A research project involving more than one language and one culture presents challenges to the research investigator. These challenges are found with cross- language quantitative studies that necessitate translating a survey instrument from a source language to the participant language, as constructs or concepts can be expressed differently from one language to another. Im et al. (2016, p. 142) summarized this challenge: “a specific linguistic group has its own social and cultural factors that influence [their] reality”. Al-Amer, Ramjan, Glew, Darwish and Salamonson (2015) suggest cross-language quantitative research involves translation from one “professional” language to another. Maneesriwongul and Dixon (2004) note a range of techniques for translating quantitative research instruments (e.g. forward-only translation, forward-only translation with testing, back-translation, back-translation with monolingual test, back-translation with bilingual test, and back- translation with both monolingual and bilingual test), and conclude there is no single perfect translation technique. Nevertheless, Cha, Kim and Erlen (2007) alert us to the problem of “vocabulary equivalence”, which occurs when there is no exact word
73 match of the original language in the target language. They suggest using a
comparable word or a group of words that convey a similar meaning may solve the problem. In short, in cross-language quantitative studies the objective is to translate the actual meaning of an instrument’s scale from one context to another context.
Cross-language qualitative research presents even more problems. These concerns are relevant for both the conduct of interviews and the translation of the interviews. Interview questions need to be conveyed in a way understood by the non-English speaking participants without losing their meaning. Conversely, interview data needs to be translated in a way the participants’ communicated meanings are understood by English speakers. Conversational-style interviews are problematic in producing transcripts when non-English slang and idioms are used. Hence direct translations into English can result in erroneous interview data. Esposito (2001) discussed these issues with focus group research, and highlighted the importance of portraying the “meaning” of the non-English speaking participant, otherwise their words (data) can be misunderstood. She argues the cultural context of the participant needs to be reconstructed into the semantics (appropriate vocabulary and syntax) of the English language for the translation to be comprehensible. This challenge is relevant for both spoken translation and producing transcripts: “[the] goal is to develop transcripts that are accurate, clear, and sound as natural as possible [while retaining] connotations and contextual meaning” (Esposito 2001, p. 571). Regmi, Naidoo and Pilkington (2010, p. 19) contend understanding or interpretation of meaning is fundamental in qualitative analysis. Yet, Al-Amer et al. (2015, p. 1152) argue the challenges of translation in cross-language qualitative research are rarely noted.
Squires (2009) reviewed 40 cross-language qualitative studies, and concluded addressing “language barriers” is not straightforward. In the studies reviewed, she found language barriers were seldom mentioned as a limitation of the study. These barriers have two elements: one, the conduct of interviews; and two, the translation of the interviews. She notes two different approaches. First, achieving “conceptual equivalence” involves interpreting concepts mentioned by the interviewee rather than just a word-for-word translation. Cultural equivalence has been described as “an insolvable problem since almost any utterance in any language carries with it a set of assumptions, feelings, and values that the speaker may or may not be aware of but
74 that the field worker, as an outsider, usually is not” (Temple & Young 2004, p. 165). Second, achieving “narrative analysis” to capture experiences of interviewees. Both approaches “require the investigator to have high-level sociocultural competence and significant background knowledge about the country or place of study” (Squires 2009, p. 280). Accordingly, methodological strengths are found when the person who conducted the interview is also the person who conducts the interview transcription (Squires, 2009, Table 2 and Table 3).
Temple and Young (2004) argue who conducts the translation and produces a transcript can shape research outcomes. They likewise suggest advantages of having the interviewer also being the person who compiles the transcript, so long as they remain objective. The sequence of interview, transcription and analysis all done by the same person has virtue in contrast to each activity being done by different persons: “It may seem that researchers who can translate themselves are
automatically best situated to do cross language data analysis” (Temple & Young 2004, p. 167). Hence, this “researcher as translator” technique aids analysis: “The researcher can use the experience of translating to discuss points in the text where she has had to stop and think about meaning. Some researchers who also act as translators regard the discussion of the translation processes as a check to the validity of interpretations” (Temple & Young 2004, p. 168).
