After reviewing 34 empirical gamification studies, Dicheva et al. (2015) suggested gamification elements fall into two categories: game design principles and game dynamics. Dicheva et al. (2015) proposed that game dynamics (the concrete elements players interact directly with, such as points, levels, and leaderboards) enhance and develop game design principles, larger structures and concepts, such as narrative, challenges, and competition (Dicheva et al., 2015). To illustrate, points and leaderboards (game mechanics) enhance and develop competition (a design principle).

Adopted from game theory, the MDA framework categorizes measurable game elements (i.e. points, levels, badges, etc.) as game mechanics, while game dynamics (e.g. reward, status, competition, etc.) refers to the results of players' interactions with game mechanics (Hamzah et al., 2015). The aesthetics category accounts for the emotional reaction (e.g. satisfaction, pleasure, belonging) experienced by the player during gameplay (Hamzah et al., 2015). In the classroom, awarding points (game mechanics) for completing a graphic organizer to teams of students (game dynamics) produces feelings of belonging and satisfaction (game aesthetics).

Similar to MDA model, Cheong et al. (2014) proposed a framework in which game elements are categorized by their levels of abstraction, from concrete components (i.e. badges or points) to game process mechanics (i.e. challenge or competition) to abstract dynamics (i.e. emotions or social relationships). In this model, components serve to develop mechanics and dynamics (Cheong et al., 2014). In terms of motivation and engagement, the components and mechanics provide the player constant feedback regarding their progress, which sustains engagement (Cheong et al., 2014).

Whereas the three frameworks discussed thus far are concerned with the function and categorization of game element, Lieberoth (2015) categorized and tested the motivational effectiveness of three levels of game element integration. Lieberoth (2015) made a distinction between game form or "frame" (i.e. game board, moving game pieces, taking turns, rolling dice) and game mechanics (i.e. game goals, competition, winning criteria). The three integration levels depend on the degree to which tasks use game frame and game mechanics: full games, such as live action or videogame simulations are high in game frame and game mechanics; deep gamification is low in game frame but high in game mechanics; and shallow gamification is low in mechanics but high in game frame (Lieberoth, 2015). Lieberoth (2015) found that shallow gamification or "framification" (high frame, low mechanics) produces almost equivalent engagement levels as deep gamification (high frame, high mechanics). The implication here is that educators may save time and effort by simply framing tasks as games with simple mechanics, rather than producing elaborate gamification materials and processes (Lieberoth, 2015).

Some studies measured motivation as an expression of the time learners spent on tasks (Abrams & Walsh, 2014; Landers & Landers, 2014), student participation (Chen et al., 2015), or the quantity of learning tasks completed (de-Marcos, Garcia-Lopez, & Garcia-Cabot, 2016; Landers & Landers, 2014; Mekler, Bruhlmann, Tuch & Opwis, 2015). In their investigation on virtual currency, Chen et al. (2015) reported that students who were awarded digital currency for participation inside and outside class time (i.e. joining student groups or engaging in online business simulations) were more engaged initially, but that engagement and participation fell dramatically over time. The falloff of interactions and engagement is consistent with de-Marcos, Garcia-Lopez and Garcia-Cabot (2016), who also examined the impact of awarding virtual currency for completing computer basics training modules (word processing, spreadsheets, presentations, and databases).

In another study, Mekler et al. (2015) used leaderboards, points and levels to track task completion in a gamified art history class. Students were asked to apply their knowledge of art history and techniques by annotating digital images of art; students were rewarded with surface level rewards for completing the annotations. Results showed that even though the quantity of annotations increased, self-reported levels of intrinsic motivation decreased over time (Mekler et al., 2015). Hanus and Fox (2015) reported similar results in their study of a gamified communications course. The researchers awarded badges, points, and leaderboards to reward participation outside class time, such as working with a partner, visiting the library, reading sections of course materials, or visiting the course web page. Results showed that in comparison to the control group, the gamified group participated more initially, with a falloff by the end of the course; and the experimental group reported lower intrinsic motivation overall (Hanus & Fox, 2015).

The dropoff in motivation and engagement in these four studies is consistent with the argument that, over time, extrinsic rewards can diminish or undermine intrinsic motivation (Kapp, 2012; Ryan & Deci, 2000). Moreover, the awards in all of these studies are primarily performance completion feedback awards, and none of the studies incorporated the use of instructional game elements. The fact that participants in Mekler et al. (2015) continued to engage in the task even when their intrinsic motivation dropped supports the theory that both intrinsic and extrinsic motivation can exist at the same time (Ryan & Deci, 2000).

Two studies, however, showed sustained intrinsic motivation, increased time-on-task and increased number of tasks completed. By integrating digital progress bars, badges, and challenges into a eighth grade literacy and vocabulary building unit, Abrams and Walsh (2014) reported increased intrinsic motivation, higher quantity of tasks completed and more time spent on vocabulary challenges. Similar results were reported by Landers and Landers (2014), who used gamified challenges and leaderboards to motivate university psychology students to express their knowledge of course concepts in the form of online wikis. In comparison to the control group, students showed more time-on-task and reported higher intrinsic motivation (Landers & Landers, 2014). Researchers noted, however, that increased motivation was reported in relation to succeeding in challenges, not in regards to the quantity of the wiki entries produced (Landers & Landers, 2014).

