It is about timing: Network prestige in asynchronous online discussions

Abstract

Lay Description
What is already known about this topic: Social interaction is important for learning, and asynchronous online discussions are broadly used to support social learning.
Teaching practices and strategies have been developed to support student participation in online discussions.
In some contexts, participation intensity predicts learning performance.
However, researchers have documented a participation gap in online discussions.
What this paper adds: We applied prestige, a sociological term, to a study of the participation gap in an online undergraduate class and revealed varied levels of prestige among students.
Lower prestige students were not inactive; rather, they initiated an equivalent volume of connections with peers but were often ignored.
The timing of discussion participation can be a probable contributor to the gap, as lower prestige students' participation was less timely and more compressed together.
Implications for practice and/or policy: Pedagogical supports are needed to help students develop advanced self‐regulation skills conducive to productive discussion participation.
The design of online discussion environments and analytics tools needs to consider students' discussion dispositions and behaviours.

Full text

Running head: PRESTIGE IN ONLINE DISCUSSIONS 1
It Is About Timing: Network Prestige in Asynchronous Online Discussions
Bodong Chen, University of Minnesota
Tianhui Huang, Yunnan University
Author Note
(in press). Journal of Computer Assisted Learning. DOI: 10.1111/jcal.12355

PRESTIGE IN ONLINE DISCUSSIONS
2
Abstract
Asynchronous online discussions are broadly used to support social learning. This paper reports
on an undergraduate class’s online discussion activities over one semester. Applying Social
Network Analysis, this study revealed a participation gap among students reflected by their
varied levels of network prestige. The low-prestige group initiated equivalent volumes of
interactions but were less reciprocated. In-depth analysis found the high-prestige group also
advantageous in other network measures such as closeness centrality and eigenvector centrality,
as well as the strength, persistence, and reciprocity of their ties. To probe potential explanations
of the revealed gap, we further contrasted post content and posting behaviors between two
groups. Results did not identify any significant differences in post content but found low-prestige
students’ participation less timely and more temporally compressed. This paper calls for
attention to the participation gap in online discussions, micro-level temporal patterns of student
activities, and practical means to scaffold student participation in asynchronous online
discussions.
Keywords: asynchronous online discussions; online learning; social network analysis;
temporal analysis; computer-mediated communication

PRESTIGE IN ONLINE DISCUSSIONS
3
It Is About Timing: Network Prestige in Asynchronous Online Discussions
1. Introduction
Asynchronous online discussions have been broadly used to support social interaction in
various contexts of learning that follow the social constructivist paradigm. In online and blended
classes, students are often instructed to post reflections, share ideas, and comment on each
other’s posts in online forums (Chan, Hew, & Cheung, 2009). In programming classes, students
embark on sophisticated information-seeking journeys on social question and answer (Q&A)
sites (e.g., Stack Overflow) to achieve learning goals (Lu, Sharon, & Hsiao, 2016). In massive
open online courses, discussion forums provide the main mechanism for learners to forge
connections with peers from around the world (Rosé & Ferschke, 2016). In less structured
learning settings, social media and online discussion spaces such as Reddit play critical roles in
connecting lifelong learners with information and learning partners (Greenhow & Robelia, 2009;
Haythornthwaite et al., 2018). Despite technological differences of these asynchronous online
discussion spaces, they share a same purpose of facilitating social interaction among learners.
Social interaction in asynchronous online discussions has attracted substantial research
attention, leading to empirical findings that have theoretical, pedagogical, and technological
implications. Prior research of asynchronous online discussions has studied the instructor’s role
(Mazzolini & Maddison, 2007), peer facilitation strategies (Chan, Hew, & Cheung, 2009),
technological design of discussion environments (Guzdial & Turns, 2000; Scardamalia &
Bereiter, 2003), interaction patterns (Dringus & Ellis, 2010; Jo, Park, & Lee, 2017; Wise & Chiu,
2011), and community structures (So, 2009; Wise & Cui, 2018), leading to various strategies for
facilitating effective asynchronous online discussions.
Despite substantial work in this area, an enduring problem facing asynchronous online
discussions is a participation gap among learners. The participation gap is not only reflected by
different rates of participation by learners, different levels of conceptual engagement, and
different circles of interacting peers (Dawson, 2010; Vaquero & Cebrian, 2013). More
profoundly, the participation gap engenders varied social configurations that can facilitate or
undermine learner performance. To support success of all learners in social constructivist
learning environments, more research is needed to understand the participation gap and probe its
possible causes.
With this recognition, the present paper attempts to reveal, characterize, and explain
network prestige in an asynchronous online discussion space of an undergraduate class. In a
directed social network, prestige reflects an actor’s success in attracting peer attention in relation
to the actor’s activity level. By focusing on the sociological construct of prestige, this study
seeks to advance our understanding of social dynamics in online discussions and to suggest
practical means to mitigate the participate gap. Below, we first survey pertinent education and
sociology literature. Then we introduce the research context and study design. After presenting
and discussing main findings, we discuss scholarly significance of the study, practical
implications, and future directions.

