CHAPTER 2 Literature Review
This chapter reviews and discusses literature relevant to our study. It is divided
into six sections: Section 1 introduces the theory of knowledge sharing and models of
knowledge transition. Section 2 introduces the instructional strategies of online
collaborative discussion. Sections 3 and 4 discuss strategies used to enhance knowledge
sharing activities: problem solving and peer-assessment and existing related studies. In
Section 5, we discuss the limitations and current status in implementing online teacher
communities. In Section 6, we discuss the current status and limitations of online
project-based learning activities. Finally, we discuss research methods for analyzing
online discussions in Section 7 and describe the methods used in this study.
2.1 Knowledge Sharing
Knowledge sharing is the process of knowledge transition among community
members. Many studies have proposed models of knowledge transition (Nonaksa &
Takeuchi, 1995; Gilbert & Gordey-Hayes, 1996; Davenport & Prusak, 1998;
Hendriks 1999). Nonaka & Takeuchi (1995) focused on the correlation between
knowledge transition and knowledge creation and discussed a model for transition,
stating that the creation of knowledge is made possible by knowledge-transition
between ontology and knowledge aspect. The former includes the individual, group,
organization, and inter-organization, while the latter includes the sub-aspects of tacit
and explicit knowledge; knowledge creation is achieved through interaction between
these two types of knowledge. The model includes the following four phases: (1)
Socialization, i.e. the transition from one tacit knowledge to another. Socialization
refers to an individual sharing experiences and the transition of mental models, not
reliant on language or words. This includes apprenticeship, observation, and imitation,
and is more focused on face-to-face observation. (2) Externalization, i.e. the transition
from tacit knowledge to explicit knowledge; knowledge is basically shared through
concepts, hypotheses, or models. For example, a person’s beliefs, know-how or
opinions are expressed via language, written materials, or images. (3) Internalization,
i.e. the transition from explicit knowledge to tacit knowledge, in which explicit
languages, documents, and images are externalized, so they can be absorbed by an
individual and turned into his or her tacit knowledge. For example, a person reads a
technical manual and understands a certain skill. (4) Combination, i.e. the transition
from explicit knowledge to explicit knowledge. Through methods such as storage,
adding, arranging, combining, classifying, and recombining, existing explicit
knowledge becomes systematic. For example, computer systems or knowledge bases
allow members to collect and combine existing explicit knowledge. With these four
aspects, knowledge transition allows socialization, externalization, internalization,
and combination, which constantly circulate and accumulate knowledge. Ontology
aspect extends from a person to the group, organization, and inter-organization, and
the process of knowledge transition forms a dynamic system.
Hendriks (1999) proposed that basic knowledge sharing includes two bodies:
knowledge owners and knowledge demanders. The former need to be willing to
“externalize”theirknowledgevialectures,writing,orothermethods, and the latter need to acquireand “internalize”theknowledgethrough listening orreading.While learning knowledge, knowledge-demanders must also reconstruct knowledge;
knowledge sharing refers to the process of knowledge internalization and
externalization.
Gilbert & Gordey-Hayes (1996) emphasized that an organization must first have
an interaction mechanism before knowledge-transition can take place. Davenport &
Prusak (1998) also pointed out the importance of using knowledge sharing strategies
to promote knowledge sharing.
Nonaksa & Takeuchi (1995) and Hendriks (1999) discussed how knowledge is
transferred; while they had their own features and focuses, they all emphasized the
importance of the two factors: knowledge “internalization” and knowledge
“externalization.”Davenport& Prusak (1998)and Gilbert& Gordey-Hayes (1996) all emphasized designing mechanisms or strategies to promote knowledge exchange
or transition.
The developmentofknowledgemanagementsystemsand a “knowledgebase”
to assist knowledge management/sharing has become pervasive (Spector, 2002; Plass,
2002; Rafaeli, et al., 2004), and appropriate use of technology promotes knowledge
interactions. However, some studies point out that many organizations believe that
technology will promote knowledge sharing (Dixon, 2000; Pfeffer & Sutton, 1999),
but, people are not necessarily willing to share their knowledge. Therefore, many
studies on knowledge sharing have discussed motives, rewards, and trust among an organization’smembers(Bock etal.,2005;Kankanhallietal.,2005;Wasko & Faraj, 2005; Hsu, et al., 2007) and the mechanism that uses technology to promote
knowledge sharing (Li, Montazemi & Yuan, 2006; Ras et al., 2005; Rafaeli et al.,
2004; Soller, 2004; Roda et al., 2003). Moreover, knowledge sharing behavior is correlated with an organization’sfeatures(Yang,2007;Yang & Chen,2007;Bock et al., 2005). Different organizations have different knowledge sharing contexts, so it is
most appropriate to design knowledge sharing activities with suitable technology based on theorganization’scharacteristics.
