Jinyan Huang

This study examined human-AI judgment in EFL speaking assessment as a measurement problem relevant to research on individual differences: how consistently do expert raters and repeated AI sessions preserve between-person differences in a complex performance? Thirty oral-performance samples from first-year Chinese university English majors were evaluated by four experienced EFL teachers and by ChatGPT-4o across four repeated sessions using the same holistic and accuracy-fluency-complexity framework. Score distributions were summarized descriptively; human-rater intraclass correlations and generalizability-theory G- and D-studies evaluated dependability. Feedback quantity and thematic coverage were summarized descriptively, and post-results interviews with the same four teacher-raters were analyzed as exploratory participant reflections. Teachers produced more dependable single ratings. Holistic phi coefficients were 0.54 for one ChatGPT-4o session and 0.65 for one teacher, increasing to 0.83 and 0.88 with four assessments. Analytic single-assessment coefficients ranged from 0.54 to 0.56 for ChatGPT-4o and 0.61 to 0.65 for teachers. Across holistic and analytic scoring, a phi coefficient of at least 0.80 was reached with four repeated ChatGPT-4o sessions but three independent teachers. ChatGPT-4o’s analytic dimensions showed limited differentiation from its holistic scoring: accuracy matched the holistic score in all 120 AI ratings, while fluency and complexity differed in only three and four ratings, respectively. ChatGPT-4o also produced more rubric-relevant comments descriptively, while the four teacher-raters emphasized human oversight for contextual and interpersonal interpretation. The study contributes to intelligence research at the level of psychometric measurement and human-AI judgment: assessor systems can differentially preserve observed individual differences in complex performance. EFL speaking was not treated as an intelligence test, and unmeasured cognitive mechanisms were not inferred from the score patterns.

Universities increasingly promote sustainability, yet environmental intentions do not consistently become private or public action. This qualitative study examined how university students and faculty in eastern China understood the personal, social, pedagogical, and institutional conditions shaping campus pro-environmental behavior. Four focus groups with 20 undergraduate and graduate students from arts and science disciplines and 10 semi-structured faculty interviews were analyzed using a hybrid deductive-inductive framework approach informed by an institutionally situated extended Theory of Planned Behavior. Six interrelated themes showed that intention was necessary but conditional on situated agency, habit, workload, convenience, health, safety, and institutional support. Perceived behavioral control was analytically distinguished from institutional reliability: the former concerned participants’ appraisals of capability and opportunity, whereas the latter concerned whether facilities, procedures, information, and response systems actually worked. Public participation additionally required collective efficacy grounded in student voice, visible outcomes, organizational continuity, and institutional responsiveness. Environmental knowledge became actionable when local, practical, and linked to meaningful opportunities. Co-designed living laboratories emerged as a participant-generated proposal, not an intervention conducted in the study. The findings explain how curriculum, operations, and governance jointly condition private and public environmental action.

Generative artificial intelligence (AI) tools, including ChatGPT and other AI-based writing assistants, have rapidly entered higher education (Mahapatra, 2024; Van Dis et al., 2023). In academic wr...

Feedback literacy concerns learners' capacities and dispositions to recognize the value of feedback, make evaluative judgments, manage affective responses, and take action. This quantitatively dominant mixed-methods study examined theory-informed associations among these four features using questionnaire data from 2,116 first- and second-year Chinese English majors, two student open-ended items, and interviews with 18 instructors. Data analysis used reflective Mode A partial least squares path modeling, 5,000 bootstrap resamples, and 10-fold PLSpredict repeated 10 times. All four hypothesized paths were positive (beta = 0.457–0.830, p < 0.001), with R ² values of 0.532–0.730. However, HTMT values ranged from 0.921 to 1.022 and all exceeded 0.90, indicating inadequate discriminant validity. SRMR was 0.071 for the saturated model and 0.079 for the estimated model. Q2_predict values were positive (0.257–0.531), but the linear benchmark had lower RMSE for 13 of 14 indicators, indicating limited out-of-sample predictive advantage. Qualitative themes emphasized targeted practice, self-assessment, feedback seeking, actionable comments, affective support, and structured peer dialogue. The findings support an intertwined rather than fixed developmental view of speaking-feedback literacy and require cautious interpretation because of construct overlap and cross-sectional self-report data.