PRESTIGE IN ONLINE DISCUSSIONS
4
2. Literature Review
2.1 Social Participation in Online Discussions
Learning is inherently social, no matter whether it is about language development of
children (Carey & Bartlett, 1978), science learning in K-12 classrooms (Chin & Osborne, 2010),
or the development of the newest surfing tricks by avid hobbyists (Hagel, Brown, & Davison,
2012). Over the past decades, this recognition has motivated an expansion of the
conceptualization of learning from acquiring knowledge as entities, to include participating in
social activities that are conducive to knowing (Sfard, 1998). Social participation is essential in
formal classrooms as well as in emerging contexts of learning, where learning happens through
legitimate peripheral participation, game plays, new media practices, and artifact fabrication (Ito
et al., 2013; Jenkins, 2006).
The recognized importance of social participation for learning has triggered widespread
efforts to foster it, especially through the design and use of digital environments. Back in the
1980s, the Computer Supported Intentional Learning Environment (CSILE) was developed to
support group learning mediated by conceptual artifacts stored in a virtual discussion space
(Scardamalia, Bereiter, & Lamon, 1994). The Web 2.0 movement has greatly expanded the
online space for social encounters, leading educators to explore social affordances on various
social media platforms (Greenhow & Robelia, 2009; Lewis, Pea, & Rosen, 2010). Gaming
environments and virtual worlds such as Civilization and Minecraft are also harnessed to support
rich modalities of social learning (e.g., Dezuanni, O’Mara, & Beavis, 2015).
Among various digital tools, asynchronous online discussions, one of the most popular
computer-mediated communication (CMC) technologies, are widely applied to supporting social
participation in education. Since its inception, CMC has been recognized as an important
learning resource that can support learner communication and shape the formation of learning
cultures (McAteer, Tolmie, Duffy, & Corbett, 1997). In particular, asynchronous online
discussions are often used to facilitate idea sharing, resolve student confusion, coordinate
cognitive actions, support negotiation, scaffold group knowledge building, and maintain social
bonds within a learner community (Fields et al., 2016; Resnick, Säljö, Pontecorvo, & Burge,
1997; Stahl, 2006). In successful cases, online discussions can engender positive student
experiences by enabling discussion that is active, interactive, constructive, and hereby conducive
to learning (Chi & Menekse, 2015).
Empirical studies of asynchronous online discussions have broadly covered interaction
patterns, pedagogical practices, potential benefits, and implementation challenges. For instance,
Hewitt (2005) studied patterns in online discussions and found the learner’s “single-pass
discourse practices” contributing to premature abandonment of discussion topics. While online
forums provide affordances for learner communication, their thread-based design often leads to
repetition and duplication in different threads (Thomas, 2000). Many pedagogical strategies have
been proposed to mitigate these challenges, such as engaging students to take different roles
(Wise & Chiu, 2011), incorporating different questioning strategies (Wang, 2005), and
scaffolding students’ socio-metacognitive skills (Borge & White, 2016). Studies probing the

PRESTIGE IN ONLINE DISCUSSIONS
5
efficacy of online discussions found participation intensity predictive of learning performance
(Jo et al., 2017; Joksimović, Gašević, Kovanović, Riecke, & Hatala, 2015; Russo & Koesten,
2005), pointing towards the potential benefits of scaffolding learner participation in online
discussions.
2.2 The Participation Gap in Online Discussions
Despite benefits of online discussions reported in the literature, social participation in
online discussions often varies among learners. For example, Vaquero and Cebrian (2013) report
a “rich club” phenomenon that encourages social interactions among high-performing students
while posing barriers to participation against low-performing students who could not effectively
blend in. Similar findings are independently reported elsewhere (Dawson, 2010), calling for
more research on the participation gap in online discussion environments. Admittedly, these
observations of the participation gap may have unique sociological determinants, given elite
players, celebrities, or “VIP-clubs” are commonly observed in various social contexts (e.g.,
Masuda & Konno, 2006). Notwithstanding, understanding the nature and mechanism of the
participation gap in online discussions is imperative for both research and practice.
We conceptualize the participation gap in online discussions as a complex, dynamic
phenomenon that is shaped by behavioral, cognitive, social, and emotional characteristics of
participants, together with social structures, rules, norms, and cultures forming within the
discussion environment. The shifting participation gap can be characterized using sociological
terms and network metrics. For instance, the participation gap in online classes can be
characterized as a rich-club phenomenon (Vaquero & Cebrian, 2013), a sociological term widely
applied and jointly developed in fields such as computer networks (Zhou & Mondragon, 2004),
sports management (Masuda & Konno, 2006), and neuroscience (Daianu et al., 2014; van den
Heuvel & Sporns, 2011). At the individual level, another sociological term prominence (Burt,
2000) can be used to characterize an actor’s favorable position in a network and advantages that
come with such positions. According to Knoke and Burt (1983), “An actor is prominent to the
extent that he is involved in relationships that make him an especially visible member of a social
system” (p. 198). Network prominence is further categorized into network centrality and
prestige—respectively for undirected/symmetric and directed/asymmetric networks. The concept
of centrality is concerned with the extent to which an actor is involved in symmetric
relationships, whereas prestige captures the directionality embedded in prominence (Torfason &
Kitts, 2011). In particular, high-prestige actors can attract more incoming connections while low-
prestige actors initiate outgoing connections but are not necessarily reciprocated (Knoke & Burt,
1983; Torfason & Kitts, 2011). Varied levels of prestige contribute to the emergence of self-
esteem and social capital of individuals (Ellison, Steinfield, & Lampe, 2007), as well as social
stratification of larger social systems (Wegener, 1992). Related to prestige, other sociological
constructs, such as preferential attachment (Barabasi & Albert, 1999) and the Matthew effect
(Perc, 2014), are also observed in real-world scholarly communication and computer networks.
In empirical studies, the prestige concept is operationalized using network metrics such as
indegree centrality (Russo & Koesten, 2005) and eigenvector centrality (Bonacich, 1987;

PRESTIGE IN ONLINE DISCUSSIONS
6
Bergstrom, 2007). For the investigation of online discussions, the multifaceted concept of
prestige and its related metrics can help us move beyond basic counts of discussion behaviors to
capture the interactional and asymmetric aspects of the participation gap (e.g., Cho, Stefanone, &
Gay, 2002).
2.3 Factors Related to Participation in Online Discussions
Previous studies have reported various factors contributing to different levels of learner
participation in online discussions (Nandi et al., 2012; Ouyang & Scharber, 2017). Activity
design and instructor presence matter, as researchers find them affecting online discourse in a
class (Dennen, 2005). Facilitation techniques also matter because the Resolving and
Summarizing facilitation techniques may lead to early thread termination and the Questioning
technique is needed to enhance thread continuity (Chan, Hew, & Cheung, 2009). Technological
features of forums also play important roles, as they help learners understand how, what, and
where to contribute (Guzdial & Turns, 2000; Warren & Rada, 1998). Grounded in an informed
synthesis of the literature, Zingaro and Oztok (2012) constructed a statistical model that predicts
the likelihood a post would receive a reply; their model is based on six quantitative predictors:
the note’s posting date, written by active or inactive participant, ease of reading, word count,
written by student or instructor, and containing a question. They found notes that were longer,
written early in the week, and containing at least one question were more likely to be replied,
while readability did not make a difference in their study context. Overall, learner participation
in online discussions is multifaceted and has been related to many factors.
2.4 Methodological Approaches to Investigating Participation in Online Discussions
Studies of learner participation in online discussions have benefited from multiple
methodologies (Jo et al., 2017). A wide range of methodological approaches, such as Social
Network Analysis, Content Analysis, and Regression Analysis, have been applied to studying
student participation and its contributing factors. For instance, both social network visualizations
and indices are widely adopted to depict and/or to measure social participation (Chen, Chang,
Ouyang, & Zhou, 2018; Wise & Cui, 2018). Measures of discussion behaviors—both writing and
reading—are derived from system log files to describe learner participation and to predict
learning performance (Jo et al., 2017; Wise, Hausknecht, & Zhao, 2014). Content analysis is also
widely used to assess the quality of post content (e.g., Chen, Chang, Ouyang, & Zhou, 2018;
Wise, Hausknecht, & Zhao, 2014).
A subset of these studies are uniquely focused on the temporal dimension—which is
important conceptually because asynchronous online discussions are transactional and unfold
over time. Hesse et al. (1988) delineate four temporal qualities: (a) temporal scale—“the scope
or duration of events and relationships”; (b) temporal sequencing—“the specific patterns of
activities within a recurrent or nonrecurrent event”; (c) pace—“the rate at which events pass in a
specified unit of time” or “an individual’s psychological experience of time”; and (d) temporal
salience—“the degree to which an individual’s thoughts, feelings, and actions are past-, present-,
or future-oriented” (p. 157). Building on this taxonomy, Wise, Zhao, Hausknecht and Chiu
(2014) further contextualize these temporal aspects in online discussions and provided specific

