Stepwise\nregression<\/strong> (data driven): This approach\nassumes that features are flat. The best-fitting model is sought using\nstatistical criteria to determine the appropriate combination of predictors\n(regardless of whether they are themes, mechanics, or genres).<\/li><\/ul>\n\n\n\nIn\nmodeling the analysis in two different ways, we sought to better understand whether\nconsidering themes, mechanics, and genres was more useful in terms of entire\ncategories of game features or in terms of individual game features that\ncombine to influence players\u2019 opinions in certain ways.<\/p>\n\n\n\n
For both models, each\nof the 208 educational games in our study had 16 measures. First,\nis the Geek Rating (or Bayes-corrected average rating), which represents\nplayers\u2019 reception of a game. Geek Ratings are continuous values with an upper\nbound of 10 and a lower bound of 1. Next, are 15 component scores, one for each\ntheme, mechanic, and genre that emerged from our PCAs. Component scores range\nbetween \u20133.05 and 13.91 and indicated the extent to which an individual game\nhas that game feature. For example, the game Master of Economy has a score of 8.45\non the trading component, indicating that its mechanics require a substantial\namount of trade and exchange between players.<\/p>\n\n\n\n
Results<\/h2>\n\n\n\n
These procedures allowed us to answer our research questions. The following sections include a description of the themes, mechanics, and genres that resulted from our PCAs and the relationship between these game features and player ratings from the BGG community. <\/p>\n\n\n\n
RQ1: Themes, Mechanics, and Genres<\/h3>\n\n\n\n
Table 2 shows the fifteen themes, mechanics, and genres that emerged from our reclassification and summary of the category and mechanic tags in the BGG database.<\/p>\n\n\n\n
Table 2 <\/strong>
Themes, Mechanics, and Genres Derived From Original BGG Categories and Mechanics <\/p>\n\n\n\n\n\n\n\tComponents<\/strong><\/th>Interpretation<\/strong><\/th>\n<\/tr>\n<\/thead>\n\n\n\tThemes<\/strong><\/td>\n<\/tr>\n\n\tMiddle Eastern Conflict<\/td> | Represents themes associated with wars in the Middle East during the 20th and 21st centuries.<\/td>\n<\/tr>\n | \n\t17th to 19th Century History<\/td> | Represents themes associated with the conflicts and other history of the 1600s through the 1800s.<\/td>\n<\/tr>\n | \n\t19th to 21st Century History<\/td> | Represents themes associated with the conflicts and other history of the 1800s through the present.<\/td>\n<\/tr>\n | \n\tMedia-Based<\/td> | Represents the themes of games based on other media, including books, television, and film. It also represents themes common to those media, including the American West.<\/td>\n<\/tr>\n | \n\tProgress and Development<\/td> | Represents themes of games that give the player the role of guiding a nation, company, or other organization from being small to being large or from the past to the future.<\/td>\n<\/tr>\n | \n\tTravel<\/td> | Represents themes involving exploration, adventure, and travel.<\/td>\n<\/tr>\n | \n\tMechanic<\/strong><\/td>\n<\/tr>\n\n\tTrading<\/td> | Represents mechanics that involve acquiring, trading, or selling one or more goods.<\/td>\n<\/tr>\n | \n\tActing and Betting<\/td> | Represents two kinds of mechanics: Those that have players act out roles or tell stories and those related to betting and gambling. These two kinds of mechanics appear together because a number of games in this study have mechanics related to both.<\/td>\n<\/tr>\n | \n\tMap<\/td> | Represents mechanics closely related to moving between or connecting places on a game map of some kind.<\/td>\n<\/tr>\n | \n\tMemory<\/td> | Represents mechanics that require players to memorize and recall information in order to succeed in the game.<\/td>\n<\/tr>\n | \n\tNumber<\/td> | Represents mechanics that require players to manipulate numbers and make calculations.<\/td>\n<\/tr>\n | \n\tStrategy<\/td> | Represents mechanics typically associated with games that put an emphasis on planning, conflict, and management.<\/td>\n<\/tr>\n | \n\tGenre<\/strong><\/td>\n<\/tr>\n\n\tStrategy Games<\/td> | Represents genres that require thinking or skill, from pursuing correct strategies to correctly balancing pieces.<\/td>\n<\/tr>\n | \n\tParty Games<\/td> | Represents genres that focus on interaction with other people.