Gamification on Mathematics Engagement and Motivation in Secondary School and Higher Education: A Systematic Review and Meta-Analysis

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Gamification has emerged as a promising pedagogical strategy with potential to capture attention, foster active learning, and enrich learning experience, particularly in subjects where disengagement and low motivation are common.

“Gamification has emerged as a promising pedagogical strategy, with the potential to capture students' attention, foster active learning, and enrich the overall learning experience, particularly in subjects where disengagement and low motivation are common (Dichev & Dicheva, 2017).”

• cites Dichev & Dicheva, 2017 sourceSupported

May combine two claims.

In mathematics, incorporating game elements into the curriculum can create more dynamic and interactive experiences, motivating students to participate more actively.

“In mathematics, incorporating game elements into the curriculum can create more dynamic and interactive experiences, motivating students to participate more actively (Sailer & Homner, 2020;Subhash & Cudney, 2018).”

• cites Sailer & Homner, 2020 sourcePartly supported

  The Motivational Learning Outcomes section reports small, significant effects of gamification on motivational and behavioral learning outcomes. The Learning Arrangement of the Comparison Group section also says gamification can create opportunities for active learning. These passages support the general point about motivation and participation, but they do not establish the claim specifically for mathematics curricula or show that mathematics lessons become more dynamic and interactive. (LLM, conf=0.86)

  Experimental — accuracy not yet measured
• cites Subhash & Cudney, 2018 sourceNot checked — no full text of the source

Classified as probably not a substantial claim.

Digital and game-inspired strategies have emerged as promising approaches to improve mathematics learning in secondary and higher education.

“Digital and game-inspired strategies have emerged as promising approaches to face these challenges, providing innovative methods to improve mathematics learning in secondary and higher education contexts (Faure-Carvallo et al., 2022).”

• cites Faure-Carvallo et al., 2022 sourcePartly supported

  Section 3.2.1 (Beneficios didácticos; page unavailable) reports that adolescents using a digital game in mathematics classes were more engaged and performed better on skills taught by the game. The review concerns secondary education; it does not establish improved mathematics learning in higher education. (LLM, conf=0.87)

  Experimental — accuracy not yet measured

May combine two claims.

Gamified designs that balance difficulty and ability can help sustain motivation and foster a flow state.

“Gamified designs that balance difficulty and ability can help sustain motivation and foster a flow state (Khoshnoodifar et al., 2023;Oliveira et al., 2021).”

• cites Khoshnoodifar et al., 2023Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Oliveira et al., 2021 sourcePartly supported

  The review’s RQ 4 discussion (p. 4) reports that a challenge-oriented gamification design positively affected flow in one study, while other studies found no significant effect. It does not establish that balancing difficulty with ability sustains motivation; the reported outcome was intention to continue use, not motivation. (LLM, conf=0.83)

  Experimental — accuracy not yet measured

May combine two claims.

Some studies show positive outcomes of gamification, while others report declines in motivation due to competition, poor design, or reduced novelty.

“While some studies show positive outcomes (e.g., Sailer & Homner, 2020), others report declines in motivation due to competition, poor design, or reduced novelty (e.g., Hanus & Fox, 2015;Santhanam et al., 2016).”

• cites Sailer & Homner, 2020 sourceSupported
• cites Hanus & Fox, 2015 sourceNot checked — no full text of the source
• cites Santhanam et al., 2016 sourceNot checked — no full text of the source

May combine two claims.

Early mathematics emphasizes basic numeracy and playful learning, whereas later stages involve more abstract, demanding and advanced content, where math anxiety and declining motivation and engagement become especially salient.

“Whereas early mathematics emphasizes basic numeracy and playful learning, later stages involve more abstract, demanding and advanced content, where issues such as math anxiety and declining motivation and engagement become especially salient (Kaur et al., 2022;Pekrun et al., 2017).”

• cites Kaur et al., 2022 sourcePartly supported

  The Mathematics education and Student self-regulation factors sections report declining attitudes toward mathematics and increased anxiety around the primary-to-secondary transition. The section on implications of transition also reports declines in motivation and engagement among secondary students. The paper does not establish the claimed contrast between early numeracy and playful learning and later abstract content; its School-related and academic factors section says teaching practices, rather than increased content or workload, pose the greater transition challenge. The supplied passages have no page numbers. (LLM, conf=0.88)

  Experimental — accuracy not yet measured
• cites Pekrun et al., 2017 sourcePartly supported

  The Discussion (page unavailable) reports that math anxiety and other negative emotions predicted adolescents’ later math grades. The grade 5–9 descriptive table (page unavailable) also shows declining enjoyment and rising boredom. The paper does not compare playful early mathematics with more abstract later content, or establish a decline in motivation and engagement. (LLM, conf=0.84)

  Experimental — accuracy not yet measured

May combine two claims.

The ARCS Model aligns with gamification: attention via challenges or visual rewards, relevance via real-world goals, confidence via scaffolded levels, and satisfaction via immediate feedback or badges.

“The ARCS Model (i.e., Attention, Relevance, Confidence, and Satisfaction; Keller, 1987) also aligns with gamification.Attention can be captured through challenges or visual rewards; relevance by linking tasks to real-world goals; confidence by scaffolded levels; and satisfaction by immediate feedback or badges (Mese & Dursun, 2019;Özhan & Kocadere, 2020).”

• cites Keller, 1987 sourceNot checked — no full text of the source
• cites Mese & Dursun, 2019 sourcePartly supported

  The Discussion and Conclusion (page unavailable) links rewards and advancing through levels to the satisfaction dimension of the ARCS-V model. The Application and Progress Bar Leaderboard sections (pages unavailable) describe instant feedback, badges, and level-based activities. The paper does not make the claimed mappings from challenges or visual rewards to attention, real-world goals to relevance, or scaffolded levels to confidence. (LLM, conf=0.86)

  Experimental — accuracy not yet measured
• cites Özhan & Kocadere, 2020 sourceNot checked — no full text of the source

May combine two claims.

Gamification can encourage mastery via personal progress tracking or performance through competitive leaderboards.

“Gamification can encourage mastery via personal progress tracking, or performance through competitive leaderboards (Durmaz et al., 2022;Hakulinen & Auvinen, 2014).”

