Gamified flipped learning and the affective domain of language learning: a qualitative case study of online English classes

1 Introduction

The COVID-19 pandemic rendered online teaching ubiquitous, as the internet became the only viable channel through which teachers could reach their students. While this transition familiarised teachers with online instruction, it simultaneously exposed a range of pedagogical challenges spanning lesson preparation, learner engagement, and assessment. Ensuring meaningful learning in the online mode proved particularly difficult: even in physical classrooms, no single teaching method consistently satisfies both teachers and students, and the constraints of the online environment compound this difficulty. Communicative language teaching, which depends on a sustained flow of interaction, is especially vulnerable in online settings, where it rarely achieves the interactional density of face-to-face classes and where lecture-based instruction offers little assurance of capturing learners’ attention.

Compounding these challenges are the characteristics of contemporary learners. Prensky (2001) argues that “digital natives” prefer multitasking, graphics, rapid and random access to information, and constant connectivity, and that instructional design must respond to these dispositions. Digital games and gamified activities are frequently proposed as one such response. Gamification refers to the incorporation of game design elements, such as points, badges, challenges, feedback, and leaderboards, into contexts that are not themselves games, with the aim of engaging and motivating users (Deterding et al., 2011). In educational settings, gamification promotes learner-centred instruction and creates conditions for meaningful learning through peer scaffolding and sustained engagement (Ghazal and Singh, 2016; Qian and Clark, 2016).

At the same time, devoting entire class sessions to gameplay is pedagogically inadvisable, as learners also require structured conceptual input. The flipped classroom model offers a potential resolution to this tension: by delivering conceptual explanations before class through videos, audio recordings, and presentation slides, class time can be reserved for active, game-based application of those concepts (Herreid and Schiller, 2013). The present study therefore combined gamification with flipped learning in a semester-long online English course and examined, through qualitative analysis of learner interviews, how this gamified flipped design influenced the affective domain of students’ language learning. The affective domain concerns ‘a feeling tone, an emotion or a degree of acceptance or rejection’ (Horne, 1980, p. 191); because affective states shape learners’ attitudes and behaviours, understanding how instructional innovations influence them is of direct pedagogical significance.

1.1 Research question

The study was guided by the following research question: How does gamified learning material, delivered within a flipped online class, contribute to the development of the affective domain of undergraduate students’ English-language learning?

2 Literature review

2.1 Gamification in language education

Although game-based learning predates digital technology, the recognition that gamified learning, both digital and non-digital, could gain broad acceptance in education emerged roughly a decade ago (Ghazal and Singh, 2016). Gamification in teaching denotes the application of game design elements or game frameworks to existing learning activities, with opportunities for feedback and interaction that enhance the value of the learning system (Deterding et al., 2011). Deterding et al. (2011) caution, however, against reducing gamification to its technological dimension, a reduction that also constrains research possibilities. Incorporating game elements into contexts ordinarily free of them augments the user experience, and design-based complex games enable students to develop expressive competence, discover strategies for managing and winning, sustain motivation through collaboration and competition, and receive immediate feedback and rewards; teachers’ preference for design-based games over purely educational or entertainment games has accordingly been associated with better learning outcomes (Qian and Clark, 2016). Gee (2003) similarly observes that the design features embedded in computer games are both available and relevant for language teaching. Research output on gamification in education has proliferated markedly over the past decade (Fernandez-Antolin et al., 2021), and contemporary digital natives have been found to draw readily on gamified approaches in second-language learning (Harvey Arce and Cuadros Valdivia, 2020).

Several psychological principles underpin effective game design (Anderson, 2011). Sequencing, the organisation of tasks from simple to complex, motivates sustained play. Appropriate challenge maintains interest by balancing boredom against excessive difficulty. Status, or one's standing relative to others and to one's own best performance, regulates the duration and intensity of interaction. Finally, the recognition of meaningful achievement propels continued engagement.

2.2 Gamification and affective outcomes in language learning

Empirical studies across disciplines document the affective benefits of gamification, including sustained motivation, interest, interaction, and engagement. Reinders and Wattana (2014) found that carefully designed gamified tasks increased learners’ willingness to communicate, while Ferrão et al. (2019) identified the promotion of interest and motivation as key success factors for gamified language teaching across platforms. Sun and Hsieh (2018) confirmed that a gamified interactive response system sustained learners’ attention and intrinsic motivation through interaction and challenge, consistent with the broader finding that contextual appropriateness triggers or curbs intrinsic motivation (Ryan and Deci, 2000). Findings are not uniformly positive: in Buckley et al. (2017), postgraduate students criticised gamification on grounds of time constraints, whereas undergraduates valued its novelty, competitive spirit, efficiency, and rewards, with the display of relative rank amplifying extrinsic motivation. Research on novice players of massively multiplayer online role-playing games likewise shows that grappling with game rules and strategies generated interactions that expanded opportunities for language development (Jinjing, 2016).

