comparison of 1:1 and 1:m cscl environment for collaborative concept mapping

Comparison of 1:1 and 1:m CSCL environment forcollaborative concept mappingjcal_421 99..113

C.-P. Lin,* L.-H. Wong† & Y.-J. Shao†*Graduate Institute of e-Learning Technology, National Hsinchu University of Education, Taiwan†Learning Sciences Laboratory, National Institute of Education, Singapore

Abstract This paper reports an investigation into the effects of collaborative concept mapping in a digitallearning environment, in terms of students’ overall learning gains, knowledge retention, qualityof student artefacts (the collaboratively created concept maps), interactive patterns, and learn-ing perceptions. Sixty-four 12-year-old students from two 6th grade classes (32 from eachclass) participated in the study. Guided by the methodology of quasi-experimental research,group scribbles 1.0 was adopted in which students carried out collaborative concept mappingactivities for social studies in two different settings: (1) 1:1 (one-device-per-student) – studentsworking in pairs with one Tablet PC assigned to each of them; and (2) 1:m (one-device-to-many-students) – multiple students sharing a Tablet PC. Both settings were evaluated and theinteractional patterns of the student groups’ concept mapping were identified. The results indi-cated that in both 1:1 and 1:m settings, students had improved their learning results and reten-tion. Nevertheless, while 1:1 groups had demonstrated more consistency in group participation,improved communication and interaction, the 1:m groups had instead generated superior arte-facts as all the notes were well discussed among the group members. The findings suggest that ahigher quality of collaborative processes does not necessarily lead to improved studentartefacts.

Keywords collaborative concept mapping, collaborative learning, computer-aided concept mapping,group scribbles, meaningful learning.

Introduction

Social Studies is regarded as an integrative subject thatapplies multiple disciplines of learning concepts andprinciples to social life (Mintrop 2004). In most of theregular Social Studies lessons in schools, teachers arestill giving didactic instructions and employing drill-and-practice routines (see, for example, Kramer-Dahlet al. 2007; Guildry et al. 2010). These tend to belearning outcome-oriented and do not focus on ensur-ing a deep understanding of what is learnt, often result-

ing in students overemphasizing the importance ofmemorizing fragmented knowledge without synthesiz-ing it. This makes learning neither interesting noreffective, and meaningless (Yu et al. 1996). In turns,we conducted a study on incorporating alternativelearning design into Social Studies lessons to addressthis problem. By synthesizing collaborative learningtheories and concept mapping, we designed a learningenvironment for collaborative concept mapping (CCM)activities to support collaboration in competitions andoffer more opportunities for discussions and interac-tions among students. Students were placed into groupsand allocated devices in two ways: each student had aTablet PC in a group [one-device-per-student (1:1)] oreach group of students shared only one Tablet PC

Accepted: 28 January 2011Correspondence: Chiu-Pin Lin, Graduate Institute of eLearning Tech-nology, National Hsinchu University of Education, 521 Nan-Da Road,Hsin-Chu City, 300, Taiwan. Email: [email protected]

doi: 10.1111/j.1365-2729.2011.00421.x

Original article

© 2011 Blackwell Publishing Ltd Journal of Computer Assisted Learning (2012), 28, 99–113 99

[one-device-to-many-students (1:m)]. We envisionedthat different affordances of devices to various groupsof students would have affected the collaboration andthe learning outcomes, as well as influenced theirthinking processes and use of inference in the construc-tion of a concept map.

In this paper, an attempt has been made to seekthe answers in 1:1 and 1:m settings to the followingresearch questions:

• Is there any difference in the learning outcomes,learning retention, and the co-constructed conceptmaps between collaborative concept mapping in 1:1and 1:m settings?

• Did students in these two settings display differentbehaviours in their in-group interaction andcommunication?

• Did students in these two settings differ in their per-ceptions in and attitudes to the CCM?

Theoretical background

Earlier literature has investigated the advantages ofCCM in learning. Previous research on 1:1 digital learn-ing has also shed light on the technology used toenhance concept mapping learning.

CCM

CCM (Roth & Roychoudhury 1994; Hoppe & Gabner2002; Komis et al. 2002; Basque & Lavoie 2006) isgrounded in the synergy of concept mapping andcomputer-supported collaborative learning (CSCL). Inthe context of CCM, the target content to be learnt andthe target concept map to be constructed are the sameacross all members within a learning group. Someresearch suggested that CCM is an effective tool thatcan lead to synchronous collaboration and effectivediscussions concerning concepts and thus enhancemeaningful learning through a structure of annotationsof various levels of abstraction, interpretation, andreflection (Fischer et al. 2002; Avouris et al. 2004).Group members can co-construct a map within thesame framework, thus reducing the cognitive load ofindividual learners (Nesbit & Adesope 2006). Further-more, through interaction and peer support, each indi-vidual continuously appends to and amends the targetconcept map based on his/her own understanding of the

learning content as well as on their judgments of peeropinion.

For each learning group, CCM is a process of col-laborative problem solving. During such processes,group members may gain new knowledge from oneanother. Even though the final maps were the effort ofthe whole group, it is deemed much more importantfor all students to have the opportunity to engage indiscussion to consider the connections between majorconcepts from the course. Individual differences inunderstanding should not be extinguished by thisprocess, but rather should be offered for comparisonwith other perspectives introduced by students inorder to promote conceptual change or consolidation(Kinchin et al., 2005). Previous practices of onlineCCM environment have reported positive outcomes ofenhancing group critical thinking and collaborativeproblem-solving skills, constructing diagrammatic con-ceptual representations or sharing solutions into acommon space (Fidas & Komis 2001; Komis et al.2002; Madrazo & Vidal 2002). Researchers found thatstudents could elaborate, refine, and improve their ownknowledge structures, using CCM to represent knowl-edge (Cañas et al. 2003).

1:1 digital learning

The term ‘1:1 digital learning’ was proposed by CathieNorris and Elliot Soloway (Norris & Soloway, 2002,2004) and refers to one device (or more) per student forlearning purposes. With a variety of affordances offeredby wireless mobile computing, information and digitalresources can be transferred between devices or storedon the Internet. That opens up endless possibilities ofthe design and enactment of innovative teaching andlearning models (Looi et al. 2010a), ranging from con-ventional personal computer labs to perpetual and ubiq-uitous learning (e.g. De Jong et al. 2008; Mifsud &Mørch 2010), authentic and contextualized learning(e.g. Lai et al. 2007; Wong & Looi 2010), seamlesslearning (e.g. Chan et al. 2006; Roschelle et al. 2007;Looi et al. 2010b), rapid knowledge co-construction(Lin et al. 2008), among others. Furthermore, Chan(2010) points out that we are now at the onset of a digitalclassroom wave which will bring significant changes toeducation.

