Exploring Social Network Structure Patterns Suitable to the Community of Inquiry Model Moderated by the Task

Abstract

There are few studies devoted to analyze the relationships between the structure of the social network and performance in the online learning process as group. In this study, the Community of Inquiry (CoI) model is used as an analytical framework, along with quantitative content analysis and social network analysis, in order to identify the effect of density, centralization, and centrality—of the coordinators—in the different categories of CoI model in self-regulated groups, according to the type of task. Over a period of 3 academic years, a total of 96 discussion forums and 7,155 units of analysis were analyzed, focusing on two types of tasks. The results showed that the effects of the density, centralization, and centrality of the coordinators on the categories of social and cognitive presence were moderated by the type of task. The findings contribute to identify the adequate structure of the social network for the CoI model categories and the learning tasks.

Keywords

communities of inquiry online learning quantitative content analysis higher education social network analysis

Introduction

Universities and other institutions of higher education are increasingly incorporating technologies in instructional settings to enhance teaching and learning experiences. The impressively fast growth of online learning is challenging higher education institutions to ensure that their online programs and courses have the same high quality as their traditional classes.

At present, it is recognized that interactions can greatly influence the performance of online learning as group (e.g., Henrie, Halverson, & Graham, 2015). Interaction is a relational, interdependent process that constitutes social structure (Sauer & Kauffeld, 2016). Likewise, researchers have highlighted the group structure as a key factor for achieving active participation, which is crucial for meeting effectiveness, and learning success (e. g., Cohen, Rogelberg, Allen, & Luong, 2011; Sauer & Kauffeld, 2016). It is important that instructors and educational managers have the knowledge that allows them to guide, assist, and encourage the participation of the group and the coordinators to improve performance (e.g., Dado & Bodemer, 2017; Shea et al., 2010).

Social network analysis (SNA) is usually used by researchers to improve the understanding of the interactions in online learning and learning communities (e.g., Tirado-Morueta, Maraver-López, & Hernando-Gómez, 2017; Cho, Gay, Davidson, & Ingraeffea, 2007; Dado & Bodemer, 2017; Dawson, Bakharia, & Heathcote, 2010; Jan & Vlachopoulos, 2018; Macfayden & Dawson, 2010). Using SNA, researchers have tried to identify structure of the social network (SSN) that facilitate optimal performance of the group (e.g., Sauer & Kauffeld, 2016). In this regard, there is evidence that showed that the relationship between the SSN and group performance is complex, depending on learning model (Dado & Bodemer, 2017) and the type of learning task to be performed (Ab Jalil & de Laat, 2014; Donaldson, 2001).

In this scenario, there are few studies dedicated to analyze the relationships between the SSN and the performance considering the online learning process (e.g., Gunawardena, Lowe, & Anderson, 1997; Wang, Han, Fisher, & Pan, 2017) and the type of task as a moderating variable (e.g., Tirado-Morueta, Maraver-López, Hernando-Gómez, & Harris, 2016; Tirado-Morueta, Maraver-López, & Hernando-Gómez, 2017; Ab Jalil & de Laat, 2014; Mohammed & Harrison, 2013).

Therefore, the goal of this study is to understand the SSN that facilitate the performance during the online learning process as a small group and self-regulated, considering the type of learning task.

To do this, first, the authors conducted a descriptive and exploratory case study of the SSN (cohesion, centralization, and centrality of the coordinator) in the online learning process of a self-regulated group. The authors chose self-regulated groups because of their effectiveness for collaborative discourse (Oh, Huang, Mehdiabadi & Ju, 2018). Likewise, the Community of Inquiry (CoI) model (Garrison, Anderson, & Archer, 2000, 2001, 2010) was used as a reference for the analysis of the online group learning process.

Second, an online course, based on the Salmon (2004) model, was designed. The online course was implemented within the MOODLE Learning Management System (LMS) during 3 academic years. The discussion forum was the interaction tool between student groups, which can promote high levels of interaction among students and teachers and is considered an effective means for engaging learners in online group learning and critical thinking (de Leng, Dolmans, Jöbsis, Muijtjens, & van der Vleuten, 2009; Richardson & Ice, 2010), among other benefits.

Theoretical Background

The theoretical background and bases of the study are presented in this section. With the purpose of identifying patterns of SSN facilitators of the group's online learning process, the structure and functioning of the CoI model are presented. Taking the CoI model as a reference for the learning process, findings are reviewed that associate certain structural characteristics (cohesion, centralization, and centrality of the coordinator) with social presence (SP) and cognitive presence (CP).

CoI as Framework for Online Learning

Although there are various models available for the ( analysis of online group learning see de Wever, Schellens, Valcke, & Van Keer, 2006), this study was based on the model proposed by Garrison et al. (2000, 2010), a framework that has provided significant insights and methodological solutions for studying online learning as group (Garrison & Archer, 2003; Garrison, Cleveland-Innes, Koole, & Kappelman, 2006). In higher education, the CoI framework is generally viewed as a dynamic model of the key core elements for both the development of community and the pursuit of inquiry, which are required for higher education (Garrison, 2011; Swan, Garrison, & Richardson, 2009). The structure of the CoI framework was confirmed through factor analysis by Garrison, Cleveland-Innes, and Fung (2004) and Arbaugh and Hwang (2006). The model assumes that learning in online environments occurs through the interaction of three core elements: SP, CP, and teaching presence (TP). These elements work together to support deep and meaningful online inquiry and learning.

The first element, SP, is defined as the ability of learners to project themselves socially and affectively into a CoI (Rourke, Anderson, Garrison, & Archer, 2007). SP is divided into the three categories of affective, interactive, and cohesive presence, which mirror a supportive context for emotional expression, open communication, and group cohesion for task resolution, respectively.

However, SP is essential for collaborative learning experiences and it is also an essential element in establishing CP (Garrison, 2011), the second core element of learning in online environments. CP refers to the extent to which online learners can construct and validate meanings based on communication and thinking processes (Garrison et al., 2000, 2001).The CoI model categorizes CP into four phases: a triggering event (an issue is identified for inquiry), exploration (exploring the issue through discussion and critical reflection), integration (constructing meaning from the ideas developed through exploration), and resolution (applying new knowledge into a real world context), with specific descriptors for each phase.

The third element of the CoI model is TP. It consists of two general functions: (a) the design of the educational experience and (b) facilitation between the teacher and the students. It is the responsibility of the teacher to design and integrate both CP and SP for educational purposes. This category is not considered in this study because it analyzes the process of learning self-regulated groups.

Recent studies have attempted to enhance the understanding of the relationships between presences, confirming (Garrison, 2017) that SP (which implies affective, complete and open communication) promotes CP (which includes collaboration and critical discourse), and CP is improved and maintained when SP is established. Likewise, TP is necessary to maintain participation. Over time, high levels of SP are replaced by CP, as participants assume different roles and responsibilities within a CoI.

