You are on page 1of 13

Virtual Reality

https://doi.org/10.1007/s10055-018-0358-z

S.I. : VR IN EDUCATION

Hybrid learning environment: Collaborative or competitive learning?


Calixto Gutiérrez‑Braojos1   · Jesus Montejo‑Gamez2 · Ana Marin‑Jimenez3 · Jesús Campaña4

Received: 7 October 2017 / Accepted: 6 July 2018


© Springer-Verlag London Ltd., part of Springer Nature 2018

Abstract
The aim of this study is to analyze the effect of different conditions of face-to-face learning on the participation and learning
of students in the virtual community. In particular, we analyze the effect of collaboration vs intergroup competition learning
on interdependence, regulation of socio-cognitive conflict, participation, and the quality of the contributions in a virtual
learning platform. This study took place in an educational research course. The participants in the full investigation were 36
(94.44% females) undergraduates enrolled in the subject of educational research, which was part of a 4-year social education
degree program. A Latin Squares design was applied to carry out the full investigation. The results show that conditions of
collaboration without competition in face-to-face environments facilitate better quality of learning to a greater extent. In
addition, there are no benefits of intergroup competition over individual competition learning. Thus, in this study, collabo-
rative learning designs without any competition could be considered more consistent with the goal that underpins the main
right to education: Each student is able to achieve what he has to achieve, taking advantage of the abilities of each and every
agent in the learning community (teacher and students) to do so. These results are not consistent with previous studies. We
suggest that characteristics of the sample could explain these differences between studies.

Keywords  CSCL · SOLO taxonomy · Socio-cognitive conflict · Positive interdependence · Intergroup competition ·
Collaborative learning

1 Introduction

The fast growth and improvement in online technologies in


* Calixto Gutiérrez‑Braojos recent years has elicited the appearance and development of
calixtogb@ugr.es
virtual learning (VL), which is usually organized through
Jesus Montejo‑Gamez virtual communities. Virtual learning communities refer to
jmontejo@ugr.es
cyberspaces in which individuals and groups of individuals
Ana Marin‑Jimenez accomplish their learning goals (Yang et al. 2007). Dillen-
anamarin@ugr.es
bourg et al. (2002) describe the main features of VL envi-
Jesús Campaña ronments. These authors also point out that virtual learn-
jesuscg@decsai.ugr.es
ing communities are not restricted to systems that include
1
Department of Research Methods in Education, Faculty 3D or virtual reality technology; instead, they can include
of Educational Sciences, University of Granada, Campus exclusively text-based interfaces. Use of VL environments in
Universitario Cartuja, s/n, 18011 Granada, Spain university degrees shows some advantages. First of all, Vir-
2
Department of Mathematical Education, Faculty tual Learning Technologies make it possible to implement
of Educational Sciences, University of Granada, Campus hybrid modalities. Thus, they constitute a useful tool for
Universitario Cartuja, s/n, 18011 Granada, Spain
distributing students’ workload between the classroom and
3
Department of Quantitative Methods, Faculty of Business home, according to European directives for higher educa-
Management, University of Granada, Campus Universitario
Cartuja, s/n, 18011 Granada, Spain
tion (European Commission 2015). These hybrid modalities
4
can be understood as ‘the thoughtful integration of class-
Department of Computer Sciences and Artificial Intelligence,
High School of Informatics and Telecommunications,
room face-to-face learning experiences with online learn-
University of Granada, Calle Periodista Daniel Saucedo ing experiences’ (Garrison and Kanuka 2004, p. 96). In this
Aranda, s/n, 18071 Granada, Granada, Spain

13
Vol.:(0123456789)
Virtual Reality

context, technology provides more opportunities for learn- can share and discuss their controversy. The coordination of
ing collaboratively, compared to purely face-to-face modali- viewpoints and their integration into a higher level represen-
ties (Gútierrez-Braojos and Salmerón-Pérez 2015). This tation than the previous one will produce learning improve-
approach increases the possibilities of creating an effective ments (Lacasa 1993).
learning environment and eliciting more positive responses However, not all socio-cognitive conflict situations nec-
from students than purely virtual learning (Westbrook 2006). essarily produce learning improvements. In fact, there are
In addition, it produces higher satisfaction levels for intro- two mechanisms of socio-cognitive conflict regulation that
verted students than face-to-face modalities (Downing and individuals can experience in social interaction situations:
Chim 2004). epistemic and relational (Doise and Mugny 1984; Mugny
Although virtual communities are useful resources for et al. 1978, 1978–1979; Mugny et al. 1999). Regulation
university formation, the use of these resources should be centered on epistemic conflict assumes that the learner’s
supported by suitable pedagogies. In other words, technol- attention is focused on the task. In this case, collaboration
ogies alone do not ensure pedagogical benefits. Teaching is directed toward building emerging and consensual knowl-
strategies are also essential for turning potential benefits edge that has more explanatory potential than knowledge
of VL into actual outcomes (Dillenbourg et al. 2002). In built separately (Butera and Mugny 2001; Doise and Mugny
particular, Computer-Supported Collaborative Learning 1984; Mugny et al. 1999). Regulation focused on relational
(CSCL) is becoming one of the most impacting learning and conflict deals with the evaluation and affirmation of self-
teaching perspectives in higher education. CSCL is a con- competence. Therefore, whereas the regulation of epistemic
structivist theory that defends the social nature of learning conflict leads to improve understanding and knowledge, rela-
(Koschmann 1996; Stahl 2006). Introducing CSCL means tional regulation leads to the defense of each actor’s own
jointly tackling the use of VL technologies and collaborative capacity or competence through the affirmation of his/her
learning. CSCL is not just employed in purely online learn- own point of view. In a subconscious effort to protect their
ing. It is also designed to extend face-to-face collaborative identity, learners try to be right, even when this means miss-
learning to virtual communities through software packages ing valuable learning opportunities. Arguments expressed
(Stahl et al. 2006). Most virtual environments overlap with in the regulation of relational conflict are directed toward
face-to-face scenes, and so virtual learning communities assessing one’s own competence, even when this stands in
are useful to enrich classroom activities (Dillenbourg et al. the way of learning. By contrast, in epistemic regulation, the
2002). Therefore, teachers face the challenge of creating interchange of opinions is directed toward conflict resolu-
hybrid environments in order to improve collaborative rela- tion. Thus, the quality of reasoning is improved, and learners
tions and, thus, the quality of students’ learning (Álvarez and who are engaged in the knowledge problem benefit from a
Mayo 2009; Basheri et al. 2013; Carter et al. 2017; Echever- broadened viewpoint (Mugny and Doise 1978). Therefore,
ría et al. 2017; Mehlenbacher 2010). progress in the learning process will be gained through the
coordination and integration of different viewpoints, lead-
1.1 Collaborative learning ing to the creation of a shared conceptualization at a higher
level of abstraction.
Collaborative learning is generally understood as a pro-
cess that fosters different forms of interaction that lead to 1.2 Making collaborative learning effective
significant advancement of shared ideas to solve learning
problems. Learning to work on a team is one of the most Research suggests that just asking students to collaborate
important skills professionals must develop to address the does not guarantee higher quality learning (Neugebauer et al.
challenges of the twenty-first century (Fischer et al. 2007). 2016). Students’ goal structures should be taken into account
However, it is not the only reason collaborative learning is (Deutsch 1949). Indeed, effective collaborative learning is
reaching a central role in educational practices. Collabora- achieved by stressing positive interdependence (Deutsch
tive learning facilitates learning to collaborate as well as 2006; Johnson and Johnson 2004), which increases whenever
collaborating to learn. Several authors have pointed out individuals move their goals closer together. Teaching strate-
the benefits of collaborative conditions for the quality of gies and tools have an impact on students’ goal structures and,
the learning (Deutsch 2006; Johnson and Johnson 1989). thus, on both attitudes and/or the quality of the learning, as
In social interaction situations, learners can exhibit oppos- discovered by Baer et al. (2010), Cheng-Huan and Chiung-
ing centrations or different levels of development that Hui (2016), Johnson and Johnson (1989), Ke and Grabowski
lead to socio-cognitive conflict. The social and cognitive (2007), or Yu (2001), for instance. The notion of positive inter-
dimensions are involved because divergent arguments are dependence appeared for the first time within Social Interde-
expressed on a relational level (Doise et al. 1975; Mugny and pendence Theory (Deutsch 1949). Society makes people set up
Doise 1983). Through constructive dialogue, the learners relationships in such a way that one person’s actions may affect

