5 per cent of Australian university students studying at
least one language, while Nettelbeck, Byron, Clyne, Hajek, Lo Bianco
and McLaren (2007, 2) reported that ‘fewer than 10% of first-year
students undertake LOTE [language other than English] study of any
kind … with overall languages enrolments stagnant over the 2005–
2007 period while student cohorts increased’. While the most recent
available Australian Government data, from 2013, does not refer
specifically to L&C study, it suggests that fewer than 5 per cent of
all students are studying in the broad field of  Society  and Culture
(Australian Government Department of Education and Training, 2014).

This content downloaded from

217.64.17.122 on Wed, 26 May 2021 07:41:05 UTC
All use subject to https://about.jstor.org/terms
 The Doubters' Dilemma

The 2014 First Year Experience Study data from the University of
Melbourne’s Centre for the Study of Higher Education encourages
some optimism: in both 2009 and 2014 (sample sizes 2,422 and 1,739
respectively), 23 per cent of the national first year respondents reported
that they ‘planned to, or were, studying a language as part of their
course’ (Baik, Naylor and Arkoudis, 2015). By way of international
comparison, language studies were reported as accounting for 8.6 per
cent of all 2009 course enrolments in US higher education institutions
(Furman, Goldberg and Lusin, 2010, 5), and 8.1 per cent of all 2013
enrolments (Goldberg, Looney and Lusin, 2015), while Byrne (2005)
reported that fully one third of tertiary students in Europe are
studying languages as an assessable part of their degree.
The attrition rate of L&C students in Australian universities is of
particular concern. In landmark research (detailed later in this
chapter), Nettelbeck et al. (2007, 3) reported that:
on average, one third of students beginning a LOTE at [an Australian]
university do not complete more than one semester; a third of those
remaining do not continue into second year; there is further attrition
after second semester of second year, and of those completing second
year, only two thirds continue into third year. Overall, fewer than
25% of students beginning a LOTE complete a third year.

Given this kind of data, and their lived experience of student attrition
during a program, all L&C teachers in Australia at tertiary level—
whatever the language they teach and whatever their institution—are
likely at some point (usually when enrolments drop and their course
is threatened) to ask themselves not only ‘What makes students
decide to study a language at university?’ but also, perhaps even more
urgently, ‘What makes those same students decide to continue, or to
stop, studying a language?’
Despite some key attempts in recent years to investigate these
questions in a sector-wide context (e.g. Nettelbeck et al., 2007;
Nettelbeck, Byron, Clyne, Dunne, Hajek, Levy, Lo Bianco, McLaren,
Möllering and Wigglesworth, 2009), Australian research in this field
has been very limited. On the basis of an extensive literature review,
Lobo and Matas (2010, 39–40) argue that there is still an inadequate
volume of research into student attrition in language learning courses,
and that the research that does exist is patchy and poorly integrated
into an overall theoretical framework. With no definitive data on

This content downloaded from

217.64.17.122 on Wed, 26 May 2021 07:41:05 UTC
All use subject to https://about.jstor.org/terms
 1. Understanding the Problem

whether the key influences on students’ decisions about continuing

or discontinuing language studies are related more to aspects of the
teaching or the workload, or to the inherent or acquired characteristics
of the students themselves, university language teachers have few
evidence-based strategies with which to confront the harsh economic
drivers that see languages with small overall enrolments or apparently
high attrition rates relegated to minimal funding options or closure.
With attrition measures given sector-wide importance for decision-
making around financial support, despite there being little evidence
that attrition really is a measure of performance or quality, we
should expect that, at the very least, universities are able to report
accurately the rates at which students leave or stay in L&C programs.
This was the goal set by the language-teaching academics at ANU in
2008, who decided to initiate an institution-wide research project to
explore the issue of attrition and retention in all the L&C programs
at ANU. (See  Acknowledgements for details of participants in this
research.) As  no other Australian institution has conducted a study
of such scope in breadth and depth, the research has become an
important case study in this field, but one that has, until now, not
been reported in its entirety, although there have been preliminary
and selective presentations and publications (e.g. Jansen, Åkerlind
and Maliangkay, 2011; Jansen and Martín, 2011; Jansen, Martín and
Åkerlind, 2009; Jansen, Maliangkay, Martín and Åkerlind, 2009;
Jansen and Schmidt, 2011; Martín and Jansen, 2011, 2012; Martín,
Jansen and Beckmann 2015).
In this book, we remedy that omission by reporting in full the relevant
methodologies, analytical processes and outcomes of the ANU case
study, and thus provide other researchers with access to what is
probably the most detailed and comprehensive institutional data set
in the field. We also explain how we were able to use this data set
to interrogate the motivations and constraints that influence tertiary
students’ decisions as to whether to continue or discontinue their L&C
study. As we take the reader forward into understanding the context
for this institutional case study, with detailed presentation, analysis
and discussion of the research findings, we will start building the
thesis of this book, namely that university language departments must
become more aware that students at risk are found at all levels of L&C
study (not just in first year or Beginner cohorts); that the ‘language

This content downloaded from

217.64.17.122 on Wed, 26 May 2021 07:41:05 UTC
All use subject to https://about.jstor.org/terms
 The Doubters' Dilemma

capital’ of students plays a role in their propensity to continue their

studies; and that policies that cater to the needs of all students are
crucial to maximise retention through all levels of L&C programs.

1.3. The impact of government, university

and school policies on language teaching
in Australia
Before engaging the reader with the rationale and methodology of the
ANU case study, however, we feel it is important to explain some
aspects of the broader context of language teaching in Australian
universities and their feeder systems (especially secondary schools).
Australian government policies, along with societal and external
factors, have clearly exerted significant influence on the levels of
enrolment, retention and attrition in L&C programs in schools and
universities (Clyne 1993, 1997; Djité, 2011; Kleinsasser, 2000; Leopold,
1986; Liddicoat, 2010; Liddicoat and Scarino, 2010; Lo Bianco and
Gvozdenko, 2006; Nicholas, 2004; Pauwels, 2002; White and Baldauf,
2006). Djité (2011, 65) provides a thoughtful historical analysis of the
way in which ‘national sentiment and ideologies have … dictated
language policy in Australia over the last 30 years’, and concludes
that ‘language policy in Australia continues to be a site for negotiation
between the monolingual ethos and the urge for linguistic pluralism’.
Two examples of influential late twentieth-century policies are the
1991 Australian Language and Literacy Policy (Australia, Department
of Employment, Education and Training, 1991), and the 1994 report to
the Council of Australian Governments (1994). The latter led directly
to the National Asian Languages and Studies in Australian Schools
(NALSAS) Strategy 1996–2002, which impacted on both the funding
and the demand for Asian languages in schools and, as a knock-on
effect, in universities.
Not surprisingly, many authors believe that the provision and uptake
of languages in the Australian tertiary sector has been directly—and,
most argue, negatively—influenced by the lack of effective language
provision in the secondary (school) education sector (Group of Eight,
2007; Liddicoat, Scarino, Curnow, Kohler, Scrimgeour and Morgan,
2007, 38–41; Liddicoat and Scarino, 2010; Lo Bianco, 2009, 48–51).
Despite multiple strategic federal and state/territory government
8

