The present study explores how the communicative purpose of reporting research findings can be realized in research article Abstract, Results, and Discussion sections.
In this final chapter, the major findings, along with a comparison with previous studies, are first summarized and discussed. After that, pedagogical implications, limitations as well as suggestions for future research are drawn.
Discussions and Summary of the Study
This study took a genre- and corpus-based approach to examining the moves and language use of reporting research findings in RAs of two disciplines. A corpus consisting of 48 RAs, 24 from two major journals in the field of applied linguistics (Applied Linguistics and TESOL Quarterly) and the other 24 from another two journals in the field of computer science (ACM Transactions on Computer-Human Interaction and ACM Transactions on Information and System Security), was constructed. Specifically, the present study aimed to investigate how similar or different research findings are presented in Abstract, which entails the most concise illustrations of the main findings; Results, which includes detailed illustration and interpretation of the findings; and Discussion, which not only signals the end of an RA but also includes possible implications, applications and comparison with other studies.
Move analysis showed that in Abstract, the moves of AS (summarizing results) and AA (indicating implications/applications) were obligatory or near-obligatory. The finding was consistent with other studies on Abstract (Bhatia, 1993; Hyland, 2000;
Lóres, 2003; Martín, 2003; Samraj, 2001, 2002). Though previous studies on Abstract
110
did not pay special attention to the move of AA, its high frequency could be attributed to the promotional nature of Abstract and its being a miniature of the whole study. In addition, the most frequently-used moves included RR (reporting findings), RI (interpreting results and findings), RL (locating data), and RS (summarizing results).
Some of the moves occurred in cycles. This finding is similar to that in one of the much-cited studies investigating RA Results—Brett (1994). In the present study, it was found that the cyclic patterns in Results included RS→RR→RI, RL→RR→RI, or RR→RI, in which writers provide an interpretation of a small piece of presented findings before they start discussing a new set of data. The move patterns identified in the present study were similar to what Brett indicated as ―pointer → statement of finding → substantiation of finding‖ (p.55) in his study on sociology RAs. Finally, move analysis of Discussion indicated that the main rhetorical moves included DS (summarizing results), DI (interpreting results and findings), DC (comparing results to literature), and DA (indicating implications/applications). This final section in RAs not only highlights what has been found in the study but also elaborates on the results, mainly interpretations, implications, and applications. Cyclic patterns were also identified in this section, including DS→DI or DS→DC, the former providing possible interpretation after presentation of findings and the latter entailing comparisons to previous studies after report of findings. The findings agreed with previous studies (Kanoksilapatham, 2005; Nwogu, 1997; Posteguillo, 1999; Yang &
Allison, 2003).
From the discussions above, it can be concluded that while –R (reporting findings) is the most frequently used in Results section, Abstract and Discussion sections entail frequent use of –S (summarizing results), indicating that in Abstract and Discussion, research findings are presented in a summarized manner whereas findings are presented in precise and factual data in the Results section. In addition, as
111
Abstract is the shortest section in an RA, it contains no move cycles whereas move cycles were identified in both Results and Discussion sections. However, the move cycles in Results and Discussion section vary. Those in Results are closely related to reporting and interpreting findings, and those in Discussion are presented in report of findings, followed by interpretation, comparison, applications, or suggestions. The three RA sections, though all of them must report research findings, focus on different moves or use different move patterns, as shown in the move analysis of this study, in order to realize their respective communicative purposes.
To take move analysis a step further, that is, to know how the three sections may vary in levels of generality in reporting results, content analysis of the three sections of a single RA was conducted. The purpose was to identify the corresponding parts reporting a specific result in the three sections and show their variation in generality.
It was found that in Abstract, the most concise section, research findings are presented in the most general manner without referring to specific data. On the other hand, in Results sections, writers usually provide a detailed illustration of specific results and use precise data to support the finding or interpretation of the findings. In the Discussion section, writers present their research findings with a level of generality between the prior two sections and focus more on applications and comparisons with other studies. Examples from both disciplines were used to showcase different levels of generality in the three sections. The implication of these results is that EAP courses of writing research articles should pay special attention to the variations in levels of generality and section focus of the three sections as novice researchers or graduate students often find it problematic to differentiate between/among the three sections of RAs in terms of reporting research results (Bitchener & Basturkmen, 2006).