SSN and CoI (Moderated by the Task)

SNA provides useful information and quantitative indicators of the quality of the learning as a way to effectively analyze the online learning process in a CoI (e.g., Tirado-Morueta, Hernando-Gómez, & Aguaded-Gómez, 2015; Shea et al., 2010; Shea et al., 2013; Wang & Li, 2007). It also allows for description, analysis, and visualization of the patterns of SSN of the online groups formed by students, offering a multitude of network indicators to evaluate the process and quality of the online learning of students from a social perspective. This study focuses on some of the variables mostly used in SNA, for both groups and individuals (see review, Dado & Bodemer, 2017): cohesion, centralization, and degree centrality.

Cohesion connections between participants are indicators of SP that are always present in a CoI (Garrison, 2017); but if more advanced stages display high levels of CP, it is reasonable to think that other structural characteristics—such as the centralization of relationships and centrality of the teacher (or coordinators)—could also be indicators of CP (Tirado-Morueta, Hernando-Gómez, & Aguaded-Gómez, 2015; Shea & Bidjerano, 2010). However, there is a lack of empirical evidence associating the structural characteristics of the connections between CoI participants with the different presences and their categories (e.g., Jan & Vlachopoulos, 2018).

Network cohesion and CP

Network cohesion describes the general degree of linkage among the participants in the network (Scott, 2000). Network cohesion refers to the forces holding participants within their groupings (Yang & Tang, 2004).

The relationship between group cohesion and group success has been widely explored. Social exchange theorists have stated that high cohesion has a positive impact on behavioral commitment, which is reflected in greater investment, more contributions to joint tasks, and better performance (e.g., Thye, Yoon, & Lawler, 2002; Yoon & Lawler, 2005). In recent studies (e.g., Evans & Dion, 2012; Gaggioli, Mazzoni, Milani, & Riva, 2015), a positive relationship between group cohesion and group performance was found. It is reasonable to think that in a very cohesive group, there is a greater probability that more resources and knowledge flow, and as a result, the group has a high performance (Wise, 2014).

However, there is evidence that showed that the relationship between the cohesion and performance depends on the type of task to be performed (Ab Jalil & de Laat, 2014), although the results are not conclusive. For example, the experimental research has demonstrated that complex tasks (less specific) are carried out more efficiently when the communication structure is decentralized (e.g., Huang & Cummings, 2011; Sparrowe, Liden, Wayne, & Kraimer, 2001).

Network centralization and CP

Network centralization refers to the extent to which a structure of relationships is centered on a participant. Centralization of the network can be of two types: centralization based on in-degree (centralization [in]) and centralization based on out-degree (centralization [out]) (Borgatti & Everett, 2006). Centralization (in) refers to the extent to which a group is focused on a participant, that is, who most group members directed their interventions toward. Centralization (out) refers to the extent to which a group is focused on a participant, that is, who among the group engages in most of the interventions.

There are studies that showed negative effects of centralization on group performance in online learning (e.g., Sauer & Kauffeld, 2013). However, there is evidence that showed that the relationship between the centralization and performance is complex, depending on the type of task.

For example, there are studies which has shown to be inversely related to categories of critical thinking (e.g., Thormann, Gable, Fidalgo, & Blakeslee, 2013) and group creative (exploitation tasks) performance (e.g., Gaggioli et al., 2015). However, in one study of 45 student project teams, it was found that network centralization was positively related to team performance (Lin, Yang, Arya, Huang, & Li, 2005). Likewise, Troster, Mehra, and van Knippenberg (2014) developed longitudinal study used data from 91 self-managed teams and concluded that complex tasks (less specific) require higher levels of centralization.

Centrality of the coordinator and SP/CP

Degree centrality is the measurement of the interaction considering the directionality of the sending or receiving of messages. There are two types of centrality: centrality-based on in-degree (centrality [in]) and centrality based on out-degree (centrality [out]). Centrality (out) is the number of messages sent by the student and Centrality (in) is the number of messages received by the student. The idea is that if a participant is central in their group, then they will be the most popular and will thus receive the most attention. Some authors have found that students who occupy central positions in the formal network were more likely to put forward ideas and suggestions on group work (e.g., Venkataramani, Zhou, Wang, Liao, & Shi, 2016) and could trigger peer interaction, which could further advance group knowledge construction (Ouyang & Chang, 2018).

Coordination in online group learning is a complex process involving distinct control modes of interaction (Mayordomo & Onrubia, 2015). Likewise, some studies have suggested that a relationship exists between the quality of work coordination and the final performance of the students engaged in the online learning as group (e.g., Engel & Onrubia, 2010). Also, there are also studies that show that the quality of work coordination is moderated by the stage and the task when analyzing the relationship between the centrality of the coordinators and learning behaviors (Nahrgang et al., 2013; Van der Haar, Koeslag-Kreunen, Euwe, & Segers, 2017; Wang et al., 2017).

Some authors have found that in the early stages of team collaboration, the centrality of the coordinators stimulates interactions between the work team—SP—although an inverse effect has been found on learning behaviors—CP—in later phases (Wang et al., 2017). Likewise, Van der Haar et al. (2017) showed that leaders of effective teams use more structuring behaviors in earlier phases but fewer of these behaviors in later phases, compared with leaders of less effective teams.

Regarding the influence of the task, other studies, when considering the degree of specification of the task, showed a high level of centrality of the coordinators in the more open tasks (less specific) and creative (exploitation tasks) (Tirado-Morueta, Maraver-López, & Hernando-Gómez, 2017), promoting communication among the members of the team (e.g., Kauffeld & Meyers, 2009), in addition to responding to the interventions of their colleagues (Ab Jalil & de Laat, 2014).

Research Questions

In short, so far the greatest certainty is that cohesion and density of connections between group members are indicators of SP. However, the extent to which it can be associated with the affective, interaction, and cohesion categories of the CoI model is unknown. Also, although there are numerous studies that indicated that cohesion facilitated CP, others have shown that centralized structures were more effective. Regarding these apparently contradictory results, there is sufficient evidence to assert that the nature of the task conditions the structure of the most efficient social network. For this reason, this study considered two types of task according to the following criteria:

(a) According to the objective of the task, exploration or exploitation tasks are used (Donaldson, 2001). The exploration tasks usually consist of data or information analysis to resolve a problem or make a decision. However, the exploitation of the tasks are usually dedicated to carrying out a project or creative collaborations.

(b) According to the degree of specification of the task (see Nahrgang et al., 2013). In this sense, the more specific tasks are those that are described step by step, while the less specific are those in which the students have to make their own decisions to solve them.

