The relationship between differentiated instruction and learner levels of engagement at university

Purpose – The aim of this paper is to explore the importance of individual learner differences and the effect differentiated instruction (DI) has on learners ’ levels of engagement. Design/methodology/approach – The author carried out this research using a small-scale action research (AR) study. Findings – The findings suggest that in acknowledging and responding to individual learner differences, especially interests, levels of learner engagement are positively affected. Research limitations/implications – This study ’ s key limitations were sample size, short-term study and potential teacher as researcher bias. Practicalimplications – Recommendations weremade fora furtherlongitudinalstudyintothe relationship between DI and language learner levels of engagement at University. An additional study into DI that looksbeyond language learning at HE, could add value to pedagogic approaches, which could make courses of greater intrinsic value to its students. Originality/value – ThisresearchstudyaimstohelpfillagapintheliteratureontheapplicationofDI,aswell as a unique perspective into its effect on learner engagement within a university context.


Introduction
What is moderately challenging and motivating for one learner may offer far too little challenge (and therefore little motivation) for a classmate.The same task may be too stressful for yet another classmate.Learning tasks must be adjusted to each student's appropriate learning zone (Tomlinson, 2014, p. 34) Learner diversity is a common aspect of many if not all classrooms, irrespective of subject, level or nature of study because learners are human beings with their own unique identities (D€ ornyei, 2009, p. 230).Teaching a class with a diverse array of learners, each bringing with them a myriad of differences is therefore typical but can make responding effectively difficult if not impossible (Macintyre et al. in Hall, 2016, p. 310).If understanding learners' differences and responding to them is needed in order to engage learners (Subban, 2006, p. 941) then this arguably needs to be the first step, before innovating teaching practices, to ensure that they are appropriately challenging (Hattie, 2012, p. 52).
Learner diversity, "learner factors" (Stern, 1983) or "individual differences" (Dornyei, 2005;Macintyre et al. in Hall, 2016) offer a "key reason why many second language learners fail-while some learners do better with less effort" (Dornyei, 2005, abstract).Therein lies the need for the teacher to reflect on learners within their context (D€ ornyei, 2009), to understand their complex individual differences and respond with appropriate contextual and pedagogic realisation (Jolly and Bolitho, 2011, p. 108).This could be a core component of Evans et al.'s "high impact" higher education (HE) strategy, where "learning opportunities [that] lead to student retention, successful completion of programmes, and encourage student behaviours that lead to meaningful learning gains" (2015, p. 7).

Background
This research study was born out of a context of teaching beginners French as a foreign language at a British university.The university's language courses were level specific, using the Common European Framework of Reference of Languages (CEFR) and a place on each course was determined by previous language learning experience.The course syllabus was designed around the CEFR and learners' ability to demonstrate "can-do" statements, which "reflect the continuum of growth in communication skills" (ACTFL https://www.actfl.org/publications/guidelines-and-manuals/ncssfl-actfl-can-do-statements).Formative and summative feedback was provided but only summative assessments included feedback using assessment criteria, learning outcomes and a grade.
On the A1, beginners' French course, there existed a low percentage of learners completing the course and sitting the summative assessments, which suggested that there were more individually desirable learning outcomes and course content than simply demonstrating CEFR A1 level of competence.Dornyei argued that, in fact, it is the combination of learners' individual differences that "answer why, how long, how hard, how well, how proactively, and in what way the learner engages in the learning process" (2009, pp. 231-232), which is the hypothesis of this research study.

