turns-00009.parquet:47931
19a327260861b3a105c9916bdegenerate_repetitionAbsentFinal dense release
USER
translate the following paragraphs into Chinese. 3.2.2. Research Foci and Theoretical Orientations In the first round, open codes were applied to the research focus of each article. The two coders closely examined the titles, abstracts, keywords, and introductions of articles and independently assigned codes as the research foci. Some articles had multiple research foci, and if necessary, the coders were allowed to assign two or more codes describing the topics of each article. In the initial coding, 15 research foci were identified, and the inter-coder reliability was 0.72, indicating low agreement between the two coders. We then compared, revised, and integrated the codes to generate a mutually agreed upon list of research foci, which included writing instruction, student writing, the composing process, writing feedback, individual differences in writing, writing teacher development, writing assessment, and materials for teaching writing. Based on this coding framework, the two coders independently recoded the articles, compared their results, and obtained a satisfactory level of agreement (0.89). Unlike Riazi et al. (2018), who identified a primary research focus for each article, we coded the articles using several codes, as necessary. We assumed that some articles would attempt to simultaneously address several dimensions of an issue (Bliuc et al., 2007). The coding of the theoretical orientations may have been largely interpretive, owing to their generally implicit manifestation in articles. To facilitate our coding and avoid biased interpretations, we initially followed Cumming’s (2016) categorisation of major theories in L2 writing: contrastive rhetoric, cognitive models of the composing process, genre theories, and sociocultural theory (for the definition of each orientation, see Table 4). The selection of a predefined framework was based on the assumption that ‘any attempt to understand a text presupposes some prior understanding on the part of the interpreter’ (Kuckartz, 2014, p. 19). The coders read through the title, abstract, introduction, and literature review of each article to identify its theoretical perspective on EFL writing. Some articles explicitly stated their theoretical orientations; an example is Lee’s (2012) study, which focused on genre-based instruction. However, the theoretical foundations of most of the studies were not overtly presented; it was necessary to interpret them based on an overall understanding. In the initial coding, we encountered several cases in which the theoretical orientations did not seem to fit the framework. For example, Chandrasegaran’s (2013) study approached writing both as a sociocultural activity and as a thinking process, making it difficult to determine whether the study followed sociocultural theory or a cognitive approach. We therefore held rounds of discussion and consulted several prominent publications on theoretical perspectives of EFL writing research (e.g., Bawarshi, 2003; Cumming, 2010; Hyland, 2019; Leki et al., 2008; Silva, 2005). After a thorough discussion and an overall examination of the cases in question, we added a new category, socio-cognitive approach, to the coding framework for the theoretical orientations. Table The socio-cognitive approach to theorisation acknowledges both the cognitive side of writing, such as planning and drafting, and the sociocultural context of writing and its influence on writing performance. With the revised coding scheme, the coder reliability in theoretical orientations increased from 0.71–0.87, attaining a satisfactory level of agreement. 