With this background in mind, the first research question is as follows:

RQ1: Taking the type of learning task as a moderating variable, what is the effect of the structure of group—cohesion and centralization—on the categories of social and cognitive presence in CoI model?

In this study, self-managed groups were used where the figure and roles of the coordinator were established to stimulate the completion of the task. Regarding the centrality of the coordinators, few studies have analyzed their effect on the performance of the group in online learning. Recent findings showed its importance in the early stages of socialization and have also emphasized its usefulness in less structured tasks. However, its effects on more specific aspects of group socialization and cognitive activity are unknown.

With this background in mind, the second research question is as follows:

RQ2: Taking the type of learning task as a moderating variable, what is the effect of the centrality (out/in) of the coordinators on the categories of social and cognitive presence in CoI model?

Materials and Methods

Participants

A total of 206 university students from different degree areas and from eight universities in Andalusia (Spain) participated in this study over a period of 3 academic years. During the first year, 71 students participated in the course. In the second year, 77 participated, and in the last year, 64 participated. Participants were randomly assigned to small groups of eight. Students ranged in age from 19 to over 51 years. Approximately 80% of the students in the sample were females. Analysis and results from the current study were based on the online group discussions of the students in the sample. A total of 96 discussion forums were conducted: eight groups in each academic year for each of the four activities (including an initial socialization activity).

Experimental Setting

This study took place at the Virtual Campus of Andalusia, Spain, where students from nine public universities from the Andalusian region of Spain participated in a common online course through the MOODLE LMS. The online course was named: “Intervention on risky behaviors.” Up to 10 students from each of the public universities of Andalusia and from different academic areas (social sciences, engineering, experimental, and health) could enroll. This course was chosen for the following because of the high-quality academic outcomes, and the low dropout rate (10.6% [Year 1], 12.2% [Year 2], and 14% [Year 3]). Likewise, the experience was evaluated as excellent and recommended by all the students and groups.

MOODLE LMS was employed to facilitate content delivery and interactions among the students. The moderating tasks had different levels of specification (see Nahrgang et al., 2013) and served different educational purposes, that is, the task had different loads of exploration and exploitation (Donaldson, 2001) (Table 1).

Table 1.

Definition of the Tasks Analyzed in the Study.

Task	Task specification (Nahrgang et al., 2013)	Task purpose (Donaldson, 2001)	Description	Materials	Role of the coordinators (Mayordomo & Onrubia, 2015)
ACS	High	Exploration	Analyzing, in a collaborative way, the real world case of a student from a secondary school.	Case description, scheme and help for the case analysis.	Jigsaw model, in which the coordinator integrated the contributions of the members of his group
CWQ	Low	Exploitation	CWQ that can serve as a tool for information-training for drug prevention.	Web sources and examples.	Chain model, in which the coordinator prepared the material and the group reviewed it

Note. ACS = analyzing case study; CWQ = creating a WebQuest.

The course design was inspired by the Salmon model (2004, 2013), because it provided a staged, practical approach to teaching and learning online. In accordance with this model, these were the following steps: In Step 1, eight working groups were established, each one placed into a discussion forum where students were given directions for how to proceed by the tutor. In Step 2, the first few actions were performed, in order for the students to socialize, exchange messages, interact, and learn. In Step 3, three types of tasks were presented, with different levels of cognitive demand that required participants in each group to share information, begin to engage in group work, and to share the same learning goal. In Step 4, the learners focused on knowledge development and discussion activities.

To conduct the assigned tasks, each of the eight working groups were provided a private forum to work together, an open forum for intermeeting (as a space for discussion between the groups), and a file space for uploading or downloading documents. They were offered theoretical and practical resources that also served as reference material and support for the implementation of the tasks. Due to the effectiveness of peers to strengthen critical thinking and collaborative discourse (Oh et al., 2018), two members were selected as coordinators for each task, two members were randomly selected in each group to coordinate the task's resolution.

Measures and Procedures

The study was based on group discussions among the participating students. A total of 96 discussion forums were conducted: eight groups per 3 academic years for each of four group activities (including the first socialization activity which was not included in the analysis). For the analysis, three analysis techniques were combined (see Jo, Park, & Lee, 2017, synergetic effects of integrating methodologies): quantitative content analysis (QCA), SNA, and statistical analysis.

Quantitative content analysis

First, data were collected using the messages in the forums created by each group, using the instruments developed by Garrison et al. (2000, 2001) and Rourke et al. (2007) as references for measuring SP and CP. For the content analysis, units of meaning were identified as units of analysis. A unit of meaning can be defined simply as a thought or idea (Rourke et al., 2007) and can include expressions, sentences, or paragraphs in which important thoughts and ideas (meanings) are conveyed. Depending on the semantic sense used, several units of meaning could be conveyed within each message. A total of 7,155 units of meaning were coded into the qualitative analysis program AtlasTi v.6. Finally, data were imported to SPSS v.21 in order to perform quantitative analysis (Table 2).

Table 2.

Units of Analysis.

		Task
Presence	Category	ACS	CWQ	Total
Social	Affective	327	512	839
	Interactive	462	1,081	1,543
	Cohesive	626	1,122	1,748
	Subtotal	1,415	2,715	4,130
Cognitive	Triggering	205	417	622
	Exploration	698	976	1,674
	Integration	241	455	696
	Resolution	15	18	33
	Subtotal	1,159	1,866	3,025
	Total	2,574	4.581	7.155

Note. Data were calculated based on the units of meaning, not on the number of messages. ACS = analyzing case study; CWQ = creating a WebQuest.

In accordance with the principles of systematicness, objectiveness, and reliability of content analysis indicated by Rourke and Anderson (2002), three coders classified the messages using the CoI model (Online Appendices A and B). Messages were coded in the chronological order in which they were posted in the discussion forums. The researchers coded the messages, engaged in continuous dialogue, and set and checked rules and procedures throughout the coding process. The statistical package macro (KALPHA) by Hayes and Krippendorff (2007) was used for the calculation of the interrater and intrarater reliability coefficient, Krippendorff's alpha (α). The global output from the macro Krippendorff's ordinal α was 0.76, a modest degree of interrater reliability. The first year Krippendorff's α was.71. The second year Krippendorff's α was .69, and the third year, it was .86. In addition, the intrarater reliability (the degree of agreement among repeated administrations of a diagnostic test performed by a single rater) was .92. The values for Krippendorff's alpha were situated within the classification of “fair to good agreement beyond chance.”

Social network analysis

To perform the SNA, the authors considered variables indicators of the structure of the groups. In Online Appendix C, a description of the metrics of degree centrality (out/in), network cohesion, and network centralitation (out/in) is shown. The data were entered into UCINET 6 for Windows (Borgatti, Everett, & Freeman, 2018), where rates of the different variables were obtained for SNA. Finally, data were imported to SPSS v.21 in order to conduct quantitative analysis.