Here we see a key difference between the two studies that showed increased intrinsic motivation (Abrams & Walsh, 2014; Landers & Landers, 2014 ) and the four that did not (Chen et al., 2015; Hanus and Fox (2015); de-Marcos, Garcia-Lopez, & Garcia-Cabot, 2016; Mekler et al., 2015): the former studies used primarily game elements that provide measurement feedback (progress bars, challenges and leaderboards) rather than completion feedback elements (badges, points, and virtual currency).

Several of the studies measured the impact of gamification in regards to how game elements fulfilled motivational or psychological needs of learners. Three studies identified self-efficacy and social needs specifically as indicators of intrinsic motivation, while others used self-determination theory (SDT) as a theoretical framework for investigating impacts of gamification on motivation. SDT argues that experiences that fulfill psychological needs of competence, autonomy, and relatedness build intrinsic motivation (Ryan & Deci, 2000). This analysis will use SDT as a lens for examining the studies' finding because competence is roughly equivalent to self-efficacy, autonomy is related to the control granted by game elements that allow choice and the freedom-to-fail, and many game elements and design features promote social interaction, which fulfills the need for relatedness.

Competency, autonomy, and relatedness. Banfield and Wilkerson (2014) argued that intrinsic motivation is contingent upon feelings of self-efficacy; and that external motivation from game elements can lead to intrinsic motivation over time, which, in turn, is internalized as increased self-efficacy. In other words, they argue that gamification has the ability to convert extrinsic motivation into intrinsic motivation, which improves learners' feelings of self-efficacy (Banfield & Wilkerson, 2014). In their experiment, researchers integrated points and leaderboards into a university computer science lesson designed around solving IT administration challenges. For example, students competed with each other online when solving a series real-world network communication calculations and issues; their points and rankings within class groups were displayed in leaderboards. Even though the researchers found the competition and completion awards were extrinsically motivating, results showed statistically significant increases in measures of both self-efficacy and intrinsic motivation (Banfield & Wilkerson, 2014).

Self-efficacy was also the key measure of intrinsic motivation in Harrold (2015), who examined the impact of competition and completion awards (points, badges, and leaderboards) on motivation in a grade 12 English composition course. Students were awarded for completing various reading and writing tasks (i.e. peer reviews, titles read, written works, etc. ) throughout units of study. Interestingly, students who reported either low or high self-efficacy and intrinsic motivation on the pretest were the ones who reported the most gains in intrinsic motivation and self-efficacy on the posttest; participants with mid level intrinsic motivation and self-efficacy remained unchanged (Harrold (2015).

While Banfield and Wilkerson (2014) and Harrold (2015) do not refer directly to SDT, their results can be explained by SDT. First, in regards to Banfield and Wilkerson (2014), the conversion of extrinsic motivation to intrinsic motivation is conceptually consistent with SDT. Deci and Ryan (2000) argued that extrinsic motivators can be effectively used without undermining intrinsic motivation. Moreover, when extrinsic motivation is autonomous, meaning there is inherent choice, or when a person can see a value of performing a task even if they do not want to engage in it, extrinsic rewards can facilitate the internalizing the value of the task; in turn that value can lead to positive self-image and competence.

Second, it is possible that the competition elements in Banfield and Wilkerson (2014) created social comparisons that fulfilled needs of relatedness, and when learners solved challenges, they internalized the value of the IT management skills, which increased their feelings of competence. Finally, learners had much autonomy, as the challenges were designed to promote experiential learning in which the students had choice, control, and the freedom-to-fail at tasks.

In regards to interpreting results of Harrold (2015) through the SDT lens, students had much autonomy in the writing course (flexible time frames, choice of titles to read, freedom to submit multiple drafts for evaluation, and many types of composition style to choose from). In fact, out of all the game elements, students favoured the freedom-to-fail and mastery grading systems the most (Harrold, 2015). As well, the increased intrinsic motivation experienced by those who initially reported low motivation may be an example of how external motivations (completion awards in the form of badges and points) can facilitate the internalization of the values connected to unmotivating tasks, which ultimately increases one's intrinsic motivation by generating feelings of success and self-efficacy.

Other studies also reported intrinsic motivation stemming from extrinsic motivators. In their study on the impacts of digital badges in elementary school geometry lessons, da Rocha Seixas, Gomes and de Melo Filho (2016) found that students who were identified as having social and collaboration motivational needs also reported the highest levels of intrinsic motivation, which suggests badges appear to have fulfilled their needs for relatedness. To be specific, when a badge was awarded for mastering a particular geometry skill or concept, it was publicly displayed in the classroom, which gave students social recognition for their achievements. Similarly, Kwon, Halavais and Havener (2015) studied why learners choose to share and publicize their digital badges awards. The researchers examined achievement badges that were awarded for variety of tasks related to fitness, leisure, programming, community involvement, and education. Results showed that digital educational badges fulfilled learners' needs for competence (as indicated by self-efficacy) and relatedness (as indicated by social comparison and personal/social identity; Kwon, Halavais & Havener, 2015).