3.2.3. Research Methodology and Data Sources In education and language research, the taxonomy of methodologies for empirical studies as quantitative, qualitative, and mixed methods is widely accepted (e.g., Creswell, 2009; Johnson & Christensen, 2008; Kuckartz, 2014; Riazi & Candlin, 2014; Silva, 2005) and can offer a useful framework for research synthesis. However, as Riazi and Candlin (2014) noted, some studies consider both quantitative and qualitative datasets but do not explicitly use the term ‘mixed methods’ to justify their methodological basis. Considering this complexity, we followed the more nuanced research methodology classification proposed by Riazi et al. (2018). In their coding framework, research methodologies were categorised as quantitative, qualitative, mixed methods, and eclectic. Compared with research utilising a mixed methods approach, studies based on an eclectic methodology do not use the term ‘mixed methods’ to denote their methodological orientations, nor do they draw on methodological support from mixed methods theorisation. Moreover, in Riazi et al. (2018) framework, studies that adopted the eclectic approach could be further grouped into eclectic (QUAL + quan), eclectic (QUAN + qual), or eclectic (QUAL + QUAN). This further classification of eclectic methodologies reflected the possibility that studies placed equal or unequal emphasis on their quantitative and qualitative datasets; the upper and lower cases were used to weight more and less emphasised components, respectively. For example, eclectic (QUAL + quan) signified that a study emphasised the qualitative component, whilst eclectic (QUAL + QUAN) was used to denote studies that placed an equal emphasis on quantitative and qualitative components. In coding the data sources, we directly adopted Riazi et al.s’ (2018) categorisation, which has proved sufficient for analysing data collected in L2 writing research (cf., Zheng & Yu, 2019). Riazi et al. (2018) operationalised Hyland’s (2016) data collection methods as data sources. Four categories were predefined for the coding of data sources: elicitation, introspection, observation, and text samples. Elicitation includes self-reports, such as questionnaires, interviews, and tests. Introspection concerns data generated through participants’ verbal or written retrospective recall, including reports, think-aloud protocols, and diaries. Observation involves recorded data on actions such as audio/video recordings and keystroke logging. Text samples focus on students’ writing itself and include single texts, text chains (i.e., a series of writing products), and corpora. Unlike Riazi et al. (2018), who used ‘multiple’ to code studies which used two or more sources of data, we assigned several codes to a single study, as necessary, to fully reflect the data collection orientations in EFL writing research in school contexts. 4.Findings 4.1. Contexts and Participants Table 1 summarises the top 10 most frequently investigated macro and micro contexts reported in the 136 research articles from 2012 to 2019. As shown in Table 1, Spain, the most significant contributor, accounted for 17.6% (n=24) of the total studies and was followed closely by Hong Kong (23, 16.9%). South Korea (16, 11.8%) ranked third in this respect. The top three countries/regions (i.e., Spain, Hong Kong, and South Korea) were followed distantly by other macro contexts such as Indonesia (6, 4.4%), Jordan (6, 4.4.%), and Taiwan (6, 4.4%). Except for Spain (24, 17.6%), Norway (5, 3.7%), and the Netherlands (4, 2.9%), the remaining top 10 most researched macro contexts were all in Asia, indicating a strong emphasis on English education in Asian countries/regions. Table 1 also indicates that the majority of research articles typically focused on EFL students in secondary schools (93, 68.4%); the micro contexts of primary schools (37, 27.2%) received less attention. Studies concerning both primary and secondary school contexts (3, 2.2%) were extremely limited, indicating the scant attention paid to students in the two micro contexts. These findings align with the previous observation made by Ortega (2009) and Lee (2016) that insufficient research on EFL writing has addressed students in primary schools or mixed school contexts (e.g., transferring from primary to secondary schools). Table 2 shows that the most common participants in the reviewed articles were learners (101, 74.3%), followed distantly by instructors (15, 11.0%). Eighteen articles involved both learners and instructors (13.2%), and four articles (1.5%) obtained their data from written texts (e.g., lesson plans). Table 2 also reveals that whilst EFL learners remained the focus throughout the entire period, a limited number of the reviewed studies investigated instructors or other stakeholders in EFL writing contexts, such as educational policymakers, proofreaders, writing instructors and raters, teaching assistants, and written texts.
