“What our children and young people learn during their time at school has never been more important.” (Donaldson, 2015, p.5)

By Donaldson’s own admission, greater focus has recently been placed upon curriculum design around the world as more attention has been paid to the importance of what and how pupils learn. As a result of Wales’ continuously declining performance in PISA accreditations, the Welsh Government commissioned Donaldson (2015) to complete a review of the curriculum in order to ensure that all pupils are afforded ample opportunities to benefit from an education which effectively prepares them for the future. Within ‘Successful Futures’(2015), Donaldson openly makes a direct connection between the quality of education and the productivity of countries as a result of being exposed to an eminent, meaningful and pertinent curriculum and his visions for the Welsh education system. 
When reflecting upon documentation in place, it is clear that Wales’ current provision in Key Stage Two is largely driven by the 2008 curriculum which promoted the importance of a subject focused design and ensured that all pupils were entitled to an education which encouraged an acquisition of designated knowledge and skills. However, in response, Donaldson suggests that the development of this style of curriculum “led to unacceptable inconsistency” (Donaldson, 2015 p.8). As a result of this inconsistency, the Welsh Government has continuously come under fire for the underperformance of its pupils and so additional areas of importance have been added to the curriculum over time. When looking at the Key Stage Two Curriculum alone, the LNF (2014) and the DCF (2017) have most recently been developed as bolt-on documentation in response to high pressure circumstances and the need for national improvement. It is therefore understandable that Donaldson’s review aims to modernize the Welsh curriculum in an attempt to bring it into the 21st Century and help all pupils to “meet the needs of today and the challenges of tomorrow” (Donaldson, 2015 p.6).
From the outset of ‘successful Futures’ (2015) Donaldson emphasizes the importance of pupil autonomy through the power of ‘choice’, actively placing pupils at the heart of his recommendations for curriculum reform.  Throughout his review, Donaldson encourages teachers to disregard didactic pedagogy in which curriculum content is merely delivered rather than collaboratively explored by pupils.  He states that meaningful experiences which place learners at the heart of curricular exploration encourage lifelong learning. Donaldson (2015) further heralds the importance of personable and purposeful opportunities for independent learning rather than mere responses to external requirements. So suggesting that the ‘New Curriculum’ for Wales should ensure that children are active drivers of their own education by developing documentation which is both flexible and responsive to the needs of children in any individualized learning context. 
When considering the concept of learning autonomy and the theory of independent learning as a whole, it is of utmost importance to further explore key literature which could endeavour to explain Donaldson’s beliefs. When initially reflecting upon the discussion of pupil independence, it is clear that Donaldson (2015) draws heavily upon the work of Hattie and Yates (2014). Their publication ‘Visible Learning and the Science of How We Learn’ (2014) openly discusses the pitfalls of passive learning experiences within modern day education and promotes the importance of active, visible and meaningful learning in which pupils are leaders of their own academic achievement. Within their research, Hattie and Yates (2014) imply that passive learning results in a decline in engagement levels and pupil progress in contrast to active learners who thrive amongst their own elements of choice.  They state that academia of today is oversaturated by a style of ‘recitation learning’ which ultimately disengages pupils and doesn’t lead to an increased acquisition of life long skills. As a result of this, Hattie and Yates (2014) argue that pupils are prepared for the content of assessments and receive a false sense of achievement through graded academic success, but actually lose out in the long run due to the lack of exposure to collaborative and dynamic problem solving thought processes and an inability to lead independently.  
In comparison, Williams’ (2003) research based upon the rivalry between dependent and independent learners directly compliments Hattie and Yates’ beliefs. Throughout her publication, ‘Promoting Independent Learning in the Primary Classroom’ (2003), Williams makes a close link between lifelong success and the importance of pupil led experience. She argues that independent children have a greater sense of wellbeing and are able to overcome failure when the main motivation for learning originates within their own experiences. Williams (2003) further implies that teachers rotate their curricular approaches around the need for pupil output - as Janet Moyles (2003) suggests, this encourages children to become quickly “disempowered and wholeheartedly dependent” (Moyles, 2003 p. xxi) upon the teacher in order to feel any sense of academic fulfilment or achievement. 