• cites Durmaz et al., 2022 sourceSupported
• cites Hakulinen & Auvinen, 2014 sourceNot checked — no full text of the source

May combine two claims.

In mathematics, particularly at secondary and higher education levels, research remains scarce and often constrained by small samples or limited methodological rigor.

“In mathematics, particularly at the secondary and higher education levels, research remains scarce and often constrained by small samples or limited methodological rigor (Elmawati et al., 2023;Ratinho & Martins, 2023).”

• cites Elmawati et al., 2023 sourcePartly supported

  The Methods and Data section (page unavailable) reports a limited set of publications on gamification in mathematics learning, offering some support for the claim that this research is scarce. It does not establish that scarcity is particular to secondary or higher education, and neither that section nor the Limitations section assesses the studies' sample sizes or methodological rigor. (LLM, conf=0.74)

  Experimental — accuracy not yet measured
• cites Ratinho & Martins, 2023 sourcePartly supported

  The Results and discussion section reports only four studies of gamification and motivation among high-school students, which supports scarcity at the secondary level for that topic. It gives sample sizes of 22–683 across the reviewed studies, but does not characterize them as inadequate or establish that mathematics research at either educational level is scarce. The Conclusions section says the review did not evaluate study quality, so it cannot substantiate the claim about limited methodological rigor. (LLM, conf=0.82)

  Experimental — accuracy not yet measured

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The review followed PRISMA guidelines and was registered in PROSPERO.

“This review followed the PRISMA guidelines (Page et al., 2021) and was registered in PROSPERO (ID-CRD42023481878).”

• cites Page et al., 2021 sourceSupported

May combine two claims.

Engagement was assessed far less frequently; only eight studies measured this construct, using either a unidimensional approach or specific subdimensions.

“Engagement was assessed far less frequently.Only eight studies measured this construct, using either a unidimensional approach (Boom-Cárcamo et al., 2024;Chen et al., 2018;Lee et al., 2023;Wang et al., 2024) or specific subdimensions such as study-skills and interaction engagement (Feroz et al., 2020;Pimentel et al., 2022), cognitive and emotional engagement (Santos-Guevara et al., 2024), or”

• cites Boom-Cárcamo et al., 2024 sourceNot checked — no full text of the source
• cites Chen et al., 2018Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Lee et al., 2023 sourceSupported

Classified as probably not a substantial claim. May combine two claims.

The quality of included studies was independently assessed by two authors using the Mixed Methods Appraisal Tool, with disagreements resolved through discussion.

“The quality of the included studies was independently assessed by two authors using the Mixed Methods Appraisal Tool (MMAT) (Hong et al., 2018).Disagreements were resolved through discussion.”

• cites Hong et al., 2018 sourceSupported

May combine two claims.

Regarding Social Cognitive Theory, most studies documented improvements in self-efficacy beliefs in mathematics, although one study reported a decline.

“Regarding SCT, most studies documented improvements in self-efficacy beliefs in mathematics (García-Hernández & González-Ramírez, 2021;Hammad et al., 2022;McIntyre, 2023;Montero-Izquierdo et al., 2024;Robinson et al., 2023), although one study reported a decline (García-López et al., 2023).”

• cites García-Hernández & González-Ramírez, 2021 sourceSupported
• cites García-López et al., 2023 sourceSupported

May combine two claims.

Most studies using the ARCS model reported increases across all four dimensions (Attention, Relevance, Confidence, Satisfaction), with particularly high attention scores in one case.

“Most studies using the ARCS model reported increases across all four dimensions, Attention, Relevance, Confidence, and Satisfaction (Chen et al., 2023;Jiménez-Hernández et al., 2020;Zabala-Vargas et al., 2019, 2021), with particularly high attention scores in one case (Hernandez-Mena et al., 2024).”

• cites Chen et al., 2023 sourcePartly supported

  The Methods > Instruments passage identifies Attention, Relevance, Confidence, and Satisfaction as the four factors in the ARCS-based motivation scale. The Results > Learning motivation analysis passage reports a significant increase in the experimental group’s overall motivation, but does not report increases for each factor separately. The particularly high attention scores are attributed to a different citation in the citing sentence, not to Chen et al. (2023). Page numbers are unavailable in our copy. (LLM, conf=0.86)

  Experimental — accuracy not yet measured
• cites Jiménez-Hernández et al., 2020 sourcePartly supported

  Table 7 (shown under §8) reports higher experimental-group means than control-group means for all four ARCS dimensions. However, §7.2 reports statistically significant differences only for Attention, Relevance, and Satisfaction; Confidence was not significant. The questionnaire was administered only at posttest (§6.3), so the paper does not establish increases over time. (LLM, conf=0.86)

  Experimental — accuracy not yet measured
• cites Hernandez-Mena et al., 2024Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

May combine two claims. The quote only partly states this claim.

Although engagement was assessed less frequently, findings were generally positive: higher engagement levels, active participation, strong cognitive and emotional engagement, and greater intentions to engage when motivated; however, lower engagement was reported among low-performing students, and one study found no significant differences between gamified and traditional groups.

“Although engagement was assessed less frequently, findings were generally positive.Higher engagement levels were observed following gamification (Boom-Cárcamo et al., 2024), including active participation in tasks (Feroz et al., 2020) and strong cognitive and emotional engagement (Santos-Guevara et al., 2024).Students also reported greater intentions to engage when they felt motivated (Aljamaan et”

• cites Boom-Cárcamo et al., 2024 sourceNot checked — no full text of the source
• cites Lee et al., 2023 sourcePartly supported

  Split into 2 atoms — 0 confirmed, 2 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.

  Experimental — accuracy not yet measured
• cites Chen et al., 2018Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

May combine two claims.

Motivation was predominantly examined through SDT, followed by SCT and the ARCS model.

“motivation was predominantly examined through SDT, followed by SCT and the ARCS model, confirming SDT's role as the dominant framework (Alsawaier, 2018;Ratinho & Martins, 2023).”

• cites Alsawaier, 2018 sourceNot checked — no full text of the source
• cites Ratinho & Martins, 2023 sourcePartly supported

  The Conclusions identify self-determination theory (SDT) as the most frequently explored theory, and the RQ1 results report its use in 13 studies. The Discussion mentions social learning theory and the ARCS model, but does not establish that SCT was the second most examined framework, ahead of ARCS; it says self-efficacy was relatively little explored. (LLM, conf=0.86)

  Experimental — accuracy not yet measured

May combine two claims.