PRESTIGE IN ONLINE DISCUSSIONS
7
guidance for temporally-aware analysis and design. Such temporally-aware analysis could either
focus on a single temporal quality, such as temporal sequencing (Chen, Resendes, Chai, & Hong,
2017), or integrate temporal analysis with other analytical approaches (Häkkinen, 2013). Despite
a rising interest in applying temporal analysis in educational research (Chen, Knight, & Wise,
2018), the potential of temporal analysis for examining the participation gap in asynchronous
online discussions remains under-tapped.
2.5 The Present Study
The present study explored student participation, and especially the participation gap, in
asynchronous online discussions. Within the context of an online undergraduate course, we
attempted to determine the extent to which network prestige was reflected in student interactions
and, if so, explore its potential contributing factors related to students’ post content and posting
behaviors. Research questions of the study included:
1. Did students demonstrate varied levels of prestige in online discussion activities?
2. How was prestige reflected in the discussion environment?
3. Which factors could have contributed to the emergence of prestige?
3. Methods
3.1 Research Context and Participants
The study was situated in an undergraduate, fully-online course at a large public
university in the United States. As of the study, this course, which covered topics related to
technology, ethics, and society, was listed as a liberal-education elective for undergraduate
students across the university. One section of this course (n = 20) participated in this study.
Participating students came from a wide range of disciplines such as psychology, medicine,
biochemistry, economics, management, and journalism; they were not concentrated in one
college so unlikely to know each other prior to this course.
Informed by social constructivist learning theories, the course design included
asynchronous online discussions participated by students weekly on a social learning platform
named Yellowdig. Except for the first and last weeks of the semester when the class
communicated through a video-based discussion tool (named Flipgrid), students discussed
readings and ideas on Yellowdig in the other 13 weeks. By the end of Monday in each week, they
were required to minimally contribute one initial post and make two comments on posts from
peers. Figure 1 shows an excerpt of the Week 4 discussion prompt provided to students.

PRESTIGE IN ONLINE DISCUSSIONS
8
Figure 1. Yellowdig discussion prompt for Week 4 of the class.
The discussion environment Yellowdig positions itself as a social learning platform that
resembles social network sites such as Facebook and Twitter. As illustrated in Figure 2, students
could initiate a post, called pins on Yellowdig, which can be commented on and reacted to (e.g.,
“Like,” “Love”) by members of the class community. Students could also explicitly mention
each other in a post or comment (see Figure 2). As of the study, Yellowdig supported only one-
level threading, and displayed a new comment on an existing comment as a comment on the top-
level post, while the actual student being commented on got mentioned in the new comment.
Like most social network sites, Yellowdig was also implemented with a news feed powered by a
sorting algorithm, which placed posts with the most recent activities on the top. In comparison
with discussion forums found in popular learning management systems (e.g., Moodle and
Canvas), Yellowdig provided richer technological features for social interaction.
The instructor’s presence on Yellowdig in this course could be characterized as “high” in
the first three weeks and “medium” in the other weeks (Dennen, 2005). To nurture a sense of
community, he spent a great deal of time during the first three weeks commenting on students
posts (making more than 10 comments per week) and connecting students with each other by
mentioning multiple names in a same comment. For the rest weeks, his presence on Yellowdig

PRESTIGE IN ONLINE DISCUSSIONS
9
was reflected by his reactions (e.g., “liking” a post or comment) and sparse comments (1-2 per
week). Overall, he did not dominate the discussion at any point of the semester.
Figure 2. Yellowdig interface. The main interface displays a post (also called pin on Yellowdig),
with its comments listed below.
3.2 Data and Analyses
Interaction data on Yellowdig, including 274 posts/pins, 514 comments, 36 mentions, and
74 reactions, were the primary data source. This dataset provided comprehensive information

PRESTIGE IN ONLINE DISCUSSIONS
10
about students’ discussion activities. Course materials, including the syllabus, weekly
announcements, and readings, were included as secondary data sources to contextualize primary
data analyses.
Social Network Analysis (SNA) was adopted as the main analytical approach. All
computational procedures were carried out in R (an open-source computational environment)
and its igraph package developed for network analysis and visualization. To examine the first
question about varied levels of prestige among students, we first constructed a directed social
network based on students’ commenting interactions on Yellowdig. In the constructed network,
each node denoted a student and each directed edge represented the existence of minimally one
comment from one student to another. We did not include mentioning and reacting interactions
due to their sparsity. To represent network prestige of each student in the network, we computed
indegree centrality of each node, which measures the number of nodes pointing towards a
particular node (Russo & Koesten, 2005).
To further characterize the manifestation of network prestige, we divided students into
two groups with high- and low-level prestige based on indegree centrality. After grouping, we
examined group differences in a number of network measures to further characterize prestige in
the studied context.
To answer the final research question about potential contributing factors of prestige, we
compared two groups in terms of their post content (including length, questioning activity, and
readability) and posting behaviors (including timing and temporal dispersion of posts). The
choices of these measures were informed by the literature. By contrasting two groups, we sought
to tentatively explain the emergence of network prestige in this particular context.
4. Findings
4.1 Descriptive Overview of the Interaction Network
Table 1 presents basic statistics of discussion activities in the whole interaction network
involving students and the instructor. On average, each participant authored 13.7 posts and 25.7
comments, and connected with 11 class members for 3.42 times each week. Figure 3(a) presents
a visualization of the interaction network, with each node denoting a participant and each edge
representing aggregated reply interactions between two participants. The instructor was not
found central in the visualized network (see Figure 3(a)); as described earlier, he actively
commented on student posts only in the first three weeks.
Network measures in Table 1, particularly network density (55%), centralization (0.28),
and mean distance (1.51) (Scott, 2012), indicated that this network was fairly decentralized and
well connected for this particular context (Ergün & Usluel, 2016). Students’ average degree,
indicating the number of peers they linked with, was consistently between 3 and 4 across all
weeks. The most frequent words used in Yellowdig posts and comments, shown in Figure 3(b),
were relevant to course content.

