Problems With Assessing Student Autonomy in Higher Education , an Alternative Perspective and a Role For Mentoring

The paper explores, from a conceptual basis, the inherent tensions in assessing student autonomy in higher education. The author argues that, despite the development of student autonomy being a key aim of higher education, there are problems in identifying with any level of precision what it is, and that its assessment is highly problematic. An alternative perspective is provided that, with the support of mentoring processes, allows for authentic assessment. The paper is intended to stimulate debate amongst university management and academic practitioners in higher education. This is a conceptual paper considering the problematic nature of learner autonomy and the inherent difficulties in assessing it, with a practical potential


Introduction
One of the espoused aims of higher education programs throughout the world is to develop undergraduate student autonomy.Autonomy has been an educational aim for many years, perhaps as far back in time as the Roman civilization (Boud, 1988).Over 20 years ago, Littlewood argued that autonomy must be a goal of education, because no learners will ever have their teacher available to accompany them throughout life (Littlewood, 1996), whilst Candy identified that it was: "almost universally proclaimed as a goal of education" (Candy, 1991, p. 119).Learner autonomy has always been important, and has become a central pillar of education in recent years (Viera, 2009).In today's mass higher education system, where students experience larger class sizes than in the past, where digital technologies and online learning replace face-to-face classroom teaching, and "flipped classrooms" are used with an increased emphasis on independent or "personalized" learning, it is even more important (Cao, 2012;Kormos & Csizer, 2014;Moore, 2016).
Yet "learner autonomy" is a problematic concept, it is multifaceted, being both a process and a product of education.It is frequently confused with independent learning, and may be extremely difficult to assess authentically (Everhard, 2011).It may also be difficult to precisely articulate what it is, other than in a generalized way.Yet it is a central goal of education, and one of the ultimate goals of higher education (Bajrami, 2015).As an educational aim, it is overarching and all-encompassing, of a greater magnitude than other aims of higher education.Almost 40 years ago, Boud argued that it was: "not just one goal among many but rather a characteristic of all of the others: it is the manner in which all skills should be displayed and all beliefs held" (Boud, 1981, p. 25).This position is still just as valid today.Many university program specifications and learning outcomes clearly specify autonomy as being an essential graduate attribute (Channock, Clerehan, Moore, & Prince, 2004) and claim that it is being developed and assessed.Yet it is the author's opinion that, in practice, it would seem to rarely be assessed in the same way that other learning outcomes are.
This paper provides a critical overview of the problematic nature of learner autonomy and its assessment and offers a potential solution through mentoring processes, which may allow universities to justifiably claim they develop student autonomy.In recognition of the diversity of higher education establishments and structures, the paper does not seek to describe a framework for the specific role of a mentor, but to make practical suggestions as to how such roles could operate within higher education.

Confused terminology
The study of autonomy has been approached from a range of disciplinary perspectives, with much of the literature originating from language education and language learning.The concept of autonomy in learning is not a simple one (Benson, 2011;Boud, 1988) and in a similar vein to terms such as "critical thinking," can mean many things to many people.It is not a simple, or easily described behavior (Dam, 2003;Little, 1991).Although there is common understanding amongst educators as to what learner autonomy in general is, there is a lack of consensus as to precisely what it is.The terms: autonomy, independent learning, autonomous learning, and independent study, are often used interchangeably in the literature to describe the same thing.Definitions and explanations of autonomy often include independence (e.g.Little, 1988), and vice-versa, definitions of independent learning EDUPIJ • Volume 7 • Issue 1 • 2018 often include autonomy (e.g.Moore, 1973).Autonomous learning is not always clearly defined (Macaskill & Denovan, 2013).And there is a lack of consensus as to what independent learning means (Broad, 2006).As Scott, Furnell, Murphy, and Goulder (2015) argued: "There exists a myriad of inter-related definitions of autonomous/independent learning" (p.946).