This content downloaded from

217.64.17.122 on Wed, 26 May 2021 07:41:05 UTC
All use subject to https://about.jstor.org/terms
 1. Understanding the Problem

policy changes, each indicating a willingness to address the issue

of language study at secondary level, there appears to have been no
increase in the last 20 years in the proportion of Year 12 students
studying a language (Liddicoat et al., 2007, 38–41; Lo Bianco, 2009,
48–51), which is the indicator most commonly used as a surrogate
for the extent of language study at secondary level. In contrast to
the 10 per cent of Australian Year 12 students studying a foreign
language in 2006 (Lo Bianco, 2009, 49), about 60 per cent of senior
secondary students in Europe in 2009/10 were learning two or more
foreign languages (Education, Audiovisual and Culture Executive
Agency, 2012). In the words of the Group of Eight (research-intensive)
universities: ‘decades of policy neglect and inaction mean Australian
school students now spend less time learning a language than students
in all other OECD countries’ (Group of Eight, 2007, 1).
What has led to this somewhat parlous state of affairs? Martín
(2004,  2005) and Lo Bianco (2009) argue that the low levels of
commitment to language study in Australia have resulted from various
historical circumstances meshed with the characteristics of ethnic
community relations, which have together discouraged the use of
languages other than English in mainstream settings. Martín (2004,
2005) identifies three key twentieth-century influences: i) in the late
1940s and 1950s, newly created Australian universities often decided
to waive knowledge of a language other than English (LOTE) as an
entrance requirement; ii) in the 1960s, curriculum reforms reduced
language provision in secondary schools; and, iii) since the late 1980s,
the predominance of Australian government economic rationalist
policies in higher education have not favoured the labour-intensive
and small-enrolment nature of L&C courses. In evidence for the
latter, for example, from 2001 to 2005 enrolments in L&C courses in
Australian universities remained relatively stable, but fewer languages
were taught, and there was an increasing reliance on casual, rather
than permanent, language teaching staff (White and Baldauf, 2006).
In a more recent review of language offerings at Australian
universities, Dunne and Pavlyshyn (2012, 15) argued that the
‘apparent health’ of tertiary language teaching in Australia, based
on the total number of less commonly taught languages on offer, was
‘illusory and potentially misleading’, because the majority of those

This content downloaded from

217.64.17.122 on Wed, 26 May 2021 07:41:05 UTC
All use subject to https://about.jstor.org/terms
 The Doubters' Dilemma

languages were only taught at ANU, and thus, in the author’s words,
and quite prophetically (Macdonald, 2015), ‘vulnerable to changes in
[that institution’s] financial climate’.
Internal university policies also play an important role in encouraging
student enrolments. In the USA, most tertiary institutions traditionally
had compulsory language requirements for all undergraduate degrees
(McGroarty, 1997, 80–83). Although this requirement has become less
prevalent in recent years (Furman et al., 2010, 5), at least half of all
US universities were still insisting on compulsory language study in
2010 (Lusin, 2012). This is especially true for institutions with highly
competitive entry. For example, Yale College—a partner of ANU in
the International Alliance of Research Universities (IARU)—requires
all students to study a foreign language, regardless of their existing
knowledge of that language or another (Yale College, 2015). At the
University of California, Berkeley, (another IARU member), every
student in the College of Letters and Science (but not in all Colleges)
must demonstrate ‘proficiency in reading comprehension, writing, and
conversation in a foreign language equivalent to the second semester
college level, either by passing an exam or by completing approved
course work’, although this can be achieved through evidence of
appropriate high school study (University of California, Berkeley,
2015). It is notable that, following increases in aggregate US higher
education enrolments in all languages consistently from 1980 to 2009,
there was a decrease in the period between 2009 and 2013 (Goldberg,
Looney and Lusin, 2015), the same period during which compulsory
requirements were becoming less common or less rigorous (Furman et
al., 2010, 5).
By contrast, no Australian university has compulsory language
requirements for all undergraduate degrees on offer, although a
limited level of compulsion may occur in some degree programs.
At ANU, degrees with compulsory language study accounted for just
10 per cent of the total student load in 2008 and 2009 (the relevant
period for the case study): this proportion has decreased even further
in the light of subsequent reforms in relevant degree structures.

10

This content downloaded from

217.64.17.122 on Wed, 26 May 2021 07:41:05 UTC
All use subject to https://about.jstor.org/terms
 1. Understanding the Problem

1.4. Historical perspectives from the

teaching coalface
The changes and difficulties experienced in Australian university
language departments in the mid to late twentieth century is
perhaps best encapsulated by the personal case study of Professor
Keith Leopold, who detailed his own experiences in the German
department at the University of Queensland over a period of some
40 years (Leopold, 1986). Identifying many factors that impacted on
course structure, teaching approaches, workload, and standards—
all of which significantly influenced what he called the ‘struggle for
students’—Leopold (1986, 9) described the outcome of his long-term
perspective from the coalface: ‘as the stress on numbers has become
greater and greater, finances have become tighter and tighter, [and]
the utilitarian aspects of education have moved more and more into
the foreground’. Leopold (1986) especially noted the significant
stresses on staff that derived specifically from the widening range in
the language competence of beginner students. Some 30 years later,
our experience is that many university language teachers still identify
very closely with Leopold’s concerns.
In the same time frame, but methodologically in diametric contrast
to Leopold’s very personal analysis, an extensive set of relevant
research data was collected by Bowden, Starrs and Quinn (1989).
Through diverse methods, including a national survey, interviews,
and observations, these authors examined the attitudes of Australian
university academics who were teaching LOTE on aspects such
as students’ entry skills, streaming, student workload, curricula,
course structures, students’ expectations, the use of audio-visual
media, teaching specialisations, language-teaching methodologies,
and the status of language teaching. Most of the university language
departments in the study complained of difficulties with staffing
levels, which impacted on the feasibility of implementing appropriate
approaches to teaching (Bowden et al., 1989).
The diversity in background knowledge and skills of students starting
L&C study has long been an issue for university teaching departments.
Bowden et al. (1989) found that first year students showed great
diversity in their previous experiences of learning a second language,
which was attributed to inadequate language provision in the