Gamification has been applied productively to specific language skills. A quasi-experimental study with 30 secondary school learners demonstrated benefits for grammar instruction (Hashim et al., 2019), and the use of Socrative and clickers improved reading instruction by enabling formative and summative assessment alongside peer interaction (Faya Cerqueiro and Martín-Macho Harrison, 2019). Bal (2019) found that the Web 2.0 tool Storium supported group work in writing instruction and accommodated individual differences, while Retherford (2020) reported gains in vocabulary growth and motivation among Spanish-speaking English language learners using Osmo Words. Well-designed games also regulate emotion: identification with the game suppresses anxiety and boredom (Csikszentmihalyi, 1990), gamification provides a “safety net” and a space for experiencing social presence while learning at one's own pace (Han, 2015, p. 265), and game-mediated interaction cultivates a “community culture” (Squire and Barab, 2004, p. 512). Nevertheless, Whitton (2013) cautions that game techniques can create inequity even as they empower learners, and Singhal et al. (2019) recommend limiting the number of games to prevent learners’ diversion from learning goals. Dashboards and learning analytics on gaming platforms further support engagement by enabling students to track their own activity (Klock et al., 2018). Affective states in digital game-based learning have themselves become an established research strand (Chen et al., 2021; Langin, 2022), and dynamic, age-appropriate game environments (Liontas, 2022) and syllabus-aligned gamified tools (Luo, 2023) have been identified as conditions for genuinely beneficial integration.

2.3 Flipped learning and gamified flipped models

Gamification alone, however, has a structural limitation: activities built solely around game mechanics can consume class time that learners still need for structured conceptual input, and they leave the sequencing of that input to chance rather than to instructional design. This limitation motivates pairing gamification with a complementary model that secures conceptual delivery before class. Flipped instruction supplies learners with videos, audio materials, and slides before class so that class time can be devoted to reinforcement and knowledge development; it permits flexible, self-paced pre-class learning, enables diagnosis of learning styles and difficulties in class, guarantees continuous access to materials, and accommodates new teaching approaches and technologies (Herreid and Schiller, 2013). Combining flipping with gamification has yielded encouraging, though mixed, results. Pozo Sánchez et al. (2020) observed improved motivation and teacher–student and peer interaction in a gamified flipped design, although the small sample precluded generalisation, a limitation common to quantitative gamified flipped studies that implicitly treat generalisability as the marker of a successful design. The present study responds to this gap differently rather than by simply enlarging the sample: it adopts a qualitative case-study design explicitly oriented towards depth of understanding within a single bounded case, with its sample justified by information power (Malterud et al., 2016) and observed thematic saturation (Section 4.3) rather than by statistical generalisability, thereby reframing, rather than repeating, the generalisability problem that constrained Pozo Sánchez et al. (2020). Aşıksoy (2018) found that integrating gamified strategies into the flipped process prolonged student motivation, and Toriz (2019) documented gains in critical thinking, metacognition, and collaborative decision-making when gamified simulations were combined with just-in-time teaching in an inverted classroom. In language education specifically, a flipped contextual game-based approach improved English as a foreign language (EFL) students’ business writing organisation and accuracy (Lin and Hwang, 2018), and a flipped online learning community built on the LINE application enhanced EFL learners’ oral proficiency through collaboration and interaction (Wu et al., 2017). Lo and Hew (2017) note, however, that unfamiliarity with the flipped format itself constitutes a challenge for many learners.

Recent studies published between 2024 and 2026 reinforce and refine these findings. In a large three-group comparison involving 327 lower-intermediate learners, Daliranfirouz et al. (2025) found that a gamified flipped English class significantly outperformed both flipped-only and control conditions in grammar, vocabulary, and reading achievement, with learners describing the class as fun, collaborative, and engaging while also reporting an increased workload, an affective trade-off directly relevant to the present study. Arsyad et al. (2024) similarly reported that integrated flipped and gamified learning in online settings enhanced test of English as a foreign language performance and student motivation. At the platform level, Özdemir's (2025) meta-analysis of 43 experimental studies concluded that Kahoot exerts a significant moderate positive effect on academic achievement, alongside benefits for retention, motivation, anxiety, and attitude. Despite this growing body of evidence, Ekici's (2021) systematic review of gamification in flipped learning identified a persistent insufficiency of qualitative studies capable of illuminating learners’ lived experience, and semester-long qualitative investigations in Indian higher education remain particularly scarce.

2.4 Critical and mixed perspectives

Notwithstanding the largely positive findings reviewed above, a substantial body of research counsels caution. In a longitudinal comparison of gamified and non-gamified sections of the same course, Hanus and Fox (2015) found that students in the gamified condition showed declining intrinsic motivation, satisfaction, and empowerment over time, together with lower final examination scores, suggesting that leaderboard- and badge-driven designs can undermine the very motivation they seek to foster. Toda et al. (2018) systematise such “dark side” effects, including indifference, loss of performance, undesired competitive behaviour, and declining engagement, while Koivisto and Hamari's (2019) review concludes that the empirical evidence on gamification remains mixed and methodologically limited, with many studies relying on short interventions and self-report data. The novelty effect constitutes a further threat to inference: Tsay et al. (2020) demonstrate that engagement gains in gamified systems can attenuate once novelty wears off. Concerns specific to flipped and online delivery include increased learner workload and potential cognitive overload from combined pre-class and in-class demands (Daliranfirouz et al., 2025), inadequate pre-class preparation and unfamiliarity with the model (Lo and Hew, 2017), inequity among learners with differing gaming dispositions and experience (Buckley et al., 2017; Whitton, 2013), learner fatigue and distraction where games are overused (Singhal et al., 2019), and in pandemic-era online contexts, unequal access to suitable devices and stable connectivity, which can disrupt gameplay itself. These critical strands informed both the design of the present study, which limited gameplay to concept-linked activities rather than whole-class play, and the deliberately cautious, exploratory framing of its claims.