A myriad of complexities
Teaching in a university context poses a myriad of complexities but taking into consideration the language learners themselves, further intensifies the complexity.If "whenever two or more people are present in a social context, there will be difference" (Adams and Nicolson, 2014, p. 25) and "diversity is the rule not the exception" (Macintyre et al. in Hall, 2016), then individual differences need to be responded to.Whole class instructional strategies and teaching to the middle with a one-size-fits-all coursebook leads to disengaged learners, where they do not involve themselves and speed through or struggle with the activities.Generic learning materials are arguably "outside the cultural experience" of most learners and "thus effectively useless" (Jolly and Bolitho, 2011, p. 108), despite teachers' creative attempts to blindly "personalise, localise and adapt" (Masuhara et al., 2008in Masuhara, 2011, p. 262).What therefore appears necessary is a greater understanding of learners' individual differences and teacher action to make the course and teaching more responsive to their individual learners.These developments could lead to a more inclusive and responsive practice and have a high impact (Evans et al., 2015, p. 7).
Differentiated instruction (DI) is a principle-based approach to teaching and learning that "advocates beginning where individuals are rather than with a prescribed plan for action that ignores student variance" (Tomlinson, 2014, p. 170).It neither advocates nor denies the use of materials, nor the use of specific instructional strategies because there is "no patented formula for creating a differentiated classroom" (Tomlinson, 2014, p. 25).It is an approach that recommends teachers "modify, adapt or design new approaches to instruction in response to students' needs, interests, and learning preferences" (Heacox, 2012, pp. 6-7) so that lessons no longer teach to the middle but are the product of understanding its learners.
It is my intention in this research study to explore how DI can be implemented within this context and to investigate its relationship with learner levels of engagement.Intent on improving teaching and learning within and without this context and regarding learner variance as a "resource, rather than an obstacle to overcome" (Richards, 2015a, p. 137), this research study intends to explore how it can be implemented and to investigate its relationship with learner levels of engagement.

