Foreign Language Teaching to Romani Students Through the Use of Differentiated Instruction and Portfolio Assessment. A Case Study.

The socio-cultural changes taking place in Greek classrooms render the modication and the revising of the existing teaching material imperative. In classes with cultural minorities students’ needs are put into the forefront and the teaching objectives are also adapted accordingly. However, such a challenging project requires trained teaching staff and the provision of relevant guidance. The aim of the present study is to investigate the effectiveness of Differentiated Instruction and Alternative Assessment in the teaching of English to diverse population classes. The research focuses on the effect the aforementioned approaches have on triggering students’ motivation and overall linguistic development. In particular, the study is an Action Research examining the outcomes of two groups of students attending the third grade of primary school. The experimental group was being taught through differentiated instruction, while the control group attended the lessons of a traditional class. For the collection of data several tools were employed regarding qualitative and quantitative information in order to render the conclusions reliable. A Needs Analysis questionnaire was distributed to the students, a pre- and post- test was also assigned, portfolio evaluation checklists were administered and student semi-structured interviews were conducted interpreted and analyzed. The ndings of the study revealed the astounding effects that the implementation of alternative hands-on tasks and holistic skills development had on Romani students’ participation and foreign language learning. More specically, the young students were more motivated and willing to become actively involved in the learning process.

Introduction welcomed by the students' community as they offer an alternative to tests. However, they are something more than just an assessment tool; they are a motivating source. Berrett (2005) discusses the multiple purposes a portfolio may have depending on the context. Portfolios motivate students, as they trigger their imagination through authentic and original tasks. Consequently, motivated students tend to engage in class activities, devote time for developing their metacognitive skills and reach out for help when needed (Jones, 2009). This type of motivation is important especially when dealing with students or classes with special characteristics, as for example culturally diverse classes or students with learning di culties. Teachers should take certain parameters into consideration in order to trigger students' involvement in the teaching/learning process. Garton et al (2011) point out that tasks should be short including visuals and music. The addition of the element of play, so that young students' interest will be enhanced, is also considered important. Furthermore, Cameron (2015) highlights the importance of students' development when designing and implementing material, as well as, the addition of interests which young students may have. When applying portfolios in the class, students are not con ned by strict time limits and they constitute the core of learning, and more speci cally autonomous learning. Tuksinvarajan and Todd (2009) refer to the fact that portfolios encourage students to document their learning and observe and estimate the strategies used, in order to acquire knowledge. It is worth mentioning that specially designed checklists contribute to this idea and facilitate the process of self-monitoring. Based on the aforementioned literature, in the particular context the portfolio application included engaging tasks and students devoted time to interact with their work and opportunities to evaluate themselves, as well as the material, through specially designed checklists.
Taking into consideration the Greek school reality and in line with the relevant literature, the researchers designed a series of lessons for the third grade of primary school according to the needs and unique cultural background of the particular group of students. The purpose, of these lessons, was to achieve optimal participation and to engage young students in the learning process. De cient attendance combined with a lack of motivation, on the part of the Romani students, constituted the core of the lessons which were designed to trigger their interest. The study underlines the importance of the instructorstudents strong academic relationship and the application of the appropriate strategies in order to encourage attendance and participation (Burnette, 1999).
Understanding the role of culture as an in uential factor regarding learning because it will facilitate as well as accelerate the teaching/learning process of the student (Hartman, 2016).
Lastly, due to the limited research availability, concerning the provision of motivation to young Romani EFL students, the current study aims at shedding light and examining the extent to which DI and alternative assessment motivate students and provide a stimulus for the development of their language skills, as well as a way to explore the different ways students of diverse populations learn a foreign language when alternative teaching practices are employed.

