Computers & Education: So Fia B. Dias, So Fia J. Hadjileontiadou, Jos e A. Diniz, Leontios J. Hadjileontiadis

Computers & Education 107 (2017) 127e146
Contents lists available at ScienceDirect
Computers & Education

journal homepage: www.elsevier.com/locate/compedu
Computer-based concept mapping combined with learning

management system use: An explorative study under the self-
and collaborative-mode
Soa B. Dias a, *, Soa J. Hadjileontiadou b, Jose
A. Diniz a,
Leontios J. Hadjileontiadis c, d
a
Faculdade de Motricidade Humana, Universidade de Lisboa, 1495-688 Cruz Quebrada, Lisbon, Portugal
b
Hellenic Open University, Praxitelous 23, 10562 Athens, Greece
c
Department of Electrical and Computer Engineering, Aristotle University of Thessaloniki, GR-54124 Thessaloniki, Greece
d
Department of Electrical and Computer Engineering, Khalifa University, PO BOX 127788, Abu Dhabi, United Arab Emirates
a r t i c l e i n f o a b s t r a c t
Article history: The effects of the combination of computer-based concept mapping with the Learning
Received 10 March 2016 Management System (LMS) Moodle use for the construction the Concept Map (CM) in self-
Received in revised form 9 January 2017 and collaborative-mode are explored here. This approach builds upon the hybrid inter-
Accepted 16 January 2017
connection of blended (b-) and collaborative (c-) learning perspectives to form an
Available online 18 January 2017
extended teaching/learning environment. The analysis axes are based on the structural
characteristics of the constructed CMs, peers' interaction during the collaborative con-
Keywords:
struction of CMs, and LMS Moodle use expressed via the user's Quality of Interaction (QoI).
Blended/Collaborative learning
Computer-based concept mapping
One hundred and twenty-eight participants were categorized in two groups, i.e., G1 and
LMS Moodle G2, of 64 students (32 pairs) each. Both groups constructed CMs in self- (1e3 weeks) and
Self-/Collaborative-mode collaborative-mode (4e6 weeks), yet G2 only was instructed to additionally use LMS
Moodle during the whole six-week period. The experimental results comply with the
ndings from previous studies regarding with the positive effect that shifting from self- to
collaborative-mode has on the constructed CMs, but extends further the notion of the
additional positive effect of the LMS Moodle use in the collaborative construction of the
CMs, as expressed via more increased peers' turn-taking, balanced collaboration, and CM-
enabled constructive, reective and organizational interactions. It is also shown that
collaboration increases LMS Moodle users' QoI. The hybrid b-/c-learning approach pro-
posed here sets new directions towards the enhancement of LMS use and computer-based
concept mapping, contributing to the enrichment of the Higher Educational Institutions
services and re-examination of educational practices.
2017 Elsevier Ltd. All rights reserved.
* Corresponding author.
E-mail addresses: sbalula@fmh.ulisboa.pt (S.B. Dias), shadjileontiadou@gmail.com (S.J. Hadjileontiadou), jadiniz@fmh.ulisboa.pt (J.A. Diniz), leontios@
auth.gr (L.J. Hadjileontiadis).
http://dx.doi.org/10.1016/j.compedu.2017.01.009
0360-1315/ 2017 Elsevier Ltd. All rights reserved.
128 S.B. Dias et al. / Computers & Education 107 (2017) 127e146
1. Introduction
Computer-based learning (CBL) systems are being deployed in many Higher Education Institutions (HEIs), suggesting
the development of means that bring learners together, reinforcing creative activities towards intellectual exploration,
metacognition and social interaction. In several contexts, CBL is interpreted as blended (b-)learning experiences through
the mediation of Information and Communication Technologies (ICTs), rather than being completely online or Face-to-
Face (F2F) (Clouder et al., 2006; Michinov & Michinov, 2008). An important issue in b-learning, when it is situated in
CBL environments (CBLEs), is how to manage the two modalities, i.e., F2F and online, to better support collaboration and
to create more holistic/integrated learning experiences. The option for a b-learning structure is justied by its exibility,
ease of access, and the possibility of integration of sophisticated and personalized technologies (Johnson, Adams Becker,
Estrada, & Freeman, 2015). In this line, Learning Management Systems (LMSs) have become very popular among edu-
cators to improve teaching-learning processes, incorporating tools, such as discussion forums, Wikis and other interactive
tools that make them especially useful for constructing enriched learning environments (Dias, Diniz, & Hadjileontiadis,
2014). At the same time, several studies have clearly demonstrated the efcacy of computer-oriented concept mapping
techniques in supporting the learning process (Bridges, Corbet, & Chan, 2015; Omar, 2015). In particular, concept-maps
(CMs) are exible cognitive tools that engage students in the learning process, promoting activities that guide stu-
dents to read, think, organize, relate with prior knowledge, draw, revise, reect, rebuilt, and communicate (Novak &
Can~ as, 2006). From this perspective, important research has focused on the micro level of collaborative learning, i.e.,
single/small groups; however, to support collaborative learning and knowledge building, a more detailed understanding
of how small groups of learners construct shared meaning using various tools/systems should be considered. Educators
who understand the advantages of b-learning can improve the impact of collaborative (c-) learning within their class-
rooms. In this way, advances within the eld of CBL will depend on better understanding and use of the hybrid of the (b-/
c-) educational approaches. The context of this work is placed exactly within this perspective, trying to explore the use of
concept mapping activities combined with the LMS use under the b-/c-learning concept, as it becomes evident from the
succeeding subsections.
1.1. Computer-based learning environments (CBLEs)
CBLEs assist individuals in learning using multiple representations of information for a specic educational purpose
(Ifenthaler, 2012). CBLEs frequently confront learners with a number of support devices (also referred as tools) in order to
enhance learning, to help learners in their learning and to provide a learning opportunity (Collazo, Elen, & Clarebout,

2015; Garcia-Alvarez,
Suarez Alvarez, & Quiroga Garcia, 2014). Nowadays, a LMS plays a central role in any online
learning environment; in particular, the LMS Moodle is a free, widely used, open-source software that supports online
learning based on a social constructionist framework of education (Seluakumaran, Jusof, Ismail, & Husain, 2011). So far,
limited efforts have been made to understand the development and use of theory in the particular domain of b-learning
research (Drysdale, Graham, Spring, & Halverson, 2013; Graham, 2013). The concept of b-learning is embedded in the idea
that learning is not just a one-time episode, but also a continuous/dynamic learning process. Blending different delivery
modes/tools can be seen as an imaginative solution in educational contexts, since it has the potential to balance out and
optimize the learning development (Dias et al., 2014). However, according to Bates and Sangra (2011): Teachers must
decide which tools are most likely to suit the particular teaching approach (pp. 44e46).
1.2. Concept mapping in CBLEs
CMs are considered effective as teaching and learning tools that assist the development of conceptual knowledge,
allowing visual observation of relationships and connections between multiple areas and pieces of information (Novak &
Gowin, 1984). Actually, the ability to recognize connections between different pieces of information or aspects of a
problem facilitates problem-based learning (PBL) (Schaal, 2010). The latter assists the development of higher order
thinking skills, helping students to become independent, self-directed learners who appropriately respond to situations
in a logical and reasonable manner (Savery & Duffy, 1995). According to Novak (2010), a CM is a (hierarchical) network,
comprised of concept-terms (nodes) and directed lines, accompanied with phrases, linking pair of nodes (so as two nodes
with the linking phrase to constitute a correct logical proposition) that lend to the network a propositional character; at
the same time, CMs provide a window into students' mind, reecting students' knowledge structures. Seen as an
instructional tool, CM encourages students to explicitly organize and make public their knowledge. Concept mapping has