13
Virtual Reality

another’s goal achievement (Table 1). Social interdependence learning relations. The question then is whether these sorts of
theory analyzes the general case: when the actions of all indi- relations have positive effects on learning.
viduals affect everyone else’s outcomes.
As Deutsch’s theory states, goal structures determine inter- 1.3 Beyond positive interdependence: intergroup
actions between individuals, whereas the interaction patterns competition
determine the outcomes of a certain situation. Here, a goal
refers to a desired future of the state of affairs; thus, a goal When intergroup competition is introduced as a learning
structure is a specification of the type of interdependence condition, individuals are usually supposed to collaborate
among individuals’ goals. Deutsch (2006) proposed two types with their group mates. Combinations of intragroup collabo-
of goal interdependence. The first one is positive interdepend- ration and intergroup competition involve a double effect:
ence, which arises when the probability of one person’s goal On the one hand, the groups must make a greater effort
attainment is correlated (positively) with the likelihood of to achieve a competitive goal. On the other hand, there is
another obtaining his goal. The second one is negative inter- simultaneous collaboration within each group, which makes
dependence, which arises in the opposite case. learning easier (thanks to socio-cognitive conflict) and may
The set of different individuals’ sequential or simultaneous lead individuals to counteract negative emotional effects
actions that affect the outcomes of other actors engaged in the (because they are part of a group). Nevertheless, it is not
situation is defined as an interaction (Johnson and Johnson clear what effects intergroup competition has on learning.
2005). Three types of interaction are defined in Social Inter- This question has concerned researchers in recent years, but
dependence Theory: (1) Actions that increase the probability their results seem to be inconclusive. A summary can be
of others’ success in achieving their goals are promotive inter- found in Table 2.
actions; (2) actions that increase the probability of one’s own Regarding the regulation of socio-cognitive conflict,
success and reduce other individuals’ possibilities of success studies about competition and collaboration reveal that stu-
are oppositional interactions; and (3) when individuals take dents who participate in competitive learning environments
actions that promote their own success without affecting the are more likely to focus on relational conflict. In this way,
goal attainment of others, there are no interactions. Positive motivation may decrease, as well as attitudes toward learn-
interdependence gives rise to promotive interactions, whereas ing and academic outcomes (Butera et al. 2010). Likewise,
negative interdependence results in oppositional interactions Mugny and his colleagues found that collaborative condi-
(Johnson and Johnson 2005). tions promote the regulation of epistemic conflict. Receiving
Different situations such as constructive/destructive pro- criticism, as long as it does not interfere with the epistemic
cesses of conflict resolution or cooperative/competitive inter- process, may lead the receiver to integrate and take advan-
action when solving a problem or learning can be analyzed tage of it and progress (Doise and Mugny 1984; Quiamzade
using Social Interdependence Theory (Deutsch 2006). Within et al. 2013). As a consequence, epistemic conflict prevails
a learning community (like a classroom), cooperative rela- over relational conflict when the interaction is not perceived
tions appear when the goals of the members are positively as a threat. In this case, individuals not only consider their
interdependent (promotive interactions), whereas competi- own standpoint, but they are also capable of empathizing
tive relations emerge when there is negative interdependence and, thus, are more likely to learn. By contrast, in competi-
among goals (oppositional interactions). When no interactions tive conditions, the climate may be perceived as a dispute. In
appear in a group, members are working individually in the this case, individuals question the validity of every figure or
community. There is one a priori goal structure of positive piece of information, regardless of the source, and so regula-
interdependence in learning situations, due to the fact that tion is more centered on relational conflict, specifically on
all the members share the same goal: to learn. However, in oneself and the interlocutor (Mugny et al. 1984, 1975–1976,
practical situations, the classroom dynamics are full of com- 1978–1979).
petitive relations. Personal concerns, disagreements between Regarding the outcomes of learning and other attitudinal
classmates, or grades are some of the causes that may lead to dimensions, a wide range of research approaches exist. Tauer
losing focus. Hence, teaching methods play an essential role and Harackiewicz (2004) found that both collaborative and
in emphasizing positive interdependence, that is, cooperative intergroup competitive conditions have positive aspects, so