Methods
The research conducted was based both on qualitative and quantitative data. In AR there isn't a wide use of quantitative methods; however, for reasons of objectivity and closer adhesion to the scienti c method, quantitative tools were also used (Burns, 2010, Taylor, 2005. Qualitative methods are more prevalent in AR; as Glaser and Strauss (2017, p. 18) claim, the advantages of qualitative data collection are numerous highlighting that it is the most "adequate" and "e cient" tool for optimal information obtaining. Consequently, a mixed data collection was used which attempted to increase the research validity and objectivity and achieve 'triangulation' (Ferrance, 2000, Sagor, 2000. More speci cally, the numerical data of the current research was collected through quantitative methods while the non-numerical with qualitative. Since the methods elicit responses which are "true" and "relevant" to the purpose of the data collection, both are valid and contribute to triangulation (Taylor, 2005, p. 5, Dörnyei, 2007. The current AR was planned in accordance with Burns' (2010) suggestions on the research features: the purpose, the data collection, the resources and the time span needed. At the beginning of the study the researchers set the following research questions: 1) Will alternative teaching methods motivate students who are learning English as a foreign language towards the teaching/learning process?
2) Το what extent will alternative assessment promote English as a foreign language learning in the classes under scrutiny?
3) Will the students be able to develop all their skills? 4) Will differentiated instruction prove to be more e cient in terms of students' progress?
The Action stage starts with a NA 1 questionnaire in order to identify students' interests and attitudes towards the English language. There follows a pre-test evaluating students' competency level before the implementation of differentiated instruction and then, a series of differentiated lessons including tasks which are in accordance with the principles of differentiated teaching. After the completion of the lessons, the students are given a post-test -identical to the one which preceded the instruction-and further data was collected through semi-structured interviews. The aim of the aforementioned methods was to acquire the fullest picture possible to re ect upon. remigrants are bilingual speaking Greek and Russian and the Albanian student's L1 is Albanian and Greek is his L2. Due to the nature of the population, attendance is not stable, and the teacher is obliged to invent ways to attract the students' interest, and improvise in order to encourage attendance and participation. There is also diversity concerning their students, as well as their level of competency. Furthermore, the majority of the students come from impoverished families relying on bene ts, and there are frequent incidents of theft, as well as violence among students. Due to the nancial problems of the students' families, the lack of materials such as notebooks, pencils, folders are a usual phenomenon. Additionally, Romani students are deprived of stimuli and they present di culty in staying in the class or listening to instructions for long. Most of the Romani students also attend reception classes, as they have ignorance of the Greek alphabet and they can barely read.
Companies and industries operating in the area donate material to students. There is also an active parents' association struggling to cope with the challenges emerging. Moreover, the school participates in volunteer programmes and the students are provided with meals by European programmes.
Regarding the school building, the classrooms are prefabricated. There is lack of basic equipment and the existing equipment is in poor condition.
The seating arrangement alternates in order to trigger students' interest and ensure regular teaching/learning. Consequently, 'horseshoe', group and frontal arrangements are applied according to the needs.

The control group
The control group is the second class of third grade in the 6th Primary School of Aspropyrgos. Regarding learners' background, it is similar to the experimental group. There are 19 students (8 boys and 11 girls). The cultural diversity exhibits resemblance to the other group, that is 15 are Romani, 2 are Albanian, 2 are Russian remigrants and there are no Greek learners. Reception classes are also provided to the Romani students of the control group.
Similarities are also observed in the frequency of attendance due to the nature of the population.

The research procedure
The present research was carried out from October 2018 to January 2019 and lasted approximately 25 teaching hours. The researchers designed the lessons in accordance with the DI principles, as well as the TBL 2 approach. Portfolio assessment was also applied, as a motivating means. A pre-and a post-test was conducted in order to evaluate the results of the DI and the concluding step was the semi-structured interviews of the young students

The lessons
The instruction to the control group The control group was instructed only through the coursebook 'Magic Book 2' which is distributed to all students in Greece, free of charge. There was no differentiation in teaching but rather 'one-size-ts-all' instruction. The alphabet was presented through pictures followed by rhyming songs. There were also amusing familiar comic-strip stories through which the young students were presented with the new vocabulary and were taught reading which was integrated with listening through the audio material. Consequently, the Phonics approach facilitates the students to familiarize themselves with the Latin alphabet (Scott and Ytreberg, 1990). The oracy and literacy skills were developed through interesting activities regarding songs, crafts or dramatizations. All the students were encouraged to participate and further support was more often than not necessary.
As far as assessment is concerned, students were not given any form of test and their assessment was based on their overall performance and mainly on their participation and effort. This choice was made because the young learners were not familiar with testing. Additionally, their background with the entailed di culties, in combination to their de cient attendance, hindered learning, thus, resulting in low performance. Consequently, failure in a test would probably result in lack of motivation and participation.