Independent learning
Independent learning and independent study are usually described as involving "selfregulated learning."Here learners are regarded as having an understanding of their own learning, of being motivated to take responsibility for it and where the learning environment is structured by a teacher (e.g.Gorman, 1998;Perry, Phillips, & Hutchinson, 2006).Candy, in his seminal text "Self-direction for lifelong learning," argues that independent learning and study are a process, a method and a philosophy of education in which a student acquires knowledge by his or her own efforts and develops the ability for inquiry and critical evaluation (Candy, 1991).This includes the learner's freedom of choice in determining those objectives, within the limits of a given project or program of study and importantly, with the aid or support of a tutor.For the purposes of this paper, we may regard a mentor as playing a similar role to that of a tutor in providing support, advice, guidance, and to a lesser extent, in structuring the learning process.Yet the roles differ in that a mentor is not directly involved in teaching.Further information about how the role could operate in practice is provided later.
Independent learning requires freedom of process for learners to be able to carry out their own objectives.This places increased responsibility on the student for achieving the learning objectives or outcomes (Forster, 1972, as cited in Candy, 1991).There are two key points to note.Firstly, that there are parameters i.e. the limits or boundaries of the program of study.Secondly, that the tutor, or mentor, provides support for the independent learner.Responsibility for learning is placed on, and lies with, the student.

Autonomous learning
Autonomous learning has a range of broadly similar definitions (e.g.Benson, 2011;Dickinson, 1993;Holec, 1981).It is variously depicted as learners taking charge of their own learning (Holec, 1981), involving learning how to learn and learners' attitudes and attitudinal values (Dickinson, 1993), and about learners' modes of learning (Benson, 2011).The most frequently cited definition is Holec's which identifies it as "The ability to take charge of one's learning" (Holec, 1981, p. 3).Little (2007) argues that the ability to take charge of one's own learning is the "single common thread" that runs through the literature (p.15).Holec's definition is a broad one.A more comprehensive picture is provided by Candy, who describes six aspects of autonomy.These are that: 1) the learner has freedom of choice; 2) can develop goals and plans independently of pressure from others; 3) a capacity for reflection; 4) the will and the capacity to "fearlessly and resolutely to carry into practice, and through to completion, plans of action…without having to depend on others for encouragement and reassurance"; 5) the exercise of self-mastery; and 6) a personal concept of him or herself as being autonomous (Candy, 1991, p. 109).Similarly, Moore defines it as "The will and ability to exercise powers of learning, to overcome obstacles for oneself, to try to do difficult learning tasks, and to resist coercion" (Moore, 1973, p. 667).Although these definitions help in identifying the nature of autonomy, they do not lend themselves easily to EDUPIJ • Volume 7 • Issue 1 • 2018 use for the purposes of assessment.Using them to articulate clear assessable learning outcomes within the credit-bearing frameworks used in universities is problematic.How, for example, could it be determined with any level of authenticity whether a person had acted fearlessly and resolutely, or that a power of learning had been exercised?
In discussing autonomy as being an approach to learning, Boud identifies that its main characteristic is that students take: "significant responsibility for their own learning over and above responding to instruction" (Boud, 1988, p. 23).This captures the essence of what learner autonomy is, though individuals will necessarily have somewhat differing interpretations of the word "significant."The statement could, however, equally well serve to define that of the independent learner.In both cases the learner is seen as exercising greater personal agency.A more comprehensive definition is provided by Little (2000), conceptualizing it as a psychological capacity, which a learner may, or may not, decide to exercise: Autonomy…depends on the development and exercise of a capacity for detachment, critical reflection, decision making and independent action: autonomous learners assume responsibility for determining the purpose, content, rhythm and method of their learning; monitoring its progress and evaluating its outcomes.(p.69) Little's view of autonomy provides us with a very useful distinction between autonomy and independent learning.Using it we may argue that independent learners do not determine the content, or purpose, of their learning.Their learning is structured and facilitated by a tutor, working within a defined structure of the curriculum and its assessment tasks.Learning lies within the specified parameters of their program of study, as determined by disciplinary boundaries, the educational institution and its structures, and those of external quality agencies such as, in the United Kingdom for example, the Quality Assurance Agency (QAA).
An interesting perspective is provided by Moore (1973), who has written extensively about independent learning (though in the context of distance learning where the students are separated from the tutor, without face-to-face teaching).He argues that: For the non-autonomous learner…the teacher's role is that of a director of learning and the learner's is to respond to the teacher's directions.The teacher tells the learner what is to be learned, how it is to be learned and when it has been learned.(p.669) If we turn this statement around and say that where the teacher does not tell the learner what to learn, nor how to learn it, nor by when it should be learned, then we may have a working definition that allows us to identify differences between the autonomous learner and the independent learner that may be used to inform assessment processes.Therefore a working definition for autonomous learning proposed by the author is that the autonomous learner has freedom of choice in determining what to learn, how they should learn it, and by when they should have learned it.Doing this allows us to identify that the autonomous learner is able to determine the what, the how, and the when of their learning.It follows that a mentor may have a key role to play in facilitating the development of autonomy through their advice, guidance and support in helping learners identify the what, how and when.This would necessarily be a different process to the instruction that may be EDUPIJ • Volume 7 • Issue 1 • 2018 provided by a tutor.One practical issue here though is that the extent of the mentor's role would need to be determined on an individual student basis, and university resources may not be sufficient for this.