been described as a technique that can increase student's learning in the traditional classroom (Alvarez-Montero, Saenz-
rez, & Vaquero-Sa
Pe nchez, 2015; Novak & Can ~ as, 2008).
Similarly, Clariana and Taricani (2010) have been researching the use of CMs in an online learning environment,
including a computer-based approach for scoring the quality of the resultant CMs and their potential as classroom study
aids. In general, the CM quality refers to the amount, depth, and breadth of information and the number of connections
made among different items included in CMs (Gurupur, Jain, & Rudraraju, 2015). In fact, a relationship between the
quality of the CM and the total acquisition of information can be identied. In other words, students that take more time
S.B. Dias et al. / Computers & Education 107 (2017) 127e146 129
in the construction of their maps and include more detailed information tend to retain more information for longer
periods of time (Clariana & Taricani, 2010).
1.3. Self- and collaborative-mode CMs
Several studies have investigated the use/potential of (online) CMs to supporting processes of self-knowledge man-
agement (Conceia ~o, Desnoyers, & Baldor, 2008; Tergan, 2005; Tergan, Keller, Gra
ber, & Neumann, 2006; Vodovozov &
Raud, 2015). Other authors, on the other hand, have investigated that collaborative CMs has the potential of facili-
tating knowledge construction as a study/collaborative tool (Gao, Thomson, & Shen, 2013; Koc, 2012; Lee, 2013; Lin,
Wong, & Shao, 2012; Molinari, 2015; Rafaeli & Kent, 2015). Although originally developed to assist individual learners,
collaborative use of CMs emphasizes brainstorming among group members, leading to visualization of new ideas and
synthesis of unique concepts (Novak, 2010). With regard to the use of CMs for educational purposes, the following ex-
amples of research studies based on individual mapping vs. collaborative mapping can be considered:
Kwon and Cifuentes (2007) revealed that individually generating CMs on computers is more effective than either inde-
pendent, unguided study, or collaboratively generating CMs. In addition, students in both individual and collaborative
concept mapping groups had positive attitudes toward concept mapping using Inspiration software. Findings indicate that
teachers should train their students in computer-based concept mapping and facilitate adoption of concept-mapping as an
independent study strategy.
Coutinho (2009), unlike the ndings of Kwon and Cifuentes (2007), stated that working in a team to develop a CM is more
benecial than working by oneself; the results have shown that the interaction in teams further helped the group in
developing their understanding of the content under study. Moreover, the comparison of the same CMs on a specic
theme designed by groups with those designed by individuals shown statistically signicant difference. The scores from
the collaboratively constructed CMs compared to the ones from each CM indicated statistically signicant improvement,
showing greater understanding of the content and higher processing of related ideas as students pulled their knowledge
together.
Kwon and Cifuentes (2009) reported that constructing/sharing a CM with others requires communication/negotiation
processes, guiding learners to growth in their conceptual understanding. Additionally, the collaborative process and the
high level of social interaction increased each student's knowledge, which would not have been possible if they would
have worked on their own.
Hwang, Shi, and Chu (2011) supported that collaborative CM achieves higher learning results. In particular, in the post-
test results, the students who used collaborative online CMs revealed signicantly better learning achievement than the
students who learned the same materials with other methods. Improved students' attitudes towards science learning,
improved condence in their peers and higher expectations of collaboratively learning were also reported. Moreover,
the collaborative work encouraged students' engagement and self-efcacy in learning, as well as their motivation to
communicate/collaborate with their peers.
1.4. Study aims and research questions
From the aforementioned, it is evident that although the benets from the collaborative construction of CMs are already
identied, no research effort has explored the effect of the concept mapping placed within the b-/c-learning environment, by
examining the alterations of the CM characteristics when the LMS use is also considered both in self- and collaborative-mode,
which denes that aim of this exploratory study. The analysis axes are based on the structural characteristics of the con-
structed CM and the analysis of CMs based peers' interaction during the collaborative mode. To this end, the following
research questions (RQ) emerge:
RQ1: How shifting from not using to using LMS Moodle affects the quality of CM in self- and collaborative-mode?
RQ2: How shifting from not using to using LMS Moodle affects the CM-based collaboration? and
RQ3: How shifting from self- to collaborative-mode affects the structural characteristics of CM and the LMS Moodle Quality of
Interaction (QoI)?
The rest of the paper is organized as follows: rst, a description of the proposed approach is presented, followed by a
description of the proposed metrics. Next, a description of the experimental validation/implementation issues, analysis of the
derived results, along with discussion and interpretation of the ndings are provided. Then, the overall perspective (limi-
tations and implications) are critically discussed. Finally, conclusions and future work conclude the paper.
2. Methodology
2.1. Rationale
The hybrid concept of b-/c-learning adopted here is in line with the European Union's perspective about the trends and
evolution of Higher Education in the next years. More specically, in the 2014 Horizon Report on Higher Education it is
mentioned that:
Education paradigms are shifting to include more online learning, blended and hybrid learning, and collaborative
models. () Hybrid models, when designed and implemented successfully, enable students to travel to campus for
some activities, while using the network for others, taking advantage of the best of both environments. (Johnson,
Adams Becker, Estrada, & Freeman, 2014, p. 10).
In fact, an increasing number of HEIs are incorporating online environments into courses of all kinds, which is making the
content more dynamic, exible, and accessible to a larger number of students. Merging these settings transforms them to
hybrid-learning ones, engaging students in creative learning activities that often demand more peer-to-peer collaboration
than traditional courses (Hadjileontiadou, Dias, Diniz, & Hadjileontiadis, 2015).
Taking into account the aforementioned perspectives, an alternative research framework is proposed here, by exploring
the ways effective teaching could be accomplished, when bridging the elds of b- and c-learning into a hybrid and enhanced
teaching-learning environment. The rationale behind is that a holistic approach of the fundamental channels from which the
educational process is conveyed could be adopted, combining cognitive and social information of the peers' behavior and
interactions. In other words, a hybrid approach is proposed and discussed here, that may serve, not only for synthesizing and
representing knowledge obtained from the data, but also for exploring possible future online learning environment (OLE)
states/scenarios. Pragmatically, this paper explores the peers' collaborative activities within a computer-based concept
mapping combined with their QoI with a LMS, both involving modeling techniques that are drawn from interaction data, as
vehicles to improve the personalization and intelligence of an OLE. Furthermore, the estimated parameters related to the
structural characteristics of the derived CMs, along with the ones related to peers' collaboration and QoI, can form a combined
basis for a more pragmatic approach of OLEs and b-/c-learning environments, within the context of Higher Education. They
can also provide dynamic feedback that accounts for the non-stationarities seen in students' learning process, capturing in a
more pragmatic way the underlying interaction/collaboration trend-shifts and/or artifacts.
2.2. The proposed analysis framework
Fig. 1 illustrates a schematic representation of the proposed analysis framework. According to Fig. 1, two modes are
employed, i.e., self- and collaborative-mode, denoted as SELF- and COLL-MODE, respectively, when realizing the CM. The
switches allow for the possible combinations considered, leading to the corresponding outputted parameters. The latter
include:
1. SELF-MODE parameters related with the structural characteristics of the actual CM (example of Peer Si in Fig. 1). These are:
a. Topological Taxonomy Score (TaxScoreSELFMODE ), which ranges from 0 to 6 and it is calculated according to ve criteria
dened in (Novak & Can ~ as, 2006), i.e.: a) use of concepts rather than of chunks of text, b) establishment of relationships
between concepts, c) degree of branching, d) hierarchical depth, and e) the presence of crosslinks. Higher topological
taxonomy scores typically indicate higher quality concept maps (Novak & Can ~ as, 2006).
b. Set of CM-based structural elements, which relate with CM construction (CON), i.e., Add, Move and Connect, expression of
user's reection (REF), i.e., Delete, Resize and Modify, and CM organization (ORG), i.e., Concept, Linking Phrase, denoted
in Fig. 1 as the fCON; REF; ORGg set.
2. COLL-MODE parameters related with the structural characteristics of the collaboratively produced CM, reecting also the
characteristics of peers' collaboration and transition from SELF- to COLL-MODE (example of Pair (Si ; Sj ) in Fig. 1). These are:
a. Topological Taxonomy Score Difference (TaxScoreDiff ) (see Appendix A-A1), which considers the difference between the
topological taxonomy score of the collaboratively produced CM from the pair (Si ; Sj ) and the lowest topological tax-
onomy score of the individually constructed CMs by Si and Sj , expressing, thus, the maximum level of improvement in
the topological taxonomy score when shifting from the SELF- to COLL-MODE.
b. fCON; REF; ORGg set, as in the SELF-MODE case, yet here it is derived from the collaboratively produced CM.
c. Turn-taking TTCOLLMODE between the peers Si and Sj across their collaboration during the construction of the
collaboratively produced CM. The TTCOLLMODE takes into account all the alterations between the peers' active role
(mouse control), when producing the fCON; REF; ORGg set. This parameter serves as an indicator of the way that the
collaboration unfolds between the collaborators through the turn in the submission of their computer mediated in-
teractions (Hadjileontiadou & Hadjileontiadis, 2004; Hadjileontiadou et al., 2015).
d. Absolute difference of the peers' balance (BalDiff (see Appendix A-A2), that takes into account the number of
fCON; REF; ORGg set contributions of each peer, normalized to the total number of the fCON; REF; ORGg set
Fig. 1. Schematic representation of the proposed analysis framework. The employed switches allow for possible transitions from SELF- to COLL-MODE.
contributions in the pair (see Appendix A-A3). This parameter is adopted as an indicator of a preferred balanced
collaborative work as opposed to diverging individual ones (Hadjileontiadou & Hadjileontiadis, 2004; Hadjileontiadou
et al., 2015).
3. SELF-/COLL-MODE parameters related with the LMS Moodle users' QoI, both in the SELF- and COLL-MODE of the CM
construction, based on the FuzzyQoI model (Dias & Diniz, 2013), outputting the QoI of Peers Si and Sj , assuming that they
participate in the corresponding group (G2) that enables the LMS Moodle use (refer to x3.1) (LMS switches close in Fig. 1).
Epigrammatically, the FuzzyQoI model estimates the QoI of the students within the LMS Moodle b-learning context, by
fuzzifying the input from 110 LMS Moodle interactions pre-categorized into 14 basic categories (C1-C14), namely: C1:
{Journal/Wiki/Blog/Form (J/W/B/F)}, C2: {Forum/Discussion/Chat (F/D/C)}, C3: {Submission/Report/Quiz/Feedback (S/R/Q/
F)}, C4: {Course Page (CP)}, C5: {Module (M)}, C6: {Post/Activity (P/A)}, C7: {Resource/Assignment (R/A)}, C8: {Label (L)},
C9: {Upload (UP)}, C10: {Update (U)}, C11: {Assign (A)}, C12: {Edit/Delete (E/D)}, C13: {Time Period (TP)}, and C14:
{Engagement Time (ET)}. Using the expert's knowledge base, as expressed through 600 fuzzy rules, the fuzzy-logic based
aggregation of all categorized interactions (C1-C14) outputs (via a defuzzication process) the nal estimation of the user's
QoI with the LMS Moodle (Dias & Diniz, 2013).
3. Experimental and implementation issues
3.1. Participants
One hundred and twenty-eight pre-service vocational education teachers undertaking an one-year Pedagogical Training
Program, completed the entire study, which lasted six weeks (W1:W6), in the second semester. All the trainees voluntarily
participated in this study and the identities of the students were kept condential.
To avoid the potential extraneous factors of vocational specialty and gender in the experiment, the following procedure
was followed. Initially, clusters of vocational specialties were formed for the male and the female participants. Then, 65
female and 63 male participants were matched on gender in two groups, i.e., G1 and G2, of 64 students (32 pairs) each,
(31(G1)/32(G2)-63male and 33(G1)/32(G2)-65 female). The matching occurred on successive assigns of the randomly listed
male participants from each cluster, e.g., the rst to the G1 the second to the G2 and so on for all the clusters. The same
procedure was followed for the female participants too. Both groups performed CM in both modes, i.e., SELF-MODE (W1-W3)
and COLL-MODE (W4-W6), yet G2 only was instructed to additionally use LMS Moodle during the whole period (W1-W6)
(see Table 1).
The participants varied little in other personal characteristics that might inuence the experiment, (age 28 2.7 yrs, all
Greeks and graduates from Greek universities). All of them had experience in diagrammatic depictions of concepts (yet,
without the use of linking phrases that form a logical proposition with specic meaning) and the use of computer tools used
in their specialties (equally distributed experiences in the G1 and G2), yet none of them of CM, and the of use of a CM-related
software, such as the CmapTools (either in the self- or in collaborative mode), and the Moodle LMS. Moreover, none of them
had any experience concerning computer-mediated collaboration.
Table 1
The groups (G1 and G2) involved in the present study considering the individual and collaborative modes, as well as the using of the CmapTool and
the LMS Moodle distributed across the six-week (W1:W6) period of the experimental study.
Groups CmapTool use LMS Moodle use

