29

2

Multi-perspective research

Brian Paltridge

Introduction

This chapter discusses multi-perspective research; that is, research which incorporates multi- ple perspectives into its study design, data collection, and analysis. Research that has examined writing for publication and the peer review of submissions to academic journals are discussed as examples of this. This includes approaches and methods which have been employed to examine authors’ experiences of writing research articles and getting published in academic journals, as well as approaches that have been drawn on to examine reviewers’ reports on submissions to academic journals and the experiences of authors dealing with the peer review process. Particular attention is given in these discussions to the data collection methods and data analysis techniques employed in the studies. Proposals are made for further studies that employ multiple perspectives in their study design and research.

The starting point for multi-perspective research is what is often termed triangulation (Bry- man, 2016; Cohen, Manion, & Morrison, 2011; Denzin, 1970; Webb, Campbell, Schwartz, & Sechrest, 1966) in discussions of research methods. Triangulation is a research strategy that involves analysing data from multiple sources (e.g., surveys and diaries), using multiple groups of participants (e.g., students, teachers, and parents) and/or multiple research techniques (e.g., observations and interviews) to examine the matter under investigation. Triangulation, thus, aims to collect multiple perspectives on an event so that the more complete understanding of the topic under examination can be gained (Paltridge & Phakiti, 2015). While triangulation has strong associations with quantitative research, it is equally a strategy than can be drawn on in qualitative research (Bryman, 2016), as well as in mixed methods research (Hashemi, this vol- ume; Riazi, 2017) where it is often used to establish whether qualitative and qualitative findings corroborate, or cross-validate, each other (Bryman, 2016; Riazi, 2016) – that is, to produce more valid findings than might be achieved by the use of a single method (Kramsch, 2015). When triangulation is drawn on in qualitative research, its goal is, rather, to explore a matter from more than a single perspective so as to provide a broader understanding of the issue being investigated.

Cohen et al. (2011) describe different kinds of triangulation which, translated into multi- perspective research, might be termed methodological perspectives, theoretical perspec- tives, investigator perspectives, time perspectives, and space perspectives. Methodological

Brian Paltridge

30

perspectives employ different data sources and methods in the same study. Theoretical per- spectives draw on a range of different theories rather than just a single theory. Investigator perspectives involve the use of two or more researchers examining the same matter. Time perspectives involve the collection of data on more than one case (cross-sectional design), and the collection of data on more than one occasion (longitudinal design). Space perspectives involve the collection of data in more than one site or location.

Methodological perspectives

Lillis and Curry’s (2010) research into second language writers’ experiences of writing for pub- lication is an example of methodological perspectives in that their data sources include texts written by the participants in their study, language, and literacy interviews; email correspond- ence with the participant writers; correspondence between participants and colleagues, jour- nal editors, and reviewers about the writers’ texts; observational field notes; research diaries; telephone interviews; discussions; and documentary data such as departmental and national policy documents. Lillis (2008) describes this use of multiple data sources as well as the (eight- year) period of involvement in the context in which the texts were produced as an example of both of ethnography as methodology and ethnography as ‘deep theorizing’. Ethnography as methodology involves using multiple data sources as well as a period of sustained involvement in the research site (see Wei, this volume). Ethnography as deep theorizing takes this a step further and considers how the use of language and orientation of the texts that were produced index (Ochs, 1992; Starfield, this volume) and connect to certain social structures, values and relations. Embedded within these two views of ethnography is also what Lillis (2008) terms ethnography as method, which employs strategies such as ‘talk around text’ to gain insiders’ per- spectives on the texts that the writers have produced. From this data, Lillis and Curry developed ‘text histories’ which charted the trajectories of writers’ texts from initial drafts through to final publication, rejection, or withdrawal. The history of the texts included original texts, revised versions of the texts, reviewers’ reports, correspondence with editors and interactions with peo- ple who supported the writers (literacy brokers), as well as data from cycles of talk around the texts carried out over an extended period. In doing this, they were able to account for the types of changes made to the texts and relate these to contextual and situational factors, drawing also on the observational and other data they collected in the sites where they carried out their study.

