The Impact Of Gamified Learning On Enhancing Metaphorical Vocabulary Acquisition And L2 Writing Quality

In advanced L2 writing, grammatical accuracy alone cannot foster natural, nuanced academic expression. Learners must develop metaphorical competence specifically, the mastery of metaphorical vocabulary (e.g., conceptual metaphors and fixed metaphorical expressions) to transform abstract concepts into vivid, persuasive discourse. Despite its importance, this aspect of language acquisition is often neglected or relegated to rote, mechanical memorization in traditional classrooms. Consequently, EFL learners frequently struggle to deeply internalize these expressions or successfully transfer them into their actual writing performance (Luu, 2025; Luu & Le, in press; Luu et al., 2025).

To bridge this pedagogical gap, gamification offers a promising alternative. By integrating core game mechanics such as points, badges, leaderboards, and quests, gamified instruction transitions language learning from passive consumption to active engagement. Grounded in Self-Determination Theory, Flow Theory, and the ARCS framework, a theory-driven gamified environment fosters intrinsic motivation by satisfying learners' basic psychological needs. Additionally, it provides immediate feedback and lowers affective barriers like writing anxiety, which heavily impedes student achievement (Luu & Nguyen, in press; Luu et al., 2025; Luu, Nguyen, Do, & Nguyen, 2025).

However, major empirical gaps persist regarding the application of gamification to abstract, high-level writing skills. While existing research validates the use of educational apps for basic vocabulary retention or surface-level grammar drills (Gamal et al., 2025; Jabbar & Usman, in press), studies on cognitively demanding lexical domains like metaphorical vocabulary remain scarce. Crucially, whether game-mediated vocabulary gains can effectively transfer to independent essay composition rather than remaining confined to isolated lexical tasks remains unanswered (Eanga et al., 2025; Luu, Nguyen, Anh, et al., 2025; Nilubol & Sitthitikul, 2025). Therefore, investigating the dual mechanisms of metaphor acquisition and productive writing transfer is essential for designing sustainable, context-sensitive EFL methodologies (Calero Sánchez et al., 2024; Muralei et al., 2025).

This study addresses these gaps by evaluating the impact of gamified instruction on EFL learners' acquisition of metaphorical vocabulary and its subsequent transfer to essay writing. Specifically, the study aims to measure learners' retention and contextual understanding of metaphorical vocabulary following a gamified intervention; evaluate improvements in overall writing quality across fluency, accuracy, and lexical diversity; and explore learners' perceptions of and experiences with this instructional model. To guide the empirical inquiry, three research questions are addressed:

• RQ1: How does gamified instruction affect learners’ acquisition and retention of metaphorical vocabulary?
• RQ2: How do gamification-mediated gains in metaphorical vocabulary manifest in the quality of learners’ writing in terms of fluency, accuracy, and lexical diversity?
• RQ3: What are the learners’ perceptions and attitudes toward integrating gamification into metaphorical vocabulary learning for academic writing?

METHODOLOGY

Research Design

To capture a comprehensive understanding of the intervention, this study employs a mixed-methods concurrent embedded design grounded in a quasi-experimental framework (Nilubol & Sitthitikul, 2025). Given the administrative constraints of the university setting, randomized assignment of individual participants is not feasible. Consequently, intact, pre-existing classrooms are utilized to maintain ecological validity.

The design features an Experimental Group (EG), which receives gamified vocabulary instruction, and a Control Group (CG), which undergoes traditional vocabulary instruction. Quantitative data from linguistic assessments and psychological scales serve as the primary source of evidence, while qualitative insights from semi-structured interviews are embedded to couch the statistical trends in a deeper contextual understanding, helping to clarify the specific cognitive and affective mechanisms at play (Eanga et al., 2025).

Participants

The study sample comprises 80 non-English major students enrolled in general English courses at a public university in Vietnam. To eliminate confounding variables, all participants were administered a standardized baseline proficiency test to ensure an equivalent linguistic starting point corresponding to the B1/B2 levels of the Common European Framework of Reference for Languages (CEFR). The two intact groups were designated as the Experimental Group (n = 40) and the Control Group (n = 40).

In terms of specific breakdown, the gender distribution consists of 18 males and 22 females in the Experimental Group, compared to 16 males and 24 females in the Control Group, yielding a total of 34 male and 46 female participants. The mean age of the sample is 19.45 years, with the Experimental Group averaging 19.4 years (SD = 0.6) and the Control Group averaging 19.5 years (SD = 0.5). Across both cohorts, participants represent three main academic majors: Information Technology (n = 29, with 15 in the EG and 14 in the CG), Business Administration (n = 25, with 12 in the EG and 13 in the CG), and Food Technology (n = 26, with 13 in each group). Finally, baseline assessments of prior gamification exposure revealed relatively balanced distributions across both classes: 15 students reported high daily app usage (8 EG, 7 CG), 42 reported moderate occasional exposure (20 EG, 22 CG), and 23 had low or no prior exposure (12 EG, 11 CG).