(W1:W6)
SELF-MODE COLL-MODE
(W1:W3)a (W4:W6)
Group 1 (G1) 64 (31M; 33F)b 32 (pairs) NO
Group 2 (G2) 64 (32M; 32F) 32 (pairs) YES
a
W: Week.
b
M: Male; F: Female.
The participants within each group were randomly matched from different clusters. In order to succeed this matching, a
slight difference in the male/female distribution among the G1 and G2 was selected (63/65). This effort was put in order to
handle the homogeneity of the collaborating groups, i.e., in all cases pairs of different specialties. No prior tests or scores were
considered in order to detect high and low achieving participants so to assign one high to one low achieving participant to the
same pair, as towards this purpose the aforementioned metrics were used as the independent variables of the RQs.
At the beginning of the second semester, all the participants were asked to write an essay based on two questions: a) what
is modeling, and b) how can it be connected to the conceptual understanding of new knowledge. In general, from the content
analysis of their texts, it was revealed that in the question (a), all of them described a model as an external tool and in the
question (b), they focused again in the merits of its use for teaching and learning, adopting a poor approach to its contribution
to the conceptual understanding of new knowledge. These results were more or less expected, since all the participants
graduated from vocational studies without any pedagogical focus. Upon these ndings, the background of the participants
was considered homogenous in relation to the text that was given to them, in order to transcribe it to CM in SELF- and COLL-
MODE, as it is presented in Appendix B.
Finally, the potential differences in teaching were controlled, as the same researcher (second author) performed the
training and the experimental procedures.
3.2. Implementation
The implementation took place according to the experimental design. The advantage of online mapping tools, like the
CmapTools (http://cmap.ihmc.us) adopted here, is the collaborative element, as it allows other users and oneself access to the
constructed CMs from anywhere/anytime, allowing to work in pair or teams on them (Hanewald & Ifenthaler, 2014). One
basic feature of the CmapTools software is the recording/replaying of the construction procedure of the CM in both modes, i.e.,
SELF- and COLL-MODE. Moreover, upon this feature, the extraction of a .txt log le is facilitated that presents time-stamped,
all the interactions that were performed by its author/s, i.e., the produced fCON; REF; ORGg set contributions. In both groups,
the indexing adopted in the analysis followed the structure of consecutive students being corresponded to the consecutive
indexing of pairs, e.g., students 1 and 2 formed the pair 1, students 3 and 4 formed the pair 2, and so on. In this way, the
association between SELF- and COLL-MODE results is facilitated.
The design of the online course was organized in the three-modules approach. In particular, the module-0 refers to the
introductory phase of acquainting the G2 students with the LMS Moodle environment. Apart from the introductory phase, the
learning activities (which is the focus of the analysis here) included module-1 (W1-W3), which refers to the construction of
the SELF-MODE CM, and module-2 (W4-W6), which refers to the construction of the COLL-MODE CM. The content in the LMS
Moodle course was organized in the way of including the text in Appendix B, which was just e-mailed to the participants of
the G1, and uploaded to the LMS Moodle for the participants of the G2. Apart from this text, a resource in the LMS Moodle
referring to a manual for LMS Moodle use was also uploaded, combined with different assignments and a timetable of the
important dates for the deliverables. In particular, the assignments included: (i) an initial acquaintance of the G2 students
with the environment of the Moodle (F2F situation) (module-0), (ii) a requirement from them to update their prole and to
upload a picture of themselves (CBL situation) (module-0), (iii) the construction of the SELF-MODE CM (module-1) and (iv)
the construction of the COLL-MODE CM (module-2).
The LMS Moodle was also prepared from the beginning to provide to its G2 users spaces for interaction that could trigger
metrics in all the 14 basic categories (C1-C14), that are presented above for the measurement of the QoI via the FuzzyQoI
model (Dias & Diniz, 2013).
The SELF- and COLL-MODE work was distantly performed from the place of each participant. In both modes, the par-
ticipants were asked to record the construction of the CM. In the case of the SELF-MODE, each participant worked in his/her
desktop and upon the end of the W3 s/he submitted to the researcher both the .cmap le and the .txt le with the logged raw
data that are exported on the basis of the recording. In the COLL-MODE, the IHMC CmapTools server was used (Shared Cmaps
in Places). In particular, a structure of a password secured folder and common workspace for each pair of the G1 and G2 was
prepared. Upon logging to the common workspace, each collaborator interacted with it in order to contribute to the common
work. The CM refreshed every time that an interaction took place and in this way the common CM was constructed. Hence,
during the collaboration, only the working CM was depicted in the common workspace. The researcher extracted from the
collaborative CMs the .cmap le and the relevant .txt le.
All the adult participants agreed that across the duration of the experiment they were not to use any extra reading apart
from Appendix B. To control the collaboration environment, the G2 students were instructed and all have agreed upon not to
communicate with Skype or Discus, or even emailing each other, i.e., to only use the LMS Moodle for communication con-
cerning their assignments. In parallel, they were instructed and agreed upon to use the IHMC CmapTools server (Shared
Cmaps in Places) as the only collaboration path for the construction of the CMs. In particular, they were instructed to shift to
the LMS Moodle for communication, and not to use the integrated communicative tools in the CmapTools e.g., chat.
Considering the voluntarily participation of the students, following of the above learning contract was considered as an action
upon the highest level of trust.
During the F2F communication that took place in a weekly basis (three hours per week), all the participants were sup-
ported by the researcher. More specically, every clarication that was asked for the use of the CmapTool was provided to
both G1 and G2 and for the use of the LMS Moodle only to the G2. The week was selected as an analysis basis to be coherent
with the time unit involved in the estimation of the QoI by the FuzzyQoI model (Dias & Diniz, 2013). Furthermore, it should be
noted that a possibility to submit up to three CMs (for possible corrections until the nal version) was also provided to the
students, in order to relax them and yet to avoid possible overtraining upon the CMs construction through their multiple
submissions. However, all of the participants delivered only the nal version of the CM in the COLL-MODE and only less than
3% submitted two CMs and less than 1% submitted three CMs in the SELF-MODE.
During the F2F communication that took place in a weekly basis, all the participants were supported by the researcher. To
avoid any erroneous CMs, in terms of their content, to be included in the analysis, an automated comparative analysis be-
tween the outputted CMs with a list of correct propositions and main concepts (dened by the expert and related with the
main content of the given text), which should not be absent in the outputted CMs, took place. The latter was realized via the
Cmapanalysis (Can ~ as, Bunch, Novak, & Reiska, 2013) plugin by the construction of a corresponding .xml le with the expert's
CM data and served as a baseline test. All the produced CMs have successfully passed this baseline test and were included in
the subsequent analysis. This was a quite expected result, as the construction of the CMs took place within the setting of a HEI;
hence, it was most unlikely from the students' side to intentionally deliver nonsense or erroneous CMs.
3.3. Data analysis tools
Data acquired from CmapTools software and LMS Moodle use set the experimental corpus of the present study. For the
estimation of the parameters dened in x2.2, the Cmapanalysis (Can ~ as et al., 2013) plugin, along with custom-made pro-
gramming code in Matlab 2015a (The Mathworks, Inc., Natick, USA), were used. The archived data in the LMS Moodle re-
pository were exported from .xml to .xlsm (Microsoft Excel format) and imported to the Matlab environment and archived as
.mat les. The .s structures of FuzzyQoI dened in (Dias & Diniz, 2013) were employed for the estimation of the LMS Moodle
users' QoI. For the between-subjects (G1 vs. G2) statistical analysis the one-way analyses of variance (ANOVA test) was
employed, whereas for the within-subjects (SELF-MODE vs. COLL-MODE) statistical analysis the two-sided Wilcoxon rank
sum test was used, both implemented in Matlab 2015a. The programming code developed for the study is freely available
upon request to the authors.
The results from the analysis of the aforementioned CMs and LMS Moodle data are described and discussed in the suc-
ceeding subsections.
4. Results and discussion
4.1. RQ1: The effect of LMS Moodle use on the quality of CMs
In order to exemplify the effect of the LMS Moodle use on the quality of the constructed CMs, two examples of constructed
CMs are given in Fig. 2, both under the COLL-MODE, corresponding to a pair of G1 (Fig. 2(a)) and a pair of G2 (Fig. 2(b)),
respectively. In particular, in Fig. 2(a), the CM of pair (P5: S9,S10) from G1 is illustrated, with corresponding values of
S9G1 S10G1 P5: S9;S10G1 P5G1
TaxScoreSELFMODE 1, TaxScoreSELFMODE 2, TaxScoreCOLLMODE 4, TaxScoreDiff 3. Furthermore, in Fig. 2(b) the CM
S45G2 S46G2
of pair (P23: S45,S46) from G2 is depicted, with corresponding values of TaxScoreSELFMODE 0, TaxScoreSELFMODE 0,
P23:S45;S46G2 P23G2
TaxScoreCOLLMODE 6, TaxScoreDiff 6. Clearly, in this case, the shift from SELF- to COLL-MODE had a positive effect
in the quality of the constructed CMs, as reected in the increase of the topological taxonomy scores in both G1 and G2
groups, complying with the ndings of (Kwon & Cifuentes, 2009). What is more profound is that here, the use of LMS Moodle
supported further this increase of the topological taxonomy score, as the corresponding TaxScoreDiff was greater for the case
of G2 compared to the one of G1.
To further examine the generalization of the aforementioned observation, the TaxScoreDiff values for both G1 and G2, i.e.,
G1;G2
TaxScoreDiff , across all pairs were calculated via (1) and depicted in Fig. 3. The results presented in Fig. 3 justify the
observation of Fig. 2, as for the case of G1 the shifting from SELF- to COLL-MODE has in general produced positive TaxScoreG1
Diff
values, yet with some negative ones (6 out 32) and some equal to 0 (8 out of 32). The TaxScoreG2
Diff
values, however, are all
positive and all 2, showing the benecial effect of the LMS use in the quality of the collaboratively constructed CMs.
When approaching the same concept from the SELF-MODE perspective, there is a similar behavior in the resulted
TaxScoreSELFMODE values between the G1 and G2 groups. This is illustrated in Fig. 4, where the estimated TaxScoreG1;G2
SELFMODE
values across all students per group are shown. From the latter it is concluded that LMS Moodle use did not affect the quality
of the CM construction reected in the relevant topological taxonomy score under the SELF-MODE.
The abovementioned results are also statistically justied as shown in Table 2. From the latter, a statistically signicant
difference between the TaxScoreG1 Diff
and TaxScoreG2
Diff
was found (p 4$109 ), whereas no statistically signicant difference
was found between the TaxScoreG1 G2
SELFMODE and TaxScoreSELFMODE p 0:3398.
4.2. RQ2: The effect of LMS Moodle use on the CM-based collaboration
Adopting the case of COLL-MODE in both G1 and G2 groups, the TTCOLLMODE values across the pairs of each group, i.e.,
G1;G2
TTCOLLMODE , were estimated and illustrated in Fig. 5. From the latter, in almost all cases (exception of 4 pairs out of 32) the
G2
TTCOLLMODE G1
values were greater than the TTCOLLMODE G2
ones, exhibiting a mean value of TTCOLLMODE almost three times higher
G1
than the one of the TTCOLLMODE . This difference is also statistically justied, as a statistically signicant difference between
G1
TTCOLLMODE G2
and TTCOLLMODE values was found p 1:79$107 , as shown in Table 2. This implies that the collaboration in the
G2 group was more active from both peers, compared to the one in the G1 group, showing that in most cases the employment
of the LMS Moodle use triggered further both peers to participate in the collaborative activities during the collaborative
construction of the CM.
The aforementioned result is further justied via the corresponding BalDiff values, estimated, via (2) and (3), across the
pairs of both G1 and G2 groups, i.e., BalG1;G2
Diff
, illustrated in Fig. 6. From the latter it is evident that the pairs of G2 group
exhibited more balanced collaboration compared to the ones from G1 group, as the BalG2
Diff
values are always less than the
BalG1
Diff
ones, lying at a mean value around 15% in contrast to the mean value of BalG1
Diff
that lies around 30%. This is also sta-
tistically justied, as a statistically signicant difference between BalG1
Diff
and BalG2
Diff
was found p 9:5$1019 , as it is shown
in Table 2. These results support the perspective that the LMS Moodle use potentially contributes to the avoidance of any
possible domination of one peer to another within the pair, in terms of more balanced collaboration during the collaborative
construction of the CM.
The derived fCON; REF; ORGg set in both SELF- and COLL-MODE for G1 and G2 groups are depicted in Figs. 7e9. In all
gures, the left column corresponds to the fCON; REF; ORGg set from SELF-MODE, whereas the right one to the
fCON; REF; ORGg set from COLL-MODE; moreover, the thin line denotes the G1 group whereas the thick one denotes the G2
group. As it is apparent from Figs. 7e9, there is a difference in the structural elements of the CMs constructed both in SLEF-
and COLL-MODE between the G1 and G2 groups. This is justied by the statistical analysis tabulated in Table 2. In particular, a
statistically difference is identied in all cases of COLL-MODE between the G1 and G2 groups
(CON fAdd; Move; Connectg : p 8:79$108 , p 9:14$106 ; p 1:25$104 ; REF {Delete; Resize; ; Modify}:
p 1:66$10 11 , p 0:0223; p 2:58$10 ; ORG fConcept; Linking Phraseg : p 9:11$109 ; p 1:86$106 ), whereas only
4
in the case of CON fMoveg : p 0:0219, REF fDelete; Resizeg : p 0:0158, p 0:0400; and
ORG fConceptg : p 0:0048; of the SELF-MODE the statistically difference is observed. These results denote that the LMS
use mostly contributes to the enhancement of the CM structural elements under the COLL-MODE and in part under the SELF-
MODE.
4.3. RQ3: The effect of shifting from SELF- to COLL-MODE upon the structural characteristics of CM and the LMS Moodle QoI
In order to respond to RQ3, a within-subjects statistical analysis between the SELF- and COLL-MODE of the derived
fCON; REF; ORGg set was conducted, with the derived results tabulated in Table 3. This is also evident from the comparison
between the SELF- and COLL-MODE in Figs. 7e9 (left vs. right columns). As it is evident from the results of Table 3 and Figs.
7e9, the shifting from SELF- to COLL-MODE created statistically signicant difference in all cases of fCON; REF; ORGg set for
the G1 group, and similarly in all cases of fCON; REF; ORGg set, except CON fMoveg and REF fDeleteg for the G2 group.
These results denote that the LMS Moodle use does not signicantly have any additional effect to the structural characteristics
of CMs, expressed through the fCON; REF; ORGg set, when shifting from SELF- to COLL-MODE, as this shift is almost equally
important within both G1 and G2 groups.
Finally, Fig. 10 depicts the estimated mean QoI when shifting from SELF-, i.e., QoIW1:W3G2 G2
, to COLL-MODE, i.e., QoIW4:W6 , for
each student of G2 group. As it is clear from Fig. 10, there is a distinct improvement in the QoI when the students of G2 started
G2
their collaboration for the construction of CMs, as in all cases, QoIW4:W6 G2
> QoIW1:W3 . This is further justied by the statistical
G2
analysis results tabulated in Table 3, where a statistically signicant difference between the QoIW1:W3 G2
and QoIW4:W6 was found
p 4:54$1021 . These results indicate that shifting from the SELF- to COLL-MODE had a positive effect in the corresponding
Fig. 2. Examples of CMs developed by a pair in the COLL-MODE drawn from (a) G1 group (P5: S9,S10) and (b) G2 group (P23: S45,S46).
student's QoI, motivating them to further interact with the LMS Moodle, responding to the demands of the collaborative
activity during the COLL-MODE of the constructed CMs.
Fig. 3. The TaxScoreDiff values across the pairs of both G1 and G2 groups, i.e., TaxScoreG1;G2
Diff
(note that comparison between G1 and G2 upon the Pair No is
meaningless, as the peers of each pair are not the same across G1 and G2).
Fig. 4. The estimated TaxScoreG1;G2