ASSISTANT
3.2.2. 研究重点和理论取向 在第一轮中，对每篇文章的研究重点进行了开放编码。两个编码者仔细审查了文章的标题、摘要、关键词和引言，并独立地分配了研究重点的编码。一些文章有多个研究重点，如果有必要，编码者可以分配两个或更多描述每篇文章主题的编码。在初步编码中，确定了15个研究重点，编码者之间的一致性可靠性为0.72，表示两个编码者之间的一致性较低。然后，我们比较、修订和整合编码，形成了一个互相协商的研究重点列表，包括写作教学、学生写作、写作过程、写作反馈、写作个体差异、写作教师发展、写作评估和教学写作材料。基于这个编码框架，两个编码者独立重新编码文章，比较他们的结果，并获得了令人满意的一致性水平（0.89）。与Riazi等人（2018）不同的是，他们为每篇文章确定了一个主要研究重点，我们使用了若干个编码，根据需要。我们假设有些文章会同时涉及问题的几个方面（Bliuc等人，2007）。对于理论取向的编码可能在很大程度上是解释性的，因为它们在文章中通常是隐含的表现形式。为了便于我们的编码并避免主观解释偏差，我们最初遵循了Cumming（2016）对L2写作中主要理论的分类：对比修辞学、写作过程的认知模型、体裁理论和社会文化理论（关于每个理论取向的定义，请参见表4）。事先确定框架的选择是基于“任何对文本的理解都预设解释者事先具有某种理解”（Kuckartz，2014，第19页）的假设。编码者阅读每篇文章的标题、摘要、引言和文献综述，以确定其关于外语写作的理论视角。有些文章明确陈述了它们的理论取向，比如Lee（2012）的研究，重点是基于体裁的教学。然而，大多数研究的理论基础并未明确呈现，需要根据总体理解进行解释。在初步编码过程中，我们遇到了一些情况，其中理论取向似乎不符合框架。例如，Chandrasegaran（2013）的研究将写作作为社会文化活动和思维过程来处理，这使得很难确定该研究是否遵循社会文化理论还是认知方法。因此，我们进行了一系列讨论，并参考了一些关于外语写作研究理论观点的重要出版物（例如Bawarshi，2003；Cumming，2010；Hyland，2019；Leki等人，2008；Silva，2005）。经过深入讨论和对相关案例的全面审查，我们在理论取向的编码框架中添加了一个新的类别，即社会认知方法。 表格 社会认知方法论承认写作的认知方面，如规划和起草，以及写作的社会文化背景对写作表现的影响。通过修订后的编码方案，理论取向的编码者一致性从0.71至0.87提高，达到了一个令人满意的一致性水平。 3.2.3. 研究方法和数据来源 在教育和语言研究中，将实证研究分类为定量、定性和混合方法的分类法被广泛接受（例如，Creswell，2009；Johnson＆Christensen，2008；Kuckartz，2014；Riazi＆Candlin，2014；Silva，2005），并可以为研究综合提供一个有用的框架。然而，正如Riazi和Candlin（2014）指出的那样，一些研究同时考虑定量和定性数据集，但并未明确使用“混合方法”这一术语来证明它们的方法论基础。考虑到这种复杂性，我们遵循Riazi等人（2018）提出的更细致研究方法分类。在他们的编码框架中，研究方法被分类为定量、定性、混合方法和折衷方法。与采用混合方法的研究相比，采用折衷方法的研究不使用“混合方法”这一术语来表示其方法论取向，也不从混合方法的理论化得到方法论支持。此外，在Riazi等人（2018）的框架中，采用折衷方法的研究可以进一步分为折衷（QUAL + quan）、折衷（QUAN + qual）或折衷（QUAL + QUAN）。这种进一步分类的折衷方法反映了研究可能对其定量和定性数据集给予平等或不平等的重视；大写和小写分别用于加重和较少重视的组成部分。例如，折衷（QUAL + quan）表示研究强调定性成分，而折衷（QUAL + QUAN）则表示研究对定量和定性成分给予了平等重视。在对数据来源进行编码时，我们直接采用了Riazi等人（2018）对数据收集方法的分类，这已经足够分析在学校环境中进行的外语写作研究的数据采集（参见Zheng和Yu，2019）。Riazi等人（2018）将Hyland（2016）的数据收集方法操作化为数据来源。对于数据来源的编码，预先定义了四个类别：引导、内省、观察和文本样本。引导包括自我报告，如问卷调查、访谈和测试。内省涉及通过参与者口头或书面回顾生成的数据，包括报告、口述思维协议和日记。观察涉及记录动作的数据，如音频/视频记录和按键记录。文本样本关注学生的写作本身，包括单个文本、文本链（即一系列写作产品）和语料库。与Riazi等人（2018）不同的是，他们使用“多重”来编码使用两个或更多数据源的研究，我们为单个研究分配了几个编码，以充分反映学校环境中外语写作研究的数据收集取向。 4.结果 4.1.背景和参与者 表1总结了2012年至2019年的136篇研究文章中被报道的前10个最常研究的宏观和微观背景。如表1所示，西班牙是最重要的贡献者，占总研究量的17.6%（n=24），紧随其后的是香港（23，16.9%）。韩国（16，11.8%）在这方面排名第三。西班牙（24，17.6%）、挪威（5，3.7%）和荷兰（4，2.9%）之外，其他前10个最常研究的宏观背景均位于亚洲。这表明亚洲国家/地区在英语