<\/td>\n<\/tr>\n | \n\tCard Games<\/td> | Represents genres common to games that use cards frequently or exclusively.<\/td>\n<\/tr>\n<\/tbody>\n | \n\n\tNote<\/em>. Component loading factors listed in Appendix.<\/th>\n<\/tr>\n<\/tfoot>\n<\/table>\n\n\n\n\nRQ2: Relationship Between Game Features and Player Ratings<\/h3>\n\n\n\nWe pursued two strategies for modeling\nthe ways these 15 principal components (i.e., game features) influenced player ratings\n\u2014 a theory-driven hierarchical model and a data-driven flat model.<\/p>\n\n\n\n Theory-driven model.<\/em><\/strong> <\/em>In this\napproach, the conceptual framework of themes, mechanics, and genres guided the\nanalysis approach. Accordingly, hierarchical regression (Table 3) was used with\nthe Bayes-adjusted Geek Rating as the dependent variable and component scores\n(grouped by themes, mechanics, and genres) as the independent variables.<\/p>\n\n\n\nTable 3 <\/strong> Theory-Driven Model of Player Enjoyment and Educational Game Features <\/p>\n\n\n\n\n\n\n\tModel<\/strong><\/th>Themes<\/strong><\/th>Mechanics<\/strong><\/th>Genres<\/strong><\/th>\n<\/tr>\n<\/thead>\n\n\n\tParameters<\/td> | B<\/em><\/td>SE<\/td> | B<\/em><\/td>SE<\/td> | B<\/em><\/td>SE<\/td>\n<\/tr>\n | \n\tStep 1: Related to Themes<\/td>\n<\/tr>\n | \n\t Middle Eastern Conflict<\/td> | .00<\/td> | .02<\/td> | .00<\/td> | .02<\/td> | .00<\/td> | .02<\/td>\n<\/tr>\n | \n\t 17th\u201319th Century History<\/td> | .12**<\/td> | .02<\/td> | .12***<\/td> | .02<\/td> | .12***<\/td> | .02<\/td>\n<\/tr>\n | \n\t 19th\u201321st Century History<\/td> | .04<\/td> | .02<\/td> | .03<\/td> | .02<\/td> | .03<\/td> | .02<\/td>\n<\/tr>\n | \n\t Media-Based<\/td> | \u2013.07**<\/td> | .02<\/td> | \u2013.07**<\/td> | .02<\/td> | \u2013.07**<\/td> | .02<\/td>\n<\/tr>\n | \n\t Progress and Development<\/td> | \u2013.01<\/td> | .02<\/td> | .00<\/td> | .02<\/td> | .01<\/td> | .03<\/td>\n<\/tr>\n | \n\t Travel<\/td> | \u2013.01 <\/td> | .02<\/td> | \u2013.06<\/td> | .03<\/td> | \u2013.06<\/td> | .03<\/td>\n<\/tr>\n | \n\tStep 2: Related to Mechanics<\/td>\n<\/tr>\n | \n\t Trading<\/td> | \u00a0<\/td> | \u00a0<\/td> | \u2013.02<\/td> | .02<\/td> | \u2013.03<\/td> | .02<\/td>\n<\/tr>\n | \n\t Acting and Betting<\/td> | \u00a0<\/td> | \u00a0<\/td> | \u2013.02<\/td> | .02<\/td> | \u2013.02<\/td> | .02<\/td>\n<\/tr>\n | \n\t Map<\/td> | \u00a0<\/td> | \u00a0<\/td> | .06<\/td> | .03<\/td> | .06<\/td> | .03<\/td>\n<\/tr>\n | \n\t Memory<\/td> | \u00a0<\/td> | \u00a0<\/td> | .00<\/td> | .02<\/td> | .00<\/td> | .02<\/td>\n<\/tr>\n | \n\t Number<\/td> | \u00a0<\/td> | \u00a0<\/td> | \u2013.01<\/td> | .02<\/td> | \u2013.01<\/td> | .02<\/td>\n<\/tr>\n | \n\t Strategy<\/td> | \u00a0<\/td> | \u00a0<\/td> | .01<\/td> | .02<\/td> | .00<\/td> | .02<\/td>\n<\/tr>\n | \n\tStep 3: Related to Genres<\/td>\n<\/tr>\n | \n\t Strategy Games<\/td> | \u00a0<\/td> | \u00a0<\/td> | \u00a0<\/td> | \u00a0<\/td> | \u2013.04<\/td> | .03<\/td>\n<\/tr>\n | \n\t Party Games<\/td> | \u00a0<\/td> | \u00a0<\/td> | \u00a0<\/td> | \u00a0<\/td> | \u2013.01<\/td> | .02<\/td>\n<\/tr>\n | \n\t Card Games<\/td> | \u00a0<\/td> | \u00a0<\/td> | \u00a0<\/td> | \u00a0<\/td> | .00<\/td> | .02<\/td>\n<\/tr>\n | \n\tTotal R<\/em>2<\/td>.21<\/td> | .23<\/td> | .24<\/td>\n<\/tr>\n | \n\tF<\/em> for change in R<\/em>2<\/td>8.70*** (6,201)<\/td> | 1.16 (6,195)<\/td> | 0.94 (3,192)<\/td>\n<\/tr>\n<\/tbody>\n | \n\n\t* p<\/em> < .05, ** p<\/em> < .01, *** p<\/em> < .001<\/p><\/th>\n<\/tr>\n<\/tfoot>\n<\/table>\n\n\n\n\nThe\nfirst step of the hierarchical regression was a model that included all of the\ntheme-related predictors. This model was significant and accounted for 21% of\nvariance. The effect size of this model \u2014 as measured by Cohen\u2019s f<\/em>2 <\/sup>\u2014 was .27, which Cohen\n(1992) suggested interpreting as a medium effect. In other words, themes have a\nsignificant influence on player ratings.