2.5 Research gap

Taken together, the literature establishes the motivational potential of gamification and of gamified flipped designs, but three gaps remain. First, most existing studies are quantitative or short-term, whereas qualitative accounts of learners’ affective experience over a full semester are rare (Ekici, 2021). Second, the affective domain, encompassing confidence, social presence, motivation, and preferences, has seldom been examined as the central construct rather than as a by-product of achievement studies. Third, evidence from Indian higher education contexts, where online gamified flipped instruction expanded rapidly during the pandemic, is limited. The present study addresses these gaps through a semester-long qualitative case study of gamified flipped online English instruction at an Indian university. In keeping with the mixed evidence reviewed above, the study is framed throughout as exploratory and context-bound, concerned with participants’ perceived affective states rather than with measured learning outcomes.

3 Theoretical framework

The study is grounded in six complementary theories that together explain why gamified flipped instruction may influence the affective domain of language learning. Rather than serving as background decoration, each theory informed a specific aspect of the study's design, analysis, or interpretation, as detailed below and mapped against the empirical themes in Section 5.9.

Uses and Gratifications Theory (UGT) posits that learners actively select media and learning platforms to satisfy cognitive, social, and emotional needs. In this study, UGT framed the expectation that students would engage with Kahoot and LearningApps to relieve the monotony of routine classwork, to obtain peer support, and to satisfy curiosity about word usage and collocation; it therefore guided interview questions probing why and when students chose to play.

Self-determination theory (SDT; Ryan and Deci, 2000) explains motivation through three basic psychological needs: autonomy, competence, and relatedness. The gamified flipped design operationalised these needs by offering choice over when and how to engage with games (autonomy), graduated challenges with feedback and rewards (competence), and group play modes (relatedness). SDT accordingly informed the analytic attention paid to motivation and self-confidence in the coding scheme.

Constructivist Learning Theory, particularly Bruner's discovery learning, holds that learners construct knowledge actively by connecting existing knowledge with new knowledge through exploration (Chand, 2024). The variety of game types employed, from cloze tests to millionaire-style quizzes, was designed to afford precisely such exploratory connection-making, and this lens shaped the interpretation of the Language Suitability theme.

Vygotsky's social constructivism emphasises internalisation through social interaction within the Zone of Proximal Development (ZPD), whereby learning from more capable peers scaffolds progress (Chand, 2024). Gamified group play creates conditions for such scaffolding: repeated exposure enables lower-proficiency students to move towards automaticity (Rachels, 2016), and interaction affords the “noticing of the gap” that Schmidt and Frota (1986) identify as critical for language improvement. This lens directly informed the group play design decision and the interpretation of the Social Presence and Playing Together themes.

Bandura's (1963) social learning and self-efficacy constructs explain how vicarious experience of peers’ success invites learners’ own attempts, and how mastery experiences and social persuasion strengthen the belief that a task can be accomplished. These constructs guided the interpretation of confidence gains reported by participants who alternated between individual and group play.

Finally, Csikszentmihalyi's (1990) Flow Theory describes the state of complete absorption that arises when challenge and skill are optimally balanced. In this state, action and awareness merge, self-consciousness is lost, feedback is immediate, and the ordinary sense of time is distorted, culminating in an autotelic experience pursued for its own sake. Flow Theory informed both the graduated design of the games and the analysis of the Motivation, Seeking Challenges, and Time Engagement themes, in which students described persisting with difficult tasks and playing beyond class hours.

4 Methodology

4.1 Research design

The study adopted a qualitative case-study design. A qualitative approach was chosen because the research question concerns learners’ subjective affective experience, feelings, motivation, confidence, and preferences, which are best accessed through participants’ own accounts rather than through pre-defined measurement scales, and because the review of related studies revealed a dearth of qualitative research on semester-long use of digital games in Indian higher education classrooms (Ekici, 2021). The case comprised one intact Foundation English cohort at a private university in Chennai, India, taught by the first author over one full semester of online instruction during the COVID-19 pandemic. The researchers deliberately remained open to unexpected responses; an inductive orientation therefore governed both data collection and analysis. Because the study reports original, previously unpublished interview data, the manuscript is presented as an original research article (qualitative case study) rather than as a review.

4.2 Setting and participants

Twenty-eight students were enrolled in the Foundation English course (ENG2000) from July to December 2020. Placement in the course was determined by performance on an institutional English Proficiency Test administered at entry, ensuring relative homogeneity of language proficiency across the cohort. Students attended a 90-min laboratory session on every weekday except Wednesday during the Fall 2020 semester. All 28 students experienced the gamified flipped instruction; from this cohort, 10 students (3 female, 7 male; mean age 18 years) volunteered to be interviewed at the end of the semester.