In my study of reviewers’ reports on submissions to academic journals (Paltridge, 2017), I employed corpus informed discourse analysis (Hyland, 2009), survey, and interview data. Ninety-seven reviewers’ reports were collected for the study and 45 reviewers completed a questionnaire. The questionnaire asked about the reviewers’ experience in doing peer reviews, how they had learnt to do write reviewers’ reports, and the issues they faced in writing them. Follow-up emails were sent to reviewers to seek further elaboration on answers that had been provided in the questionnaires. The reports were analysed from the points of view of genre, con- text, pragmatics, politeness, and the use of evaluative language (see ‘Theoretical perspectives’ later in this chapter). The questionnaire and email data were analysed by a process of identifying themes and cross-referencing themes with the analysis of the data as it proceeded. Finally, the questionnaire and email data were analysed to identify the main issues the reviewers faced as they wrote their reviews, as well as how it is that the reviewers had learnt to write their reports.

In Li’s study (2007) of a Chinese chemistry doctoral student’s attempts to get published, she collected process logs, drafts of the student’s writing, email exchanges she had with him, and interview data to examine how he went about writing for publication and the engagement he had with others as he did this. The process logs were online blogs written by the student

Multi-perspective research

31

which Li and the student’s friends and fellow students had access to and commented on. Li also exchanged personal emails with the student with follow-up questions. She was able to track the student’s draft text via the blog site and connect his blog reflections to his evolving paper. Li discusses how the student benefitted from ‘textual mentoring’ (Casanave, 2014) as he examined published articles by leading scholars in his field and analysed their rhetorical features to learn how to better position the findings in his own research papers.

A further study which combines data sources is Flowerdew’s (2000) study of a PhD student struggling with the process of getting published. The data sources for his study included analy- ses of the student’s draft and final texts; interviews and email communication with the student; communications between the student and the journal editor, reviewers, and in-house editor; field notes; and discussion with an editor who helped the student with his writing. Li’s (2005) study of a Chinese physics student trying to publish his work included conversations and emails with the student, drafts of his manuscripts, correspondence with journal editors, and reviewers’ reports as well as interviews with the student’s’ supervisor. Mur Dueñas’s (2012) examination of the writing for publication practices of four Spanish scholars uses text trajecto- ries (drafts and subsequent versions of their papers), editors’ letters, referees’ reports, authors’ responses to the reports, and interviews with the authors as its data sources. Cheng’s (2005) study of graduate students learning to write research articles (published as Cheng, 2006, 2007, 2008a, 2008b) uses students’ genre analysis tasks, student–instructor conversations around these tasks, a needs and expectations survey, literacy autobiographies, students’ reflections on their learning, curriculum materials, and recordings of classroom interactions in its investiga- tion. Curry’s (2014) examination of how writers in engineering employ graphics in their schol- arly writing (which employs data such as focus group discussions, semi-structured interviews, audio and video recorded participant observations, field notes and drafts of students’ texts) is a further example of a study which brings together different research methods and perspectives.

Theoretical perspectives

A feature of my study of reviewers’ reports (Paltridge, 2017) is the theoretical perspectives I draw on to analyse the texts. First, I examine the texts as instances of the genre (Swales, 1990, 2004) of reviewers’ reports. I discuss the reports in relation to the chain of genres (Devitt, 2004; Swales, 2004) of which reviewers’ reports are part, then consider the discourse structures and content of the reports. The reports are also considered from the point of view of performance (Devitt, 2015) which entails reviewers sharing and understanding the norms, values, and expec- tations of the particular disciplinary community as they ‘behave in certain ways and make par- ticular discourse choices’ (Hyland, 2012, p. 20) in their reports. Drawing on speech act theory (Austin, 1962; Sadock, 2004; Searle, 1969) I examine the ways in which reviewers ask for changes to be made to submissions and the extent to which they do this explicitly and implicitly. The reports are then considered from the point of view of politeness (Brown & Levinson, 1987; Kádár & Haugh, 2013) and, in particular, politeness as social practice (Haugh, 2013) in which participants (in this case reviewers and authors) understand what is polite and what is impolite (Watts, 2003) in the particular context of a reviewer’s report. The use of evaluative language in the reports is also considered by examining the stance (Hyland, 2005a, 2005b) the reviewers take towards the submissions they are evaluating. In another study (Paltridge, 2019a) I exam- ine the stance taken by reviewers on non-native speaker writers’ submissions to the journal in comparison with that given to native speaker authors. I also examine (Paltridge, 2017) the roles the reviewers assume as they write their reviews through an analysis of transitivity patterns (Halliday, 1994; Hart, 2014) in the texts (see Paltridge, 2019b for a summary of this research).

