Abstract Writing is one of the most challenging skills to teach in English as a Foreign Language (EFL) contexts. With the rise of artificial intelligence tools, ChatGPT has emerged as a potential assistant in writing instruction. This mixed-methods study explores the impact of integrating ChatGPT into narrative writing classes, focusing on learners’ performance and experiences throughout the writing process. Sixty EFL learners at two proficiency levels (Pre-intermediate and Upper-intermediate) were divided into a control group with traditional instruction and an experimental group using ChatGPT-assisted instruction over five weeks. Quantitative findings showed that the ChatGPT group outperformed the control group in grammatical accuracy, lexical range, and cohesion. Qualitative results highlighted increased engagement, better feedback, and improved vocabulary and grammar use, though some participants noted reduced creativity and dependence on the tool. The study suggests that ChatGPT can enhance narrative writing when used alongside pedagogical strategies that promote critical thinking and learner autonomy.
Many factors are involved in determining how and when mindsets affect learning outcomes and recent second language (L2) research has begun to explore the impact of language mindset (LM) on a variety of learning outcomes, including willingness to communicate in a second language (L2 WTC). However, it still remains unclear how LM may contribute to L2 WTC. The present study addresses this gap and explores how learners' engagement, particularly emotional engagement (EE) and behavioral engagement (BE), mediate the relationship between LM and L2 WTC. Two-hundred and five English as a foreign language learners took part in an online survey containing questionnaires measuring the research variables. The construct validity of the questionnaires was confirmed by a confirmatory factor analysis, ensuring their accuracy and relevance. Additionally, structural equation modeling was used to evaluate the anticipated connections between the variables and analyze the proposed structural relationships. Findings showed that growth language mindset (GLM) is positively associated with L2 WTC, and their association was mediated by EE and BE. This study presents practical suggestions for implementing classroom-ready strategies that operationalize GLM to enhance BE, EE, and L2 WTC.
Writing is one of the most challenging skills to teach in English as a Foreign Language (EFL) contexts. With the rise of artificial intelligence tools, ChatGPT has emerged as a potential assistant in writing instruction. This mixed-methods study explores the impact of integrating ChatGPT into narrative writing classes, focusing on learners' performance and experiences throughout the writing process. Sixty EFL learners at two proficiency levels (Pre-intermediate and Upper-intermediate) were divided into a control group with traditional instruction and an experimental group using ChatGPT-assisted instruction over five weeks. Quantitative findings showed that the ChatGPT group outperformed the control group in grammatical accuracy, lexical range, and cohesion. Qualitative results highlighted increased engagement, better feedback, and improved vocabulary and grammar use, though some participants noted reduced creativity and dependence on the tool. The study suggests that ChatGPT can enhance narrative writing when used alongside pedagogical strategies that promote critical thinking and learner autonomy.
This study examined the impact of sound-and-spelling consistency of L2 English words on written vocabulary knowledge among Japanese learners studying English as a foreign language (EFL). In Study 1, 162 Japanese participants completed a multiple-choice meaning recognition test based on the updated Vocabulary Levels Test [VLT]; in Study 2, 107 Japanese participants completed a spelling test, adapted from the VLT used in Study 1. Two directions of consistency at the rime level (i.e., phonology to orthography and orthography to phonology) were analyzed, while controlling for other word-related variables (e.g., word frequency and cognateness). Although exploratory analyses revealed that the consistency effect emerged under controlled conditions for specific word types (i.e., noncognates) and frequency bands (i.e., high-frequency vocabulary), no reliable consistency effect was observed irrespective of test formats. These findings suggest that consistent words are not necessarily easier to learn, particularly for EFL learners whose L1 writing system is not alphabetic.