Statistical analysis

As already indicated, statistical analysis of data was performed using measurements of QCA and SNA. For answering the first and second research questions, a linear regression analysis was conducted. With this analysis, the influence of the social connections of the group (density), centralization (out/in), and the level of centrality (out/in) of the group's coordinator on the different categories of SP and CP (at the group level) was identified. Also, with the R² analysis, the predictability of the density, centralization (out/in), and centrality (out/in) of the group's coordinator was tested.

To observe the moderating effect of the task, the differences between covariance parameters were conducted. The recommendations of Byrne and Van de Vijver (2010) were followed, and the critical ratio difference method offered by AMOS was used. If the critical ratios exceed 1.96, the parameter is significantly different between the two groups at a level of p < .05.

Results

The descriptive statistics results—mean and standard deviation—of the activity of the groups (see Table 3), on the one hand, showed that the less specific the task, the greater the cognitive activity of the group. On the other hand, the cognitive activity of the coordinators increased, especially in the less specific task (creating a WebQuest [CWQ]).

Table 3.

The Descriptive Results of the Activity of the Groups.

			Social presence			Cognitive presence
Task			Affective	Interaction	Cohesion	Triggering	Exploration	Integration	Resolution
ACS	Member (N = 173)	Mean	1.35	1.64	2.16	.55	2.76	.87	.02
	Member (N = 173)	SD	2.24	2.54	2.90	1.10	4.76	1.52	.15
	Coordinator (N = 32)	Mean	2.81	5.53	7.66	3.34	6.84	2.81	.34
	Coordinator (N = 32)	SD	3.27	5.80	6.39	4.66	9.46	5.61	.70
	Total (N = 205)	Mean	1.58	2.24	3.02	.99	3.40	1.17	.07
	Total (N = 205)	SD	2.48	3.54	4.16	2.31	5.91	2.69	.32
		F	9.83**	35.51***	60.82***	48.23***	13.65***	15.09***	29.46***
CWQ	Member (N = 167)	Mean	1.93	3.93	3.93	1.35	3.54	1.63	.05
	Member (N = 167)	SD	3.15	7.52	5.48	2.24	6.53	2.71	.24
	Coordinator (N = 39)	Mean	4.85	10.87	11.95	4.90	9.85	4.69	.26
	Coordinator (N = 39)	SD	4.19	8.24	8.13	4.22	9.48	4.53	.49
	Total (N = 206)	Mean	2.49	5.25	5.45	2.02	4.74	2.21	.09
	Total (N = 206)	SD	3.55	8.11	6.82	3.05	7.57	3.35	.31
		F	23.52***	25.93***	55.21***	53.52***	24.36***	30.12***	14.71***

Note. ACS = analyzing case study, CWQ = creating a WebQuest.

**p≤.01. ***p≤.001

For RQ1, based on the analysis of linear regressions, the following results were obtained (see Table 4 and Figures 1 and 2).

Figure 1.

Effect (Beta) of the density on the categories/phases of social and cognitive presence in CoI model. ACS = analyzing case study; CWQ = creating a WebQuest.

Figure 2.

Effect (Beta) of the centralization (out/in) on the categories/phases of social and cognitive presence in CoI model. ACS = analyzing case study; CWQ = creating a WebQuest.

Table 4.

Linear Regression; Dependent Variables: Group Performance in the Categories/Phases of the CoI Model.

	Beta
	ACS	CWQ	CR
Centralization (out) → Affective	−.34***	.00	3.866***
Centralization (out) → Interaction	−.41***	−.00	2.714***
Centralization (out) → Cohesion	−.34***	.09	4.230***
Centralization (out) → Triggering	−.05	.19**	3.301***
Centralization (out) → Exploration	−.29***	−.21***	.408
Centralization (out) → Integration	−.12	−.01	.958
Centralization (out) → Resolution	−.26***	−.32***	−1.847
Centralization (in) → Affective	.14**	.07	−.299
Centralization (in) → Interaction	.27***	.00	−1.607
Centralization (in) → Cohesion	.10*	.06	.093
Centralization (in) → Triggering	−42***	.41***	9.086***
Centralization (in) → Exploration	−.03	−.08	−.712
Centralization (in) → Integration	−.15*	−.29***	−2.343**
Centralization (in) → Resolution	.18**	−.17**	−3.825***
Density → Affective	.52***	.61***	.112
Density → Interaction	.68***	.63***	3.802***
Density → Cohesion	.55***	.39***	−1.873
Density → Triggering	.46***	.16**	−2.571**
Density → Exploration	.23***	.63***	3.021***
Density → Integration	.20**	.06	−1.839
Density → Resolution	.29***	.03	−2.909***
R² Affective	.38	.39	−2.683**
R² Interaction	.67	.40	−1.488
R² Cohesion	.40	.18	1.324
R² Triggering	.38	.25	−6.299***
R² Exploration	.13	.42	−1.478
R² Integration	.06	.07	2.469**
R² Resolution	.16	.12	4.636***

Note. ACS = analyzing case study; CWQ = creating a WebQuest ; CR = critical ratio.

*p≤.05. **p≤.01. ***p≤.001.

First, regarding the predictive ability of the social structure of groups (density and centralization) on SP and CP, the results of the linear regressions are as follows:

(a) The predictive ability of the social structure of groups on SP and CP depended (partially) on the type of learning task.

(b) In this regard, based on the values of R², it was observed that group structure most predicted the variance of triggering in the analyzing case study (ACS) task (R²[Tri] = .38).

(c) In contrast, in the CWQ task, the group structure had greater predictive value on the affective category (R²[Aff] = .39) and on cognitive categories of integration (R² = .07) and resolution (R²[Res] = .16).

Second, the results regarding the influence of the group structure on the SP and CP of the group were as follows:

(a) The density of the group directly influenced all categories of SP and CP in the two learning tasks. Likewise, generally density had more influence in the categories of SP than in the categories of CP.

(b) However, the strength of this influence depended on the type of learning task. First, density more strongly influenced interaction (β[Int] = .68, p<.01), triggering (β[Tri] = .46, p < .001), and resolution (β[Res] = .29, p < .001) in ACS task than in CWQ task. Second, in CWQ, the influence of exploitation was stronger (β[Exp] = .63, p<.001).

(c) Moreover, centralization (out) usually inversely influenced all categories of SP and CP, with a more negative influence on the ACS task than on CWQ task.