Table 1. Participants’ demographic (N=80)

Experimental Procedure

The instructional intervention spans a total of eight weeks, structured across three distinct operational phases to monitor both progressive development and immediate learning outcomes.

During Phase 1, which takes place in Week 1, both groups undergo identical baseline testing procedures. This phase requires students to complete a metaphorical vocabulary pre-test to gauge their initial receptive and productive lexical knowledge, followed immediately by a diagnostic pre-writing prompt designed to assess their baseline academic writing quality prior to any instructional treatment.

Phase 2 encompasses the core instructional intervention running from Week 2 through Week 7. During this six-week period, both cohorts cover an identical vocabulary syllabus focused on specific target conceptual metaphors, such as ARGUMENT IS WAR and TIME IS MONEY, alongside their associated formulaic expressions (Luu, 2025; Luu, in press). However, the pedagogical delivery varies sharply between the classes. Students in the Experimental Group interact with the target vocabulary via digital gamified platforms like Classcraft and Quizizz, where lexical acquisition is driven by structural game mechanics, including quest-based progressions, narrative challenges, points accumulation, digital badges, and live leaderboards to support intrinsic motivation (Luu et al., 2025; Luu, Nguyen, Do, & Nguyen, 2025). Conversely, students in the Control Group study the exact same metaphorical sets through conventional, paper-based methods, relying on contextual fill-in-the-blank drills, explicit definitions, traditional sentence writing, and rote memorization without digital or game-based incentives.

Phase 3 marks the post-intervention stage conducted in Week 8. Both groups are required to complete parallel forms of the vocabulary post-test and a post-writing task to measure empirical performance gains. To capture the qualitative and psychological dimensions of the intervention, the Experimental Group also completes an attitude survey, and a representative subset of these students participates in semi-structured interviews.

Data Collection Instruments

To ensure robust data triangulation, multiple instruments are mapped to specific research variables, ensuring that both linguistic output and learner psychology are evaluated through validated metrics. The quantitative data stream relies on four distinct instruments: a 30-item multiple-choice and contextual cloze test to measure receptive and productive metaphorical vocabulary knowledge; time-constrained argumentative essay writing prompts addressing abstract themes like Success and Time Management to prompt metaphor utilization; a standardized analytical writing rubric to evaluate lexical diversity and writing quality; and a 24-item student psychological survey using a 5-point Likert scale adapted from the Technology Acceptance Model (TAM) and Self-Determination Theory (SDT). Complementing these instruments, the qualitative data stream utilizes 15–20 minute semi-structured digital audio-recorded interview protocols designed to explore deeper cognitive engagement, writing anxiety, and perceived learning barriers.

To operationalize writing quality and monitor lexical transfer, the student essays are blind-scored by two independent inter-raters utilizing a specialized rubric matrix (Gamal et al., 2025; Jabbar & Usman, in press). Under the metric of Lexical Diversity and Transfer, raters evaluate Metaphorical Accuracy, which assesses the precision of embedding targeted conceptual metaphors within correct registers, alongside Lexical Density, measured as the ratio of metaphorical and idiomatic items to the overall word count. Under the metric of Writing Quality and Structure, raters evaluate Coherence and Cohesion, focusing on the logical sequencing of arguments and the functional use of cohesive devices, alongside Grammatical Complexity, which evaluates sentence structure variation and overall syntactic control.

Data Analysis

Quantitative Analysis

All numerical data will be cleaned and processed using SPSS. First, tests of normality (Shapiro-Wilk) and homogeneity of variance will be conducted to confirm parametric assumptions. To measure intra-group lexical growth from baseline to post-intervention, a Paired-Samples t-test will be applied to the pre- and post-test scores of each cohort.

To determine the comparative efficacy of the gamified model over traditional instruction, an Analysis of Covariance (ANCOVA) will be performed on the post-test and post-writing scores. In this model, the treatment type (EG vs. CG) serves as the independent variable, while the respective pre-test scores are integrated as covariates to statistically control for any initial baseline variations. Descriptive statistics (M, SD) will be computed to summarize the psychological constructs from the survey.

Qualitative Analysis

The qualitative data from the semi-structured interviews will be transcribed verbatim and analyzed using Braun and Clarke’s (2006) six-phase framework for thematic analysis. The workflow moves systematically from raw audio to verified themes, beginning with transcription and check-backs for accuracy, followed by line-by-line open coding to capture initial student reactions regarding task authenticity, game mechanics, and cognitive load.