Instruction to the experimental group
The lessons for the experimental group were designed in accordance with the NA questionnaire completed by the third graders. Thus, there was an effort to re ect on the students' interests, and aimed at massive participation. The approach considered most appropriate for the current context was DI. The lessons followed the guidelines of IFLC and were in accordance with the objectives set by the textbook. Since the purpose of the lessons was motivation, the tasks follow the principles of DI and included the elements of game and sensory-motor development (Cameron, 2015). Furthermore, TBL was applied; the core of the process is the student who develops an understanding of language as a tool rather than a goal (Task Based Learning, 2009). Tasks convert classes into amusing and interesting experiences while students use their language skills at the current level (ibid.).
Lesson 1 entitled The Alphabet aimed at the enhancement of the English Alphabet through group-work and craftwork. According to Case (2012), the implementation of lessons with craft encourages students' motivation and has a positive effect on discipline problems. In the pre-task stage students were motivated through audiovisual material; they watched and listened to the Alphabet song and they also sung along. During the Task Cycle students, worked in groups, and were asked to make an alphabet 'bean soup'. More speci cally, they wrote letters on beans and put them on a piece of cardboard. Then, they practiced the song they heard in the rst part of the lesson and presented it to the rest of the class. Through this craft students practiced their writing skills while cooperating. Language focus was not presented in the strict sense; because of the nature of the class, as well as the students' level, this part focuses on feedback provision and further enhancement of the Alphabet letters through a matching activity. Lessons 2 and 3 are entitled Colours and extended into two teaching hours. The aim was to activate prior knowledge on colours and encourage students' participation and cooperation through group-work. During the rst class, the pre-task and part of the Task-cycle phase were realized. With the use of ashcards and strips with words, the instructors attracted the students' interest and promoted the development of the skills of reading, listening and speaking. Kinesthetic students were encouraged to participate with the next activity which required students to read the word-strips and arrange them appropriately on the board. Decision-making was also promoted through a group activity in which students were asked to make an 'ice-cream' colouring paper-scoop and to write the correct colour. Students voted for the best ice-cream and their work was hung on the classroom wall. In the language focus phase, there was skills' integration. The students were handed-out black and white pictures and were asked to colour the picture according to the teachers' instructions.
Lessons 4 and 5 aim at introducing the textbook heroes, as well as practising speaking by introducing, and writing names. In order to achieve the set objectives, sock-puppets were employed as a motivating force. Through the use of puppets, the instructors communicated with and motivated the students regardless of their age, language level and cultural background (Peck, 2005). Marshall (2013) suggests that puppetry encourages and promotes the development of imagination and creativity. In the pre-task stage, the teachers introduced their puppets in order to set an example and to trigger theirs students' imagination. In the following stages the students created their own puppets, presented them to their classmates and "discussed" with other puppets imitating the questions produced by their teachers in the rst stage. In the Language focus stage, the students were asked to practise writing their names on worksheets and also to form their own names using clay.
Lessons 6 and 7 dealt with the enhancement of numbers (1-10). Apart from visual aids ( ashcards), the educators employed the element of rhythm by using maracas in order to rhythmically pronounce the numbers. The students were encouraged to take the role of their teachers and to play the maracas while their classmates called out the numbers in English. Next, students were engaged in a board-game task. Through the popular game of 'Snakes and Ladders' the young students developed a sense of cooperation and practice counting in English. The pedagogical value of board-games application is discussed in Mehregan (2014) who highlights the stress-free conditions created and discusses the enjoyable way in which students learn, thus, contributing to better vocabulary enhancement. The language focus is realized through computer games which offer further practice. According to Vasileiadou and Za ri (2017), online computer games play a motivating role and contribute to optimal vocabulary understanding.
Lesson 8 entitled Kelly (My name) focused on introducing reading textbook stories by using the puppets made in a previous lesson. Firstly, students practised listening to the lesson in order to arrange a set of pictures. Afterwards, the young students simulated the dialogues using their puppets and presented it to their classmates. The last stage aims at practising simple expressions found in the text (Who's this? I'm strong! etc.) and promoted a relaxing atmosphere through mime playing. Lessons 9 and 10 had the title Kelly: A magic key and aimed at the development of reading, speaking and listening through dramatization. Initially, the teachers encouraged vocabulary learning with the use of realia. Craftwork and dramatization were combined in order to trigger motivation and achieve participation. The use of drama in the EFL class presented multiple bene ts (Boudreault, 2010); students developed a sense of con dence and performed better. Additionally, the new vocabulary was fully contextualized and consequently more meaningful (ibid.). In order to perform the dialogues, the students created masks depicting the heroes of the lesson. The last part of the session focused on understanding prepositions of place (in, on, under) through a hiding game.