PRESTIGE IN ONLINE DISCUSSIONS
11
Table 1
Descriptive statistics of the complete commenting network.
Variable
Value
Total posts/pins
274
Total comments
514
Mean degree
20.90
Mean weighted degree
45.40
Mean tie strength
2.07
Network density
0.55
Mean distance
1.51
Degree centralization
0.28
Figure 3. Overview of the interaction network. (a) Sociogram of the whole network. Node size
denotes prestige measured by indegree centrality, and node color denotes the group membership
(including Group A, Group B, and the Teacher). (b) A word cloud of top words that appeared in
post content on Yellowdig.
4.2 Did Students Demonstrate Varied Levels of Prestige in Online Discussion Activities?
Indegree centrality was computed to measure each student’s network prestige. The class
was then divided into two groups: Group A (n = 10) constituted by students with higher indegree
centrality (M = 15.1, SD = 2.28), and Group B (n = 10) with students having lower indegree
centrality (M = 5.8, SD = 2.97). As indicated by the varied levels of indegree centrality, Group A
students received comments from 15 peers on average, whereas Group B students were
commented on by approximately 6 classmates. The average number of comments received by
Group A students was 4.28 times as many as Group B students (Ms = 36.8 and 8.6).
Despite different levels of indegree centrality, two groups had equivalent outdegree and
outgoing interactions, indicating higher volumes of unreciprocated interactions from Group B
students and hence their lower prestige status (Knoke & Burt, 1983). Paradoxically, Group B did

PRESTIGE IN ONLINE DISCUSSIONS
12
not demonstrate less efforts in forging connections with peers, given their equivalent volume of
outgoing interactions, but somehow achieved lower network prestige (lower indegree and fewer
incoming interactions) than Group A.
4.3 How Was Prestige Reflected in the Discussion Environment?
Further analyses sought to characterize students’ varied levels of prestige from different
angles. First, we compared posting activities of two groups. On average, two groups contributed
equivalent volumes of posts (Ms = 14.0 and 12.6) and comments (Ms = 25.4 and 21.0). However,
85.96% of initial posts from Group A were responded, significantly higher than 51.32% of Group
B (t(11.56) = 6.18, p < .001). Posts created by Group A also received more comments (M = 2.66
vs. Group B’s M = 0.71) and sustained longer discussion activities (M = 33.15 hours vs. Group
B’s M = 8.69 hours). Notably, posts shared by two groups received different levels of attention in
this class.
Table 2
Independent sample t-tests of SNA measures between two groups.
Note: **—p<.01, ***—p<.001.
Group A
(n = 10)
Group B
(n = 10)
SNA measures
M
SD
M
SD
t
df
p
In-degree
15.1
2.28
5.8
2.97
7.84
16.87
0.000***
Out-degree
10.0
1.83
10.9
3.21
-0.77
14.26
0.454
Incoming interactions
36.8
14.0
8.6
5.06
6.01
11.31
0.000***
Outgoing interactions
24.6
3.53
20.8
7.67
1.42
12.66
0.179
Avg. tie strength
2.43
0.51
1.71
0.37
3.62
16.48
0.002**
Betweenness
12.0
8.34
7.22
7.87
1.31
17.94
0.205
Closeness
0.05
0.00
0.04
0.00
4.41
14.61
0.000***
Eigenvector
0.74
0.21
0.33
0.15
5.04
16.10
0.000***
Authority
0.61
0.26
0.12
0.08
5.66
10.51
0.000***
Word count of pins
178
32.6
160
37.8
1.13
17.62
0.274
Word count of
comments
34.9
5.05
38.3
14.2
-0.70
11.23
0.499
Percentage of posts
being responded
86%
6%
51%
17%
6.18
11.56
0.000***
Average count of
comments to a post
2.82
0.87
0.79
0.34
6.89
11.63
0.000***

PRESTIGE IN ONLINE DISCUSSIONS
13
Second, we contrasted several social network indices between two groups. Conceptually,
network indices applied in this study may correlate with the prestige measure (i.e., indegree
centrality) but would still reveal different characteristics of the network. As shown in Table 2, a t-
test of the average tie strength demonstrated significant difference between two groups (t(16.48)
= 3.62, p < .01); Group A did not only have more incoming connections but also built stronger
ties than Group B. Further comparisons of closeness and eigenvector centralities also
demonstrated significant differences between two groups, favoring Group A; Group A students
were closer to peers and connected to peers with higher prestige. Group A students did not have
significantly higher betweenness centrality, however, showing that they did not do a superior job
on bridging peer interactions than Group B students. This finding indicated that those denser and
stronger connections of Group A students were likely to be concentrated among themselves
instead of benefiting lower-prestige peers.
Given the importance of directionality for network prestige, we further inspected the
direction of student interactions and divided the interactions into 4 categories: (1) A→A
interactions within Group A, (2) B→B interactions within Group B, (3) A→B interactions from
Group A to Group B, and (4) B→A interactions from Group B to A. Results showed that among
all directed interactions, 45.20% were A→A interactions, followed by 35.15% B→A, 10.26%
B→B, and merely 8.52% A→B. This pattern held true when we inspected temporal changes
across weeks (see Figure 4). Notably, Group B students were actively initiating contact with
Group A but were seldom reciprocated, and Group A students were more likely to comment on
peers with similar prestige.
Figure 4. Four different types of interactions across 13 weeks.