Table 1  Interrelations among Others’ actions facilitate one’s goal achievement


individuals Source: Johnson and
Johnson (2005) Yes No

Others’ actions facilitate one’s goal achievements Yes Interdependence Dependence


No Independence Helplessness

13
Virtual Reality

Table 2  Summary of research about collaborative (C) and intergroup competitive (IC) learning
C improves cognitive competencies or Cognitive competen- IC improves cognitive competencies or
creativity cies are not measured creativity

C improves attitudes or climate Butera et al. (2010), Johnson and John- Wang et al. (2017) Ke and Grabowski (2007)
son (1989), Tauer and Harackiewicz
(2004)
Attitudes or climate are not measured Quiamzade et al. (2013) Baer et al. (2010), Regueras et al.
(2009), Arnab et al. (2015)
IC improves attitudes or climate Yu (2001) Cheng-Huan and Chiung-Hui (2016),
Tauer and Harackiewicz (2004)

that proper instruction that includes both can lead to high the most effective for producing greater participation and
motivation and quality learning. In turn, Baer et al. (2010) learning in the virtual community? Thus, the objective of
found better levels of creativity for low and high competi- this study is to analyze the effect of different face-to-face
tion and worse levels for intermediate competition. From learning conditions (cooperative, intergroup competition,
the CSCL approach, there are studies that show better stu- and inter-individual competition; see condition parameter
dent engagement, learning achievement, and creativity when in methodological issues section) on the participation and
groups compete with each other (Cheng-Huan and Chiung- learning of students in the virtual community.
Hui 2016 in higher education and Chen and Hwang 2016 The hypothesis in this study is that students working in
in elementary school). Likewise, some papers reveal better competitive intergroup conditions in the small-group sphere
attitudes toward the subject and more positive interpersonal show higher levels of positive interdependence than in other
relationships in the intergroup competitive case (Yu 2001). learning conditions. In addition, competitive intergroup
Other investigations indicate that competitive situations pro- conditions in the small-group sphere turn into cognitively
mote learning, whereas purely cooperative relations improve deeper contributions in the virtual community sphere. In
positive attitudes (Ke and Grabowski 2007). When hybrid turn, students in the collaborative condition tend to tackle
modalities are implemented, Regueras et al. (2009) show epistemic conflict more, show greater positive interdepend-
that collaborative learning with intergroup competition is ence, and make cognitively deeper contributions than stu-
more effective than pure collaborative learning. Arnab et al. dents in the inter-individual competition condition.
(2015) also reach similar conclusions, although their results
are less significant. Other studies in hybrid environments
explore the comparison of different levels of competition. 2 Methodological issues
Yu et al. (2008) found that students in anonymous team
competition conditions exhibited significantly better levels 2.1 Participants
of motivation and satisfaction than those in the face-to-
face condition. In turn, Wang et al. (2017) observed greater This study took place at the University of Granada (Spain).
engagement in collaborating in an unrewarded competitive The participants in the full investigation were 36 (94.44%
regime than in a zero-sum competitive situation. Finally, females) undergraduates enrolled in the subject of educa-
no previous studies were found that analyze the relation- tional research, which was part of a 4-year social education
ship between learning conditions in a face-to-face environ- degree program. The students’ participation in the virtual
ment and the quality of the contributions in virtual learning environment was obligatory, making up 40% of the final
communities. grade.
Our research question arises at this point. We wonder
whether learning conditions in the face-to-face environment 2.2 Learning environment
influence the quality of learning students show in the virtual
community. This paper intends to provide clues that help to Analogous to professional research activity, Gutiérrez-
understand the effects of different types of interdependence Braojos and Salmerón (2012) suggested that work in the
on hybrid learning environments. classroom should be structured around three spheres of con-
structive activity in order to optimize the effectiveness of
1.4 Problem, objective, and hypothesis a learning environment: individual, collaborative in small
groups, and communitary. This proposal has been followed
We propose the following question: In a hybrid environment, in this study, and so students worked in every sphere. In the
what learning condition in the face-to-face environment is individual sphere, activity is carried out by a single student

13
Virtual Reality

in one portfolio per content unit. The portfolio includes per- each phase. Condition A (collaboration): The groups col-
sonal reflections about the topics and some ideas of other laborated with each other to resolve the activities proposed.
schoolmates that the student perceived as the most impact- Each student who participated in this condition got a bonus
ing. The choice of each of the selected ideas must be justi- point when every group correctly resolved the unit activi-
fied. Results for this sphere are not included in the follow- ties. Condition B (intergroup competition): Groups com-
ing section because they are not related to the researchers’ peted with each other to resolve activities. The first group to
objective. In the face-to-face sphere, activity is based on correctly resolve the activities got a bonus point. Condition
small-group collaboration. There were four small groups per C (inter-individual competition): This was a ‘paradoxical’
content unit in the classroom. Each group had three mem- condition. Students could only collaborate with students in
bers in order to facilitate comprehension of the subject. In their group. However, only those students who were the first
the virtual community sphere, collaborative activities were to correctly solve the activities in the face-to-face sphere got
designed to reflect on and solve several tasks or knowledge bonus points, regardless of the group membership base. In
problems per content unit. In each unit of content, there were other words, we proposed that these degrees of collaboration
activities with three levels of complexity: comprehension, can be understood under a continuum from high interde-
application, and reflection. For example, in the second unit pendence to zero interdependence in a learning environment:
about mixed-methods research, students had to identify high interdependence (collaboration among the whole class
conceptual mistakes from propositions about main ideas group, i.e., without competition); close to medium interde-
from the unit and identify types of mixed methods from pendence (collaborative learning intragroup, and competi-
an authentic case (comprehension task). Later, they had to tion intergroup); and close to lack of interdependence (col-
provide an adequate mixed method for a given authentic laborative learning intragroup with competition among all
research problem. Finally, they had to discuss and elabo- the students).
rate a shared reflection about the mixed-method unit. The
unit was developed in the Knowledge Forum (KF) virtual 3.1 Information‑gathering instruments
platform to facilitate asynchronous communication (Scarda- and application procedure
malia 2004). The Knowledge Forum environment includes:
(1) a main menu with tools designed to edit collaborative To achieve the objectives of this study, it is not necessary
knowledge building. For this purpose, the KF offers an inter- to analyze and present results on the sphere of individual
face that provides interaction scaffolds to facilitate collective activity. Thus, useful information was collected from two
knowledge building (e.g., ‘I need to understand,’ ‘a better sources: the small-group (face-to-face) sphere and the virtual
theory,’ ‘putting our knowledge together’); (2) a configura- community sphere.
tion menu to set up the basic functions of the virtual envi-
ronment (type of scaffold to use, display of contributions’ 3.1.1 Small‑group sphere
authorship, etc.); (3) tools to browse and create discussion
windows (Fig. 1); and (4) a menu to analyze and assess The students’ perception of positive interdependence was
activity within the platform. obtained by applying the positive interdependence subscale
(Ghaith et al. 2007). This subscale is part of a longer scale,
the Classroom Life Measure (Johnson and Johnson 1996).
3 Method The subscale applied is composed of six Likert-type items
rated on a 5-point scale, where a rating of ‘1’ indicates that
A Latin Squares design was applied to carry out the full the item is very untrue and ‘5’ that it is very true, for their
investigation. It took two parameters into account (Table 3). group. In the present study, the Cronbach’s alpha values were
The ‘phase’ parameter refers to the time spent working good in two of the learning conditions (collaborative learn-
on each unit of content. This study analyzes three content ing condition: .96; competition intergroup condition: .94),
units in the subject of research methods in education. These and Cronbach’s alpha value was acceptable in the individual
content units correspond to three phases. All the groups have competition condition (.694). The instrument was adminis-
the same tasks in the same phase. In each of these phases, tered in the presence of the teacher during tutorial classes.
students work through individual, small-group, and virtual
community spheres. The ‘Condition’ parameter refers to 3.1.2 Virtual community sphere
conditions of collaboration and assessment of the learners
in the small-group (face-to-face) sphere, where students The students’ perceptions of socio-cognitive regulation were
were distributed in four small groups of three members gathered by applying the socio-cognitive regulation scale
each. There were three goal structures or conditions, and (Darnon et al. 2006). This scale is composed of two sub-
all the small groups worked under the same condition in scales (epistemic and relational subscale) with three items