<\/p>\n\n\n\nIn the second step, we added all of the predictors related to mechanics. This second model was still statistically significant and now explained 23% of variance. The change in explained variance is practically small but statistically insignificant (f<\/em>2<\/sup> = .03). That is, adding mechanics to the model did not significantly change the influence of game features on player ratings. <\/p>\n\n\n\nAdding\nall of the genre predictors in the third step contributed even less: The model \u2014\nwhich is still significant \u2014 now explained 24% of variance. The change in\nexplained variance (f<\/em>2<\/sup> =\n.01) was below what Cohen would consider small and was also statistically\ninsignificant. As was the case for mechanics, adding genres to the model did\nnot significantly change the influence of game features on player ratings. <\/p>\n\n\n\nIn the final model \u2014 indeed, in all three models \u2014only two specific game features were shown to have a significant impact on player ratings. These features were both themes: 17th to 19th century history<\/em>, which had a positive effect on ratings, and media based<\/em>, which had a negative effect on ratings. <\/p>\n\n\n\nThe comparison of these models suggests that the most parsimonious model is one based solely on predictors related to themes. That is, themes on their own explained as much of the variance in player enjoyment in this model as did themes, mechanics, and genres together. Adding mechanic- and genre-based predictors to the model increased the amount of variance that it explained, but adding additional variables to a model always results in an increased R<\/em>2<\/sup> (Field et al., 2012). <\/p>\n\n\n\nFurthermore,\nthese increases are not statistically significant, suggesting that adding these\nother sets of predictors to the model does not improve the model. In short, the\ncollective influence of all of these game features was not significantly\ngreater than the influence of themes alone. <\/p>\n\n\n\n Data-driven model.<\/em><\/strong> <\/strong>In this\napproach, the individual feature components were considered in their own right.\nThat is, although each component drew uniquely from either themes, mechanics,\nor genres, these three categories were otherwise set aside. In this approach,\nthe data itself were used to derive the structure of the best-fitting model. <\/p>\n\n\n\nAccordingly, we carried out an all-subsets\nstepwise regression, which uses statistical criteria to determine the\nappropriate combination of predictors. Kelley and Bolin (2013) warned against\nstepwise regression for any purpose besides \u201cresearch [that] is completely\nexploratory\u201d (p. 93) and urged that researchers rely instead on theory. However,\nthis paper is among the early attempts to use features to predict player\nreception of educational games, so such an exploratory analysis was\nappropriate.<\/p>\n\n\n\n Using an all-subsets stepwise regression of the 15 mechanics, themes, and genres components, the 40 best models for each possible number of variables (i.e., from one predictor to 15 variables) were compared. We used Mallows\u2019s C<\/em>p<\/sub>,a measure of fit, to identify the most appropriate model. Table 4 shows that model, which had a Mallows\u2019s C<\/em>p<\/sub> value of 1.81.<\/p>\n\n\n\nTable 4<\/strong> Statistics-Driven Model of Player Enjoyment and Educational Game Features<\/p>\n\n\n\n\n\n\n\tParameters<\/strong><\/th>B<\/em><\/strong><\/th>SE<\/em><\/strong><\/th>p<\/em><\/strong><\/th>\n<\/tr>\n<\/thead>\n\n\n\tF<\/em>(4, 203) = 13.84; p<\/em> < .001<\/td>\u00a0<\/td> | \u00a0<\/td> | \u00a0<\/td>\n<\/tr>\n | \n\t 17th\u201319th Century History<\/td> | .12<\/td> | .02<\/td> | ***<\/td>\n<\/tr>\n | \n\t 19th\u201321st Century History<\/td> | .04<\/td> | .02<\/td> | \u00a0<\/td>\n<\/tr>\n | \n\t Media-Based<\/td> | \u2013.07<\/td> | .02<\/td> | ***<\/td>\n<\/tr>\n | \n\t Strategy Games<\/td> | \u2013.03<\/td> | .02<\/td> | \u00a0<\/td>\n<\/tr>\n<\/tbody>\n | \n\n\t* p<\/em> < .05, ** p<\/em> < .01, *** p<\/em> < .001 <\/th>\n<\/tr>\n<\/tfoot>\n<\/table>\n\n\n\n\n As is the case for the hierarchical regression, theme-based predictors played a dominant role here. Although the fourth predictor was based on game features coded as genres, the other three predictors were derived from themes. The model was statistically significant (F<\/em>[4, 203] = 13.84; p<\/em> < .001) and had an effect size \u2014 as measured by Cohen’s f<\/em>2 <\/sup>\u2014 of .27, which can be interpreted as a medium effect (Cohen, 1992). Furthermore, the only two predictors to have a statistically significant impact on player enjoyment were the same the predictors that had such an impact in our hierarchical regression: 17th to 19th century history<\/em> and media-based<\/em>. <\/p>\n\n\n\nDiscussion<\/h2>\n\n\n\nThis section describes noteworthy implications of this study\u2019s results. <\/p>\n\n\n\n Game\nFeatures as a Framework for Teachers<\/h3>\n\n\n\nThe\nresults of this study \u2014 particularly the first research question \u2014 may serve as\nbasic guidelines for teacher candidates, in-service teachers, and teacher educators.