Learner engagement
Engagement is rooted in the constructivist belief that learning is "influenced by how an individual participates in educationally purposeful activities" (Coates, 2005, p. 26) and has been described as "the time and energy students devote to educationally sound activities" (Kuh, 2003, p. 25); "the connection between person and activity" (Russell et al., 2005, in Appleton et al., 2008, p. 380) and "the psychological process, specifically, the attention, interest, investment, and effort students expend in the work of learning" (Marks, 2000, p. 154).As teachers, we notably monitor engagement through learner attendance, participation, effort and achievement, and it is therefore an optimal outcome for both teachers and learners.We want learners to be engaged because "successful (language) learning depends upon active participation and involvement by learners" (Richards, 2015a, p. 136).It is therefore easy to believe the positive link between engagement and achievement (Marks, 2000).Janosz (2012, p. 695) distinguishes between determinants and outcomes of engagement, i.e. what affects engagement (determinants) and the effects of engagement (outcomes).Appleton et al. (2008, p. 383) employs facilitators of engagement and indicators of engagement in a similar way.To achieve the latter, Janosz argues that "we must privilege age-appropriate interventions, educational environments, and learning situations that respond to fundamental individual needs" (in Christenson et al., 2012, p. 699).Tinto (2012) goes on to present four conditions that promote engagement: expectations, support, assessment and feedback and involvement.This research is echoed by Rumberger and Rotermund (in Christenson et al., 2012, p. 503), who reiterate the importance of setting high educational expectations which are reinforced by assessment and feedback, arguably because they are integral to the positive cyclical effects of engagement.Osterman's (1998) research study further suggests this, where "engaged students perceive more support from teachers and peers and that this perception leads to a beneficial cycle of increased levels of engagement and increased adult support" (in Appleton et al., 2008, p. 374).This link has been highlighted in a report by Thomas, which links "belonging, retention and success" (2020, p. 20) and recommends institutional approaches that promote belonging of its students with an "HE experience that is relevant to interests and future goals" (Thomas, 2020, p. 72).Teaching at HE, therefore, should meet the individual needs of students (Chipchase et al., 2017, p. 40) in order to be relevant, to engage and to support success.
However, HE has historically been known for having "less personalised or less studentcentred approaches" (Evans et al., 2015), which are arguably the antithesis of determiners of engagement, such as work not being intrinsically or extrinsically motivating and too abstract instructional strategies (Wehlage et al. in Christenson et al., 2012, p. 495).However, although "one of the most direct and visible indicators of engagement is attendance (. ..) identifying the causes of dropping out is extremely difficult" (Rumberger and Rotermund in Christenson et al., 2012, p. 500)," and demands further investigation.Nonetheless, although attendance is not the only determiner of engagement, "student engagement can be influenced by the ways we teach" (Darr, 2012, p. 708), and this is the basis for this research study.
Differentiated instruction: its guiding principles DI is said to be "heuristic or principle driven" (Tomlinson, 2014, p. 25) as opposed to a fixed method or formula that breaks away from the "one size fits all" approach.Heacox (2012, p. 5) defines DI with five main attributes: rigorous, relevant, flexible and varied and complex, whereas Blaz offers ten: choice, collaboration, communication, connections, learning how to learn, multiple learning modes, open-endedness, routine, variety in instruction and assessment and collegiality (2016, pp. 4-5).Both Heacox and Blaz's key defining terms are embedded within Tomlinson's five guiding "underpinnings" (2014, p. 14) of DI: creating environments that are catalysts for learning; building on a foundation of a quality curriculum; using assessment to inform teaching and learning; tailoring instruction to assessment-indicated student needs and leading and managing a flexible classroom (2014, p. 20).These underpinnings or principles are inherent in Tomlinson's differentiation of content, process and product (1995,1999,2001,2014), which are flexible and responsive; based on three categories of learner individual differences: readiness, interest and learning profile (Tomlinson, 2014).
Differentiating content, refers to "what pupils learn" (Rogers, 1976in Johnson, 1989, p. 26) or more accurately what "teachers want students to learn" (Tomlinson, 2014, p. 18).Rock et al. in their framework for DI refer to it as the "content variable" and propose that teachers ask themselves: "what content is there?Why should they care?"(2008, p. 35).DI proposes that in order to answer these questions with genuine effect, teachers must know their learners.Knowing what learners already know about a topic, their readiness, will inform their choice of content and challenge and "if material is presented at or below the mastery level, there will be no growth.If presented well above the zone, children will be confused and frustrated" (Byrnes, 1996, p. 33).In order to apply this theory, teachers will need to know their learners' level of mastery.
The second element process refers to how learners will "make their own sense of the content or input" (Theisen, 2002, p. 2) and "how they learn it, and how teachers help them learn" (Rogers, 1976in Johnson, 1989, p. 26).In order to differentiate process, the teacher uses knowledge of the learners' identity; their investment and motivation and learning profile; how they learn best, i.e. learning strengths (Gardner, 2006) and styles (Rubin, 1975;Fleming and Mills, 1992;Reid, 1998;Skehan, 1998;Dornyei, 2005;Fleming and Baume, 2006).Understanding how they learn best and what skills they want to improve, provides a foundation of knowledge on which process can be designed.
Differentiating the third element: product, involves differentiating the "styles and methods of assessment" (Rogers 1976in Johnson, 1989) or output (Theisen, 2002, p. 4), where learners demonstrate and extend their learning (Tomlinson, 2014, p. 18).The product could be based on interest and learning profile, and the difficulty of the product could be linked to readiness, rooted in Bloom's Taxonomy (1956) and underpinned by Vygotsky's ZPD (1986).The type of product can reflect their strengths (Multiple Intelligences, Gardner, 2006) and styles (Rubin, 1975;Fleming and Mills, 1992;Reid, 1998;Skehan, 1998;Dornyei, 2005;Fleming and Baume, 2006), and presenting an output could encourage learner motivation and investment (Norton, 2013).The output of "constructing a public entity" can encourage further learning and gaining peer feedback can help to "build knowledge structures" (Papert and Harel, 1991).Such peer review, according to Petty, reinforces learners' motivation as "students are more motivated when peers are the audience rather than the teacher" (2006, p. 242).
Understanding learner readiness, interests and learning profile and using this knowledge to modify content, process and product, will help teaching be on its way to being of "high interest and high relevance" (Tomlinson, 2014, p. 33).What therefore seems of paramount importance is the role of the teacher.Described as diagnostician (Tomlinson, 2014, p. 4), change agent (Hattie, 2012, p. 162) and more knowledgeable other (Vygotsky, 1978), the teacher's role is instrumental in identifying learner needs and making the teaching relevant to their interests and future goals (Thomas, 2020, p. 72).In using this information to better their teaching they will be innovating their practice and in doing so, securing a positive and supportive climate for learning that will positively impact learner engagement.This idea is not new, as Subban stated: "curricula should be designed to engage students, it should have the ability to connect to their lives and positively influence their levels of motivation" (2006, p. 941).

