Professional Documents
Culture Documents
Problem Solving and Game-Based Learning Effects of
Problem Solving and Game-Based Learning Effects of
net/publication/228819306
CITATIONS READS
105 2,565
4 authors, including:
Bradford W. Mott
North Carolina State University
114 PUBLICATIONS 1,792 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by Hiller Spires on 01 June 2014.
Author Note
Hiller A. Spires, Friday Institute for Educational Innovation, North Carolina State
This research was supported by the National Science Foundation under Grants
recommendations expressed in this material are those of the authors and do not
Friday Institute for Educational Innovation, North Carolina State University, Raleigh, NC
Problem Solving and Game-Based Learning 2
Abstract
Targeted as a highly desired skill for contemporary work and life, problem
solving is central to game-based learning research. In this study, middle grade students
achieved significant learning gains from gameplay interactions that required solving a
science mystery based on microbiology content. Student trace data results indicated that
task was predictive of student learning and in-game performance. Students who selected
completed fewer in-game goals. Although there was no relationship observed between
providing correct explanations for hypothesis selections and learning gains, students
providing incorrect explanations completed fewer goals within the game. Finally, there
was no significant gender effect observed on the relationship between hypothesis testing
narrative-centered learning
Problem Solving and Game-Based Learning 3
In 2006, the educational research community received a strong mandate from the
how and under what conditions games can be used to maximize learning potential. One
of the assumptions that members of the Summit made was that students “acquire new
knowledge and complex skills from game play, suggesting gaming could help address
one of the nation’s most pressing needs–strengthening our system of education and
preparing workers for 21st century jobs” (Federation of American Scientists, 2006, p. 3).
Earlier work by Levy and Murnane (2004) supports this assumption. They conclude that
the nation’s challenge is to prepare youth for the high-wage/high-skilled jobs that are
rapidly growing in number—jobs that involve expert problem solving skills and complex
communication. Games are heralded as important tools for teaching an array of 21st
century skills because: 1) they accommodate various learning styles and promote a
complex decision-making context (Squire, 2006); 2) skill sets and dispositions embedded
in well-designed games are a good match for the contemporary, technology-rich worlds
students inhabit (Gee, 2003; Spires, 2008); and 3) games have the potential to promote
the 21st century skills recognized as critical for all citizens (NRC, 2010). Increasingly,
educators, private industry as well as policy makers agree that the bottom line for success
in contemporary life and work is the ability “to learn rapidly and efficiently and to go into
almost any situation and figure out what has to be learned” (Morrison, 2001). Games can
be viewed as potential tools for learning since they can simulate real-world complexity
Problem Solving and Game-Based Learning 4
and fast-paced processing in ways that traditional school learning scenarios cannot
approximate.
Empirical support for how games can improve learning outcomes for academic
content in schools has been slow to emerge (Hayes, 2005). Mayer & Johnson (2010)
conducted a recent review where they synthesized learning outcome research in three
categories: “cognitive consequences, media comparisons, and value added” (p. 246).
They concluded that the value added research, (i.e., when researchers ask which features
add value in terms of educational effectiveness of the game) is most important when the
goal is academic learning. Their research synthesis, providing a much needed utilitarian
Historically, games have been associated with play, but in light of low testing
scores and increasing rates of high school dropouts, many are looking to games for
educational benefits in schools. Pelletier (2009) has challenged educators to resist the
urge to frame games as a way to salvage education by claiming that the value of games
should be re-thought in terms of “the situated signification of ‘game’ rather than games
causing learning” (p. 83). Klopfer, Osterweil, and Salen in Moving Games Forward:
Obstacles, Opportunities, & Openness (2009), on the other hand, provide a conceptual
path for people and organizations interested in fostering the development of games for
learning purposes. Bypassing the mutual exclusion argument, they make “a case for
learning games grounded in the principles of good fun and good learning” (p. 1) and
devote their efforts to motivating and informing educators and researchers who want to
Problem Solving and Game-Based Learning 5
One of the most promising goals for games is to foster problem-solving skills
related to scientific inquiry. For example, Chris Dede and his colleagues have used multi-
to middle grade students. They argue that rather than learning by listening to lectures or
reading textbooks, students can learn science by exploring and solving problems in
realistic environments. Their early MUVE, River City, is set in a historical town in the
late 1800's, where residents are becoming ill. The students take on the role of 21st century
scientists who travel back in time in order to help the mayor identify the cause of the
illness. Kettlehut (2007) found that embedding science inquiry curricula in River City
could act as a catalyst for change in middle grade students’ self-efficacy and learning
processes. Nelson (2007) found differences in overall learning outcomes by gender and in
patterns of guidance use by boys and girls, with girls outperforming boys across a
Kamarainen, 2010).