In addition, in ‘Effective Learning in Classrooms’; Watkins et al (2007) discuss the concept of ‘learner-driven’ academic experiences. They argue that ‘interdependence’ amongst learners is key as interactivity promotes deeply embedded learning experiences and push children to reach their fullest potential.  Furthermore, when considering Maslow’s Hierarchy of Needs (1943) and the emphasis he places upon the human need for ‘self-actualisation and self-esteem’ in order to formally experience and sustain achievement, it seems difficult to deny the theoretical link that also exists between autonomy and life-long success. Having considered the perspectives and beliefs of researchers such as those discussed above therefore, it seems clear that there is a plethora of academic research that could support Donaldson’s desire to develop a curriculum that places pupil voice, autonomy and self-efficacy at its very core. In addition to this, it seems vital to also consider exemplar curriculums from around the world as important perspectives. 
When reflecting upon the core curriculum in Finland (2014), it is clear that project based immersive learning grants its pupils utmost flexibility in the way they convey their learning outcomes and plays to the strengths of pupil individuality.  In comparison, the core curriculum in New Zealand (2007) also emphasizes the importance of actively involving participants in their learning and wider community and encourage teachers to take the stance of facilitators of pupils’ learning in their academic endeavours. It is a curriculum which promotes shared learning experiences between pupils and ensures the development of curricular opportunities for deep level thinking that rotate around oral discussion and personal contribution. Similarly, upon discussion with Welsh Pioneer schools such as School C in which independent learning sits at the heart of their established curricular approach, School O which has been heralded by Estyn for its approach to child led experiences and School L which has received an excellent recommendation as a result of its use of ‘EPIC Boards’ which allow pupils to lead all aspects of curriculum design, it is evident that Donaldson’s (2015) suggested ways forward have already been applauded on a local level for apparent success. 
In contrast to the positivity displayed above however, some critics disagree with Donaldson’s promotion of pupil led learning. When reflecting upon the work of Priestly and Minty (2013) for instance, they directly criticize Donaldson’s ‘Curriculum for Excellence’ in Scotland (2004) and openly condemn the placement of pupils at the centre of the curriculum suggesting it’s “lacking in theoretical underpinnings” and state that in spite of the focus placed upon promoting active learning throughout CFE documentation, Donaldson allegedly fails to “articulate any form of learning theory, including the social-constructivism upon which [teachers can] implicitly draw” (Priestly and Minty, 2013 p.41).  In similarity, Drew and Mackie (2011) also suggest that there is “a lack of clarity and consensus” (Drew and Mackie, 2011 p.1) to the meaning of pupil autonomy and independence which can impact upon both pupils and teachers negatively. 
In spite of the apparent critique mentioned above and given the attention that has been paid to the integration of child centred learning in curriculums around the world and at home in Wales, it is evident that Donaldson’s suggestions for a ‘New Curriculum’ draw upon a wealth of current academic knowledge and well established curricular achievement. It is understandable therefore the main focus of the small scale research project in this paper was to explore the impact Donaldson’s suggested ways forward increasing pupil independence could have. All research completed was based in a local Primary School and was conducted with pupils in a small set maths class as the focus of the intervention. Full authorization was granted by the head teacher for the suggested research to commence. The research was aimed at 10 primary school age pupils who were identified as disengaged in maths and its aim  was to ascertain whether teaching as a facilitator of learning rather than didactically and establishing a choice driven learning environment would impact upon pupil engagement, attitudes and progress in maths. 
	From the outset, a mixed methods approach was chosen to be the core driver of research design with the main aspects of data collection being: a pupil questionnaire including an initial baseline Likert Scale Standardised assessment, an observational diary which included graded observations (0-5) based upon key characteristics and narrative in its function and a concluding pupil questionnaire which included both Likert Scale standardized assessments and questions which prompted an emotive response. The selection of the methods mentioned above was based upon a plethora of academic reading prior to establishing the design of the study. It seems important therefore to present and discuss some of the most pivotal above mentioned research in greater depth prior to discussing the findings of the project and ways forward in terms of both chosen methods and school based practice.
