Mathematical skills play an essential role in our daily lives, and they a provide solid foundation for numerous disciplines. Historical records show that humans have been experimenting with mathematics for over 4000 years (Huang et al., 2014; Ku et al., 2014).

Despite the importance of mathematics, the majority of students in K-12 education view mathematics unfavorably and recognize it as a frustrating and difficult subject that causes learning fatigue, pressure, and anxiety (Deng et al., 2020; Geist, 2010; Offer & Bos, 2009; Luhan et al., 2013; Sun et al., 2021). Therefore, mathematics is the subject with the highest student failure rate (Huang et al., 2014).

A number of factors have been attributed to students’ mathematics problems, among these factors, the traditional method of teaching has been receiving growing criticism. For example, Paul (1992) argued that the traditional method promotes memorization and encourages students to keep practicing what they already know. Additionally, Skinner and Belmont (1993) stated that this method is incapable of motivating students to learn and engage in the learning process. Further, students’ exposure to complex problems is very limited, hence, this method does not advance the development of students’ problem-solving skills, conceptual understanding, or critical analysis (Bikić et al., 2016; Kinard & Kozulin, 2008; Voskoglou & Salem, 2020).

Due to substantial technological developments during the past decade, a growing number of researchers and educators are incorporating technology in education (Hussein, Ow, Cheong, & Thong, 2019; Hwa, 2018). Among the various technology-supported learning methods, Digital Game-Based Learning (DGBL) is perceived as an effective way to learn mathematics. For instance, Ke (2008) and Tsai et al. (2012) indicated that the use of DGBL applications in mathematics promotes students’ perseverance and improves their engagement in the learning process. Martinovic et al. (2014) showed that gaming could contribute to young learners’ cognitive development. However, despite DGBL’s potential, prior research has not supported the notion that its use in mathematics education guarantees positive outcomes. For example, Ke (2008) reported that there is no significant evidence to support the premise that DGBL improves students’ comprehension of mathematics or metacognitive skills. Similarly, Hung et al. (2014) articulated that DGBL applications are not always effective in managing students’ anxiety towards mathematics.

Given the prevalence of gaming in education and the inconclusive findings in the literature, a growing number of scholars have conducted meta-analyses and reviews pertaining to the effectiveness of the DGBL approach. Abdul Jabbar and Felicia (2015), Boyle et al. (2016), Connolly et al. (2012), Hainey et al. (2016), and Vogel et al. (2006) argued that gaming results in improved academic performance, motivation, and attitudes towards learning. However, there are some other studies that do not show that the literature consistently supports DGBL’s empirical effectiveness. For example, Young et al. (2012) stated there is limited evidence to support the effectiveness of gaming applications in the domain of mathematics. Additionally, in their meta-analysis, Wouters et al. (2013) stated no evidence suggests that computer games are better motivators in learning than the traditional teaching methods.

Although the earlier studies provide crucial insights into the effects of gaming in the context of learning; their scope was broad (e.g., covered several curricular subjects) and their findings were inconsistent. As a consequence, Divjak and Tomić (2011), Byun and Joung (2018), and Tokac et al. (2019) conducted domain-specific reviews to examine DGBL’s effectiveness in mathematics. Although their analyses showed that playing games contribute positively to students’ knowledge comprehension and motivation, there are issues associated with the timelines of the studies or the type of publications reviewed. For example, Divjak and Tomić (2011) surveyed papers published between 1995 and 2010, while Byun and Joung (2018) examined papers published between 2000 and 2014. Therefore, there is a need to update their findings and demonstrate how this area of research has evolved after 2014. Further, out of the 24 research studies reviewed in Tokac et al., 2019) meta-analysis, nine publications were gray literature (including eight dissertations, and one conference paper). While the inclusion of such publications may significantly improve the comprehensiveness of findings and reduce publication bias (Paez, 2017), it is often a challenge to evaluate the rigor of the reviewing process, due to the lack of information on how this process has been performed with these publications (Garousi & Rainer, 2020; Noroozi et al., 2020). Furthermore, recent DGBL meta-analyses in K-12 mathematics education lack an in-depth analysis of individual papers. This causes different critical aspects of the learning process and different angles of students’ learning outcomes to remain unclear. In addition, McLaren et al. (2017) stated that there is limited empirical evidence concerning the DGBL effectiveness, especially in the context of mathematics education. Therefore, McLaren et al. (2017) called for more studies to investigate whether the excitement behind using this method of instruction in mathematics education is justified.