11

This content downloaded from

217.64.17.122 on Wed, 26 May 2021 07:41:05 UTC
All use subject to https://about.jstor.org/terms
 The Doubters' Dilemma

secondary sector. The researchers concluded that teaching staff could

no longer expect all students to have specific skills or capabilities
on entry, and that budgetary constraints made it difficult to provide
enough staff to meet the consequently varied and divergent needs of
these students (Bowden et al., 1989). Smaller departments in particular
have to balance the learning benefits and staff costs of finely tuned
streaming versus the need to ensure adequate assistance to individual
students in more broadly streamed, and thus staffing-effective,
placements (Bowden et al., 1989, 132). Dealing with this by streaming
students according to their level of language attainment on entry
(i.e. placing language-competent first year students into second or
third year L&C classes) has, in our experience, met with only limited
success, because of budget constraints that impact on the required
class sizes and teaching methods.
In addition, since the mid-1980s, most Australian institutions have had
more students enrolling in languages at Beginner Level (Level 1) than
at other entry points (Hawley, 1982; Nettelbeck et al., 2007), because of
the reduction of language teaching in high schools. Naturally, this has
implications for undergraduates’ potential levels of achievement: for
example, Australia’s Group of Eight research-intensive universities—
which includes ANU—openly acknowledge that students who start
out as beginners at university are unlikely to achieve a sophisticated
level of language competence in three years of classroom-based study
alone (Group of Eight, 2007, 4).
At this stage, readers may be wondering why this chapter is referring
extensively to research that dates back more than 25 years. The sad
truth is that there is still a disappointing relevance to this data, and
to the conclusion reached by Bowden et al. (1989, 129), namely that
‘tertiary language teaching bristles with sensitive and contentious
issues … subject to conflicting opinion and practice’. At least four
of the key issues highlighted all those decades ago by Bowden et al.
(1989) are still highly relevant in Australia today.
First, problematic degree structures are still making it difficult for
students to combine the study of a language with other subjects:
despite some universities implementing changes that improve
the situation (e.g.  Diploma of Languages at several universities;
the  creation  of flexible double degree opportunities at ANU;

12

This content downloaded from

217.64.17.122 on Wed, 26 May 2021 07:41:05 UTC
All use subject to https://about.jstor.org/terms
 1. Understanding the Problem

and  the  degree curriculum changes at the University of Melbourne

and the University of Western Australia), degree structures remain
a key barrier to L&C studies in many universities.
Second, the increasing number of international students enrolling
in higher education who wish to study their own first language, or
another language that they have previously studied or spoken, has
increased still further the range of entry skill levels among students,
with a consequent emphasis on the need for ever more effective
placement tests and streaming. This aspect is crucially important
to learning outcomes: Bowden et al. (1989, 131) found high levels
of dissatisfaction among first year students who were placed in
unstreamed classes, not least because complete beginner students
were reluctant to speak their ‘new’ language in the presence of more
advanced students.
A third issue that remains highly topical is the disparity between
the expectations of students and those of teachers and university
administrators. Bowden et al. (1989, 139) found ‘a close correlation
between the level of student satisfaction with the course, and the
level of [oral] fluency achieved’ in all the language departments they
visited. Students were strongly in favour of communicative approaches
to language teaching: most reported a desire to speak the language of
study, while relatively few wanted to read literature in that language
(Bowden et al., 1989, 141). The researchers concluded that Australia
was ‘witnessing a major shift in orientation away from a traditional
humanist view of university language teaching … towards a very
pragmatic emphasis on practical communicative competence’ (Bowden
et al., 1989, 145).
Leal et al. (1991) soon confirmed this perception in a government-
sponsored Australia-wide review, revealing with concern that
outcomes sought by students did not always correspond to those
sought by teaching staff. While heads of departments focused first
on students’ linguistic and reading performance, and next on their
cultural knowledge, a large majority of students were primarily
seeking a high level of oral/aural proficiency, with an appreciation
of the relevant society and culture being quite a secondary objective:

13

This content downloaded from

217.64.17.122 on Wed, 26 May 2021 07:41:05 UTC
All use subject to https://about.jstor.org/terms
 The Doubters' Dilemma

although … third year students were happy or very happy with their
courses, they did not hesitate to propose changes in the curriculum
… [most frequently] a request for more oral input into the mode of
teaching (38%). As in most past surveys, oral command of the language
was what students most wanted to achieve (Leal et al., 1991, 120).

The ‘contest’ between spoken and other forms of language learning

remains a concern of curriculum designers today.
The final issue identified by Bowden et al. (1989) that is still highly
relevant today concerns student perceptions that workloads are
different in different languages (for example, that European languages
require relatively less effort than Asian languages). The researchers
found students reporting that they had withdrawn from L&C courses
with (perceived) heavy workloads because the students felt that such
workloads would prevent the attainment of the relatively high marks
needed to assure scholarships and jobs.
In closing this section, we note with disappointment that key issues
about L&C teaching in universities that were raised some 25 years
ago by Leopold (1986) and Bowden et al. (1989)—degree structures
unsupportive of L&C study; high diversity in student cohorts;
curriculum design conflict from the perspective of students and
staff around the relative importance of spoken and written language;
and perceptions of workload—are still highly relevant, empirically
identified in the findings of the ANU case study (considered in depth
in Chapter 4) as well as in other studies described below.

