Moreover, applying a multi-approach behavior pattern analysis
may enable advance triangulation of the results of learning processes
and behavior pattern analysis. Compared to single-method
analyses, such an analytical method can explain in-depth ‘‘why’’
or ‘‘how’’ simulation games promote learning while exploring the
limitations of the game design. Therefore, this study seeks to
explore in-depth learners’ behavioral patterns in a simulation
game with situated-learning context in science education.
1.2. Multi-approach pattern analysis on flow and learning behavior in
simulation games with situated-learning context
Compared to players of popular casual games, players of educational
simulation games perform complex procedural knowledgerelated
operations; thus, greater focus and involvement in the
game and its learning content during the gaming process are
required. Flow refers to a person’s mental state when he is fully
immersed in an activity and filtering out irrelevant emotions
(Csikszentmihalyi, 1975). Flow and the intrinsic motivation of
the activity participants are highly correlated (Moneta, 2012),
which helps to reflect learners’ intensity of motivation and focus
on educational games. During educational gaming, the relation
between learners’ motivation, focus and learning behaviors is a significant
issue worthy of exploration; furthermore, flow also serves
as an important indicator of game-based learning (Bressler &
Bodzin, 2013). Therefore, understanding the relation between flow
and learning behaviors will facilitate the design of educational
simulation games that elevate students’ motivation and allow
them to acquire greater procedural knowledge.
Recent research on game-based learning has been devoted to
learners’ flow experience, including research on learners’ flow state
and the association between learners’ flow, acceptance and learning
effectiveness (Hou & Li, 2014). Some studies indicate that the
challenge and clarity of the objectives of a game have a significant
effect on learners’ flow experience (Hou & Li, 2014; Wang & Chen,
2010). Liu et al. (2011) observed that students participating in digital
simulation games are more prone to flow state than those
engaged in traditional methods of learning; studies also note that
the players’ flow state and their fidelity to the game are positively
correlated (Faiola, Newlon, Pfaff, & Smyslova, 2013). In game-based
learning, the manifestation of flow is closely related to the players’
prior knowledge and the game’s interactive mechanisms
(Admiraal, Huizenga, Akkerman, & ten Dam, 2011; Hwang, Hong,
Hao, & Jong, 2011). Liu et al. (2011) further explored the correlation
between learners’ flow experience in simulation software operations
and problem-solving strategies. However, sufficient studies
have not yet been conducted exploring the behavioral pattern analysis
of the application of simulation games with situated-learning
context in science education. Compared to conventional learning
performance assessment and self-report questionnaires, behavioral
pattern analysis can conduct a more in-depth exploration of
manipulation learning processes in scientific experiments. Presently,
analyses on learning performance, acceptance, and flow in
simulation games with situated-learning context are available
(Hou & Chou, 2012; Hou & Li, 2014) but lack a detailed examination
of these behavioral patterns. Therefore, this study utilizes a
multi-approach analysis for in-depth results.
Simulation games with situated-learning context require both
story scenarios and virtual manipulations; therefore, a wide range
of analyses of players’ behavioral patterns such as observation,
exploration, analysis, and experimental manipulation should be
conducted simultaneously in this study. Moreover, the latent
learning behavioral patterns of learners with different degrees of
flow should also be examined. By applying cluster analysis (e.g.,
Hou, 2012; Hou & Li, 2014), the potential cluster patterns of learners’
various behaviors can be explored (for example, by analyzing the overall learning process of a group of students, questions can
be raised: How many potential clusters of learners with similar
behavioral traits are being formed? What are the characteristics
of each cluster?). However, because cluster analysis can only disclose
the potential cluster patterns of learning behavior, a deeper
understanding of learning processes can be attained and behavioral
patterns can be visualized if sequential analysis is further
applied (Bakeman & Gottman, 1997; Hou, 2010a; Hou, 2010b) to
analyze the learners’ behavioral sequence patterns in each cluster
(for example, whether a behavior sequence from A behavior to B
behavior in the overall learning process of a student cluster
achieves statistical significance).