Overall, the majority of the 45 studies reported positive effects of gamification on motivation and engagement in mathematics; only four studies reported negative effects, relating to intrinsic/extrinsic motivation or engagement.

“Overall, the majority of the 45 studies included in this review reported positive effects of gamification on motivation and engagement in mathematics.Only four studies reported negative effects, relating to intrinsic/extrinsic motivation (Reyssier et al., 2022;Vandercruysse et al., 2015) or engagement (Lee et al., 2023).”

• cites Reyssier et al., 2022 sourceNot checked — no full text of the source
• cites Vandercruysse et al., 2015Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Lee et al., 2023 sourcePartly supported

  Lee et al. report in the Conclusions that the increase in interest in mathematics was minimal and that the course was less effective for lower-level students. The results also describe low interest in gamification-only modules, but do not establish a negative effect on engagement; game-based modules generally drew higher interest. The count of four studies is the citing review’s own finding, not a finding attributed specifically to Lee et al. (LLM, conf=0.82)

  Experimental — accuracy not yet measured

May combine two claims. The quote only partly states this claim.

Collaborative rankings/team tasks can be experienced as challenging/competence by social learners and learners with high initial amotivation, and by intrinsic motivators.

“| Collaborative rankings/ | Social learners and | Can be demotivating to | García-López et al. |”

• cites García-López et al. (2023) sourcePartly supported

  The Introduction (page unavailable) says collaborative work can increase students’ motivation through social relationships, and that levels and scores can support competence. It does not establish how collaborative rankings or team tasks are experienced by social learners, learners with high initial amotivation, or intrinsically motivated learners. (LLM, conf=0.82)

  Experimental — accuracy not yet measured
• cites Marinho et al. (2025) sourceNot checked — no full text of the source

May combine two claims.

Leaderboards boosted motivation for some while reducing autonomy for others (Ortiz-Rojas et al., 2025).

“or leaderboards boosting motivation for some while reducing autonomy for others (Ortiz-Rojas et al., 2025).”

• cites Ortiz-Rojas et al., 2025 sourcePartly supported

  The Results (page unavailable) include a student saying the leaderboard prompted them to study and report interview themes suggesting competition was detrimental to autonomous motivation. But the study did not establish that autonomy declined for a distinct group of students; the Discussion (page unavailable) says its leaderboard design did not produce a significant change in autonomous motivation. (LLM, conf=0.85)

  Experimental — accuracy not yet measured

May combine two claims.

Timer/time pressure can be demotivating to learners high in intrinsic motivation, possibly increasing anxiety or undermining autonomy.

“| | initial amotivation; | motivation (may increase | Reyssier et al. |”

• cites Lee et al. (2023) sourceCould not verify

  The Literature Review discusses autonomy as a factor in motivation, and the Conclusions discuss how students with different motivation levels respond to gamification. Neither section addresses whether timers or time pressure demotivate highly intrinsically motivated learners, increase anxiety, or undermine autonomy (searched: timer, time pressure, anxiety, autonomy). (LLM, conf=0.74)

  Experimental — accuracy not yet measured
• cites Reyssier et al. (2022) sourceNot checked — no full text of the source
• cites Saleh Alabdulaziz (2023) sourceCould not verify

  The Discussion (page unavailable) describes timed escape-room tasks as exciting and reports improved motivation; it does not examine learners high in intrinsic motivation. The Results and Discussion (pages unavailable) address anxiety and autonomy, but do not link either to timer pressure. Searches for intrinsic motivation and time-pressure effects found no passage establishing the claimed possibility. (LLM, conf=0.74)

  Experimental — accuracy not yet measured

May combine two claims.

Rewards can be demotivating to learners high in intrinsic motivation, possibly increasing extrinsic motivation behaviors.

“| | extrinsic motivation | motivation (may increase | (2023); Vander- |”

• cites Robinson et al. (2023) sourceCould not verify

  The Introduction describes a reward hidden in a red envelope, and §3.2.2 measures students’ interest and enjoyment. Neither passage addresses whether rewards demotivate learners with high intrinsic motivation or increase extrinsically motivated behavior. Searches for those terms found no passage making that claim; page numbers were unavailable in our copy. (LLM, conf=0.76)

  Experimental — accuracy not yet measured
• cites Vandercruysse et al. (2015)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Zabala Vargas et al. (2019) sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

Rewards can be experienced as challenging/competence by learners with high extrinsic motivation.

“| Rewards | Learners with high | Learners high in intrinsic | Robinson et al. |”

• cites Robinson et al. (2023) sourcePartly supported

  The Introduction describes a reward unlocked by completing challenges, and the Results report participants’ perceived competence after the exercise. Neither passage distinguishes learners by extrinsic motivation or shows that highly extrinsically motivated learners experienced the reward as a challenge or a source of competence. (LLM, conf=0.75)

  Experimental — accuracy not yet measured
• cites Vandercruysse et al. (2015)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Zabala Vargas et al. (2019) sourceNot checked — no full text of the source

May combine two claims.

The systematic review indicates that gamification supports motivation and engagement, with only four studies reporting negative effects.

“The systematic review indicates that gamification supports motivation and engagement, with only four studies reporting negative effects.”

May combine two claims.

Most interventions used digital platforms (e.g., Kahoot!; Classcraft) and common game elements such as points, leaderboards and instant feedback.

“Most interventions used digital platforms (e.g., Kahoot!; Classcraft) and common game elements such as points, leaderboards and instant feedback.”

May combine two claims.

Motivation was assessed more consistently than engagement, which could not be included in the meta-analysis due to lack of validated measures.

“Motivation was assessed more consistently than engagement that could not be included in the meta-analysis due to the lack of validated measures.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The review identified 45 studies for qualitative synthesis and 11 for meta-analysis.

“A literature search (April 2025) followed by an updated search (November 2025) across ten databases identified 45 studies for qualitative synthesis and 11 for meta-analysis.”

May combine two claims.

Among the 45 studies, the combined sample comprised 4,345 students, including 2,435 from higher education and 1,910 from secondary school, with ages ranging from 13 to 38 years.

“Among the 45 studies included in the qualitative analysis, the combined sample comprised a total of 4,345 students, including 2,435 from higher education and 1,910 from secondary school, with ages ranging from 13 to 38 years.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Regarding education level, 17 articles focused on secondary school students, while 28 articles targeted higher education populations.