1.5. Research into attrition in university

Language & Culture programs
Despite an increasing focus on retention and attrition as surrogate
measures of performance, including as funding indicators (Gabb et al.,
2006), until the late 2000s there was virtually no systematic research
on L&C courses in Australian universities, and only a little relevant
research focused on the school sector. While school and university
perspectives are by no means equivalent—school students are often
strongly influenced by their parents, and schools are affected by state

14

This content downloaded from

217.64.17.122 on Wed, 26 May 2021 07:41:05 UTC
All use subject to https://about.jstor.org/terms
 1. Understanding the Problem

and territory government policies as well as national ones—language-

teaching staff in universities have often been able to relate anecdotally
to some of the findings and conclusions of school-focused studies.
For example, there was interest in the research by Curnow and Kohler
(2007) on why high school students continued studying a language:
the most important reasons were academic achievement, personal
interests, and relationships, with other notable factors including
bonus schemes that rewarded language study in university entry
schemes; students having travelled to, or having connections with,
a country where the target language was spoken; and the influence of
friends. Important reasons for discontinuing study included the lack
of availability of their preferred language; the perception that learning
languages gave rise to a relatively higher workload than learning other
subjects; and the belief that language learning was not meaningful,
because other subjects carried more value (Curnow and Kohler, 2007).
Another schools-focused discussion worthy of note is the more
philosophical, rather than empirical, review of language education
in Australia’s schools by Scarino (2012). Noting the complexity of
measuring language-learning outcomes among individual students
and cohorts, especially in the light of the ‘highly diverse teaching,
learning and assessment practices and diverse expectations about
learner achievements’ created by the diverse policy contexts across
Australia, Scarino (2012, 240) identified the need nationally for
a ‘curriculum and assessment framework that acknowledged the
diversity of student achievements’ to provide baseline and reference
points for monitoring and planning.
In university-based language education, the continuing lack of
reliable  and valid data on any aspect of tertiary L&C programs
(Leal  et  al., 1991; Murray, 2010), let alone on the key aspects of
retention  and attrition (Lobo and Matas, 2010), means that even
informal  findings from universities have been valued. One such
study, given much attention at the time, was the internal review
of the University of Melbourne’s Diploma in Modern Languages
(DML), documented (but not published) by Rover and Duffy in 2005.
The DML was a supplementary program that allowed undergraduate
students to study a language in addition to their degree. The internal
review was triggered by high discontinuation rates (about 60 per cent
of enrolled students) occurring in the early 2000s. The review

15

This content downloaded from

217.64.17.122 on Wed, 26 May 2021 07:41:05 UTC
All use subject to https://about.jstor.org/terms
 The Doubters' Dilemma

examined six of the program’s languages (French, German, Italian,

Chinese, Japanese and Indonesian) through semi-structured individual
interviews with 13 staff and 50 current or past students, across wide-
ranging topics, including students’ motivations for enrolling, their
learning experiences during the program, and, where applicable,
their reasons for discontinuing. The qualitative data collected in this
way were complemented by internal statistical data.
The findings showed that the DML was regarded positively by
both staff and students, even when those students had withdrawn
from the program (Rover and Duffy, 2005). Students reported many
reasons for enrolling in the DML, including (in descending order of
frequency) wanting greater ‘flexibility’ in their degree; ‘continuing
language study’ (beyond the Year 12 certification level); ‘learning
a language’ (for beginners); and the ‘opportunity to learn two
languages’. Students most commonly left the program in its earlier
stages: although there were many reasons given, the most common
were ‘high workload’, ‘wrong placement level’, and ‘personal reasons’
(Rover and Duffy, 2005).
While it was clear in the mid-2000s that strategically designed and
evidence-based policies were crucial to the quality of future language
teaching in the higher education sector, it was equally obvious that
Australia lacked valid and reliable empirical data on which to base such
policies. In consequence, the Australian Academy of the Humanities
decided to fund first one, then a second, national investigation
into Beginner (Level 1) courses in university L&C programs. These
Australian Research Council Linkage Learned Academies Special
Projects (LASP) studies delivered broadly scoped and wide-reaching
findings, including specific consideration of issues related to retention
and attrition, documented in two reports referred to hereafter as
LASP1 (Nettelbeck et al., 2007) and LASP2 (Nettelbeck et al., 2009).
The timing of these two LASP studies—which respectively involved
data collection in 2007 and 2008, and reports in 2007 and 2009—largely
paralleled the in-depth institutional case study being conducted from
2008 to 2009 at ANU on the nature of retention in L&C programs.
This coincidence of timing meant that ANU language-teaching staff
were contributing to LASP1, and facilitating the involvement of ANU
students in LASP2, at the same time as supporting the institutional
research reported in this book.

16

This content downloaded from

217.64.17.122 on Wed, 26 May 2021 07:41:05 UTC
All use subject to https://about.jstor.org/terms
 1. Understanding the Problem

We ask readers to recognise that this concurrency means that the LASP
findings were not available when the ANU data collection research was
being planned, implemented and analysed. Nevertheless, the fact that
the LASP2 data were being collected in the same year as those of the
second phase of research at ANU meant that the ANU research team
was able to support its fellow researchers by maximising the collection
of LASP2 data at ANU, thus ensuring that the ANU case study—
focused in depth on one institution—would complement and give
more resonance to LASP2, which was focused on many institutions.
Bearing this timing in mind, we will now review the methodologies,
findings and implications of the LASP1 and LASP2 research, as well as
some more recent research on retention strategies for L&C at the course
level, again unknown to the ANU researchers at the time (e.g. Lobo
and Matas, 2010, 2011; Hanley and Brownlee, 2013), before delving
deeply into the ANU institutional case study in future chapters.

1.6. The LASP1 study

The LASP1 research provided an audit survey of Beginner (Level 1)
courses in L&C university programs Australia-wide, derived
from an intense study of 10 universities and at least 10 distinct
languages (Nettelbeck et al., 2007). Data collection methods included
questionnaires, classroom observations and interviews with staff
‘interlocutors’ (but notably, not with students: their voices did not
come into play until LASP2). Unfortunately, the LASP1 report did not
identify all the languages that it had covered, but it did identify a
focus on six languages with increasing enrolments, and four languages
with decreasing enrolments (Nettelbeck et al. 2007, 12). With respect
to retention and attrition, LASP1 requested and analysed retrospective
longitudinal enrolment data over five semesters from two cohorts,
namely those students who had started studying in a Beginner3 L&C
course in Semester 1 (February) 2005 and those who had started in
Semester 2 (July) 2006.

3 We use the nomenclature of Beginner, Intermediate and Advanced courses to denote Level
1, 2 and 3 courses respectively. These are usually expected to equate to first, second and third
year enrolments at university, but—as we explain in detail in later chapters—this form of
contextualisation actually creates an excessive oversimplification that hinders, rather than helps,
an understanding of the complexities involved.