PRESTIGE IN ONLINE DISCUSSIONS
14
Disparity did not only figure in the presence of connections but also the quality of them,
such as reciprocity and persistence (Vaquero & Cebrian, 2013). Here, we define a connection
between X and Y as being reciprocal within one week when an X→Y interaction is reciprocated
by a Y→X interaction during the same week. Analysis of the network identified 59 reciprocal
connections across all weeks, which could be further divided into 28 reciprocal connections
within Group A, 26 reciprocal connections between Group A and Group B, and the rest 5 within
Group B. Among these reciprocal connections, only 16% reciprocal connections were initiated
by Group B students, whereas the rest 84% were initiated by Group A. Considering students
from two groups initiated equivalent volumes of interactions, it became clear that connections
initiated by Group A students were more likely to be reciprocated.
Analysis of persistent connections arrived at similar findings. Given this course’s time
span, persistent connections in this context were operationalized as connections that occurred in
at least five weeks. Among 23 persistent connections found in the student network, 65.22% were
within Group A, 30.43% between two groups, and the rest 4.35% within Group B. Further
analysis found two unique persistent ties were pointing from Group B towards Group A, and no
persistent connections was directing from Group A to Group B. In other words, persistent
interactions mostly happened within the high-prestige group, and when they did occur between
groups, they were persisted by students in the low-prestige group and not reciprocated by the
high-prestige group.
To summarize, prestige manifested in this online class in multiple ways. A series of
comparisons found the high-prestige group had more incoming interactions, stronger ties, higher
centrality scores, higher odds of being responded, and higher-quality ties—regardless of
equivalent measures of outdegree, outgoing interactions, number of posts, and betweenness
centrality of two groups. Namely, students from two groups initiated equivalent amount of
interactions with equivalent number of peers, but Group A students received significantly more
comments than others, occupied more favorable positions in the network, and maintained
stronger and higher-quality ties with peers they were connected with—all indicating higher
prestige in this network. Their favorable positions were reflected in closeness with peers and
being connected with higher-prestige peers (i.e., eigenvector centrality) but not in bridging
connections among peers (i.e., insignificant difference in betweenness centrality).
4.4 Which Factors Could Have Contributed to the Emergence of Prestige?
The third research question was concerned with factors that might have contributed to the
gap in network prestige. Under the guidance of literature, we inspected a number of factors
related to students’ post content and posting behaviors, including post length, questioning activity,
post readability, and temporal features of posting behaviors.
4.4.1 Post length
Was the disparity in prestige a function of the informativeness of student posts? Did
Group A students receive more comments because their posts were longer? Group comparisons
of post length found nonsignificant differences in the average length of either initial posts or
comments (ps > .05).

PRESTIGE IN ONLINE DISCUSSIONS
15
4.4.2 Questioning activity
Did Group A students receive more replies because their posts were soliciting responses
by posing questions? We used the presence of a question mark as an approximate indicator of a
question. Results indicated posts shared by two groups did not differ in terms of questioning
activity (χ2(1) = 0.02, p = .88).
4.4.3 Post readability
Prior studies also examined the role played by text readability in online discussion
(Zingaro & Oztok, 2012). To gauge the linkage between readability and network prestige, we
computed several readability indices, including the Flesch Reading Ease score, the Average
Grade Level, and the Automated Readability Index. Contrasts of these readability scores between
two groups did not identify any significant difference (ps > .10).
Figure 5. Time difference between student pins and the weekly deadlines. The vertical lines
indicate Group A students were on average 49.9 hours earlier, while Group B students were 20.7
hours earlier than the deadline.
4.4.4 Temporal features of discussion behaviors
Timeliness of discussion behaviors. Informed by the literature (Hesse, Werner, & Altman,
1988; Wise et al., 2014), we compared the timeliness of student posts in light of weekly
deadlines between two groups. As indicated by density curves in Figure 5, most Group B
students’ posts were barely on time, whereas most Group A students posted one day in advance.
Group A and B students significantly differed in the “time delta” of their posts (i.e., the time

PRESTIGE IN ONLINE DISCUSSIONS
16
difference between a post and the deadline) (t(246) = -6.15, p < .001); the same trend was
observed for comments (t(350) = -6.42, p < .001).
Dispersion of discussion behaviors. Conceptually, dispersion means the extent to which
students’ posting behaviors were clustered or dispersed compared to the weekly participation
timeline (Hesse, Werner, & Altman, 1988). Grounded in this understanding, we computed the
index of dispersion (Cox, 1966) and compared it between two groups. Results indicated Group A
students’ discussion behaviors were marginally more dispersed than Group B (t(151) = 1.85, p
= .06). When inspecting the dispersion index across weeks, we found two groups started from a
similar level of dispersion and Group B students’ dispersion index declined over the semester,
meaning their posting behaviors became increasingly jammed together in later weeks.
Availability of prior posts for commenting. Due to individual differences in the timing
and dispersion of participation, each week students might have had different number of available
posts to comment on. By reconstructing the discussion context of each individual comment, we
found that on average a Group B student would have 9.50 Group A posts and 5.22 Group B posts
to comment on, whereas a Group A student would have 6.61 Group A posts and 2.05 Group B
posts to respond to. Lower availability of posts from Group B students could have contributed to
their lower incoming connections. Because of their different temporal participation patterns,
Group A students did not have abundant supplies of Group B students to respond to.
Taken together, while two groups did not significantly differ in post content, they
demonstrated significantly different posting behaviors that could have contributed to their varied
prestige levels. Discussion contributions made by Group B students were less timely and less
temporally dispersed (i.e., more pressed together). These temporal participation patterns could
have led to lower availability of Group B posts for Group A students to comment on.
5. General Discussion
In this study, we asked three major questions regarding: (1) Did students demonstrate
varied levels of prestige in online discussions? (2) How was prestige reflected in the discussion
environment? and (3) Which factors could have contributed to the emergence of prestige? Using
indegree centrality as a measure of prestige (Knoke & Burt, 1983; Russo & Koesten, 2005), we
found varied levels of prestige reflected in the student network. When students were divided into
two groups based on the prestige measure, even if they had equivalent outdegree and outgoing
interactions, they differed significantly in incoming interactions. The lower-prestige group did
not demonstrate less efforts in connecting with peers but somehow achieved lower network
prestige.
To answer the second question, further in-depth analysis was conducted from different
aspects. Results showed multiple ways in which prestige manifested in this online class. In
particular, posts shared by two groups received different levels of attention. Even though two
groups had equivalent measures of outdegree, outgoing interactions, number of posts, the high-
prestige group had more incoming interactions, stronger ties, higher centrality scores, higher