“Regarding education level, 17 articles focused on secondary school students, while 28 articles target higher education populations.”

May combine two claims.

Of these, 11 studies (9 journal articles and 2 conference papers) met the criteria for the meta-analysis, while the remaining records were excluded due to absence of control groups (n = 30) or incomplete statistical data (n = 4).

“Of these, 11 studies (9 journal articles and 2 conference papers) met the criteria for the meta-analysis, while the remaining records were excluded due to the absence of control groups (n = 30) or incomplete statistical data (n = 4).”

May combine two claims.

In total, 45 studies were included in the qualitative synthesis: 30 peer-reviewed journal articles, 14 conference papers, and one book chapter.

“In total, 45 studies were included in the qualitative synthesis: 30 peer-reviewed journal articles, 14 conference papers, and one book chapter (Fig. 1).”

May combine two claims.

Most studies scored 60% (n = 26), followed by 80% (n = 10) and 40% (n = 7); one study scored 100% and another scored 20%.

“Most studies scored 60% (n = 26), followed by 80% (n = 10) and 40% (n = 7); one study scored 100% and another scored 20%.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Publication dates for included studies ranged from 2015 to 2025, showing a significant increase beginning in 2021, with the highest number in 2023.

“Publication dates for the included studies ranged from 2015 to 2025, showing a significant increase beginning in 2021 (Fig. 2).The highest number of publications was recorded in 2023.”

Probably not a substantial claim: may not pass S3 (load-bearing).

For studies that reported mean age, the mean participant age was 17.92 years (SD = 2.35).

“Although several studies did not report the mean age, for the studies that reported it, the mean participant age was 17.92 years (SD = 2.35).”

May combine two claims.

The ARCS model was common (n = 6), and Social Cognitive Theory was used to assess self-efficacy beliefs (n = 8).

“The ARCS model was also common, capturing Attention, Relevance, Confidence, and Satisfaction (n = 6), while SCT was used to assess self-efficacy beliefs (n = 8).”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Most articles addressed general Mathematics (n = 27), followed by Statistics (n = 8) and Calculus (n = 5).

“Most articles addressed general Mathematics (n = 27) followed by Statistics (n = 8) and Calculus (n = 5) subjects.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Kahoot! emerged as the most used tool, appearing in 9 studies.

“Kahoot!emerged as the most used tool, appearing in 9 studies (n = 9).”

May combine two claims.

Within Self-Determination Theory, most studies found increases in both intrinsic and extrinsic motivation, but two studies reported declines in both types, and one found high motivational levels but no significant gains compared with a control group.

“Within the SDT framework, most studies found increases in both intrinsic and extrinsic motivation (Khalid & Zainuddin, 2020;Kickmeier-Rust & Niggli, 2023;Ooge et al., 2024;Robinson et al., 2023).However, two studies reported declines in both types of motivation (Reyssier et al., 2022;Vandercruysse et al., 2015), and one found high motivational levels, but no significant gains compared with a”

• cites Khalid & Zainuddin, 2020Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Reyssier et al., 2022 sourceNot checked — no full text of the source
• cites Dicheva et al., 2021 sourceNot checked — no full text of the source

May combine two claims.

The most frequent game elements were points (n = 34), leaderboards (n = 31), feedback (n = 26), and levels (n = 23); badges (n = 12), timers (n = 14), and rewards (n = 11) also appeared consistently, while missions (n = 3), narrative (n = 8), and progress bars (n = 6) were less common.

“According to that, the most frequent elements were points (n = 34), leaderboards (n = 31), feedback (n = 26), and levels (n = 23).Badges (n = 12), timers (n = 14), and rewards (n = 11) also appeared consistently, while missions (n = 3), narrative (n = 8), and progress bars (n = 6) were less common.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The included studies originated from 23 countries, with largest representation from Mexico and Spain (n = 5 each), followed by the United States of America (n = 4).

“The included studies originated from 23 countries, with the largest representation from Mexico and Spain (n = 5, each), followed by the United States of America (n = 4).”

Probably not a substantial claim: may not pass S3 (load-bearing).

An updated search conducted on November 10 (2025) identified one additional eligible study.

“An updated search conducted on November 10 (2025) identified one additional eligible study.”

May combine two claims.

The most frequent methodological issues involved lack of randomization, limited control of confounders, incomplete reporting, and insufficient methodological detail.

“The most frequent methodological issues involved lack of randomization, limited control of confounders, incomplete reporting, and insufficient methodological detail (see Online Resource 3).”

May combine two claims.

Peer-reviewed articles showed the full range of MMAT scores (20%-100%), whereas conference papers were mostly rated 60% (n = 8) or 80% (n = 2), with one paper at 40%; the single book chapter scored 60%.

“When examined by publication type, peer-reviewed articles showed the full range of scores (20%-100%), whereas conference papers were mostly rated 60% (n = 8) or 80% (n = 2), with one paper at 40%.The single book chapter scored 60%.”

May combine two claims.

Motivation was assessed either as a unidimensional construct (n = 16) or as a multidimensional construct grounded in established motivational theories.

“Motivation was assessed either as a unidimensional construct, treated as a general drive to learn and measured through a single global factor (n = 16), or as a multidimensional construct grounded in established motivational theories.”

Probably not a substantial claim: may not pass S3 (load-bearing).

A snowballing procedure yielded no additional studies.

“A snowballing procedure yielded no additional studies.”

May combine two claims.

Varying the assumed within-study correlation (rho = 0.30, 0.60, 0.90) produced similar estimates, indicating results are stable across different dependence assumptions.

“First, varying the assumed within-study correlation (ρ = 0.30, 0.60, 0.90) produced similar estimates (Table 3), indicating that the results are stable across different dependence assumptions.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Table 2 shows that behavioral (n = 6) and emotional (n = 4) engagement indicators are the most frequent.

“Table 2 summarizes these indicators and their theoretical categorization and shows that behavioral (n = 6) and emotional (n = 4) are the most frequent.”

May combine two claims.

Three moderators were tested using the CHE model with RVE: educational level, methodological quality (MMAT), and presence of competitive elements; none explained a significant proportion of between-effect variation.