Research design
This study aimed to explore the relationship between DI and levels of learner engagement in a university teaching context.By combining a cycle of action and reflection (Kolb, 1984), the aim was to improve my teaching practice by interventionist means.Although small-scale due to time, which arguably only "permits answers to short-term issues" (Cohen et al., 2007, p. 80), the issue of learner engagement is a long-term issue in language teaching and more generally across HE.
Rooted in my desire to improve my quality of teaching and learning, this study involved a mixed method approach; incorporating quantitative and qualitative questionnaires, group interviews, peer observations and "teacher as researcher" observations to allow for both inductive and deductive reasoning (Bryman, 2001, p. 20).The participants were all invited, self-selected learners of French level A1, constituting university students, public and staff.The number of self-selecting participants that took part in this study was eight.The participants' ages ranged from twenty to sixty-seven.
A combination of instruments for data collection were employed, including a pre-assessment pack (Heacox, 2012;Blaz, 2016;Tomlinson, 2014) (Appendix 1) that was designed and disseminated to participants and a "triangulation" (Denscombe and Dawsonera, 2010, p. 154) of methods, which considered researcher observations, teacher observations and participant feedback.

Framework for data analysis
A framework for data analysis has been developed for each instrument, based on Tomlinson's principles.The data will be presented in response to each sub-question and where appropriate, data from different instruments will be analysed.

Findings and discussion
This findings and discussion section aims to answer the principle question: what is the relationship between DI and learner levels of engagement at university level?Findings and discussion will be presented under four sub-questions.It is important to state that due to the amount of data recorded, only the most salient findings that addressed each sub-question will be presented and discussed.
How can individual differences be identified and responded to through DI? Readiness.Information on participants' readiness was assessed in order to determine what should be taught and at what level of challenge.Participants' responses demonstrated that four out five participants felt "somewhat" confident in competencies related to the subject topic and one participant felt competent.Participants' readiness was further evaluated by attendance records as well as professional observations (Heacox, 2012, p. 26).
Although measurements of participants' readiness seemed consistent before the AR class took place, in-class observations and interviews showed it to be inaccurate.Despite aiming to begin "where individuals are rather than with a prescribed plan for action that ignores student variance" (Tomlinson, 2014, p. 170), the inconsistency of the pre-assessments with participants' actual readiness highlighted the difficulty of accurately setting challenges to their actual level of mastery (Vygotsky, 1986).
Interest.Assessing participants' interests was manageable, insightful and cogent.The data (Table 1) demonstrated motivations that support the belief that there are more desirable learning outcomes to our course than simply demonstrating CEFR A1 level of competence.
These findings allowed me to design tasks that were truly relevant to the students' motivations and interests, where the content, process and product were all differentiated.Task 1, for example, involved participants choosing one of four "learning centres" that each had a specific holiday focus: the beach, the countryside, the city and adventure and carry out a vocabulary activity followed by creating a holiday character and using the past tense and vocabulary to tell the others what they did on their holiday.Planning the lesson in accordance with their interests does "motivate students' exploration of topics" (Tomlinson, 2014, p. 76) because the content was "dynamic, intellectually intriguing and personal" (Tomlinson, 2014, p. 53).Knowing participants' interests and hobbies and designing responsive content allowed me to truthfully answer "what content is there?Why should they care?"(Rock et al., 2008, p. 35)  Learning profile.Data collated on participants' learning profiles were varied, eclectic and occasionally contradictory (Table 2), supporting Graham Hall's assertion that learning styles are "not wholly innate and therefore not completely fixed in nature" (2011, p. 4).That being said, how they learn and their strengths were taken into consideration with task 1 involving flexible instructional groups, which took into consideration their learning style preferences, with matching images with text that aimed at the four participants with naturalistic strength.Task 3 again involved flexible instructional groups, as well as participants choosing content, process and product (Appendix 2).Ironically, in task 1, despite P3's preference for working independently, he worked in a pair.Likewise, in task 3, despite participants stating they wished to produce a role play (P1 and P2), produce a drawing (P4) and a film (P5), they all wrote down a script in prose format.How they engaged in their learning and what they produced was ultimately different to their self-assessed learning styles and strengths, an incongruity that again supports Hall's assertion that learning styles are "not wholly innate and therefore not completely fixed in nature" (2011, p. 4).However, participants' feedback suggests that this response was rooted in their own learning strategies, writing it down "just to get our head around the grammar and to work out the grammar", reflecting their linguistic abilities.
Participants' learning strengths and preferences for learning were assessed at the beginning of the AR in order to identify "the ways in which a learner learns" (Tomlinson, 2014, p. 19) so that instructional strategies could be tailored to each participant (Blaz, 2016, p. 160).However, the eclectic and sometimes complex responses demonstrated how their ideas on how they learn best and their perceived strengths were not completely reflected in "the characteristic manner in which an individual chooses to approach a learning task" (Skehan, 1998, p. 237).Therefore, although advantageous in terms of encouraging reflection and designing an array of instructional strategies with a variety of choices, participants did not always choose the most suited process and product.However, this "fairly complete knowledge of the student's learning styles and preferences can provide an effective basis for differentiated instruction" (Bender, 2008, p. 4) because it encourages a variety of choice of process and product, where choice can lead to empowerment and engagement.
Do learners engage with this approach to teaching and learning equally?i.e.Is it inclusive?Academic engagement.Video based and subsequently teacher observation found that P3 did not academically engage in two of the tasks.This could have multiple explanations: reflective of the complex nature of individual internal variables that "interact with each other" (Dekeyser 2013in DeKeyser, 2016, p. 353), including anxiety and willingness to communicate (Dornyei, 2005), P3's lack of active participation could have been reflective of his readiness, being that the tasks were above his level of mastery (Vygotsky, 1986, p. 33) or alternatively, that he did not like the tasks, having a preference for working independently.This task

