Keywords

1 Introduction

Games have become prevalent in education for its potential to provide a motivational and meaningful learning systems that tailor learner interactions with instructional materials and tasks. In a typical game design, the user’s success depends on overcoming various challenges in the games which offer a highly self-directed and immersive experiences [1]. By allowing learners to explore and experience the context, learners can be more engaged [2]. Compared to traditional instructional media, previous study [3] supported that games led to higher involvement, higher motivation and higher entertainment. Serious games were defined as the combination of the educational purposes with the form of game-based learning. Serious games could accommodate the complex dynamic interactions of multiple people, contexts, subject knowledge and learning outcomes in the educational environment [1], and grant enjoyable opportunities for formal and informal learning environment. The process of playing games often involves problem solving, reading comprehension, and social skills [3], more educational studies conducted in informal learning environment, such as libraries and museums, have paid attention to serious games for information literacy education to motivate and enhance patrons’ problem identification skill, thinking skill, analysis skill, expression skill and ability to communicate [4].

Serious games are used in a wide variety of occasions. Michael and Chen divides serious games into eight categories according to the market: educational games, business games, medical games, government games, political games, military games, religious games, and art games [5]. Regarding the character design of serious games, Bonsignore, Hansen, Kraus, Visconti and Fraistat pointed out that there are three important roles in the immersive serious game, the protagonist-by-proxy, the protagonist-mentor and the antagonist. The protagonist-by-proxy is the avatar that the learner himself manipulates in the game or other characters which have similar skill. The design of protagonist-by-proxy needs to be similar to the learner’s age to simulate the efficient information search and solution that the player can emulate. The protagonist-mentor provides learner with training and advice. In the serious game of information literacy, the protagonist-mentor often appears as a librarian. The antagonist, on the other hand, can cause conflicts and promote user participation.

However, when learners interact with one another together in the game-based learning environment, the interaction could be complicated with a variety of connections and relationships occurred. Cooperation and competition, for example, were explored by previous studies as they were typical in many games, yet sometimes counteracted with each other in terms of motivation and learning. Cooperative learning is considered to minimize the unpleasantness of group work and improve learning outcomes and satisfaction in both formal and informal educational situation [6]. While competitive learning is considered to enhance learning outcomes by competing with each other [7], yet it also causes negative effects such as disruption of relationships and anxiety [8]. Competition and collaboration are both common motivational factors for business multiplayer games and online games [9], but the impact of competition and collaboration on learning motivation, learning participation and learning effectiveness in serious games are not consistent in past research works.

In addition, the current empirical research on the impact of collaboration or competitive interaction on learning outcomes is limited to serious games without simulations, virtual characters, and tasks. Most of the empirical studies were conducted in formal education contexts, and there is a lack of in-game communication channel. While the simulation scenario is helpful to promote learners’ engagement [2] and learners’ emotions and motivations could largely differ in formal and informal education, empirical studies would be necessary to investigate the impact of collaboration and competition on learning in a simulation games with learning tasks.

Motivated by the aforementioned issues, this study intends to explore the elements of collaboration and competition in a multi-player serious game. An online game-based learning environment and materials were designed and developed for college students to understand the topic of citrus fruits, and participants’ interaction behaviors and performance were recorded and analyzed.

2 Research Design

This study compared the impact of different forms of learning on learning outcomes. A quasi-experimental design was adopted. Twenty-four college students were recruited to participate in the study. And they were randomly assigned into three groups with 8 valid samples in each group. From the review of the related studies, it was hypothesized that the game could train users to find, use and judge information. Helpful elements such as virtual characters, backgrounds and plots, tasks in games were therefore introduced and designed in this study to carry out practical operations and exercises.

2.1 Game Design and Development – The Citrus

For the instructional purpose to develop college students’ awareness and literacy about the library special collection, the topics of plant taxonomy that related to a series of special collection called “Tanaka Collection [10]” was selected for the serious game. The learning progression that outlined the sequence of the sequence of the instruction was defined with reference to the literature, librarians, and subject matter experts with a background in botany. Tanaka Collection consisted of 3,856 titles of books, journals, magazine, and pamphlets written in more than 15 languages, which could be viewed as a major library in the world in the area of botany studies. This collection was made by Prof. Tyôzaburô Tanaka during his term of the first University Librarian of Taihoku Imperial University, the predecessor of National Taiwan University. The Collection itself stood for an exceptional case of colonialism and modernity, which was believed to benefit Taiwanese students to develop cultural legacy. Based on the historical importance of the collection and the rich contextual resources provided in the literature of this collection, the game was set in historic time period of 1928–1945 when Prof. Tanaka was a professor at the Department of Horticultural of the College of Science, Thaihoku Imperial University. Players were asked to visit the simulated setting of Shilin Horticultural Experimental Branch, where Prof. Tanaka often conducted his field research, and use the botanical concepts to help Prof. Tanaka identify certain citrus fruits.