Data collection Needs Analysis Questionnaire
The process of language learning is not always linear; it involves gradual development and unexpected outcomes (Long, 2005). As a result, curriculum adaptations and special material designing could become integral parts of the teaching/learning process. Such a step presupposes the collection of appropriate information related to students' attitudes and needs through NA (Richards, 2001). Putting the student in the forefront of teaching and embracing student-centeredness, stakeholders employ tools for the identi cation of needs, attitudes and perceptions related to language learning (Nunan, 1988). Amongst the methods used (meetings, observations, interviews, questionnaires, etc.) the questionnaire exhibits a wide range of bene ts. The preparation of a NA questionnaire could be considered easy and the analysis of the information obtained can be realized through simple steps. Additionally, it can examine numerous issues and the participants can complete it easily (Richards, 2001).
The NA questionnaire designed for the current research was distributed to the experimental group. Although the students were provided both with an English and a Greek version of the questionnaire, they preferred to complete the one in Greek, as it was less demanding in terms of language. Due to the students' level and age, there were only three parts and most of the items were structured. The students were also provided with an unstructured form of questions where they were encouraged to add or even draw further suggestions. The completion demanded approximately 10 minutes. In order to ensure the time needed, the questionnaire was pilot tested by two colleagues.

The Pre-and Post-Test
Pre-and post-tests are considered useful tools to diagnose knowledge prior and after the implementation of an approach, and they indicate the degree of the improvement accomplished (Dimitrov and Rumrill, 2003). More speci cally, pre-tests are used in classes with diverse students and aim at the identi cation of problem areas and ine cient teaching methods. Students are not expected to excel in a pre-test. Its main function is diagnostic; it unveils topics that have been conquered or knowledge which still needed further elaboration (Kuehn, 2019). Post-tests are usually identical to the pre-test which preceded. Through a post-test the teacher can measure the e ciency of a method which is being followed and provide help to the students who need it the most.
In the particular teaching context where the students' level is low and they have never been tested before, it could be argued that the level of competency is pre-A1 (CEFR, 2018). Consequently, it was considered appropriate to design a test using the pre-A1 Starters 3 , as its framework. The test consisted of three sections. In the rst section the students were tested on their listening skills; the rst part included two listening activities: part 1 was a multiple-choice task where students were asked to listen to the teacher reading short and simple dialogues and were then expected to circle the correct answer. In part 2 the testtakers were asked to listen to a short description and to colour a set of items according to the instructions.
The next section assessed reading and writing skills. In part 1, students were given pictures followed by words. The students were asked to look carefully at each picture and to read the word. If the word was correct, students were asked to tick the box, otherwise they were expected to put an X. Part 2 included a set of pictures. Next to each picture there were words with scrambled letters. Students were expected to put the letters in the correct order and to form words.
The last section of the test was a speaking task. Here, students were given a picture and were asked simple questions regarding their age and the number of objects in the picture. All the sections included pictures with vivid colours as they attract students' interest and encourage motivation (Shabiralyani et al, 2015). Consequently, in the case of this research, pictures seemed to contribute to a more relaxed atmosphere for the young inexperienced students.