(d) The influence of centralization (in) on the categories of SP and CP was usually variable, depending on the type of learning task. Its influence on the categories of SP was direct and significant in the ACS task, but it was not significant in the CWQ task. Also, while there was a direct effect of centralization (in) on the triggering category of the CWQ task (β[Tri] = .41, p<.001), its effect on the same category was reversed in the ACS task (β[Tri] = −.42, p<.001).

For RQ2, the following results were obtained (see Table 5 and Figure 3).

Figure 3.

Effect (Beta) of the centrality (out/in) on the categories/phases of social and cognitive presence in CoI model. ACS = analyzing case study; CWQ = creating a WebQuest.

Table 5.

Linear Regression; Dependent Variables: Group Performance in the Categories/Phases of the CoI Model.

	Beta
	ACS	CWQ	CR
Centrality (out) → Affective	.62***	.62***	0.652
Centrality (out) → Interaction	.47**	.68***	3.454***
Centrality (out) → Cohesion	.47**	.36*	−0.081
Centrality (out) → Triggering	.70***	.35*	−0.428
Centrality (out) → Exploration	.58***	.71***	0.605
Centrality (out) → Integration	.49**	−.24	−3.334***
Centrality (out) → Resolution	.31	−.33	−2.845***
Centrality (in) → Affective	−.08	.12	1.038
Centrality (in) → Interaction	.15	.05	−0.378
Centrality (in) → Cohesion	.02	.26	0.910
Centrality (in) → Triggering	−.19	.18	1.979*
Centrality (in) → Exploration	−.10	−.08	0.351
Centrality (in) → Integration	.23	.16	−0.658
Centrality (in) → Resolution	−.08	.08	0.695
R² Affective	.39	.40	−0.405
R² Interaction	.25	.47	−0.479
R² Cohesion	.22	.19	0.759
R² Triggering	.53	.15	1.026
R² Exploration	.35	.51	−0.580
R² Integration	.29	.08	3.228***
R² Resolution	.10	.11	1.573

Note. ACS = analyzing case study; CWQ = creating a WebQuest; CR = critical ratio.

*p≤.05. **p≤.01. ***p≤.001.

First, regarding the predictive ability of the centrality of coordinators on the SP and CP of the groups, the results are as follows:

(a) The predictive ability of the activity of the coordinators on the SP and CP of the group not depended (generally) on the type of learning task.

(b) In this regard, based on the values of R², it was observed that in the ACS task, activity coordinators predicted the stronger variance in the integration category (R²[Inte] = .29).

(c) Moreover, in the CWQ task, the activity of coordinators was particularly relevant, both with regard to SP (R²[Aff] = .40; R²[Int] = .47) and CP (R²[Exp] = .51).

Second, the results regarding the influence of the coordinators on the SP and CP of the group are as follows:

(a) The centrality (out) of coordinators often had a direct influence on the two types of task, although the nature of the learning task modulated its effect on the SP and CP categories. First, in ACS task, the influence on integration and resolution was positive, while in CWQ the influence was reversed.

(b) In ACS, the influence of the centrality (out) of coordinators was positive and significant in almost all categories of SP (β[Aff] = .62, p<.001; β[Int] = .47, p<.01; β[Coh] = .47, p<.01) and CP (β[Tri] = .70, p<.001; β[Exp] = .58, p<.001; β[Inte] = .49, p<.01).

(c) Moreover, in the CWQ task, the influence was positive and significant on the interaction and cohesion categories of SP (β[Aff] = .62, p<.001; β[Int] = .68, p<.001; β[Coh] = .36, p<.05) and on the cognitive category of exploration (β[Exp] = .71, p<.001).

(d) A significant influence regarding the centrality (in) of coordinators was not seen in any of the two learning tasks.

Discussion

SNA is a sufficiently valid technique to identify the interaction structures of groups that facilitate their performance and provide feedback to teachers and educational managers to optimize online group learning processes. Seemingly contradictory findings have led to the conclusion that task characteristics determine the most optimal structures for online group performance. Likewise, after years of validation, the CoI model has been established as an adequate approach for the design and analysis of online learning in higher education. However, there are no findings that help to know the adequate online interaction structures for the CoI model categories taking into account the type of task.

Regarding RQ1, the density of the connections positively influenced all the categories of the CoI model, although the strength of its influence was moderated by the type of task. On one hand, density had a greater influence on group social activity than on cognitive activity in all tasks. In contrast, density had a strong influence on the exploration phase in CWQ tasks (less specified). These data reinforce the results of other studies that suggested that density of group promotes the SP (Garrison, 2017). Also, the density favored learning behaviors in the early phases of group work, both the triggering in more specified tasks and the exploration in less specified tasks. These results reinforce other studies that have correlated high-density levels of interactions with group creativity (e.g., Gaggioli et al., 2015).

Regarding the centralization (out) of the group, data showed an inverse effect of centralization (out) on all categories of SP, which was more pronounced in the most specific task (ACS task). On the other hand, in the less specified task (CWQ), centralization had a positive effect on the triggering category—identification of the problem and its approach. On the other hand, centralization (in), defined by the messages sent to the forum, had a positive effect on the triggering phase in the least specified task (CWQ).

Regarding RQ2, the data suggested that the interventions of the coordinators (centrality [out]) often had a direct influence on the two types of tasks, although the type of the learning task moderated its effect on the SP and CP categories (see Ceri-Booms, Curşeu, & Oerlemans, 2017).

Regarding the effects of centrality (out) on SP, data showed positive effects on all categories and on two tasks, but especially on interaction—open communication—in the least specified and creative task (CWQ task).

With respect to the effects on CP, the data showed a clear moderating effect of the task. In the more specified ACS task, in which the coordinators were dedicated to distributing the tasks and integrating the final result, the interventions (centrality [out]) were especially relevant in almost all phases of the cognitive activity of the group, that is, in the definition of the case problem, in the exploration of suggestions, and in the integration of the contributions of group members. On the other hand, in the less specified CWQ task, open and creative (exploitation), in which the coordinators dedicated themselves to the construction of the prototype and the group validated it, the interventions of the coordinators were relevant in the exploration phase, that is, for the exchange of ideas and suggestions for the construction of the final result.

These results suggest that in the more open and creative tasks, the coordinators promoted communication between team members toward the task objective (e.g., Kauffeld & Meyers, 2009), as well as responding to suggestions from their peers (Ab Jalil & de Laat, 2014). On the other hand, in the more specified and analytical (exploratory) tasks, the coordinators centralized the process of solving the task in which group members worked by themselves for a longer period of time and the coordinators dedicated themselves to receiving and integrating the contributions of group members.

Online Appendix D shows a summary of the moderating influence of the type of task on the SP and CP, as well as on the SSN of the groups.