After the initial hype, however, there have beenvoices within the researcher community to reassess the

100 C.-P. Lin et al.

© 2011 Blackwell Publishing Ltd

notion of 1:1 computing for technology-transformedlearning (Looi et al. 2010a), such as whether 1:1 set-tings may impact peer collaboration and teachers’ roles(e.g. Carlson 2007; Dillenbourg 2010), the issues ofstudent and social readiness (e.g. Wong et al. 2010), aswell as the explorations of alternative settings ofmany-to-one (e.g. Dieterle & Dede 2007), 1:m (e.g.Wong et al. 2009), many-to-many (e.g. Rogers et al.2010) configurations, or hybrid settings of theseconfigurations (e.g. Liu & Kao 2007). Clearly, each ofthese computer–student ratios provides differentdynamics of interaction and collaboration that sup-port a myriad of learning designs. From the learner’spoint of view, the individual herself is the invariantand there needs to be a sense of seamlessness inswitching contexts between these different designs(Wong 2010).

In the current study, the Tablet PC was chosen as thelearning tool for face-to-face communications, andmembers of learning groups leveraged on wirelessaccess to interact with each other in the collaborativeactivities. Nevertheless, as a 1:1 setting (one Tablet PCper student in a group) tended to give individual stu-dents greater autonomy in making personal contribu-tions and in editing the group concept map, conceptualor process (such as concurrent mapping) conflicts wereoften inevitable (e.g. Cheng, Kruger & Daniels 2003).Managing conflicts is always deemed a great challengein CCM activities (Chiu 2003). The 1:m setting (oneTablet PC per group), on the contrary, is potentially acompromise. A literature scan revealed many studieswhich compare individual concept mapping and CCM(e.g. Chiou 2009; Coutinho 2009; Kwon & Cifuentes2009), and the results generally favoured CCM. Forexample, the study of face-to-face 1:m networkedconcept mapping conducted by Chiu and his colleagues

suggested the significant positive correlation betweengroup performance and the level of group interaction(Chiu et al. 2000). However, a research gap still existson the comparative studies between 1:1 and 1:m set-tings in the context of CCM which the empiricalexperiment of the current report seeks to fill. It isargued that differences exist in the interaction of groupmembers between these 1:1 and 1:m face-to-face butnetworked settings for CCM, and a review is presentednext on studies of interactional patterns for moreinsights of group interaction.

The interactional patterns of learning groups

Over the decades, a number of researchers have beenseeking ways to categorize interactional patterns insmall learning groups, with or without computersupport, through their ethnographic studies on theirlearning activities. In search of an analysis tool touncover and classify the student group communicativepatterns in the study, three relevant frameworks wereidentified and synthesized, namely, the categorizationsproposed by Milson (1973), Roth (1995) and Huang(2001). Milson’s categorization has offered one of themost influential interaction analysis tools for collabora-tive learning, and recent CSCL studies are still adopt-ing this technique (e.g. Chen et al. 2003; Liu & Kao2007). Examples of recent CSCL studies that adoptedRoth’s framework are Li (2006) and Hu and Chiou(2008).

Some parallels can be observed among the categori-zations of the three interactional patterns, which werewell grounded in their empirical studies and analyses asreported in the respective publications. In the presentstudy, they have been consolidated and mapped into fivecommon patterns (Table 1).

Table 1. Consolidation of interactional patterns in small groups.

Milson (1973) Roth (1995) Huang (2001)

Ideal Ideal Symmetric interaction Turn takingLeader Dominant leader Asymmetric interaction Leader

Shifting symmetric interaction FocusTete-a-tete Tete-a-tete Parallel occasional interactionFragmented Fragmented, cliquish Asymmetric interaction Fragmented

StiltedNo participation Unresponsive No participation

Unsocial

1:1 and 1:m CSCL environment comparison 101


Methodology

This study sought to investigate the effects of CCM in a1:1 digital learning environment, as compared with 1:m.Sixty-four students, aged 11 or 12, from two 6th gradeclasses (32 from each class) in a primary school in Tai-chung County participated in the study. They weretaught the second unit of the Social Studies lesson,‘Investment, Financial Management and EconomicActivities’, as well as taking a competency pre-test onthe subject, prior to the study. Based on their results inthe pre-test, a homogeneity test was performed on thetwo groups of students by classifying them into the fol-lowing heterogeneous groups.

Students were split into eight ‘1:1 groups’ and eight‘1:m groups’ with similar numbers of high-, medium-,and low-ability students across the groups to ensure ahomogeneous subject matter competency among thegroups. Whereas each of the members of the 1:1 groupwas assigned a Tablet PC to perform CCM, in the 1:mgroup four students shared one Tablet PC. Sets of quan-titative and qualitative data were collected for analysis.The quantitative data consisted of the results of pre-,post-, and postponed-tests, with the ‘N-G score’(see thesection under ‘Research findings and discussion’ formore details) proposed by Novak and Gowin (1984) asthe scoring rubric of the students’ concept maps, and astudent questionnaire investigating student attitudes tocollaborative learning, the usability of the software, andlearning by concept mapping. The qualitative data con-sisted of students’ interviews which sought their per-ceptions in learning, and video recordings made duringthe intervention in order to analyse their in-group inter-action. All the data were triangulated and the findingswere formulated in the nature of 1:1 computer-supported CCM activities.

Computer-supported CCM activities

We adopted group scribbles (gs) (Chaudhury et al.2006), a CSCL system developed by SRI International,to conduct small group CCM activities. The typical gsinterface presents each student with a three-panewindow, as shown in the student’s interface in Fig 1.The lower pane is the user’s personal work area, or‘private board’, with a virtual pad of fresh ‘scribblenotes’ (in the context of concept mapping, each ‘note’carries a concept) on which the student can draw or type

with different colours. Students can use the ‘Link’ func-tion to connect two scribble notes. A scribble can bevisible to others by dragging it into the ‘public board’(or their own ‘group board’) in the upper pane and theentire board can be rearranged if necessary, synchro-nized across all devices. The reverse drag makes theitem private again. Users may interact with publicscribbles in a variety of ways, such as browsing theircontent, repositioning them, or moving one from thepublic board into their private space. New public boardscan be created to support multiple activities or spacesfor small groups to work (Looi et al. 2008).