Differentiated instruction
offered no choice and may have brought on an affective filter (Krashen, 1982) such as anxiety, which can negatively impact engagement and consequently language achievement (MacIntyre, 1999).Alternatively, if learners "are welcomed and valued as they are"' (Tomlinson, 2014, p. 15) then perhaps this finding should simply reflect this participant's individual choice to not actively engage, despite its potential negative effect on learning.As Dornyei asserts, it is the combination of learners' individual differences that "has been seen to answer why, how long, how hard, how well, how proactively, and in what way the learner engages in the learning process" (2009, pp. 231-232).
Behavioural engagement.Similar to the academic engagement findings, P3 demonstrated the least behavioural engagement (Table 3) throughout the AR class.Interestingly, P5 behaviourally engaged the most and despite having initially responded over confidently with his readiness, his confidence arguably positively affected his behavioural engagement.Griffiths theory that "learners' sense of identity is seen as a major contributor to motivation" (2015, p. 430) is supported by P5 who commented "what I found good about that is I could focus on speaking ((hand gestures from mouth)) because that's my weakness".Reiterating his awareness of self and motivations for investing time in the class, P5's sense of identity has therefore been accurately responded to by DI's "learning situations that respond to fundamental individual needs" (Janosz in Christenson et al., 2012, p. 699).
The data also demonstrated that task 3, where content, process and product were differentiated, generated the most behavioural engagement.Having an allocated 30-40 min to complete, which although did not accurately implement DI's flexibility of time, did give all participants an extended amount of time to complete the task; allowing participants to collaborate and communicate; to apply their knowledge and use their own resources (Ellis, 2009) to "make their own sense of a topic" (Petty, 2006, p. 234).The combination of differentiating all curriculum elements and increasing time spent on the task seems to have increased participants' behavioural engagement.
Cognitive engagement.Findings from the student post-questionnaire (Table 4) showed that all felt engaged, and the three main determinants (Janosz in Christenson et al., 2012) or facilitators of engagement (Appleton et al., 2008) were relevance of schoolwork to future endeavours, value of learning and self-regulation.Relevance is one of five main attributes of DI, as defined by Heacox (2012, p. 5) that is equally a characteristic of a quality curriculum that "taps into learners' feelings and experiences" (Tomlinson, 2014, p. 33).Arguably, participants felt engaged because the content, process and product held their attention, interested them and encouraged expending effort because it was deemed relevant.
Value of learning reflects participants' positive perception of the content, process and product and self-regulation, the opportunity to self-direct (Zimmerman, 2002), arguably reiterates what Janosz describes as "learning situations that respond to fundamental individual needs" (in Christenson et al., 2012, p. 699).With opportunities to be active in their own learning, participants were equally encouraged to be autonomous.
Findings on P3's cognitive engagement could be explained by the individual difference variable of anxiety (Dornyei, 2005), affective filters (Krashen, 1982), or alternatively, his lack of contributions to the group interview could demonstrate reactivity (Cohen et al., 2007).P3's observed lack of active participation and engagement could reflect the reverse of Osterman's (1998) findings of a "beneficial cycle of increased levels of engagement and increased adult support" (Appleton et al., 2008, p. 374).Not as engaged as the others, perhaps P3 did not perceive the support available.In contrast, P4 and P5, who had the highest frequency of behavioural engagement, also referenced the most internal indicators of cognitive engagement, suggesting the "beneficial cycle of increased levels of engagement" (Appleton et al., 2008, p. 374).Table 3.