	When considering the use of questionnaires in formalized research Burton et al (2005) herald the approach suggesting that it is an easy way to ascertain large amounts of data in a short space of time. They also suggest that the data collated can be expressed in a variety of appealing ways and statistical analysis of the information that is gathered can provide opportunities to make generalizable statements. In contrast, however, Corbetta (2003) argues that being able to design an appropriate questionnaire without researcher bias is a difficult task which requires great skill. He implies that questionnaires provide ample opportunity for human error and fail to afford participants with the ability to provide an answer with any form of emotional complexity. In spite of the suggested strengths and weaknesses of questionnaires explored by Burton (2005) and Corbetta (2003), I believed this mode of assessment would be most appropriate within this research context as it was deemed to be an effective way to gather a wide range of information. When considering the age of the participants involved in the study, the accessibility of the assessment was exceptionally important. As pupils within the selected primary school setting were well versed in the use of questionnaires as a mode of expressing their own opinions on a regular basis, it seemed to be an appropriate method in spite of its apparent limitations as the core foundations of participant understanding were already established and would therefore minimize the opportunity for human error as a result of lack of understanding.  
	When reflecting upon the use of a quantitative system within questionnaires, it was apparent that there was literature that both supported and condoned the use of the selected Likert Scale. When discussing the use of a five point scale of ‘Strongly Disagree, Disagree, Neutral, Agree, Strongly Agree’, Likert (1932) argues that the scale system provides a balance in the positivity and negativity of the scale. By including a five point system, the researcher allows participants with the opportunity to feel comfortable as a response is not enforced by including a ‘neutral’ option. Jamieson (2004) also comments that the Likert Scale and the ordinal values that can be assigned to participant responses is very versatile and is utmost effective as it can produce statistical data from an emotive and otherwise qualitative assessment. 
	On the contrary however, Johns (2010) argues that effective Likert Scale data is limited as in order to ascertain data from a scale system it is of utmost importance to ensure that all questions posed are neutral and do not lead participants to respond in any particular fashion. A suggestion is supposedly corrupted by the scale system as he states “Likert Scales by their very nature break all rules of coercion as they are built on the tendency to agree or disagree” (Johns, 2010 p. 5). Johns further discusses the possibility for ‘acquiescence bias’ in which participants have a human tendency to display agreement with given statements irrespective of the contextual information surrounding the posed question. A concept which was clearly depicted and discussed in the works of Schuman and Presser (1996) in which the same question was posed twice within their research scale system, one exaggerating negative aspects of the given context and the other exaggerating the positive with both questions ultimately leading to the same answer. When assessing their results, Schuman and Presser (1996) discovered that participants had altered their answers according to the level of negativity in the question in an attempt to avoid being perceived to be agreeing with a negative statement. In comparison, Johns (2010) also suggests that providing a ‘neutral’ marker in a Likert System promotes opportunities to be ambiguous, ambivalent and indifferent’ and implies that it provides participants with the opportunity to abuse the option to remain neutral in an attempt to ensure that the answers they provide are always deemed to be socially acceptable - this can therefore result in measurement error during the research process.
	Having considered the perspectives above, it was evident that there were reasons both for and against the use of the Likert Scale System. However, in spite of the points made by Johns (2010) and Schumer et al (1996), it was believed the simplicity, versatility and generalizability of the Likert Scale greatly outweighed its negative implications. When reflecting upon its popularity in the world of research and the fact that Likert originally developed the scale to ascertain data from attitudinal behaviours, it seemed like an appropriate choice.  A five scale system was also chosen in order to avoid causing distress to the respondents by enforcing a committed response. Credence was however, paid to the points made by Johns (2010) to a certain extent as the Likert Scale System within the designed questionnaire was complimented by some basic questions that allowed participants to express deeper reasons for their choices. 