The teacher’s interface has the additional administra-tor board added in the left window frame, as illustratedin Fig 2. This frame shows the list of the students andthe grouping of the students. In the student listing, agreen dot means that the student is online while a red dotmeans that that student is offline or absent. On the top ofthe pane, the function of the group composition is pro-vided, which facilitates the teacher in changing thecomposition of the group by simply dragging namesfrom the student list.

Before the learning activity, the teacher explained tothe students the essence and the components of conceptmaps, and facilitated a hands-on activity to becomefamiliar with the gs software and to establish groupwork skills. The group concept mapping activitiesencouraged the students to collaborate in buildingconcept maps on the theme of ‘Investment, Financial

Fig 1 Students’ interface.



Management and Economic Activities’ covered in theSocial Studies textbook. Figure 3 depicts the 1:1 and1:m CCM activities.

During the CCM phase, members of each 1:1 groupcarried out their discussion and posted ideas or con-nected various concepts to their Group Board, usingtheir individual Tablet PCs. As they continued torefresh their displays for the up-to-date group conceptmaps, individual students would instantly raise theirthoughts or edit the concept map from their ownTablet PC. Conversely, having only one sharedTablet PC, each 1:m group identified a team memberto assume the responsibility of creating and editingthe concept map, while the rest provided only verbalopinions.

Data collection

To evaluate and compare the processes and outcomes ofgroup concept mapping in 1:1 and 1:m settings, a datacollection plan was executed throughout the course ofthe study as summarized next.

• Pre- and post-tests: An instrument was developedfor the Social Studies competency test on the‘Investment, Financial Management and EconomicActivities’ topic. The instrument consisted of 27multiple-choice questions and true–false questions,with a total score of 100. The instrument was dulyverified and validated through a pilot study and therevision of three teachers.

• Learning attitude questionnaire: The questionnairewas intended to investigate the student attitudes in theCCM activities. The 37-item scale consisted of threecomponents: attitudes in collaborative learning (12items), the usability of the software (10 items), andlearning by concept mapping (15 items). Studentsresponded to a four-point Likert Scale (1 = ‘stronglydisagree’, 2 = ‘disagree’, 3 = ‘agree’, 4 = ‘stronglyagree’). The questionnaire was administeredpromptly at the end of the intervention.

• Interviews: Semi-structured interviews were con-ducted with both groups of students. All the inter-views were audio recorded and then transcribed andcoded for further analysis and data triangulation.

• Video recordings of the intervention: To ascertain theattitudes of the students and their learning processduring the mapping activities, photos were takenand the intervention was filmed; this also helped toinvestigate the interaction patterns within individualstudent groups.

Experimental process

• Pre-test (20 min): The paper-and-pencil test wasadministered in each of the two participating classes inturn.

• gs training (20 min): The teacher introduced to the stu-dents the electronic notes, the shared space, and theconcept mapping tool to familiarize them with the gsenvironment. The students then carried out a conceptmapping exercise in order to establish their group workskills.

• Group concept mapping phase (80 min): The conceptmapping theme was ‘Investment, Financial Manage-ment and Economic Activities’. The students generatedgroup concept maps collaboratively, based on their indi-vidual understanding of the topic.

• Post-test and questionnaire (40 min): The 20-min post-test was administered promptly at the end of the inter-vention, which was followed by the administration of thelearning attitude questionnaire, which lasted another20 min.

• Postponed-test (20 min): The 20-min postponed-testwas administered 1 month after the end of the interven-tion in order to assess how well they had retained theirlearning.

• Student interviews: One-to-one interviews were con-ducted after the end of the intervention.

Fig 2 Teacher’s interface.



Research findings and discussion

This section presents findings from the study. In thispaper, focus has been placed on (1) the analysis of theconcept mapping results and (2) qualitative analysis ofgroup interactional patterns. We summarize the analy-ses of the pre- and post-test data and questionnairesnext; for more details of these aspects, refer to Lin et al.(2010).

Analysis of the pre-, post-, and postponed-test andquestionnaire data

• Comparison of the prior competencies of the studentsinvolved in 1:1 and 1:m settings: An independent samplet-test was conducted on their pre-test results, yielding thesignificance value P = 0.157 > 0.05 (Levene test:P = 0.299 > 0.05), and indicating that both sets of stu-dents had compatible competencies in Social Studiesprior to the study.

• Comparison of the learning effects after the intervention:The analysis of covariance test was carried out on theirpre- and post-test results and yielded F = 0.231,P = 0.632 > 0.05 (test of homogeneity of regressionslopes: P = 0.398 > 0.05), showing that the post-test

results between both sets of students were not signifi-cantly different.

• Analysis of the intervention feedback questionnaire:Students involved in the 1:1 setting scored significantlyhigher than those in the 1:m setting in all three sub-scales(perceptions) of the questionnaire, namely, ‘collabora-tive learning’ (P = 0.033 < 0.05), ‘usability of the soft-ware’ (P = 0.044 < 0.05), and ‘learning by conceptmapping’ (P = 0.028 < 0.05).

The analysis of the concept mapping results

As previously stated, an ‘N-G score’ was adopted tograde the students’ group concept maps for furtheranalysis. The scoring scheme is described as follows:(1) one point is awarded for each valid and meaningfullink; (2) five points are awarded for each hierarchy inthe map; (3) 10 points are awarded for each cross-linking; (4) one point is awarded for each valid example.The total score of a concept map is the sum of the scoresearned for these four components.

Table 2 depicts the artefact results (scores) yielded bythe two groups, which were very close. However, thestandard deviation (sd) of the 1:1 groups was promi-nently higher than that of the 1:m groups, indicating that

a b

Fig 3 (a) 1:1 and (b) 1:m collaborative concept mapping activities.