17

This content downloaded from

217.64.17.122 on Wed, 26 May 2021 07:41:05 UTC
All use subject to https://about.jstor.org/terms
 The Doubters' Dilemma

The authors noted that determining attrition was ‘particularly arduous

because many institutions merge Beginners’ streams with others at
various points’, so that the research required ‘close analysis of actual
class lists as distinct from enrolment numbers’ (Nettelbeck et al.,
2007, 14). Moreover, while some of the 10 institutions surveyed could
provide such detailed source data, others could not. Nevertheless, the
authors believed that ‘sufficient data was collected overall … to make
some important observations’ (Nettelbeck et al., 2007, 14).
The key LASP1 findings (Nettelbeck et al., 2007, 14) relevant to the
theme of this book were that:
• retention from Beginner (Level 1) to Advanced (Level 3) courses
averaged just 25 per cent for the 2005 cohort, with a similar pattern
found in the 2006 cohort
• retention rates varied considerably among institutions and
• retention rates varied considerably within institutions for different
languages.
Why were some three quarters of students who started an L&C course
at university giving up on those L&C studies before completing their
degree? The staff interlocutors who were surveyed suggested four
reasons for the high attrition rates, namely that many students:
• had problems with the (perceived) heavy workload
• were frustrated with their slow progress
• were experiencing timetabling problems and/or
• were starting a language as an elective in later years (Nettelbeck
et al., 2007, 15).
With such high attrition rates in L&C programs Australia-wide finally
revealed, the LASP1 researchers identified an urgent need for a large-
scale national study, which was soon realised in LASP2 (Nettelbeck
et al., 2009).

1.7. The LASP2 study

The LASP2 research essentially involved a follow-up study of 11
universities: the original 10 universities examined in LASP1, plus one
more. Like LASP1, the focus of LASP2 was Beginner (Level 1) students.
18

This content downloaded from

217.64.17.122 on Wed, 26 May 2021 07:41:05 UTC
All use subject to https://about.jstor.org/terms
 1. Understanding the Problem

The data collection focused on a semi-longitudinal student survey

to explore retention strategies and the use of technology-enhanced
language learning, via two hard-copy questionnaires completed by
2,968 students in Semester 1, 2008 and 1,810 students in Semester 2,
2008 (Nettelbeck et al., 2009). The questionnaires contained 14
structured questions—with mostly predetermined response choices—
regarding students’ academic profile, language background, intended
length of study and motivations. (It is important to note here that,
although LASP2 sought a longitudinal dimension to its data analysis,
this was not possible in terms of statistical validity, as the surveys
did not control for individual student identity in the two sets of
responses.)
Among a wealth of results, the LASP2 study had seven key findings
of particular relevance to this book’s theme: four of these findings
identified factors that are potentially confounding for those studying
student attrition, while the remaining three findings related to student
motivation, data collection issues and policy recommendations
(Nettelbeck et al., 2009).
The first relevant LASP2 finding—and one which confirmed data from
LASP1—was that many first year L&C students are late enrolments,
‘taking up a language too late in their studies to be able to complete
a major or even a minor sequence in the language’ (Nettelbeck et al.,
2009, 11). Whereas traditional measures of L&C attrition assume that
all students begin language study in their first year at university,
no less than half the students who responded to the LASP2 surveys
reported starting a language after their first year of study at university.
Although even short-term language learning has significant value—as
Nettelbeck et al. (2007) had argued in the LASP1 report—there are
significant implications for course and program planning and design
if students are enrolling later than expected in their university career
(Nettelbeck et al., 2009). We draw readers’ attention to this issue,
and will explore the full import of ‘Late Starter’ language students
in Chapters 2 and 5, where we present a detailed analysis of the Late
Starter phenomenon as explored in the ANU case study.
A second important LASP2 finding from our perspective was that
even Beginner (Level 1) classes contain mixed levels of proficiency:
indeed, among the so-called Beginner students surveyed in LASP2,
just 38 per cent actually had no previous background in the target

19

This content downloaded from

217.64.17.122 on Wed, 26 May 2021 07:41:05 UTC
All use subject to https://about.jstor.org/terms
 The Doubters' Dilemma

language, while most Beginners had diverse previous language-

learning experiences, including a ‘not insignificant number … who
had successfully completed [that language at school in] year 12’
(Nettelbeck et al., 2009, 12). As identified decades earlier by Leopold
(1986) and Bowden et al. (1989), the impact of such mixed proficiency
groups in Beginner classes was problematic for teachers, potentially
creating ‘perceptions of disadvantage’, and hence negative impacts on
motivation, among genuine Beginners (Nettelbeck et al., 2009, 12).
In addition, cross-cultural, related-language issues were evident: for
example, some 30 per cent of students enrolling in Beginner Spanish
had previously studied French, which might confer some familiarity
advantage, while about half the students who were enrolled in
Beginner Japanese identified themselves as native Chinese speakers,
which might confer some advantage in terms of character recognition
(Nettelbeck et al., 2009, 12). This situation again arose in the context
of the ANU case study, and is explored in detail in chapters 2 and 4.
The LASP2 study also found a mismatch between students’
expectations of workloads and the reality—or, rather, students’
perceptions of reality—with many students reporting that their L&C
workload was ‘higher than expected’ (Nettelbeck et al., 2009, 19),
again echoing the earlier research findings by Bowden et al. (1989).
Perceiving the workload as high was not a clear disincentive, however:
many students reported that they would be studying the language
for longer than they had originally planned because they had found
the learning ‘more interesting’ and/or the teaching ‘better’, than
expected (not because they had experienced ‘less work than expected’
or other reasons). Finding students apparently pleasantly surprised
by the quality of language teaching at university, Nettelbeck et al.
(2009, 19) concluded that ‘high attrition does not appear to be caused
(and  may in  fact be mitigated) by perceived quality of teaching or
course interest’.
Crucially, LASP2 respondents valued language speaking skills most
highly, followed by understanding (Nettelbeck et al., 2009, 19).
This  finding—again consistent with previous research (e.g. Bowden
et al., 1989; Leal et al., 1991)—has profound implications for course
design. Unfortunately, as Nettelbeck et al. (2009, 19) explained—and
again echoing the research findings of 20 years earlier—‘the dominant

20

This content downloaded from

217.64.17.122 on Wed, 26 May 2021 07:41:05 UTC
All use subject to https://about.jstor.org/terms
 1. Understanding the Problem

motivations [of students] could hardly be clearer, [but] the degree to

which [these motivations] are taken into account in course planning
and design is less evident’.