example of an academic innovation that targets science education for 8th grade middle
school students. Taking their cues from Bruner (1990, p. 35), who observed that the way
people organize their experience and knowledge with the social world “is narrative rather
than conceptual,” the CRYSTAL ISLAND designers devised a narrative centered learning
Problem Solving and Game-Based Learning 6
environment for students to explore concepts related to microbiology (Mott & Lester,
2006).
on Valve Software’s SourceTM engine, the 3D game platform for Half-Life 2. The
Carolina’s standard course of study for eighth-grade microbiology. Students play the role
of the protagonist, Alyx, who is attempting to discover the identity and source of an
infectious disease plaguing a research station. Several of the team members have fallen
gravely ill, and it is the student’s task to discover the nature and cause of the outbreak.
recently discovered tropical island. As students explore the camp, they investigate the
and testing hypotheses. Throughout their investigations, students interact with non-player
characters offering clues and relevant microbiology facts via multimodal “dialogues”
delivered by characters through student menu choices and characters’ spoken language.
The dialogues’ content is supplemented with virtual books, posters, and other resources
they have access to a personal digital assistant to take and review notes, consult a
Problem Solving and Game-Based Learning 7
microbiology field manual, communicate with characters, and report progress in solving
the mystery. To solve the mystery, students complete a diagnostic worksheet to manage
their working hypotheses and record findings about patients’ symptoms and medical
history, as well as any findings from tests conducted in the camp’s laboratory. Once a
student enters a hypothesized diagnosis, cause of illness, and treatment plan into the
diagnosis worksheet, the findings are submitted to the camp nurse for review and possible
revision.
Over the past several years, CRYSTAL ISLAND has been the subject of multiple
based learning technologies. Previous studies have investigated the impact of narrative on
learning and presence (McQuiggan, Rowe, Lee, & Lester, 2008), as well as the
environments (Rowe, Shores, Mott, & Lester, 2010b). Recent work has explored
individual differences in students’ gameplay behaviors and learning (Rowe, Shores, Mott,
& Lester, 2010a), student note-taking practices (McQuiggan, Goth, Ha, Rowe, & Lester,
2008), and off-task behavior (Rowe, McQuiggan, Robison, & Lester, 2009). An
work devising and evaluating empathetic virtual agents (McQuiggan, Rowe, & Lester,
2008; Robison, McQuiggan, & Lester, 2009), predictive models of student affect and
self-efficacy (McQuiggan, Mott, & Lester, 2008), models for interactive narrative
Problem Solving and Game-Based Learning 8
generation (Mott and Lester, 2006a; Mott and Lester, 2006b) and models for automatic
Three theories provide the foundation for the development of CRYSTAL ISLAND
and the specific study at hand: narrative-centered learning, activity theory, and cognitive
load theory. These theories are discussed briefly in relationship to the game world of
CRYSTAL ISLAND and specifically the problem-solving task learners engage in to solve
the mystery. We propose a blending of these three theories to illustrate how within the
artifacts, and use tools to offset cognitive load during processing as they test their
based learning environments and virtual worlds by building on Gerrig’s (1993) two
transported, i.e., they are somehow taken to another place and time in a manner that is so
compelling it seems real. Second, they perform the narrative. Simulating actors in a play,
readers actively draw inferences and experience emotions prompted from interactions
with the narrative text, or what Deslandes (2004) refers to as emoting by proxy. In the
same way that good readers employ a particular stance in order to achieve their reading
Problem Solving and Game-Based Learning 9
purpose and goals, a game player may employ a stance, not unlike those in the efferent-
participate successfully in the game (Spires, Turner, Rowe, Mott & Lester, 2010). With
the recent explosion in game-based learning environments and virtual world creation,
narrative is being appropriated as a dynamic tool for exploring the structure and
processes of game-based learning related to engagement and meaning creation (Mott &
Lester, 2006).