“Three moderators were tested using the CHE model with RVE (CR2-adjusted standard errors): educational level, methodological quality (MMAT), and the presence of competitive elements.None explained a significant proportion of betweeneffect variation.”

• AssumptionRests on an assumption: The 11 studies included in the meta-analysis are independent.

May combine two claims.

Across 45 studies with 4,345 participants, most interventions targeted core mathematical topics such as algebra, calculus, statistics and used quasi-experimental designs.

“Across 45 studies with 4,345 participants, most interventions targeted core mathematical topics such as algebra, calculus, statistics and used quasi-experimental designs.”

May combine two claims.

A random-effects model using a single aggregated effect per study (g = 0.416, 95% CI [0.160, 0.672]) and a model restricted to peer-reviewed studies (g = 0.364, 95% CI [-0.009, 0.736]) were closely aligned with the primary CHE-RVE estimate.

“Additionally, two alternative model specifications were examined: a random-effects model using a single aggregated effect per study (g = 0.416, 95% CI [0.160, 0.672]) and a model restricted to peer-reviewed studies (g = 0.364, 95% CI [-0.009, 0.736]), both were closely aligned with the primary CHE-RVE estimate.”

May combine two claims.

Gamification type (competitive vs non-competitive) did not moderate the effects (beta = 0.02, SE = 0.31, t(7.89) = 0.05, p = .959); the intercept for competitive designs showed a moderate effect (beta = 0.38, SE = 0.17) but did not reach statistical significance (p = .081).

“Finally, gamification type (competitive vs non-competitive), coded based on explicit references to leaderboards, ranking systems, or social comparison mechanisms, did not moderate the effects (β = 0.02, SE = 0.31, t(7.89) = 0.05, p =.959).The intercept for competitive designs showed a moderate effect (β = 0.38, SE = 0.17), although it did not reach statistical significance (p =.081).”

• AssumptionRests on an assumption: The 11 studies included in the meta-analysis are independent.

May combine two claims.

The three-level CHE model indicated substantial heterogeneity: 31.56% of total variance was between-study (Level 3), 49.18% within-study (Level 2), and overall I2 was 80.74%.

“The three-level CHE model indicated substantial heterogeneity.Variance decomposition showed that 31.56% of the total variance was attributable to between-study differences (Level 3), whereas 49.18% reflected within-study heterogeneity (Level 2) arising from multiple correlated outcomes reported within the same studies.The overall I 2 was 80.74%, indicating that most of the observed variability”

• AssumptionRests on an assumption: The 11 studies included in the meta-analysis are independent.

May combine two claims.

Two of the negative-effect studies were conducted in secondary education and two in higher education, with negative outcomes observed in general mathematics (n = 2), calculus (n = 1), and algebra (n = 1).

“Two of these studies were conducted in secondary education and two in higher education, with negative outcomes observed in general mathematics (n = 2), calculus (n = 1), and algebra (n = 1).”

May combine two claims.

Excluding lower-quality studies (MMAT 60% or less) yielded a less precise but statistically consistent estimate (g = 0.107, 95% CI [-0.368, 0.581]) relative to the main model, with substantial overlap in confidence intervals.

“Second, to examine the influence of methodological quality, we excluded the lower-quality studies.As no study scored below 50% on the MMAT, the next meaningful threshold (MMAT ≤ 60%) was applied.Removing these studies yielded a less precise but statistically consistent estimate (g = 0.107, 95% CI [-0.368, 0.581]) relative to the main model, with substantial overlap in confidence intervals.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Game elements most frequently implemented were points, leaderboards, and rewards, often through platforms such as Kahoot!.

“Game elements most frequently implemented were points, leaderboards, and rewards, often through platforms such as Kahoot!(Feroz et al., 2020;Stoyanova et al., 2018)”

• cites Feroz et al., 2020 sourceNot checked — no full text of the source
• cites Stoyanova et al., 2018 sourceNot checked — no full text of the source

May combine two claims.

Educational level did not differentiate effect sizes between secondary (n = 6) and higher education (n = 5) studies, with a nonsignificant contrast term (beta = 0.10, SE = 0.30, t(7.08) = 0.33, p = .751); the intercept for higher education was also nonsignificant (beta = 0.33, SE = 0.23, p = .240).

“Educational level did not differentiate effect sizes between secondary (n = 6) and higher education (n = 5) studies, with a nonsignificant contrast term (β = 0.10, SE = 0.30, t(7.08) = 0.33, p =.751).The intercept representing higher education studies was also nonsignificant (β = 0.33, SE = 0.23, p =.240).”

• AssumptionRests on an assumption: The 11 studies included in the meta-analysis are independent.

May combine two claims.

Across the 45 included studies, gamified strategies were often associated with improvements in students' motivation and engagement, although a small number of studies (n = 4) reported negative effects.

“Across the 45 included studies, gamified strategies were often associated with improvements in students´ motivation and engagement, although a small number of studies (n = 4) reported negative effects.”

• AssumptionRests on an assumption: The quasi-experimental designs of most included studies provide valid estimates of gamification's causal effect on motivation and engagement.

May combine two claims.

Boom-Cárcamo et al. (2024) observed increased collaboration and persistence during mathematical problem-solving.

“Boom-Cárcamo et al. (2024) similarly observed increased collaboration and persistence during mathematical problem-solving.”

• cites Boom-Cárcamo et al. (2024) sourceNot checked — no full text of the source

May combine two claims.

Substantial heterogeneity was observed across studies, and limited data were available to assess engagement effects.

“However, substantial heterogeneity was observed across studies, and limited data were available to assess engagement effects.”

May combine two claims.

Most studies relied on non-validated (e.g., ad hoc surveys) or indirect measures of engagement, which prevented its inclusion in the quantitative synthesis.

“Most studies relied on non-validated (e.g., ad hoc surveys) or indirect measures of engagement, which prevented its inclusion in the quantitative synthesis.”

May combine two claims.

Based on 11 studies, the meta-analytic CHE model using RVE indicated a statistically significant positive effect of gamification on motivation (g = 0.383), representing a small to moderate increase in motivational outcomes for students exposed to gamified mathematics instruction.

“Based on 11 studies, the meta-analytic CHE model using RVE indicated a statistically significant positive effect of gamification on motivation (g = 0.383), representing a small to moderate increase in motivational outcomes for students exposed to gamified mathematics instruction.”