As a consequence, the aims of this study are to extend the findings of previous reviews and provide an updated synthesis of evidence regarding the effectiveness of DGBL applications in the domain of K-12 mathematics education. This is achieved by utilizing a multi-dimensional framework to gain an in-depth understanding of the current status of DGBL research in K-12 mathematics education. To accomplish the above-mentioned objectives, this review poses the following research questions:

  1. 1.

    Do students learn mathematics more effectively when it is presented via DGBL applications, compared to other teaching methods?

  2. 2.

    How do students learn mathematics more effectively via DGBL applications?

The present study has two key contributions: first, providing an evidence-based discussion with regard to the effectiveness of DGBL in K-12 mathematics, synthesized from rigorously-reviewed academic journals, and second, providing detailed insights into the current trends in K-12 mathematics education. Therefore, it is believed that this review will provide useful information to aid researchers, educators, and game designers who are interested in using DGBL in K-12 mathematics education. This paper could also assist researchers and instructors from other knowledge disciplines in obtaining additional evidence pertaining to the impact of DGBL as a teaching method.

1 DGBL definitions

DGBL refers to a student-centered approach where educational objectives and material are embedded in gaming activities in an attempt to motivate students to learn and improve their skills and knowledge by providing them with an enjoyable and interactive learning environment (Prensky, 2001; Qian & Clark, 2016; Sung & Hwang, 2013).

However, some studies, particularly those employing a value-added design, have addressed at least two types of DGBL applications: value-added application and a base or simplified application. The former refers to a DGBL application enriched by a specific design or learning feature (Mayer, 2019), and the latter is similar to the value-added application with the sole exception that it does not have this specific design feature (Mayer, 2019). The significance of value-added comparisons lies in attempting to understand and investigate how a specific game design feature could foster students’ understanding of a certain mathematical concept. This, in turn, could lead to an improved game design that might enhance the learning process.

2 Methodology

This study utilizes the classification method proposed in Connolly et al. (2012), which is useful for identifying similarities, differences, and limitations in DGBL studies. This method classifies outcomes into four categories:

  • Knowledge acquisition

  • Perceptual and cognitive skills

  • Affective and motivational outcomes

  • Behavior change outcomes

According to Connolly et al. (2009) and Hainey et al. (2014), the knowledge acquisition category is primarily about gains in factual knowledge and improvements in students’ performance as a result of the DGBL intervention. The perceptual and cognitive category encompasses the learners’ perceptions such as their flow experience or cognitive load; this assesses the effects of DGBL interventions on the students’ cognitive competencies (e.g., problem-solving). The affective and motivational category concerns a number of aspects such as the learners’ particular motivations for using the intervention as well as their level of interest in participation. Finally, the behavioral change category, as the name implies, relates to the effects of the DGBL interventions on the occurrence of change in the learners’ behaviors and attitudes (All et al., 2016; Stewart et al., 2012).

3 Database

The Social Sciences Citation Index (SSCI) database of Web of Science was used to search for DGBL interventions in mathematics education in the Educational Research category. This online repository contains rigorously reviewed research and high impact studies (Zydney & Warner, 2016). In addition, to ensure wider coverage of high-quality journals, the researchers employed the Google Scholar metrics in a manner similar to that used in Nikou’s and Economides’ (Nikou & Economides, 2018) review to identify educational technology journals with the highest impact factors. Among the 20 publications in this subcategory, the following 13 journals were considered:

  • Computers and Education

  • British Journal of Educational Technology

  • Educational Technology and Society

  • Journal of Computer Assisted Learning

  • Education and Information Technologies

  • Educational Technology Research and Development

  • Interactive Learning Environments

  • Tech Trends

  • Turkish Online Journal of Educational Technology

  • Learning at Scale

  • Learning, Media and Technology

  • IEEE Transactions on Learning Technologies

  • Australasian Journal of Educational Technology

Seven publications were not considered because their scope did not align with the scope of this study, as these publications either did not target students in K-12 education or did not include DGBL interventions in the subject of mathematics.