13
Virtual Reality

Fig. 1  A discussion window (top) and a course menu (bottom) in the Knowledge Forum platform

13
Virtual Reality

Table 3  Diagram of the Latin Squares design


Phase

Condition PHASE 1 PHASE 2 PHASE 3


Condition A Group 1 Group 3 Group 2
Condition B Group 2 Group 1 Group 3
Condition C Group 3 Group 2 Group 1
Groups composition Random Consistent Consistent
combina- combina-
tion 1 tion 2

Configuration of phases 2 and 3 was selected depending on phase 1,


which was chosen by chance

each. Likert-type items on these subscales are also based on


a 5-point scale, where a rating of ‘1’ indicates that the item
is very untrue and ‘5’ that it is very true. The alpha values
were good for each epistemic regulation subscale in the three
conditions (collaboration learning condition: .93; competi- Fig. 2  SOLO taxonomy
tion intergroup condition: .87; and individual competition
condition: .91). However, in the case of the relational regu-
lation scale, Cronbach’s alpha value was good in the condi- (extended abstract level). Finally, other contributions, such
tion of collaborative learning without competition (.93); and as ‘good contribution, I agree with your idea’ or ‘thanks for
acceptable in the individual competition condition (.69); but your explanation, now I understand it better’ were classified
it was not acceptable in the intergroup competition condi- as socio-affective contributions.
tions (.51). The instrument was administered in the presence
of the teacher during tutorial classes.
3.2 Variables and analytical procedure
Information about participation and learning was col-
lected through automatized records on the Knowledge
To fulfill the objectives, the analyses were divided into three
Forum platform. In all, 379 contributions were recorded.
main stages. In the first stage, we applied an exploratory
A contribution refers to a student’s written message in the
descriptive analysis of the level of positive interdependence,
KF with the purpose of participating in the discourse on a
socio-cognitive conflict regulation, and activity in the VL
specific topic, and it was considered a unit of analysis in
environment. In the second stage, we carry out a Multivari-
this study. Every contribution was rated using the Structure
ate General Linear Model analysis in order to analyze differ-
of Observed Learning Outcome taxonomy (SOLO, Biggs
ences in the dependent variables regarding both factors, i.e.,
and Collis 1982), which has been widely used to evaluate
phases and conditions of learning. In addition, we applied
the learning on virtual platforms, specifically to analyze
a Bonferroni post hoc test of the number of contributions
the correction, structural complexity, and originality of the
in order to compare the pairs of learning conditions. Sub-
knowledge reflected in the contributions (e.g., Brown et al.
sequently, in the third stage, another descriptive analysis is
2006; Hatzipanagos 2006; Holmes 2005; Gútierrez-Braojos
included for the factor that explains the statistically signifi-
and Salmerón-Pérez 2015; Schrire 2006). The SOLO tax-
cant differences, including the variation coefficient. Data
onomy has five levels of complexity, meta-categorized into
were analyzed using SPSS (Version 18).
two levels (Fig. 2). On the one hand, the superficial level
comprises three type of contributions:(1) Pre-structural con-
tributions refer to incorrect or disconnected contributions to
relevant knowledge for the community that are overly sim- 4 Results
ple; (2) unistructural contributions are overly simple; and
(3) multi-structural contributions provide a large number 4.1 Exploratory descriptive analysis
of content elements, but in a disorganized way, lacking a
coherent conceptual structure (multi-structural level). On To fulfill the objectives, the analyses were divided into two
the other hand, the deep level includes relevant contribu- main stages. In the first stage, an analysis was performed of
tions that coherently integrate important aspects of the task the students’ perceptions of positive interdependence and
requirements (relational level) and/or contributions that socio-cognitive conflict regulation. The second stage cor-
involve generalizations, knowledge transference, and novelty responded to the analysis of the number of contributions