Is autonomy innate, or does it have to be taught, or learned?
There is a tension within the literature between two views.One regards autonomy as something that students need to be taught, or to learn; the other regards it as pre-existing, something that all learners have and are able to exercise to some extent.Holec's work, for example, a commissioned report to the Council of Europe, often cited as a key text, supported the former view, arguing that learner autonomy was not inborn, but must be acquired through formal or informal learning (Holec, 1981).Holec's definition was written in the context of the Council of Europe's work on the theory and practice of adult education in the context of language learning (Little, 2007), it was not specifically aimed at undergraduate students.Moore, on the other hand, argues that it is innate and related to a person's state of development; that at some point a person "acquires" autonomy (Moore, 1973).This tension would seem to be not dissimilar to the nature-nurture debate, with no definitive satisfactory answer currently, perhaps ever, being available.Yet it may be inferred from both positions that autonomy may be linked to a learner's level of maturity.
It is the author's experience of working with higher education undergraduate students that some level of learner autonomy is innate, but that it can, and should, be developed; moreover, that facilitation of the development of autonomy may be most effective when enacted through a supportive and structured relationship.Yet, academic staff often have insufficient time to provide this support and students may have little out-of-class contact with tutors, particularly so in modularized university programs where they may be taught by different staff for each module studied.A formal mentoring system utilizing dedicated staff could be a solution to this problem, and is discussed later.

Autonomy, product or process?
Little argues that autonomy comprises two elements; the complexity of knowledge and skills that constitute the aim of learning, and the processes by which they are acquired.
In the domain of formal learning, autonomy entails both the capacity to apply the knowledge and skills learned in the classroom to appropriate contexts in the world beyond the classroom, and the capacity to update that knowledge and those skills in response to the demands of changing circumstances.The explicit pursuit of autonomy in formal educational contexts is thus a matter both of learning and of learning how to learn.(Little, 1997, p. 94, emphasis added) This quote illustrates the differing attributes of learner autonomy.Some (e.g., Candy, 1991;Thanasoulas, 2000) argue that it is a process, not a product and that a learner does not actually become autonomous, only that they work towards autonomy.Others, however, regard it as being both a process and an outcome.It is advocated as both a means to an end, and as an end in itself.This is problematic for educators, assessors and for learners, as it may be unclear which of these is the aim.Is it what Boud (1988) labels as being a productorientation; that is, producing an autonomous person?Or, is it a process orientation, i.e. introducing activities to the teaching and learning process which require students to act EDUPIJ • Volume 7 • Issue 1 • 2018 autonomously (e.g.Dam, 1998)?Or is it, as Holec (1981) argues, both?For the purposes of this paper, the author's position is that it is about both and that the latter could facilitate the former.Unfortunately, a process-oriented approach does not necessarily, or automatically, lead to a satisfactory product outcome.A capacity for behavior is not the same as demonstrating that behavior.A student may believe that they are, and that they act as, an autonomous learner, yet in practice their behavior may exhibit few signs of this.The mentor could have an important role to play here in supporting the learner to help them identify differences between what they believe, what they say, and what they do, or helping them acknowledge, and develop strategies to challenge, their own confirmation bias.