SELFMODE values across all students per group (note that comparison between G1 and G2 upon the Student No is meaningless, as
the students are not the same across G1 and G2).
4.4. Overall perspective
The proposed approach, when placed within the panorama of the works that combine hybrid perspectives with the
educational contexts, lls a gap that relates to the way the users interact with LMS and collaborate with CMs within a b-/c-
learning context. When compared with the previous works, the ndings here comply with the works of Coutinho (2009),
Hwang et al. (2011), and Kwon and Cifuentes (2009), fostering the positive effect of shifting from SELF- to COLL-MODE in
the construction of the CM. Nevertheless, none of these works extend the vision of combining the CM with the LMS Moodle
use, as it was examined here, adding to more alternative teaching/learning practices and strategies (e.g., by using different
tools). This can help educators to produce an effective constructivist environment, promoting a dynamic and effective
teaching-learning process/atmosphere. From a theoretical perspective, two main types of constructivism can be considered,
namely: i) cognitive or individual constructivism depending on Piaget's theory; and ii) social constructivism depending on
Vygotsky's theory (Powell & Kalina, 2009). In fact, social constructivism is historically parallel to Piaget's theory and evolved
to include a more dynamic and social interactive environment for learning. Moreover, while social and cognitive construc-
tivist theories are different there is a major similarity, i.e., both Piaget and Vygotsky agree that teachers play a role of facil-
itator/guide, instead of a director/dictator. In this line, educators must develop/use psychological or strategic tools to promote
a constructivist, guided and inclusive learning environment.
For instance, unlike other online learning environments/platforms, the LMS Moodle does not emerge from the engineering
context but, on the contrary, it has an educational background (Cole & Foster, 2007; Olomos, Juanjo, Torrecilla, & Iglesias,
2015) based on different theories on learning. Several authors refer to LMS Moodle as being based on social
Table 2
The Statistical results from between-subjects analysis: G1 (without LMS) vs. G2 (with LMS). The boldface numbers denote statistically signicant
differences (signicance level p < 0:05).a
Parameter (related gure) One-way ANOVA test