This research targets the acquisition of copular constructions by Chinese learners of European Portuguese (EP), evaluating the impact of the potential absence of the copular verb in Chinese in the acquisition of EP and determining whether L2 learners associate the copular verbs ser and estar in EP with their [+/-stage-level] semantic values. Moreover, it intends to assess whether non-native speakers acquire non-default features of ser when it combines with a PP predicate locating events. 102 participants completed an acceptability judgment task including APs and PPs. The results demonstrate that learners readily acquire the features that configure the presence of the verb in EP and associate ser and estar with the adequate semantic values, but show greater difficulty in acquiring non-default features of ser ([+s-level, +dynamic]). We relate this progression in the acquisition path to the hierarchy of meso-, micro- and nanoparameters proposed by the updated Bottleneck Hypothesis (Slabakova, 2019).
This study investigates the predictive power of motivational orientations on second language (L2) writing task performance across different proficiency levels. Drawing on the framework of motivational orientations, including Promotion, Prevention, Assessment, and Locomotion, the research explores how these orientations predict various aspects of L2 writing, such as syntactic complexity, accuracy, lexical complexity, and fluency (CALF). The study also examines the interaction between motivational orientations and learners' writing proficiency levels. To do so, 120 undergraduate English as a foreign language (EFL) learners participated in this research, categorized into three proficiency groups (low, mid, and advanced) based on their writing placement test scores. They completed a motivation questionnaire and performed an argumentative writing task. The essays were then analyzed using CALF measures. Results revealed that Promotion positively predicted phrasal complexity and lexical sophistication. Although Promotion was a positive predictor of overall syntactic complexity for the Upper Intermediate and Advanced groups, it negatively predicted this metric for the Intermediate group. Prevention positively predicted accuracy, particularly for the Intermediate group, while negatively predicting phrasal complexity and lexical diversity in lower proficiency groups. Assessment positively predicted syntactic subordination, especially for Intermediate and Advanced learners, while negatively predicting accuracy in the Advanced group. Finally, locomotion predicted fluency, especially for the Intermediate learners, but negatively predicted accuracy in the Advanced group. These findings suggest that motivational orientations predict L2 writing task performance, with varying effects depending on L2 learners' writing proficiency levels.
This study adopts a corpus-assisted approach to examine differences in interactional metadiscourse (IM) between TED Talks and L2 student digital persuasive speeches. Two corpora were compiled for analysis: a TED corpus and a STU corpus comprising English speeches delivered by L2 students in a public speaking course at a Hong Kong university. Quantitative results revealed significant differences across all IM categories except hedges, with the TED corpus showing higher frequencies of self-mentions and boosters, and the STU corpus featuring more directives and audience pronouns. Qualitative analysis further indicated that L2 students employed a narrower range of IM forms, often overusing or underusing specific types, resulting in less persuasive stance and weaker emotional appeal. The rhetorical divergences between the two genres offer valuable insights for L2 public speaking pedagogy, highlighting the importance of explicit instruction in stance and engagement through the effective use of IM in digital oratory.
This study examines filled pauses and prolongations in Mandarin Chinese, Russian, and Hebrew by comparing monolingual and bilingual speakers to identify both universal and language-specific disfluency patterns. Data were collected from monologues produced by monolinguals and two bilingual groups: Russian-Hebrew speakers who acquired both languages in early childhood, and Mandarin Chinese-Russian speakers who learned Russian later as a second language (L2). Analyses focused on the frequency and types of disfluencies. Monolinguals showed similar disfluency rates across languages, suggesting some universal patterns. Early bilinguals mirrored monolingual patterns in both languages, likely due to balanced early exposure. In contrast, Mandarin-Russian bilinguals exhibited higher disfluency rates in L2-Russian, likely due to increased cognitive load during speech planning. Additionally, they produced unique filled pause types not found in monolinguals, reflecting cross-linguistic transfer. These findings highlight how factors such as language proficiency, language exposure onset, and typological differences shape disfluency patterns in bilingual speech.
The present study tested the effects of interleaving versus blocking practice on contextualized grammar learning. An unfamiliar structure for the learners, the pronoun "y" in French, was used in meaning-focused activities, which are more challenging than existing studies, at three different tenses (past, present, and future) according to an AAA-BBB-CCC schedule in one group and an ABC-ABC-ABC schedule in another. Two groups from two intact classes ( n=22 and n=23 ) of first-year Chinese students studying French participated in the study. A pretest-training phase-posttest design was adopted as in existing studies. The blocked group used the structure with greater fluency (reduction of mid-clause pauses) during the training phase and the posttest while the interleaving group used the structure more accurately, but at the expense of fluency. Blocked practice seems to promote an initial stage of proceduralization in the application of the rule, but with more errors produced than in the interleaved group.