Portfolios
The new curriculum promotes alternative methods of assessment as more suitable for young learners. Despite the fact that portfolios are time-consuming to prepare, they can also prove to be valuable tools which encourage a stress-free assessment of students' competence and development. The use of portfolios was also considered ideal regarding the teaching context under scrutiny, as portfolios encouraged motivation and helped Romani students to acquire a more positive attitude towards learning (Axton, 2012). As Ryan and Deci (2000) so rightfully put it, intrinsic motivation stems from the appropriate circumstances; more speci cally, students do not have to run into failure but they are responsible for their work, its selection and organization, this was also the case of the experimental group of this research.
Coming to the present research, the students have never previously been assessed in the foreign language. Additionally, the frequency of their class attendance could be characterized as low due to students' lack of motivation.
Regarding the procedure followed for the application of portfolios, it lasted for about two months in the rst school trimester. The steps followed were those de ned by Kemp and Toperoff (1998), namely: 1. Teaching goals identi cation: The main objective of portfolio application is motivation encouragement. However, linguistic competence is not neglected.
Skills development is also encouraged through the tasks which are applied in the class.
2. Format speci cation: Each student had his/her own folder which they decorated. According to Georgousis (1990, p. 30-36), this is a way to promote the "feeling of ownership", and, more extensively, to engage students in the process. Consequently, the format is class-based.
3. Portfolio introduction in class: This step required one teaching hour. Students were given information about the process to be followed, as well as the purpose of the application. A fruitful class discussion was realized where students asked for information and clari cations.
4. Content speci cation: In a new session, the students were presented with the contents to be included in their portfolios. The base of the portfolio format is the ELP 4 but with several modi cations and simpli cations, thus, rendering it suitable for the students' age, language level and cognitive development. Consequently, it was divided into different sections: a. Contents page.
b. My Language Passport. Here, students presented simple personal information about themselves.
c. My Favourite activity. In this section students attached material regarding an activity performed in class which was found amusing. It was followed by a checklist where students re ected on the activity. These forms reinforce autonomous learning and self-evaluation.
5. Presentation of guidelines: For this step focus was placed on the importance of organization and layout of the portfolios. Students were encouraged to take care of their work. Due to frequent incidents of theft, portfolios were gathered and kept safely away, after each session.
6. Noti cation of other parties: The school principal was informed about the new method to be applied in class and the latter was fully approved. Moreover, a letter of consent was sent to the parents so that they could be informed about the application of portfolios and so that they could give their permission.
7. Preparation period: The time required for the application was demanding. Although it was principally arranged to devote one session per week, more often than not students asked for their portfolios in order to complete something or in order to organize them. The teachers' who were also the researchers worked as facilitators providing help and feedback when needed.
8. Assessment and feedback: As far as assessment was concerned, the portfolio entries, selected by the students themselves, were assessed through checklists. Through this process, students were given the opportunity to self-re ect, and to take responsibility of their own learning (McDonald, 2011). The forms were in Greek as the students' English language level would hinder the completion of a form written in the English language. Furthermore, the answers were mainly presented visually as visuals are considered motivating factors.
Regarding feedback students were encouraged and advised throughout the process. The objective was to activate students to become engaged in portfolio application. At the end of the set period, the students were nominated with certi cates so that they could develop a much-needed feeling of achievement.
Parent-teacher communication was oral as the majority of the parents could not read because they were illiterate. Additionally, this was found to be a more direct and fruitful means of information provision.
Semi-structured interviews were also conducted for the realization of data collection. Qualitative methods enable researchers to explore the purpose, the context, and the meaning of the research (Van Teijlingen, 2014). Edwards and Holland (2013) claim that semi-structured interviews are exible, in the sense that, no structure is required. The researchers addressed a set of open-ended questions accompanied by further queries which worked as facilitators (Piercy, 2004). These questions allowed the interviewees the freedom to talk about whatever they considered as more important following the general guidelines (Robson, 2005).
In the current research, the semi-structured interview questions were selected in order to diagnose the attitude the young students developed for English language learning after the realization of DI, and identify the stronger and weaker points of its application. In designing this research method, the steps suggested by Laforest (2009) were taken into consideration. The researchers identi ed the participants and de ned the number of interviewees. The interviews were conducted individually so that the shier or less willing students could participate more eagerly and comfortably and would also not mind if they were recorded on a mobile phone.
During the preparation of the interviews the respondents were informed about the process to be followed (Laforest, 2009). In this case, the school principal took cognisance and approved the realization and the recording of the interviews. The students were also fully informed in class, as well as before the interviews showing great enthusiasm and willingness.
Due to the lack of room availability, the interviews were conducted in the 'Teachers' O ce' when it was vacant; this resulted in occupying the participants beyond the English classes. The duration of the interviews did not exceed 5 minutes for each participant. Regarding the questions, they were asked in Greek so as to avoid intimidating the students and because the language level of the interviewees would not allow a fruitful discussion in English (Cohen et al, 2007