JRIT
What are the learners' attitudes to the DI lesson?Relevance of content and learning opportunities.Relevance of content.Consensus of participants' interest in the topic and how the types of holidays and activities reflected their individual preferences, reiterated the importance of relevance, an attribute of DI (Heacox, 2012, p. 5), a determinant of learner engagement (Janosz in Christenson et al., 2012) and a key component to creating a sense of belonging at HE (Thomas, 2020, p. 72).P4's comment on wanting to "talk and write about what matters to us!" further reiterates the importance of relating content to their interests and future goals (Thomas, 2020) and indicates how DI related to interests alone can encourage learner engagement.Relevance appears to be the glue of engagement; "the connection between person and activity" (Russell et al. in Appleton et al., 2006, p. 380).The pre-questionnaire pack allowed the design of content that reflected each participants' unique and individual identities, demonstrating "high respect" (Hattie, 2012, p. 26) for the learner-participants.In doing so, the relevant content helped create a supportive learning climate (D€ ornyei, 2001) where participants felt engaged, interested and their needs met.
Relevance of learning opportunities.Despite the inaccuracy of assessing readiness, participants left positive feedback regarding the relevance of the learning opportunities.The tasks were designed in response to their self-reported and observed readiness to meet their "appropriate learning zone" (Tomlinson, 2014, p. 34).Even P5 who stated that he felt "confident" in the pre-assessment pack with all three can-do statements, welcomed the opportunity to practise speaking, and P3 positively reflected on the challenge.Due to learning profile data being varied, eclectic and occasionally contradictory, the processes and products were varied in choice and allowed participants to respond in a way that was relevant to them, in the moment, which seemed to encourage autonomy, another contributing factor to engagement (Janosz, 2012).
Variety of choice.Variety of choice was perceived, in the most part, as positive, especially with regard to choosing content or "what pupils learn" (Rogers, 1976in Johnson, 1989, p. 26).Responding to "fundamental individual needs" (Janosz in Christenson et al., 2012, p. 699), participants had choice of content, process and product and their feedback, combined with observational analysis, evidenced how choices "enhance their motivation to learn" (Subban, 2006, p. 938).However, participants' feedback relating to choice of process and product reinforces Blaz's suggestion of "not doing it exclusively" because "students don't always know (or choose) what is best for them" (2016, p. 12).There appears a need for a balance of choice, especially with regards to process and product and for the teacher to be the diagnostician (Tomlinson, 2014, p. 4); identifying and giving more structured support to learners to make informed choices and increase learning opportunities.
Time.Participants' positively responded to having more time to apply and learn what had been taught.One participant commented that the original teaching was "too short it

Differentiated instruction
didn't register ((P1 nods)) (. ..)I had no memory of them" (P2), but the class allowed her to "try to work them out (. ..)Actually wrote them down then I understood the rule" (P2).This point of view was verbally agreed with by P4 who reflected: "actually do connected pieces ((P1 and P2 nod)) rather than simply I know the phrase for that, or I know the phrase for that" (P2).These comments highlight how time allows for the application of knowledge to "make their own sense of a topic" (Petty, 2006, p. 234) and thereby participate in active learning where they "build knowledge structures" (Papert and Harel, 1991).Time paired with relevant and appropriately challenging content, process and product could therefore contribute to a "high interest curriculum" (Tomlinson, 2014, p. 33), where learners have more time to attempt to "make sense and meaning of the world they inhabit" (Tomlinson, 2014, p. 78).
What are the teachers observer's (TO) perceptions of learner engagement?
The TO perceptions of learner engagement were positive, stating that they were "absolutely engaged" and "completely emerged", reflecting participants' own admissions.The TO noted the positive impact of responding to individual learner differences, noting they felt "valued" and were "learning at their own pace", comments that resonate with Tomlinson's (2014)principle of creating environments that are catalysts for learning.Valuing individual learner differences also suggests how DI can outwardly portray "high respect" (Hattie, 2012, p. 26) for its learners, a contributing factor to the "beneficial cycle of increased levels of engagement and increased adult support" (Appleton et al., 2008, p. 374).Despite the overarching positive feedback, the TO did make two critical observations: preparation time and class size.The comment on preparation time reflected findings from other research studies (Willard-Holt, 1994;Rock et al., 2008) who found that DI takes too much time, with many teachers opting against it.However, its positive effect on learning suggests that perhaps "time may actually be saved as students engage in learning that responds to their needs" (Heacox, 2012, p. 14).The longer tasks, reflecting elements of Task Based Learning, could reduce teacher preparation time and simultaneously encourage learners to "engage dynamically with the language" (Bygate, 2016, p. 382).
The TO's second critical comment regarding how a larger class size could negatively affect teacher and learner interaction did raise concerns as the study's class size was admittedly smaller than average and much smaller than HE lectures and seminar classes.However, Blaz states that DI requires "focus on one aspect at a time" (2016, p. 15) and therefore with practice, development and greater awareness of DI by the learners and teacher, class size might not be an issue.With further practice and development, DI could be a long-term solution to the "one size fits all" approach (Tomlinson, 2014, p. 25) that stultifies learners (Wehlage et al., 1989in Christenson et al., 2012, p. 495).Concluding that DI is "brilliant for retention I think this is the way forward", the TO's over-arching optimism positively reflects on DI as "heuristic or principle driven" (Tomlinson, 2014, p. 25) and identifies its potential role in augmenting learner levels of engagement and potentially, retention of language learners in HE.