SDT can also explain why in some of the studies engagement increased while intrinsic motivation fell or remained unchanged. For example, Merkler et al. (2015) reported increased quantity of image annotation tasks (engagement) but lower intrinsic motivation overall; in this case, extrinsic motivation (points) maintained task engagement by fulfilling some needs for relatedness (leaderboards), but the lesson lacked autonomy (15 images annotations were required, with no choice or control offered to the learner) which may not have facilitated the valuing of task. Had the lesson continued past 15 images, it is possible that the task engagement would decreased as well. Also, Lieberoth (2015), who studied the effect of framing a participant survey like a board game, observed increased time-on-task engagement but no improvements to intrinsic motivation. Conceivably, the social and competition game elements (relatedness) in Lieberoth (2015) and Landers and Landers (2014) kept participants engaged at first, but the participants did not find value in the tasks, which were primarily simply answering questions or performing arbitrary participation tasks.

Only three of the studies specifically accounted for demographic variables in their findings. Cheong, Filippou and Cheong (2014) found no correlation between demographic data (age or sex) and positive feelings toward game elements. This was consistent with Mekler et al. (2015) and Harrold (2015); both studies showed no relationship between motivation and age or sex. It should be noted, however, that in many of the studies there were more male than female participants.

There is a tendency in gamification literature to describe the learner as a "player." Since gamification attempts to plug into the engagement afforded by games, some researchers asked if participants' videogame exposure and gaming habits impacted their experiences in gamified learning. Two studies (Cheong, Fillipou, & Cheong, 2014; Hanus & Fox, 2015) specifically measured if prior gaming experiences and gaming inclinations (i.e. preferences for adventure games, simulations, puzzles, etc.) impacted motivation in gamified learning; both found that prior exposure to videogames or game preferences had no correlation with motivation in gamification. However, a limitation noted in many studies was that since learning tasks involved the use of computers or mobile devices, participants' motivation may be impacted the novelty of technology.

Some studies examined how participants' existing motivational orientations or psychological needs impacted motivation in gamified learning. First, as indicated earlier, Harrold (2015) found that students who were identified in a pretest as having a self-efficacy needs orientation (as opposed to self-regulation or self-determination) reported the highest increases in intrinsic motivation in gamified learning in the posttest. As well, participants in da Rocha Seixas, Gomes and de Melo Filho (2016) were grouped into four levels of engagement (low to high), and then each group's engagement was measured against a number of motivation indicators, such as needs for social interaction, autonomy, collaboration, task completion, and enjoyment. In all four groups, participants who showed social and collaboration needs indicators were also the participants who showed the greatest positive impact on their motivation (da Rocha Seixas, Gomes, & de Melo Filho, 2016). Finally, Kwon, Halavais and Havener (2015) found that the need to express one's self-efficacy, social identity and personal identity were the dominant reasons for sharing educational badges. Interestingly, reasons for displaying educational awards were the most similar to the reasons for displaying fitness badge awards; as well, the need for self-efficacy was the least reported reason for sharing badges awarded for leisure activities, such as engaging in social networking challenges, being first to "check-in" at an event, or to write restaurant reviews (Kwon, Halavais, & Havener, 2015). The researches do not address this difference, but this may suggest that leisurely gamification tasks (perhaps even videogames themselves) are not well suited to learning, as they may not fulfil needs of self-efficacy. On the other hand, the gamification of fitness may be a promising avenue for gamification of learning research because it appears to be similar in regards to self-efficacy needs fulfilment, perhaps because like fitness, education shares the goal of self-improvement.

Kapp (2012) noted that instructional designers of educational software and gamification should use different types game elements depending on the type of knowledge being delivered or practiced. However, this analysis did not find any correlation between types of knowledge domains and the use of particular game elements, or between knowledge domains and impacts on motivation. Moreover, the analysis did not uncover specific subject areas particularly well suited to gamification.

Nevertheless, the analysis revealed a relationship between learning objectives and motivation. Studies that asked participants to learn new curricular content (Abrams & Walsh, 2014), to create or demonstrate knowledge creations (Harrold, 2015; Landers & Landers, 2014) to solve challenges (Banfield & Wilkerson, 2014; Harrold, 2015), or practice new skills (da Rocha Seixas, Gomes and de Melo Filho, 2016) also reported increases in intrinsic motivation. Conversely, studies that asked participants to complete modules (de-Marcos et al., 2016), repeat tasks (Mekler et al, 2015), share opinions (Lieberoth, 2015) or simply engage in course related behaviors (i.e. work with a partner, visit a course web site, or visit the library; Chen et al., 2015; Hanus & Fox, 2015) reported initial engagement with a falloff in motivation or, worse, decreases in intrinsic motivation.