Autonomy as a "wicked" competency
Learner autonomy may well be an example of a "wicked" competency."Wicked" competencies are achievements that cannot be neatly pre-specified, take time to develop and resist measurement-based approaches to assessment; in that there is a lack of agreement about what is being assessed (Knight & Page, 2007, p. 2).Knight (2007) argues that many problems in life are "wicked," they resist definition, shift shape and are never "solved" as such.Wicked competences though, do not easily lend themselves to being assessed (Knight & Page, 2007).There is, therefore, an argument that it may be more appropriate to make an assumption that curricula and good learning and teaching arrangements are likely to promote their development and achievement.Yet, in today's mass marketized higher education system, it is unacceptable to stakeholders for universities to merely argue that their good teaching and learning is likely to lead to the development of autonomy; proof is required.Yet should a graduate employer request proof from a university that a prospective employee was an "autonomous learner," could an institution provide such evidence?Of course, doing so may be difficult because assessing learner autonomy is problematic.
There are questions that may be asked about the nature of learner autonomy.Is it a specific skill, or a set of skills?Is it a competence, an attitude, or an understanding?Is it a disposition, a behavior or set of behaviors?Or, is it some combination of any, or all of these?The author's position is that it is an "essentially contested concept" (Gallie, 1956) i.e. something that is impossible to conclusively define, but perfectly possible and rational for people to discuss and justify their holding of one interpretation rather than a competing one.Yet, if we acknowledge that autonomy may not be possible to conclusively define, and that there are different interpretations, then we necessarily will have difficulty in assessing it.This does not imply that identifying the attributes and characteristics of autonomy is impossible, nor that these could be not be developed for assessment purposes.There are, however, a number of challenges in assessing autonomy.