4.3 Sampling and sample size justification

Volunteer sampling was employed for the interviews, with an element of convenience arising from the proximity of the semester examinations, which limited the window for recruitment. All 28 members of the cohort were invited to participate through an end-of-semester announcement in the class communication channel; participation carried no academic credit, incentive, or penalty, and 10 students came forward. Although 10 participants constitute a small sample, the size is defensible on established qualitative grounds. First, the study pursued depth of understanding within a single bounded case rather than statistical generalisation. Second, the sample possessed high “information power” (Malterud et al., 2016): the participants were highly specific to the study aim, having all experienced the same semester-long intervention, and the analysis was supported by an explicit multi-theoretical framework. Third, empirical work on saturation indicates that core themes in relatively homogeneous samples typically emerge within the first 6–12 interviews (Guest et al., 2006); in the present data, no new thematic nodes emerged after the eighth transcript, suggesting that thematic saturation was reached for this cohort. The sample nevertheless represents roughly one-third of the intact class, and the implications of its size for transferability are addressed in Section 8.

4.4 The gamified flipped intervention

The intervention combined two components. For the flipped component, the researchers uploaded teacher-prepared PowerPoint slides, videos, and audio recordings covering every syllabus topic to the class WhatsApp group before each session, so that students could study the underlying concepts prior to class. For the gamified component, students played digital games on those concepts during the subsequent online class session, throughout the entire semester.

Two platforms were used. Kahoot, a game-based learning platform available on iOS and Android, was used to introduce rules of grammar and discourse, to provide practice across language skills, and to conduct quizzes assessing progress; students could play collaboratively in class or asynchronously at home. LearningApps (learningapps.org), a free online platform developed at the Bern University of Teacher Education (Switzerland), Johannes Gutenberg University Mainz, and the Zittau/Görlitz University of Applied Sciences, allows teachers to use default templates or create their own. Game types designed by the researchers on LearningApps included Guess the Word, Where Is What?, image matching pairs, pairing games, audio/video with notices, cloze tests, the Millionaire game, number line, multiple-choice quizzes, simple ordering, and group assignment. These games targeted picture description, résumés and cover letters, parts of speech, tense, modal verbs, conditionals, dangling modifiers, parallelism, instructions, technical proposals, and, principally, vocabulary.

4.5 Data collection

Data were collected through semi-structured, one-on-one interviews conducted online at the end of the semester. An interview guide of open-ended questions invited participants’ views on (a) the gamified learning material designed for the daily classes, (b) their preferred modes (individual versus group play) and preferred timing of gameplay within the class, and (c) the flipped delivery of conceptual input prior to the online sessions; follow-up probes were guided by participants’ responses. The semi-structured format permitted the pursuit of unexpected responses while maintaining comparability across interviews. Each interview lasted approximately 20 min. Interviews were audio-recorded with participants’ consent and transcribed verbatim by the first author, and each transcript was checked against its recording for accuracy before analysis. The interview guide is provided in Appendix A.

4.6 Data analysis

Transcripts were analysed in NVivo 14 using a codebook approach to thematic analysis, grounded in the phases described by Braun and Clarke (2006) but employing a structured codebook and consensus-based review rather than the fully organic procedures of reflexive thematic analysis. This designation is deliberate: the analysis combined an inductive orientation to the data with a documented codebook (Table 1), operational node definitions, co-author consensus review, and a multi-theoretical framework that sensitised, but did not constrain, the coding. The study did not employ grounded theory procedures: no theory generation, constant comparative method, or theoretical sampling was undertaken, and terminology associated with grounded theory has been avoided accordingly. Analysis proceeded through three phases: (1) initial coding, in which meaning units in the 10 transcripts were assigned inductively derived codes; (2) theme development, in which related codes were consolidated into 10 thematic nodes under the parent node “Affective Domain of Gamified Learning” and each node's operational definition was recorded in the codebook; and (3) theme review and refinement, in which the node structure was checked and refined against the full corpus. Coding was performed by the first author. The two co-authors, who did not teach the cohort, were provided with the codebook (Table 1) and the full set of coded transcripts and independently reviewed the coding of all 10 scripts against the node definitions; they did not independently re-code the data from scratch, and their role was that of critical reviewers rather than second coders. Discrepancies in coding were identified in a minority of the 10 transcripts and were resolved through discussion until consensus was reached; where discussion exposed ambiguity in a node definition, the codebook was revised and the affected segments were re-checked against the revised definition. No formal intercoder agreement statistic (such as percentage agreement or Cohen's kappa) was computed. This is consistent with a codebook approach to thematic analysis, in which coding quality is pursued through documented definitions, an audit trail, and consensus dialogue that treats coding as an interpretive rather than purely mechanical act; the limitations of relying on consensus review rather than independent double-coding are acknowledged in Section 8. Thematic saturation was monitored by tracking the emergence of new nodes across successive transcripts; no new nodes emerged after the eighth transcript, and the final two transcripts served to confirm and enrich existing themes. An audit trail comprising the NVivo project file, node definitions, coding memos, and successive versions of the node hierarchy was maintained throughout.

Table 1. Study codebook: node definitions and indicative excerpts.

Table 1 presents the study codebook, comprising each node's operational definition and an indicative excerpt. A segment was coded to a node only where the participant's utterance matched the node definition, and a single segment could be coded to multiple nodes where warranted.