Players joined the game by their avatars to play the game alone, or to interact with others under different social contexts: the one-on-one collaborative learning and one-on-one competitive learning modes. The protagonist-by-proxy character design proposed by Bonsignore et al. [11] was adopted because it prompted the learners to engage in the role and naturally develop a collaborative or competitive relationship. Gaming tasks and challenges included the readings of the literature collection, and citrus species classification. The educational purpose of this serious game is to enhance learners’ understanding of citrus species, the use of literature, and the value of library special collections.

The game was implemented by Unity®. According to the designed roles and relationships between learners, three modes of interaction were developed: “Individualistic”, “One-on-One collaborative” and “One-on-One Competitive” delivered in identical story theme and learning task in terms of information quantity and difficulty. The task is to observe the appearance of the plant, including the shape of trees, fruit and leaf in order to classify 3 types of different citrus fruits including Tankan, Ponkan, and Citrus Unshiu with the literature at their hands. The game design contains immersive background settings and contexts, including modern citrus exhibition scenes and the real citrus garden in the 1920s. In the citrus garden user can choose to take a closer look at the tree, the fruit, and the leaf. Also they could cut off the fruit to see the sectional view as shown in Fig. 1.

Fig. 1.
figure 1

Users can choose to observe the citrus fruits by different points of view

Full size image

Participants were assigned to play the two virtual characters of Mei and Satoshi to interact with Pr. Tanaka in the game, as shown in Fig. 2. With critical information highlighted for the learners, participants needed to read the literature, determine which type of observations to perform, observe and compare the types of features to classify them (see Fig. 3). The game supported different level of investigation from the basic watch-and-remember, to more complicated judging tasks such as cross-referencing all available resources to the observation and problem-solving.

Fig. 2.
figure 2

Users played the virtual roles to interact with the leading character Pr. Tanaka

Full size image
Fig. 3.
figure 3

Handbook of literature (left); A closer and sectional view of the citrus fruits

Full size image

A between-subject experiment was conducted. Participants were randomly assigned into three groups of interaction modes of individual, collaborative and competitive, with their learning outcomes, interactions and behaviors recorded for analysis. For “One-on-One collaborative” and “One-on-One Competitive” groups, to enable participants’ conversation with each other, a chatbot was implemented by the social networking site Slack and Flow XO settings. The chatbot played the virtual character of “Mystery Citrus” in the game for participants to talk with and ask questions to (see Fig. 4).

Fig. 4.
figure 4

The Mystery Citrus in the game (left) and in the chat room (right) for context extension.

Full size image

2.2 Interaction Modes

According to the designed roles and relationships between learners, three modes of interaction were developed: Group 1 (Individualistic), Group 2 (One-on-One collaborative) and Group 3 (One-on-One Competitive). Group 1 (Individual) participants were assigned to play the role Mei who meet Pr. Tanaka and his student Satoshi from the past, and need to observe the fruit trees and reading the literature at the same time herself to classify the citrus. Participants were given a tablet and a printed handbook of literature, as shown in Fig. 5 to complete the game challenges.

Fig. 5.
figure 5

Individualistic version experiment setting

Full size image

For Group 2 (Collaborative) participants, they were assigned to play the role of Mei(A) and Satoshi(B) who were located in two different physical space. They work together online to complete the game tasks of classifying citrus fruits. The character Mei(A) is responsible for explore the field by observing the trees, fruits and leaves, while Satoshi(B) is responsible for reading the handbook of literature. Two participants and the chat robot (Mystery Citrus) communicate with one another by text through online chat rooms to classify citrus species as shown in Fig. 6.

Fig. 6.
figure 6

One-on-one collaborative version experiment setting

Full size image

Group 3 (Competitive) participants were assigned to play the same character of Mei(A) and Satoshi(B), but they compete for limited resources and hints provided by the chatbot (Mystery Citrus) online. There are 9 items of literature in total, 4 are in Chinese and 5 are in Japanese, each of which can only be used by one person. Through the competition of resources, the two participants must explore the citrus garden faster than the other party and judge the required documents earlier. Figure 7 show how the users could get the resources in electronic forms through the online chatroom.

Fig. 7.
figure 7

One-on-one competitive version experiment setting

Full size image

2.3 Instruments and Procedures

A questionnaire to investigate participants’ prior experiences and knowledge toward the collections was developed and distributed during the recruitment. Volunteers with similar background were sampled, and randomly assigned into three groups of different interaction modes. Formative assessment in quiz type were integrated in the game tasks to understand participants’ learning performance. Their interaction behaviors during the experiment were observed and screen-recorded. After they completed the game, the participants were asked to fill in a questionnaire of game attitude on a semantic differential scale and a post-test inventory to measure participants’ flow experiences. The instrument consisted of 14 aspects to investigate critical factors that influence users’ engagement in a game.

3 Preliminary Findings

There were 27 college students volunteered to participate in the user experiment. Three of them did not complete the experiment due to technical problems. The valid sample of 24 college students consisted of 11 males (45.8%), 13 females (54.2%). 15 of them majored in liberal arts and social science (58.3%), 7 in engineering (29.2%) and 2 in agriculture (8.3%). And they were randomly assigned into three groups with 8 valid samples in each group.