Table 2. The concept mapping scores ofthe 1:1 and 1:m groups.Group Group (No)/scores Mean SD

1:1 (1) (7) (3) (4) (8) (5) (2) (6) 42.63 11.4959 58 46 42 40 37 33 26

1:m (2) (5) (6) (1) (8) (3) (7) (4) 42.25 6.7451 51 47 42 39 38 37 33



the differences among the 1:1 group members weregreater. The artefacts created by the groups were com-pared to find out who scored the highest and the lowestpoints (Group 1 with 59 points and Group 6 with 26points); it was observed that Group 1 had produced amore structured concept map, with more appropriatedescriptions for the concepts and the links. However,the concept map of Group 6 was deficient in omittingkey points and giving colloquial, sentence-styledescriptions of the links, which resulted in a lack ofstructure and many invalid links in the concept map. Itwas therefore concluded that members of Group 6needed more training in the method of conceptmapping. Figure 4 depicts that the concept mapscreated by Group 1 is more structured and hierarchicalthan Group 6 of the 1:1 groups.

In contrast, within the 1:m groups where one TabletPC was shared in each group, the group members wereless likely to interfere with one another in the process ofconcept mapping. The interactional patterns were moreconsistent, and so were the quality of their artefacts(which is reflected by smaller differences in theirconcept mapping scores). It can be argued that the 1:1groups are likely to generate greater differences in theirscores than the 1:m groups.

Furthermore, in the 1:1 setting group, work skills andinteractional patterns might have exerted greater influ-ence on their scores compared with those of the 1:mgroups. It was observed that a student group with effec-tive group work skills might perform even better in a 1:1setting – for example, Groups 1 and 7 of the 1:1 groups

produced the most organized concept maps with thehighest scores (59 and 58, respectively), and the greatestnumbers of nodes (37 and 36, respectively). Observa-tions from video records revealed that each member ofthese groups engaged well in their discussions. Thepost-interviews also revealed that these two groups hadindeed exercised more effective group work skills – forexample, a student in Group 7 remarked that, ‘Eachmember proactively participated, as we followed theleader’s instruction at times, and engaged in discussionsat other times’.

However, it was also observed that as all members ofthe 1:1 groups were at liberty to add new nodes and links,it had nevertheless been difficult for them to anticipatewhat and when their fellow group members would beposting next. Such virtually synchronous mode ofconcept mapping might have resulted in relatively disor-ganized concept maps, as in 1:1 Group 3. In those groupswith weaker bonding, members had been interfering ordeleting each other’s contributions, causing weakconcept map contents and lower scores. For example,the student from 1:1 Group 2 argued that his group hadlacked collaborative skills and often interfered with eachother’s efforts. He further criticized a team member fordeleting other people’s notes without permission, result-ing in fewer notes made in their group concept map.Figure 5 depicts the concept maps of 1:1 Groups 2 and 3.

Table 3 presents the amount of notes generated by the1:1 and 1:m groups. As the 1:1 groups had apparentlygenerated more notes compared with the 1:m groups, itcan be argued that by adopting the 1:1 setting, students

a b

Fig 4 Students’ concept maps [1:1 Group 1 (a) vs. 1:1 Group 6 (b)].



would be more proactive in offering opinions and par-ticipating in concept mapping. This reasoning tallieswith the questionnaire results on ‘attitudes on collabo-rative learning’, namely, Q8 ‘I think such a learningmethod has motivated me in expressing my thoughtsand opinions to the group,’ that yielded significant dif-ference (P = 0.041 < 0.05) in the means between the 1:1(mean = 3.72) and 1:m groups (mean = 3.41). Further-more, the sd among the 1:1 groups’ notes (sd = 0.457)was greater than that of the 1:m groups (sd = 0.712).The reasoning for this phenomenon is similar to theexplanation for the greater sd among the scores yieldedby the 1:1 groups as compared with the 1:m groups,although the group work skills might have also played apart in this.

Table 4 lists the ‘scoring ratios’, i.e. to divide thescores by the number of notes of all the groups. It shows

that the scoring ratios of the 1:m groups are higher thanthose of 1:1 groups – the lowest scoring ratio yieldedamong the 1:m groups, 1.76 (Group 2), is higher thanthose of 1:1 Group 7, Group 1, and Group 6. The impli-cation is that although the mean of the amount of notesfrom the 1:m groups is less than that of the 1:1 groups by4, the former set of groups usually held group discus-sions prior to concept mapping, and therefore theirconcept maps were more relevant and accurate, result-ing in higher scores.

In general, more notes should result in higher scores.Nonetheless, Table 4 seems to suggest otherwise -1:1Group 6 and 1:m Group 8 are positive and negativeexemplars of this. The 1:1 Group 6 posted 24 notes andshould have scored around 40 marks by the same stan-dard of 1:1 Group 4 with 23 notes; instead, the groupscored the lowest (26) marks, which produced a scoring

a b

Fig 5 Students’ concept maps [1:1 Group 2 (a) vs. 1:1 Group 3 (b)].

Table 3. The amount of notes of the 1:1and 1:m groups.Group Group (No)/amounts of notes Mean SD

1:1 (1) (7) (3) (6) (4) (8) (5) (2) 24.38 8.46737 36 26 24 23 18 17 14

1:m (2) (5) (1) (6) (3) (7) (4) (8) 20.75 5.65129 27 23 23 18 18 14 14

Table 4. The ‘scoring ratios’ of the 1:1 and1:m groups.Group Group (No)/scoring ratios Mean SD

1:1 (2) (8) (5) (4) (3) (7) (1) (6) 1.83 0.4172.36 2.22 2.18 1.83 1.77 1.61 1.59 1.08

1:m (8) (4) (3) (7) (6) (5) (1) (2) 2.10 0.3332.79 2.36 2.11 2.06 2.04 1.89 1.83 1.76



ratio of 1.08 – the lowest among all the groups. Theconcept map of 1:1 Group 6 was examined and it wasfound that the concept descriptions were colloquial, insentence style, and failed to capture the key points.Hence, the overall concept map was unstructured andcontained many invalid notes.

Conversely, 1:m Group 8 posted 14 notes and scored39 points (thus yielding a scoring ratio of 2.79), whichwas superior to 1:1 Group 2 and 1:m Group 4 whoscored 33 points, respectively, with the same amount ofnotes. The concept map of 1:m Group 8 was studied,and it was noted that there was an additional hierarchyin the map (five hierarchies in total, scoring 25 marks,which was five marks more than the previously men-tioned two groups, respectively).