Challenges in assessing learner autonomy
One of the problems both with assessing, and helping to develop learner autonomy, is that it does not comprise a single, easily described, behavior or set of behaviors (Little, 1991).It is a multi-dimensional construct (Benson, 2011) with multiple and frequently unclear meanings.There are many differing interpretations of autonomous learning.Macaskill and Taylor (2010) argue that "there does not seem to be a single consensual definition" and that many academic journal articles "appear to discuss autonomous learning without defining exactly what they mean by it" (p.351).Ecclestone (2000) argues that it is EDUPIJ • Volume 7 • Issue 1 • 2018 genuinely difficult to articulate with any level of precision what the criteria and outcomes are for autonomy, particularly so for higher levels of learning.This causes tensions for educators, making the assessment of autonomy highly problematic for the tutor/assessor.Defining and clearly describing autonomy is also problematic because it may manifest in many different forms, at different ages, and in different situations (O'Leary, 2007).It is thus difficult to clearly articulate, and ergo, to assess.It is a "slippery concept" (Ecclestone, 2007, p. 321).
One of the challenges of assessing autonomy is that, because it is something exercised by the learner, autonomous behavior will be manifested in different learners, in different forms, and at different stages of their education (O'Leary, 2007).It will depend on, for example, age, prior learning experiences, perceived learning needs, and the student's current learning context.This implies different forms of assessment may be required, for different learners, at different stages.This is highly problematic, because unless program and module specifications are able to clearly articulate what different forms of autonomy are, its attributes and how these may be assessed at different stages, then they may not be able to be assessed authentically in a way that could be graded.However, the argument for assessing autonomy is that, in addition to stakeholder requirements, if we do not, then, as the literature on assessment would suggest (e.g., Gibbs, 2006;Race, 2014;Torrance, 2012), some students may not regard it as being important and not strive, or make any effort, to become autonomous.
Unfortunately, if we make it clear to students that their autonomy is going to be assessed this may create further problems.Benson argues that measuring autonomy is, of itself, "problematic," that autonomous behavior should be initiated by the student, rather than in response to an assessment task (Benson, 2001).This creates a serious challenge, because if a tutor/assessor explicitly identifies to learners that they are assessing their autonomy some will start to demonstrate behavior that they believe the assessor will perceive as being autonomous, although it will not be genuinely autonomous behavior.Almost by default, as soon as it is indicated to learners that their autonomy will be assessed, an assessor may be unable to do so with any authenticity.Notwithstanding the general argument that students may change their behavior once they know they are being observed and assessed (Rowntree, 1987), this raises a question about universities' espoused aims of developing autonomous learners.If lecturers are unable to assess genuine autonomy, then should universities make that claim?It is perhaps unethical to do so and may be a falsehood.As Benson argues: "If we aim to help learners to become more autonomous, we should at least have some ways of judging whether we have been successful or not" (Benson, 2001, p. 54, emphasis added).
How are assessors to judge success, if, as soon as they try to do so, it may no longer genuine autonomous behavior that is being assessed?One potential solution to this problem could be the information provided by a mentor who has worked closely with a learner over a period of time.They should be in a much better position, able to recognize genuine behavior, that is repeated and developed over a period of time, than an assessor who may only have one or two opportunities to do so at a specified point of assessment within a program of study, or a lecturer who has taught students for only one module.

Structural challenges in assessing autonomy
Institutional structural factors may prevent the development of autonomy.Students are required to display their understanding of a subject, within the confines of a curriculum determined by national quality assurance agencies, the individual higher education institution, program specifications with pre-specified learning outcomes and assessment criteria, all of which are mediated by an academic tutor.A learner may wish to exercise their autonomy but has to demonstrate that they have achieved a module's intended learning outcomes through meeting pre-specified assessment task(s).For some students though, prespecified learning outcomes may act to constrain learning (Sadler, 2007) as they may believe that, once an outcome has been achieved, that no further learning is necessary.
In almost all university programs the learner is unable to decide when something has to be learned by, as the tutor or the institution stipulates assessment submission deadlines.If the development of learner autonomy is a goal, as universities universally proclaim, then, arguably, the curriculum structure and pedagogical process should provide for, and allow, individual students to be able to negotiate what assessment evidence they will produce, and allow them to decide when to do so.In practice, learners, particularly in the latter stages of a degree, are often able to negotiate the assessment evidence they will produce, yet they are rarely able to discuss and negotiate submission deadlines, i.e. the "when" something has to be learned by.It follows that university educators may be unable to genuinely make a claim that a student is truly autonomous when studying any credit-bearing module or unit of study leading to a formal qualification, because the learner has not been able to decide "when" they have had to learn something by.How then may higher education institutions legitimately claim that their students are developing autonomy?There may be a solution available, using a framework for learner autonomy that has been developed, and a mentor could have an important role to play.