Limitations and Future Studies

Several limitations are noted in this study. First, this study has shown that in an online context, there are certain group behaviors that are not recorded through discussion forums. This often occurs during the resolution phase, so it will be appropriate in future studies to include other technological tools to facilitate collaborative resolution of the task and to explore relevant learning outcomes of the resolution phase (e.g., Gutiérrez-Santiuste & Gallego-Arrufat, 2014). To do this, it seems appropriate to use analytics learning (e.g., Agudo-Peregrina, Iglesias-Pradas, Conde-González, & Hernández-García, 2014; Siemens et al., 2011).

Second, we also consider it appropriate to consider other variables to contrast the observed activity, through content analysis, with learning outcomes (e.g., Akyol & Garrison, 2011), satisfaction, and self-reported values of social, cognitive, and teaching activity (e.g., Garrison, Cleveland-Innes, & Fung, 2010; Garrison & Vaughan, 2011; Wicks, Craft, Mason, Gritter, & Bolding, 2015). Note that the citation “Garrison et al., 2010” has been changed to “Garrison, Cleveland-Innes, & Fung, 2010” as per the reference list, as the latter is otherwise uncited. If this is inaccurate, please update the citation and the reference.]

Finally, we believe that the timing factor during the process of social construction of knowledge and the social structure of the groups (through SNA; Shea et al., 2010; Wang & Li., 2007) will provide important inputs for regulation of the processes.

Conclusions

Understanding the mechanisms that favor successful online learning experiences in higher education remains a prime target for current research. Although many pedagogical models designed to exploit the virtues of technology in higher education have been developed, they have provided insufficient empirical evidence to guide educators and students in developing online learning processes. The results demonstrate the relevance of the CoI model and SNA as conceptual and analytical resources, in order to improve the understanding of online learning processes.

One of the main strengths of the study is the combination of direct interaction observations using methods such as QCA and SNA. This combination of analytical techniques has allowed us to answer the research questions posed. Another advantage is the strength of the sample. To find stable patterns of student behaviors during the online group learning process, the interactions carried out in a graduate course across 3 academic years were observed and analyzed.

The data from this study showed that the influence of the network density, centralization (out/in) and the interventions/reply of/on the coordinators (centrality [out/in]) on the performance in the categories or phases of the CoI model are moderated by the type of task, in this case the degree of specification and objective of the task.

In addition, this study identified SSN patterns appropriate to the categories or phases of the CoI model (in self-managed groups) and to two types of tasks frequently encountered in pedagogical studies, such as case studies and production of educational materials.

This study provides inputs that will allow instructors and educational managers to guide the social participation and connections of members and coordinators in self-managed learning groups in a university context. The results have shown social network structures that have a good fit with learning based on communities of inquiry. Likewise, this type of study can help in the development and improvement of learning analytics that use the structure of group relationships as inputs.

Supplemental Material

Supplemental material for Exploring Social Network Structure Patterns Suitable to the Community of Inquiry Model Moderated by the Task

Supplemental Material for Exploring Social Network Structure Patterns Suitable to the Community of Inquiry Model Moderated by the Task by Ramón Tirado-Morueta, Pablo Maraver-López, Amor Pérez-Rodríguez and Ángel Hernando-Gómez in Journal of Educational Computing Research

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Ramón Tirado-Morueta, PhD, is a professor of Educational Technology at the University of Huelva. He is a member of the Group of Educational Studies and Research in Communication Technologies, Orientation and Sociocultural Intervention Ágora (HU-M648), as well as international advisor to different universities and journals.

Pablo Maraver-López, PhD, is a research member of the Agora Research Group HUM 648 of the University of Huelva. Collaborative researcher in projects of the National R&D Plan of the Ministry of Science and Technology. Teaching experience at the Andalusian Virtual Campus.

Amor Pérez-Rodríguez, PhD, is an associate professor in the Department of Philology at the University of Huelva. She is member of the Agora Research Group (HUM 648). Her research interests are media literacy, media competence, the new communication languages and narratives, and their application in educational contexts.

Ángel Hernando-Gómez, PhD, is an associate professor in the Department of Social, Evolutionary and Educational Psychology at the University of Huelva and an associate editor of the journal Comunicar (Psychology Section). His interests as a researcher are focused on the promotion of positive adolescent development and education.

References

Agudo-Peregrina

Á. F.

Iglesias-Pradas

Conde-González

M. Á.

Hernández-García

Á.

(2014) Can we predict success from log data in VLEs? Classification of interactions for learning analytics and their relation with performance in VLE-supported F2F and online learning. Computers in Human Behavior 31: 542–550. doi:10.1016/j.chb.2013.05.031.

Ab Jalil

de Laat

(2014) Discovering the pattern of interactions in online discussions using the social network analysis principle. Malaysian Journal of Distance Education 16(2): 83–108.

Arbaugh

J. B.

Hwang

(2006) Does ‘teaching presence’ exist in online MBA courses? The Internet and Higher Education 9(1): 9–21. doi:10.1016/j.iheduc.2005.12.001.

Akyol

Garrison

D. R.

(2011) Assessing metacognition in an online community of inquiry. The Internet and Higher Education 14(3): 183–190. doi:10.1016/j.iheduc. 2011.01.005.

Berkowitz, S. D. (2013). An introduction to structural analysis: The network approach to social research. Toronto, Canada: Butterworths.

Byrne

B. M.

van de Vijver

F. J. R.

(2010) Testing for measurement and structural equivalence in large-scale cross-cultural studies: Addressing the issue of nonequivalence. International Journal of Testing 10(2): 107–132. doi:10.1080/15305051003637306.

Borgatti, S. P., Everett, M. G., & Freeman, L. C. (2018). UCINET. Encyclopedia of Social Network Analysis and Mining, 3243–3249. https://doi.org/10.1007/978-1-4614-6170-8_316.

Ceri-Booms

Curşeu

P. L.

Oerlemans

L. A.

(2017) Task and person-focused leadership behaviors and group performance: A meta-analysis. Human Resource Management Review 27(1): 178–192. doi:10.1016/j.hrmr.2016.09.010.

Cho

Geri

Davidson

Ingraffea

(2007) Social networks, communication styles, and learning performance in a CSCL community. Computers & Education 49(2): 309–329. doi:10.1016/j.compedu.2005.07.003.

10.

Cohen

M. A.

Rogelberg

S. G.

Allen

J. A.

Luong

(2011) Meeting design characteristics and attendee perceptions of staff/team meeting quality. Group Dynamics: Theory, Research, and Practice 15(1): 90–104. doi:10.1037/a0021549.

11.

Dado

Bodemer

(2017) A review of methodological applications of social network analysis in computer-supported collaborative learning. Educational Research Review 22: 159–180. doi:10.1016/j.edurev.2017.08.005.

12.