3.1 Learning Performance of Different Groups

To understand whether the design of collaboration and competition will affect the learning outcome, we compared the learning outcomes of the three groups by the number of correct answers: Group 1(Individualistic), Group 2(Collaborative), and Group 3 (Competitive).

The three groups were significantly different in their learning outcomes. As shown in Table 1, both the one-on-one collaborative learning outcomes and one-on-one competitive learning outcomes were significantly greater than the individual learning outcomes, while the one-on-one collaborative and one-on-one competitive learning results were not significant (Group 2 > Group 1**, Group 3 > Group 1*, Group 2 > Group 3). The possible reasons why collaborative learning is better than individual learning is because of distinguishing the important parts of literatures during the discussion, and the Group-to-individual transfer. In addition, due to the need of communication and confirmation, or the internal reasons for solving the problem, the participants will spend more time in learning and have complicated learning behavior, so they can achieve better learning results. For example, Group 2 (Collaborative) participants seemed more open-minded than Group 3 (Competitive) participants. They reported that because of the discussion, all the information was double checked, which increased their exposures and rehearsals of the information.

Table 1. Comparison of the learning performance among the three groups
Full size table

On the other hand, participants who play the game together tended to feel more satisfied toward the game experiences (Group 2 > Group 1, Group 3 > Group 1, Group 2 < Group 3). Although the difference is not statistically significant, the interviews with the participants suggested that competition made the game more fun and satisfied because of participants’ motivation to win their counterpart.

3.2 Flow State in Different Groups

The results echoed previous studies and suggested that situational simulation, role setting, collaboration and competition in serious games improved participants’ engagement. The flow state of the three groups of participants was measured and compared accordingly. The results suggested that generally participants were more engaged with higher flow state when they played alone for their game tasks in competitive and individual modes. Compared with single-person learning, multi-person learning is better in several sub aspects of flow state such as self-awareness (F4), perceived importance of the game to the player (F10), perceived importance of the game to others (F11), suggested the benefits of peers, whether collaborative or competitive, to participants’ learning experiences.

However, Group 2 (Collaborative) and Group 3 (Competitive) participants’ scores of the flow in general were lower than Group 1 (Individual), although not statistically significant. Participants who played the game with peers reported lower scores in the sub-aspects such as perceived difficulty to remain concentrated (F2), perceived control over the progression (F5), whether the game meet my expectation (F6) and other’s expectation (F7), and if the game is meaningful to me (F12). The findings suggested the participants tended to follow the game passively without thinking too much, but the social interaction with their partners or competitors distracted them from the game. Furthermore, Group 3 (Competitive) participants reported generally higher scores of flow state than Group 2 (Collaborative) participants. But participants who played the game with a partner were more aware of what they were doing in the game (F3), and perceived the game more meaningful (F12) to their learning than those who played the game with a competitor.

According to the interviews after the experiment, some participants reported they paid more attention to their partners over the game tasks in collaborative mode, it was possible that their social attention distracted their engagement in the learning content. While the situational simulation and role setting provide in collaborative and competitive modes had brought a certain degree of flow experiences, but the operation of online chat room communication could be complicated and interfered by the abilities of the partner. Therefore, the participants may feel less controllable over the learning environment, which could also decrease their scores of flow state.

4 Conclusion and Future works

Based on the historical importance of a special collection in a university library and the rich contextual resources provided in the literature of this collection, this study developed a simulation game about Tanaka Collection to investigate participants’ experiences and performance in individual, collaborative and competitive learning contexts. The main mechanics of the game “The Citrus” was developed and implemented by Unity® with integration of protagonist-by-proxy character design and chatbots, to enable social interaction of the participants. The valid sample of 24 college students consisted of 11 males (45.8%), 13 females (54.2%) with similar prior knowledge and experiences participated in the experiment and were randomly assigned into three groups of individual (Group1), one-on-one collaborative (Group 2), and one-on-on competitive (Group 3) to complete the tasks of identifying 3 types of different citrus fruits.

The preliminary findings suggested that participants performed better and were more satisfied within multiple player interaction modes, whether it be with collaborative partners or competitive opponents. The results of participants’ flow state also supported that the existence of peers affected participants’ learning experiences positively. It was found that when participants collaborated with their partners in the game, their exposures to the related knowledge and information were increased and enhanced during the frequent communication and confirmation between them. The frequent exposures to the learning subjects therefore led to better learning performance. On the other hand, participants in the competitive relationship with their opponents were intrinsically motivated by their psychological pursuit to win before actually playing the game. They were more active in taking challenges by frequently thinking over, therefore they also could achieve better learning performance.

Based on the findings of the study, it is suggested that interweaving the elements of collaboration and competition in serious games was able to provide learners with sufficient exploration time and rich interaction to improve their learning performance. The current study investigated the history game implemented on tablet with simulation technologies. While the integration of collaborative and competitive learning in serious games require more empirical studies for the design guideline, different types of game and alternative instructional media will be explored and interwoven continuously among different storyline and game types in the future.