4 Search terms

The Boolean operator “OR” was utilized to combine all the keywords related to DGBL (i.e., serious game*”, “game-based learning”, “educational game*”, “computer game*”, “online game*”, “digital game-based learning”, “Digital game*”, “gaming”, “MMORPG”, “augmented reality”, “video games”, “video gaming”, “electronic games”). In a similar way, the Boolean operation “OR” was employed to combine all the keywords related to mathematics learning (i.e., “math* learning”, “learning math*”, “math* teaching”, “teaching math*”, “math* education”, “math* instruction”, “math* evaluation”, “math* outcome”, “math* skills”). Finally, the Boolean operator “AND” was used to combine DGBL and mathematics learning keywords.

5 Inclusion and exclusion criteria

To ensure the retrieved studies are relevant to the scope of the review, the researchers applied the following inclusion and exclusion criteria:

  • Published in English, in a peer-reviewed journal, and between 2008 and 2019 (as DGBL research in mathematics education experienced a notable upsurge in 2008 (Byun & Joung, 2018)).

  • Each intervention is related to mathematics learning and includes participants from K-12 education.

  • Each DGBL application is clearly designated as a game and the term game must be included in the title or abstract of the publication†.Footnote 1

  • Eligible studies must incorporate at least one comparison of a DGBL application to a simplified game design, other e-learning tools, or the traditional method of teaching†.

  • Upon completion, students received feedback (e.g., score points and/or a progress report)†.

The researchers applied the following exclusion criteria:

  • Non-English publications.

  • Interventions published in conference papers, book chapters, or PhD dissertations.

  • Findings of qualitative studies.

  • Findings of single-group interventions.

6 Article selection

The search process resulted in the retrieval of 1347 research articles, after excluding duplicates, inaccessible publications, and studies unrelated to DGBL in K-12 mathematics education. Two researchers conducted two in-depth screening rounds to finalize the search results according to the inclusion and exclusion criteria mentioned above.

During the first screening process, the two researchers independently rated the papers, they had inter-rater reliability of 88.5%, which was brought to 100% agreement after discussion. This process resulted in the inclusion of 70 research articles. During the second screening process, the same two researchers read the full text of the 70 articles to ensure that these studies truly satisfied the inclusion and exclusion criteria. The inter-rater reliability was 90.7% and then was brought to 100% after addressing all the inconsistencies and disagreements through discussion. Finally, a total of 43 research articles were included in the present review. Figure 1 details the various stages that this study followed during the data collection and evaluation processes.

Fig. 1
figure 1

Articles selection process

Full size image

7 Results

7.1 Analysis of game and study variables

This section highlights the findings of the analysis of variables related to the studies included in this review with a specific focus on the educational level of participants, the year of publication, and the genre of the DGBL application.

7.1.1 Participants

Figure 2 identifies the educational levels of students who participated in DGBL interventions in the domain of K-12 mathematics education. Twenty-eight of the 30 studies were focused on students at the primary level, while two studies had participants from primary and junior high schools. Eight studies were conducted at the high school level and seven were conducted at the junior high school level, while one study included participants from both junior and senior high school. Furthermore, four studies were performed at pre-vocational levels and one study was carried out at the vocational level. It should be noted that primary education refers to students who are in grades 1–6, junior high school refers to students in grades 7–9, and senior high school refers to students in grades 10–12. Prevocational and vocational learning refer to students who are in a less advanced level of high school education that specifically prepares them for vocational learning (ter Vrugte, de Jong, Vandercruysse, et al., 2015).

Fig. 2
figure 2

Distribution of articles in K-12 mathematics education, according to participants’ educational level.

Full size image

7.1.2 Number of studies with respect to publication year

Figure 3 shows the number of articles published between 2008 and 2019, presented by the year of publication. The study of DGBL in mathematics learning has experienced two phases. In the first phase, from 2008 to 2013, the number of published articles was relatively limited. This indicates that the use of DGBL in K-12 mathematics was still in its early stages. In the second phase, from 2014 until the present time, the number of papers rapidly increased, showing a growing interest among researchers in utilizing DGBL applications in K-12 mathematics.