		When also reflecting upon the use of observations to collect research data consideration was paid to Bernard’s (1994) reasons for including observation in research design. Firstly, he states that observation allows the researcher to better understand the context of the participant and in so doing develops a greater knowledge of the reasons for their actions or beliefs thus having a direct impact on the validity of research findings. Bernard (1994) also implies that unknowingly observing participants reduces the chances of performativity as people tend to act differently when they feel like they are being watched. He suggests that first hand observation also encourages the secondary questions that are poignant to individual experiences and this therefore makes data collection more purposeful as more appropriate/ contextualized questions can be formed and subsequently answered.
When considering the weaknesses within this method however, DeWalt and DeWalt (2002) present the problems that can arise from the level of researcher interpretation within the context of an observation. They exaggerate the impact biases in terms of ‘gender, sexuality, ethnicity, class, and theoretical approach’ may in fact have upon both the process of observation and the outcome of the chosen method and, in so doing, encourage researchers to understand their own biases and the impact they may have. Moreover, Schensul et al (1999) also argue that in order to observe an ‘unaffected’ representation of any given situation, participants must trust the observer in their entirety as those who experience any form of discomfort or displeasure may respond differently in observations to those who have an established level of trust which would deem them to be ‘observable’. Throughout the literature of both DeWalt and DeWalt (2002) and Schensul et al (1999), great emphasis is placed upon the importance of the role of the observer and the necessity to make accurate notes from the outset in order to avoid the fragmentation of any findings as a result of researcher or participant bias or prejudice. 
Having reflected upon the views of Bernard (1994), DeWalt and DeWalt (2002) and Schensul (1999), it seemed that they had very strong arguments for and against the use of observation in academic research. When thinking about the context of this particular research project, especially the age of the participants, it was decided that observation was to be made use of as it was of utmost importance to observe the participants in their natural environment without the pressure of ‘performing’ which could impact negatively upon results. When considering the core purpose of this study which was to see how engagement, progress and attitudes to learning would be impacted upon by increased pupil independence, it seemed difficult to comprehend a research design that didn’t include observation. By its very nature, observation was seemingly the main research method which allowed participants to be completely free of the strictures of formality and to thrive in their own environment and therefore directly aligned itself with both the curricular and pedagogical facets being explored.
A further method of research to consider was the use of interviews and video based discussions of which Kajornboon (2005) openly discusses their advantages.  She suggests that video based discussions and interviews allow opportunities for other researchers independent to the study to provide analysis after the event and allows analysts the opportunity to revisit responses on a word to word level. Kajornboon (2005) argues that this mode of research is most effective as it directly involves the participant in the data collection process and in so doing provides a more meaningful account of human experience as it is based upon individualized perspectives. In direct contrast however, Mays (1995) implies that one of the most common criticism of this method would in fact be that “it is strongly subject to researcher bias” (Mays, 1995 p.109) and argues that qualitative research methods require extensive record keeping and time for individual scrupulous observation. In spite of the positive attributes presented above, it was decided that the use of this method would not be a valid option. This decision was based upon the suggestion that too much time would be needed to ensure an appropriate meticulous level of attention was paid to the interview transcripts in order to ascertain important details fairly and effectively. In addition, it was felt that the margin of error from the perspective of the researcher was too high and a lack of knowledge in interview coding and analysis was considered. When also considering the age of the participant it seemed important to acknowledge that primary pupils associate being filmed or interviewed with performance and assessment and this could therefore impact upon information collated during the process. 
Having chosen the main methods required to carry out this research project and identifying the ways in which they aligned themselves methodologically, it seemed appropriate to also consider the different perspectives associated with mixed methods as an overall approach to research design. On the one hand, Denzin (1970) implies that mixing methods is the only way to truly emit a detailed understanding of ones findings, going as far as to suggest that a mixture of both quantitative and qualitative results increases the validity of and researcher confidence in their results. Chen (1997) further suggests that as every type of research methodology includes a mixture of both strengths and weaknesses, it is only by mixing approaches that the ‘best of both worlds’ can truly be achieved. As a result of combining multiple methods, researcher gain ‘Triangulation’. Denzin (1970) comments that the triangulation of results that is afforded by a mixed method approach allows for increased validation and overall confidence in research findings and can motive researchers to broaden their levels of originality in terms of research design in order to create enriched ways of ascertaining data, inspiring researchers step outside the confines and strictures of a single methodological approach. 