Brian Paltridge

32

Putting different theories together is not, however, a matter of ‘anything goes’. There are some perspectives that might not work together because of their very different views of ‘the nature of knowledge and how we know it’ (Heller, Pietikäinen, & Pujolar, 2018, p. 7). Cam- eron (2005) discusses problems associated with what she calls ‘theoretical and methodological eclecticism’ (p. 125). She points out that sometimes this carries a high risk of superficiality as the researcher may be trying to do too many things at once and not end up doing any of them properly. It is not impossible, of course, to mix research perspectives (and methods). What this requires, however, is ‘a clear rationale for putting approaches together, a sophisticated understanding of each approach, and an account of how the tensions between approaches will be handled in [the] study’ (Cameron, 2005, p. 127). This is something that all multi-perspective research needs to be mindful of. Even though people are now talking about ‘the end of the para- digm wars’ (Choi & Richards, 2016, p. 3), when people argued against the bringing together of quantitative and qualitative perspectives that is a core feature of mixed methods research, the underlying basis and assumptions of different theoretical perspectives (and methods) are still matters that need to be considered when planning, carrying out, and making claims in multi- perspective research (see Kramsch, 2015; Mirhosseini, 2017 for further discussion of this),

Investigator perspectives

Investigator perspectives involves the use of two or more researchers examining the same matter. Examples of this can be seen in the work of Hewings (2004), Belcher (2007), Paltridge (2017) and Samraj (2016), each of whom have examined reviewers’ reports written on sub- missions to the journal English for Specific Purposes. Hewings (2004) examined 228 reviews for the journal English for Specific Purposes. Belcher (2007) examined 29 reviewers’ reports, also submissions to the journal English for Specific Purposes. Paltridge (2017) examined 97 reports written for submissions to the same journal, and Samraj (2016) analyses 50 reviews, drawing on a subset of the same data set that I employed in my study (Paltridge, 2017). These studies build on each other, using similar research procedures to previous research on the topic and, as a result, provide a cumulative understanding of the aspects of the texts being examined.

The focus of Hewings’ (2004) study was what matters were most frequently evaluated in the reviews that he examined. To analyse his data, Hewings first created an electronic corpus from his set of reviews. He then used Wordsmith Tools (Scott, 1996) to produce a list of the most frequent words in the corpus. After this, he manually identified evaluative adjectives in the corpus (such as, succinct and interesting), then generated concordance lines for each of the adjectives and the entities that were evaluated. Hewings then counted each type of evaluated entity (such as research and writing style), after which he grouped the evaluative adjectives according to the broad type of quality they represented (such as, interest and suitability). He then grouped the adjectives into those that made positive judgements and those that made negative judgements on the submissions that were being reviewed. The frequency of occur- rence of all the evaluative adjectives and what was evaluated was then recorded.

Belcher (2007) examined positive and negative commentaries in reviews of rejected and accepted papers for the journal English for Specific Purposes. To analyse her data, Belcher first read the reviews to get a global impression of the frequency of commentary types. She did this to get a sense of which were focussed on more, and less, in the reviews. She then identi- fied what seemed to her, as an experienced editor, to be the most common and salient of the comments in the reviews. She looked at the reviewer guidelines employed by the journal, but found that the categories she was observing were more extensive and specific than the areas that were covered in the guidelines. This resulted in the set of categories she used for her analysis.