The above literature review shows that to promote knowledge sharing in online
teacher/learner communities, we should consider their organizational culture and
design strategies and online discussion mechanisms that promote inter-member
knowledge internalization/externalization. The strategy must also incorporate a
suitable online knowledge sharing environment to achieve the best results. This is
exactly what we are exploring in this study. In order to develop suitable knowledge
sharing activities, we review strategies that promote community collaborative
learning and integrate these strategies into activities. In the following sections, we review literature regarding “problem solving” and “peer assessment” –the two strategies frequently applied in the educational technology.
2.2 Instructional Strategies of Online Collaborative Discussion
Ever since the Internet was introduced, education paradigms and e-Learning
tools have developed rapidly for the past twenty years. With the arrival of Web 2.0
which allows users to actively participate and interact with each other (Musser &
O’Reilly,2006), the applications of the Internet are also changing, which also affects the development of educational technologies and makes e-Learning more interactive
and diverse. The interactive features of the Internet that focus on users are able to
help learners achieve knowledge construction through knowledge sharing.
The theory and paradigm of knowledge construction have been widely
discussed for the past decades (Malinowski, 1967; Mary, & Cook, 1991; von
Glaswesfeld, 1993; Allchin, 1999). Constructionists focus on the user-centered
process of knowledge construction (von Glaswesfeld, 1993; Allchin, 1999), in which
learning is achieved through knowledge sharing under social interactions (Lochhead
& Yager, 1996; Leach & Scott, 2000), and knowledge is gradually constructed
through the cognitive conflicts and consensus reached through social interactions
(Driver, et. al, 1994).
The Internet’s high interactivity helps remove the temporal and spatial
limitations in education, allows for more efficient knowledge interactions, and
promotes teaching through online collaborative learning (Chang & Chen, 1997).
There are currently many studies on online collaborative learning as well as
interactions in online discussions (Hewitt, 2005; Fahy, Crawford, & Ally, 2001;
Sudweeks & Simoff, 1999; Gunawardena, Lowe & Anderson, 1997; Newman, Webb
& Cochrane, 1995; Levin, Kim, & Riel, 1990). Many studies also suggest important
correlations between the design of online discussion mechanisms and the depth of
discussions (Patricia & Dabbagh, 2005; Hewitt, 2003; Vonderwell, 2003; Swan, at el.,
2000; Vrasidas & McIsaac, 1999). Therefore, an important topic today is how to
design a good teaching activity involving online discussions. Currently, instructional
designers can apply common interactive instructional strategies on the Internet in
order to design online discussion teaching activities. The followings are interactive
instructional strategies that are more commonly seen:
(1) Problem solving (Gagne, 1980; Mayer, 1985; Hatch, 1988; Sternberg, 1996;
Gagne & Briggs, 1979; Henna, Potter & Hagaman, 1995) : Teachers or students
ask questions and achieve knowledge exchange or construction by discussing
on the same topic and develop problem solving skills.
(2) Peer assessment (Topping, 1998; Falchikov & Goldfinch, 2000 Cizek, 1997;
Shepard, 2000 Lin, Liu, & Yuan, 2001; Sung, Chang, Chiou & Hou, 2005):
Students look at each others’work and comment on/assess on them, and this
allows them to think critically, develop cognitive skills, and construct
knowledge.
(3) Peer tutoring (Annis, 1982; Cohen, Kulik, & Kulik, 1982; Greenwood, Carta,
& Hall, 1988; Miller, 1995; Fantuzzo, King & Heller, 1992): Students guide
and tutor each other in order to reorganize and express their thoughts, allowing
them to develop expression skills and achieve knowledge construction.
(4) Role playing (Kirs, 1994; Bell, 2001): Students play the roles in certain
situations, allowing them to think based on the situations, interact, and achieve
knowledge construction.