On the contrary however, Driscoll et al (2007) states that the extensive time and budget needed to invest in a mixed methods approach greatly outweighs any benefits that could be experienced. He argues that studies which make use of mixed methods instantly resign themselves to the fact that extensive analysis will not be possible as the time necessary to analyse multiple sources of information would negatively impact upon a researchers capability to complete deep analysis. He implies that the main aspect of research design that would have to be diminished would be sample size in order to manage the workload of collecting multiple sets of information. Moreover, Saunders et al (2003) also suggest that means of ascertaining qualitative data can in fact undermine the validity of quantitative data as qualitative information is more prone to being affected by bias as part of the research process. 
Having reflected upon the differing views regarding mixed methods as an overall approach to research design, it seemed clear that there was credence in both the reasons for and against its use. However, when considering the size and nature of the research proposed within this context it was evident that the variety available in a mixed approach allowed greater opportunities to identify discrete changes in human behaviour which was vital to the success of the project and that also including a quantitative avenue of assessment, such as the Likert scale, would encourage more opportunities to identify a formal numerical pattern within results in spite of it being small scale research. It was hoped that by mixing both quantitative and qualitative research methods and achieving triangulation of results therefore, the information ascertained during the process would complement one another and grant both the findings and methods used increased validity. 
As a result of deciding upon the methods to be used in this research project, an ethics approval form was submitted to the Ethics Committee and their approval for the suggested design was sought. During this process, great consideration was paid to the ethical foundations of the research proposed and close examination of the BERA (2011) guidelines were established with particular focus being placed upon: anticipated risks to participants, the researcher and institution, the possibility of coercion, modes of consent, levels of confidentiality and the means by which retrieved data or information would be stored and analysed. Approval was subsequently granted by the Ethics Committee as the process detailed within the submitted approval form (Appendix A) was deemed to be ethically sound. 
Prior to commencing the study a representative participant sample was chosen. Ten primary school age pupils who had been identified as disengaged with maths due to underachievement, lacking in confidence but able and over achievement. All pupils were afforded the same opportunities throughout the research process and were made aware that they could opt out at any time. Consent was sought and granted by the Governing Body and the Head Teacher. The project was then discussed with parents of participants at which point written parental consent was gathered. The project was further discussed with pupils in full and verbal and signed consent was acquired and the study took place over a four week period, the design of which was as follows. 
During the initial week of the study, pupils were asked to complete a paper based questionnaire which included both Likert Scale questions and emotive questions (Appendix B) in order to assess their attitudes to the didactic learning styles already embedded in classroom practice and their enjoyment of a specified subject, namely maths.  Within the context of this initial week of the study, the approach to the curriculum encouraged participants to continue with didactic learning activities in maths sessions. Pupils were not afforded with any opportunities to choose their own approach to learning or question types and were led by teacher instruction in their totality. At this time a teacher observation log book was also established in which comments were made based upon pupil engagement, responses to the learning environment and progress through the research period in the mode of scale style questions and observational commentary.
The second and third week of the study saw the implementation of direct pedagogical changes and approaches to the curriculum in an attempt to develop a learning environment in which pupils were promoted to be more active. During this time, ‘Chilli Challenges’ were established as the main driver for learning in the classroom. ‘Chilli Challenges’ were simply differentiated pieces of work under the themes of ‘extra-hot, hot, mild’  and pupils had the opportunity to choose which strength of chilli they completed and the resources by which they intended to do so. Throughout this time pupils were afforded opportunities to challenge themselves and guide their own learning process as teacher input was minimalistic. 