PRESTIGE IN ONLINE DISCUSSIONS
17
odds of being responded, and higher-quality ties. They were closer to peers and were connected
with higher-prestige peers but did not help bridge connections among peers.
Analyses responding to the third question attempted to explain the identified gap in
network prestige. Factors related to post content, including length, readability, and questioning
activity, were nonsignificant predictors. In contrast, two groups were found to differ in micro-
level temporal patterns including the timing of posts each week, dispersion of posting behaviors,
and availability of posts to be commented on for each group. The high-prestige group generally
posted earlier each week, spaced weekly posting activities more evenly, and therefore left peers
ample opportunities to react to their ideas. The lower prestige group did the opposite: they posted
closer to the deadline, posted hastily, and did not provide peers (especially Group A students)
enough time to engage with their posts. Understandably, when the high-prestige group attempted
to leave comments, they would have a small number of low-prestige group posts to choose from.
5.1 Scholarly Significance of Investigating Network Prestige and Temporality
Significance of this study ought to be discussed with a few limitations in mind. First, this
study is contextualized within a small online class using a uniquely designed discussion
environment. Findings about learner participation should be generalized with great caution.
Second, this study relied on Social Network Analysis as the guiding methodology, which has its
own limits for researching online discussions. For example, while we used Yellowdig log data to
construct student networks, prior work found it useful to integrate content analysis to more
accurately capture the direction of learner interactions (Manca, Delfino, & Mazzoni, 2009).
Third, when analyzing post content we used the presence of a question mark as an approximate
indicator of a question. This analytical treatment could be enhanced by manually coding
discussion posts or constructing a machine learning model to more accurately identify questions
in student posts. Regardless of these limitations, we believe this study has accomplished its goal
of calling for attention to network prestige and temporality in asynchronous online discussions.
In the social constructivist paradigm, social participation is treated as an important
component of learning, motivating the use of asynchronous online discussions to support student
participation in blended and online classes. The study of network prestige in online discussions
build on prior work and contributes to the literature in several manners. First, building on prior
work on the social dimension of online discussions (e.g., Dawson, 2010; Vaquero & Cebrian,
2013), this study incorporates sociological constructs of prominence and prestige to characterize
the participation gap. The use of sociological concepts adds conceptual richness to the literature
and encourages future efforts to look into micro-level social patterns and “structural signatures”
in online discussions. For instance, while social network metrics such as degree centrality are
often used to describe a pattern or to predict learner performance, a closer look at social
interaction in group processes based on more sophisticated sociological terms (such as inertia,
reciprocity, and transitivity) can further explain the emergence of interaction patterns of interest
(Leenders, Contractor, & DeChurch, 2016). This study demonstrates the promise of deeply
integrating sociology and network science perspectives in the study of micro patterns in online
discussions.

PRESTIGE IN ONLINE DISCUSSIONS
18
Second, the multifaceted depiction of varied prestige has provided a rich picture of the
participation gap that merits further investigation. Earlier research has found social participation
in online forums predictive of learning performance (Jo et al., 2017; Joksimović, Gašević,
Kovanović, Riecke, & Hatala, 2015; Russo & Koesten, 2005) and low-performing students more
likely to connect with similarly low-performing peers (Dawson, 2010; Hu & Zhao, 2016). This
study provides detailed depiction of the participation gap. In short, lower-prestige students were
attempting to connect with higher-prestige peers but their attempts were not reciprocated.
Higher-prestige students were more connected, had more reciprocal and persistent connections,
closer to peers, and connected with similarly high-prestige peers; however, they did not occupy
more favorable positions in terms of bridging peer connections. Describing the participation gap
from different angles is critical for understanding the phenomenon and developing scaffolding
strategies for facilitating student participation.
Finally, the uncovered differences in temporal behavioral patterns between high- vs. low-
prestige groups merit special attention. Research of online discussions, and education research in
general, needs to treat the temporal dimension more seriously (Chen, Knight, & Wise, 2018). The
finding that high-prestige students generally posted earlier is in line with the “first-mover
advantage” in scholarly communication (Newman, 2009). The temporal patterns that high-
prestige students posted early and spaced their participation more evenly also reflect the
importance of self-regulation for discussion participation found in prior research (Shea et al.,
2013). Overall, the recognition of temporal behavioral patterns as potential contributors to varied
prestige contributes to the growing area of temporal analysis in educational research (Chen,
Knight, & Wise, 2018).
5.2 Practical Implications
Findings from this study have practical implications for online teaching. The study’s
multifaceted depiction of prestige—a sociological construct that captures the directionality and
reciprocity of student interaction—offers new angles for instructors and instructional designers to
support student participation. For example, current social network analytics tools often provide
network visualizations for the instructors to identify students in the peripheral for timing
intervention (Chen, Chang, Ouyang, & Zhou, 2018). However, network visualizations in these
tools cannot fully capture prestige differences uncovered in this study. In particular, students who
actively initiate unreciprocated contacts with others and hold less prestige will actually appear in
the center of a typical “force-directed” network visualization. This study urges instructional
attention paid to not only peripheral students who are less connected but also low-prestige
students whose attempts to connect are not responded favorably by peers.
The finding that high-prestige students tend to space their efforts more evenly implies the
possibility of offering additional scaffolds such as suggesting all students to check-in for a
couple of times each week. Being able to tactfully plan forum participation reflects strong self-
and co-regulation skills that are linked with more central positions in discussion networks (Shea
et al., 2013). Stronger supports are needed to help students develop such advanced skills
conducive to productive participation behaviors and dispositions.