• AssumptionRests on an assumption: The effect sizes from the 11 studies included in the meta-analysis measure the same underlying construct of motivation and are comparable enough to be pooled.
• AssumptionRests on an assumption: The quasi-experimental designs of most included studies provide valid estimates of gamification's causal effect on motivation and engagement.

Probably not a substantial claim: may not pass S3 (load-bearing).

Some included studies (e.g., conference papers, book chapters) may not have undergone full peer review.

“Some included studies (e.g., conference papers, book chapters) may not have undergone full peer review.”

May combine two claims.

Engagement was rarely measured using validated tools and comparing an experimental group to a control group, preventing a meta-analytic synthesis of this construct.

“Engagement was rarely measured using validated tools and comparing an experimental group to a control group, preventing a meta-analytic synthesis of this construct.”

May combine two claims.

Chen et al. (2018) found higher participation but no meaningful change in perceived engagement.

“In contrast, Chen et al. (2018) found higher participation but no meaningful change in perceived engagement”

• cites Chen et al. (2018)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

May combine two claims.

Overall, gamification appears promising for enhancing motivation and engagement in mathematics when designs are aligned with students' needs, balancing competition with mastery and cooperation.

“Overall, gamification appears promising for enhancing motivation and engagement in mathematics when designs are aligned with students' needs, balancing competition with mastery and cooperation.”

May combine two claims.

Mathematics is perceived by many students as challenging and demotivating, associated with high anxiety, disengagement, and lack of motivation.

“Mathematics has long been noticed by many students as a challenging and demotivating subject, regularly associated with high levels of anxiety, disengagement, and lack of motivation (Liebendörfer & Schukajlow, 2017).”

• cites Liebendörfer & Schukajlow, 2017 sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The review followed PRISMA 2020 guidelines with a pre-registered protocol, and study quality was appraised with the Mixed Methods Appraisal Tool.

“The review followed PRISMA 2020 guidelines with a pre-registered protocol, and study quality was appraised with the Mixed Methods Appraisal Tool.”

Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

These gaps make it difficult to draw firm conclusions and highlight the need for a systematic synthesis of available evidence in these contexts.

“These gaps make it difficult to draw firm conclusions and highlight the need for a systematic synthesis of available evidence in these contexts.”

May combine two claims.

Numerous studies have examined gamification in education, but findings remain mixed; some report improvements in intrinsic motivation and engagement, others limited or inconsistent effects.

“Numerous studies have examined the effects of gamification in education, but findings remain mixed.While some report improvements in intrinsic motivation and engagement, others suggest that these effects are limited or inconsistent (e.g., Hanus & Fox, 2015).”

• cites Hanus & Fox, 2015 sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

According to Self-Determination Theory, satisfaction of autonomy, competence, and relatedness is essential to fostering intrinsic motivation.

“According to SDT, the satisfaction of autonomy, competence, and relatedness is essential to fostering intrinsic motivation (Deci & Ryan, 2000;Ryan & Deci, 2000, 2017).”

• cites Deci & Ryan, 2000 sourceNot checked — no full text of the source
• cites Ryan & Deci, 2000, 2017Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S3 (load-bearing).

According to Social Cognitive Theory, self-efficacy beliefs play a central role in shaping motivation, influencing choices, effort, and persistence.

“According to SCT (Bandura, 1977), self-efficacy beliefs play a central role in shaping motivation, influencing the choices individuals make, the effort they invest, and their persistence in the face of challenges.”

• cites Bandura, 1977 sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

Achievement Goal Theory distinguishes mastery goals focused on learning from performance goals focused on comparison.

“AGT (Dweck, 1986) distinguishes mastery goals focused on learning, from performance goals focused on comparison.”

• cites Dweck, 1986 sourceNot checked — no full text of the source

May combine two claims.

Leaderboard features emphasizing constant peer comparison may undermine self-efficacy, particularly for lower-ranked students, unless carefully framed to support personal progress and collaboration.

“In contrast, leaderboard features that emphasize constant peer comparison may undermine self-efficacy, particularly for lower-ranked students, unless carefully framed to support personal progress and collaboration (Santhanam et al., 2016).”

• cites Santhanam et al., 2016 sourceNot checked — no full text of the source

May combine two claims.

Gamified design elements can be aligned with Self-Determination Theory psychological needs: points and levels may enhance competence, avatars and team leaderboards may strengthen relatedness, and task or difficulty choice may support autonomy.

“Gamified design elements can be aligned with these psychological needs, as points and levels may enhance perceptions of competence through progress feedback; avatars and collaborative features such as team leaderboards can strengthen relatedness; and opportunities to select tasks or difficulty levels may support autonomy (Hakulinen & Auvinen, 2014;Hanus & Fox, 2015;Jones et al., 2023).”

• cites Hakulinen & Auvinen, 2014 sourceNot checked — no full text of the source
• cites Hanus & Fox, 2015 sourceNot checked — no full text of the source
• cites Jones et al., 2023 sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

Flow Theory suggests students achieve optimal engagement when challenges match skills.

“Finally, Flow Theory (Csikszentmihalyi, 1990) suggests students achieve optimal engagement when challenges match skills.”

• cites Csikszentmihalyi, 1990Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S3 (load-bearing).

The most accepted framework for engagement is the multidimensional model including behavioral, emotional, and cognitive components.

“The most accepted framework for engagement is the multidimensional model (Fredricks et al., 2004), which includes three interrelated components: behavioral engagement (participation and effort), emotional engagement (feelings like interest or frustration), and cognitive engagement (mental investment and learning strategies).”

• cites Fredricks et al., 2004 sourceNot checked — no full text of the source

Classified as probably not a substantial claim.

For inclusion in quantitative analysis, studies had to compare a gamified group with a control group receiving traditional non-gamified instruction.

“Additionally, for inclusion in the quantitative analysis, studies had to compare a gamified group with a control group receiving traditional non-gamified instruction (e.g., standard exercises).”

May combine two claims.

Tests for publication bias accounted for dependency among effect sizes, using a funnel plot, an Egger-type meta-regression with RVE, and PET-PEESE estimators with RVE.

“Tests for publication bias accounted for dependency among effect sizes.A funnel plot was generated using one aggregated effect per study, followed by an Egger-type meta-regression with RVE.PET-PEESE estimators (with RVE) were also used to obtain bias-adjusted effects under scenarios of zero or non-zero true effects (Stanley & Doucouliagos, 2014).”