\nThis study used principal components analysis to identify 15 themes, mechanics,\nand genres that emerged from the BGG community\u2019s classification of games. These\n15 design features contribute to a vocabulary that teachers can use to quickly\nand easily communicate or evaluate the design of a game. <\/p>\n\n\n\n Indeed,\nthe summarized game features identified in Table 2 can serve as an initial\nframework that these populations can use to guide their consideration of (a)\nthe basic categories of game design features they ought to consider when\nevaluating a game, (b) key features in these categories that are common to\neducational games, and (c) whether and how those features correspond with the\ncontent and pedagogical considerations involved in a particular teaching\ncontext. <\/p>\n\n\n\n In\nthe following paragraphs, we demonstrate the utility of these results by\ndescribing how this basic framework (i.e., Table 2) could be used by a hypothetical\ngroup of teacher candidates from a range of grade levels and subject areas\nwhose professor has organized an activity intended to help them look for and identify\nanalog games that may be helpful in their teaching context. <\/p>\n\n\n\n One\nteacher candidate in this class is preparing to be a social studies teacher and\nwonders whether she could find a game that could be used to support a history curriculum.\nLike many teachers (see Sardone & Devlin-Scherer, 2009; Copeland et al.,\n2013), she begins searching for educational games that are thematically related\nto the subject matter that she will be teaching and begins evaluating games\nbased on their themes. <\/p>\n\n\n\n She\nreviews some of the common themes in educational games and notes two periods of\nhistory that educational games are frequently associated with. Based on this\ninsight, she begins identifying potential lesson plans corresponding to these\nperiods and then looking for games that might fit those lesson plans. <\/p>\n\n\n\n Another\nsocial studies teacher candidate is more interested in the potential of games\nto support lessons on globalization. Seeing his classmate\u2019s success in finding thematically\nrelevant games, he sets aside the framework and begins with the intuitive step\nof searching for educational games \u201cabout globalization.\u201d After a few minutes\nof fruitless searching, he consults the list of common themes in educational\ngames and realizes that none of these themes have an explicit connection with\nthe concept he is interested in. <\/p>\n\n\n\n Although\nthis discovery does not rule out the possibility that such a game exists, he begrudgingly\nacknowledges that this theme is not common. Seeing his frustration, another\nclassmate (who has been looking for math-related games) reminds him that theme\nis only one category of design features he could consider and explains that her\nsuccess has come from looking for mechanics related to the mathematical\nconcepts she teaches. The two teacher candidates work together to identify\nmechanics that might be related to globalization. A few minutes later, they are\nsearching for trading games that the social studies-focused candidate could\nadapt for a globalization lesson.<\/p>\n\n\n\n A\nthird candidate, preparing to be an elementary school teacher, feels\noverwhelmed by the exercise. He has little personal experience with analog\ngames. In fact, a once-and-never-again attempt at learning a collectible miniatures\ngame popular with his roommates has left him with the impression that games are\ntoo complicated for him \u2014 and if too complicated for him, definitely too\ncomplicated for his young future students. <\/p>\n\n\n\n He\nexpresses this concern to his professor, who asks him to try his best to\ncomplete the activity and reminds him that there are multiple kinds of analog\ngames. The candidate takes this encouragement to heart and begins examining\ncommon genres of educational analog games. To his relief, collectible\nminiatures games are not one of the genres commonly associated with educational\ngames, and although he feels like strategy games would probably still be\noverwhelming for him and his students, he begins searching for party and card\ngames he might be able to use, recognizing them as genres that may be less\ncomplicated and more age-appropriate.