Results
The results arose from the data provided by the NA questionnaires, the pre-and post-tests, portfolios and semi-structured interviews which ensured the triangulation of the research and the validity of the study (Guion et al., 2011).

The Needs Analysis Questionnaire Results
The NA questionnaires aimed at collecting the information needed in order to identify students' preferences, strengths and weaknesses. The results were used, so as to proceed to appropriate curriculum adaptations and to improve their learning process.
The quantitative data collected were divided into three sections according to the parts of the questionnaire distributed.
Part A: In the rst part, there were six questions deriving information on the students' stance in English, as well as their opinion on its usefulness.
Ten students of the experimental group participated in the process. Although the total number of the students who enrolled was 18, there was great attendance out ow. The majority of the students (80%) answered that they like English and the specimen, in total, showed their preference in using their L1 in class. As far as the di culty of the target language was concerned, the results portrayed a dichotomy; half of the participants answered that they faced di culty in learning English while the other half claimed the opposite. Moreover, the vast majority of the specimen (80%) answered that they remembered knowledge which they had acquired in previous years, and most of the students expressed the opinion that English will prove to be useful for them in the future.
Part B: The second section of the NA questionnaire contained further details, on the parts of the learning process, which the students found di cult and their preferences in the way they wished to learn the target language.
The reading, writing and listening skills were viewed as equally sophisticated, by the young students. Acquisition of new vocabulary was considered to be less di cult, while two students commented that none of the elds of EFL were di cult for them. Tzotzou (2014), in a NA research which concerned 5th grade students' motivation, preferences, and needs, it was found that, amongst students of primary school, the listening skill was the least favourite, while they identi ed no di culties in reading, writing and vocabulary. In the present study, since the students' population was radically different from the students who participated in the aforementioned study, the ndings are dissimilar, showing that all the skills are met with lack of con dence.
Next, students were asked about their preferences on the ways they like to learn the target language and showed great interest in alternative teaching methods ( Figure 1). Learning through stories was not as popular as learning with the use of New Technologies, which was met with enthusiasm (21%). Games, songs, videos and painting were also selected, with the latter being more preferable (17%). It is worth mentioning that two students also liked writing as a way of learning the foreign language. The results presented by Tzotzou (2014) portrayed a massive preference towards learning through games and using New Part C: The last part of the questionnaire asked the participants to comment, in writing, or through drawings, on what they enjoyed the most while they were learning English. This question was asked because we wanted the students to express their opinion freely on the particular topic. Many students drew colourful pictures which were irrelevant to the question while others commented (or drew) that they liked painting with friends or writing. Yaman (2018) conducted a similar research to primary school students in Turkey, and studied their perceptions of English as a foreign language through drawings. Their ndings showed that primary school students perceived English as objects and, thus, they drew objects -mostly -as regarding the vocabulary which they were taught. In the current research, although the aim of Part C of the questionnaire was also to identify the general attitude towards EFL, there were quite a few examples of students depicting objects. This may have stemmed from the fact that the students, under scrutiny, had not connected English language learning with vocabulary learning but with situations and experiences.