Fig. 3
figure 3

Distribution of articles in K-12 mathematics education, according to year of publication.

Full size image

7.1.3 DGBL genres

Game genres are constantly evolving and changing (Lee et al., 2006), hence one DGBL application may belong to more than one genre (Ke, 2016), this study categorized the game genres based on the classifications suggested by Bontchev and Vassileva (2010), Carmigniani et al. (2011), Herz (1997), Ke (2016) and Minkkinen (2016). A brief description of these genres is provided in Table 1.

Table 1 Definitions of gaming genres
Full size table

In this review, nine genres were identified, as shown in Fig. 4. These genres can be categorized into two distinct groups. The first group consists of frequently employed genres such as simulations (n = 13), puzzle and adventure (n = 8), role-playing (n = 7), and strategy (n = 4), whereas the second group involves board game, virtual reality, platform, and construction only once. There was also one study that did not present clear details regarding the DGBL application they used, hence, it was difficult to recognize its genre.

Fig. 4
figure 4

Distribution of articles in K-12 mathematics education, according to genre of the DGBL application

Full size image

7.2 Analysis of methodologies and learning outcomes

This section reviews the articles with regard to the following considerations:

  • Research design is primarily concerned with the type of methodology being used (e.g., Randomized Controlled Trial (RCT) and quasi-experimental).

  • Effects of the intervention on learning outcomes (e.g., positive, neutral, mixed, or negative).

  • The comparison approach refers to the educational activities that students in the control group received during the intervention (e.g., traditional method, value-added, or other computer-enhanced learning methods).

  • Areas of interest were concerned with which area of mathematics education a particular DGBL application is targeting (e.g., arithmetic operations, algebra, etc.).

  • Learning dynamics were examined to explore which learning dynamic (e.g., individual, collaborative/cooperative, or competitive) is more effective in promoting students’ mathematical knowledge in K-12.

  • Design characteristics aimed at highlighting the design features and learning theories that guided the development of some DGBL applications.

After reviewing the articles that met the inclusion criteria, it was revealed that the studies that investigated the effects of DGBL applications on students’ affective and motivational outcomes and behavioral change were very limited. Therefore, these two categories were combined together in one category known as affective, motivational, and behavioral change. It should be noted that similar approaches have been implemented by other reviews that have used the multi-dimensional framework to classify the learning outcomes of DGBL applications in science education (Hussein, Ow, Cheong, Thong, & Ebrahim, 2019) and digital storyline-enhanced learning (Novak, 2015).

7.2.1 Knowledge acquisition

It was revealed that 27 papers have investigated the effects of DGBL applications on students’ construction of mathematical knowledge and concepts. Eight studies implemented an RCT design; among them, six papers reported positive outcomes (Barzilai & Blau, 2014; Denham, 2015; Kebritchi et al., 2010; Kim & Ke, 2017; Lin et al., 2013; ter Vrugte et al., 2017). Wouters et al. (2017) reported mixed findings from two experiments. The first demonstrated neutral outcomes, while the second showed positive outcomes. There was only one collaborative study that reported neutral outcomes (Gresalfi et al., 2018). Table 2 lists the studies that followed the RCT research design along the dimensions of the platform, comparison approaches, areas of interest, and learning dynamics.

Table 2 Summary of RCT studies investigating the DGBL effects on mathematics knowledge acquisition
Full size table

In this category, 19 studies were included because they followed a quasi-experimental research design. Of these 19 studies, 16 papers revealed positive learning gains (Barros et al., 2019; Beserra et al., 2014, 2017; Huang et al., 2014; Hung et al., 2015; Hwa, 2018; Ke, 2008a; Kolovou et al., 2013; Masek et al., 2017; Pareto et al., 2012; Shi et al., 2019; ter Vrugte, de Jong, Vandercruysse, et al., 2015a; ter Vrugte, de Jong, Wouters, et al., 2015; Vandercruysse et al., 2017; Wang et al., 2018 and Yang et al., 2018). The results of the remaining studies were mixed. Two studies Ke (2008b) and Panoutsopoulos and Sampson (2012) reported neutral outcomes, and one study, Fokides (2018), reported mixed outcomes. The researcher in Fokides (2018) provided students with 15 learning exercises, and students in the experimental group who used the DGBL application significantly outperformed their counterparts in control group (A) who received instruction via the traditional method in all the 15 exercises. However, students in the experimental group achieved significantly better learning gains than the students in control group (B) who received instruction via contemporary teaching methods based on collaborative learning of the textbook as well as worksheets designed by the teachers in only four out of the 15 learning exercises.