The above-mentioned interactive strategies all have their features and can serve
as the basis for designing online discussion teaching activities. Since problem solving
and peer assessment have been widely used on the studies on online teaching for
many years, we will explore these two strategies in this study first in order to analyze
the behavioral process when they are applied in online knowledge-sharing
discussions.
2.3 Problem Solving
Problem solving is an instructional strategy frequently used in collaborative
learning (Gagne & Briggs, 1979); there have been many studies on instructional
strategies that use online problem solving methods. Different scholars define
“problem solving” in different ways (Gagne, 1980; Mayer, 1985; Hatch, 1988;
Sternberg, 1996); Gagne (1980) treats problem solving as the process of recombining previousknowledgeand solving anew problem.Mayer(1985)believesthat“problem solving” is the process of transforming an initial status to a targeted status, and proposed that problem solving is a cognitive process, the behavior of an individual
seeking a solution, and a process of using previous experience. Hatch (1988) defines
“problem solving” as the process of finding a suitable solution to a question.
Sternberg (1996)believes“problem solving”isaprocessofremoving obstacles when finding solutions.
Theabovescholars’opinionsindicatethat“problem solving”focuseson using past experience and knowledge, thinking deeply, and using cognitive skills to solve
new problems. This kind of process not only helps solve problems but can also
encourage learners to interact/discuss with peers and develop their cognitive skills
when applied in group learning settings. This is why it has long been used as a
teaching strategy (Gagne & Briggs, 1979).
Many studies propose procedures and models of problem solving. Isaksen &
Parnes (1985) describe six elements of problem solving: (1) determine the target, (2)
look for information, (3) determine the problem, (4) find the cause, (5) look for a
solution, and (6) accept the solution. Glaser & Holyoak (1986) proposed a more
detailed model: The problem-solver discovers the problem and then attempts to find a
solution. If he cannot find one, he redefines the problem, and must redefine again and
clarify the problem if the problem remains. If the solution works, he executes the plan
and verifies the results. The problem is solved when the solution executes effectively,
and the attempt is repeated until successful. The model proposed by Henna, Potter &
Hagaman (1995) defines six steps of problem solving: (1) define the problem, (2)
analyze the problem and form a hypothesis, (3) collect related data, (4) analyze,
organize, and classify the data, (5) form problem solving strategies, and (6) apply the
problem solving strategies. Sternberg (1996) proposed seven procedures of problem
solving: (1) problem identification, (2) definition of a problem, (3) constructing a
strategy for problem solving, (4) organizing information about a problem, (5)
allocation of resources, (6) monitoring problem solving, and (7) evaluating problem
solving.
To design an online problem solving knowledge sharing discussion activity that
promotes knowledge interaction, we applied previous problem-solving studies to an
online problem solving discussion context. We thus summarize the four
interaction/discussion behaviors that may emerge during online problem solving
knowledge sharing discussion activities: (1) propose, define, or clarify problems
(propose new questions, define problems, and clarify the meaning of a question); (2)
analyze the questions and give solutions (formulate strategies for solving problems,
gather data from the Internet, and share them with others); (3) compare and analyze
the more likely solutions (analyze, compare, and evaluate the comments posted in the
forum); and (4) choose the best solution and draw a conclusion (draw a conclusion
based on the analysis).
2.4 Peer-Assessment
Peer-assessment refers to the assessment conducted by a learner against his or
her peers (Topping, 1998; Falchikov & Goldfinch, 2000). This kind of strategy has
received more attention (Cizek, 1997; Shepard, 2000) as a method of class assessment
becauseitinvolvesobserving others’work and asking questions,which promotes critical thinking and meta-cognitive skills (Topping, 1998; Lin, Liu, & Yuan, 2001),
improves the quality of learning, and encourages learners to be active (Falchikov,
1995). Since peer-assessment requires evaluations and discussions between learners,
it can be used to promote knowledge sharing.
Recently, peer assessment mechanism that utilizes the Internet technology has
been widely discussed (Sung, Chang, Chiou & Hou, 2005; Lin, Liu, & Yuan, 2001).
Uploading,observing,and evaluating each other’swork via theinternet allows faster evaluations; we wish to includetheuseofweb technology in learnercommunity’s
knowledge-interactions in order to understand the process and limitations community
members face when using online peer assessment.