In summary, this analysis shows that the mean scoresof the 1:1 groups and the 1:m groups were very close.Nevertheless, the 1:1 groups revealed a greater differ-ence among their scores, probably because of the largerinfluence of the group work skill factor on the results. Incontrast, the 1:1 setting resulted in more notes beingposted, although not necessarily transforming intohigher scores. The video recordings were examined,and it was discovered that some of the notes might havebeen posted without being discussed beforehand, andmay not therefore be valid. However, the 1:m groupsheld discussions prior to posting notes, thus producingconcept maps with greater validity and higher scoringratios.

Closer observation of the videos of the groups’ col-laborative process revealed that most of the 1:1 and 1:mgroups had been able to carry out discussions and col-laborations smoothly, perform analysis and classifica-tion effectively, and give examples on the theme. Withthe exception of Groups 3 and 6, all of the concept mapsof the 1:1 groups were neatly executed and structured.The phenomenon shows that the 1:1 setting would notnecessarily result in disorganized concept maps; it issuggested that the groups who scored lower should beable to improve their concept maps by receiving addi-tional instruction in collaborative learning and conceptmapping skills.

Another important observation is that although mostof the students had been able to demonstrate theirknowledge of investment and financial management intheir concept maps, their classifying structure and mostof the examples raised were based on the textbookcontent, with the exception of a few groups who

managed to extend their concept maps beyond the text-book. In addition, most of the groups did not considercross-linking – an important activity to achieve scores –in their map constructions, thus no score was earned inthis aspect. The possible reasons for this phenomenonare: (1) the time limit for CCM had resulted in the stu-dents not being able to incorporate their own conceptsbeyond the textbook and determine the sets of noteswith potential relations of cross linking; and (2) the stu-dents were rarely reminded to note the possibilities of‘cross linking’ throughout the intervention. The studentinterviews also revealed that they needed more time toconstruct their concept maps, as a student from 1:1Group 3 stated, ‘We would be able to produce a more“complete” concept map should we be given more timeto discuss’. To address the problem of the lack of cross-linking, it is therefore advisable to allow students moretime for in-depth discussion, as well as further instruc-tion in the skills of observation and reflection.

Analysis of the interactional patterns

One of the major goals of this study was to find outwhether there was a significant difference between theinteractional patterns of the 1:1 and 1:m groups. Ananalysis was made of the video recordings and groupobservation diagrams created, as presented in Table 5,guided by the consolidated categorization of interac-tional patterns presented in Table 1. In the diagrams, H,M, and L denote high-, medium-, and low-ability stu-dents, respectively; the two-way arrow ↔ refers to two-way communication that occurred between twostudents; and the one-way arrow → denotes one-waycommunication. No arrow between two studentsimplies occasional or no interaction.

In correspondence with Table 1, four types of interac-tional patterns were observed in the video records ofstudents’ activities: ‘ideal’, ‘leader’, ‘tete-a-tete’, and‘fragmented’, occurring during the CCM activities ofthe 1:1 groups. The most common interactional patternassumed by 1:1 groups was the ‘ideal’ pattern. Amongthe three 1:1 groups with the ‘ideal’ pattern, Group 3and Group 7 performed well (ranked second and third,respectively, among the 1:1 groups in terms of theconcept map scores) while Group 6 yielded the lowestscore among all 16 1:1 and 1:m groups. This finding wasunexpected, which suggested that not all learning in the‘ideal’ pattern would lead to good learning outcomes.



The reason for this is not clear, but it also indicates thatthere could be factors other than interactions that influ-ence learning. It may also be connected with their man-agement of intra-group tasks and their organization ofdiscussion, and this needs further investigation.

In contrast, the 1:m groups demonstrated ‘leader’ or‘fragmented’ patterns but no ‘ideal’ or ‘tete-a-tete’ pat-terns. The ‘fragmented’ pattern was overwhelminglypopular – seven groups undertook this pattern. Group 1,the only group that followed the ‘leader’pattern, ranked

Table 5. The interactional patterns of the1:1 and 1:m groups.1:1 groups 1:m groups

Ideal Group 3 (46) Group 6 (26)

Group 7 (58)

Leader Group 1 (59) Group 8 (40) Group 1 (42)

Tete-a-tete Group 4 (42)

Fragmented Group 2 (33) Group 5 (37) Group 2 (51) Group 3 (38)

Group 4 (33) Group 5 (51)

Group 6 (47) Group 7 (37)

Group 8 (39)

The * in 1:m group indicates the student who operated the computer. The numbers inbrackets represent their group concept map scores.



fourth among the eight 1:m groups. This indicates thatin the pattern of 1:m groups, better interactions also ledto more effective learning outcomes. It was apparentthat the interactions among 1:m group members weregenerally not as effective as the 1:1 groups. Isolated stu-dents were spotted in those groups, and the intra-groupcommunication was far from ideal. It was observed thatmost of these isolated students were of low abilitywhose post-test results did not show significantimprovement over their pre-test results. In 1:m groups,therefore, there could be a risk that the individual iso-lated students who were not engaged in the group worklearned very little and might have lost their enthusiasm.This could be a disadvantage for this form of collabora-tive learning.

Nevertheless, some of these 1:m groups with theperceived inferior interactional patterns had insteadachieved a higher score for their concept maps thanmost of the 1:1 groups. Unlike 1:1 groups which oftensuffered from the interference by group members withone another’s contributions, the dominating 1:m groupmembers were able to become more focused in conceptmapping, thus producing concept maps of betterquality. However, it can be suggested that members ofthose groups who demonstrated the ‘fragmented’pattern might not learn and benefit from one another(some members were left out by the core group of con-tributors) and therefore their ‘better’ concept maps didnot necessarily reflect the collective learning gains forall the members. On the contrary, whereas assuming the‘ideal’ pattern did not guarantee the delivery of betterartefacts for the 1:1 groups, the group members’ level ofparticipation tended to be more balanced and thereforecreated a more fulfilling learning experience, as indi-cated by the results of the intervention feedback ques-tionnaire. In this sense, this finding agrees with theargument proposed by Kinchin et al. (2005) based ontheir paper-based CCM study.

Analysis of student interviews

The constant comparative method inspired by groundedtheory (Strauss & Corbin 1990) was applied to code thestudent interview data; it was divided into three axes,namely:

• Interactional behaviours among students: Studentsfrom the 1:1 groups claimed that they were able to

participate proactively in discussions and in the CCMactivities, as well as helping each other; their levels ofcollaborative skills were influential to the learningactivities. Students from the 1:m groups, conversely,observed that some of their fellow group members didnot participate often in the learning activities.