13
Virtual Reality

registered in the virtual sphere; i.e., number of deep cog- Table 4). Moreover, a discriminant analysis was performed
nitive, surface cognitive, and socio-affective contributions. to measure the degree of success of the classification made
On the one hand, the descriptive analysis of the students’ by the cluster analysis. The results revealed that 98.1% of
perceptions shows that the mean for the variable positive the original cases were correctly classified. The first clus-
interdependence is 3.65 (SD .86), for socio-cognitive rela- ter, made up of 27 students, presents medium–high scores
tional regulation, 3.67 (SD .64), and for socio-cognitive on positive interdependence, epistemic regulation, and
epistemic regulation, 3.52 (SD .83). Based on the variation relational regulation. These students showed a moderate
coefficients, .24 for the variables positive interdependence number of contributions: deep, surface, and socio-affective,
and socio-cognitive epistemic regulation, and .17 for the compared to the rest of the clusters. We call these individu-
variable socio-cognitive relational regulation, it can be said als ‘students moderately engaged with epistemic improve-
that the latter variable is more homogeneous than the others. ment.’ The second cluster, made up 15 students, presents
Similarities in the students’ perceptions of these variables high scores on positive interdependence and epistemic regu-
can be observed in Fig. 3 (left). lation, but low scores on relational contributions. Moreover,
On the other hand, a basic descriptive analysis of activity students who belong to this pattern show a high number of
in the Knowledge Forum platform shows that students gen- contributions: a high number of deep and socio-affective
erated 338 contributions. The content analysis of the con- contributions, but a low level of surface contributions. We
tributions using the SOLO taxonomy reveals that 57.69% call these people ‘students highly engaged with epistemic
are deep cognitive complexity contributions (relational or improvement.’ Finally, the third cluster, made up 66 stu-
abstract-extended contribution), whereas 26.04% are surface dents, shows low scores on positive interdependence and
cognitive complexity contributions. The rest of the contribu- epistemic regulation, but high scores on relational regula-
tions are socio-affective (Fig. 3, right). tion. This pattern presents students who show the lowest
In addition, a cluster analysis using the K-mean proce- number of deep contributions, but a high number of socio-
dure was performed, revealing three learning patterns (see affective contributions, compared to the rest of the patterns;

Fig. 3  Summary of the exploratory descriptive analysis: attitudes (left) and contributions in the KF platform (right)

Table 4  K-mean cluster Variables Cluster 1 Cluster 2 Cluster 3 Associated values


analysis: pattern of engagement
with collective learning Mean SD Mean SD Mean SD F p η2

Positive interdependence 3.87 .86 4.2 .81 3.42 .8 6.794 *** .115
Number of contributions 5.48 1.15 9 2.10 1.45 1.15 239.667 *** .820
Nº deep contributions 2.33 1.3 5.93 1.16 .65 .86 164.651 *** .758
Nº surface contributions 2.33 1.3 1.6 1.72 .59 .76 26.109 *** .332
Nº socio-affective contributions .81 .92 1.46 1.18 .22 .45 20.252 *** .278
Relational regulation 3.75 .53 3.06 .64 3.78 .61 9.140 *** .148
Epistemic regulation 3.62 .56 4.37 .68 3.28 .83 13.095 *** .2

13
Virtual Reality

however, they also show the highest level of surface contri- test was applied to find out which means showed statistically
butions. We call these individuals ‘students poorly engaged significant differences.
with epistemic improvement.’ Results do not show significant differences depending on
These results led us to test the hypothesis that the learn- the phase parameter, and so we excluded this factor from
ing group may be responsible for the differences between further analyses. By contrast, learning conditions seem to
the patterns found. be factors that have an influence on the dependent variables.
In particular, positive interdependence and the number of
4.2 Multivariate general linear model surface and deep contributions show significant differences
depending on the learning conditions, as do relational and
By applying multivariate generalized linear models, we epistemic regulations. However, we did not find these sig-
compared the means of our variables, depending on the nificant differences in the variables number of contributions
learning condition and phase factors. A post hoc Bonferroni and number of socio-affective contributions (Table 5).
A pairwise analysis of the learning conditions for each
dependent variable (Table 6) shows that perception of pos-
itive interdependence is significantly greater in the collab-
Table 5  Analysis of differences depending on learning conditions orative learning condition and significantly smaller in the
Variables Mean square F p η2p learning condition with individual competence. Regarding
the variables number of contributions and contributions of
Positive interdependence 21.36 58.501 *** .542
socio-affective type, no significant differences were found
Nº contributions 8.398 .871 .422 .017
between the groups. Nevertheless, we found significant
Nº deep contributions 33.33 8.802 *** .151
differences between the collaborative condition and both
Nº surface contributions 12.03 7.749 *** .135
of the competitive conditions on the other variables. In
Nº socio-affective contributions .481 .636 .531 .013
particular, when students work in collaborative learning
Relational regulation 7.621 29.503 *** .373
conditions, they generate a significantly higher number of
Epistemic regulation 11.54 22.748 *** .315
contributions with high cognitive complexity, and fewer

Table 6  Multiple comparisons between pairs of learning conditions


95% confidence interval
Dependent variables (I) First condition (J) Second condition Difference of p Lower limit Upper limit
means (I–J)

Positive Interdependence Condition A Condition B .681 *** .334 1.027


Condition C 1.537 *** 1.194 1.883
Condition B Condition C .856 *** .51 1.203
Nº contributions Condition A Condition B .916 .64 − 0.865 2.698
Condition C 0.722 .98 − 1.059 2.504
Condition B Condition C − 0.194 1 − 1.976 1.587
Nº deep contributions Condition A Condition B 1.677 .001 .549 2.783
Condition C 1.677 .001 .549 2.783
Condition B Condition C 0 1 − 1.117 1.117
Nº surface contributions Condition A Condition B − .972 .004 − 1.687 − .257
Condition C − 1.028 .002 − 1.742 − .312
Condition B Condition C − .055 1 − .771 .659
Nº socio-affective contributions Condition A Condition B .222 .843 − .277 .722
Condition C .056 1 − .443 .554
Condition B Condition C − .167 1 − .666 .332
Relational regulation Condition A Condition B − .778 *** − 1.069 − .486
Condition C − .814 *** − 1.107 − .523
Condition B Condition C − .037 1 − .329 .255
Epistemic regulation Condition A Condition B .926 *** .517 1.335
Condition C 1.028 *** .618 1.437
Condition B Condition C .102 1 − .307 .511