2.5 Online Teacher Community
Communities of organizational learning are classified as task-oriented,
knowledge-oriented, and practice-oriented (Brarb, Kling, & Gray, 2004). Since the
teacher community focuses more on sharing education-related experiences, it tends to be a “community of practice,” and may also contain elements of professional knowledge exchange. Most teachers do not exchange their experiences (Barab et al.,
2001; Tyack & Cuban, 1995), and are used to designing instructional activities in
isolation (Goodlad, 1984; Rosenholtz, 1991; Tyack & Cuban, 1995). Since
knowledge of teaching is often tacit (Carroll et al., 2003), it is difficult for teachers to
share and exchange their knowledge of teaching. As a result, they do not effectively
share their resources, hindering their teaching performance. Thus, the teacher
community plays a critical role in helping teachers exchange their knowledge.
Many studies discuss the teacher community and attempt to enhance
inter-teacher interactions by establishing interactive mechanisms or technological
interventions (Snow-Gerono, 2005; McCotter, 2001; Olson & Craig, 2001; Hsu,
2004; Carroll, Choo, Dunlap, Isenhour et al., 2003; Stigler & Hiebert, 1999; Gibson,
Neale, Carroll, & VanMetre, 1999). Due to the spatial and temporal limitations of
face-to-face community interactions, more and more studies have started using
online forums to create online teacher communities (Hobson & Smolin, 2001; Dana
& Yendol-Silva, 2003; Sing & Khine, 2006), and some communities use online
teaching films to promote knowledge sharing among teachers (Brarb, Kling, & Gray,
2004; Brarb, Barnett, Squire, 2002).
However, many studies have shown limitations in teacher interactions in
community activities, including a lack of motives, interactions, and depth (Carroll, et
al., 2003; Fishman & Pinkard, 2001; Barab, MaKinster, Moore, Cunningham, & The
ILF Design Team, 2001; Chancy-Cullen & Duffy, 1999). Studies that explore teacher communities’onlineinteractionsareoften limited to caseanalyses and interviews,
and do not offer enough information about the limitations of interactions and the
overall behavior/content pattern of teacher communities. We therefore designed
online knowledge sharing discussion activities and used a larger number of samples
to conduct an empirical analysis and explore potential limitations and solutions.
2.6 Online Project-based Learning
Project-based Learning (PBL) is a teaching method widely discussed and used
in educational technology. Originating from scholars such as Dewey, PBL focuses on
the process by which a learner defines and proposes topics, gathers and analyzes
information, communicates with others, solves problems, and shares the results (Marx,
Blumenfeld, Krajcik, & Soloway, 1997; Blumenfeld, et al., 1991; Blumenfeld, et al.,
1994; Thomas, Mergendoller & Michaelson, 1999). PBL is a representative sample of
constructive learning, and consists of: (1) inductive teaching, (2) preparing and
announcing surveys and projects, (3) interactions through collaborative learning, and
(4) the use of technology as a supplementary tool (Blumenfeld, et al., 1991;
Blumenfeld, et al., 1994). This kind of teaching strategy is pervasively integrated with
technology (Land & Greene, 2000).
Thomas (2000) pointed out that a learner in the PBL process often faces
limitations when analyzing data (Krajcik, et al., 1998; Edelson, Gordon, & Pea,
1999), as learners often fail to think deeply or interact/discuss with peers and treat
onlineresourcesdirectly as“answers”to theirproject(Wallace & Kupperman, 1997;
Chang & McDaniel, 1995). This kind of problem may lead to inappropriate
conclusions of a learning topic (Krajcik, et al., 1998), and possible causes include
the learner lacking systematic data collection and meaningful compilation (Krajcik,
et al., 1998). Another possible reason is the failure to interact with peers (Achilles &
Hoover, 1996), as this leads to inappropriate data analysis and conclusions in a PBL
setting that focuses on collaborative learning and knowledge interactions. This is
closely related to the mechanism of online discussion and interaction, and many
studies have shown that the mechanism of online discussion greatly affects the
quality of discussions (Patricia & Dabbagh, 2005; Hewitt, 2003; Vonderwell, 2003;
Swan, et al., 2000; Vrasidas & McIsaac, 1999).
A well-designed online knowledge sharing discussion activity may effectively
solve the above mentioned bottlenecks. Since PBL focuses on collaborative learning
and the use of technology, a PBL online community is formed when online discussion
is used in the process, allowing users to gather and share relevant information and
communicate with each other via online forums.