• Student perceptions in learning: Students from the 1:1groups thought that the learning activities were con-ducive to boosting group discussion and increasingindividuals’ willingness to express their ideas. Theyfound the class enjoyable and engaging.

• Views on concept mapping: The students believedthat concept mapping helped them in retaining andunderstanding the course contents. The strategy couldbe applied to the learning of other subjects as well asin their daily lives.

Conclusions and recommendations

A Social Studies course emphasizes the skills of synthe-sis and application. The key to mastery of Social Studiesis to move away from conventional rote learning andinstead become actively engaged in meaningful learn-ing and knowledge acquisition. In this study, CCM wasinvestigated as an effective method of developinglogical thinking and enhancing engagement. Both 1:1and 1:m settings were implemented and compared, withthe following conclusions being drawn:

• Learning outcomes: The analysis of student perfor-mances showed that both 1:1 and 1:m settings couldimprove the students’ results and also demonstratedgood retention. Neither setting resulted in significantdifferences in the improvement of these two aspects.As no control group was involved, it is not the aim ofthe study in the first place to find out whethercomputer-assisted CCM will result in better learningeffectiveness compared with, for instance, paper-based CCM or individual concept mapping.

• Results of concept mapping: Despite there being littledifference in the concept map scores between stu-dents engaged in the two settings, the sds of the 1:1groups were greater than those of the 1:m groups.According to the analysis, the disparity between theperformances among the 1:1 groups could beexplained by the greater influence of the levels ofgroup work skill on the effectiveness of their collabo-rations. Conversely, the statistics show that 1:m



groups yielded higher scoring ratios (1.76 marks pernote posted in average) than the 1:1 groups. Although1:1 groups posted four notes (per group) more than1:m groups on average, the higher number of notesdid not necessarily transfer to higher scores. In 1:1groups, many of the notes may have been postedwithout being discussed beforehand, thus they werenot necessarily valid. In contrast, most of the 1:mgroups made well-discussed notes and were thereforemore valid and consistent on theme.

• Interactional patterns: When the interactions of all thestudent groups’ were analysed, it was discovered thatthe 1:1 groups had practised four interactionalpatterns: ‘ideal’ (the most popular), ‘leader’, ‘tete-a-tete’, and ‘fragmented’. The 1:m groups, on the otherhand, assumed the ‘ideal’ and ‘fragmented’ (the mostpopular) patterns but not the other two. It is arguedthat the 1:1 groups demonstrated better quality inter-actions compared with those of the 1:m groups, as theformer setting facilitated greater autonomy for indi-vidual students, thus enhancing their level of partici-pation in collaborative learning. In contrast, therewere isolated students in some of the 1:m groups,resulting in far-from-ideal group interactions.

• Student learning attitudes and perceptions: Thestudent interviews and the questionnaires revealedtheir perception that computer-supported CCM isconducive to enhancing group collaboration and peersupport, as well as increasing their interest in theSocial Studies subject. Most students believed thatconcept mapping could be applied to other subjects.Furthermore, they felt that the gs software was easyto learn and could assist them in expressing theirviews.

Table 6 summarizes the differences between the 1:1 and1:m settings that were observed in this study.

Informed by the said findings, the following recom-mendations are offered to educators who are keen toadopt the design of this learning activity. First, if there isinsufficient equipment available for student access andtherefore the 1:m setting is the only viable option, theteacher is advised to monitor all the student groups tocheck whether or not any student dominates the TabletPC or is isolated. To avoid such situations, the teachermay design additional activities or measures in order totrain the students in established collaborative manage-ment strategies, such as taking turns to use the Tablet Ta

ble

6.Su

mm

ary

ofth

ean

alys

isof

the

1:1

and

1:m

sett

ings

inth

est

udy.

1:1

gro

up

s1:

mg

rou

ps

Sett

ing

sO

ne

Tab

let

PCp

erst

ud

ent

for

colla

bo

rati

veco

nce

pt

map

pin

gO

ne

Tab

let

PCp

erst

ud

ent

gro

up

for

colla

bo

rati

veco

nce

pt

map

pin

gLe

arn

ing

effe

cts

No

sig

nifi

can

td

iffe

ren

ceLe

arn

ing

rete

nti

on

No

sig

nifi

can

td

iffe

ren

ceA

vera

ge

sco

res

of

con

cep

tm

aps

No

sig

nifi

can

td

iffe

ren

ceSt

and

ard

dev

iati

on

s(S

D)o

fth

eco

nce

pt

map

sco

res

amo

ng

the

gro

up

sLa

rger

SDSm

alle

rSD

Co

llab

ora

tive

con

cep

tm

app

ing

pro

cess

Gre

ater

dif

fere

nce

inth

ep

roce

ssam

on

gth

eg

rou

ps

Gre

ater

con

sist

ency

inth

ep

roce

ss;g

reat

eref

fici

ency

ing

rou

pd

iscu

ssio

ns

Inte

ract

ion

alp

atte

rns

‘Idea

l’,‘L

ead

er’,

‘Tet

e-a-

tete

’,‘F

rag

men

ted

’‘L

ead

er’,

‘Fra

gm

ente

d’

Perc

epti

on

sin

colla

bo

rati

vele

arn

ing

Hig

her

mea

nsc

ore

inth

eq

ues

tio

nn

aire

Low

erm

ean

sco

rein

the

qu

esti

on

nai

rePe

rcep

tio

ns

inth

eso

ftw

are

Hig

her

mea

nsc

ore

inth

eq

ues

tio

nn

aire

Low

erm

ean

sco

rein

the

qu

esti

on

nai

rePe

rcep

tio

ns

inco

nce

pt

map

pin

gH

igh

erm

ean

sco

rein

the

qu

esti

on

nai

reLo

wer

mea

nsc

ore

inth

eq

ues

tio

nn

aire

Inte

ract

ion

sB

ette

rco

mm

un

icat

ion

and

inte

ract

ion

s;g

rou

pm

emb

ers

par

tici

pat

ing

mo

reco

nsi

sten

tly

inth

eac

tivi

ties

Less

-th

an-s

atis

fact

ory

com

mu

nic

atio

nan

din

tera

ctio

ns;

som

eg

rou

pm

emb

ers

wer

eis

ola

ted

Suit

able

for

Suffi

cien

teq

uip

men

tIn

suffi

cien

teq

uip

men

tR

eco

mm

end

atio

ns

for

imp

rove

men

tEn

han

ceco

llab

ora

tive

skill

s;p

reve

nt

gro

up

mem

ber

sfr

om

inte

rru

pti

ng

each

oth

erIn

corp

ora

teg

rou

pm

on

ito

rfo

rsh

ared

dis

pla

yin

ord

erto

pre

ven

tg

rou

pm

emb

ers

fro

mb

lock

ing

each

oth

er’s

view



PC. Furthermore, as a compromise solution, we recom-mend additional projecting equipment or an externalmonitor for the 1:m setting for all members of eachstudent group to view their work-in-progress conceptmaps at the same time. Such an arrangement may boostthe cost of the learning environment but is still relativelyless expensive than 1:1 setting. It is hoped that such anarrangement could reduce the potential problem ofstudents becoming isolated and thereby increasegeneral participation within the groups.