In the final week of the study, research participants were asked to complete the initial questionnaire once more in order to ascertain comparative finding and all scale scores from the teacher observation log book were also collated. Information collected within the process was treated with the utmost confidentiality and when discussing participants or data retrieved from the research process, all pupils were anonymized through the use of terms such as ‘Pupil A’ and ‘Pupil B’.  Tables and graphs featured on the subsequent pages and in the Appendix of this paper, depict research findings and will hereby be formally analysed and discussed with initial focus being placed upon the Questionnaire that was distributed. 
The Likert Scale used within the questionnaire was assigned an ordinal value score which allowed for data to be collated in tables (Appendix C/ D). When initially looking at the collated totals for each of the Likert Scale Questions (see figure 1), the trend in data which is undoubtedly visible is that the intervention had a totally positive impact. Namely impacting upon pupils’ ability to engage with maths, overcome challenges, explain their learning, their overall enjoyment of maths, and their ability to lead their own learning. The biggest impact recorded was equal in kind and was in relation to the pupils’ being able to overcome challenges within the subject and their sense of independence within their own learning. This finding therefore supports the views of Donaldson (2015) and many other academic publishers discussed within this paper which imply that affording pupils with the opportunity to take ownership of their learning and guide their own academic process encourages resilience. A trait which is said to rid pupils of the strictures which are enforced by learning didactically (Dweck, 2012, p 44).  
In addition, and in spite of revealing positive impact, it seems interesting that the two categories which were least impacted by the intervention were pupils’ enjoyment within the subject and their levels of engagement. It is clear that when anticipating the influence of the intervention prior to putting it in place, the largest progress was expected to be in the categories which evidently displayed the least.  As a result of this ambiguity therefore, a more in depth analysis of the ascertained data was carried out and revealed that the depiction of results shown above, although accurate when conveying overall trends, didn’t reveal an in depth picture as only 70% of participants made progress on the basis of individual totals with 30% of participants actually regressing in their final score. 
	When reflecting more closely upon the table of results for the Likert questionnaire before intervention (Appendix C) and after intervention (Appendix D), it seems clear that the ambiguity identified above could in fact be caused by aspects of research design and the impact of unanticipated human behaviour. Upon closer analysis, Likert question one (which focused upon pupil engagement) and Likert question four (which focused upon the level in which pupils liked maths) were the two questions in the questionnaire which pupil’s responded with the largest number of ‘unsure’ or ‘neutral’ responses. Providing a five point scale evidently allowed pupils to be indifferent to certain questions as an enforced response was not implemented. As Johns (2010) suggested therefore, the five point scale could have afforded too many opportunities for ambivalence and that maybe implementing a four point scale could have been more effective within this case as pupils would not have had the option to avoid a committed response – an aspect which undoubtedly impacted upon the final scores within this research process. 
Furthermore; when analysing the emotive responses attached to Likert question three in particular it is evident that, as Cohen et al (2000) imply, the values on the Likert Scale did hold a greater value within the context of a given question as the pupils assigned the level of their answer to the relationship they had with their teacher. For instance, in their emotive response pupils A and C both commented that they didn’t like maths but that it had nothing to do with the performance of their teacher. Given that both pupil A and pupil C chose the ‘unsure’ or ‘neutral’ response it could therefore be suggested that bias occurred due to an unanticipated human emotional response to the given question. Pupils A and C didn’t like maths but also didn’t want to put an answer which could impact upon the relationship they have with their teacher and therefore, as Johns (2010) suggested, were willing to provide a ‘neutral’ response rather than one that they felt would be socially unacceptable. 
