3 METHOD
3.4 Data Collection
My entry into this school began in the fall of 2008, when I conducted a small qualitative study focusing on two students and their mothers. One was a HLL and the other a non-HLL, though both were girls in the seventh grade. This study involved interviews with each student and her mother together (Parent-Child), an interview with the two students together (Child-Child) and one classroom observation in each of their Arabic classes. The findings of this study offered insights into the motivations of these learners and their mothers, their perceptions of Arabic and the language program at Mawaarith, and the relationships between language learning, identity, and community for each of these families. Although
data from this preliminary investigation will not appear in the current study, these findings provided the basis for the development of the initial survey questionnaires and interview guides used in the current study.
As mentioned above, the current study draws on data from surveys, interviews, classroom observations, documents, and other sources including images, informal conversations, and researcher reflections. Table 3.2 summarizes the principal data collection methods in general chronological order. The sections below will describe and provide a brief rationale for my approach to each of the data collection methods.
Table 3.2
Overview of data collection procedure
Time Surveys Interviews Classroom Observation
Preliminary study Fall 2008 --- 2 Parent-Child 1 Child-Child 1 observation in each of 2 proficiency levels Current study Beginning of 2009- 2010 school year
First student survey Parent survey Fall 2009 Recruitment Initial Parent-Child interviews Weekly observations, rotating visits to 6 classes Spring 2010 Follow-up interviews with parents and children separately Interview with Arabic teachers and humanities teacher Weekly observations, rotating visits to 6 classes End of 2009-2010 school year Second student survey
Spring & Summer 2012
Follow-up interviews with some focal families
3.4.1 Surveys. In this study, the surveys served two purposes. One was to provide a basis for selecting a purposeful sample of informants for the more intensive process of interviews (Morse, 2007; Strauss & Corbin, 1998). The other, related purpose was to glean data on language backgrounds, language use, attitudes, and motivations from a larger number of participants than I could feasibly interview (Dörnyei, 2007; McKay, 2006). The results of the first student survey and the parent survey support a description of demographic characteristics of the population as a whole. The second student survey, conducted at the end of the year, was intended to reveal the dynamics of motivation (Gardner, 2006), showing change in the students’ self-report of language ability over the year and responses to the instructional experience.
3.4.1.1 Survey questionnaires. The questionnaires for these surveys were designed specifically for this population and context. Surveys in applied linguistics ask three primary types of questions: factual questions, behavioral questions, and attitudinal questions (Dörnyei, 2007). These questionnaires included all three. Factual questions asked about languages spoken, language learning history, self- assessments of proficiency, and history with this school. Behavioral questions covered the ways in which Arabic or another home language is used. Attitudinal questions asked about motivations for learning Arabic, attitudes about learning, and intentions for Arabic and language learning in the future. The student and parent questionnaires included a variety of question types, including specific open questions, closed questions with a number of options, Likert scale items asking respondents to agree with various statements, questions regarding frequency of behaviors, and open clarification questions (Dörnyei, 2007). The complete questionnaires are included in Appendix C.
3.4.1.2 Survey procedures. The initial student questionnaire was administered during Crew meetings (the school’s equivalent of homeroom classes). The questionnaires were distributed to teachers who led Crews, who gave them out to their students and made themselves available to answer questions. The completed surveys were returned to an administrator, who gave them to me. The parent
questionnaire was sent home in folders with other school materials, and parents were asked to return the completed surveys to school in these folders. A statement on these questionnaires explained in non-
technical terms that the results of the surveys would be shared in aggregate and that their names would not be associated with their responses. These surveys were piloted with the Arabic teachers, which was not an ideal choice but provided opportunities for clarification.
The second student survey was conducted near the end of the school year during Arabic class time. I was present to distribute the questionnaires and answer questions. For the second survey, which asked students to express attitudes about the learning experience that they may not have wanted to reveal to teachers, I identified the questionnaires by number and maintained a list that associated these numbers with each student’s name. Hereafter, I will refer to the first student survey as the “initial” survey and the second student survey as the “final” survey.
3.4.1.3 Survey participants. There were approximately 100 students enrolled in the middle school in the 2009-2010 school year, of whom 83 responded to the initial student survey. Ages ranged from 10 to 14. Numbers of male and female respondents were almost equal, with 40 girls and 43 boys. About a quarter (n = 19) of these learners had been attending Mawaarith since the school opened in 2005, but 27 learners (32.5%) were new to the school in 2009.