Summary of findings
This small-scale AR study suggests the positive relationship between DI and learner levels of engagement in response to meeting the needs and wants of a small number of diverse language learners at university.After discussing the data gathered for the four research questions (RQs), the following key findings emerge: Individual learner differences are varied and can be identified and responded to through DI Interests seem the easiest for participants to accurately identify and of most value when designing relevant content, whereas readiness seems more difficult to accurately assess.That being said, DI raised my awareness of learner readiness and what seems necessary is further investment, development of formative assessments and better two-way feedback.Assessing learning profiles gave insight into learners' multiple learning strengths and preferences, a knowledge that perhaps unbeknown to the learners themselves, was used to optimum effect when designing instructional strategies that offered choice that varied "systematically to cater to individual learner differences" (Long, 2016, p. 7).Variety and choice of content, process and product were positively received by learners who felt engaged.However, it seems apparent that "students don't always know (or choose) what is best for them" (Blaz, 2016, p. 12) and therefore, the teacher's role is paramount in the success of DI and its effect on learner engagement.
Two-way feedback is pivotal in analysing and securing learner engagement through DI Although this study did take learners' individual differences into significant consideration, through the pre-assessment pack, as well as course outcomes, it did not "provide learners with guidance of how to manage them" (Macintyre et al. in Hall, 2016, p. 319).More constructive and dialogic feedback on their individual differences could prevent inaccurate reports of readiness and promote engagement in all tasks (Tinto, 2012, p. 7), supporting learners to "believe they are capable of achieving success" (Rumberger and Rotermund in Christenson et al., 2012, p. 503).Advocating such a student-centred approach might not only encourage dialogue, improve the teachers' understanding of learners' individual differences and learner engagement but could also "lead to meaningful learning gains" (Evans et al., 2015, p. 7).DI appears inclusive although not all learners engage with this approach equally Analysing findings from all three "dynamically interrelated" (Fredricks et al., 2004, p. 61) measurements of engagement makes apparent that learners engaged with DI differently, based on their own individual differences.All participants gave positive feedback on their engagement, reflecting "their attention and interest in the psychological process" (Marks, 2000, p. 154) of learning.DI allowed all participants to feel engaged and respond to the tasks with individualised levels of engagement.
Participants and teachers' share positive attitudes towards DI and its opportunity for tasks Three key themes emerged from participants' feedback: relevance, variety of choice and time, all of which are embedded within DI's guiding principles (Tomlinson, 2014), recognising the effective nature of DI.The emphasis on the main task and time allocation positively reflect attributes of task-supported language teaching (TSLT) "not so they think like target speakers, but simply so that they engage dynamically with the language" (Bygate, 2016, p. 382).The tasks provided "an opportunity for additional communicative language use" (Bygate, 2016, p. 387) and allowed learners to develop and demonstrate communicative competence at level A1; to "understand and use familiar expressions" (Council of Europe), where the familiar expressions were situated within each participants' own wants and needs.