Dawson

Bakharia

Heathcote

(2010) SNAPP: Realizing the affordances of real-time SNA within networked learning environments. In: Dirckinck-Holmfeld

Hodgson

Jones

de Laat

McConnell

Ryberg

(eds) Proceedings of the 7th international conference on networked learning, Lancaster, England: Lancaster University, pp. 125–133.

13.

de Leng

B. A.

Dolmans

D. H.

Jöbsis

Muijtjens

A. M.

van der Vleuten

C. P.

(2009) Exploration of an e-learning model to foster critical thinking on basic science concepts during work placements. Computers & Education 53(1): 1–13. doi:10.1016/j.compedu.2008.12.012.

14.

de Wever

Schellens

Valcke

Van Keer

(2006) Content analysis schemes to analyze transcripts of online asynchronous discussion groups: A review. Computers & Education 46: 6–28. doi:10.1016/j.compedu.2005.04.005.

15.

Donaldson

(2001) The contingency theory of organizations, Thousand Oaks, CA: SAGE.

16.

Evans

C. R.

Dion

K. L.

(2012) Group cohesion and performance a meta-analysis. Small Group Research 43(6): 690–701. doi:10.1177/1046496412468074.

17.

Engel

Onrubia

(2010) Patrones de organización grupal y fases de construcción del conocimiento en entornos virtuales de aprendizaje colaborativo [Patterns of small group organisation and phases of collaborative knowledge construction in CSCL environments]. Infancia y Aprendizaje 33: 515–528. doi:10.1174/021037010793139608.

18.

Gaggioli

Mazzoni

Milani

Riva

(2015) The creative link: Investigating the relationship between social network indices, creative performance and flow in blended groups. Computers in Human Behavior 42: 157–166. doi:10.1016/j.chb.2013.12.003.

19.

Garrison, D. R. (2011). E–Learning in the 21st century: A framework for research and practice (2nd ed.). London, England: Routledge.

20.

Garrison, D. R. (2017). E‐learning in the 21st century: A community of inquiry framework for research and practice (3rd ed). London, England: Routledge.

21.

Garrison

D. R.

Anderson

Archer

(2000) Critical inquiry in a text-based environment: Computer conferencing in higher education. The Internet and Higher Education 2: 87–105. doi:10.1016/S1096-7516(00)00016-6.

22.

Garrison

D. R.

Anderson

Archer

(2001) Critical thinking, cognitive presence, and computer conferencing in distance education. American Journal of Distance Education 15(1): 7–23. doi:10.1080/08923640109527071.

23.

Garrison

D. R.

Anderson

Archer

(2010) The first decade of the community of inquiry framework: A retrospective. Internet and Higher education 13: 5–9. doi:10.1016/j.iheduc.2009.10.003.

24.

Garrison

D. R.

Archer

(2003) A community of inquiry framework for online learning. In: Moore

(ed) Handbook of distance education, New York, NY: Erlbaum.

25.

Garrison

D. R.

Cleveland-Innes

Fung

T. S.

(2004) Student role adjustment in online communities of inquiry: Model and instrument validation. Journal of Asynchronous Learning Networks 8(2): 61–74.

26.

Garrison

D. R.

Cleveland-Innes

Koole

Kappelman

(2006) Revisiting methodological issues in the analysis of transcripts: Negotiated coding and reliability. The Internet and Higher Education 9(1): 1–8. doi:10.1016/j.iheduc.2005.11.001.

27.

Garrison

D. R.

Vaughan

N. D.

(2011) Blended learning in higher education: Framework, principles, and guidelines, San Francisco, CA: Jossey-Bass.

28.

Gunawardena

Lowe

Anderson

(1997) Analysis of a global online debate and the development of an interaction analysis model for examining social construction of knowledge in computer conferencing. Journal of Educational Computing Research 17(4): 379–431. doi:10.2190/7MQV-X9UJ-C7Q3-NRAG.

29.

Gutiérrez-Santiuste

Gallego-Arrufat

M. J.

(2014) Internal structure of virtual communications in communities of inquiry in higher education: Phases, evolution and participants‘ satisfaction. British Journal of Educational Technology 46(6): 1295–1311. doi:10.1111/bjet.12218.

30.

Hayes

A. F.

Krippendorff

(2007) Answering the call for a standard reliability measure for coding data. Communication Methods and Measures 1(1): 77–89. doi:10.1080/19312450709336664.

31.

Henrie

C. R.

Halverson

L. R.

Graham

C. R.

(2015) Measuring student engagement in technology-mediated learning: A review. Computers & Education 90: 36–53. doi:10.1016/j.compedu.2015.09.005.

32.

Huang

Cummings

J. N.

(2011) When critical knowledge is most critical centralization in knowledge-intensive teams. Small Groups Research 42(6): 669–699. doi:10.1177/1046496411410073.

33.

Jan, S. K., & Vlachopoulos, P. (2018). Social network analysis: A framework for identifying communities in higher education online learning. Technology, Knowledge and Learning. Advance online publication. doi:10.1007/s10758‐018‐9375‐y.

34.

Park

Lee

(2017) Three interaction patterns on asynchronous online discussion behaviours: A methodological comparison. Journal of Computer Assisted Learning 33(2): 106–122. doi:10.1111/jcal.12168.

35.

Kauffeld

Meyers

R. A.

(2009) Complaint and solution-oriented circles: Interaction patterns in work group discussions. European Journal of Work and Organizational Psychology 18(3): 267–294. doi:10.1080/13594320701693209.

36.

Lin

Yang

Arya

Huang

(2005) Structural versus individual perspectives on the dynamics of group performance: Theoretical exploration and empirical investigation. Journal of Management 31(3): 354–380. doi:10.1177/0149206304272150.

37.

Macfayden

L. P.

Dawson

(2010) Mining LMS data to develop an “early warning system” for educators: A proof of concept. Internet & Higher Education 54: 588–599. doi:10.1016/j.compedu.2009.09.008.

38.

Mayordomo

R. M.

Onrubia

(2015) Work coordination and collaborative knowledge construction in a small group collaborative virtual task. The Internet and Higher Education 25: 96–104. doi:10.1016/j.iheduc.2015.02.003.

39.

Mohammed

Harrison

D. A.

(2013) The clocks that time us are not the same: A theory of temporal diversity, task characteristics, and performance in groups. Organizational Behavior and Human Decision Processes 122(2): 244–256. doi:10.1016/j.obhdp.2013.08.004.

40.

Nahrgang

J. D.

DeRue

D. S.

Hollenbeck

J. R.

Spitzmuller

Jundt

D. K.

Ilgen

D. R.

(2013) Goal setting in teams: The impact of learning and performance goals on process and performance. Organizational Behavior and Human Decision Processes 122(1): 12–21. doi:10.1016/j.obhdp.2013.03.008.

41.