Activity Theory
Activity theory suggests that learning is shaped by practice, with people and
artifacts mediating the learner’s relationship with reality (Vygotsky, 1978; Cole &
Engeström, 1993; Kaptelinin & Nardi, 2006). Derived from Vygotsky’s cultural-
people and cultural tools. Meaning and subsequent learning are produced in the
enactment of activity with other people and things, rather than being something confined
1987; Cole & Engeström, 1993) provides a theoretical lens for gameplay, specifically the
tools available to complete the activity and the relationships within the game that create
the interactivity leading to action and subsequent meaning. The basic triangle consists of
the following elements: subject, object, and mediating tools--all in the service of learning
outcomes. Activity theory and the meditational triangle can be used to evaluate students’
problem solving paths and specifically the architecture of how and under what conditions
Problem Solving and Game-Based Learning 10
the number of elements it can handle concomitantly, Sweller (1998) proposed cognitive
load theory, which treats schemata as the cognitive structures that make up an
structures that permit us to perceive, think, and solve problems," (p. 20) rather than
isolated facts (Sweller, 2005). These structures, referred to as schemata, are what permit
the learner to treat multiple elements as a single element. Mayer (2005) related cognitive
extraneous processing does not enhance the desired instructional objective, and in fact
CRYSTAL ISLAND. The interplay of the three theories outlined above is illustrated in
relationship to the problem-solving task, and specifically the potential hypothesis testing
line creates a sense of urgency by drawing the learner into the problem as explained by
the character Kim, the nurse, who says: “Thank goodness, you're here. I'm Kim, the
camp nurse. People on the island are getting sick, and we don’t know why. Please, can
you help us?” This interaction quickly prompts the student to a level of action.
Throughout the game Kim serves the function of increasing intensity within the narrative
as she directs the learner to possible avenues to gather pertinent information. A dialogue
branch listing the sub-activities necessary for completing the task includes: “You can
Problem Solving and Game-Based Learning 11
gather clues by talking to other team members, exploring the camp, and using the
laboratory’s testing equipment. Complete a diagnosis worksheet, then come talk to me.”
At this stage, students are free to engage the resources within the game: they may
choose to speak with other characters, explore the island and read books and posters, or
begin conducting tests in the laboratory. Activity theory is enacted in the problem space
as students become involved in shared tasks with cultural tools that potentially can
produce learning, i.e., in this context, microbiology content. The activity system
which he defined as the difference between what one could do alone and what one could
do with assistance. Additional pedagogical supports for problem solving in the character
dialogues are provided by Elise, the lab technician character. Activity theory also comes
in to play when something unanticipated happens within the problem space. For example,
some students take a scatter shot approach to testing hypotheses about what is causing the
illness rather than a strategic approach; this choice is a result of social and cultural
practices related to personal learning trajectories that students bring to the problem
solving context.
At the next stage of the game, students are encouraged to use the lab’s testing
equipment to investigate objects for contaminants, and to speak with Elise if any
contaminated objects are found. Hypothesis options include: 1) I believe this object is
believe this object is contaminated with Carcinogens. (See options in Figure 2).
Problem Solving and Game-Based Learning 12
If the students have exhausted all five of the initial ‘tests’ on the equipment, they
received a phone call from the camp nurse. The nurse offers an opportunity to earn back
more tests, but requires that they answer microbiology quiz questions in order to
demonstrate that they are able to make informed decisions about which tests to run in the
future. Students complete the diagnosis worksheet, which scaffolds problem solving by
explicitly outlining four sub-components of the game’s overall task: recorded symptoms,
laboratory testing results, beliefs about various candidate diagnoses, and a final diagnosis.
The worksheet serves as a tool to offset cognitive load as the students are making
decisions about which objects to test as part of their hypothesis strategy. In essence the
tool was designed to help reduce working memory load and increase schema construction
students should understand that the team members are sick from a pathogen, and more
specifically are suffering from a bacterial infection that was spread via the milk. Next,
they must determine a specific diagnosis and treatment plan, again using the worksheet as
knowledge about science and microbiology concepts (e.g., definition of pathogen, sizes
and structures of bacteria and viruses, treatments, scientific method, symptoms of various
Problem Solving and Game-Based Learning 13
Research Context
Since problem solving is a targeted 21st century skill that is highly sought after in
contemporary society, we were curious how well students who played CRYSTAL ISLAND
could problem solve within the game. To eliminate potential confounds and to ensure a
clear analysis with the trace data, we focused on one aspect of the problem-solving
process, namely, students’ strategies for hypothesis testing. Because of the level of
students’ strategies for testing food items for contamination and their learning outcomes.