The majority (n = 14) of studies were single-player and non-collaborative, with regard to different learning dynamics (e.g., individual, collaborative/cooperative, or competitive). Ke (2008, 2008a) looked at the same learning dynamics and reported interesting findings; Ke (2008) revealed that students from economically disadvantaged backgrounds who learned collaboratively and competitively made greater gains than their counterparts in the individualized group. Conversely, students from better socio-economic backgrounds gained significantly more from individualized learning than other learning dynamics. In addition, with respect to the mathematics learning attitude, Ke (2008) reported that cooperative learning significantly improved students’ learning attitude towards mathematics more than the individual and competitive learning dynamics. Ke (2008a) found that students who learned individually significantly outperformed their counterparts who learned cooperatively and competitively. As for learning attitude, students learning cooperatively demonstrated a better attitude than those who learned competitively, but not than their counterparts who learned individually.

In addition, Fokides’ (2018) study found that collaborative learning via a DGBL application is significantly more effective than individual learning administered via the traditional method. Further details regarding the individual characteristics of quasi-experimental studies that focused on the students’ knowledge acquisition category are demonstrated in Table 3.

Table 3 Summary of quasi-experimental studies investigating the DGBL effects on mathematics knowledge acquisition
Full size table

7.2.2 Perceptual and cognitive skills

In the perceptual and cognitive skills category, nine studies were included. Among them, four studies (Castellar et al. (2015), Hulse et al. (2019), Ke (2019), and Lee and Ke (2019) have followed the RCT research design and reported positive outcomes. Table 4 highlights the key features of RCT studies that targeted the students’ perceptual and cognitive skills.

Table 4 Summary of RCT studies investigating the DGBL effects on perceptual and cognitive skills*
Full size table

Five studies followed a quasi-experimental research design, among which three studies (i.e, Brezovszky et al. (2019), Kiili et al. (2018), and Lee et al. (2014)) revealed positive outcomes. Eseryel et al. (2011) reported mixed findings from two learning interventions. Kiili and Ketamo (2017) is the only study that reported negative outcomes, whereby the effects of a DGBL application was compared to the traditional method. Table 5 details the key components of the quasi-experimental studies that emphasized promoting the students’ perceptual and cognitive skills.

Table 5 Summary of quasi-experimental studies investigating the DGBL effects on perceptual and cognitive skills*
Full size table

7.2.3 Affective, motivational, and behavioral change

This category included seven studies with only two articles following an RCT research design. One of these studies Habgood and Ainsworth (2011) reported positive learning outcomes, while the other one Kim et al. (2017) reported neutral outcomes.

The remaining five studies Chang et al. (2016), Chen et al. (2012), Garneli et al. (2017), Ku et al. (2014), and Mavridis et al. (2017) have followed a quasi-experimental design, and all of them obtained positive results. Due to the limited number of studies investigating affective, motivational, and behavioral change, both RCT and quasi-experimental studies are combined in one table. Thus, Table 6 summarizes the main aspects of this category and demonstrates the research design they followed.

Table 6 Summary of studies investigating the DGBL effects on affective, motivational, and behavioral change*
Full size table

7.3 Analysis of design features and learning theories

Among the 43 studies included in this review, only 20 studies detailed how the DGBL applications they used were designed. These 20 papers can be divided into two groups. The first group comprises 17 studies that focused on the effects of specific design features, and the second group consists of three studies that examined the impact of general learning theories.

In the context of K-12 mathematics, a number of design features were employed, which can be further classified into three groups based on the functions they perform and services they provide, namely, learning content representation, scaffolds, and question representation features. First, seven studies focused on how mathematical learning content is presented to students within the game world. For example, Denham (2015), Habgood and Ainsworth (2011), Ku et al. (2014), and Vandercruysse et al. (2017) examined intrinsic and extrinsic features. Chen et al. (2012) incorporated game quests, Garneli et al. (2017) embedded storyline narratives, and Wouters et al. (2017) investigated surprising events. Table 7 briefly defines these mechanisms and highlights their effectiveness.