This study examines filled pauses and prolongations in Mandarin Chinese, Russian, and Hebrew by comparing monolingual and bilingual speakers to identify both universal and language-specific disfluency patterns. Data were collected from monologues produced by monolinguals and two bilingual groups: Russian-Hebrew speakers who acquired both languages in early childhood, and Mandarin Chinese-Russian speakers who learned Russian later as a second language (L2). Analyses focused on the frequency and types of disfluencies. Monolinguals showed similar disfluency rates across languages, suggesting some universal patterns. Early bilinguals mirrored monolingual patterns in both languages, likely due to balanced early exposure. In contrast, Mandarin-Russian bilinguals exhibited higher disfluency rates in L2-Russian, likely due to increased cognitive load during speech planning. Additionally, they produced unique filled pause types not found in monolinguals, reflecting cross-linguistic transfer. These findings highlight how factors such as language proficiency, language exposure onset, and typological differences shape disfluency patterns in bilingual speech.
Like other questionable research practices (QRPs) discussed in this issue, self-citation can range from fitting and appropriate to self-serving and unethical ( Ioannidis, 2015 ). The present study sought to estimate self-citation patterns using a large, representative sample of applied linguistics research articles ( K = 969). Our results indicate a median of 1 self-citation per paper (2% of all references) at the individual author level (median = 3 or 5% at the author-team or article level). However, much higher rates of self-citation were also observed among individual authors and author-teams (max = 23 and 31, respectively). We explore these and other results in the context of QRPs and in light of bibliometric research from other disciplines. We also consider our findings in relation to the incentive structures in academia. Recommendations for future research are provided along with suggestions for preventing and addressing excessive self-citation for different stakeholders (e.g., journals, institutions, learned societies).
Questionable research practices (QRPs) comprise a gray area of researcher decisions that may be reasonable in some situations but dubious in others. Recent works in this area have sought to catalog and estimate the presence of QRPs in applied linguistics (e.g., Isbell et al., 2022; Larsson et al., 2023). Building on those studies, we reintroduce the notion of QRPs in this introductory article to the present special issue, linking recent work in this area to related movements toward research ethics, open science, and methodological reform. We then outline the remainder of the special issue that follows. In doing so we highlight the ways in which these papers - both individually and in the aggregate - expand and advance the conceptual and methodological scope of research on QRPs. We conclude by reflecting on and considering next steps for this vibrant research agenda.
The focus on ensuring quality control measures and ethics is gaining more and more prominence in the field of applied linguistics; and in this avenue, researchers are examining Questionable Research Practices (QRPs). In this study, we set out to examine self-citations within QRPs as a form of self-promotion that may unjustifiably boost the scientometrics of academics and may lead to unfair advantages. To this end, we created a database of recent open-access articles from the last five years (2019–2023) published in the five leading journals in applied linguistics ( k = 359). Our findings suggest that there is a high extent of self-citations, and there are significant differences in the total counts of self-citations in the selected journals. Based on the COPE (2019) guidelines, the range of excessive self-citations is relatively high in the selected journals. Policy is needed to be included in author guidelines regarding what excessive self-citation means.
The field of applied linguistics is currently undergoing a methodological shift, specifically in raising accountability towards ethical research practices (Plonsky et al., 2024, Yaw et al., 2023). The current study aimed to discern the perceptions of research ethics of intensive English practitioners affiliated with a university and are positioned to contribute to the scholarly record. A 40-item Q-sort was used to ascertain participants' (n = 51) perceptions of ethical, unethical and questionable research practices, which resulted in 6 distinct perceptions: ethically informed, unethically informed, uninformed, misinformed, ethically inclined, and QRP misinformed. Cross-referencing participants' previous research experiences showed no correlation between educational attainment and perceptions of ethics but indicated a trend of theoretical and practical research as a coursework requirement at all degree levels. Overall, instruction of ethics can be supported in multiple ways to encourage and engage practitioner-researchers to contribute to the academic record.