If a 1:1 setting is feasible, the teacher should managethe potential problems of lack of group work skills andinconsistent results of concept mapping by providingadditional training in general collaborative strategiesand, specifically, CCM and conflict resolution skillswith the1:1 setting.

Finally, in the aspect of the software system, theincorporation of an asynchronous discussion board ordatabase is recommended for storing the students’concept maps for after-class revision and follow-up dis-cussions, so that student learning and reflections can beextended beyond the classroom hours. Furthermore, weenvisage that additional functionalities will be incorpo-rated into the new gs version, namely: (1) a feature fortidying up the visual representations of concept maps,which would be beneficial to student learning andteacher evaluation; and (2) a feature for logging thelearning process for the teacher’s own reference, forexample, to record the notes posted and deleted by everystudent. Whereas the first proposed feature is mainlyfor addressing practical needs, the second one wouldbenefit both the researchers and teachers for analysingstudents’ CCM processes and identifying their deficien-cies in learning or collaborations.

Acknowledgements

This study was supported by the National ScienceCouncil, Taiwan (NSC 97–2511-S-134–002-MY3).The authors wish to thank SRI International for provid-ing them with a license to use and adapt the gs softwarein the present study.

References

Avouris N., Komis V., Margaritis M. & Fiotakis G. (2004)An environment for studying collaborative learningactivities. Educational Technology & Society 7, 34–41.

Basque J. & Lavoie M.C. (2006) Collaborative conceptmapping in education: major research trends. Proceedingsof the International Conference on Concept Mapping,79–86, San José, Costa Rica.

Cañas A., Coffey J.W., Carnot M.J., Feltovich P., HoffmanR.R., Feltovich J., & Novak J.D. (2003) A summary of lit-erature pertaining to the use of concept mapping techniquesand technologies for education and performance support.Report to the Chief of Naval Education and Training,Pensacola, Florida, IHMC.

Carlson C.B. (2007) A case study of pilot one-to-one laptopinitiative in a high computing Catholic high school. DEdThesis, Wichita State University, Wichita, KS, USA.

Chan T.W. (2010) How East Asian classrooms may changeover the next 20 years. Journal of Computer AssistedLearning 26, 28–52.

Chan T.-W., Roschelle J., Hsi S., Kinshuk M., Sharples M.,Brown T., Patton C., Cherniavsky J., Pea R., Norris C.,Soloway E., Balacheff N., Scardamalia M., Dillenbourg P.,Looi C., Milrad M., & Hoppe U. (2006) One-to-onetechnology-enhanced learning: an opportunity for globalresearch collaboration. Research and Practice inTechnology-Enhanced Learning 1, 3–29.

Chaudhury S.R., Roschelle J., Schank P., Brecht J. & Tatar D.(2006) Coordinating student learning in the collaborativeclassroom with interactive technologies. Paper presented atthe 3rd International Society for the Scholarship of Teach-ing and Learning Conference (ISSOTL 2006), WashingtonD.C., USA.

Chen G.D., Wang C.Y. & Ou K.L. (2003) Using groupcommunication to monitor web-based group learning.Journal of Computer Assisted Learning 19, 401–415.

Cheng A.S., Kruger L.E. & Daniels S.E. (2003) “Place” as anintegrating concept in natural resource politics: proposi-tions for a social science research agenda. Society andNatural Resources 16, 87–104.

Chiou C.C. (2009) Effects of concept mapping strategy onlearning performance in business and economics statistics.Teaching in Higher Education 14, 55–69.

Chiu C.H. (2003) Exploring how primary school studentsfunction in computer supported collaborative learning.International Journal of Continuing Engineering Educa-tion and Lifelong Learning 13, 258–267.

Chiu C.H., Huang C.C. & Chang W.T. (2000) The evaluationand influence of interaction in network supported collabora-tive concept mapping. Computers & Education 34, 17–25.

Coutinho C.P. (2009) Individual versus collaborativecomputer-supported concept mapping: a study with adultlearners. Proceedings of World Conference on E-Learningin Corporate, Government, Healthcare, and Higher Educa-tion 2009, 1173–1180, Vancouver, Canada.



De Jong T., Specht M. & Koper R. (2008) A reference modelfor mobile social software for learning. InternationalJournal of Continuing Engineering and Lifelong Learning18, 118–138.

Dieterle E. & Dede C. (2007) Building university faculty andstudent capacity to use wireless handheld devices forlearning. In Ubiquitous Computing in Education: InvisibleTechnology, Visible Impact (eds M. van ‘t Hooft & K.Swan), pp. 303–328. Lawrence Erlbaum Associates,Mahwah, NJ.

Dillenbourg P. (2010) Keynote speech. IEEE InternationalConference on Wireless, Mobile, and Ubiquitous Technol-ogy in Education 2010, Kaohsiung, Taiwan.

Fidas C. & Komis V. (2001) The architectural design of areal time collaborative concept-mapping environment fromdistance. P. Dillenbourg & A. Eurelings, Proceedings ofEuro Computer Supported Collaborative Learning, 656–659.

Fischer F., Bruhn J., Gräsel C. & Mandl H. (2002) Fosteringcollaborative knowledge construction with visualizationtools. Learning and Instruction 12, 213–232.

Guildry A., Cuthrell K., O’Connor K. & Good A. (2010) Fromthe green mile to the yellow brick road: using a practicalmodel to fill in the social studies content gaps. The SocialStudies 101, 22–29.

Hoppe U. & Gabner K. (2002) Integrating collaborativeconcept mapping tools with group memory and retrievalfunctions. Proceedings of Computer Support for Collabo-rative Learning 2002, 716–725, Colorado, USA.