13
Virtual Reality

contributions with a low level of complexity, than the


other conditions. In addition, students perceive that dis-
cussions focus more on improving knowledge (epistemic
regulation of conflict) and less on the defense of personal
views (relational regulation of conflict). On all of these
variables, there are no differences between the learning
conditions with competition (collaborative learning condi-
tion with intergroup competition, and collaborative learn-
ing condition with individual competition).
On the other hand, the mean of contributions with cog-
nitive complexity in the collaborative learning condition is
3.28, exceeding the conditions with competition by almost
one point (Table 7). By contrast, the mean for the surface
contributions in the collaborative condition is .5, a point
below the conditions with competition. Likewise, in the
condition of collaboration without competition, the stu- Fig. 4  Distributions of the different kinds of contributions depending
on learning conditions
dents manifest a preference for the epistemic resolution
of socio-cognitive conflict, whereas in the competitive
conditions, there is a preference for the relational resolu- 5 Discussion and conclusions
tion. The perception of positive interdependence is greater
in the collaborative condition, with the lowest perception This empirical study was designed to understand the influ-
appearing in the condition of inter-individual competition. ence of goal structures on university students’ learning in a
In addition, it can be observed from the variation coeffi- hybrid collaborative learning environment. To achieve this
cients that when the work condition is collaborative, the aim, an exploratory analysis was performed that focused on
students’ perception of positive interdependence is more measuring positive interdependence, participation, resolu-
homogeneous, as occurs with the perception of the epis- tion of socio-cognitive conflict, and complexity of students’
temic regulation and the number of complex contribu- contributions, within a virtual environment. The existence of
tions. However, in the case of superficial contributions, significant differences between contributions across distinct
the least homogeneous condition occurs when the work learning conditions was also analyzed.
is collaborative. The results showed that students presented higher levels
As Fig. 4 shows, the number of complex contributions of positive interdependence when they worked under col-
is greater in all of the conditions, as it is 72.4% when the laborative learning conditions in the face-to-face sphere.
condition is collaborative, 46.39% for intergroup com- Regarding participation on the virtual learning platform,
petition, and 46.87% in the inter-individual competition although there were no significant differences in the num-
modality. Superficial contributions only represent 11.7% ber of contributions and socio-affective contributions, it was
when the work is collaborative, rising to 40.20% when observed that students who worked in collaborative learning
there is intergroup competition, and 33.3% when there is conditions in the face-to-face sphere provided more cogni-
individual competition. tively complex contributions and less surface contributions
than the other conditions. These results are not consistent

Table 7  Descriptive analysis: Condition A Condition B Condition C


activity on the KF platform
using SOLO taxonomy X SD M Cv% X SD M Cv% X SD M Cv%

Positive interdependence 4.38 .56 5 12.7 3.70 .73 3a 19.7 2.85 .46 3 16.1
Number of contributions 4.06 2.81 3 69 3.14 2.79 0 88.8 3.33 3.60 1 108
Nº deep contributions 3.28 1.97 2 60 2.25 1.62 1 72 2.81 2.19 1 77.9
Nº surface contributions .50 .65 0 130 1.47 1.34 0 91.2 1.53 1.54 1 101
Nº socio-affective contributions .64 .87 0 136 .42 .73 0 174 .58 .94 0 162
Relational regulations 3.15 .68 3 21.6 3.92 .34 4 8.67 3.96 .50 4 12.6
Epistemic regulation 4.18 .77 4 18.4 3.25 .64 3 19.7 3.15 .68 3 21.6

‘M’ stands for mode

13
Virtual Reality

with some of the studies mentioned above (Arnab et al. intragroup and competition intergroup); and close to lack
2015; Baer et al. 2010; Cheng-Huan and Chiung-Hui 2016; of interdependence (i.e., a ‘paradoxical’ condition in which
Ke and Grabowski 2007; or Regueras et al. 2009). The dis- students were allowed to collaborate with their learning
tribution of the cognitive load in the virtual community is intragroup and also competed with all the students. Students
a dimension of special interest in this study. The results could only collaborate with their group mates. However,
showed that the coefficient of variation in cognitively deep only those students who finished first and correctly solved
contributions within the virtual community is smaller than the activities in the face-to-face sphere got bonus points,
in the other conditions. This indicates that there is lower regardless of their group membership).
relative dispersion of knowledge throughout the community To sum up, designing learning environments in condi-
when its members collaborate to learn; i.e., collaborative tions of collaboration without competition in face-to-face
conditions seem to promote a more homogeneous distri- environments seems to facilitate better quality of learning to
bution of knowledge, a situation that suggests the notion a greater extent. We did not find any benefits of intergroup
of collective cognitive responsibility (Scardamalia 2002). competition or individual competition over collaborative
The results also reveal that when students participate in col- learning without competition. Thus, teaching methods based
laborative learning environments without competition in a on goal structures that stress cooperative relations foster
face-to-face environment, a greater number of participants higher quality learning than competitive conditions. There-
generate more contributions with cognitive quality and fewer fore, collaborative learning designs without any competi-
contributions with low cognitive quality in the virtual learn- tion could be considered more consistent with the goal that
ing environment. Although our findings do not support pre- underpins the main right to education: Each student is able to
vious empirical studies, they do agree with the conclusions achieve what he/she has to achieve, taking advantage of the
of Butera et al. (2010) and Quiamzade et al. (2013), who abilities of each and every agent in the learning community
found that collaborative conditions help to improve learning (teacher and students) to do so. This could provide relevant
better than competitive orientations. Likewise, this research insights about how to increase participation and improve
allows us to conclude that collaboration without competition achievement of students in the CSCL approach. However,
was more effective than other conditions where competition the size of the sample employed does not allow us to gen-
and collaboration were balanced. eralize. In this regard, we suggest that this study should be
In addition, and contrary to what we expected, we did replicated in different contexts, including a broader profile
not observe significant differences between the types of of students (various ages and degrees). The characteristics
competition. Thus, intergroup competitive conditions led to and size of the sample could explain these differences. In
more positive interdependence than inter-individual com- other words, applying competitive learning conditions in
petitive conditions. We did not find significant differences the entrepreneurial and economic degree could be a more
in the rest of the dependent variables. On the contrary, in the effective teaching strategy than other conditions without
comparison of intergroup competition and inter-individual competition. However, creating learning environments with
competition, no significant differences were found in the competitive learning conditions could be a bad strategy for
quality of learning. These results contrast with those of Yu students who are enrolled in social and educational careers.
et al. (2008), who noticed higher motivation in anonymous Thus, some students may feel stressed or uncomfortable with
rivals, and Wang et  al. (2017), whose results suggested learning assignments that require competition. It is relevant
greater engagement in the unrewarded competition than in to point out that our sample is mostly composed of women.
the rewarded one. We could infer that these findings partially Our results coincide with previous studies indicating that
support Johnson and Johnson’s (1989) and Deutsch’s (2006) women show a collaborative learning profile, whereas men
studies. These differences between our discoveries and the tend to exhibit a more individualist learning style (e.g., Yaz-
previous literature are meaningful. We suggest that these dif- ici 2005; Inkpen et al. 1995). In addition, our sample was
ferences could be explained by the conditions of our learning small, and so we wonder if the small number of participants,
environment. We created three learning conditions that could rather than the characteristics of the sample, could explain
be explained from the parameter called ‘degree of learning these differences. Therefore, other studies would be neces-
collaboration parameter,’ which differs from Deutch’s expla- sary in order to analyze bias due to gender and the size of
nation of interdependence. Deutch proposes two poles of the sample. In addition, some future lines of research can be
interdependence: positive and negative. We proposed that explored to add to the understanding of positive interdepend-
these degrees of collaboration can be understood on a con- ence and collaborative learning. One of them is related to
tinuum from high interdependence to zero interdependence Deutsch’s theory, where goal structures define a ‘one dimen-
in a learning environment: high interdependence (collabora- sional’ continuum of possibilities whose ends are positive
tion among the whole class group, i.e., without competition); and negative interdependence. When intergroup competition
close to medium interdependence (collaborative learning comes into play, this framework seems to break down. It is