The study addressed the following question: What is the relationship between students’
Method
153 eighth grade students ranging in age from 12 to 15 (M = 13.3, SD = 0.48). Since
eight of the participants were eliminated due to incomplete data, and eight additional
students were eliminated because they had prior experience with CRYSTAL ISLAND, the
number used for data analysis was 137 (77 males and 60 females). Approximately 3% of
the participants were American Indian or Alaska Native, 2% were Asian, 32% were
African American, 13% were Hispanic or Latino, and 50% were White. The study was
conducted prior to students’ exposure to the microbiology curriculum unit of the North
Carolina state standard course of study in their regular classes. The study took place in a
Problem Solving and Game-Based Learning 14
public magnet school, in which 35% of the students received free or reduced lunches; the
Procedure
Students had completed the majority of pre-test materials one week prior to the
intervention. Participants were randomly assigned to treatments before they entered the
intervention classroom. Each participant was seated in front of a laptop computer with
general details about the CRYSTAL ISLAND mystery and game controls. The demonstration
was conducted by one of the researchers. Second, students completed the remaining pre-
test materials. Third, students were provided with several CRYSTAL ISLAND
After the introductory session was completed, participants were given sixty
minutes to play the game and solve the mystery. Solving the mystery consisted of several
objectives, including the following: learning about pathogens, viruses, and bacteria;
compiling the symptoms and recent history of the sick researchers; recording details
about diseases believed to be afflicting the team members; and formulating and testing a
variety of hypotheses concerning sources for the disease. The final task was to report the
solution to the mystery (which included the cause, source, and treatment) to the camp
nurse. After the designated amount of time had elapsed (60 minutes) or the students had
Problem Solving and Game-Based Learning 15
content questions. The total duration of the introduction, gameplay, and post-test sessions
did not exceed 120 minutes. Students who finished before the allotted time were directed
to complete a related work sheet; all students left the testing area at the same time.
Dependent Measures
curriculum specialists. Two eighth-grade science teachers critiqued the content test to
establish content validity. There were 8 factual questions that were designed to be direct
and literal in nature and 8 application questions that required an application of knowledge
to a situation.
gameplay and logged to a database for post-hoc analysis. The logging facility recorded
millisecond intervals. The present study examined one aspect of these traces, the number
goals; however, not all students completed all the goals in the 60 minutes allotted.
Problem Solving and Game-Based Learning 16
Results
students’ hypothesis testing strategies, problem solving, and learning. The results of a
one-tailed paired samples t-test showed that students achieved significant learning gains
from their interactions with CRYSTAL ISLAND, (t(136) = 9.34, p < .0001). Students
answered an average of 2.26 more questions correctly on the post-test (M = 8.60) than the
pre-test (M = 6.34). The additional analyses are organized in three categories and
demonstrate the relationship between: 1) hypothesis testing strategies and learning gains,
2) hypothesis testing strategies and in-game performance, and 3) gender and learning
The association between in-game problem solving and learning on the content test
was investigated. In-game problem solving was measured in terms of the number of in-
game goals students completed during interactions with CRYSTAL ISLAND. Multiple
regression analysis was performed in order to determine the association between number
of goals completed and learning outcomes. The post-test score was used as the dependent
variable and both pre-test score and the number of goals completed were treated as
post-test scores, F(2, 134) = 33.17, p < .0001. R2 for the model was .33, and adjusted R2
was .32. Both pre-test score (β = .323, p < .0001) and number of goals completed (β =
Problem Solving and Game-Based Learning 17
between students’ hypothesis selection practices and learning. Students conducted tests in
the laboratory using a virtual computer terminal shown in Figure 2. Before each test,
carcinogens, and select a short justification for the test they were about to conduct. Early
in the mystery, students were expected to infer that the sick team members were suffering