When focusing upon pupils F, G and H all of which regressed in total score in the final questionnaire, it is clear that there could be a suggested link between more able pupils and an inability to adapt to change in addition to a reiteration of the power of participant relationships. Pupils F, G and H all scored the highest baseline score in the initial questionnaire with all 3 participants scoring at least one question as ‘strongly agree’. This therefore suggests that, from their perspectives, the original didactic style of learning was effective in granting them a sense of happiness and achievement in maths. Throughout the process all three pupils, who were an exceptionally close nit friendship group and often work together on a daily basis, showed very similar traits and characteristics during the implemented change, all of which compliments the fact that their score from the concluding questionnaire was a depreciation from their initial baseline. When also looking at the final scores for Pupil G and Pupil H it is clear that their final scores are exceptionally similar and that Pupil F scored most of their categories in the same way as the initial baseline. With this in mind therefore, it could be implied that Johns (2010) concept of participants in any given research context feeling the need to avoid socially unacceptable outcomes could in fact be coming into play once again. The contrast in data presented by pupils F, G and H is a reminder of the importance of acknowledging the individuality of all pupils within the learning process. The three pupils identified clearly felt that the curricular and pedagogical change that was implemented during the research process did not suit their learning needs and therefore emphasized the fact that a blanketed approach to change, as was implemented with this context, may not be as successful for some as it would for others. 
When reflecting upon the use of the Likert Scale as a whole, it is clear that the many reasons both for and against its usage are well founded. On the one hand it was an exceptionally quick and easy way to ascertain data from a young audience but given the ambiguities surrounding its design it could be suggested that using the Thurstone Scale (1929) would be a more viable option as it provides an opportunity for each item to be equally weighted whilst ensuring that shades of responses are unable to be given by participants. The emotive questions that were also present in the Questionnaire were effective in providing an opportunity for greater understanding for the participant scoring as they encouraged pupils to justify the reasons for their choices. As Johns (2010) suggested, they provided the participants with a greater opportunity to express the complexity of their emotions in relation to the different question themes and in so doing, a greater sense of understanding was gained from the process. This aspect of the assessment that was created for the project would not change if the research were to be repeated in the future as it was invaluable to the interpretation of the final data gleaned from the intervention.  
	Throughout the research progress, the level of challenges completed by each participant was monitored in order to ascertain whether any progress would be made in terms of Curriculum Years during the implementation of the ‘Chilli Challenges’. The method by which this information was collected was a sound process and one that would be completed in the future – it was completed without the awareness of the participants in order to avoid coercion and was merely an assessment of which ‘Chilli Challenge’ had been completed within maths lessons for each participant.  When considering the findings, Figure 2 shows 60% of pupils made progress in the year group skill covered across maths lessons during the formal intervention stages of which 20% made excellent progress. However, 40% of participants made no progress of which 10% actually regressed in skill level during the research process. With this in mind therefore, it is very difficult to suggest that the implementation of independent learning lead within this context provided any coherent evidence to suggest it promotes greater success in pupil attainment and academic progress. In spite of the fact that a majority of participants made some form of progress throughout the research, the minority that didn’t were equivalent to 40% of the sample. With the inconclusive and juxtaposing nature of these findings in mind it could therefore be suggested that the connection between independent learning and academic progress is not as reliable as was originally considered.  
	When reflecting upon the overall progress made in the observational diary scale assessments, it is clear that a very similar pattern emerges as of that in the Likert Scale graph (see figure 3). Upon consideration, when looking at the trend across the total scores associated with pupil engagement, ability to overcome challenges in maths and the ability to explain learning, it seems that it is one of positive progression. All total scores relating to the main areas of observation progressed in a very similar fashion across the process which is indicative and supportive of the change as shown in Figure1. However, upon closer inspection of the individual data collected (see Appendix E/F/G), it seems that some variation was present. 
When analysing the individual findings, it is apparent that 90% of participants made some form of progress throughout the process with only one participant showing a regression in observational score - pupil H.  However, 40% of participants, namely pupils C, I, F and G all made limited progress throughout the observational period, some of which can be explained and explored further. As previously mentioned, pupils F and G were mainly disengaged with the process and openly displayed their dislike for the new approach to the curriculum and so their limited progress was to be expected after their regression in score in the questionnaire.  When analysing the scores displayed by pupil I and pupil G it is clear from the observational comments made that poor confidence levels massively impacted upon their overall score progression as it wasn’t until the final observation week that they had adapted to the change in curricular and pedagogical approach and started to embrace the opportunities afforded to them by increased learner independence. In spite of the small progress made by some pupils, however, it is evident that the behavioural and attitudinal attributes displayed by the majority of participants during the research project changed as a result of the introduction of an independent learning environment and increased pupil responsibility.  