1.8. Implications of the LASP1 and LASP2

research
Both LASP studies were unequivocal in their call for action as a result
of their research findings. The LASP1 authors identified ‘an urgent
need for governments and universities alike to recognise languages
as a strategic and essential sector and to support them accordingly’
(Nettelbeck et al., 2007, 6).
This view was reiterated by the LASP2 report’s primary
recommendation:
That universities, at the policy level, give explicit and urgent
recognition of the strategic importance of the study of languages and
cultures; and that they develop appropriate strategies and provide
adequate resources for the promotion and effective maintenance of
these studies (Nettelbeck et al. 2009, 6).

A direct outcome of these recommendations—and a policy action

of great significance with regard to L&C programs in Australian higher
education—was the creation in 2011 of the Languages and Cultures
Network for Australian Universities (LCNAU; see www.lcnau.org;
Hajek, Nettelbeck and Woods 2013). This network, which aims to
raise the profile of language educators and public awareness of the
cultural, strategic and economic importance of language education
in, and for, Australia, is already having an impact as a central voice
and focus for research outcomes and policy development (John Hajek,
pers. comm., 2014).
Of even more significance to the central theme of this book, both
LASP1 and LASP2 researchers identified the lack of accurate data
as a significant hindrance to the calculation of realistic rates of
attrition in university L&C programs. The LASP1 authors placed
the onus for better data collection onto universities, recommending
the ‘creation of processes to ensure that universities collect data in
a readily accessible form on the LOTE experience of their students,
including formal secondary training and background experience’
21

This content downloaded from

217.64.17.122 on Wed, 26 May 2021 07:41:05 UTC
All use subject to https://about.jstor.org/terms
 The Doubters' Dilemma

(Nettelbeck et al., 2007, 6). The LASP2 authors focused more on the

complexities of collecting accurate and useful data on comparative
L&C enrolments in Australian universities, and recommended that ‘the
university sector … work towards a uniform and nuanced definition
of what constitutes attrition, and that the relevant faculties generate
and make readily available comparative statistics about attrition in
languages and other humanities and social sciences areas’ (Nettelbeck
et al., 2009, 6). This crucially important issue is one that we address in
depth in Chapter 2, in the context of the ANU case study, where we
report on an innovative approach to calculating retention rates based
on detailed institutional and collected data.

1.9. Attrition at the course level: Risk factors

So far, the findings described in this chapter have largely focused on
research at the university or program level. For students, of course,
the decision to discontinue their formal L&C studies generally occurs
during a specific course. For this reason, Lobo and Matas (2010,
2011) looked specifically at attrition at the level of an individual L&C
course (i.e. students withdrawing from a course before its end), by
investigating the reasons why students withdrew from a first-year ab-
initio (Beginner/Level 1) Spanish course at an Australian university.
The authors’ approach was prognostic/remedial rather than diagnostic/
explanatory: instead of aiming to explain attrition from the perspective
of the students who had left the course, the authors attempted first
to identify students at risk of withdrawing, and then to provide an
intervention that would reduce that risk (Lobo and Matas, 2010, 155–
161; 2011, 305–306). Through an extensive review of the Australian
and international literature on attrition and retention focused on
first year students, the authors identified 17 key factors known to
influence the likelihood of a student not completing a course (Figure
1.2, which can be seen to be an extension of Figure 1.1; Lobo, 2012).
The authors next developed a student ‘risk’ questionnaire based
on these factors, phrasing questions such that responses could be
scored, and totals ranked, to provide cut-off values that identified an
individual student’s risk of withdrawing before the end of the course
in one of three categories (‘very little risk’, ‘fair risk’ and ‘high risk’).
To add a qualitative dimension to the scoring, some students were also
interviewed.

22

This content downloaded from

217.64.17.122 on Wed, 26 May 2021 07:41:05 UTC
All use subject to https://about.jstor.org/terms
 1. Understanding the Problem

1. Students’ expectations and perceptions of university life and study (of the course,
degree or programme, the people and the university itself)
2. Social and academic student integration
3. Teaching and learning styles
4. Assessment strategies used in courses
5. Lack of student mentoring
6. Students’ living arrangements (on campus, with friends, at home, among others)
7. Student age
8. Student gender
9. Work and issues with employment
10. Financial concerns
11. Student lack of preparation for university life and study
12. Family responsibilities and obligations
13. Dissatisfaction with the university
14. Academic difficulties
15. Health and personal reasons
16. Course or Program unsuitability
17. Learning anxiety (in particular foreign language learning anxiety)

Figure 1.2. Factors associated with course attrition, as identified from

the literature
Source: After Lobo and Matas (2010, 14–40).

With this knowledge of the risk factors of attrition, and the

information from their students’ risk questionnaires, the researchers
then developed a two-page ‘First Year Student Guide’ specifically
designed ‘to facilitate the social inclusion and academic connection
of [each] student’ (Lobo and Matas, 2011, 311). This guide, which
provided students with relevant information about how to study and
the university’s support services, was given to students after they
had completed the risk questionnaire, and then followed up with
class discussions half-way through the semester. Remarkably, given
the 85 withdrawals from the previous year’s cohort, no student from
the ‘Guide’ cohort withdrew from the course. From this outcome and
student feedback, the authors argue that their approach to maximising
retention was successful, although they acknowledge that both survey
and guide require validation with a larger sample (Lobo and Matas,
2011, 312).

23

This content downloaded from

217.64.17.122 on Wed, 26 May 2021 07:41:05 UTC
All use subject to https://about.jstor.org/terms
 The Doubters' Dilemma