from some form of pathogen that had contaminated the camp’s food supply. One measure
students should be able to eliminate before conducting any tests, and therefore an
hypotheses was calculated for each student by counting the number of tests conducted for
mutagens and carcinogens, divided by the total number of tests conducted. Multiple
regression analysis was performed with post-test score as the dependent variable, and
F(2, 124) = 15.93, p < .0001. The adjusted R2 was .19. Pre-test score had a significant
positive association (β = .382, p < .0001) with post-test score, whereas proportion of
Problem Solving and Game-Based Learning 18
with post-test score. The full regression model is shown in Table 1. The finding indicates
explanations provided for laboratory tests. Students could justify a test by selecting one
of five possible explanations: the sick team members recently ate the item, the sick team
members recently drank the item, the sick team members recently touched the item, the
item usually carries disease, or the item looks dirty (see Figure 2). The only explanations
that were consistent with the virtual characters’ dialogues were the first two; the latter
three were considered incorrect for the scenario. The proportion of incorrect
explanations was calculated for each student by counting the number of tests conducted
with incorrect explanations, divided by the total number of tests conducted. Multiple
regression analysis was performed with post-test score as the dependent variable, and
the model explained a significant proportion of the variance in post-test score, R2 = .16,
F(2, 124) = 12.02, p < .0001, only pre-test score was found to be a significant predictor
of post-test score (β = .399, p < .0001). Proportion of incorrect explanations was not a
performed with number of in-game goals completed as the dependent variable, and pre-
Problem Solving and Game-Based Learning 19
model significantly predicted number of goals completed, F(2, 124) = 8.17, p < .001. R2
for the model was .12, and the adjusted R2 was .10. Investigating each of the individual
with number of goals completed, (β = -.292, p < .001). Pre-test score had a marginally
significant positive association with number of goals completed, (β = .151, p < .08).
was number of in-game goals completed, and the independent variables were pre-test
score and proportion of incorrect explanations. The model significantly predicted number
of goals completed, F(2, 124) = 4.12, p < .02. R2 for the model was .06, and the adjusted
R2 was .05. Both independent variables were significant predictors of number of goals
completed. Pre-test score had a marginally significant positive association with the
dependent variable, (β = .169, p < .06), whereas proportion of incorrect explanations had
investigate the association between gender and post-test score. The analysis failed to
observe a significant effect of gender, t(135) = .05, p > .95. A two-tailed t-test was also
performed to investigate the association between gender and number of in-game goals
Problem Solving and Game-Based Learning 20
t(135) = 1.81, p < .08. On average, males (M = 9.47) accomplished a slightly greater
testing strategies and learning gains, as well as the models investigating hypothesis
testing strategies and in-game problem solving. However, gender was not observed to
contribute significantly to any of the four models. This suggests that gender may not be a
performance.
Discussion
The study addressed the question: What is the relationship between students’
hypothesis testing strategies and their learning gains in a game environment? Given that,
predictive of learning gains. The findings of the study revealed three families of
relationships:
performance.
Problem Solving and Game-Based Learning 21
3. Gender Effects
performance.
factor between hypothesis testing strategies and learning, and gender was not
characteristics of pathogens, mutagens and carcinogens, and have inferred that the
spreading illness on CRYSTAL ISLAND stems from a pathogen, then they should be able to
infer that they need only conduct laboratory tests for pathogen-based contaminants. This
inference requires precisely the type of content knowledge that is assessed by the post-
Problem Solving and Game-Based Learning 22
test. It is plausible that students who have the ability to apply their content knowledge to
effectively select hypotheses during laboratory tests have a stronger mastery of the
relevant content than students who are unable to apply content knowledge to effectively
select hypotheses during laboratory tests. It is therefore likely that students who are
unable to effectively select hypotheses will exhibit decreased performance on the post-
test.