<\/p>\n\n\n\n These examples show how a framework of analog game design concepts may be helpful for guiding teacher candidates, in-service teachers, and teacher candidates in evaluating and choosing games for specific educational contexts. Indeed, the experience of these hypothetical teacher candidates shows not only that the separate categories of themes, mechanics, and genres all have implications for educators\u2019 evaluation of analog educational games but also that knowing common themes, mechanics, and genres in educational games can help these educators know where to start looking and what to look for when searching for analog games to use in their particular professional contexts. <\/p>\n\n\n\n We\nalso recognize the need for expansion of and further nuance in such a\nframework. Indeed, the summarizing effect of our principal components analysis\nnecessarily leaves out less common design features present in educational games\nand less systematic relationships between these design features.<\/p>\n\n\n\n Anticipating\nPlayer Reception<\/h3>\n\n\n\nJust as the results to our first research question in this study can be used to guide educators\u2019 thinking when considering different analog games, the results to our second research question may provide additional support. That is, the hypothetical teacher educator organizing the activity described in the previous section could also use the results presented in Tables 3 and 4 to guide students in considering the likelihood that a particular game will be received well by students. <\/p>\n\n\n\n Indeed,\nour findings suggest that themes play a role in predicting players\u2019 reception\nof a game. Both of the models that we tested found themes to be statistically significant\npredictors for player ratings (with a medium effect size). Furthermore, as\nevidenced by the two statistically significant predictors in each model, themes\nhave the potential to influence player reception for better (as was the case\nwith the 17th to 19th century history <\/em>theme)\nor for worse (as with the media-based<\/em>\ntheme). Based on these results, educators with a concern for player reception\nshould consider themes carefully when selecting and creating games intended to\nteach.<\/p>\n\n\n\nUpon\nfurther examination, however, the reasons themes such as 17th to 19th century history<\/em> and media-based<\/em> had such an impact on player ratings are not yet clear.\nThe respective positive and negative influences of these themes do not\nnecessarily indicate that they are inherently \u201cgood\u201d or \u201cbad\u201d for games. Indeed,\nimagining that a player\u2019s reception of a game is based entirely on the time\nperiod that it is set in or the media property that it is inspired by is\ndifficult. <\/p>\n\n\n\nRather,\nthese results are more likely to illustrate the complex interplay between\nthemes, other design features (such as mechanics and genres), and factors such\nas player expectations (which were not examined in this study). That is,\nalthough we tried to distinguish analog games\u2019 themes from their genres, there\nmay be some overlap in how the BGG community conceives of each category.<\/p>\n\n\n\n For\nexample, the media-based<\/em> theme\ndescribes games based on established movies, books, and brands, meaning that\nthe theme itself is likely to be well received \u2014 at least by those who already\nappreciate the media franchise. However, the common wisdom in games circles is\nthat players tend to respond negatively to media-based games whose themes\nappear to be \u201cpasted\u201d onto unsatisfying gameplay (Grayson, 2014). <\/p>\n\n\n\nThe\nlow ratings of the media-based<\/em> theme\nmay reflect frustration with the combination of that theme with other elements\nof the game\u2019s design (and not the theme itself). Conversely, players tend to\nrespond positively to a theme that complements other game features and\ncontributes to a \u201cpolished game experience\u201d (Squire, 2011, p. 5). That is, the\npositive reputation of games having a 17th\nto 19th century history<\/em> theme may be associated just as much with other\nparts of the games\u2019 design as with the period of history represented by that\ntheme.