Throughout this article, “coverage” denotes the percentage of the total coded transcript text (measured in characters) assigned to a given node. Because a single segment could be coded to multiple nodes, coverage percentages are not mutually exclusive and sum to more than 100% across the 10 nodes (see Table 2). Supplementary NVivo procedures were employed to describe the coding and interrogate the robustness of the themes: a word frequency query (stemming enabled, minimum word length of four characters, stop words excluded), a coding summary of sources, references, and coverage per node, a cluster analysis of theme co-occurrence, and sentiment coding of platform-referenced segments. These outputs are reported as descriptive summaries of the coding, not as statistical evidence; interpretive weight rests on the coded excerpts themselves, presented in Sections 5.7 and 5.8.

Table 2. NVivo coding summary by theme (n = 10 participants).

Coverage denotes the percentage of the total coded transcript text assigned to each node. Because a single segment could be coded to multiple nodes, coverage values are not mutually exclusive and sum to 110.0% rather than 100%; the total therefore exceeds 100% by design, and no average is reported.

4.7 Trustworthiness, reflexivity, and ethical considerations

Trustworthiness was pursued through the criteria of Lincoln and Guba (1985). Credibility was supported by prolonged engagement (the first author taught the cohort for the full semester) and by anchoring every theme in verbatim participant quotations; dependability by the documented three-phase coding procedure, the codebook (Table 1), and the NVivo audit trail; and confirmability by co-authors' review of the coding, with disagreements resolved through discussion. Transferability is supported through thick description of the setting, intervention, and participants in Sections 4.2 and 4.4, while its limits are acknowledged in Section 8.

Reflexivity was an explicit concern. The first author occupied an insider position as the course instructor, which facilitated rapport and contextual understanding of participants’ references, but which also created a risk of socially desirable responding. To mitigate this risk, participants were assured that their views would have no bearing on their assessment, and the co-authors, who did not teach the cohort, challenged interpretations during analysis and pressed for negative or discrepant evidence, which is reflected in the mixed-sentiment themes reported in Section 5. The interviews were conducted at the end of the semester, after final grades had been submitted, so that participation could have no bearing on assessment.

The study was conducted during the COVID-19 lockdown period (2020), when all teaching, learning, and data collection were conducted online. At the time the study was carried out, formal Institutional Ethics Committee approval was not required for this classroom-based educational activity at our institution, and no ethics approval was therefore obtained. Participation in the interviews was entirely voluntary. All participants were informed about the purpose of the study, provided informed consent electronically before participating, and were assured that their responses would remain confidential and would be used only for research purposes. No personally identifiable information was collected, and all interview data were anonymised prior to analysis. To minimise any perception of coercion, participation was voluntary, students could withdraw at any time without consequence, and the interviews were conducted only after completion of the course and submission of final grades. Perceived pressure to participate was minimised in three ways: recruitment proceeded through a single general announcement to the whole cohort rather than through individual approaches, so that non-participation was not visible to the instructor; participation carried no academic credit, incentive, or penalty; and participants were explicitly assured, both at recruitment and at the start of each interview, that their decision to participate and the views they expressed would have no bearing on their assessment. All participants provided informed consent prior to the interviews; participation was voluntary with the right to withdraw at any time, and confidentiality was protected by pseudonymising participants as P1–P10 and storing recordings and transcripts on a password-protected device accessible only to the research team.

5 Findings

5.1 Thematic structure

Open coding of the 10 interview transcripts yielded 10 thematic nodes as first-order constructs under the parent node “Affective Domain of Gamified Learning (Figure 1).” These nodes collectively capture the emotional, motivational, and social dimensions of students’ experience with the Kahoot and LearningApps platforms.

The 10 nodes are organised into three clusters: (1) social-affective nodes (Self-Confidence, Social Presence, Preference for Playing Together); (2) engagement nodes (Engaging Nature, Language Suitability, Motivation); and (3) instructional-pragmatic nodes (Readiness, Time Engagement, Seeking Challenges, Preferences). This architecture reflects the multi-dimensional influence of gamified flipped learning on students’ inner states and learning behaviours.

5.2 Word frequency analysis

A word frequency query was run across all 10 transcripts (Figure 2) as a descriptive orientation to the corpus. The prominence of “gamification,” “learning,” “students,” and “games” indicates that participants spoke substantively to the research topic, and terms such as “motivation,” “skills,” “vocabulary,” “confidence,” and “interaction” recur in ways broadly consistent with the coded themes. Frequency of occurrence is, however, not treated as evidence of thematic importance; the interpretive weight of the analysis rests on the coded excerpts presented in Sections 5.7 and 5.8.

5.3 Coding summary

Table 2 and Figure 3 presents the NVivo coding summary for the 10 thematic nodes, showing the number of sources coded, total references, percentage coverage of the corpus, and overall sentiment.

Motivation (17.9% coverage, 22 references) and Engaging Nature (16.3%, 20 references) emerged as the two dominant nodes, indicating that sustained motivation and the intrinsically engaging quality of the games were the most pervasive affective outcomes. Readiness to Gamified Learning had the lowest coverage (7.8%), reflecting that unfamiliarity with gamified instruction was acknowledged but was not the primary affective driver.

5.4 Distribution of themes across participants

Table 2 also shows how widely each theme was distributed across the 10 participants. Motivation and Engaging Nature were coded in all 10 transcripts (22 and 20 references, respectively) and Language Suitability in nine, consistent with their position as the most prominent affective themes. Preference for Playing Together (eight transcripts), Social Presence, and Suitability of Time Engagement (seven each) were present across the majority of participants, reflecting the collaborative affordances of the platforms. By contrast, Self-Confidence and Preferences were each coded in only five transcripts, and Readiness to Gamified Learning and Seeking Challenges in six, suggesting participant-level variability in confidence gains and prior gaming experience.