13
Virtual Reality

not clear what level of positive interdependence should be Commission E (2015) ECTS users’ guide. Publications Office of the
expected a priori in an intergroup competitive (with internal European Union, Luxembourg
Darnon C, Muller D, Schrager SM, Pannuzzo N, Butera F (2006)
collaboration) goal structure. Our findings do not help to Mastery and performance goals predict epistemic and relational
understand whether there is a significant intermediate level conflict regulation. J Educ Psychol 98:766–776
between pure collaboration and inter-individual competi- Deutsch M (1949) A theory of cooperation and competition. Hum
tion. Moreover, as mentioned above, some studies present Relat 2:129–152
Deutsch M (2006) Cooperation and competition. In: Deutsch M,
a different picture where intergroup competitive conditions Coleman PT, Marcus EC (eds) The handbook of conflict resolu-
achieve the highest levels of cognitive competencies (right tion: theory and practice. Jossey-Bass, San Francisco, pp 23–42
row in Table 2). Thus, the question is how to fit intergroup Dillenbourg P, Schneider D, Synteta P (2002) Virtual learning envi-
competition within Social Interdependence Theory (is it ronments. In: Dimitracopoulou A (ed) 3rd Hellenic conference
‘information & communication technologies in education’,
some kind of ‘neutral’ condition? Should the theory open Kastaniotis editions, Rhodes, Greece, pp 3–18
up more dimensions?), or equivalently, how to interpret the Doise W, Mugny G (1984) The social development of the intellect.
notion of positive interdependence in these types of goal Pergamon Press, Oxford
structures. These questions must be elucidated in future Doise W, Mugny G, Perret-Clermont A (1975) Social interacction
and the development of cognitive operations. Eur J Soc Psychol
studies. 5:367–383
Downing K, Chim TM (2004) Reflectors as online extraverts? Educ
Acknowledgements  This study has been made possible by the pro- Stud 30(3):265–276
gram Jose Castillejo of the Ministry of Education, Culture and Sports; Echeverría L, Cobos R, Machuca L, Claros I (2017) Using collabora-
and the collaboration of the IKIT; and the department of Curriculum, tive learning scenarios to teach programming to non-CS majors.
Teaching and Learning at OISE, University of Toronto. Comput Appl Eng Educ 25:719–731
Fischer G, Rohde M, Wulf V (2007) Community-based learning: the
core competency of residential, research-based universities. Int
J Comput Support Collab Learn 2(1):9–40
Garrison DR, Kanuka H (2004) Blended learning: uncovering its
References transformative potential in higher education. Internet High Educ
7:95–105
Álvarez RB, Mayo IC (2009) Las tecnologías de la información y la Ghaith GM, Shaaban KA, Harkous SA (2007) An investigation of
comunicación en la educación superior. Estudio descriptivo y de the relationship between forms of positive interdependence,
revisión. Revista Iberoamericana de Educación 50(7):1–12 social support, and selected aspects of classroom climate. Sys-
Arnab S, Nalla M, Harteveld C, Lameras P (2015) An inquiry into tem 35:229–240
gamification services: practices, experience and insights. In: Pro- Gutiérrez-Braojos C, Salmerón P (2012) Building knowledge: theo-
ceedings of the international gamification for business conference, retical model to analyze knowledge builders. C. 16th annual
21–22 Sept 2015, Aston, Birmingham, UK. Knowledge Building Summer Institute, Toronto
Baer M, Leenders RTAJ, Oldham GR, Vadera AK (2010) Win or lose Gútierrez-Braojos C, Salmerón-Pérez H (2015) Exploring collective
the battle for creativity: the power and perils of intergroup com- cognitive responsibility and its effects on students’ impact in a
petition. Acad Manag J 53(4):827–845. https​://doi.org/10.5465/ knowledge building community. J Study Educ Dev 38:327–367
amj.2010.52814​611 Hatzipanagos S (2006) HOT and flaming spirals: learning and
Basheri M, Munro M, Burd L, Baghaei N (2013) Collaborative learn- empathic interfaces in discussion forum text-based dialogues.
ing skills in multi-touch tables for UML software design. Int J Adv European Journal of Open Distance and ELearning, Retrieved
Comput Sci Appl 4(3):60–66 06 July 2013 from http://www.eurod ​ l .org/mater ​ i als/contr​
Biggs JB, Collis K (1982) Evaluating the quality of learning: the SOLO ib/2006/ Stylianos_Hatzipanagos.htm
taxonomy. Academic Press, New York Holmes K (2005) Analysis of asynchronous online discussion using
Brown N, Smyth K, Mainka C (2006) Looking for evidence of deep the SOLO taxonomy. Aust J Educ Dev Psychol 5:117–127
learning in constructively aligned online discussions. In: Banks Inkpen K, Booth KS, Klawe M, Upitis R (1995) Playing together
S, Hodgson V, Jones C, Kemp B, McConnell D, Smith C (eds) beats playing apart, especially for girls. In: Proceedings of
Proceedings of the fifth international conference on networked CSCL ‘95, pp 177–181.
learning. Lancaster University, Lancaster, pp 315–322 Johnson DW, Johnson RT (1989) Cooperation and competition:
Butera F, Mugny G (2001) Conflict and social influences in hypothesis theory and research. Interaction, Edina
testing. In: De Dreu CKW, De Vries NK (eds) Group consensus Johnson DW, Johnson R (1996) Meaningful and manageable assess-
and minority influence implications for innovation. Blackwell, ment through cooperative learning. Interaction, Edina
Oxford, pp 160–182 Johnson DW, Johnson RT (2004) Cooperation and the use of technol-
Butera F, Darnon C, Mugny G (2010) Learning from conflict. In: Jetten ogy. In: Jonassen DH (ed) Handbook of research on educational
J, Hornsey M (eds) Rebels in groups: dissent, deviance, difference communications and technology, 2nd edn. Lawrence Erlbaum,
and defiance. Wiley, Oxford, pp 36–53 Mahwah, NJ, pp 785–811
Carter SP, Greenberg K, Walker MS (2017) The impact of computer Johnson DW, Johnson RT (2005) New developments in social
usage on academic performance: evidence from a randomized interdependence theory. Genet Soc Gen Psychol Monogr
trial at the United States Military Academy. Econ Educ Rev 131(4):285–358
56:118–132 Ke F, Grabowski B (2007) Gameplaying for math’s learning: Coopera-
Cheng-Huan C, Chiung-Hui C (2016) Employing intergroup com- tive or not? Br J Educ Technol 38(2):249–259
petition in multitouch desing-based learning to foster student Koschmann T (1996) Paradigm shifts and instructional technology.
engagement, learning achievemet, and creativity. Comput Educ In: Koschmann T (ed) CSCL: theory and practice of an emerging
103:99–113 paradigm. Lawrence Erlbaum, Mahwah, pp 1–23