PRESTIGE IN ONLINE DISCUSSIONS
19
In terms of technological design, temporal patterns identified in this study hint the
importance of temporal salience—ways in which discussion participation is temporally-oriented
(Hesse, Werner, & Altman, 1988)—that gets built in technological environments like Yellowdig.
Optimally, technological designs are embedded within themselves intelligence to support human
interactions and help distribute cognitions between human and computers (Pea, 1993). In this
study, we recognized that technological features of Yellowdig carried out important cognitive
acts for the class. For instance, Yellowdig’s sorting algorithm, which placed posts with the most
recent activities on the top of the “news feed,” might foster more dynamic interactions than
traditional threaded forums (which require users to track discussions by themselves). In the same
time, however, we conjecture that this sorting algorithm could have also contributed to varied
prestige. In particular, the sorting algorithm’s emphasis on recency can lead to the dying of
inactive posts (Hewitt, 2005), especially undermining Group B students in our case because their
posts were commented on less frequently and their participation was compressed together. Thus,
we call for stronger reflexivity of tool designers on the temporal aspect of user experiences and
future research to consider the hybridity of temporal features in a discussion environment and
learners’ discussion dispositions and behaviors (e.g., being inclined to comment on more recent
posts).
5.3 Future Directions
Much future work needs to be done in order to advance the analysis of network
prominence and prestige in asynchronous online discussions. First of all, the temporal aspect of
student contributions examined in this study needs further investigation to understand why
students demonstrated different timing behaviors and how a student chooses a particular post to
respond to in Yellowdig’s recency-based sorting algorithm. Even though the present study drew
from Social Network Analysis (a predominantly quantitative methodology), qualitative methods
would be extremely valuable for answering these questions. Second, findings from this study
also hint possible fruitfulness of studying students’ self- and co-regulation strategies in online
discussions. It could be especially fruitful to design formative feedback tools to help students
regulate their discourse participation (Chen, Chang, Ouyang, & Zhou, 2018). Finally, we also
question whether the traditional weekly-based course design is doing justice to discussion-centric
online learning and call for innovative course designs that can break temporal boundaries set by
the week-based academic calendar. If new cultures of learning emphasize social participation in
knowledge “flows” instead of the acquisition of knowledge as “entities,” structures imposed by
traditional institutions would better be challenged.

PRESTIGE IN ONLINE DISCUSSIONS
20
References
Barabasi, A. L., & Albert, R. (1999). Emergence of scaling in random networks. Science,
286(5439), 509–512.
Bergstrom, C. (2007). Eigenfactor: Measuring the value and prestige of scholarly journals.
College & Research Libraries News, 68(5), 314–316.
Bonacich, P. (1987). Power and centrality: A family of measures. The American Journal of
Sociology, 92(5), 1170–1182.
Borge, M., & White, B. (2016). Toward the development of socio-metacognitive expertise: An
approach to developing collaborative competence. Cognition and Instruction, 34(4), 323–
360. https://doi.org/10.1080/07370008.2016.1215722
Brown, R., Lynch, C., Eagle, M., Albert, J., Barnes, T., Baker, R. S., … McNamara, D. S. (2015).
Good Communities and Bad Communities: Does membership affect performance? In
EDM (pp. 612–613).
Burt, R. S. (2000). The network structure of social capital. Research in Organizational Behavior,
22, 345–423. https://doi.org/10.1016/S0191-3085(00)22009-1
Carey, S., & Bartlett, E. (1978). Acquiring a single new word. In Papers and Reports on Child
Language Development (Vol. 15, pp. 17–29). Retrieved from
https://eric.ed.gov/?id=ED198703
Chan, J. C. C., Hew, K. F., & Cheung, W. S. (2009). Asynchronous online discussion thread
development: examining growth patterns and peer-facilitation techniques: Influence of
peer facilitation. Journal of Computer Assisted Learning, 25(5), 438–452.
https://doi.org/10.1111/j.1365-2729.2009.00321.x
Chen, B., Chang, Y.-H., Ouyang, F., & Zhou, W. (2018). Fostering student engagement in online
discussion through social learning analytics. The Internet and Higher Education, 37, 21–
30. https://doi.org/10.1016/j.iheduc.2017.12.002
Chen, B., Knight, S., & Wise, A. F. (2018). Critical issues in designing and implementing
temporal analytics. Journal of Learning Analytics, 5(1), 1–9.
https://doi.org/10.18608/jla.2018.53.1
Chen, B., Resendes, M., Chai, C. S., & Hong, H.-Y. (2017). Two tales of time: Uncovering the
significance of sequential patterns among contribution types in knowledge-building
discourse. Interactive Learning Environments, 25(2), 162–175.
https://doi.org/10.1080/10494820.2016.1276081
Chi, M. T. H., & Menekse, M. (2015). Dialogue patterns in peer collaboration that promote
learning. In L. B. In Resnick, C. Asterhan, & S. Clarke (Eds.), Socializing intelligence
through academic talk and dialogue (pp. 263–274). Washington, DC: AERA.
Chin, C., & Osborne, J. (2010). Students’ questions and discursive interaction: Their impact on
argumentation during collaborative group discussions in science. Journal of Research in
Science Teaching, 47(7), 883–908. https://doi.org/10.1002/tea.20385
Cho, H., Stefanone, M., & Gay, G. (2002). Social information sharing in a CSCL community. In
Proceedings of the Conference on Computer Support for Collaborative Learning:

PRESTIGE IN ONLINE DISCUSSIONS
21
Foundations for a CSCL Community (pp. 43–50). Boulder, Colorado: International
Society of the Learning Sciences. Retrieved from
http://dl.acm.org/citation.cfm?id=1658616.1658623
Cox, D. R. (1966). The statistical analysis of series of events (1st ed.). Springer Netherlands.
Retrieved from https://www.springer.com/us/book/9789401178037
Daianu, M., Jahanshad, N., Villalon-Reina, J. E., Mendez, M. F., Bartzokis, G., Jimenez, E. E.,
… Thompson, P. M. (2014). Rich club network analysis shows distinct patterns of
disruption in frontotemporal dementia and Alzheimer’s disease. Mathematics and
Visualization, 2014, 13–22. https://doi.org/10.1007/978-3-319-11182-7_2
Dawson, S. (2010). “Seeing”the learning community: An exploration of the development of a
resource for monitoring online student networking. British Journal of Educational
Technology, 41(5), 736–752.
Dennen, V. P. (2005). From message posting to learning dialogues: Factors affecting learner
participation in asynchronous discussion. Distance Education, 26(1), 127–148.
https://doi.org/10.1080/01587910500081376
Dezuanni, M., O’Mara, J., & Beavis, C. (2015). “Redstone is like electricity”: Children’s
performative representations in and around Minecraft. E-Learning and Digital Media,
12(2), 147–163. https://doi.org/10.1177/2042753014568176
Dringus, L. P., & Ellis, T. (2010). Temporal transitions in participation flow in an asynchronous
discussion forum. Computers & Education, 54(2), 340–349.
https://doi.org/10.1016/j.compedu.2009.08.011
Ellison, N., Steinfield, C., & Lampe, C. (2007). The benefits of Facebook “friends”: Exploring
the relationship between college students’ use of online social networks and social capital.
Journal of Computer-Mediated Communication, 12(3).
Ergün, E., & Usluel, Y. K. (2016). An analysis of density and degree-centrality according to the
social networking structure formed in an online learning environment. Educational
Technology & Society, 19(4), 34–46.
Fields, A., Lai, K.-W., Gibbs, J., Kirk, A., & Vermunt, J. (2016). The transformation of an online
learning community from an organised facility to an organic fraternity. Distance
Education, 37(1), 60–72. https://doi.org/10.1080/01587919.2016.1158769
Greenhow, C., & Robelia, B. (2009). Old communication, new literacies: Social network sites as
social learning resources. Journal of Computer-Mediated Communication: JCMC, 14(4),
1130–1161. https://doi.org/10.1111/j.1083-6101.2009.01484.x
Guzdial, M., & Turns, J. (2000). Effective discussion through a computer-mediated anchored
forum. The Journal of the Learning Sciences.
https://doi.org/10.1207/s15327809jls0904_3
Haythornthwaite, C., Kumar, P., Gruzd, A., Gilbert, S., del Valle, M. E., & Paulin, D. (2018).
Learning in the wild: Coding for learning and practice on Reddit. Learning, Media and
Technology. https://doi.org/10.1080/17439884.2018.1498356