This cross-sectional study examined how teaching presence (TP), social presence (SP), and cognitive presence (CP) were associated with undergraduates’ perceived task performance (PTP) and online learning satisfaction (OLS), with self-regulated learning (SRL) as a mediator and knowledge-seeking intention (KSI) as a first-stage moderator. Survey responses from 726 students enrolled at multiple Chinese universities were analyzed using partial least squares structural equation modeling and conditional process analysis. SRL was positively associated with both PTP and OLS. TP and CP were positively associated with SRL, whereas SP showed a significant negative partial coefficient after the three presences were modeled simultaneously. The indirect associations through SRL were positive for TP and CP but negative for SP. KSI significantly conditioned the TP- and SP-based indirect associations but not the CP-based indirect associations. The negative SP coefficient should not be interpreted as evidence that social presence is harmful; high inter-construct correlations, a marginal Fornell–Larcker violation, and elevated full-collinearity variance inflation factors indicate possible construct overlap, suppression, and common method variance. The findings therefore support a cautious interpretation of SRL as a potential pathway and KSI as a boundary condition, while highlighting the need for longitudinal, multi-source, and measurement-refined replication.

Feedback literacy describes the understandings, capacities, and dispositions through which learners recognize the value of feedback, make evaluative judgments, manage emotional responses, and take action. This mixed-methods study examined these four interrelated features in English-as-a-foreign-language (EFL) listening among 1,901 Chinese non-English-major undergraduates and explored action strategies reported by students and 18 college English instructors. The student questionnaire comprised 20 researcher-developed items contextualized to EFL listening; 19 indicators were retained after one taking-action item showed a loading below 0.50. Partial least squares structural equation modeling was used to examine theory-informed associations, and student open-ended responses and instructor interviews were analyzed through qualitative content analysis. Internal-consistency and convergent-validity indices were acceptable, but the Fornell–Larcker criterion was not met, indicating substantial overlap among the four domains. Accordingly, structural results are interpreted as exploratory rather than as evidence of distinct constructs or causal progression. The reported paths were positive, with managing affect most closely associated with taking action. Student reports clustered around engaging with feedback, using technology-mediated practice, seeking authentic listening exposure, and self-assessing progress. Instructor reports emphasized specific and actionable feedback, structured reflection and goal setting, peer-supported learning, and technology-enabled follow-up. The findings suggest that feedback use in EFL listening involves intertwined cognitive, affective, and behavioral tendencies, but stronger scale validation, longitudinal designs, and direct observation of feedback uptake are required before causal or developmental claims can be made.

This mixed-methods study examined three distinct dimensions of peer assessment in Chinese L2 writing classrooms: scoring reliability, feedback quality, and teacher acceptance. Twelve advanced Chinese L2 learners and six experienced teachers scored 20 writing samples holistically using the HSK-6 rubric and provided feedback on content, organization, and language. Generalizability-theory analyses showed that a single peer assessor was less dependable than a single teacher assessor. Two peer assessors matched or exceeded one teacher for relative score interpretations, whereas three peers were needed to reach a comparable level for absolute interpretations. Coded feedback from both groups met criteria for specificity, relevance, and actionability, but their emphases differed. Welch tests indicated that peers produced more suggestions about emphasizing key points, whereas teachers produced more suggestions about streamlining details and grammatical accuracy. Interviews with three teachers revealed conditional acceptance: teachers valued peer assessment as a learning process but emphasized the need for explicit criteria, training, multiple peer judgments, and teacher oversight. Because the study did not examine revisions or subsequent writing gains, feedback quality should not be interpreted as demonstrated pedagogical effectiveness. These findings are exploratory and context-specific.

This mixed-methods study examined feedback literacy in English-as-a-foreign-language (EFL) reading among Chinese non-English-major undergraduates. Drawing on Carless and Boud’s four-dimensional framework, the study investigated the structural relationships among appreciating feedback (AF), making judgments (MJ), managing affect (MA), and taking action (TA), and explored how students and teachers described actionable feedback practices in reading contexts. Survey data were collected from 1,596 students using a 20-item questionnaire adapted to EFL reading, together with two open-ended questions. Semi-structured interviews were also conducted with 20 college English instructors. Quantitative data were analyzed using partial least squares structural equation modeling (PLS-SEM), multigroup analysis, and necessary condition analysis (NCA), while qualitative data were thematically coded. The results showed that AF, MJ, and MA were positively associated with TA, with MA emerging as the most proximal predictor of action-taking. AF also influenced TA indirectly through MJ and MA. Multigroup analysis indicated small but significant differences across gender, year of study, and CET-4 status. NCA further suggested that AF, MJ, and MA functioned as necessary conditions at different points in the feedback process. Qualitative findings showed that students acted on feedback by reviewing comments, setting goals, seeking clarification, using supplementary resources, and reflecting on progress, while instructors promoted action-taking through clear feedback, follow-up tasks, resource provision, and structured reflection. The study extends feedback-literacy research to the underexplored domain of EFL reading and offers context-sensitive implications for supporting feedback uptake in Chinese university classrooms.