Table 7 Summary of studies investigating the learning content representation features
Full size table

Similarly, seven studies focused on scaffolding features, which are assistive mechanisms that aim to help students when they face difficulties or challenges during gameplay by providing prompts, clues, and hints (Barzilai & Blau, 2014). For example, Barzilai and Blau (2014) employed external conceptual scaffolds, Eseryel et al. (2011) embedded dynamic modeling in the DGBL application they used, Huang et al. (2014) utilized a diagnostic mechanism, and Lee et al. (2014) investigated the effectiveness of four types of scaffolds in two phases: first, single or multiple-solution, second, specific or generic prompts. Collaboration and competition facilities were examined by ter Vrugte, de Jong, Vandercruysse, et al. (2015), while ter Vrugte et al. (2017) addressed the faded worked examples. Finally, Yang et al. (2018) focused on progressive prompting. Table 8 briefly defines these mechanisms and highlights their effectiveness.

Table 8 Summary of studies investigating scaffolding features
Full size table

Finally, three studies have examined the question representation features. Such features are primarily concerned with how questions within the game world will be formatted and delivered to students. For example, Beserra et al. (2017) assessed fine-grained multiple-choice response format, Lee and Ke (2019) examined the effectiveness of iconic and symbolic prompting, and ter Vrugte et al. (2015a) focused on reflection prompts, procedural information, and reflection prompts plus procedural information. Table 9 briefly explains these mechanisms and highlights their effectiveness.

Table 9 Summary of studies investigating the question representation features
Full size table

With regard to learning theories, there were three studies focusing on the impact of general learning theories in their DGBL interventions. For example, Kebritchi et al. (2010) addressed the experiential learning theory, Lin et al. (2013) employed the remedial mastery learning, and Pareto et al. (2012) used the master-apprentice theory. Table 10 briefly explains these learning theories and highlights their effectiveness.

Table 10 Summary of studies investigating specific learning theories
Full size table

8 Discussion

This study aimed to examine and provide insights into the DGBL effects on K-12 mathematics education. To this end, the Web of Science database and top-ranking educational technology journals indexed by Google Scholar were selected for review and analysis. This paper reviewed empirical research published between 2008 and 2019. The keywords used in this study produced 1347 research articles, which shows that DGBL in mathematics education is frequently addressed in high impact journals. The inclusion and exclusion criteria followed by this study resulted finally in the inclusion of 43 studies. The multi-dimensional framework proposed by Connolly et al. (2012) was used to identify and summarize the key trends of research in this domain.

The distribution of papers based on their publication year revealed that research in this area went through two periods. The first was from 2008 to 2013, when research into the DGBL effects on K-12 mathematics was receiving limited interest from the academic community. From 2014 onwards, it was noted that scholars’ interest grew considerably and steadily, as more studies were found focusing on investigating the effectiveness of DGBL applications in K-12 mathematics education.

The findings showed that different researchers have used a number of game genres with a notable interest in simulation games. The frequent utilization of simulations is justified due to their ability to provide students with rich and compelling narratives (Abdul Jabbar & Felicia, 2015). Another justifiable explanation is that simulation games have been found beneficial in promoting students’ mathematics learning experience (Wang et al., 2018) and enhancing their mathematical knowledge and problem-solving skills (Ke, 2019).

Twenty-nine studies followed a quasi-experimental research design, and only 14 implemented the RCT methodology. The quasi-experimental method was the most utilized research design in all three categories. The limited number of RCTs could be attributed to the challenges associated with finding an equally engaging and motivating learning activity for students in the comparison group (Gauthier & Jenkinson, 2016). This observation is in line with reports of previously conducted reviews that have recognized the quasi-experimental design as the most employed form of research in DGBL (Boyle et al., 2016; Connolly et al., 2012; Hainey et al., 2016; Hussein, Ow, Cheong, Thong, & Ebrahim, 2019).