TaxScoreG1;G2
Diff
(Fig. 3) Group: fSS 72:25; df 1; MS 72:25; F 46:81; p 4$109 g
Error: fSS 95:68; df 62; MS 1:54g
TaxScoreG1;G2
SELFMODE (Fig. 4) Group: fSS 1:75; df 1; MS 1:75; F 0:91; p 0:3398g
Error: fSS 241:23; df 126; MS 1:91g
G1;G2
TTCOLLMODE (Fig. 5) Group: fSS 2:8$104 ; df 1; MS 2:8$104 ; F 34:54; p 1:79$107 g
Error: fSS 5:1$104 ; df 62; MS 828:92g
BalG1;G2
Diff
(Fig. 6) Group: fSS 4$103 ; df 1; MS 4$103 ; F 158:72; p 9:5$1019 g
Error: fSS 1:5$103 ; df 62; MS 25:22g
Construction Add (Fig. 7)
SELF-MODE Group: fSS 3:1$103 ; df 1; MS 3:1$103 ; F 0:58; p 0:4467g
Error: fSS 6:8$103 ; df 126; MS 5:4$103 g
COLL-MODE Group: fSS 2:3$104 ; df 1; MS 2:3$104 ; F 32:27; p 8:79$108 g
Error: fSS 8:7$104 ; df 126; MS 691:70g
Construction Move (Fig. 7)
Error: fSS 7:3$105 ; df 126; MS 5:8$103 g
Error: fSS 1:5$104 ; df 126; MS 1:2$103 g
Construction Connect (Fig. 7)
Error: fSS 1$106 ; df 126; MS 8:1$103 g
Error: fSS 3:1$105 ; df 126; MS 2:4$103 g
Reection Delete (Fig. 8)
Error: fSS 7:6$104 ; df 126; MS 604:05g
COLL-MODE Group: fSS 6$103 ; df 1; MS 6$103 ; F 54:80; p 1:66$1011 g
Error: fSS 1:3$104 ; df 126; MS 110:51g
Reection Resize (Fig. 8)
SELF-MODE Group: fSS 23:86; df 1; MS 23:86; F 4:30; p 0:0400g
Error: fSS 697:95; df 126; MS 5:53g
COLL-MODE Group: fSS 39:34; df 1; MS 39:34; F 5:35; p 0:0223g
Error: fSS 925:58; df 126; MS 7:34g
Reection Modify (Fig. 8)
Error: fSS 1:85$105 ; df 126; MS 1:47$103 g
COLL-MODE Group: fSS 2$103 ; df 1; MS 2$103 ; F 14:14; p 2:58$104 g
Error: fSS 1:82$104 ; df 126; MS 144:9g
Organization Concept (Fig. 9)
Error: fSS 6$105 ; df 126; MS 4:8$103 g
Error: fSS 1:4$105 ; df 126; MS 1:1$103 g
Organization Linking Phrase (Fig. 9)
Error: fSS 2:4$105 ; df 126; MS 1:9$103 g
COLL-MODE Group: fSS 1:5$104 ; df 1; MS 1:5$104 ; F 25; p 1:86$106 g
Error: fSS 7:6$104 ; df 126; MS 608:62g
a
G1 and G2 correspond to groups without and with the use of LMS, accordingly; TaxScoreG1;G2 Diff
; TaxScoreG1;G2 G1;G2 G1;G2
SELFMODE ; TTCOLLMODE and BalDiff
denote the difference of the taxonomy score of the pair (Si ; Sj ) from the minimum taxonomy score between the two peers Si and Sj , the taxonomy
score of each peer (SELF-MODE) of each group (G1, G2), the turn-taking in the pairs of each group (G1, G2) and the absolute difference of the balance
between each peer Si and Sj from each pair (Si ; Sj ) of each group (G1, G2), respectively; SS: Sum of Square; df : Degrees of freedom; MS: Mean Square;
F : critical F value from a F distribution.
constructivist principles/theory (Chao, 2008; Duff & Jonassen, 1992; Sclater, 2008; Wainwright, Osterman, Finnerman, & Hill,
2007). According to Chavan and Pavri (2004): While most [VLEs] are instructor-oriented and largely concerned with how
course content is delivered, Moodle is based on a learner-oriented philosophy called social constructionist pedagogy, in which
students are involved in constructing their own knowledge (p. 129). On the other hand, CM as a method/technique is usually
associated with a theory from David Ausubel, a proponent of Piaget's theories, supporting that CMs can be used as a cognitive
tool to assist learners' knowledge construction (Novak & Gowin, 1984).
In this way, the effective understanding of the management of different communicative tools (e.g., LMS Moodle, CM) can
help teachers to develop individual learning methods, such as, discovery learning and social interactive activities to develop
peers' collaboration, by successfully incorporating Piaget's and Vygotsky's theories into the teaching-learning practices,
which can positively impact learner's achievement. At the end, in constructivist b-/c-learning environments, teachers should
use/adopt a variety of management tools (synchronous/asynchronous) to facilitate and encourage communication,
G1;G2
Fig. 5. The TTCOLLMODE values across the pairs of both G1 and G2 groups, i.e., TTCOLLMODE (note that comparison between G1 and G2 upon the Pair No is
meaningless, as the peers of each pair are not the same across G1 and G2).
Fig. 6. The BalDiff values across the pairs of both G1 and G2 groups, i.e., BalG1;G2
Diff
(note that comparison between G1 and G2 upon the Pair No is meaningless, as the
peers of each pair are not the same across G1 and G2).
collaboration, interaction, knowledge construction and sharing among students (Penland, 2015); however, there is a need to
develop a better understanding of constructivist Internet-b-learning environments (Zualkernan, 2006). Actually, there are
different perspectives on learning and what constitutes an effective constructivist learning environment; however, acquiring
knowledge, experience or understanding is a common room to all.
Overall, the blendedness of media or pedagogies, as the combination of tools employed in an online and c-learning
environment, or the combination of different educational approaches, should be seen as the thoughtful integration of
classroom F2F learning experiences with the combination of online learning experiences and as a real tool capable for
transformational (sociocultural) change. In general, based on a hybrid approach proposed, we intended to investigate the
effects of the LMS Moodle use in the quality of CM construction in SELF- and COLL-MODE. Moreover, when looking from an
overall perspective at this approach, the following educational contributions could be identied: i) development of an
educational/innovative framework around the online instructional environments, by exploring the potentialities of b-/c-
learning/teaching in the context of higher education; ii) contribution to educational improvement on teaching practice
supported in the LMS Moodle and/or CM tool (or OLEs in general), providing new tools more suited to users' QoI and
enhancement of collaboration; iii) introduction of extended means, integrated tools and pathways for shifting from the
typical form of LMS to the iLMS (Dias et al., 2014); and iv) construction of new recommendations about the enhancement of
OLE-based teaching/learning processes, contributing to the enrichment of the HEI services and reformulation of educational
practices.
From the results presented in this study it was made clear that the involvement of the LMS Moodle use was quite effective
in the increase of the quality of the constructed CMs, especially under the COLL-MODE. This was based on the fact that LMS
Moodle boosted the role of CM as a kind of template or scaffold to help to organize knowledge and to structure it, even though
Fig. 7. The derived fCONg set in both SELF- and COLL-MODE for G1 and G2 groups. In all subplots, the left column corresponds to the fCONg set from SELF-MODE,
whereas the right one to the fCONg set from COLL-MODE; moreover, the thin line denotes the G1 group whereas the thick one denotes the G2 group (note that
comparison between G1 and G2 upon the Student No is meaningless, as the students are not the same across G1 and G2).
the structure must be built up piece by piece with small units of interacting concept and propositional framework (Novak,
1990).
Based on the features provided by CmapTools, the CM can be used by the student as an initial step towards learning the
pieces of knowledge that s/he needs to better understand, as the basis on which to perform the research that leads to this
understanding, as a way to organize the various sources from which the student will construct this understanding, as the
artifact with which to collaborate with peers, and as the means to present his/her ndings at the end of the unit (Vitale &
Romance, 2000). When combined with the LMS Moodle underlying theory, i.e., the individual development of meaning
through construction and sharing of texts and other social artifacts (Gergen, 1995), it seems that the COLL-MODE is further
enhanced under the perspective of the common CM construction. The latter fosters collaboration and sharing processes,
which increase the quality of dialogue between participants as a tool to construct knowledge (Dougiamas, 2000) and make
learners get apprenticed into communities of practice which embody certain beliefs and behaviors (Lave & Wenger, 1991).
Overall, the ndings of this study support the combination of the LMS Moodle use with the collaborative construction of
the CMs, as an integrated learning environment that fosters hybrid b-/c-learning towards the merging of the social
constructivist approach of LMS Moodle with the knowledge scaffolding of CMs. To our knowledge, this is the rst attempt to
combine these two elds as a new means of fostering the students' learning and collaborating processes within the hybrid b-/
c-learning environment.
4.4.1. Limitations and implications