Multi-perspective research

33

Because, however, of the difficulty of deciding where each of the comments began and ended in the reviews, Belcher decided to count the number of reviews that commented on particular features of the papers rather than the number of individual comments. The findings Belcher pre- sents, however, she cautions, need to be seen as indicative and only one possible reading of her data. Notwithstanding, they do suggest she says, ‘why manuscript authors may respond quite emotionally to reviews’ (p. 7), especially when the comments made by reviewers are negative. Continuing on from the work of Belcher (2007), Paltridge (2017) examined reviews from the same journal which she drew on for her research (English for Specific Purposes) by examining the content of reviewer comments in relation to the review outcomes of accept, major revisions, minor revisions, and reject. Following Belcher, the analysis focussed on the number of reviews that commented, positively and negatively, on submissions and the features Belcher had identi- fied in her study (although not assuming there may not have also been others).

Samraj’s (2016) analysis of the discourse structure of reviewers’ reports for the journal English for Specific Purposes found similar organizational patterns to those proposed by Fortanet (2008) in her analysis of reviewers’ reports in the areas of business organization and linguistics. Samraj also examined commentary sections (as opposed to the introductory and concluding sections) in major revisions and reject reviews. Paltridge’s (2017) analysis of the discourse structure of reviews was somewhat broader than that presented by Samraj (2016) as her study did not include reports written for all categories of recommendation made by the reviewers, only major revisions and reject reviews. Notwithstanding, the analysis in Paltridge’s (2017) study supports Samraj’s (2016) view that major revisions reviews appear to have a largely directive function while reject reviews are more mixed, having more negative commentary and less recommendation units than major revisions reviews (see Paltridge, 2017, Chapter 4, where this is discussed further).

Each of these studies, carried out by different investigators, builds on the other and, while not examples of replication (Marsden, this volume; Polio & Friedman, 2017; Porte & Richards, 2012), do follow similar analytical procedures and orientations. The aim of the studies, however, is not to test or validate the findings of previous studies by comparing them to each other, as is the case with replication, but to build on each study’s results ‘so as to expand knowledge pro- duction beyond what can be accomplished in a single study’ (Polio & Friedman, 2017, p. 268).

Time perspectives

Lillis and Curry’s (2010) study of second language writers’ writing for publication is an exam- ple of cross-sectional design in that they collected data from 50 professional scholars, in 12 aca- demic institutions, across four different countries. Their study was also longitudinal in design in that it was conducted over a period of eight years. The data they collected included approxi- mately 1,192 texts written by the study’s participants; 208 talk around text sessions; 50 lan- guage and literacy history interviews’ approximately 2,000 emails between the researchers and the participants; approximately 500 pieces of correspondence between participants and their colleagues, reviewers, and editors; 60 sets of observational field notes/research diaries; approxi- mately 15 telephone conversations; network diagrams drawn by participants; and documentary data such as policy documents from each of the national sites. The participants’ text histories that Lillis and Curry developed explored ‘what happened as texts moved from one context to another’ (p. 3). The collection of text histories, however, they point out, is extremely time- consuming and can take years to produce. Through them, Lillis and Curry (and their readers) were able to see ‘important moments within the texts’ trajectories toward publication’ (p. 5).

Carnell, MacDonald, McCallum, and Scott (2008) Passion and politics: Academics reflect on writing for publication, a further example of cross-sectional design, reports on the

Brian Paltridge

34

experiences of published authors writing for publication in the same academic discipline (edu- cation) and at the same academic institution (University College London). The study is based on semi-structured interviews with 18 academic writers collected over a period of six months. The interviews lasted about an hour and were audio recorded and transcribed, then sent back to the each of the writers to ensure their accuracy. The transcripts were read several times by the researchers, during which recurring themes were identified in the data. From this, a tentative coding for the data was developed. The authors then exchanged ideas through email and tele- phone conversations to refine their analyses at the same time as they wrote drafts of their texts. They also met several times to discuss the transcripts, their analyses, and their texts. Before the book was published, Carnell and her colleagues sent the edited transcripts back to the people they interviewed as well as a draft of the introduction to the book to be sure that the quotations they had used reflected the authors’ views and that the analyses and arguments were based on accurate understandings of what the authors had intended. Carnell et al. (2008) include the full transcripts of the interviews in their book so that readers can see what the authors had to say, in their own words. They also provide suggestions for further research on this topic.