With regard to the areas of interest, the results showed that DGBL was used to teach a wide range of mathematics-related topics, with a specific interest in arithmetic operations (14 studies). A possible explanation for this emphasis is that arithmetic operations are essential for advanced mathematical achievements (Geary, 2011; van der Ven et al., 2017). Another justification for this trend could be that students’ interactions with DGBL may help them develop a better understanding of arithmetic operations (Denham, 2015).

Twenty-eight DGBL studies were delivered via computers, nine were web-based, and hand-held devices were used only in eight studies. One possible reason for the extensive use of computers is that the majority of schools and educational institutes are already provided with computer laboratories. Therefore, researchers and educators often utilize computers to deliver learning content to students.

With regard to participants’ age, the majority (n = 30) of studies were devoted to primary education, with 28 studies conducted on the primary education level and two studies on students from primary and junior high school grades. One possible reason for this trend is that games are inherently joyful and pleasing, particularly for young learners; therefore, combining mathematics and games could successfully reduce students’ fear of learning this subject (Chang et al., 2012; Huang et al., 2014; Mayer et al., 2002). According to Khan et al. (2017), high school teachers have acknowledged the potential benefits of using DGBL in improving students’ learning and engagement. However, Proctor and Marks (2013) observed that the utilization of DGBL application at the high school level was limited compared to its application the primary education. The lack of DGBL studies at the high school level could be attributed to barriers and hindrances that teachers experience when they attempt to adopt DGBL in the classroom. Such barriers include a mismatch between the knowledge and skills embedded in the game and those explicitly identified by the curriculum, lack of ICT skills, and negative attitude towards using games in a classroom environment (Papadakis, 2018; Sánchez-Mena & Martí-Parreño, 2017). Further, Romero and Barma (2015) stated that educators in primary education enjoy greater flexibility in their teaching activities, hence, they are more willing to integrate serious games in their classrooms than educators at the high school levels. Furthermore, one could argue that researchers prefer to focus on students in primary education as learners in this age group are easier to please and less demanding than students in high school grades who might require sophisticated and advanced DGBL applications.

Only 20 out of 43 studies provided background information concerning the design features and theories behind the design of the DGBL applications they utilized in their learning interventions. It could be argued that this is a small percentage, and a possible explanation for this is that earlier DGBL research has mainly focused on using DGBL applications in a proof-of-concept manner to examine whether this method of instruction could lead to better academic achievements (Young et al., 2012), and only recently researchers have called for investigating how specific design features and learning theories could influence students’ learning outcomes (Clark et al., 2016). Another plausible explanation is that most studies that have explored the effects of a specific design feature utilized a value-added comparison, and such comparisons require designing at least two DGBL applications, which is a time-consuming and financially demanding task. As previously mentioned, numerous design features were utilized by the studies reviewed in this article, which suggests that the developers of these DGBL applications are experimenting with these design features and learning theories to ascertain their impact on promoting students’ mathematical knowledge. However, there was some interest in intrinsic and extrinsic design features as they were employed three times. Although their utilization did not produce conclusive findings, the results suggested that students in two out of these three studies preferred DGBL applications with intrinsic design features. This observation is supported by a similar finding reported by Westera (2015) who argued that playing such games is intrinsically satisfying for students.

The main purpose of the first research question was to investigate whether students in K-12 mathematics education can learn mathematics more effectively via DGBL compared to other methods of instruction. Findings revealed that in the category of knowledge acquisition, 22 out of 27 studies reported positive outcomes. In perceptual and cognitive skills, seven out of nine studies obtained positive outcomes. While in the category of affective, motivational, and behavioral change, six out of seven studies reported positive outcomes. Taking all the results into consideration, we can conclude that there is a promising potential to use DGBL in K-12 mathematics education, especially at the primary level, and in the category of knowledge acquisition. These findings are in agreement with previous reviews that revealed that DGBL applications were mainly used to foster students’ knowledge acquisition (Boyle et al., 2016; Connolly et al., 2012; Hainey et al., 2016; Hussein, Ow, Cheong, Thong, & Ebrahim, 2019; Li & Tsai, 2013).