• cites Stanley & Doucouliagos, 2014 sourceNot checked — no full text of the source

Classified as probably not a substantial claim.

Studies were selected using predefined inclusion and exclusion criteria framed by the PICO model.

“Studies were selected using predefined inclusion and exclusion criteria, framed by the PICO model: Population-secondary and higher education students in mathematics; Intervention-gamification strategies; Comparison-traditional, non-gamified instruction; Outcomes-effects on motivation and engagement.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Influence diagnostics (Cook's distance, DFBETAS, hat values) were computed to evaluate whether individual effects disproportionately affected model estimates.

“Influence diagnostics (Cook's distance, DFBETAS, hat values) were computed to evaluate whether individual effects disproportionately affected model estimates (Viechtbauer & Cheung, 2010).”

• cites Viechtbauer & Cheung, 2010 sourceNot checked — no full text of the source

Classified as probably not a substantial claim.

Moderator analyses were limited to single-moderator models due to the small number of clusters (k = 11).

“Moderator analyses were limited to single-moderator models due to the small number of clusters (k = 11), as recommended for RVE applications (Harrer et al., 2021).”

• cites Harrer et al., 2021 sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

Between and within-study heterogeneity were quantified using variance decomposition, Cochran's Q, and the I2 statistic.

“Between and within-study heterogeneity were quantified using variance decomposition, Cochran's Q, and the I 2 statistic (Higgins & Thompson, 2002).”

• cites Higgins & Thompson, 2002 sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Sensitivity analyses were performed using alternative within-study correlations (rho = .30, .90) and three alternative model specifications: a random-effects model with a single aggregated effect per study, a model restricted to peer-reviewed journal articles only, and a quality-based sensitivity analysis excluding studies scoring 60% or less on the MMAT.

“The working covariance matrix assumed ρ =.60, with sensitivity analyses conducted using ρ =.30 and ρ =.90 to assess robustness.Additional sensitivity analyses were performed using three alternative model specifications: a random-effects model with a single aggregated effect per study; a model restricted to peer-reviewed journal articles only; a quality-based sensitivity analysis excluding studies”

Classified as probably not a substantial claim.

Searches were conducted in ten databases: Scopus, ERIC, IEEE Xplore, MathSciNet, PsycINFO, Web of Science Core Collection, Zentralblatt MATH, ArXiv, Psychology and Behavioral Sciences Collection, and JSTOR.

“Searches were conducted in Scopus, ERIC, IEEE Xplore, MathSciNet, PsycINFO, Web of Science Core Collection, Zentralblatt MATH, ArXiv, Psychology and Behavioral Sciences Collection, and JSTOR, ensuring broad coverage of gamification research in mathematics education.”

Classified as probably not a substantial claim. May combine two claims.

The studies were independently assessed in multiple stages by two authors, with discrepancies resolved through consensus.

“The studies were independently assessed in multiple stages by two authors, with discrepancies resolved through consensus.”

May combine two claims.

The meta-analysis was well powered to detect moderate-to-large effects but underpowered for small effects.

“Thus, the meta-analysis was well powered to detect moderate-to-large effects but underpowered for small effects, a common limitation in meta-analyses with relatively few studies (Valentine et al., 2010).”

• cites Valentine et al., 2010 sourceNot checked — no full text of the source

May combine two claims.

An a priori power analysis based on 11 included studies and average group sizes (n1=44; n2=35) estimated power of 61.82% for small effects (d=0.20), 99.43% for moderate effects (d=0.50), and 100% for large effects (d=0.80).

“An a priori power analysis was conducted to estimate the statistical power of the meta-analysis (Harrer et al., 2021), based on the 11 included studies and the average group sizes (n₁ = 44; n₂ = 35).Assuming a random-effects model with moderate heterogeneity, power was 61.82% for small effects (d = 0.20), 99.43% for moderate effects (d = 0.50), and 100% for large effects (d = 0.80).”

• cites Harrer et al., 2021 sourceNot checked — no full text of the source

May combine two claims.

Because effect sizes within studies are statistically dependent, a three-level Correlated and Hierarchical Effects model with Robust Variance Estimation was used, estimating random effects at study and within-study levels and providing small-sample-adjusted standard errors and confidence intervals.

“Because effect sizes within studies are statistically dependent, a three-level Correlated and Hierarchical Effects (CHE) model with Robust Variance Estimation (RVE) was used.This approach estimates random effects at both the study and within-study levels and provides small-sample-adjusted standard errors and confidence intervals (Harrer et al., 2021;Pustejovsky & Tipton, 2022).”

• cites Harrer et al., 2021 sourceNot checked — no full text of the source
• cites Pustejovsky & Tipton, 2022 sourceNot checked — no full text of the source

May combine two claims.

Egger's regression with RVE found a non-significant intercept (b = .767, SE = 1.98, t(4.71) = .39, p = .715) and a non-significant slope (b = -1.597, SE = 8.78, t(4.65) = -.18, p = .863).

“The regression intercept was non-significant (b =.767, SE = 1.98, t (4.71) =.39, p =.715), and the slope was also non-significant (b = -1.597,SE = 8.78, t (4.65) = -.18,p =.863).”

May combine two claims.

Gamification has meaningful potential in mathematics, where motivation is often difficult to sustain.

“This indicates that gamification has meaningful potential in mathematics, where motivation is often difficult to sustain.”

May combine two claims.

These results suggest no evidence of funnel plot asymmetry or small-study effects.

“These results suggest no evidence of funnel plot asymmetry or smallstudy effects.”

May combine two claims.

The funnel plot based on aggregated data per study appeared symmetric, although some dispersion was noted among studies.

“A funnel plot based on aggregated data per study appeared symmetric, although some dispersion was noted among studies (Fig. 5).”

May combine two claims.

PET-PEESE analysis yielded non-significant intercepts (p = .715 and p = .585), suggesting no clear evidence of small-study effects or publication bias.

“Both models yielded non-significant intercepts (p =.715 and p =.585, respectively), suggesting no clear evidence of small-study effects or publication bias.”

May combine two claims.

Because engagement measures could not be synthesized meta-analytically due to absence of validated instruments and lack of experimental-control comparisons, reported outcomes were organized into behavioral, emotional, and cognitive engagement following multidimensional frameworks.