knowledge about the scenario. Specifically, providing explanations for laboratory tests is
associated with students’ ability to apply knowledge about the virtual characters’ prior
histories and eating habits. Students need only conduct tests on objects from which the
virtual characters recently ate or drank. None of the characters specifically report
touching objects, and explanations about objects looking dirty or objects often carrying
diseases are weak. Knowledge about the virtual characters’ prior histories is distinct from
the content knowledge that is assessed by the post-test. One would not expect to observe
that students’ ability to apply such knowledge would transfer to the post-test and yield
The activity system within the game offered an opportunity for learning through
the provision of additional resources; specifically, reading books and posters and
interacting with knowledgeable characters using multiple modalities was available for
students to acquire and review the necessary microbiology concepts. These embedded
supports were designed to induce “zones of proximal development,” so that the students
could process the microbiology content with appropriate assistance. The level of support
Problem Solving and Game-Based Learning 23
and the incentives for taking advantage of the supports possibly were not adequate for the
interacted with CRYSTAL ISLAND for a fixed amount of time (approximately sixty
identifying the source of CRYSTAL ISLAND’s spreading illness. Identifying the source of
CRYSTAL ISLAND’S illness required students to test a specific food object for pathogens
that the sick team members recently consumed. This was the only laboratory test that
would yield a positive result for contaminants. Students who practiced less effective
hypothesis testing strategies, and were therefore less efficient at identifying the source of
the spreading illness, would have less time available to complete other goals in the game
(e.g., goals related to reporting their final diagnosis). Therefore, it is likely that less
processing of the content. As Mayer (2005) has noted extraneous processing competes
with cognitive processing resources and can be a liability when a learner is acquiring new
content material. Problem solving in CRYSTAL ISLAND relied on students' ability to apply
their knowledge about science and microbiology concepts (e.g., definition of pathogen,
sizes and structures of bacteria and viruses, treatments, scientific method, symptoms of
Problem Solving and Game-Based Learning 24
various diseases). The narrative format and accompanying features may have contributed
Gender findings. The effect of gender on in-game problem solving may actually
students had less prior game-playing experience in general, as well as less experience
with the specific game genre exemplified by CRYSTAL ISLAND (first person role-playing
game). It is also plausible that students with less prior game-playing experience would be
less adept at using the game’s controls, navigating the game environment, and
accomplishing in-game goals because they have not yet developed mature game-playing
schemata. Therefore, the effect of gender on number of in-game goals accomplished may
Because computer game playing could be linked to computer literacy, and the
assumption that computer literacy is an important 21st century skill, the gender imbalance
in computer game playing has been a research focus. Gros (2007) observed that gender
differences do not necessarily influence interest in games, but do influence different lines
of play or preferences. For example, when playing The Sims, girls devoted more time to
the decoration of the house and to the determination of the physical aspects of people;
boys initiated the game more quickly. In creating a critical discourse about games,
Problem Solving and Game-Based Learning 25
There are several limitations to this study. First, there are limitations to using
multiple choice response items to measure complex cognition and inquiry processes
within an environment like CRYSTAL ISLAND. New models for assessment are emerging
and we are interested in adapting more authentic assessments to the CRYSTAL ISLAND
learning need to focus on performance in context rather than on tests of abstracted and
isolated skills and knowledge” (p. 34). The capacity to use trace data for analysis
between learning theory and assessment method. Additionally, transfer measures will
assess how well students can apply the information learned in the game to a new context.
Second, while the game is a narrative-centered learning environment and key features of
narrative are embedded in the pedagogical apparatus, the game play of CRYSTAL ISLAND
does not approximate the action and visual engagement offered by high-end commercial
games. Students are generally accustomed to action and visual stimulation when playing
games and can be disappointed when academic games do not mimic the level of
engagement and entertainment for which they are accustomed. Third, the study was
conducted as an experimental intervention that by and large took place outside of the
instructional context of the student’s class. The curriculum content was aligned with the
North Carolina Standard Course of Study for eighth grade science and was vetted by
science teachers who are part of the project; however, due to logistical issues, students
were not able to play the game at the appropriate juncture as a supplement to enhance
their classroom instruction focused on microbiology. The timing of the study may have
Problem Solving and Game-Based Learning 26
potentially impacted the students’ motivation to engage with the academic content since
there was no apparent connection at the time of the intervention between the game and
school requirements.
Conclusion
Results indicated that the effective exploration and navigation of the hypothesis
test and completed fewer in-game goals. There was no relationship between providing
incorrect explanations for hypothesis selection and learning gains; additionally, students
providing incorrect explanations completed fewer in-game goals. Finally, there were no
Results indicate that hypothesis testing strategies play a central role in narrative-
problem solving. These results are significant in that they contribute to a growing body of
research that explores the educational benefits, both practical and theoretical, of problem
solving tasks within games in and out of school settings. Given the central role problem
solving has as a 21st century skill combined with the challenges schools encounter as they
implement problem solving curricula, CRYSTAL ISLAND offers one example of how to
engage students in narrative-centered learning that also takes into account social aspects
Future research with CRYSTAL ISLAND will focus on more in depth analyses of in-
Problem Solving and Game-Based Learning 27
pedagogical game features. Plans are underway to create the next iteration of CRYSTAL
ISLAND, including a web-based interface and additional modules related to eighth grade
science curriculum.
and for all students. As members of the 2006 Summit suggest, gaming research needs to
continue to focus on what works with whom and in which context. When the research
community adequately addresses this concern, games will become more compatible with
school learning contexts and potentially have a greater impact on developing students’
Problem Solving and Game-Based Learning 28
References
Chen, H-H., & O’Neil., H. F. (2008). A formative evaluation of the training effectiveness
Davidson, J., & Sternberg, R. (2003). (Eds.). The psychology of problem solving.