1.10. Learning anxiety as a specific risk

factor in language learning
Although the literature reviewed by Lobo and Matas (2010) suggests
many generic reasons why a student may discontinue his or her
enrolment in a first year course (Figure 1.2), three factors appear
particularly relevant to language learners. These are i) the perception
of a high (and higher than expected) workload in the course;
ii) the student having ‘less serious’ reasons for enrolling in the course
(for example, thinking that language learning would be ‘fun’, and
so being less prepared for the realties of workload and assessment);
and iii) the important notion of students being burdened by ‘foreign
language learning anxiety’ or ‘second language anxiety’ (Lobo, 2012).
While learning anxiety is by no means unique to languages (it has also
been identified as a problem faced by students of mathematics and
science, capable of negatively impacting on performance—Ashcraft
and Kirk, 2001; Ma and Xu, 2003; Sherman and Wither, 2003; Nunez-
Pena, Suarez-Pellicioni and Bono, 2013), it is notably the only risk
factor specific to the language-teaching context (Lobo, 2012, 207).
‘Second language anxiety’—defined as ‘the feeling of tension and
apprehension associated with second language contexts, including
speaking, listening and learning’ (Onwuegbuzie, Bailey and Daley,
1999, 222), and closely linked to performance in oral examinations or
other forms of language production in the classroom—is considered
one of the major factors in foreign language attrition (e.g. Horwitz,
2010; Horwitz, Horwitz and Cope, 1986; Scovel, 1978; Wesely, 2010).
Lobo and Matas (2010, 127) found language anxiety—primarily
derived from concern about an oral interview assessment task—was a
key factor reported by students who withdrew from the Spanish L&C
course under study. This anxiety was associated with perceptions that
other students were better at languages, and that the classes were too
fast-paced, perceptions that all contributed to a feeling of inadequacy
in class (Lobo and Matas, 2010, 102–110). However, the situation is
not simple: some students who reported anxiety related to speaking,
listening, and especially the oral interview nevertheless persisted in
their studies, to the admitted bemusement of the researchers (Lobo and
Matas, 2010, 124).

24

This content downloaded from

217.64.17.122 on Wed, 26 May 2021 07:41:05 UTC
All use subject to https://about.jstor.org/terms
 1. Understanding the Problem

It may appear somewhat surprising that the relative level of learning

anxiety among individual students is not an effective predictor of
the likelihood of those students continuing or discontinuing their
language study. The ANU findings give some insights to the relevant
differences in student characteristics that may in future provide this
kind of predictive capacity for teachers: these findings also suggest
why the Student Guide produced by Lobo and Matas (2011) had its
excellent outcome in reducing attrition.

1.11. Asking the difficult questions: Attrition

as a research problem
We began this chapter with a list of questions about student
attrition that concern university L&C teachers. As we have seen,
some significant attempts have been made in recent years to tackle
these questions (e.g. Nettelbeck et al., 2007, 2009; Lobo and Matas,
2010,  2011). While the relevant research is still far too limited to
provide trustworthy answers across the sector, it does show that raw
enrolment numbers provide extremely crude measures of attrition
in university L&C programs. Nevertheless, such raw data are still
commonly used by administrators and budget planners as surrogate
indicators of a program’s success, and, critically, may be used as
performance indicators to guide the allocation of institutional and
government funding, as occurred from 2006 to 2009 nationally with
the Learning and Teaching Performance Funds (Gabb et al., 2006).
The need for a more effective approach to calculating and
understanding retention in L&C programs was crystal clear to L&C
teachers at ANU in the mid-2000s. The apparently low retention rates
being experienced across the 18 ANU L&C programs on offer at that
time—especially in terms of students discontinuing after Beginner
level—was a key discussion point among teaching staff. This led
to the establishment of an internally funded research program to
explore in detail the best ways of calculating and comparing student
retention rates, and the motivations of students in making decisions
about their L&C studies, using ANU as a case study. This was not
just an opportunistic choice of institution, but a strategic one: ANU
has long had a tradition of teaching many L&C programs, and was
not only identified as teaching the greatest diversity of languages

25

This content downloaded from

217.64.17.122 on Wed, 26 May 2021 07:41:05 UTC
All use subject to https://about.jstor.org/terms
 The Doubters' Dilemma

of any university included in the LASP1 research (Nettelbeck et al.,

2007), but also as the only university teaching many of Australia’s less
commonly taught languages (Dunne and Pavlyshyn, 2012).
In designing the case study, the research team was mindful of two
perspectives that had not yet been voiced, but have subsequently
been well expressed by other researchers. First, we understood the
overall complexity of researching L&C education. In the context of a
review of languages in the school sector, Scarino (2012, 244) explained
this complexity:
In the Australian context of languages education, descriptions that
do not take into account acknowledged differences across languages,
across groups of students with diverse linguistic and cultural
backgrounds and affiliations with the target language, and across
program conditions such as time-on-task, are too generalised to be
meaningful and of value to the diverse users.

We also understood that this complexity would be increased, rather

than lessened, with our focus on retention.
Secondly, we understood, in the more recent words of Hanley and
Brownlee (2013), that ‘investigating questions of attrition and student
motivation in tertiary language programs is not simply a matter of
asking students why they do or do not continue in their area of study’.
We thus approached the concept of student retention in university
L&C programs not just as an educational issue but also as a social
phenomenon, and thus deliberately set out to collect data on our
students’ social characteristics through institution-wide surveys of
L&C students in 2008 and 2009, with our approaches designed to
maximise response rates, and—again as Hanley and Brownlee (2013)
later reinforced—to avoid inadvertently allowing results to be skewed
by methodological defects.
The research was thus designed as a highly structured case study
that would generate valid and reliable empirical data to enlighten our
understanding of attrition and retention in the context of a specific
institution, while also suggesting appropriate methodologies for
future studies on a broader scale.

26

This content downloaded from

217.64.17.122 on Wed, 26 May 2021 07:41:05 UTC
All use subject to https://about.jstor.org/terms
 1. Understanding the Problem

Our approach was strongly influenced by Bourdieu’s concluding

comments in the methodological appendix to his seminal work,
Distinction:
The epistemological obstacles which social science has to overcome
are initially social obstacles. One of these is the common conception of
the hierarchy of the tasks which make up the sociologist’s job, which
leads so many researchers to disdain humble, easy yet fertile activities
in favour of exercises that are both difficult and sterile. Another is
an anomic reward system which forces a choice between a safe thesis
and a flash in the pan, pedantry and prophecy, discouraging the
combination of broad ambition and long patience that is needed to
produce a work of science. Unlike the sometimes illuminating intuitions
of the essay form, the sometimes coherent thesis of theoreticism and
the sometimes valid observations of empiricism, provisional systems
of propositions which strive to combine internal coherence and
adequacy to the facts can only be produced by a slow, difficult labour
which remains unremarked by all hasty readings. These will only see
repetitive reaffirmations of theses, intuitions or already known facts
in the provisional conclusion of a long series of totalizations, because
they ignore what is essential, namely, the structure of the relations
between the propositions (Bourdieu, 2010 [1984], 513).