<\/p>\n\n\n\nAlthough\nthe relationship between themes and player reception remains ambiguous, these\nfindings, nonetheless, have some implications for educators. For example, imagine\nthat the hypothetical social studies teacher candidate searching for a\nhistory-related game learned that educational games set in the 17th through 19th\ncenturies appear to be particularly well-received by players. Given this\ndiscovery \u2014 and knowing that she will likely not be able to use games in all of\nher units \u2014 she may look for some of these highly rated games in order to make\nthe most of the games that she does employ. <\/p>\n\n\n\n On\nthe other hand, the teacher candidate preparing to use games in an elementary\nschool classroom, if aware of the negative reputation of media-based games, might\nhesitate before choosing an educational game based on its association with a\nmedia franchise popular among his students. Rather than make the decision based\non that association alone, he may take the additional time to evaluate other\nparts of the game\u2019s design and whether it will meet his pedagogical objectives\nand students\u2019 expectations.<\/p>\n\n\n\n The\nImportance of Holistic Evaluation<\/h3>\n\n\n\nUltimately, the findings of this paper emphasize the importance of holistic evaluation of analog games and thorough consideration of their intended pedagogical purpose. The design features identified in this study are not comprehensive, and the relationship of themes with player reception of games is somewhat ambiguous. These design features should be seen as a starting point for those educators interested in comparing existing educational games and not as a comprehensive description of what educational games may look like or afford. <\/p>\n\n\n\n In\nother words, our results suggest that a history teacher has many games to\nchoose from and that games set in the 17th through 19th centuries have been\nparticularly well-received in the BGG community. Our hypothetical history-focused\nteacher candidate, however, is not guaranteed pedagogical success or student\nenjoyment simply by following these results. <\/p>\n\n\n\n Indeed,\nthe mechanics of a game have been found to have an important effect on players\u2019\nenjoyment (Yee, 2006; Williams et al., 2008), and a number of educators have\nfound it to be a useful indication of what is taught in a game (e.g., Berland\n& Lee, 2011; Farber, 2016; Mayer & Harris, 2008). If this teacher\ncandidate were to select a game simply based on its being set in the 18th\ncentury without considering the rest of its design, she may find that it has\nonly a tenuous connection to the learning objectives she has in mind or that\nher students find it to be less engaging than other kinds of activities. <\/p>\n\n\n\n Conversely,\nalthough the themes and mechanics identified in this study do not appear to\nlend themselves well to world language education (for example), a preservice\nFrench teacher may still be able to find a helpful analog game for that\ncontext. The design elements listed in Table 2 suggest that there are few\ncommon mechanics, themes, or genres with explicit and obvious connections to\nthe French language, indicating that this teacher candidate may have more\ndifficulty than some of his colleagues finding an obvious educational game for\nhis classroom. <\/p>\n\n\n\n Carefully considering what themes, mechanics, and genres might support a French language or culture lesson, however, would have considerable benefit for him as he looks for analog games. Furthermore, if any of these design features were found to have parallels in the list of common features, the common list of features may be helpful in identifying starting points for adapting games. For example, if the teacher candidate decides he wants mechanics that promote conversation and speaking among his students, he may find that the common acting and betting<\/em> mechanic has more potential for his classroom than he initially thought.