	When reflecting upon the use of observation as a research method, it was evident that it was an invaluable tool when trying to assess the impact of independent learning upon pupils in the Primary setting. Observing participants provided opportunities to analyse real human emotion and behaviour and even though it wasn’t fully quantifiable, it was essential to understanding the ways in which curricular and pedagogical changes could impact pupils. If this research project were to be revisited, it would be useful to also introduce and element of recording in order to, as suggested by Kajornboon (2005), re-visit participant actions when trying to analyse the qualitative information ascertained from observations. As argued by DeWalt & DeWalt (2002), there were some limitations to the ways in which the qualitative data could be analysed and it was difficult to draw a conclusive pattern of behaviour for each individual participant from observation notes. However, in spite of its limitations, it would have been difficult to truly assess the impact independent learning had upon the attitude and behaviour of the participants without the use of observation. 
	When reviewing the methods used, it seems appropriate to also consider the effectiveness of the mixed methods approach. When reflecting upon Figure 4 it is evident that employing mixed methods mainly provided information that was complimentary with the only anomaly being equivalent to one point which is minimal to say the least. It is clear that the ability to triangulate results by using mixed methods in fact granted the findings of the total score trends more credibility. However, it could be argued that scores primarily formed on an individual basis could in fact undermine this suggested coherence as they present greater variability. With this in mind therefore, it is clear that the debate that surrounds the use of mixed methods in research has evident grounding. 
On reflection, epistemologically and ontologically speaking, this research project initially displayed aspects of both the positivist and interpretivist view of the research process and was therefore complimented by the use of mixed methods approach. On reflection, choosing one avenue such as the positivist approach could have resulted in data that was less varied and in turn, more definitive. If further quantitative information was gathered the margin for variation presented by qualitative approaches could have been minimalized. However, it is questionable whether the same level of understanding and knowledge of how independent learning impacts upon the behavioural and attitudinal aspects of engagement and enjoyment of a subject would have been able to be gleaned from the research process if a mainly positivist stance was adopted throughout. 
	In spite of the ambiguities discussed above therefore, it could in fact be argued that that a correlation between independent learning and increased pupil engagement and enjoyment has some validity. The total data collected through both the questionnaire and observational scales noted a common trend of consistent improvement in the majority of participants. When reflecting upon academic progress however, it is evident that the data collected was more inconclusive and it is therefore difficult to suggest the research findings have helped to prove or disprove the concept of independence leading to academic success. 
Having completed this research project and critically analysed the chosen methods and its findings, it is clear why Donaldson (2015) places great emphasis upon the need for a curriculum which promotes independent learning. The findings of this study, although not wholly conclusive, do indicate towards the power of pupil independence and begin to explain the  reasons why such focus has been placed upon it, not only with the context of a new developing curriculum for Wales but in the wider world of international curriculum design. The body of knowledge and understanding of the impact of independent learning that has been gleaned from this research project has led to the development of the following recommendations for future practice. All recommendations have been designed in accordance with suggestions made by Donaldson (2015) in his review of Welsh education and growing knowledge of the ‘New Curriculum for Wales’.  
Recommendations for Future Practice: 
1.	Continue to promote opportunities that allow pupils to develop independence and encourage teachers to be facilitators of learning rather than directors of it.
2.	Allow more time to promote a creative curriculum rather than a didactic one but never lose sight of the fact that all learners are different. Even in the wake of a movement away from didactic teaching, such change cannot result in blanket teaching as all pupils are independent in their own learning styles and will demand support in different ways. 
3.	Still provide teacher support for pupils that need help to adapt to the change, mainly more able pupils, as we don’t want to disengage learners in the name of independence.
4.	Consider the development of classroom areas across Key Stage Two which promote independent learning holistically and allow children the opportunity to acclimatize to the change independence will bring prior to the formal introduction of the ‘New Curriculum’ in 2020. 
5.	Continue to develop the role of teachers as researchers in school by sharing findings and experiences of this research project.