Space perspectives

Lillis and Curry’s (2010) study is an example of space perspectives in that it involves the collection of data in different countries, namely Slovakia, Hungary, Spain, and Portugal. There have been few other studies in the area of writing for publication which have done this. There are, however, edited collections which aim to do this, such as Bennett’s (2014a) The semiperiphery of academic writing, and Curry and Lillis’s (2018) Global academic publishing. In her book, Bennett (2014a) brings together studies carried out in ten countries in Southern and Eastern Europe which focus on matters such as the dominance of English in writing for publication, the difficulties multilingual researchers face in dealing with the dual identities of being members of global and local academic discourse communities, and discrepancies between policies and practices in areas such as academic appointments and intellectual property. The countries represented in Bennett’s book are Portugal, Czechoslovakia, Greece, Poland, Spain, Romania, Italy, Turkey, Serbia, and Croatia. The chapters of the book describe tensions between academic cultures (such as between the sciences and the humanities, the East and the West, the local and the global, and the traditional and modern, as well as between official discourses and actual practices). This semiperiphery, Bennett (2014b) argues, can be seen ‘as a place of tension, a contact zone where different attitudes, discourses and practices meet and merge’ (p. 7), where scholars are trying to succeed in their academic writing.

In their edited collection, Global academic publishing, Curry and Lillis (2018) present research carried out in countries such as China, Columbia, France, Hungary, Iceland, Kazakh- stan, Mexico, Norway, South Africa, and Taiwan. The book, in particular, looks at the causes and consequences of the domination of English in academic publishing. Methods employed in the studies reported on in the book include interviews, surveys, questionnaires, bibliometric analysis, document analysis, and website analysis. Comparisons that can be made from stud- ies of this kind are especially important for understanding commonalities and differences in participants’ experiences of writing for publications, the different conditions in which they do this, as well as the challenges these differences create.

Future directions for multi-perspective research

A further strategy which involves the combination of research perspectives is what Swales (1998) termed textography. Textography combines elements of discourse analysis with

Multi-perspective research

35

ethnographic techniques such as interviews, observations, and document analysis. It is, thus, something more than a traditional piece of discourse analysis, while at the same time less than a full-blown ethnography. In relation to writing research, textography aims to get an inside view of the worlds in which texts are written, why texts are written as they are, what guides the writing, and the values that underlie the texts that have been written. A particular goal of a textography is to examine the contextualization and the situatedness of written texts. It aims to do this through an exploration of the texts’ contextually embedded discursive practices. A tex- tography, thus, aims to provide a ‘thick description’ (Geertz, 1973) of the context in which texts are produced to gain an understanding of why the texts are written as they are.

In his book Other floors, other voices, Swales (2018) carried out a textography of the kinds of writing that people who worked on three different floors of his building were engaged in at the time he was director of the English Language Institute at the University of Michigan. Over a period of three years, he looked at the texts that people produced, gathered observation data, carried out document and correspondence analysis, and conducted text-based interviews with employees on each floor of the building. He also examined the spaces in which people worked, providing photographic images of these in his book. From this material, textual life histories were put together of seven of the people who worked in the building, including Swales himself, as well as a broader discussion of the notion of ‘discourse community’ in relation to the academy.

A further study which employed textography was a project that examined the exegeses that art and design students write in their master’s degrees (Paltridge, 2004). An exegesis is a written text that accompanies a visual project submitted as the research component of the student’s degree. These texts are similar in some ways to what is called the thesis genre, but in many ways are also quite dif- ferent. An analysis of the exegeses was combined with an examination of texts that surrounded the exegeses, such as the postgraduate student handbook, the guide to examiners, examiners’ reports, and the annual report on the master’s degree. Interviews were also carried out with students, advi- sors, and examiners of the exegeses. This was done in order to explore the particular nature and character of the texts, the values that underlay the texts, and the role the texts played in the particular academic setting; that is, to examine the texts, role, and context of the students’ texts ( Johns, 1997).