Considering the structure and the functionality of the proposed approach, some limitations could be identied. In
particular, the sample size could be increased and transferred to other context, in terms of participants' age and CM-related
content. Moreover, since the data used here refer to the LMS Moodle usage from six weeks only, similar data from longer
periods (e.g., academic years instead of one semester) could be employed, revealing possible macroscopic causal de-
pendencies, converged/dispersed interaction trends, periodicities/specic patterns dominance in the interaction attitude.
This could shed light upon more objective interpretation of the way LMS Moodle-based b-learning affects the CM con-
struction in a long run. Finally, more repetitive testing under different educational settings should be employed, in order to
justify the usefulness of the adoption of the combined CM/LMS Moodle perspective within the teaching/learning processes of
HEIs.
Fig. 8. The derived fREFg set in both SELF- and COLL-MODE for G1 and G2 groups. In all subplots, the left column corresponds to the fREFg set from SELF-MODE,
whereas the right one to the fREFg set from COLL-MODE; moreover, the thin line denotes the G1 group whereas the thick one denotes the G2 group (note that
CMs are an effective instructional strategy across various ages of learners within a variety of disciplines. They support
learning in both cognitive and constructivist fashions and could be effectively used by teachers/faculty if they are willing to
shift their style of teaching and/or their beliefs about teaching and learning. CM can be used in online and hybrid course, as
well as a means to explore how concept mapping may facilitate learning when paired with other platforms, such as LMS
Moodle. Based on our ndings, we suggest more research in the area of collaborative learning with CMs, fostering collabo-
ration within the b-learning context, employing the combined use of LMS. Shared mental models can develop through
collaborative learning and from the interaction of individuals in the development of a collaboratively constructed CM. The
hybrid approach proposed here could be used as a basis for a strategic or academic plan that can guide strategies and policies
for learning technologies/tools, coping with ongoing changes and needs to develop and review strategies, policies and
support for new learning enriched environments. However, the processes that groups use to build from individual CM to a
group CM need more investigation. Additionally, we need to understand more fully how teams and groups construct
knowledge and then how that process of knowledge construction facilitates the performance of the group, especially when
CM/LMS online tools are employed. As a step further, the adoption of the LMS use could provide an additional source of
information towards such direction, as it could reveal characteristics of the user's knowledge acquisition that is not always
reected in a clear way during the CM construction (both in SELF- and COLL-MODE).
Furthermore, it would be interesting to understand more about how different racial or ethnic groups use CMs in their
learning, investigating further the use of CMs as research tools in both quantitative and qualitative studies. This could also be
combined with the noticed differences in the derived users' LMS QoI, showing possible dependences on the cultural char-
acteristics and social beliefs. Validity and reliability testing of CMs compared to other measures of learning should be
employed. Additionally, assessment of CMs and their potential to contribute to quality control in qualitative studies could
further be examined. Finally, research in the areas of knowledge development, knowledge modeling, and knowledge man-
agement systems need to be expanded, especially within a variety of HEIs, taking into account the depth of the LMS usage in
their functionality. This type of research has the potential to help HEIs function at increasingly higher levels of teaching/
learning quality, supported by the exploitation of the new technological means.
Fig. 9. The derived fORGg set in both SELF- and COLL-MODE for G1 and G2 groups. In all subplots, the left column corresponds to the fORGg set from SELF-MODE,
whereas the right one to the fORGg set from COLL-MODE; moreover, the thin line denotes the G1 group whereas the thick one denotes the G2 group (note that
5. Conclusions and future work
The hybrid approach presented here has shown that the combination of the LMS use with the collaborative construction of
CMs further enhances the peers' QoI, resulting in CMs with higher quality, in terms of the topological taxonomy scores, and
more productive collaboration, as it is reected in peers' active participation and balanced collaboration during the collab-
oratively constructed CMs. The proposed approach fosters to the enrichment of the HEI services that stimulate re-
examination of educational practices. In particular, concrete scenarios of potential future use of the proposed approach
can be derived from the following design considerations of the presented approach:
1. Combination of LMS and CM. CMs can encourage higher-level thinking and can be used for evaluation, curriculum planning,
and the capturing and archiving of expert knowledge Use of CM combined with LMS, instead e.g., of traditional testing
instruments, may be proved an interesting initiative regarding the exibility of these tools and the commitment and
acceptance of the educators involved in such an endeavor.
2. Enhancement of communicative/Collaborative skills. It seems that the HEIs should undertake the responsibility to foster
students' interaction/collaboration skills, so they are better prepared to confront the problems of the globalized world.
Many educators believe that OLEs can be used to facilitate group problem-solving and build communication and
collaboration skills, apart from advancing students' knowledge of the specic task they are working on. To encourage
collaboration and reinforce real world skills, HEIs are experimenting policies that allow for more freedom in interactions
between students when working on projects and assessments. The approach presented and discussed here, fosters such
freedom. In particular, the analysis framework proposed in this work exhibits characteristics that could portray it as a key-
discriminator amongst hybrid learning environments, taking into consideration that emerging digital tools make it easier
for students to ask and respond to each other's questions adopting the socio-cultural approach to learning.
3. Transparency. The approach presented and discussed here, fosters the above freedom, as it is totally transparent to the user
during the time when the CM-based collaborative and/or LMS-based interactions take place, supporting and enriching, in
this way, OLEs and promoting, at the same time, peer-to-peer collaboration within the computer-based concept mapping
environments.
Table 3
Statistical results from within-subjects analysis: SELF-MODE vs. COLL-MODE. The boldface numbers denote statistically sig-
nicant differences (signicance level p < 0:05).a
Parameter (related gure) Two-sided Wilcoxon rank sum p value

Construction (Fig. 7) G1 G2
Add 3:69$107 1:05$104
Move 0:0081 0:7083
Connect 1:1$107 0:0018
Reection (Fig. 8) G1 G2
Delete 1:4$1012 0:7804
Resize 0:0167 0:0012
Modify 0:0028 0:0009
Organization (Fig. 9) G1 G2
Concept 1:2$105 0:0234
Linking Phrase 3:6$107 0:0192
G2
QoIw1:w3 G2
vs. QoIw4:w6 (Fig. 10) 4:54$1021
a G2 G2
G1 and G2 correspond to groups without and with the use of LMS, accordingly; QoIw1:w3 and QoIw4:w6 denote the mean
quality of interaction with LMS (QoI) of the G2 subjects across W1:W3 (SELF-MODE period) and W4:W6 (COLL-MODE period),
respectively.
G2
Fig. 10. The estimated mean QoI when shifting from SELF-, i.e., QoIW1:W3 G2
, to COLL-MODE, i.e., QoIW4:W6 , for each student of G2 group.
4. Large dataset/personalized feedback. Personalized feedback to each student/group based on the estimated parameters
drawn from the CM construction and LMS use could also help instructors to leverage components of online learning and to
make personalized learning scalable in large introductory classes. Compared to the traditional model of learning, in which
space is needed to accommodate hundreds of students, the hybrid b-/c-learning examined here can address the learning
path of each individual student.
As a future work, based on a hybrid approach and fuzzy-logic (FL) modelling techniques, it is our intention to propose a
unied basis in the FL domain that could model the structural characteristics of CM, such as the element of fCON; REF; ORGg
set, transferring them to fuzzy entities, in order to achieve higher exibility and more pragmatic evaluation of the CM
structure in terms of its quality (i.e., QoCM). First promising results towards such direction can be found in an authors' recent
paper (Dias, Dolianiti, Hadjileontiadou, Diniz, & Hadjileontiadis, 2016). As an extension of the latter, a global FL-based
inference system could be designed, so to combine the FL-based LMS use QoI, with the QoCM and the quality of collabora-
tion (QoC), derived from the fuzzication of the parameters dened in x2.2, so to result in an efcient evaluation tool that
could infer about the global quality of student's learning/collaboration paths. Work towards such direction is already on the
way.
Conicts of interest
There is not any actual or potential conict of interest including any nancial, personal or other relationships with other
people or organizations within three years of beginning the submitted work that could inappropriately inuence, or be
perceived to inuence, our work.
Acknowledgments
The rst author has been supported by the Foundation for Science and Technology (FCT, Portugal) (Postdoctoral Grant
SFRH/BPD/496004/2013). Moreover, the authors would like to thank all participants of this study for their smooth collabo-
ration and support.
Appendix A
Some of the COLL-MODE parameters used in the proposed analysis framework (x2.2) are mathematically expressed below:
1. Topological Taxonomy Score Difference (TaxScoreDiff ):