During the axial coding phase, these codes are gathered into broader, conceptual themes that illustrate the psychological effects of the intervention, such as contrasting the intrinsic drive generated by leaderboard validation against the cognitive friction experienced during productive composition transfer. Finally, the synthesized themes are reviewed, refined, and verified against the broader dataset and existing second language acquisition literature to ensure theoretical cohesion (Luu, Nguyen, Anh, et al., 2025; Luu & Nguyen, in press).

RESULTS

Results for RQ1: Effects of Gamified Instruction on Metaphorical Vocabulary Acquisition and Retention

To evaluate the impact of gamified instruction on the acquisition and retention of metaphorical vocabulary, a series of Paired-Samples t-tests and an Analysis of Covariance (ANCOVA) were executed. Table 1 presents the descriptive statistics and intra-group developmental progress from the pre-test to the post-test and the delayed post-test administered two weeks after the treatment concluded.

Table 2. Paired-Samples t-tests

The intra-group analysis indicates that both instructional conditions yielded statistically significant improvements in vocabulary scores immediately following the six-week intervention. The Experimental Group exhibited an increase in mean score from 14.25 to 24.60, representing a substantial effect size (d = 2.91). The Control Group also registered a significant gain from 14.40 to 19.15, though with a markedly smaller effect size (d = 1.44).

Regarding long-term lexical retention, the Paired-Samples t-test comparing the post-test and delayed post-test scores for the Experimental Group revealed no statistically significant decay (M = 24.60 vs. M = 23.85, p = .068), indicating robust cognitive permanence of the targeted metaphorical items. Conversely, the Control Group suffered a statistically significant drop in performance over the two-week non-intervention period (M = 19.15 vs. M = 16.90, p < .001, d = 0.78).

To determine whether the post-intervention differences between the two groups were statistically meaningful when accounting for baseline variance, a one-way ANCOVA was conducted. The independent variable was the treatment type (EG vs. CG), the dependent variable was the vocabulary post-test score, and the pre-test score was entered as the covariate. The preliminary testing confirmed that the assumptions of normality and homogeneity of regression slopes were fully satisfied.

Table 3. Analysis of Covariance

The ANCOVA results in Table 2 confirm a highly significant main effect for the instructional intervention, demonstrating that students who learned through the gamified model achieved significantly higher immediate post-test scores than those exposed to traditional print materials, even when initial scores were statistically controlled. The partial eta squared value (ηp2 = .631) signifies that the gamified model accounted for 63.1% of the post-intervention variance in metaphorical vocabulary acquisition, verifying its pedagogical efficacy.

Results for RQ2: Manifestation of Lexical Gains in Essay Writing Quality

To investigate how gamification-mediated vocabulary developments manifested in the productive academic writing performance of the learners, the parallel pre- and post-writing essays were evaluated by two independent, blind raters. Inter-rater reliability was confirmed via a Pearson correlation analysis, yielding an r value of .89 (p < .001), which establishes high scoring consistency. Essays were explicitly scored on Lexical Diversity and Transfer (the precision and density of metaphorical structures) and global Writing Quality (coherence, cohesion, and grammatical complexity).

Table 4. Pearson correlation analysis

As shown by the ANCOVA outcomes summarized in Table 3, the experimental group achieved substantially higher scores across all evaluated productive metrics. The manifestation of gamification-mediated gains is most pronounced within the Lexical Diversity and Transfer dimension. For Metaphorical Accuracy, the EG expanded its mean score from a baseline of 1.85 to 4.10, whereas the CG only progressed from 1.90 to 2.85 (F = 114.28, p < .001, ηp2​ = .597). A similar margin was noted for Lexical Density (F = 98.44, p < .001, ηp2​ = .561), confirming that experimental participants injected a higher concentration of appropriately contextualized metaphorical structures into their essays.

Furthermore, the integration of these lexical resources visibly influenced global writing indices. The EG outperformed the CG in Coherence and Cohesion (M = 18.20 vs. M = 14.65, p < .001, ηp2​ = .385), as the target metaphors served as cognitive frameworks that streamlined paragraph transitions and conceptual framing. The smallest, yet still statistically significant, benefit was noted in Grammatical Complexity (F = 14.22, p < .001, ηp2​ = .156), showing that while syntax improved, the primary impact of the intervention remained tied to lexical and conceptual organization.

Results for RQ3: Learner Perceptions and Attitudes toward Gamified Lexical Learning

To chart the psychological profile of the intervention, the 40 participants in the Experimental Group completed a post-experimental survey measuring constructs from SDT and the Technology Acceptance Model (TAM). Table 4 lists the items, means, and standard deviations.

Table 5. Descriptive statistics