<\/p>\n\n\n\nLimitations and Future Research<\/h3>\n\n\n\nAlthough\nBGG provides a rich source of data for this kind of research, using this data also\nimposes limitations on the conclusions that can be drawn from this study\u2019s\nresults. Perhaps most obviously, our conclusions are limited to board games (or\nother kinds of analog games). As previously described, board, card, and other\ntabletop games are growing in popularity, and more research of the kind seen in\nthis paper is needed. However, the fact remains that digital games remain the\nemphasis of much research on games and education, and our findings should not\nbe hastily applied to these other areas of research.<\/p>\n\n\n\n There\nare further limitations to our conclusions even when they are only applied to\nboard games. For example, because the user ratings and information on themes,\nmechanics, and genres on BGG are all crowdsourced, we do not know much about\nthe identity of the players who are classifying and reviewing these games.\nFurthermore, what little information can be inferred implies still further\nlimitations: It can be reasonably concluded that the BGG community is composed\nlargely of adults. However, the target audience for educational board games is\nmore likely to be children and young adults, many of whom may not have the same\ntastes as adults. <\/p>\n\n\n\n Likewise,\nthe BGG community is largely geared towards hobby\ngaming<\/em>, a niche group of games that is often distinguished by its fans from\nthe mass-market board games familiar to most people. What hobby gamers find\nappealing in a game may not be what the general population of learners finds\nappealing in a game.<\/p>\n\n\n\nFurthermore, the number and kinds of games included in this data need to be taken into account when interpreting the results. BGG likely has one of the most comprehensive lists of educational board games in existence, but having more data for this study would still have been useful. Although we took the size of the BGG data into appropriate consideration when performing the PCAs in this study, including more data would further increase confidence in the resulting components. <\/p>\n\n\n\n It is also unclear how the BGG community identified and distinguished educational games: Some games in this study have conceivably never been used in classrooms and other games that have been used to teach may not have been included.<\/p>\n\n\n\n In\nshort, while this study provides useful information, it is largely a study of\nthe BGG community, and further research is needed to produce more generalizable\nresults about game classification and reception. Researchers interested in\npursuing these questions should expand this kind of work to larger sets of data\nand to more transparent settings. <\/p>\n\n\n\n In\nthe first case, BGG data may still prove useful. After all, the distinction\nbetween an educational game and an entertainment game is not always easy to\nmake. In the hands of a knowledgeable teacher, any board game that is carefully\nselected, framed, and scaffolded (see Shah & Foster, 2015) could\nconceivably become an educational resource. Future research may benefit from\nanalyzing all of the games in the BGG database to determine what game features\nemerge from a broader study and how those features make board games generally\nmore (or less) enjoyable. These findings could then be applied to educational\ngames. <\/p>\n\n\n\n In\nthe second case, researchers should spend time in classrooms where educational\ngames are being played and ask students and teachers how they make distinctions\nbetween games when choosing them (whether in terms of pedagogical value or\nreception). In addition to providing helpful guidelines for educators, this\nresearch could also be valuable in highlighting differences between how\nstudents and teachers categorize games and in exploring whether and how certain\nthemes or mechanics appeal differently to different groups of students (e.g.,\nelementary school students vs. high school students). <\/p>\n\n\n\n Conclusion<\/h2>\n\n\n\nIn this study, we examined data from the website BGG to investigate the design features that describe and distinguish educational games and determine how they influence players\u2019 reception of those games. We used principal components analysis to summarize the themes, mechanics, and genres being used to make distinctions between educational games. We then tested two different models to explore the relationship between these design