A possible framework for learner autonomy
In her study of advanced vocational educational learners, Ecclestone (2000) proposed a framework whereby learner autonomy can be Procedural (technical) through a behaviorist transmission model of learning, Personal (personal practice) through a transaction model of learning, and Critical Emancipatory through a transformation model of learning (Ecclestone, 2000).Ecclestone argues that autonomy can take the form of a learner being more proactive within an existing predefined system of rules.In this situation, the autonomy is predominantly procedural or technical.Personal critical autonomy may be difficult to develop without having what, initially, may be rigidly defined program, or module assessment specifications.Yet, if the specifications are too tightly defined this may hinder its development.For learners to develop a command of procedures and systems, as a precursor to greater autonomy, they may need to develop a confidence and a willingness to take risks.It is the student's confidence gained through mastering technical predefined tests that may allow them to develop further confidence, skills, and autonomy.It follows that "procedural autonomy may actually be a prerequisite or a co-requisite for more sophisticated forms of autonomy" (Ecclestone, 2000, p. 148).The author believes that the implication here is that pedagogical processes, can develop autonomy, without specifically assessing it per se.This is important, as it allows for and recognizes that, learner autonomy can be developed without it having to be assessed.

EDUPIJ • Volume 7 • Issue 1 • 2018
There is a clear role here for both structured, and less formal, mentoring processes, specifically, in supporting and guiding the learner and facilitating the development of their confidence through the process of acting as an intermediary between, and interpreter of, rigidly defined institutional and program requirements, and in helping to identify and define the limits for a procedural or technical specification.Supportive developmental mentoring processes, which are gradually withdrawn over time, would help students in moving from procedural, to personal, to critical emancipatory learning and greater personal autonomy.Implementing processes such as these would allow universities to legitimately claim that they were facilitating the development of learner autonomy.Yet, the question could still be raised by stakeholders that if autonomy was not specifically assessed, then how may a university educator prove that learners are becoming autonomous?

A possible solution to the assessment problems
There is a position developed by Boud (1981) that autonomy is not an absolute standard to be met, but is a goal to be pursued and that what is important, is the direction towards student responsibility for learning, not the magnitude of change.Building on earlier work by Lewis (1978), he argues that the: "only realistic goal for higher education is that students should be more autonomous when they leave a course than when they enter" (Boud, 1981, p. 24, emphasis added) not that they have reached an arbitrary point that can be measured.This is an interesting position, implying that we should not focus on evaluating the extent to which a student has become an autonomous learner, but, on whether they are becoming one.Using this perspective would be a pragmatic approach that potentially provides a realistic and achievable way of overcoming some of the many difficulties associated with assessing learner autonomy.If the focus is on whether a learner is in the process of becoming autonomous, then assessment processes would not have to demonstrate that a learner had become autonomous, only that they were in some way "more autonomous" when they left a program of study compared with when they started.Using such a focus would allow universities to legitimately claim that they were fulfilling one of higher education's central aims.
Three in-practice tensions, however, arise from this.Firstly, that universities may need to measure each learner's autonomy when they commence a program of education, so that it may later be demonstrated that they are more autonomous when they graduate, and as already discussed, doing so is problematic.Secondly, assessment, for summative purposes, focuses on student achievement of pre-specified intended learning outcomes.Learners are assessed, graded and awarded academic credit based on their demonstration that they have done something, (i.e.met a learning outcome), not on whether they are "moving towards" having done something (i.e. in the process of meeting a learning outcome).It may require a considerable change to current assessment practices, and external-body quality criteria, for higher education institutions to realistically assess whether a learner was "becoming" more autonomous.Certainly, for grading purposes, this would be challenging, but not impossible.Yet, if moving towards autonomy is not assessed, then some students may not see it as being important (Boud & Falchikov, 2006).Thirdly, within any modularized program with discreet modules taught by different staff, a student could demonstrate that they have become "more autonomous" in a first-semester module, but then regress and make no further progress in semester two.Yet they may still be able to demonstrate, to a different tutor/assessor, that they were more autonomous at the end of the second semester than EDUPIJ • Volume 7 • Issue 1 • 2018 they were at the start; even though their overall autonomy was no higher than at the end of the first semester.One way of addressing this would be at program level, perhaps through annual student progress meetings that could identify if the student had demonstrated progress during the whole academic year.A role here for a mentor, working across both semesters, could be to provide some form of evidence that the student had become more autonomous overall.