E. G.

Huang

W. H. D.

Mehdiabadi

A. H.

(2018) Facilitating critical thinking in asynchronous online discussion: Comparison between peer-and instructor-redirection. Journal of Computing in Higher Education 30: 489–509. doi:10.1007/s12528-018-9180-6.

42.

Ouyang

Chang

Y. H.

(2018) The relationships between social participatory roles and cognitive engagement levels in online discussions. British Journal of Educational Technology. Advance online publication. doi:10.1111/bjet.12647.

43.

Richardson

J. C.

Ice

(2010) Investigating students' level of critical thinking across instructional strategies in online discussions. The Internet and Higher Education 13(1–2): 52–59. doi:10.1016/j.iheduc.2009.10.009.

44.

Rourke

Anderson

(2002) Using peer teams to lead online discussions. Journal of Interactive Media in Education 1: 1–21. doi:10.5334/2002-1.

45.

Rourke

Anderson

Garrison

D. R.

Archer

(2007) Assessing social presence in asynchronous text-based computer conferencing. International Journal of E-Learning & Distance Education 14(2): 50–71.

46.

Salmon

(2004) E-moderating: The key to teaching and learning online, London, England: Routledge.

47.

Salmon

(2013) E-tivities: The key to active online learning, London, England: Routledge.

48.

Sauer

N. C.

Kauffeld

(2016) The structure of interaction at meetings: A social network analysis. Zeitschrift für Arbeits-und Organisationspsychologie A&O 60: 33–49. doi:10.1026/0932-4089/a000201.

49.

Scott

(2000) Social network analysis: A handbook, London, England: SAGE.

50.

Shea

Bidjerano

(2010) Learning presence: Towards a theory of self-efficacy, self-regulation, and the development of a communities of inquiry in online and blended learning environments. Computers & Education 55(4): 1721–1731. doi:10.1016/j.compedu.2010.07.017.

51.

Shea

Hayes

Uzuner

Vickers

Bidjerano

Gozza-Cohen

Tseng

(2013) Online learner self-regulation: Learning presence viewed through quantitative content and social network analysis. The International Review of Research in Distance Learning 14(3): 427–461. doi:10.19173/irrodl.v14i3.1466.

52.

Shea

Hayes

Vickers

Gozza-Cohen

Uzuner

Mehta

Rangan

(2010) A re-examination of the community of inquiry framework: Social network and content analysis. Internet and Higher Education 13(4): 10–21. doi:10.1016/j.iheduc.2009.11.002.

53.

Siemens

Gašević

Haythornthwaite

Dawson

Shum

S. B.

Ferguson

Baker

R. S. J. D.

(2011) Open learning analytics: An integrated modularized platform (Doctoral dissertation, Open University Press).

54.

Sparrowe

Liden

Wayne

Kraimer

(2001) Social networks and the performance of individuals and groups. Social networks and the performance of individuals and groups. The Academy of Management Journal 44(2): 316–325.

55.

Swan

Garrison

D. R.

Richardson

(2009) A constructivist approach to online learning: The community of inquiry framework. In: Payne

C. R.

(ed) Information technology and constructivism in higher education: Progressive learning frameworks, Hershey, PA: IGI Global., pp. 43–57. doi:10.4018/978-1-60566-654-9.ch004.

56.

Thye, S. R., Yoon, J., & Lawler, E. J. (2002). The theory of relational cohesion: Review of a research program [Electronic version]. Retrieved from http://digitalcommons.ilr.cornell.edu/articles/1220.

57.

Thormann

Gable

Fidalgo

Blakeslee

(2013) Interaction, critical thinking, and social network analysis (SNA) in online courses. The International Review of Research in Open and Distributed Learning 14(3): 294–318. doi:10.19173/irrodl.v14i3.1306.

58.

Tirado-Morueta, R., Hernando-Gómez, Á., & Aguaded-Gómez, J. I. (2015). The effect of centralization and cohesion on the social construction of knowledge in discussion forums. Interactive Learning Environments, 23(3), 293–316. doi:10.1080/10494820.2012.745437.

59.

Tirado-Morueta, R., Maraver-López, P., Hernando-Gómez, Á., & Harris, V. W. (2016). Exploring social and cognitive presences in communities of inquiry to perform higher cognitive tasks. The Internet and Higher Education, 31, 122–131. doi:10.1016/j.iheduc.2016.07.004.

60.

Tirado-Morueta, R., Maraver-López, P., & Hernando-Gómez, Á. (2017). Patterns of Participation and Social Connections in Online Discussion Forums. Small Group Research, 48(6), 639–664. doi:10.1177/1046496417710726.

61.

Troster

Mehra

Van Knippenberg

D. L.

(2014) Structuring for team success: The interactive effects of network structure and cultural diversity on team potency and performance. Organizational Behavior and Human Decision Processes 124(2): 245–255. doi:10.1016/j.obhdp.2014.04.003.

62.

Van der Haar

Koeslag-Kreunen

Euwe

Segers

(2017) Team leader structuring for group effectiveness and group learning in command-and-control groups. Small Group Research 48(2): 215–248. doi:10.1177/1046496417689897.

63.

Venkataramani

Zhou

Wang

Liao

Shi

(2016) Social networks and employee voice: The influence of group members' and group leaders' social network positions on employee voice. Organizational Behavior and Human Decision Processes 132: 37–38. doi:10.1016/j.obhdp.2015.12.001.

64.

Wang, Y. & Li, X. (2007). Social network analysis of interaction in online learning communities. Paper presented at Seventh IEEE international conference on advanced learning technologies, Niigata, Japan. doi:10.1109/ICALT.2007.228.

65.

Wang

Han

Fisher

C. M.

Pan

(2017) Learning to share: Exploring temporality in shared leadership and group learning. Small Group Research 48(2): 165–189. doi:10.1177/1046496417690027.

66.

Wicks

D. A.

Craft

B. B.

Mason

G. N.

Gritter

Bolding

(2015) An investigation into the community of inquiry of blended classrooms by a Faculty Learning Community. The Internet and Higher Education 25: 53–62. doi:10.1016/j.iheduc.2014.12.001.

67.

Wise

(2014) Can a team have too much cohesion? The dark side to network density. European Management Journal 32(5): 703–711. doi:10.1016/j.emj.2013.12.005.

68.

Yang

H. L.

Tang

J. H.

(2004) Team structure and team performance in IS development: A social network perspective. Information & Management 41: 335–349.

69.

Yoon, J., & Lawler, E. J. (2006). Relational cohesion model of organizational commitment [Electronic version]. In O. Kyriakidou & M. F. Özbilgin (Eds.), Relational perspectives in organizational studies (pp. 138–162). Northampton, MA: Edward Elgar Publishing.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.52 MB