Si ;Sj Sj
TaxScoreDiff TaxScoreCOLLMODE min TaxScoreSSELFMODE
i
; TaxScoreSELFMODE ; (A1)
S
where TaxScoreSSELFMODE
i j
and TaxScoreSELFMODE denote the topological taxonomy score of the CMs constructed by peers Si
i j S ;S
and Sj under the SELF-MODE, respectively, whereas the TaxScoreCOLLMODE denotes the topological taxonomy score of the CM
constructed by the pair Si ; Sj under the COLL-MODE; min$ denotes the minimum value and indices i; j range from 1 to the
maximum number of peers participated in each group G1 and G2 (here equal to 64).
2. Absolute difference of the peers' balance (BalDiff :

BalDiff BalSi BalSj %; (A2)
COLLMODE
where j$j denotes the absolute value and Bal corresponds to the peer's balance within the pair, dened as the number of
fCON; REF; ORGg set contributions of each peer, normalized to the total number of the fCON; REF; ORGg set contributions in
the pair, i.e.:

card fCON; REF; ORGgSi
Si
Bal 100 %; (A3a)
card fCON; REF; ORGgSi ;Sj

card fCON; REF; ORGgSj
Sj
Bal 100%; (A3b)
card fCON; REF; ORGgSi ;Sj
where card denotes the cardinality of the fCON; REF; ORGg set contributions.
Appendix B
The following text excerpt (original in Greek), extracted from (Jimogiannis & Siorenta, 2007, pp. 242e244), was used as a
content basis for the concept map construction task in all cases, i.e., self- and collaborative-mode:
The concept of modeling
Educational research and modern scientic theories of learning have demonstrated that understanding and learning of
new concepts is an internal process for each student. New knowledge does not transfer but is built based on previous
knowledge and individual experiences. Cognitive theories of learning have as a basic principle that new knowledge is the
result of conceptual changes that occur as a result of continuous and intense interaction between pre-existing knowledge and
formal knowledge of teaching. The conceptual change and learning, as well as the development of high level skills are
progressive procedures performed over time. Cognitive sources used by students for the development of knowledge include:
their daily experience and the environment where they operate and grow,
the mental processes that they develop, and
the representation systems and mental models they use.
The models are a technical representation of behavioral conditions, procedures or systems of the natural, technological or
imaginary world. A model is dened as a mental schema or tool that is familiar or accessible on existing knowledge and helps
individuals to reach a conceptual understanding of new knowledge. It is an internal representation that plays the role of a
substitute for reality. Between the model and the reality it describes there is continuous communication, both during the
construction phase and during the operational use, in order to interpret and understand new concepts, processes, systems or
phenomena. A model allows:
the representation of a system to solve problems,

the forecasting of the development and changes of the system, without the need to observe the same reality, and
the explanation of the evolution of the system associating several factors that contribute to it.
Models are usually conceptually complex systems consisting of elements, relations, functions and rules, which dene the
interactions of the analysis system using an external representation system. They grow in the students' mind and incorpo-
rated into equations, diagrams, software or other means, which are used by students to express their ideas and knowledge,
regardless of their relationship with reality.
In building new concepts and knowledge we can distinguish three different worlds:
the real world, where the various objects or systems and happening of events, processes, and the correlations between
them live,
the conceptual world, created based on scientic theories and models and includes concepts, objects and processes of the
real world, and
the mental world, consisting of subjective, individual perceptions created by people for the real world.
Accordingly, we can distinguish three kinds of models portraying the knowledge of the real world:
conceptual models: created by the activity and cooperation of experts in the eld under study; they are objective but not
visible representations of real or imaginary world and lead to scientic knowledge,
mental models: created in the minds of students and determine their individual theories about reality; they are usually
incomplete and reect misunderstandings and conceptual difculties that students encounter in a cognitive area; it
should be noted that a person reaches the understanding a conceptual model only if s/he can generate a sufcient mental
model for each representation, and
external models: they represent mental or conceptual models and are used as teaching and/or learning tools.
There is a dynamic and reciprocal relationship between internal (conceptual, cognitive) and external models. These
representations reect the content and structure of knowledge in a given cognitive area and give to the students the op-
portunity to communicate and externalize their own knowledge. The use and development of these models by the students
themselves contribute to the overcoming of cognitive barriers and in approaching scientically acceptable conceptual
models.
Appendix A. Supplementary data
Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.compedu.2017.01.009.
References

Alvarez-Montero, enz-Pe
F. J., Sa rez, F., & Vaquero-Sa nchez, A. (2015). Using datalog to provide just-in-time feedback during the construction of concept
maps. Expert Systems with Applications, 42(3), 1362e1375.
Bates, W. T., & Sangra , A. (2011). Managing technology in higher education: Strategies for transforming teaching and learning. San Francisco, CA: Jossey-Bass.
Bridges, S. M., Corbet, E. F., & Chan, L. K. (2015). Designing problem-based curricula: The role of concept mapping in scaffolding learning for the health
sciences. Knowledge Management & E-Learning: An International Journal, 7(1), 119e133.
Can~ as, A. J., Bunch, L., Novak, J. D., & Reiska, P. (2013). Cmapanalysis: An extensible concept map analysis tool. Journal for Educators, Teachers and Trainers,
4(1), 36e46.
Chao, I. (2008). Moving to Moodle: Reections two years later. Educause Quarterly, 3, 46e52.
Chavan, A., & Pavri, S. (2004). Open-source learning management with Moodle. Linux Journal, 128, 66e70.
Clariana, D., & Taricani, E. M. (2010). The consequences of increasing the number of terms used to score open-ended concept maps. International Journal of
Instructional Media, 37(2), 163e172.
Clouder, L., Dalley, J., Hargreaves, J., Parkes, S., Sellars, J., & Toms, J. (2006). Electronic [re] constitution of groups: Group dynamics from face-to-face to an
online setting. International Journal of Computer-Supported Collaborative Learning, 1(4), 467e480.
Cole, J., & Foster, H. (2007). Using Moodle: Teaching with the popular open source course management system. London: O'Reilly Media, Inc.
Collazo, N. A. J., Elen, J., & Clarebout, G. (2015). The multiple effects of combined tools in computer-based learning environments. Computers in Human
Behavior, 51, 82e95.
Concei~ ao, S. C., Desnoyers, C. A., & Baldor, M. J. (2008). Individual construction of knowledge in an online community through concept maps. In A. J. Can ~ as,
P. Reiska, M. hlberg, & J. D. Novak (Eds.), Concept Mapping: Connecting educators, proceedings of the 3rd international conference on concept mapping (pp.
24e32). Finland.
Coutinho, C. P. (2009). Individual versus collaborative computer-supported concept mapping: A study with adult learners. In Proceedings of world conference
on e-learning in corporate (pp. 1173e1180). Vancouver, Canada: Government, Healthcare and Higher Education 2009.
Dias, S. B., & Diniz, J. A. (2013). FuzzyQoI model: A fuzzy logic-based modelling of users' quality of interaction with a learning management system under
blended learning. Computers & Education, 69, 38e59.
Dias, S. B., Diniz, J. A., & Hadjileontiadis, L. J. (2014). Towards an intelligent learning management system under blended learning: Trends, proles and
modelling perspectives. In J. Kacprzyk, & L. C. Jain (Eds.), Intelligent systems reference library (Vol. 59). Berlin/Heidelberg: Springer-Verlag, ISBN 978-3-
319-02077-8.
Dias, S. B., Dolianiti, F. S., Hadjileontiadou, S. J., Diniz, J. A., & Hadjileontiadis, L. J. (2016). FISCMAP: A fuzzy logic-based quality of concept mapping modelling
approach fostering reective feedback. In Proc. Of the 7th international conference on software development and technologies for enhancing accessibility
and ghting info-exclusion (DSAI2016). Conference proceedings series. ACM, ISBN 978-1-4503-4748-8.
Dougiamas, M. (2000). Improving the effectiveness of tools for Internet based education. Teaching and Learning Forum 2000. Proceedings Contents http://lsn.
curtin.edu.au/tlf/tlf2000/dougiamas.html.
Drysdale, J. S., Graham, C. R., Spring, K. J., & Halverson, L. R. (2013). An analysis of research trends in dissertations and theses studying blended learning. The
Internet and Higher Education, 17, 90e100.
Duff, T., & Jonassen, D. H. (1992). Constructivism: New implications for instructional technology. In T. M. Duffy, & D. H. Jonassen (Eds.), Constructivism and the
technology of instruction: A conversation (pp. 1e16). Hillsdale, NJ: Lawrence Erlbaum.
Gao, H., Thomson, M. M., & Shen, E. (2013). Knowledge construction in collaborative concept mapping: A case study. Journal of Information Technology and
Application in Education, 2(1), 1e15.