Pre-and Post-Test Results
Before the application of DI, both the experimental and the control group were given a pre-test so as to identify their competency level. The same test was provided to both groups after the intervention so that their progress could be evaluated. Both groups exhibited a similar performance in all the parts of the test ( Figure 2). More speci cally, in the Reading and Writing parts the control group performed slightly better (7.25%) while in the Listening and Speaking parts, the experimental group presented higher results (6.63% and 8% respectively). In general, the performance of both groups could be characterized as low with instances of candidates who did not reply to any of the items or who replied wrongly in all the items of a speci c part. However, this might have happened due to the students' unfamiliarity with testing processes as well as their low language level. Moreover, their low performance in speaking and listening could be attributed to the fact that teaching emphasis was not given to these two skills as greater focus was placed on their literacy skills (Ulum, 2015).
The post-test ndings present great interest as both groups showed improvement in almost all the sections which were being tested ( Figure 3). However, it was evident that the number of the participants had decreased. Reading and Writing was the part with the clearest development. More speci cally, the performance of the experimental group exceeded that of the control group (23%) while in the pre-test the latter showed better results. Moreover, as far as the Speaking test was concerned, there was a slight decline in the results of the control group. This could be explained by the lack of adequate spontaneous oral practice which was triggered in the experimental group through task presentations and group-work. Kalogeri (2016) conducted similar research with Romani students who attended the third grade of primary school and who were also studying English as a foreign language. After the application of DI, she also observed that the students of the experimental group showed remarkable progress in the test undertaken. However, the test assigned by Kalogeri presents differences compared with the pre-and post-test of the present study.
The Results of the Semi-Structured Interviews The nal part of the research was consisted with the semi-structured interviews which were conducted only with the students (10) of the experimental group. The language used was Greek because of the students' English language level and for reasons of motivation. Alshenqeeti (2014) claims that data which is derived from semi-structured interviews cannot be analyzed through any established method. Creswell (2014) suggests that such an analysis could be feasible if the researcher classi ed the questions through axial coding; more speci cally, the interviewees' experiences could be classi ed and presented according to their degree of importance and similarity. In the present research the interviews were transcribed, classi ed, and studied. The results were coded into three different axes: personal information, attitude towards the target language, perceptions concerning the students' skills and usefulness of the target language in real life.
In the rst axis, the participants were asked general questions about personal information (Laforest, 2009). This section worked as an ice-breaking part so that the young interviewees would become familiar with the process. The questions did not require particular skills and did not trigger discussion. The students replied on questions concerning their gender and age, in one word only. According to the answers provided 60% of the participants were females and their average age was eight and a half.

Portfolio Application Results
The application of portfolios in the class provided the researchers with qualitative data. However, at the end of the research period the students were handed out a checklist in order to evaluate their portfolios. The design, as well as the content of the checklist was an adaptation from Kemp and Toperoff's (1998) suggestions. All of the students claimed that they had included everything needed in their portfolios. The vast majority (70%) answered that they have kept their folders organized. The participants, in total, regarded portfolios as motivating tools for learning while about the same percentage (90%) claimed that they were encouraged to engage in the learning process through the use of portfolios. In a research, which dealt with ELP application to young English students, Perclova (2006) presented ndings on the students' attitudes towards alternative assessment methods. More speci cally, the majority of the participants evaluated the ELP positively and characterized it as interesting. It is worth-noticing that portfolios, in the aforementioned research, were particularly popular amongst primary school students in comparison to lower-high school students. Finally, Barabouti (2012), reported on a research concerning portfolio application to primary school EFL students, and more particularly he suggested that the students considered good organization and clear presentations as important features of their portfolios. These ndings come to reinforce the results of the present study presenting a wide acceptance of portfolios by the young learners' community.
The second part of the checklist included two open-ended questions. Students were expected to comment on the parts of the portfolio they liked and to make suggestions for improvement. The majority of the participants did not provide any answer. From the answers which were given, the most popular was that the best part of the portfolios were the paintings which they did. Regarding further suggestions, students claimed that they would like to include more pages in their folders which leads to the conclusion that the speci c assessment tool encouraged students to participate in the learning process by taking control of the material and the organization of their work. It is worth mentioning that during the research period the students were particularly willing to devote time on their portfolios. Barabouti (2012) explains that this acceptance may be due to the fact that portfolios provide students with fairer learning opportunities.