An alternative perspective and potential solution
The author proposes that there is an alternative perspective that could be used to allow the assessment of whether a student was in the process of becoming autonomous.If learning outcomes are articulated as thresholds of achievement, and the minimum threshold is some evidence of some of the characteristics of autonomy (however they are specified at disciplinary level) then a student who is able to demonstrate this could be said to have achieved.With supporting evidence from a mentor, who has worked closely with the student over a period of time, assessors should thus be able to determine whether they were "becoming" autonomous.
As has been discussed, defining autonomy is problematic.Yet articulating the characteristics and attributes of autonomy may not be.It should be possible at a disciplinary level to clearly identify and articulate its attributes, and how these may be demonstrated in that specific field.Following this it would be possible to develop appropriate threshold learning outcomes.Doing so would allow higher education programs to justifiably make a claim that they are assessing learner autonomy.How though would these characteristics, whatever a discipline decides they may be, be best facilitated?Here the mentor has an important role to play.

The potential role of mentoring in facilitating the development of learner autonomy
Mentoring may be regarded as an ongoing relationship, over a period of time, which is more or less formalized (Connor & Pokora, 2007).It relates to a learner's broad, holistic development (Clutterbuck, 2008) and the identification and nurturing of potential for the person (Megginson & Clutterbuck, 2005).Importantly students have ownership of the goals and process of mentoring; they are more of a self-directed learner.The benefits of coaching and mentoring processes for higher education students are well known (e.g., Carnell, MacDonald, & Askew, 2006;Laverick, 2016;Schulz, 1995).And there is support from O'Leary's work (2006,2007) with language learners suggesting that assessing their autonomy can help students in their development as autonomous learners.Yet there would seem to be little research activity that has focused specifically on developing autonomy outside of the field of language learning.One piece of research which has done so is that by Ng, Confessore, andAbdullah (2012a, 2012b), using a Learner Autonomy Profile (LAP) which uses four constructs for autonomy (desire, resourcefulness, initiative, and persistence).Their study demonstrated that a five-week program of group coaching in Malaysian pre-diploma university students produced a statistically significant increase in autonomy.It recommended the coaching period for further research should be extended to the full length of a university semester.The study also suggested that the LAP may be a useful diagnostic tool for identifying areas for coaching interventions that can positively impact students' behavioral intentions to learn.That a short period of group coaching was effective suggests that over a longer period of time is likely to be even more so.And in line with other work EDUPIJ • Volume 7 • Issue 1 • 2018 (e.g., Bettinger & Baker, 2014), perhaps that individualized coaching or mentoring may be even more effective.In the context of higher education students, mentoring would be an appropriate process for helping develop autonomy.Through a relationship that is timebound by the academic year, or structure of a three-year or four-year degree program, a mentor could work over a period of time with the mentee to help them identify and develop areas where they are already demonstrating autonomy, where they could demonstrate autonomy, and how they may become "more autonomous."A long-term sustained relationship between mentor and student would be the most suitable for this process.Ideally, this relationship should involve the same mentor for the duration of the student's studies, yet it is acknowledged that, in practice, this may be difficult to achieve.
Learner autonomy is contextual and situational (Stefanou, Perencevich, DiCintio, & Turner, 2004).The same learner may exhibit high autonomy in one situation and low in another.A role for a mentor or coach here could be to work with students to identify different contexts and situations in which they behave with more, or less, autonomy and why.Helping them to reflect on this and consider how they could become more autonomous, or consider how the autonomy developed in one situation could be applied to another.
The mentor could also be involved in helping students to identify the amount of autonomy they may, or may not, have the capacity for developing in a specific learning context.This is an important role, because if a lecturer is not particularly interested in helping a learner develop their autonomy, then, despite what is articulated in learning outcomes, without support they may be unlikely to be able to clearly identify the level of autonomy available to them.The student may strive for a level of autonomy that may well not be available to them; leading to a lack of progress and frustration.This may negatively impact upon their learning.The mentor, not directly involved in teaching, yet with a remit for encouraging or facilitating autonomy, may play an important role here.
Practical considerations: how would the mentor role operate in practice?Cotterall (1995) argued that autonomy cannot just be added to an existing program of learning, but that its development should be implied throughout the whole curriculum.How mentoring is organized and coordinated within institutions would necessarily be different from one university to another in order to take account of different structures.Yet it will be necessary to embed it, rather than bolting it on to existing programs of study, and this will require time and resourcing.Resourcing includes not only the funds to operate an effective scheme and administer it, but the selection and where required, training of suitable mentors.