13
Virtual Reality

Lacasa P (1993) La construcción social del conocimiento: Desarrollo y Scardamalia M (2004) CSILE/knowledge forum. In: Kovalchick A,
conflicto sociocognitivo. Una entrevista a Willem Doise. Infancia Dawson K (eds) Education and technology: an encyclopedia.
y aprendizaje 61:5–28 ABC-CLIO, Santa Barbara, pp 183–192
Mehlenbacher B (2010) Instruction and technology, designs for every- Schrire S (2006) Knowledge building in asynchronous discussion
day learning. The MIT Press, Cambridge, p 495 groups: going beyond quantitative analysis. Comput Educ 46:49–
Mugny G, Doise W (1983) La construcción social de la inteligencia. 70. https​://doi.org/10.1016/j.compe​du.2005.04.006
Trillas, México Stahl G (2006) Group cognition: computer support for building col-
Mugny G, Doise W, Perret-Clermont AN (1975–1976) Conflit de cen- laborative learning. The MIT Press, Cambridge
trations et progrès cognitif[Conflict of centrations and cognitive Stahl G, Koschmann T, Suthers D (2006) Computer-supported col-
progress]. Bulletin de Psychologie 29:199–204 laborative learning: An historical perspective. In: Sawyer RK
Mugny G, Levy M, Doise W (1978) Conflit socio-cognitif et dével- (ed) Cambridge handbook of the learning sciences. Cambridge
oppement cognitif [Socio-cognitive conflict and cognitive devel- University Press, Cambridge, pp 409–426
opment]. Revue Suisse de Psychologie 37:22–43 Tauer JM, Harackiewicz JM (2004) The effects of cooperation and com-
Mugny G, Giroud JC, Doise W (1978–1979) Conflit de centrations et petition on intrinsic motivation and performance. J Pers Soc Psy-
progrès cognitif II: nouvelles illustrations expérimentales [Con- chol 86(6):849–861. https​://doi.org/10.1037/0022-3514.86.6.849
flict of centrations and cognitive progress II: new experimental Wang X, Wallace PW, Wang Q (2017) Rewarded and unrewarded com-
illustrations]. Bulletin de Psychologie 32:979–985 petition in a CSCL environment: a coopetition designado with a
Mugny G, De Paolis P, Carugati F (1984) Social regulations in cogni- social cognitive perspectiva using PLS-SEM analyses. Comput
tive development. In: Doise W, Palmonari A (eds) Social inter- Hum Behav 72:140–151
action in individual development. Cambridge University Press, Westbrook V (2006) The virtual learning future. Teach High Educ
Cambridge, pp 127–146 11(49):471–482
Mugny G, Tafani E, Butera F, Pigière D (1999) Contrainte et Yang SJH, Chen IYL, Kinshuk NS, Chen NS (2007) Enhancing the
de´pendance informationnelles: influence sociale sur la quality of e-learning in virtual learning communities by finding
repre´sentation du groupe d’amis ide´al [Informational constraint quality learning content and trustworthy collaborators. Educ Tech-
and dependence: soial influence on the representation of the ideal nol Soc 10(2):84–95
group of friends]. Connexions 72:55–72 Yazici HJ (2005) A study of collaborative learning style and team
Neugebauer J, Ray DG, Sassenberg K (2016) When being worse helps: learning performance. Educ Train J 47(3):216–229
the influence of upward social comparisons and knowledge aware- Yu F-Y (2001) Competition within computer-assisted coopera-
ness on learner engagement and learning in peer-to-peer knowl- tive learning environments: cognitive, affective, and social
edge exchange. Learn Instr 44:41–52 outcomes. J Educ Comput Res 24(2):99–117. https ​ : //doi.
Quiamzade A, Mugny G, Butera F (2013) Psychologie sociale de la org/10.2190/3u7r-dcd5-f6t1-qkrj
connaissance. Fondements théoriques. [Social psychology of Yu F-Y, Han C, Chan T-W (2008) Experimental comparisons of face-
knowledge Theoretical foundations]. Presses Universitaires de to-face and anonymous real-time team competition in a networked
Grenoble, Grenoble gaming learning environment. CyberPsychol Behav 11(4):511–
Regueras LM, Verdú E, Muñoz MF, Pérez MA, de Castro JP, Verdú MJ 514. https​://doi.org/10.1089/cpb.2007.0171
(2009) Effects of competitive e-learning tools on higher education
students: a case study. IEEE Trans Educ 52(2):279–285
Scardamalia M (2002) Collective cognitive responsibility for the
advancement of knowledge. In: Smith B (ed) Liberal educación
in a knolwedge society. Open Court, Chicago, pp 67–98

13

You might also like