In addition to move and content analyses, we still need to know how reporting results are realized linguistically in these sections. This study investigated linguistic
112
realizations in terms of verbs, modal verbs, lexical bundles, and use of voice. First of all, we would like to know verbs that are frequently used to report results in these sections. It was found that show, see, find, and use occur in all three lists of the top 10 high-frequency verbs of the sections. Some verbs occur in two sections such as suggest, provide, support, see and make, and some occur in only one section like perform, present, describe, do, expect, mean, report, and need. Overall, it could be observed that these words are not discipline-specific verbs, but popular verbs with meanings related to the various moves of reporting research results. In addition, lexical bundles in each section were identified to know if there are fixed expressions for reporting results. They were classified into general bundles, RA-related bundles, and discipline-specific bundles. It was found that general bundles were used in a similar way in RAs to that in general English texts. Most RA-related bundles, such as in this paper, as shown in Table/Figure, results show that, and for future research, could be related to specific moves of the section they occur. Although a few discipline-specific bundles were identified, it was found that they result from high occurrences in a single RA. This is probably because even within a single discipline, the research topics of the RAs can still be very different and for different topics there could be a wide variety of possible linguistic expressions for reporting research results.
Results from the analysis of voice in the three sections were quite unexpected since active voice was used by far more frequently than the passive voice. A closer examination of the sentences revealed that, as Martín (2003) indicated, many active sentences were constructed ―by means of a sentence initiated with an inanimate noun‖
in subject position and followed by verbs used in active voice instead of passive voice.
Also, the use of passive voice in clauses was not included in the count of passive sentences. Moreover, examination of frequency of modal verbs showed that
113
Discussion entails more modal verbs than the other two sections, indicating that when writers discuss findings and make claims, they use modal verbs to qualify statements and to show tentativeness.
Examination of disciplinary variations showed that whereas rankings of rhetorical moves in both disciplines showed similar results, move patterns showed slight variations in Abstract and Discussion. In Abstract, it was found that move pattern of AS→AA has a higher frequency in computer science RAs while move pattern of AS→AI was found more often in applied linguistics RAs. In addition, move pattern of DS→DI was utilized in applied linguistics RAs whereas DR→DI had a higher frequency in computer science RAs, a phenomenon that might be related to the nature of papers selected in the present study. As studies in applied linguistics often need to discuss implications of results for pedagogical purposes and materials development, many papers selected contain descriptions of pedagogical implications.
On the other hand, studies in the field of computer science are usually aimed to improve an existing model or theorem, thus the performance of a study is usually the main focus in computer science RAs. With respect to the use of verbs in reporting results, writers of both disciplines rely much on general result-related verbs, such as use, show, and find, rather than discipline-specific verbs. Some verbs that may show disciplinary difference would be learn and study in applied linguistics and sense and control in computer science.
Implications and/or Applications of the Study
Analyses in this study were mainly based on the construction of an RA corpus, which was further divided into subcorpora according to sections or disciplines.
Frequency of rhetorical moves, move patterns, verbs, lexical bundles were all derived from the corpora with the use of NLP tools so that the findings of the present study
114
can be used for pedagogical purposes since data in the present study were based on authentic materials. As reporting research findings is the most crucial communicative purpose of RAs, it is essential that learners should know how it is realized in the three sections of RAs. The moves and move patterns identified in each section in this study could help both teachers and learners clarify the differences and similarities between and among the sections.
Furthermore, the findings can provide instructors with pedagogical implications.
In addition to indicate the common or obligatory rhetorical moves and move patterns in Abstract, Results, and Discussion, micro-level features could be specifically pinpointed as well. For example, when providing overview of the main findings in both Abstract and Discussion, micro-level expressions like show, find, use, in + article + noun, and this + noun + showed to make general illustrations. In addition, as RA writers need to focus on specific findings or make reference to factual data in Results section, expressions of see, suggest, shown in + noun can be utilized. In addition, as Discussion further entails possible applications or suggestions derived from retrieved data, expressions like suggest, for future research could be used. All the expressions, including use of main verbs and RA-related bundles could be explicitly taught.
As writing RAs has become a must for both graduate students and researchers, they are often obliged to take academic writing classes. Though instructors may not specialize in the academic disciplines of these learners, they could provide learners with materials, such as frequency lists of verbs or lexical bundles collected from RAs in learners‘ disciplines. In addition, as findings in the present study revealed that levels of generality differ across various sections, instructors should put emphasis on this feature. They can use authentic examples to showcase how reporting results is realized differently, as reported in the section of content analysis.
115
Limitations and Suggestions for Future Research
Results of the present study showed that, by applying a genre- and corpus-based approach, reporting research findings in various RA sections can be fruitfully explored. However, because of limitations in time and scope, some aspects were not investigated in the present study. Therefore, a number of suggestions could be drawn for future research. First, as the present study used only 48 RAs of two disciplines, the findings of our study may be constrained due to the small size of the RA corpus.
Therefore, a larger corpus should be constructed for future studies to obtain more meaningful results. In addition, only RAs of two academic disciplines—applied linguistics and computer science—were collected for analysis in the present study.
Future studies could collect RAs from more disciplines to generalize the findings for academic writing. Finally, comparison of RA corpus of native speakers and RA corpus of non-native speakers or novice researchers may further reveal the differences between them in reporting research results, thereby providing helpful suggestions for curriculum design and materials development.
116
APPENDIX