Since project-writing is a common and practical method in education settings
and online PBL can be widely applied in a blended learning context, we focus on how
knowledge sharing discussion activities promote online knowledge interaction, the
depth of data analysis, and the discussion process of online learner communities.
2.7 Analytical Methods for Online Discussions
To explore the knowledge sharing discussions of teacher communities and PBL
learner communities, we must use appropriate analytical methods. Many studies have
explored interactions in online asynchronous discussions (Hewitt, 2005; Bodzin &
Park, 2000; Fahy, Crawford, & Ally, 2001; Sudweeks & Simoff, 1999; Gunawardena,
Lowe & Anderson, 1997; Newman, Webb & Cochrane, 1995; Levin, Kim, & Riel,
1990; Zhu, 2006); some were qualitative, in that they analyzed the original protocol
of discussion content or conducted interviews or case observations, while others were
quantitative content analyses or used other quantitative analysis methods. Different
coding schemes were designed for analysis (Henri, 1992; Gunawardena, Lowe &
Anderson, 1997). The procedures of quantitative content analysis are:
(1) Formulate coding schemes: This includes existing or newly formulated coding
schemes relevant to a given research purpose. The use of existing schemes
should be encouraged if they are relevant to the same research topic to ensure
analytical validity; new ones should be prepared by reviewing and summarizing
related studies if no appropriate ones are available (Rourke & Anderson, 2004).
(2) Conduct coding process: Categorize and code each article posted on the forum
based on the coding scheme.
(3) Examine inter-rater validity: In order to ensure the validity of coded data, a
part of the data is usually coded by a second coder to measure the Kappa
reliability of coded data.
(4) Follow-up analysis: Conduct an appropriate quantitative analysis after data are
coded to explore the characteristics of the overall discussion.
Quantitative content analysis is targeted at the frequency and ratio of different
coded content, but this does not really allow us to infer behavioral patterns or
determine the sequential patterns in the learner community during online knowledge
sharing discussions (e.g., What kind of discussion behavior follows another behavior in the overall students’ interaction process? Does the continuity of eachtype of discussion sequence reach a level of significance? What kind of sequential
discussion behavioral patterns exist during the discussions?). This kind of sequential
behavioral pattern not only helps us compare the differences between actual practice
and theories, allowing us to reference prescriptive theories (e.g., problem solving or
instructional-design models proposed by previous studies), but also helps us better explore the actual process, limitations, and possible causes in students’ online knowledge sharing discussions (e.g., why did a learner carelessly reach a conclusion
or stop the discussion?). These patterns also allow us to determine possible
intervention strategies to promote further knowledge sharing; thus, behavioral
patterns deserve to be analyzed and explored.
Lag sequential analysis (Bakeman & Gottman, 1997) allows us to better determine the sequential patterns in students’ online discussions as it lets us accurately examine whether the sequential relationship between any two discussion
behaviors is statistically significant. Some studies have already applied this method
(Jeong, 2003; England, 1985; King & Roblyer, 1984; Erkens, et al., 2003). Just as in
quantitative content analysis, coding schemes are required. The following five
procedures of sequential analysis are then carried out (Bakeman & Gottman, 1997):
(1) Calculate the frequency transition matrix: The frequencies of transitions
between codes are calculated to form a frequency matrix.
(2) Calculate the sequential transition conditional probability matrix: Based on
the above matrix, we calculate the conditional probability of inter-code
transitions to generate a conditional probability matrix.
(3) Calculate the expected-value matrix: Based on the above sequential frequency
matrix, we calculate the expected-values of inter-code transitions to form an
expected-value matrix.
(4) Calculate Adjusted Residuals Table: Based on the Z-score values yielded from
the above three matrices, we examine whether the continuity of each sequence
has reached the level of significance. A Z-value greater than +1.96 indicates yes
(p<0.05).
(5) Draw the behavioral transition graph: The sequences in the Z-score matrixes
that are significant are extracted to make a correlation graph, in which each
coded behavior is a node and the arrow represents its connection directions.
The thickness of the arrow heads represents the strength of significance.
Z-score values are provided for further analysis.
Since sequential analysis is more inferential in terms of analyzing
knowledge-interaction behaviors, we not only used quantitative content analysis, but
also used it to analyze behavioral patterns in online knowledge sharing discussions.
Besides the above content and sequential analyses, we also extracted a relevant case
and analyzed its original protocol of discussion content in order to provide both
qualitative and quantitative analyses and allow cross-referencing.