Who would be the most suitable mentors?
Academic staff workload pressure from the demands of teaching, research and administration may prevent them from having the necessary time to act as mentors, and institutional promotional structures may not reward a lecturer for engaging in student mentoring.A long-term sustained mentoring relationship provided by someone with neither a strictly administrative, nor pure academic role may offer a potential solution here.It is proposed that the mentor could be a person who bridges the academic-administrative divide, someone with experience of providing advice and guidance, with good listening and observational skills, and with experience of working one-to-one with learners.There are a EDUPIJ • Volume 7 • Issue 1 • 2018 number of such roles already in existence within higher education, for example, careers advisors, study and learning advisers, and student support advisors.Staff in these roles may already have skills and experience to be able to act as student mentors, and with suitable training could be very effective.It is acknowledged that the characteristics of effective mentors are identifiable and generally agreed upon (Laverick, 2016), thus identifying suitable mentors within a university should not be problematic, and training should improve their effectiveness (Clutterbuck, 2014;Thornton, 2014).Such a model should not take up additional academic staff time, yet it would still require institutional resourcing.In today's higher education system it is, unfortunately, unlikely that any existing member of staff, whether academic, managerial or support staff will have the necessary time available that may be necessary to carry out such a role.

Resourcing
What may well be an issue for any university is the funding required for a mentoring scheme.Yet, if student autonomy is an educational goal that is genuinely valued, it follows that if institutions want a successful mentoring scheme, they will need to ensure that it is appropriately resourced (Klasen & Clutterbuck, 2011).Any institution has to be able to resource mentoring, yet whether institutions decide to do this or not, relates ultimately, to how important student autonomy is regarded both by the individual university and relevant stakeholders.Ecclestone's framework involving a movement from Procedural to Personal to Critical Emancipatory (Ecclestone, 2000) could perhaps be used as a basis for developing structured mentoring processes that are gradually withdrawn as the learner progresses, and demonstrates increasing autonomy.As students moved through their degree and progressed through the stages, with the support of a mentor, this would demonstrate that they were "moving towards" autonomy.As they achieved certain specified minimum thresholds of achievement, articulated as learning outcomes, the mentor could provide evidence, or witness testimony, to support assessment processes.As previously discussed, the assessment would not be required to categorically identify that a student had become an autonomous learner, only that they were in some way "more autonomous." With a process such as this in place universities would then be able to legitimately claim that they were developing student autonomy.They would also be able to provide evidence to external stakeholders that they were fulfilling a key espoused aim of higher education.

Conclusion
It is evident that whilst developing learner autonomy is seen to be a key and increasingly important aspiration of higher education, it is not currently as clearly defined as would be ideal, nor being assessed.It is likely that existing stakeholders may have differing interpretations and understandings of what learner autonomy is and may comprise.The author acknowledges that this itself may need addressing prior to institutions investing in and implementing mentoring schemes or enacting changes to existing schemes; further that this might require a lengthy period of time to carry out.There is also much-needed work to be conducted at disciplinary level in order to clarify what the attributes of learner autonomy are within a discipline, and this too may well require action and further research prior to the introduction of mentoring schemes that contribute to facilitating learner autonomy.