Paltridge et al. (Paltridge, Starfield, Ravelli, & Tuckwell, 2012a, 2012b; Ravelli, Pal- tridge, Starfield, & Tuckwell, 2013; Starfield, Paltridge, & Ravelli, 2012) also employed tex- tography in a study of texts that visual and performing arts doctoral students write as part of the submission requirements for their doctoral degrees. Data collected for this study included a nationwide survey, 36 dissertations, 36 supervisor questionnaires, 15 student interviews, 15 supervisor interviews, university prospectuses, information given to students in relation to their candidature, published research on doctoral research and examination in these areas of study, in-house art school publications, discussion papers, and attendance at roundtable dis- cussions and exhibition openings. The study found that there was a range of ways in which stu- dents could write their texts as well as reasons for this range, some of which were institutional, and some of which were due to the influence of key figures in the field, rather than conventions of the particular disciplines. Textographies, thus, provide insights into the production and interpretation of texts that would not be gained if texts or contexts on their own were exam- ined. They are able to show diversities of practices across fields of study and across domains, providing understandings for why this might be the case (Paltridge, Starfield, & Tardy, 2016).

Conclusion

Multi-perspective research has largely been discussed in this chapter with reference to work in the area of English for research publication purposes (Flowerdew, 2014, 2015).

Brian Paltridge

36

Multi-perspective research strategies can, of course, be drawn on in many other areas of research as well. Different types of multi-perspective research strategies, as this chapter has shown, can be brought together in a single study. Studies such as those referred to in this chapter, then, recognise that ‘complex issues have to be viewed from multiple perspectives’ (Grabe, 2006, p. xi). It is not a matter, Grabe (2006) argues, of choosing one perspective over another, but one of nested perspectives in which multiple layers of evidence inform the find- ings of the studies. There will, though, be occasions where data sources are contradictory and in disagreement with each other (Baker, 2018; Heller et al., 2018). For example, different participants may give conflicting views on an event, and what the participants say may not be what the researcher observed. In cases like this, Polio and Friedman (2017) argue, ‘rather than discarding the data as unreliable, one may attempt to reconcile these disparities’ (p. 63). Ultimately, in their view: ‘researchers who aim to incorporate multiple perspectives into their work may sometimes have to accept that these viewpoints will diverge and to see this evidence of the diversity of human experience rather than as threats to reliability’ (Polio & Friedman, 2017, p. 63).

References

Austin, J. L. (1962). How to do things with words. Oxford: Clarendon Press. Baker, P. (2018). Reflecting on reflective research. In C. Taylor & A. Marchi (Eds.), Corpus approaches

to discourse (pp. 281–292). London: Routledge. Belcher, D. D. (2007). Seeking acceptance in an English-only research world. Journal of Second Lan-

guage Writing, 16, 1–22. Bennett, K. (2014a). The semi periphery of academic writing: Discourses, communities, and practices.

Basingstoke, UK: Palgrave Macmillan. Bennett, K. (2014b). Introduction: The political and economic infrastructure of academic practice: The

semi periphery as a category for social and linguistic analysis. In K. Bennett (Ed.), The semi periphery of academic writing: Discourses, communities, and practices (pp. 1–12). Basingstoke, UK: Palgrave Macmillan.

Brown, G., & Levinson, S. (1987). Politeness: Some universals in language usage. Cambridge: Cam- bridge University Press.

Bryman, A. (2016). Social research methods (5th ed.). Oxford: Oxford University Press. Cameron, D. (2005). Review of Allyson Jule: Gender, participation and silence in the language class-

room: Sh-shusing the girls. Applied Linguistics, 26, 125–138. Carnell, E., MacDonald, J., McCallum, B., & Scott, M. (2008). Passion and politics: Academics reflect

on writing for publication. London: Institute of Education, University of London. Casanave, C. P. (2014)