5.5 Cluster analysis: theme co-occurrence

A cluster analysis based on node co-occurrence identified which themes consistently appeared together within the same transcript segments. Co-occurrence was operationalised as the number of transcript segments coded to both nodes, using NVivo cluster analysis with Jaccard's coefficient as the similarity metric, coding references within the same segment as the unit of association, and a minimum co-occurrence threshold of [X] for displaying an edge. Motivation occupies the most central position in the resulting network (Figure 4), co-occurring frequently with Engaging Nature, Seeking Challenges, Social Presence, and Preference for Playing Together. Because no formal network centrality statistics were computed, this pattern is presented as a descriptive display of co-occurrence in the coding, suggesting, rather than statistically confirming, that Motivation functioned as the affective hub connecting the other themes. The tight coupling of Social Presence and Playing Together reflects students’ strong preference for collaborative gameplay, while the connection between Language Suitability and Engaging Nature indicates that students perceived the gamified content as authentically relevant to their language learning goals.

5.6 Affective influence by platform

A comparative reading of the coded segments referencing each platform, together with their NVivo sentiment coding, indicated complementary affective profiles rather than the superiority of either platform. Segments referencing LearningApps were more often associated with the Social Presence and Language Suitability themes and carried predominantly positive sentiment, which participants attributed to its diverse game templates, including cloze tests, matching pairs, and vocabulary games, that directly target language skills. Segments referencing Kahoot were more often associated with the Engaging Nature and Motivation themes, with participants describing the arousal and excitement generated by its competitive, timed quiz format. Not all participants commented on both platforms, and the number of references per platform was uneven across transcripts; these comparisons are therefore reported as descriptive tendencies in the coded data rather than as numerical platform ratings, and interpretive weight rests on the participant accounts presented in Sections 5.7 and 5.8.

5.7 Detailed thematic findings

The 10 emergent themes are presented below with supporting participant quotations drawn from NVivo-coded transcripts.

5.7.1 Theme 1: self-confidence

Students reported that both individual and group gameplay contributed to gains in self-confidence. Playing individually allowed students to test their abilities without peer pressure, while group play scaffolded weaker students’ confidence through peer support. This finding aligns with Bandura's (1963) self-efficacy constructs, wherein vicarious learning and mastery experiences reinforce students’ belief in their own language competence.

“… to improve the quality of our interaction, I prefer to play games in groups … I know how to help others … to gain confidence, I play individually too …” (P4).

5.7.2 Theme 2: social presence

Gamification provided a digital social space that substituted for face-to-face classroom interaction during the pandemic. Students expressed a need for peer connection and actively sought collaborative game modes. Fischer and Barabasch (2020) similarly note that gamification not only strengthens knowledge through application, synthesis, analysis, and evaluation, but also helps students socialise, confirming the importance of social presence, interaction, and collaboration among players.

“… sometimes played in a group … based on topic, I played in a group … if familiar with the topic, I play alone … Otherwise, I learn from them …” (P7).

5.7.3 Theme 3: preference for playing together

A clear preference for collaborative gameplay emerged, particularly for unfamiliar or challenging topics. Group play facilitated peer-mediated learning, consistent with Vygotsky's Zone of Proximal Development.

“… needed help from others, that is why I played in groups … group members made positive comments about games …” (P3, P6).

5.7.4 Theme 4: engaging nature

The novelty and interactivity of the gamified tasks were powerful engagement drivers. Students reported voluntarily engaging with games beyond class hours, indicating intrinsic motivation, and several described the games as demanding playful, imaginative engagement rather than rote recall of material. This kind of absorbing, imaginative engagement plausibly explains why games also alleviate the stress otherwise associated with traditionally taught classes (Ang et al., 2018): when attention is captured by the enjoyable, imaginative demands of a task, learners have correspondingly less cognitive space available for the evaluative anxiety that formal classroom assessment can provoke.

“… I [have] never seen games like these for any subject … I like to play beyond the class time too … very useful …” (P2).

“… picture description demands our imagination … very interesting …” (P5).

5.7.5 Theme 5: language suitability

Students explicitly connected the gamified activities to improvements in vocabulary, grammar, speaking, writing, and reading, validating the pedagogical design of the game-embedded syllabus.

“… a lot of words I learnt … difference between similar words I learnt …” (P6).

“… grammar also I improved … speaking and writing skill have improved …” (P9).

5.7.6 Theme 6: readiness for gamified learning

Most participants had no prior experience of gamified classroom instruction, which represented both a challenge and an opportunity. Initial unfamiliarity gave way to engagement as students acclimatised to the platforms.

“… at school I did not play … first time I play … Never before I played games in the class …” (P1, P8).

5.7.7 Theme 7: motivation

Kahoot's timed, competitive format created heightened motivation even under challenging game conditions, and students reported persisting with difficult activities rather than abandoning them. Flipping the class with gamified material encourages active online participation (Lo and Hew, 2017), and a well-designed game sustains motivation (Takatalo, 2011). If a game provides few opportunities for meaningful action, it is likely to result in student boredom; by contrast, continuous, immediate feedback sustains the flow of action, produces absorption, and forestalls anxiety.