Discussion
The present research suggests that the application of DI and portfolio assessment can have a positive effect on the motivation of culturally diverse students, and more speci cally on the Romani students. Since this unique cultural group is given stimuli and motivation to participate, the results are encouraging.
Similar conclusions were presented by Koraki (2017), who applied alternative teaching techniques (i.e., stories and games) to Romani students. The researcher observed a signi cant improvement in terms of students' motivation which resulted from the provision of original and challenging experiences.
Taking into consideration the research ndings, the students of the experimental group exhibited greater interest towards the EFL lessons. Additionally, there was more regular attendance compared with the control group. This could lead to the conclusion that the learners of the experimental group were given more motives and opportunities in combination with the creation of a stress-free environment (Sougari & Sifakis, 2007). In the current study, the teachers/researchers had designed a series of tasks which included crafts, puppetry, board-games, drama. The young learners welcomed the unprecedented methods with joy and, despite the di culties arising, the ndings portrayed progress. Finally, most of the tasks required students working in small groups or pairs; group-work encouraged students to practice the interactive foreign language in an affective climate (Karim, 2015).
On the other hand, the traditional class did not exhibit the same progress. This leads to the conclusion that in a culturally diverse class the adhesion to the curriculum may result in limited progress. The students of this group did not show the same enthusiasm as the students of the experimental group and the attendance was more irregular. As stated in Kalogeri (2016), when the teaching context is not motivating, the (Romani) students exhibit disruptive behaviour and lack of discipline.
Moreover, through portfolio application young students developed a sense of ownership and did not experience the fear of failure as happens with traditional forms of assessment (Georgousis, 1990, Ryan & Deci, 2000. Students' autonomy is also promoted since each student has the control of his/her work a fact that plays a signi cant role in establishing trust. Despite the positive results which emerged from the current study, the analysis of the limitations is important. The current research was a small-scale action research conducted to a limited sample of the population. According to Dörnyei (2007), the ndings of the AR could not be generalized, a fact that in uences the external validity. In addition to this, the validity is also affected by the fact that the population of the research was not stable. Due to the students' cultural background, there was great attendance out ow; the number of students, were never the same. Moreover, there were students who came very sparsely and could not keep up with the rest of the class. Nevertheless, certain adaptations were made so that all the learners could participate.
Another limitation in the current research was the lack of equipment and materials. The school could not provide all the equipment needed. In addition to this, due to the low nancial status of the learners' families, there was di culty in material provision. In order to conduct the research appropriately, the researchers had to burden themselves with both nding the appropriate material and the cost of the materials which were needed for the lessons (as for instance folders, markers, materials for puppetry).
The positive effects of DI application in diverse population classes could foster further research in the eld of EFL. The young students coming from neglected and marginalized social groups need stimuli and new ideas; in order to render the teaching process meaningful, the teaching community should provide the students with fair opportunities and exible lessons (Raines at al., 2016). It is necessary that the stakeholders meet the challenge with a fresh mind.
Finally, it is worthwhile mentioning that teaching to classes coming from a culturally diverse background could be rather demanding and tedious; it requires special handling and an open mind. The stakeholders should provide special training to teachers who undertake teaching to minorities like the Romani, so that they could use the most appropriate instructional approach. According to a case study conducted by Comfort (2016), the Roma learners are di cult to approach. They have suffered racial discrimination, and, consequently, the teaching staff should approach them with positivity and understanding (ibid.).

Conclusion
The study constitutes an action research within which differentiated instruction was applied to diverse population students of the third grade of primary school. In order to examine the effectiveness of the approach followed, the results of two groups of students were elaborated upon, in terms of the progress observed amongst the experimental and the control group. The qualitative, as well as the quantitative data collected, led to the conclusion that differentiated instruction bene tted the students as they exhibited greater progress in the post-test and appeared more motivated than the students attending the traditional classes.
Despite the limitations discussed, the research experience proved to be fruitful and the positive results rendered it rewarding. The di culties which emerged provided a challenge which worked constructively and they will be taken into consideration for future improvement.
Further and more long-term research could examine additional ways to motivate less-privileged students and provide those opportunities which would alter the dominating perceptions on cultural minorities.

Declarations
Availability of data and materials This is an original action research and all the data collected are presented as they are.
Competing interests Not applicable.

Funding
No funding was obtained for this study.