The second research question focused on how students learn mathematics more effectively via DGBL. The results demonstrated that there is a reason to believe that DGBL applications based on specific design features and learning theories are more effective than the base version of the same application. This finding is supported by a similar finding reported by Wouters and Van Oostendorp (2017) and Young et al. (2012). Results also suggested that embedding a specific design feature or a learning theory into a DGBL application does not guarantee improved performance, as other factors such as students’ prior knowledge, gender, and cognitive abilities could influence how they perceive and interact with games.

8.1 Future directions

It is essential to provide more evidence regarding the effectiveness of this relatively new method of teaching. Researchers and educators are advised to consider the following research directions:

  • Additional research is necessary to investigate how different learning dynamics (e.g., collaborative/cooperative, and competitive) could influence students’ learning of mathematics, as findings from these studies did not produce conclusive results.

  • Only a fraction of the studies included in this review was focused on cognitive skills such as problem-solving, while other twenty-first century skills have received no scholarly interest from the academic community. Thus, to address this concern, upcoming studies are encouraged to focus on utilizing DGBL applications to promote students’ creativity and critical thinking skills in the domain of K-12 mathematics education.

  • DGBL studies investigating the effects of a specific design feature could play a key role in improving the instructional effectiveness of DGBL. However, evidence suggests that using such features does not always improve learning. Therefore, to maximize their effectiveness, future studies are recommended to consider a number of factors such as students’ gender, cognitive development, and prior knowledge.

  • Although this review covered all K-12 educational levels, there was a notable lack of research investigating the academic value of DGBL on senior high school students. To provide additional empirical evidence pertaining to the effectiveness of this approach, future studies are recommended to explore the effects of DGBL on the mathematical learning of senior high school students.

8.2 Limitations

As with previous reviews and meta-analyses, the findings reported here were limited by the search terms, journals indexed by Web of Science, and Google Scholar metrics, and papers published between 2008 and 2019. The second limitation concerns the research design, as only RCT and quasi-experimental studies were considered in this paper; therefore, the outcomes reported by surveys and one group studies were excluded. The final limitation concerns the publication type. As this study focused only on peer-reviewed journals, the concluding remarks of proceedings, book chapters, or PhD dissertations were excluded.

8.3 Implications

This study is important from both theoretical and practical perspectives. Theoretically, the present article addressed the paucity of qualitative analysis focusing on DGBL in K-12 mathematics; therefore, the findings present a more complimentary understanding to researchers and educators who are interest in DGBL research, particularly, in the context of K-12 mathematics. Practically, schools could utilize the popularity of video games among K-12 students and the recent wide employment of technologies in education to develop training programs for teachers. These programs are required to familiarize teachers with using DGBL and the potential advantages of utilizing this method in mathematics education. In addition, the traditional method does not take the individual differences among students into account; consequently, a considerable number of students feel isolated and left out when learning via the traditional method. Conversely, when schools and teachers implement the DGBL approach, they could provide students with the opportunity to learn at their own pace and the freedom to reexamine the learning material multiple times at their own leisure.

9 Conclusion

The current paper reviewed the literature regarding the implementation of DGBL in K-12 mathematics education. The initial search process resulted in 1347 research articles, suggesting that significant research attention has been given to DGBL in mathematics education. However, despite this surge in interest, there is still a paucity of research offering a thorough analysis of the current state of literature. To address this issue, the present study employed the multi-dimensional framework to systematically review, classify, and analyze the findings of 43 studies published between 2008 and 2019. These articles were sourced from the SSCI database of Web of Science and major educational technology journals according to Google Scholar metrics. In K-12 mathematics education, most studies have reported positive learning gains, and researchers have primarily focused on using DGBL applications in the knowledge acquisition category. Although there is cause for optimism, other areas exist that still require additional research to provide more evidence on the effectiveness of DGBL in K-12 mathematics education. For example, more research is needed to examine the impact of different learning dynamics on students’ mathematical achievements. Moreover, there is a notable lack of research addressing various twenty-first-century skills. Therefore, more research is required to ascertain if the DGBL approach is effective in promoting students’ twenty-first-century skills such as creativity and critical thinking. It is expected that this review will offer useful guidance to researchers, educators, and game developers in the area of K-12 mathematics education as well as the scholars working in other domains who are also interested in DGBL.