“… Kahoot challenging … arranging the sentences is a bit difficult … but I didn't give up …” (P2).

5.7.8 Theme 8: suitability of time engagement

Students expressed varied preferences regarding the optimal timing of games within the class structure. A common preference was to receive conceptual explanation before playing, revealing the partial ineffectiveness of the pre-class flipping strategy. A pleasant ambience sustains motivation, but games must not displace attentiveness to the application of concepts (Ang et al., 2018); the desire for in-class explanation indicates insufficient pre-class preparation with the materials uploaded to WhatsApp.

“… after introducing the concept and let students settle, I could play …” (P4).

“… teach the concept and then allow us to play… after introducing the points…” (P7).

5.7.9 Theme 9: seeking challenges

Students actively sought challenges in the games, including playing outside class hours to find answers and resolve the cognitive dissonance created by difficult tasks, evidence of deep engagement. At the same time, uninterrupted continuous play produced boredom for some, a challenge attributable in part to unfamiliarity with the flipped format (Lo and Hew, 2017).

“… outside the class time too … I wanted to find the answers …” (P5).

“… within the class time … tenses … without any break … I felt bored to play continuously …” (P3).

5.7.10 Theme 10: preferences

LearningApps was somewhat preferred for sustained independent learning, while Kahoot was favoured for excitement and competitive play; both platforms were valued for distinct pedagogical functions. Regular play of Kahoot games has been shown to alter class dynamics over time (Wang and Tahir, 2020).

“… Learning App is more useful … card games in Learning App & Kahoot games I liked …” (P9).

5.8 Representative participant quotations

Table 3 presents selected participant quotations for key themes, as extracted and coded in NVivo; these quotations serve as verbatim evidence anchoring the thematic interpretations.

Table 3. Representative NVivo-coded participant quotations by theme.

5.9 Mapping findings to the theoretical framework

The analysis confirms the resonance of the study's multi-theoretical framework with the emergent themes. Table 4 maps each theoretical lens to the themes it best explains, together with supporting evidence from the coded data.

Table 4. Theoretical frameworks mapped to emergent themes.

6 Discussion

The central finding, that motivation and the engaging nature of the games were the most pervasive affective outcomes, extends a substantial quantitative literature into the qualitative register. The presence of Motivation in all 10 transcripts, and its centrality in the co-occurrence network, are consistent with the sustained motivation reported in gamified flipped designs by Aşıksoy (2018), Pozo Sánchez et al. (2020), Arsyad et al. (2024), and Daliranfirouz et al. (2025), and with Özdemir's (2025) meta-analytic evidence that Kahoot positively influences motivation and attitude. Interpreted through self-determination theory, participants’ accounts of choosing when, how, and with whom to play (autonomy), of persisting through appropriately difficult tasks with immediate feedback (competence), and of learning with and from peers (relatedness) suggest that the design satisfied the three basic psychological needs that underpin intrinsic motivation (Ryan and Deci, 2000).

Participants’ reports of persisting with difficult Kahoot tasks, seeking answers outside class hours, and losing track of class time correspond closely to the components of flow, an optimal challenge–skill balance, immediate feedback, and autotelic absorption, described by Csikszentmihalyi (1990). Similarly, the report of boredom during uninterrupted, repetitive play (P3) illustrates the fragility of the flow state when challenge ceases to be calibrated and supports Singhal et al.'s (2019) recommendation to limit continuous gameplay. The novelty effect documented by Tsay et al. (2020) offers a complementary caution: some of the enthusiasm observed here may attenuate as gamified instruction becomes routine, which underscores the need for the longitudinal designs proposed in Section 8.

The social-affective cluster of themes is best interpreted through the constructivist lenses. The tight coupling of Social Presence and Preference for Playing Together indicates that group gameplay functioned as a ZPD in Vygotsky's sense, with more capable peers scaffolding others, precisely the mechanism through which repeated exposure is theorised to move learners towards automaticity (Rachels, 2016) and to afford the noticing of gaps in one's interlanguage (Schmidt and Frota, 1986). The pattern whereby participants played alone on familiar topics but in groups on unfamiliar ones (P7) shows learners self-regulating their position within the ZPD, alternating between mastery experience and vicarious learning in a manner directly predicted by Bandura's (1963) self-efficacy constructs and reflected in the Self-Confidence theme.

The most pedagogically consequential finding concerns the flipped component. Participants’ consistent preference for in-class conceptual explanation before gameplay reveals that the pre-class materials distributed via WhatsApp were inconsistently used, whether through an inability to understand them independently or through a failure to prepare, making in-class recapitulation unavoidable. This qualifies the optimistic accounts of flipping in Herreid and Schiller (2013) and aligns with Lo and Hew's (2017) identification of student unfamiliarity and inadequate pre-class preparation as principal challenges of flipped instruction; it also resonates with the increased workload reported by learners in Daliranfirouz et al. (2025). The absence of formative assessment at the start of each class, which would have verified pre-class preparation, is acknowledged as a design limitation. For practice, the finding argues for deploying games after brief in-class conceptual input, or for gating gameplay behind short readiness checks, rather than relying on unmonitored pre-class study.