Choi, Kushner, Mill and Lai (2012, p. 654) define translation – in the research context – as “the transfer of meaning” from one language to another. Agreeing with Temple and Young (2004) and Squires (2009), they find advantages of the researcher as translator method if the investigator “fully understands” the participant’s culture and not just language. Mere proficiency of a translator in a non-English language without cultural understanding has the risk of overlooking the meaning of metaphors, slang and popular idioms (Choi et al., 2012, p. 659). Put another way, the translator adds their own meaning to a transcript when several interpretations of the
participant’s words are possible (Al-Amer et al. 2015, p. 1157). Appreciating not all concepts or constructs are shared among languages and cultures, they do however caution that the transfer of meaning needs to conform to English vocabulary,
grammar and syntax so that the translation (transcript) “is comprehensible to anyone” (Choi et al. 2012, p. 654).
75 Regmi, Naidoo and Pilkington (2010) similarly note the importance “meaning” for cross-language and cross-culture qualitative research. They emphasize ideas of “semantic equivalence” and “content equivalence”. The first notion concerns similar meanings held by two languages or cultures, and the second notion concerns shared concepts among difference languages or cultures. They argue simple translation may not be adequate for either notion, and suggest instead the process of “transliteration” might be more appropriate. They define transliteration as “a process of replacing or complementing the words or meanings of one language with meanings of another as sometimes the exact equivalence or exact meaning might not exist” (Regmi, Naidoo & Pilkington 2010, p. 18). In simple terms, the process involves: verbatim (word-for- word) transcription of interviews in the original language; followed by translation of the non-English transcription into English; followed by scrutiny and rechecking of the transcripts against the translated interpretation; and ending with a transcription that only includes the key themes or issues that emerge in this process of translation (Regmi, Naidoo & Pilkington, 2010 p. 20-21). Al-Amer et al. (2015, p. 1159) agree that transliteration is desirable in cross-cultural research rather than just verbatim translation, and also see merit of using translators acquainted with the culture and values of participants for this can assist with contextualization of interviewee narrations.
Al-Amer, Ramjan, Gleww, Darwish & Salamonson (2016) discuss the translation concerns when research participants speak Arabic. They note different dialects are spoken in the Gulf States (including Kuwait) and Levantine counties (e.g. Lebanon). Arabic’s linguistic structure, its grammar, syntax and intonation, is different to English. For instance, unlike English there is no definitive grammatical way to convey the present tense from the past tense. As a result, the actual meaning of an Arabic speaking interviewee can be misinterpreted, and cannot be easily translated into English if the context of the spoken words is overlooked. Further, Arabic metaphors have cultural foundations, which can be nonsensical when directly translated into English (Al-Amer et al. 2016, pp. 153-154). For example, several of the interviewees in the present study explained their approach to leadership as “holding the stick in the middle”. When producing English transcriptions of these
76 interviews the wording was rephrased so that the meaning of the culturally founded remark was conveyed.
The approach adopted in this thesis was informed by the academic literature discussed in this section. All the research participants had a Middle Eastern origin, with Arabic being their first language. The majority of both the Experimental and Control group participants were Kuwaitis. Each interview commenced in English. However, as English is not the first language of either the interviewer or any of the interviewees, all the interviews were partly – the extent varied – also conducted in Arabic. Each interview was audio recorded. The “researcher as translator” technique was followed.
Given the English language challenges of the interviewees (for reasons mentioned above), and the occasional transition to Arabic, the spoken expression was not as precise as it might have been if English was the first language. This was not helped by the participants’ use of colloquial – conversational – English and Arabic from time to time. Consequently, the morphology, terminology, grammar, punctuation, rhythm, stress and intonation of the spoken interviews, and the translation from Arabic to English, created challenges when transcribing the interview data into English. This also generated challenges when analyzing the interview data. The objective was to convey the meaning of the experiences and perceptions of each interviewee rather than directly reporting the actual words spoken.
These linguistic issues were managed in two ways. First, the initial transcripts were read several times by the author (in consultation my thesis supervisors) and any syntax irregularities were identified and corrected. Second, the revised transcripts were then considered in the context of the topic for each of the interview questions, and where necessary the audio recording of the interview was further consulted. This process resulted in additional minor changes made to the English transcripts so that the issue-specific information contained in the interviewee’s response was more apparent. The outcome of these procedures is reflected in the verbatim quotes reported in the thesis. The minor grammatical changes made to the transcripts were intended – as far as possible – to retain the “voice” of the interviewees. In sum, the emphasis was on producing interview transcriptions that convey the meaning of the
77 participants’ expressions rather than a simple translation of their words. This method is consistent Regmi, Naidoo and Pilkington’s (2010) concept of transliteration.