Türkçenin Yabancı Dil Olarak Öğretiminde Dil Becerilerine Özgü Kaygının Ölçülmesi
Measuring Skill-Specific Anxiety in Teaching Turkish as a Foreign Language

Summary/Abstract: In the field of second language acquisition, the “affective turn” of the 1990s shifted scholarly focus from purely cognitive processes to the emotional factors that shape the learning experience (Dewaele & Li, 2020). Among these factors, foreign language anxiety has been consistently identified as one of the most powerful negative predictors of success and performance (MacIntyre, 2017). The conceptualization of this anxiety has evolved from general psychological constructs like trait and state anxiety (Spielberger, 1983) to a situation-specific phenomenon, most notably captured by Horwitz et al.’s (1986) Foreign Language Classroom Anxiety Scale (FLCAS). While the FLCAS has been a foundational tool, its treatment of anxiety as a monolithic classroom construct overlooks the nuanced affective responses that specific linguistic tasks can elicit. This limitation led to the development of skill-specific anxiety research, which posits that interacting with the four primary language skills (listening, speaking, reading, and writing) can produce distinct emotional reactions (MacIntyre & Gardner, 1991). Consequently, a range of scales targeting individual skills has emerged. In the context of teaching Turkish as a Foreign Language (TFL), a similar body of literature has developed, with various scales created to measure anxiety in specific skills (e.g., Melanlıoğlu & Demir, 2013, for speaking; Altunkaya & Erdem, 2016, for reading). However, this has resulted in methodological fragmentation, with each scale based on different theoretical foundations and psychometric processes. This makes it impossible to holistically and comparably assess a single student’s anxiety profile across all four skills. To address this gap, Cheng (2017) developed a set of four brief, parallel scales grounded in Lang’s (1971) tripartite model of anxiety, which defines anxiety in terms of cognitive, somatic, and behavioral components. This study therefore aimed to adapt and psychometrically validate this comprehensive scale for the TFL context. The study was conducted with two independent samples of international university students learning Turkish in Türkiye. The first sample, comprising 250 students (145 female, 105 male), was used for Exploratory Factor Analysis (EFA). The second sample of 268 students (155 female, 107 male, based on available data) was used for Confirmatory Factor Analysis (CFA) and criterion-related validity analyses. The participants’ proficiency levels ranged from A2 to C1. The adaptation of Cheng’s (2017) four scales (Listening, Reading, Speaking, and Writing Anxiety) followed the rigorous guidelines of the International Test Commission (ITC, 2017). This involved a multi-stage process including forward translation by two bilingual experts, a reconciliation phase, back-translation by an independent expert, and a final review by an expert panel to ensure linguistic, semantic, and conceptual equivalence. The final step involved contextually adapting the items from “English” to “Turkish” to suit the TFL learning environment. For criterion-related validity, two additional scales were administered: the ShortForm Foreign Language Classroom Anxiety Scale (S-FLCAS) and the Short-Form Foreign Language Enjoyment Scale (S-FLES). Academic achievement was measured using students’ last-course exam scores and a self-reported Turkish proficiency rating (1-10). Data was collected in classrooms under the supervision of instructors using Google Forms. The analysis included EFA to explore the factor structure, followed by CFA to confirm the proposed three-factor model. Convergent and discriminant validity were assessed using Composite Reliability (CR), Average Variance Extracted (AVE), and the HeterotraitMonotrait ratio (HTMT). Internal consistency was evaluated using Cronbach’s Alpha and McDonald’s Omega. The psychometric analyses provided strong evidence for the validity and reliability of the adapted scales. The Kaiser-Meyer-Olkin (KMO) measures and Bartlett’s Test of Sphericity confirmed the suitability of the data for factor analysis for all four scales. The EFA, using Parallel Analysis to determine the number of factors, revealed a clear threefactor structure for each of the four skill-specific anxiety scales, explaining a substantial amount of the total variance (59.3% for Listening, 65.4% for Reading, and 70.2% for both Speaking and Writing). The subsequent CFA confirmed that the three-dimensional model (Cognitive, Somatic, Behavioral) provided a good fit to the data for all four scales. The model fit indices were within acceptable to excellent ranges (e.g., CFI = .959-.983; RMSEA = .067-.089). Convergent validity was established, with CR and AVE values largely exceeding the recommended thresholds of .70 and .50, respectively. Discriminant validity was also confirmed, as HTMT values were below the .90 criterion. For criterion-related validity, all four adapted anxiety scales showed a moderate, positive, and statistically significant correlation with the general classroom anxiety measure (S-FLCAS) (r ranging from .509 to .593). They also exhibited a significant, low-to-moderate negative correlation with language enjoyment (S-FLES) and both measures of academic achievement. Finally, all four scales demonstrated high internal consistency, with Cronbach’s Alpha and McDonald’s Omega coefficients ranging from .868 to .921. The findings of this study confirm that the Turkish-adapted versions of Cheng’s (2017) four-skill anxiety scales are valid, and reliable instruments for use in the TFL context. The study successfully replicated the original three-factor structure, affirming the applicability of Lang’s (1971) multidimensional model of anxiety to TFL learners. The primary contribution of this work is providing the TFL field with a standardized and theoretically coherent toolset that overcomes the issue of methodological fragmentation. This allows researchers and educators to assess a learner’s anxiety profile holistically across all four language skills. The scales can serve as a valuable diagnostic tool for teachers to identify which skills are most anxiety-provoking for students, and the nature of their anxiety (i.e., cognitive, somatic, or behavioral), paving the way for more targeted and effective pedagogical interventions. While the study is limited to university students, we suggest including different age groups and employing longitudinal designs to track the dynamics of skill-specific anxiety over time.

• Details
• Contents