features and player ratings on BGG: a hierarchical, theory-driven model and a statistical, data-driven model. Both approaches produced statistically significant results, suggesting that themes can play a role in player enjoyment. In contrast, mechanics and genres were not shown to significantly predict enjoyment. <\/p>\n\n\n\n The results of this study provide initial guidelines for teacher candidates, preservice teachers, and teacher educators who are comparing the design of educational analog games and lend empirical support to the argument that themes play an important role in predicting players\u2019 reception of a game. In light of the fact that many educational games are perceived to be less enjoyable than their entertainment counterparts, researchers and educators would benefit from a better understanding of the ways specific features of games can contribute to positive or negative player receptions. Themes should be carefully considered in the design, selection, and use of games for learning. <\/p>\n\n\n\n Our interpretation of these findings nonetheless suggests that, above all, a holistic examination of the design of analog games is necessary for their effective use in education. The design features identified in this study suggest which content areas have the most educational games available to them and provide a means of distinguishing between these games but do not comprehensively describe the design or pedagogical affordances of analog games. Furthermore, the themes in this study may be broader than just superficial \u201cdressing\u201d (Koster, 2004, p. 85), suggesting that educators should consider more than just the apparent subject material of an analog game when evaluating its potential for their classroom. <\/p>\n\n\n\n References<\/h2>\n\n\n\nAbt,\nC. C. (1970). Serious games<\/em>. New\nYork, NY: The Viking Press.<\/p>\n\n\n\nAmaro, S., Viggiano, A., Di\nCostanzo, A., Madeo, I. Viggiano, A., Baccari, M. E., . . . De Luca, B. (2006).\nKal\u00e8do, a new educational board-game, gives nutritional rudiments and\nencourages healthy eating in children: A pilot cluster randomized trial. European Journal of Pediatrics<\/em>, 165<\/em>, 630-635. https:\/\/doi.org\/10.1007\/s00431-006-0153-9<\/a><\/p>\n\n\n\nArnold,\nR. (2015). Where\u2019s the diplomacy in Diplomacy? Using a classic board game in\n\u201cIntroduction to International Relations.\u201d Political Science &\nPolitics<\/em>, 48<\/em>(1), 162\u2013166. https:\/\/doi.org\/10.1017\/S1049096514001711<\/a><\/p>\n\n\n\nBedwell, W. L., Pavlas, D., Heyne,\nK., Lazzara, E. H., & Salas, E. (2012). Toward a taxonomy linking game\nattributes to learning: An empirical study. Simulation\n& Gaming<\/em>, 43<\/em>, 729\u2013760. https:\/\/doi.org\/10.1177\/1046878112439444<\/a><\/p>\n\n\n\nBerland, M., & Lee, V. R.\n(2011). Collaborative strategic board games as a site for distributed\ncomputational thinking. International\nJournal of Game-Based Learning<\/em>, 1<\/em>(2),\n65\u201381. https:\/\/doi.org\/10.4018\/ijgbl.2011040105<\/a><\/p>\n\n\n\nBogost, I. (2007). Persuasive games: The expressive power of\nvideogames<\/em>. Cambridge, MA: MIT Press.<\/p>\n\n\n\nBrathwaite, B., & Schreiber, I.\n(2008). Challenges for game designers<\/em>.\nRockland, MA: Charles River Media.<\/p>\n\n\n\nBreuer,\nJ., & Bente, G. (2010). Why so serious? On the relation of serious games\nand learning. Eludamos: Journal for Computer Game Culture<\/em>, 4<\/em>, 7\u201324.<\/p>\n\n\n\nBridge,\nD. (2014). You sunk my constitution: Using a popular off-the-shelf board game\nto simulate political concepts. Journal\nof Political Science Education<\/em>, 10<\/em>,\n186\u2013203. https:\/\/doi.org\/10.1080\/15512169.2014.894363<\/a><\/p>\n\n\n\nBruckman,\nA. (1999, March). Can educational be fun?<\/em> Paper presented at the Game\nDevelopers Conference, San Jose, CA.<\/p>\n\n\n\nCannice, M. V. (2013). The right\nmoves: Creating experiential management learning with chess. International\nJournal of Management Education<\/em>, 11<\/em>,\n25\u201333. https:\/\/doi.org\/10.1016\/j.ijme.2012.11.002<\/a><\/p>\n\n\n\nCharsky, D., & Mims, C. (2008).\nIntegrating commercial off-the-shelf video games into school curriculums. TechTrends<\/em>, 52<\/em>(5), 38\u201344. https:\/\/doi.org\/10.1007\/s11528-008-0195-0<\/a><\/p>\n\n\n\n | | | | | | | | | | | | | | | | | | | | | |