“Because engagement measures could not be synthesized meta-analytically, given the absence of validated instruments and the lack of studies with experimental-control comparisons, the reported outcomes were organized into behavioral, emotional, and cognitive engagement, following multidimensional frameworks (Fredricks et al., 2004).”

• cites Fredricks et al., 2004 sourceNot checked — no full text of the source

May combine two claims.

Available evidence points that mastery-oriented structures, gradual progression through increasing levels of difficulty and the use of individual progress indicators may be more effective for sustaining motivation.

“Available evidence point that mastery-oriented structures, gradual progression through increasing levels of difficulty and the use of individual progress indicators may be more effective for sustaining motivation (see Table 4).”

May combine two claims.

Perceptions of competition, autonomy and failure can differ across cultural and demographic groups, implying that gamification may require cultural adaptation, gender-sensitive adjustments or consideration of sociodemographic factors to ensure equitable benefits.

“Perceptions of competition, autonomy and failure can differ across cultural and demographic groups, implying that gamification may require cultural adaptation, gender-sensitive adjustments or consideration of sociodemographic factors (e.g., socioeconomic conditions) to ensure equitable benefits (Asrifan et al., 2025).”

• cites Asrifan et al., 2025Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

May combine two claims.

Competitive features can be motivating, but if overemphasized, they risk discouraging less competitive students.

“Competitive features can be motivating, but if overemphasized, they risk discouraging less competitive students.”

May combine two claims.

Game elements such as points, badges and leaderboards are unlikely to be effective when added superficially; their effectiveness depends on alignment with students' psychological needs.

“Game elements such as points, badges and leaderboards are unlikely to be effective when added superficially; rather, evidence support that their effectiveness depends on alignment with students' psychological needs.”

May combine two claims.

This review suggests that gamification can enhance students' motivation in mathematics, although its effectiveness appears to depend on design features.

“This review suggests that gamification can enhance students' motivation in mathematics, although its effectiveness appears to depend on design features.”

May combine two claims.

The findings support the central role of SDT while showing the value of integrating complementary perspectives such as the ARCS model, AGT, SCT and Flow Theory.

“The findings support the central role of SDT while showing the value of integrating complementary perspectives such as the ARCS model, AGT, SCT and Flow Theory, which capture motivational processes not fully addressed by SDT alone.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Score/points can be experienced as challenging/competence by learners with high initial intrinsic motivation, possibly supporting perceived competence.

“| Score/Points | High initial intrinsic | Learners with high extrin- | Celis et al. (2023); |”

• cites Celis et al. (2023) sourceNot checked — no full text of the source
• cites Lozano et al. (2024) sourceNot checked — no full text of the source

May combine two claims.

Most studies were conducted in Western higher-education settings, so it is unclear whether these findings generalize to more culturally diverse classrooms.

“Because most studies were conducted in Western higher-education settings, it is unclear whether these findings generalize to more culturally diverse classrooms.”

May combine two claims.

Reviews such as Zeybek and Saygı (2023) report similar motivational benefits and similar risks of neutral or negative effects across disciplines.

“Reviews such as Zeybek and Saygı (2023) report similar motivational benefits and similar risks of neutral or negative effects across disciplines.”

• cites Zeybek and Saygı (2023) sourceNot checked — no full text of the source

May combine two claims.

The meta-analysis was based on a relatively small set of 11 studies, several of which used heterogeneous or non-validated instruments, limiting the stability of effect-size estimates.

“The meta-analysis was based on a relatively small set of 11 studies, several of which used heterogeneous or non-validated instruments, limiting the stability of effect-size estimates.”

May combine two claims.

The limited exploration of engagement and the overall quality of existing studies suggest that, while gamification appears to be a promising approach for enhancing motivation and engagement in mathematics education, the field remains at a developmental stage.

“Taken together, the limited exploration of engagement and the overall quality of existing studies suggest that, while gamification appears to be a promising approach for enhancing motivation and engagement in mathematics education, the field remains at a developmental stage.”

• AssumptionRests on an assumption: The effect sizes from the 11 studies included in the meta-analysis measure the same underlying construct of motivation and are comparable enough to be pooled.
• AssumptionRests on an assumption: The quasi-experimental designs of most included studies provide valid estimates of gamification's causal effect on motivation and engagement.

May combine two claims.

The absence of validated engagement measures constitutes a substantive theoretical gap and limits the ability to examine how motivation and engagement operate together in gamified learning.

“Similarly, the absence of validated engagement measures constitutes a substantive theoretical gap and limits the ability to examine how motivation and engagement operate together in gamified learning.”

May combine two claims.

The small number of eligible studies restricted the moderator analyses to simple, single-moderator models, and these results should be interpreted cautiously due to limited statistical power.

“The small number of eligible studies also restricted the moderator analyses: only simple, single-moderator models could be estimated, and these results should be interpreted cautiously due to limited statistical power.”

May combine two claims.

In general, engagement improved when gamification supported students' autonomy, competence, and enjoyment, whereas highly competitive or performance-oriented designs tended to reduce engagement.

“In general, engagement improved when gamification supported students' autonomy, competence, and enjoyment, whereas highly competitive or performance-oriented designs tended to reduce engagement.”

May combine two claims.

These patterns are consistent with SDT: controlling or comparative elements may undermine autonomy and competence, whereas designs that provide immediate feedback, scaffolding and opportunities for self-paced progression tend to support basic psychological needs.

“These patterns are consistent with SDT: controlling or comparative elements may undermine autonomy and competence, whereas designs that provide immediate feedback, scaffolding and opportunities for selfpaced progression tend to support basic psychological needs (Deci & Ryan, 2000;Reeve, 2012;Ryan & Deci, 2017).”

• cites Deci & Ryan, 2000 sourceNot checked — no full text of the source
• cites Reeve, 2012 sourceNot checked — no full text of the source
• cites Ryan & Deci, 2017 sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

Score/points can be demotivating to learners with high extrinsic motivation and high amotivation, as they could be perceived as controlling.

“| | motivation (may | sic motivation, high amoti- | Lozano et al. (2024) |”

• cites Celis et al. (2023) sourceNot checked — no full text of the source
• cites Lozano et al. (2024) sourceNot checked — no full text of the source