335-372.
Federation of American Scientists. (2006). Harnessing the power of video game for
Gee, J.P. (2003). What video games can teach us about literacy and learning. New York:
Palgrave-McMillan.
Problem Solving and Game-Based Learning 29
discussion. Naval Air Warfare Center Training Systems Division (No 2005-004).
Hong, N.S., Jonassen, D.H., & McGee, S. (2003). Predictors of well-structured and ill-
Kaptelinin, V. & Nardi, B. (2006). Acting with technology: Activity theory and
10.1007/s10956-006-9038-y
Klopfer, E., Osterweil, S. & Salen, K. (2009). Moving learning games forward:
Levy, F. & Murnane, R. (2004). The new division of labor: How computers are creating
Mayer, R.E. & Johnson, C. (2010). Adding instructional features that promote learning in
241-265.
Mayer, R.E. (2005). (Ed.). The Cambridge handbook of multimedia learning. New York:
Problem Solving and Game-Based Learning 30
Macmillan.
McQuiggan, S., Rowe, J., Lee, S., & Lester, J. (2008). Story-based learning: the impact
Canada.
McQuiggan, S., Goth, J., Ha, E. Y., Rowe, J., & Lester, J. (2008). Student note-taking in
McQuiggan, S., Rowe, J., & Lester, J. (2008). The effects of empathetic virtual characters
McQuiggan, S., Mott, B., & Lester, J. (2008). Modeling self-efficacy in intelligent
Miller, G.A. (1956). The magical number seven, plus or minus two: Some limits on our
Problem Solving and Game-Based Learning 31
Morrison, T. (2001). The work of nations—Interview with Robert Reich. Aurora Online:
http://aurora.icaap.org/archive/reich.html
Mott, B., Callaway, C., Zettlemoyer, L., Lee., S. and Lester, J. (1999). Towards
Mott, B., Lee, S., & Lester, J. (2006). Probabilistic goal recognition in interactive
Mott, B., & Lester, J. (2006). Narrative-centered tutorial planning for inquiry-based
National Research Council. (2010). Hilton, M. (Ed.) Exploring the intersection of science
Pelletier, C. (2009). Games and learning: What’s the connection? International Journal
Problem Solving and Game-Based Learning 32
Robison, J., McQuiggan, S., & Lester, J. (2009). Evaluating the consequences of
Rowe, J., Shores, L., Mott, B., & Lester, J. (2010). Individual differences in game play
Monterey, California.
Rowe, J., Shores, L., Mott, B., & Lester, J. (2010). Integrating learning and engagement
Pittsburgh, Pennsylvania.
Rowe, J., McQuiggan, S., Robison, J., & Lester, J.(2009). Off-task behavior in narrative-
Schafer, D.W., Hatfield, D., Svarovsky, G., Nash, P., Nulty, A., Bagley, E., Frank, K.,
Rupp, A., Mislevy, R. (2009). Epistemic network analysis: A prototype for 21st
Problem Solving and Game-Based Learning 33
Spires, H., Turner, K., Rowe, J., Mott, B., & Lester, J. (2010, April/May). Effects of
Spires, H.A. (2008). 21st century skills and serious games: Preparing the N generation. In
L.A. Annetta (Ed.), Serious educational games (pp. 13-23). Rotterdam, The
Sweller, J., (1988). Cognitive load during problem solving: Effects on learning.
University Press.
Sweller, J. (1994). Cognitive load theory, learning difficulty, and instructional design.
Warren, S., Dondlinger, M., Stein, R., & Barab, S. (2009). Educational game as
supplemental learning tool: Benefits, challenges, and tensions arising from use in
(4), 487–505.
Problem Solving and Game-Based Learning 34
Figure 1. CRYSTAL ISLAND learning environment.
Problem Solving and Game-Based Learning 35
Problem Solving and Game-Based Learning 36
Table 1
Linear Regression of Pre-Test Score and Proportion of Inappropriate
Hypotheses on Post-Test Score