We thus sought to collect comprehensive data that we could subject

to patient analysis as we sought ‘internal coherence’. In addition, our
work focused on students being at the heart of language teaching,
and their experiences and ‘individual differences’ (Dörnyei, 2005) as
learners being at the heart of language-teaching research. In so doing,
we hoped to advance the ideas on student motivation presented by
Joe Lo Bianco, Professor of Language and Literacy Education at the
University of Melbourne Graduate School of Education:
Ultimately language learning is the preoccupation of individual
students, in the same way as language teaching is the preoccupation of
language teachers. In recent policies, written with the hand of diplomats,
trade officials and other elites, there has been far less consideration of
the practical issues involved in schooling, and therefore a tendency
towards stressing accountability and imposition of numerical targets,
with less focus on capacity-building, acknowledgment of the learner
population, issues of motivation, resource constraints, personal
aspirations, experiences and motivation, identity issues and family
background. All too often it is assumed that the motivations learners
have available to them are the prospects of employment and other
material advantage that attach to language learning.

27

This content downloaded from

217.64.17.122 on Wed, 26 May 2021 07:41:05 UTC
All use subject to https://about.jstor.org/terms
 The Doubters' Dilemma

This outsider perspective on motivation is less tenable today in light

of the powerful shifting of emphasis towards the internal perspective
and experience of learners, and on the quality of micro-school
experiences in influencing motivation, persistence and interest among
language students … This research is important to language education
planners because it shows that even in the face of negative attitudes
students might inherit from the wider society, or from their parents,
about languages being unimportant, or that ‘everyone speaks English’,
micro-motivation effects (good teaching, concrete perceptible sense
of achievement, success) can override negativity and sustain student
interest. Here policy is practice, in the hands of individual teachers
and schools (Lo Bianco, 2009, 27).

Accordingly, rather than to present any particular policy or strategy

solution ourselves, our focus in reporting the ANU case study research
in this book is to present detailed data and analytical methodologies
that we believe will be highly relevant to the development of any
future evidence-based language policies intended to increase student
participation and retention in university L&C programs.

1.12. A reader’s guide to this book

The ANU case study had three data-related aims. First, it sought to
document the nature and rates of student retention and attrition
in ANU L&C programs, and, for comparative purposes, in other
discipline areas taught at ANU. Second, the study sought to explore
ANU students’ motivations for, and experiences of, studying a
language at university. In particular, the researchers investigated
students’ motivations for continuing, discontinuing, or thinking
about discontinuing/deferring their language studies. Third, the
case study was designed to identify the incentives or disincentives
that influence students either to continue language studies to the
completion of an undergraduate degree major,4 or to discontinue
those studies before completing a major. In addition, the researchers’
awareness of the potential for generalisations from the institutional
findings to inform future sector-wide policies related to increasing the

4 At the time of this research, an ANU undergraduate student seeking to complete an L&C
major had to complete either seven or eight courses, usually at a load of one course per semester.
Students aiming to complete a major in three years usually enrolled in additional L&C courses in
their final semester of study (generally the second semester in any given year).

28

This content downloaded from

217.64.17.122 on Wed, 26 May 2021 07:41:05 UTC
All use subject to https://about.jstor.org/terms
 1. Understanding the Problem

rates of student retention in L&C higher education programs led to a

realisation, during the analyses of the collected data, that what was
required was a reconceptualisation of the concept of attrition with
regard to L&C programs.
In presenting the methodologies, analytical processes and outcomes
of the ANU case study, therefore, this book positions the significant
findings of this single-institution research in the broader context of
retention and attrition of university language students Australia-
wide, and suggests some implications of these findings in terms of
future research and policies. Through Chapter 1, readers should now
have a good grasp of some of the key issues relevant to research into
attrition in L&C programs in Australian universities. Chapter 2 takes
readers into the contentious world of calculations, as we explore how
meaningful retention rates can best be computed, and compare rates for
L&C programs with those of other disciplines. The chapter introduces
a novel and comprehensive approach to calculating retention rates
from student data that universities already collect. Using ANU data,
we will show how this approach negates the potential distortions
of having several cohorts active at the same time, while respecting
the specific and unusual nature of L&C enrolments and allowing fair
comparisons with other disciplines.
In Chapter 3, the reader will meet the ‘Doubters’ of the book’s title
for the first time, as we explain the rationale and detail of the single-
institution research methodology, and the analytical approach we
adopted in dealing with the student survey data. The crucial impact
of this approach was that it demonstrated unequivocally that the simple
dichotomous classification traditionally used in retention studies—
that is, comparing those who continue studies (Continuers) to those
who stop studies (Discontinuers)—did not explain the core issues at
the heart of the discussion of retention in L&C programs. Even with
in-depth statistical analyses, this simple dichotomous classification
did not provide satisfying explanations of attrition, because we simply
could not find statistically significant differences between Continuers
and Discontinuers in terms of students’ background, motivations,
perceptions or behaviour related to their L&C studies. The chapter
takes readers on the researchers’ journey in seeking a new, data-based
approach to the grouping of students. The reader will here meet the
four descriptively named groups into which students were clustered
in terms of their characteristics—Committed Students, Doubters,
29

This content downloaded from

217.64.17.122 on Wed, 26 May 2021 07:41:05 UTC
All use subject to https://about.jstor.org/terms
 The Doubters' Dilemma

Reluctant Quitters and Voluntary Quitters—and understand why the

nature of the ANU data required a merging of the latter two categories
into an inclusive grouping of Quitters (essentially equivalent
to ‘Discontinuers’).
In Chapter 4, the three functional groupings—Committed Students,
Doubters and Quitters—come to the fore, as we explore how they differ
across a range of demographic, attitudinal and educational variables.
As we describe the intergroup differences between the three, we show
just how different are their reasons for continuing or discontinuing
L&C studies.
In Chapter 5, we take this characterisation of the three student
archetypes further, as we interpret the empirical research findings in a
way that provides an overarching explanation of the differences among
the groups. This is where we develop our argument based around the
construct of ‘language capital’, and show how the students categorised
as Doubters are easily identifiable as the students that other retention
studies classify as those most ‘at risk’ of discontinuing. The chapter
also includes an exemplar discussion of students’ perceptions about
learning spoken language as an illustration of the capacity of the
language capital construct to explain the empirical findings commonly
found in studies of L&C students.
Chapter 6 brings the book to its conclusion by presenting an overview
of the findings from the single-institution case study, and suggests
ways in which the methodologies and the construct of language capital
could benefit researchers and those developing language policies
in the future.

30

This content downloaded from

217.64.17.122 on Wed, 26 May 2021 07:41:05 UTC
All use subject to https://about.jstor.org/terms