Hu H.W. & Chiou G.F. (2008) Interaction characteristics ofstudent dyads in solving network-supported collaborativetasks. National Taiwan Normal University 53, 131–154.

Huang H.H. (2001) A study on the interactive process ofonline learners: a case study on text-based communication.Master Dissertation, National Chung Cheng University,Chia-Yi, Taiwan (in Chinese).

Kinchin I.M., De-Leij F.A.A.M. & Hay D.B. (2005) The evo-lution of a collaborative concept mapping activity forundergraduate microbiology students. Further and HigherEducation 29, 1–14.

Komis V., Avouris N. & Fidas C. (2002) Computer-supportedcollaborative concept mapping: study of synchronous peerinteraction. Education and Information Technologies 7,169–188.

Kramer-Dahl A., Teo P. & Chia A. (2007) Supporting knowl-edge construction and literature talk in Secondary SocialStudies. Linguistics and Education 18, 167–199.

Kwon S.Y. & Cifuentes L. (2009) The comparative effect ofindividually-constructed vs. collaboratively-constructedcomputer-based concept maps. Computers & Education52, 365–375.

Lai C.-H., Yang J.-C., Chen F.-C., Ho C.-W. & Chan T.-W.(2007) Affordances of mobile technologies for experientiallearning: the interplay of technology and pedagogical prac-tices. Journal of Computer Assisted Learning 23, 326–337.

Li J.Y. (2006) A study on the interactive processes in Internet-based project-based learning. Chinese Journal of ScienceEducation 14, 101–120.

Lin C.-P., Liu K.-P. & Niramitranon J. (2008) Tablet PC tosupport collaborative learning: an empirical study ofEnglish vocabulary learning. Proceedings of IEEE Interna-tional Conference on Wireless, Mobile, and UbiquitousTechnology in Education 2008, 47–51, Beijing, China.

Lin C.P., Wong L.H., Shao Y., Niramitranon J. & Tong C.J.(2010) 1:1 learning technology to support collaborativeconcept mapping: a case study of Social Studies lesson inelementary school. Proceedings of IEEE InternationalConference on Wireless, Mobile, and Ubiquitous Technolo-gies in Education 2010, 3–10, Kaohsiung, Taiwan.

Liu C.C. & Kao L.C. (2007) Do handheld devices facilitateface-to-face collaboration? Handheld devices with largeshared display groupware to facilitate group interactions.Journal of Computer Assisted Learning 23, 285–299.

Looi C.K., Lin C.P. & Liu K.P. (2008) Group Scribbles tosupport knowledge building in jigsaw method. IEEE Trans-actions on Learning Technologies 1, 157–164.

Looi C.-K., Ogata H. & Wong L.-H. (2010a) Technology-transformed learning: going beyond the one-to-one model?Workshop Proceedings of the 18th International Confer-ence on Computers in Education, 175–176, Putrajaya,Malaysia.

Looi C.-K., Seow P., Zhang B., So H.-J., Chen W. & WongL.-H. (2010b) Leveraging mobile technology for sustain-able seamless learning: a research agenda. British Journalof Educational Technology 41, 154–169.

Madrazo L. & Vidal J. (2002) Collaborative concept mappingin a web-based learning environment: a pedagogic experi-ence in architectural education. Journal of EducationalMultimedia and Hypermedia 11, 345–363.

Mifsud L. & Mørch A.I. (2010) Reconsidering off-task: acomparative study of PDA-mediated activities in fourclassrooms. Computer Assisted Learning 26, 190–201.

Milson F. (1973) An Introduction to Group Work Skill. Rout-ledge and Kegan Paul, London.

Mintrop H. (2004) Fostering constructivist communities oflearners in the amalgamated multiple-discipline of socialstudies. Curriculum Studies 36, 141–158.

Nesbit J.C. & Adesope O.O. (2006) Learning with conceptand knowledge maps: a meta-analysis. Review of Educa-tional Research 76, 413–448.

Norris C. & Soloway E. (2002) Keynote speech. Paperpresented at the IEEE International Workshop on Wireless



and Mobile Technologies in Education 2002, Växjö,Sweden.

Norris C. & Soloway E. (2004) Keynote speech. Paper pre-sented at the International Conference on Intelligent Tutor-ing Systems 2004, Maceio, Brazil.

Novak J.D. & Gowin D.B. (1984) Learning How to Learn.Cambridge University Press, London.

Rogers Y., Connelly K., Hazlewood W. & Tedesco L. (2010)Enhancing learning: a study of how mobile devices canfacilitate sensemaking. Personal and Ubiquitous Comput-ing 14, 1–14.

Roschelle J., Patton C. & Tatar D. (2007) Designing net-worked handheld devices to enhance school learning.Advances in Computers 70, 1–60.

Roth W.M. (1995) Authentic School Science Knowing andLearning in Open-Inquiry Science Laboratories. Kluwer,Dordrecht.

Strauss A., Corbin J. (1990) Basics of Qualitative Research:Grounded Theory Procedures and Techniques. Sage,Newbury Park, CA .

Roth W.M. & Roychoudhury A. (1994) Science discoursethrough collaborative concept mapping: new perspectivesfor the teacher. Science Education 16, 437–455.

Wong L.-H. (2010) Themed-based invited speech(CUMTEL). International Conference on Computers inEducation 2010, Putrajaya, Malaysia.

Wong L.-H. & Looi C.-K. (2010) Vocabulary learning bymobile-assisted authentic content creation and socialmeaning making: two case studies. Journal of ComputerAssisted Learning 26, 421–433.

Wong L.-H., Chin C.-K., Chen W. & Gao P. (2009)V.S.P.O.W.: an innovative collaborative writing approach toimprove Chinese as L2 pupils’ linguistic skills. Proceed-ings of the International Conference on Computer Sup-ported Collaborative Learning, 651–661, Rhodes Island,Greece.

Wong L.-H., Chin C.-K., Tan C.-L. & Liu M. (2010) Students’personal and social meaning making in a Chinese idiommobile learning environment. Educational Technology &Society 13, 15–26.

Yu M.-N., Pan Y.-F. & Lin W.-W. (1996) Concept mapping: tobe collaborative or individual. Journal of Educational &Psychological 19, 93–124. (In Chinese).



comparison of 1:1 and 1:m cscl environment for collaborative concept mapping

Documents