Garcia-Alvarez, M. T., Sua
rez Alvarez, E., & Quiroga Garca, R. (2014). ICTs and learning: A challenge in the engineering education. International Journal of
Engineering Education, 30(3), 636e643.
Gergen, K. J. (1995). Social construction and the educational process. In L. P. Steffe, J. E. Gale, & J. Gale (Eds.), Constructivism in education (pp. 17e39).
Hillsdale: Lawrence Erlbaum.
Graham, C. R. (2013). Emerging practice and research in blended learning. In M. G. Moore (Ed.), Handbook of distance education (pp. 333e350). New York:
Routledge.
Gurupur, V. P., Jain, G. P., & Rudraraju, R. (2015). Evaluating student learning using concept maps and Markov chains. Expert Systems with Applications, 42(7),
3306e3314.
Hadjileontiadou, S. J., Dias, S. B., Diniz, J. A., & Hadjileontiadis, L. J. (2015). Fuzzy logic-based modeling in collaborative and blended learning. Hershey, PA: IGI
Global. http://dx.doi.org/10.4018/978-1-4666-8705-9.
Hadjileontiadou, S. J., & Hadjileontiadis, L. J. (2004). A complexity analysis of collaborative turn-taking patterns evolving during computer-mediated
collaboration. World Transactions on Engineering and Technology Education, UICEE, 3(1), 155e158.
Hanewald, R., & Ifenthaler, D. (2014). Digital knowledge mapping in educational contexts. In D. Ifenthaler, & R. Hanewald (Eds.), Digital knowledge maps in
education: Technology-enhanced support for teachers and learners (pp. 3e16). New-York: Springer, ISBN 978-1-4614-3177-0.
Hwang, G. J., Shi, Y. R., & Chu, H. C. (2011). A concept map approach to developing collaborative Mindtools for context-aware ubiquitous learning. British
Journal of Educational Technology, 42(5), 778e789.
Ifenthaler, D. (2012). Computer-based learning. In N. M. Seel (Ed.), Encyclopedia of the sciences of learning (Vol. 3, pp. 713e716). New York: Springer.
Jimogiannis, A., & Siorenta, A. (2007). Modeling as a tool for development of critical and creative thinking (in Greek). In B. Koulaidis (Ed.), Modern teaching
approaches to critical development-creative thinking on secondary education (pp. 241e268). Athens, Greece: Organization of Educational Training.
Johnson, L., Adams Becker, S., Estrada, V., & Freeman, A. (2014). NMC horizon report: 2014 higher education edition. Austin, Texas: The New Media Consortium.
Johnson, L., Adams Becker, S., Estrada, V., & Freeman, A. (2015). NMC horizon report: 2015 higher education edition. Austin, Texas: The New Media Consortium.
Koc, M. (2012). Pedagogical knowledge representation through concept mapping as a study and collaboration tool in teacher education. Australasian Journal
of Educational Technology, 28(4), 656e670.
Kwon, S. Y., & Cifuentes, L. (2007). Using computers to individually-generate vs. collaboratively-generate concept maps. Educational Technology & Society,
10(4), 269e280.
Kwon, S. Y., & Cifuentes, L. (2009). The comparative effect of individually-constructed vs. collaboratively constructed computer-based concept maps.
Computers & Education, 52(2), 365e375.
Lave, J., & Wenger, E. (1991). Situated learning: Legitimate peripheral participation. New York: Cambridge University Press.
Lee, Y. (2013). Collaborative concept mapping as a pre-writing strategy for L2 learning: A Korean application. International Journal of Information and
Education Technology, 3(2), 254e258.
Lin, C. P., Wong, L. H., & Shao, Y. J. (2012). Comparison of 1: 1 and 1: m CSCL environment for collaborative concept mapping. Journal of Computer Assisted
Learning, 28(2), 99e113.
Michinov, N., & Michinov, E. (2008). Face-to-face contact at the midpoint of an online collaboration: Its impact on the patterns of participation, interaction,
affect, and behavior over time. Computers & Education, 50(4), 1540e1557.
Molinari, G. (2015). From learners' concept maps of their similar or complementary prior knowledge to collaborative concept map: Dual eye-tracking and
concept map analyses. Psychologie Franaise. http://dx.doi.org/10.1016/j.psfr.2015.11.001.
Novak, J. D. (1990). Concept maps and Vee diagrams: Two metacognitive tools for science and mathematics education. Instructional Science, 19, 29e52.
Novak, J. D. (2010). Learning, creating, and using knowledge: Concept maps as facilitative tools in schools and corporations. New York: Taylor and Francis.
Novak, J. D., & Can ~ as, A. (2006). The origins of the concept mapping tool and the continuing evolution of the tool. Information Visualization, 5, 175e184.
Novak, J. D., & Can ~ as, A. J. (2008). The theory underlying concept maps and how to construct and use them. Technical Report IHMC Cmap Tools. Pensacola,
Florida: Institute for Human and Machine Cognition http://cmap.ihmc.us/docs/theory-of-concept-maps.php.
Novak, J. D., & Gowin, D. B. (1984). Learning how to learn. Cambridge, UK: Cambridge University Press.
Olomos, S., Juanjo, M., Torrecilla, E., & Iglesias, A. (2015). Improving graduate students learning through the use of Moodle. Educational Research and Reviews,
10(5), 604e614.
Omar, M. A. (2015). Improving reading comprehension by using computer-based concept maps: A case study of ESP students at Umm-Alqura University.
British Journal of Education, 3(4), 1e20.
Penland, J. L. (2015). Constructivist internet-blended learning and resiliency in higher education. In J. Keengwe (Ed.), Handbook of research on educational
technology integration and active learning (pp. 48e61). Hershey, PA: IGI Global.
Powell, C., & Kalina, C. J. (2009). Cognitive and social constructivism: Developing tools for an effective classroom. Education, 130(2), 241.
Rafaeli, S., & Kent, C. (2015). Network-structured discussions for collaborative concept mapping and peer learning. IBM Journal of Research and Development,
59(6), 7, 1.
Savery, J. R., & Duffy, T. M. (1995). Problem based learning: An instructional model and its constructivist framework. Educational Technology, 35(5), 31e38.
Schaal, S. (2010). Enriching traditional biology lectures digital concept maps and their inuence on cognition and motivation. World Journal on Educational
Technology, 2(1), 42e54.
Sclater, N. (2008). Large-scale open source e-learning systems at the Open University UK. Research Bulletin, 12. Boulder, CO: EDUCAUSE Center for Analysis and
Research.
Seluakumaran, K., Jusof, F. F., Ismail, R., & Husain, R. (2011). Integrating an open-source course management system (Moodle) into the teaching of a rst-year
medical physiology course: A case study. Advances in Physiology Education, 35(4), 369e377.
Tergan, S. O. (2005). Digital concept maps for managing knowledge and information. In S.-O. Tergan, & T. Keller (Eds.), Vol. 3426. Knowledge and information
visualization, lecture notes in computer science (pp. 185e204). Berlin Heidelberg: Springer Verlag.
ber, W., & Neumann, A. (2006). Concept map-based visualization of knowledge and information in resource-based learning. In C.
Tergan, S. O., Keller, T., Gra
Crawford, R. Carlsen, K. McFerrin, J. Price, R. Weber, & D. Willis (Eds.), Proceedings of society for information technology & teacher education international
conference 2006 (pp. 2425e2429). Chesapeake, VA: Association for the Advancement of Computing in Education (AACE.
Vitale, M. R., & Romance, N. R. (2000). Portfolios in science assessment: A knowledge based model for classroom practice. In J. J. Mintzes, J. H. Wandersee, &
J. D. Novak (Eds.), Assessing science understanding: A human constructivist view (pp. 168e197). San Diego, CA: Academic Press.
Vodovozov, V., & Raud, Z. (2015). Concept maps for teaching, learning, and assessment in electronics. Education Research International. http://dx.doi.org/10.
1155/2015/849678.
Wainwright, K., Osterman, M., Finnerman, C., & Hill, B. (2007). Traversing the LMS terrain. In Proceedings of the 35th annual ACM conference on user services
SIGUCCS 2007 (pp. 355e359). Orlando New York: ACM Press.
Zualkernan, I. A. (2006). A framework and a methodology for developing authentic constructivist e-learning environments. Journal of Educational Tech-
nology & Society, 9(2), 198e212.

Computers & Education: So Fia B. Dias, So Fia J. Hadjileontiadou, Jos e A. Diniz, Leontios J. Hadjileontiadis

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Computers & Education: So Fia B. Dias, So Fia J. Hadjileontiadou, Jos e A. Diniz, Leontios J. Hadjileontiadis

Uploaded by

Copyright:

Available Formats

Computers & Education 107 (2017) 127e146

Contents lists available at ScienceDirect

Computers & Education

Computer-based concept mapping combined with learning

1.1. Computer-based learning environments (CBLEs)

1.2. Concept mapping in CBLEs

1.3. Self- and collaborative-mode CMs

1.4. Study aims and research questions

2.2. The proposed analysis framework

3. Experimental and implementation issues

Groups CmapTool use LMS Moodle use

3.3. Data analysis tools

4. Results and discussion

Fig. 4. The estimated TaxScoreG1;G2

4.4. Overall perspective

Parameter (related gure) One-way ANOVA test

4.4.1. Limitations and implications

5. Conclusions and future work

Parameter (related gure) Two-sided Wilcoxon rank sum p value

1. Topological Taxonomy Score Difference (TaxScoreDiff ):

2. Absolute difference of the peers' balance (BalDiff :

The concept of modeling

the representation of a system to solve problems,

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.compedu.2017.01.009.

You might also like