Finally, the platform comparison indicates complementarity rather than superiority: Kahoot's competitive, timed format generated arousal, excitement, and attention re-capture, whereas LearningApps’ diverse templates supported sustained, skill-specific independent practice. This complementarity, together with the mixed sentiment recorded for Readiness, Time Engagement, and Seeking Challenges, mirrors the mixed reactions reported by Buckley et al. (2017) and suggests that the affective benefits of gamification are conditional on design choices, timing, variety, grouping, and challenge calibration, rather than automatic.

These findings must also be weighed against the critical literature reviewed in Section 2.4, and the study's claims are moderated accordingly. The positive affect observed here derives from a single semester of exposure and from self-reported perceptions. Hanus and Fox (2015) demonstrate that motivational gains from gamification can decline, and even reverse, over longer exposures; Toda et al. (2018) catalogue negative effects ranging from indifference to undesired competition; and Koivisto and Hamari (2019) caution that much of the field rests on short, self-report designs of precisely the present kind. The study therefore makes no claim about language learning outcomes: in the absence of pre/post proficiency measures, performance scores, or triangulated data, the findings speak to participants’ perceived affective states, exploratory and bound to one course, one institution, and one pandemic-era semester, and to the design conditions under which those states arose. Machajewski's (2017) case-study observation that gamifying essential syllabus components engages contemporary students is consistent with, but not established by, the present data.

7 Conclusion and pedagogical implications

This semester-long qualitative case study examined the influence of gamified flipped online instruction on the affective domain of undergraduate English language learning in an Indian university. The NVivo-assisted thematic analysis of 10 interview transcripts indicates that participants perceived the intervention as positively influencing the affective domain across all ten emergent themes, with Motivation, Engaging Nature, and Language Suitability as the most consistently reported. These conclusions are exploratory and bound to the context of a single course at a single institution during the pandemic; they concern perceived affective states, not measured learning outcomes. Students engaged with Kahoot and LearningApps to satisfy needs for entertainment, interaction, and knowledge acquisition; experienced autonomy through self-paced play, competence through feedback and rewards, and relatedness through group activity; and reported gains in confidence, motivation, and participation. Group-based gameplay and peer discussion facilitated knowledge sharing and social interaction, and the balance of challenge and skill in quizzes, matching exercises, and interactive games sustained attention and interest. The study thereby contributes qualitative evidence to a field predominantly served by quantitative designs, addressing the gap identified by Ekici (2021), and demonstrates that each of its six theoretical lenses has direct corresponding evidence in the coded data.

Four tentative implications follow for language teachers considering gamification. First, games should be deployed after conceptual introduction, rather than relying solely on pre-class preparation, since learners in this study inconsistently used the flipped materials. Second, competitive platforms (Kahoot) and skill-oriented platforms (LearningApps) should be used in a complementary manner. Third, both individual and group gameplay modes should be included to develop confidence alongside social presence. Fourth, games should offer appropriately graduated challenge levels, interspersed with breaks, to sustain flow state engagement and avert boredom.

8 Limitations and directions for future research

The findings must be read against several limitations. First, although the sample size is justified on grounds of information power and observed thematic saturation (Section 4.3), 10 volunteer participants from a single course at a single institution constrain transferability; enrolment in the course was itself reduced because, during the initial pandemic lockdown, the course was offered in two stages and this study covered Stage I. Second, the study relied on a single data source (end-of-semester interviews); all data are self-reported perceptions, and the first author's dual role as instructor and interviewer may have encouraged socially desirable responding despite the mitigation measures described in Section 4.7, and coding was performed by a single coder subject to co-author consensus review rather than independent double-coding. Triangulation with observation data, platform analytics, or pre- and post-intervention proficiency measures would strengthen future work. Third, the gamified component targeted knowledge of language (grammar, vocabulary, and discourse conventions) rather than the four macro skills of listening, speaking, reading, and writing. Fourth, the absence of formative assessment at the start of each class limits the reliability of inferences about the flipped component. Future research should therefore employ larger and more diverse samples across arts, science, and engineering streams; triangulate interview data with pre-/post-language proficiency tests and platform analytics; extend the observation period to capture long-term affective and linguistic outcomes beyond the novelty phase (Tsay et al., 2020); and investigate the efficacy of gamified flipped designs for the four macro skills in Indian higher education contexts.

Generative AI statement

The author(s) declared that generative AI was used in the creation of this manuscript. Figures 1–4 were prepared with the assistance of Microsoft Copilot (GPT-4, accessed October 2025), which was used to generate the graphical layouts from author-provided data/format and author-prepared visualisations. The tool did not generate, analyse, or interpret any data: the values, node structures, and coding statistics depicted in all figures derive from the authors’ own NVivo 14 analysis of the interview data, and every figure was checked against the underlying NVivo outputs and verified for accuracy by the authors. Generative AI Microsoft Copilot was additionally used to assist with language editing and formatting during revision of the manuscript; it was not used to generate content, findings, or interpretations. All content, interpretations, and conclusions were reviewed, verified, and approved by the authors, who take full responsibility for the work.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence, and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Jose Belda-Medina, University of Alicante, Spain

Reviewed by: Francis G. Balazon, Batangas State University, Philippines

Rosidah Rosidah, Yogyakarta State University, Indonesia