These 83 learners were spread across the three proficiency levels, with 21 beginners, 25 in the intermediate class, and 25 in the advanced class. The remaining 12 survey respondents reported that they were not currently taking Arabic. The reasons for this were not requested on the survey, but in most cases these learners were receiving some kind of remedial support related to English reading or math skills. Very few, if any, middle school students were categorized as English language learners in that school year.14
The discussion of family backgrounds in Section 4.1 below includes all 83 learners who
responded to the initial survey, including those who were not taking Arabic, because they are nevertheless part of the middle school population and their parents also made active decisions to send them to this school. In the remainder of Chapter 4, however, the sections that consider values regarding Arabic
14 It was difficult to pinpoint an exact number of middle schoolers receiving ESL support because this number over-
lapped with those receiving other types of remediation, but based on discussion with the ESL teacher the number seemed to be two or less.
language learning, reported proficiency, and practices only include the 68 students who were taking Arabic and who completed the entire survey. The second survey was distributed in Arabic classes, so all the participants in the final survey were enrolled in Arabic at the time and no data were collected for learners who were not taking Arabic. A total of 60 students completed the final survey, and data from all of these learners are included in the discussion of items from the final survey in Ch 4.
Thus, the initial survey provides an overview of 68 Arabic language learners and the second survey yielded information on 60 learners. There was substantial turnover in the school population between the two surveys, with the greatest change taking place among students who were new to the school in 2009, most of whom were in the beginner level. No attempt was made in this analysis to compare responses from the same individual learner on the initial and final surveys, though future analyses may discuss these connections.
Participation among parents was lower than participation among the students, likely due to the extra effort involved in completing a survey at home and returning it to school. The 37 parents, including 22 mothers and 15 fathers, who responded to the survey represented 24 families and 26 children. This means that parent data was available to correspond with about one in four learners.
3.4.2 Interviews. As in most ethnographic research, interviews are used in this study as a means of learning about participants’ experiences and accessing their emic perspectives on the context of the investigation and the concepts and categories within it (Hammersley & Atkinson, 2007; Lazaraton, 2003; Purcell-Gates, 2006). They are semi-structured or unstructured and allowed me, the interviewer, and the interviewees to collaborate in directing the flow of talk as we delved into the participant’s life history, current practices, and perceptions (Spradley, 1979).
Although interviews may appear to be neutral encounters that occur for the purposes of gathering data about a natural context, it is important to keep in mind that the interview itself is a context in which reality is socially constructed (Baynham, 2006; Coffey & Street, 2008). This point is particularly relevant to situations like the interviews in this study, in which the researcher belongs to a different cultural group from most of the informants (Kvale & Brinkmann, 2009). That said, the impact of this difference is an
empirical question. Due to the repeated interviews and frequent observations, the interviewees and I were able to build rapport over a period of months that likely contributed to the openness and breadth of these interviews.
3.4.2.1 Selection of focal participants. Early in the 2009-2010 school year, I sent a recruitment letter (see Appendix D) and the parent questionnaires home with students in a folder that they regularly used to transport important documents to their parents. Parents were asked to complete the contact information form at the bottom of the letter and return it to school in the folder if they were willing to participate in interviews and allow their child to participate. Initially, families were selected from this group of willing respondents, but it became clear that more focal participants would be needed in order to understand the context more fully. As a result, two students and their families were recruited later, as will be discussed below. Although I asked teachers and administrators for some guidance as to which parents would likely be willing to devote the necessary time to the study, I did not specifically tell them which participants I contacted and chose. My choice of informants also rested to some extent on my early interaction with the students and my sense that these learners would be able to describe the context and their experiences articulately.
Based on the literature, it seemed clear from the outset of this study that an appropriate sample would include heritage and non-heritage learners of Arabic, based on the definition of HLLs as learners who speak Arabic at home. Also, since boys and girls attend separate classes in the middle school, I wanted to include an informant of each gender. As a result, my initial goal was to recruit a purposeful sample consisting of one male and one female HLL and one male and one female non-HLL who would be willing to participate in interviews along with their mother or father. In an effort to locate informants who were well acculturated in the school environment, I also tried to choose participants who had been at Mawaarith for longer than one year. At this stage, I recruited the Yusuf family, the Zaki family, the Brooks family, and the Kimball family, all of whom I hoped would be good informants15: currently
involved in and knowledgeable about the context, willing and able to participate in a series of interviews, and able to speak articulately about themselves and their experiences (Spradley, 1979).
Relying on a priori categories, however, may mean that we ignore categories that are meaningful to the context, that we fail to recognize variation within categories, or that we expect all members of a given category to be good informants (Morse, 2007). The frequency of bilingual families in the survey results suggested that my sample also needed to include a student who had a HL other than Arabic; to that end, I asked Martin Hamid and his mother, a native speaker of Polish, to participate, but I was only able to interview them once before they moved back to Egypt, where Martin’s father was still living.