Applied linguistics has been showing increased interest in research ethics, including discussion of authors’ questionable research practices (QRPs). However, less attention has been given to how organizations may engender QRPs. To address this, here we discuss how neoliberal systems of academic publishing are implicated in QRPs. Through our collaborative autoethnography as two author-editors, we jointly explore such practices’ influences. Three key findings emerge: 1. journal reviewers’ and editors’ bias towards Anglocentric writing norms; 2. the influence organizations such as publishing houses, Ministries of Education, and universities exert over academic publication; and 3. metrification of research output leading authors to disproportionately focus on journal indexing. We argue that these factors hinder faculty ability to balance publishing, teaching, and administrative responsibilities. By widening the discussion concerning QRPs, we highlight how authors’ publication practices are influenced by external factors, pushing back on the narrative of individual responsibility for QRPs.
Poor sampling practices can constitute a questionable research practice when conducting L2 inferential quantitative research. The current study, a methodological synthesis (N = 433 Scopus/Web of Science (WoS) reports: cluster random sampling) of sampling practices, revealed that L2 inferential quantitative researchers rarely employed randomized and/or effect size-driven sampling processes with only eight (1.8%) and ten (2.3%) of the reports being respectively satisfactory. Furthermore, just 33.9% of the reports featured multisite (convenience) samples. In models assessing what predicted multisite sampling, whether the report was ISLA-focused (r(s) = -.33, p < .001) or single-authored (r(s) = -.15, p < .001) incurred moderate and weak negative associations. Citation analysis metric values and the Scopus/WoS contrast had no associations. The findings of this study suggest the field's sampling practices have room to improve and guidance for future improvement is offered.
This study investigates whether adult learners of Italian as a second language (L2) acquire Obligatory Control (OC) in non-finite gerundive adjuncts, which is the local relationship between the empty subject PRO in the adjunct and the subject of the main clause. Adjunct OC in Italian is neither frequent in input nor explicitly taught. In a self-paced reading experiment with a sentence similarity-judgment task, 34 participants with various L1s and proficiency levels were tested seven months apart. By the second session, participants were more accurate and faster at identifying the main clause subject as the PRO controller and showed faster reading times at the non-finite form in the target-probe matching condition. We explore whether proficiency, exposure, L1, input distribution, or a combination of factors may explain our results.
Beginning in 2022, the field of applied linguistics has increasingly approached the evaluation of research quality through an ethical lens, with a particular emphasis on Questionable Research Practices (QRPs). Notably, the majority of existing investigations into QRPs have concentrated on mono-method studies, especially those employing quantitative methodologies, thereby neglecting the realm of mixed methods research (MMR). The present study seeks to illuminate the problematic areas that may contribute to QRPs within MMR studies. To this end, we analyzed 60 MMR studies published between 2011 and 2020 in leading journals within the domain of applied linguistics (AL). Our findings reveal a range of issues pertaining to MMR rhetoric and references, study purpose and design, as well as the integration of methodologies, all of which pose risks to the transparency and foundational principles of MMR. This study concludes with recommendations aimed at enhancing the quality of MMR studies.
In quantitative applied linguistics research, the ethical grey zone between responsible conduct of research and blatant misconduct covers numerous researcher practices that may be more or less ethical depending on situational variables (e.g., context, researcher intent). Known as questionable research practices (QRPs), these actions coincide with the day-to-day decision points that occur throughout the research process. Building on Larsson et al.’s (2023) investigation of the prevalence and severity of 58 field-specific QRPs among researchers in the quantitative humanities, the current study presents a thematic analysis of the 2,261 qualitative comments left by 167 of these survey respondents. Five overarching themes were identified in these comments: Roughly half of the responses were justifications of QRP actions, while others highlighted the contextually-dependent nature of QRPs and pointed to potential ambiguity in the wording of these items. These findings offer implications for how we as a field discuss QRPs, as well as researcher training practices.