Limitations and recommendations
A major limitation of this study was its small-scale nature and what is therefore recommended is a longitudinal study into the relationship between DI and learner levels of engagement.Over a greater amount of time and different courses, with the analysis of potentially richer data, a longitudinal study could give further insight into DI and learner

Differentiated instruction
engagement.It could also give further insight into teacher preparation time and if it will be saved (Heacox, 2012, p. 14) or whether responding to an even greater array of individual difference variables will be insurmountable (Rock et al., 2008, p. 34).Findings from future studies could far exceed what has been found in this small-scale AR study and offer greater external validity (Cohen et al., 2007, p. 186).

Conclusions
This AR study, within its own specific context of a university adult language course, has raised poignant issues relating to how we as HE teachers acknowledge and respond to learners' individual differences and its effect on learners' morale and engagement within the classroom.What has proved apparent is learners' positive acknowledgement of the conscious effort made to respond to their individual differences (Hattie, 2012) and their subsequent engagement in learning (Subban, 2006, p. 941), which could have further positive cyclical effects on engagement (Appleton et al., 2008, p. 374)

Differentiated instruction
Appendix 2 Corresponding author Robyn Moallemi can be contacted at: r.steer2@brighton.ac.uk For instructions on how to order reprints of this article, please visit our website: www.emeraldgrouppublishing.com/licensing/reprints.htm Or contact us for further details: permissions@emeraldinsight.com best interac ng with others _____ I enjoy informal chat and serious discussion _____ The more the merrier _____ I o en serve as a leader among peers and colleagues Differentiated instruction _____ I value rela onships more than ideas or accomplishments _____ Study groups are very produc ve for me _____ I am a "team player" _____ Friends are important to me _____ I belong to more than three clubs or organisa ons _____ I dislike working alone _____ TOTAL for Sec on 5 Sec on 6 _____ I learn by doing _____ I enjoy making things with my hands _____ Sports are a part of my life _____ I use gestures and non-verbal cues when I communicate _____ Demonstra ng is be er than explaining _____ I love to dance _____ I like working with tools _____ Inac vity can make me more red than being very busy _____ Hands-on ac vi es are fun _____ I live an ac ve lifestyle _____ TOTAL for Sec on 6 Sec on 7 _____ Foreign languages interest me _____ I enjoy reading books, magazines and web sites _____ I keep a journal _____ Word puzzles like crosswords or jumbles are enjoyable _____ Taking notes helps me remember and understand _____ I faithfully contact friends through le ers and/or e-mail _____ It is easy for me to explain my ideas to others _____ I write for pleasure _____ Puns, anagrams and spoonerisms are fun _____ I enjoy public speaking and par cipa ng in debates _____ TOTAL for Sec on 7 JRIT Sec on 8 _____ My a tude affects how I learn _____ I like to be involved in causes that help others _____ I am keenly aware of my moral beliefs _____ I learn best when I have an emo onal a achment to the subject _____ Fairness is important to me _____ Social jus ce issues interest me _____ Working alone can be just as produc ve as working in a group _____ I need to know why I should do something before I agree to do it _____ When I believe in something I give more effort towards it _____ I am willing to protest or sign a pe on to right a wrong _____ TOTAL for Sec on 8 Sec on 9 _____ Rearranging a room and redecora ng are fun for me _____ I enjoy crea ng my own works of art _____ I remember be er using graphic organisers _____ I enjoy all kinds of entertainment media _____ Charts, graphs and tables help me interpret data _____ A music video can make me more interested in a song _____ I can recall things as mental pictures _____ I am good at reading maps and blueprints _____ Three dimensional puzzles are fun _____ I can visualise ideas in my mind _____ TOTAL for Sec on 9 Source(s): McKenzie, W (1999) Multiple Intelligences Inventory online.Available at http://surfaquarium.com/MI/inventory.htm[accessed 9 May 2017] and promote active learner engagement.

Table 2 .
Participants' data relating to learner profile . The participants' feedback positively acknowledged the importance of inclusivity; respecting, recognising and valuing individual differences, both visible and invisible, by innovating instructional strategies that are relevant, responsive and will maximise learner engagement.Therefore, strongly suggesting DI could offer the learning opportunities that lead to achieving the "high impact" of Evan's et al.'s HE strategy and encourage both teacher and "student behaviours that lead to meaningful learning gains" (2015, p. 7).Complete each sec on by placing a "1" next to each statement you feel accurately describes you.If you do not iden fy with a statement, leave the space provided blank.Then total the column in each sec on.

Table A1 .
DI Lesson Plan