Survey results and early observations also suggested that I should include a student who was Muslim but whose family spoke a language other than Arabic, of which there were several in the
intermediate and advanced classes. I recruited the Rowther family, who are Muslims from India, in order to include informants with this experience. Also, the Yusuf family declined to participate after the first interview16, which led me to recruit another female HLL. By then I had also realized that my HLL informants in the informal pilot study and the current study all had mothers who had not been educated in Arabic. I wanted to include a HLL family in which both parents were born in the Middle East. By this time, I had met Farah El-Gendy, a sixth-grade girl whose family had moved from Egypt to the U.S. when she was two, and she had begun to tell me stories about prior schools and her reasons for valuing
Mawaarith. After I contacted Farah’s mother by phone, her family agreed to participate in the study as well.
As a result, these informants represent a theoretical sample that began with assumptions about the relevant categories in the population based on the existing literature but came to include additional informants based on “the emerging categories, and the researcher’s increasing understanding of the developing theory” (Morse, 2007, p. 240). Table 3.3 summarizes the seven families who served as
16 The Yusufs did not actively withdraw from the study, but as Mrs. Yusuf ceased to return phone calls it seemed
informants for this study. This analysis focuses on the five families who completed a series of interviews: the Zaki, El-Gendy, Rowther, Brooks, and Kimball families.
Table 3.3
Overview of focal families
Includes parents’ countries of origins, languages parents speak or have studied, and countries in which children have lived as well as languages children speak or have studied. 17
Mother Father Child Family
Native country Languages Spoken Native country Languages Spoken Countries lived in Languages Spoken Religion
Zaki U.S. English,
Arabic Syria Arabic, English U.S. English, Arabic Muslim El-Gendy Egypt Arabic,
English Palestine, Egypt Arabic, English U.S. Arabic, English Muslim Rowther India Tamil,
English, Hindi, (Arabic) India Tamil, English, Hindi, (Arabic) U.S., India English, Tamil, Arabic Muslim
Brooks Malaysia English, Malay, Mandarin U.S. English, (Mandarin) U.S. English, Arabic, Spanish, Japanese Christian
Kimball U.S. English, Spanish U.S. English, Spanish, German U.S. English, Arabic Christian
It should be stated here that this study deals with families who are invested in Arabic language learning and in the language program at Mawaarith. Unfortunately, this was not true for all families at the school, particularly among beginners. Also, this study does not include any families whose children could be considered HLLs but who have not chosen, or have not been able, to provide their children with formalized instruction outside the home in Arabic. While classroom observations did provide some insight into those learners who are sitting in Arabic classes but have not become invested in Arabic, I have only had informal conversations with HLL parents who are not investing in Arabic learning outside the home. Their stories are valuable, as well, but this study aims to illuminate the values, intentions, and
17 In order to maintain focus on the five focal families who participated more fully in the interview process, data
practices of parents and children who have actually invested in the language, while leaving these other experiences for future research.
3.4.2.2 Interview procedure. With each focal family, I aimed to conduct five interviews, moving from semi-structured to unstructured as the process went on. Each interview lasted about an hour and was recorded with a Sony digital audio recorder. The first interview included both the student and a parent. One reason for this choice was to facilitate the process of explaining the study and obtaining informed consent for the parent’s and child’s participation and to establish rapport with all the participants from the beginning. Furthermore, this interview allowed me to see how the parents interacted with the child, and in some cases to get a sense of the dialogue that might take place between them around family language policy, educational decisions, and desire to learn Arabic. The subsequent four interviews, two with the student and two with a parent, were conducted with parents and children separately, though usually on the same day. The same parent could participate in all three interviews, or the mother and father could alternate, but the total number of interviews remained the same. The Zaki, Brooks, and Rowther families participated in five interviews, the Kimballs in four, and the El-Gendys in three during the 2009-2010 school year. Two years later, I followed up with the El-Gendy and Brooks families, asking about changes in their learning and use of Arabic since the prior interviews.
The locations of the parent and student interviews varied depending on space and the informants’ preferences. Most of the initial interviews and many of the later interviews were conducted at the school, in a small, bright, isolated room near the teachers’ break area that is normally referred to as the
conference room. This room had two low tables and a few plastic chairs at the beginning of the year, but later in the year a committee of parents added a conference table and paint to make it more appealing as a conference room or teachers’ lounge. Although I often mentioned that I was willing to meet with
participants at their homes, for their convenience and to allow me to observe the home environment, only two interviews were conducted in homes.
The majority of the data regarding the three Arabic teachers comes from field notes based on interviews and informal conversations, but all three teachers were interviewed as a group in May of 2010.
Through observations and these conversations, I was able to glean insight into their intentions for the Arabic program, their reflections on the students and their progress in general, and their perspectives on certain practices that I observed and their relationship to their beliefs about Arabic and language teaching. Further research should include a greater focus on teachers’ beliefs and practices, though these beliefs and practices might also be informed by the results of this study.
3.4.2.3 Interview questions. In designing the interview questions, I tried to find a balance between defining the topics that I thought would be of interest and allowing the interviews to take shape