PRESTIGE IN ONLINE DISCUSSIONS
22
Hagel, J. I., Brown, J. S., & Davison, L. (2012). The Power of Pull: How Small Moves, Smartly
Made, Can Set Big Things in Motion. Basic Books.
Häkkinen, P. (2013). Multiphase method for analysing online discussions. Journal of Computer
Assisted Learning, 29(6), 547–555. https://doi.org/10.1111/jcal.12015
Hewitt, J. (2005). Toward an understanding of how threads die in asynchronous computer
conferences. Journal of the Learning Sciences, 14(4), 567–589.
Hesse, B. W., Werner, C. M., & Altman, I. (1988). Temporal aspects of computer-mediated
communication. Computers in Human Behavior, 4(2), 147–165.
https://doi.org/10.1016/0747-5632(88)90023-4
Hu, Y., & Zhao, F. (2016). A social network analysis of online collaborative learning aspects in
an online course. In 2016 International Symposium on Educational Technology (ISET)
(pp. 3–7). Beijing, China: IEEE.
Ito, M., Gutiérrez, K., Livingstone, S., Penuel, B., Rhodes, J., Salen, K., … Watkins, S. C. (2013).
Connected learning: An agenda for research and design (p. 99). Irvine, CA, USA: Digital
Media and Learning Research Hub.
Jenkins, H. (2006). Confronting the challenges of participatory culture: Media education for the
21st century. MIT Press.
Jo, I., Park, Y., & Lee, H. (2017). Three interaction patterns on asynchronous online discussion
behaviours: A methodological comparison. Journal of Computer Assisted Learning, 33(2),
106–122. https://doi.org/10.1111/jcal.12168
Joksimović, S., Gašević, D., Kovanović, V., Riecke, B. E., & Hatala, M. (2015). Social presence
in online discussions as a process predictor of academic performance. Journal of
Computer Assisted Learning, 31(6), 638–654. https://doi.org/10.1111/jcal.12107
Knoke, D., & Burt, R. S. (1983). Prominence. In R. S. Burt & M. J. Minor (Eds.), Applied
network analysis: A methodological introduction (pp. 195–222). Beverly Hills, CA: Sage
Publications.
Leenders, R. T. A. J., Contractor, N. S., & DeChurch, L. A. (2016). Once upon a time.
Organizational Psychology Review, 6(1), 92–115.
https://doi.org/10.1177/2041386615578312
Lewis, S., Pea, R., & Rosen, J. (2010). Beyond participation to co-creation of meaning: Mobile
social media in generative learning communities. Social Sciences Information.
Information Sur Les Sciences Sociales, 49(3), 351–369.
https://doi.org/10.1177/0539018410370726
Lu, Y., Sharon, I., & Hsiao, H. (2016). Seeking programming-related information from large
scaled discussion forums, help or harm? In T. Barnes, M. Chi, & M. Feng (Eds.),
Proceedings of the 9th International Conference on Educational Data Mining (pp. 442–
447). Raleigh, North Carolina, USA: International Educational Data Mining Society.
Retrieved from
http://www.educationaldatamining.org/EDM2016/proceedings/paper_151.pdf

PRESTIGE IN ONLINE DISCUSSIONS
23
Mazzolini, M., & Maddison, S. (2007). When to jump in: The role of the instructor in online
discussion forums. Computers & Education, 49(2), 193–213.
Manca, S., Delfino, M., & Mazzoni, E. (2009). Coding procedures to analyse interaction patterns
in educational web forums: Analysing interaction in web forums. Journal of Computer
Assisted Learning, 25(2), 189–200. https://doi.org/10.1111/j.1365-2729.2008.00296.x
Masuda, N., & Konno, N. (2006). VIP-club phenomenon: Emergence of elites and masterminds
in social networks. Social Networks, 28(4), 297–309.
McAteer, E., Tolmie, A., Duffy, C., & Corbett, J. (1997). Computer-mediated communication as
a learning resource. Journal of Computer Assisted Learning, 13(4), 228–235.
https://doi.org/10.1046/j.1365-2729.1997.00024.x
Nandi, D., Hamilton, M., & Harland, J. (2012). Evaluating the quality of interaction in
asynchronous discussion forums in fully online courses. Distance Education, 33(1), 5–30.
https://doi.org/10.1080/01587919.2012.667957
Newman, M. E. J. (2009). The first-mover advantage in scientific publication. Europhysics
Letters. Retrieved from http://iopscience.iop.org/article/10.1209/0295-
5075/86/68001/meta
Ouyang, F., & Scharber, C. (2017). The influences of an experienced instructor’s discussion
design and facilitation on an online learning community development: A social network
analysis study. The Internet and Higher Education, 35, 34–47. Retrieved from
http://www.sciencedirect.com/science/article/pii/S1096751616302585
Perc, M. (2014). The Matthew effect in empirical data. Journal of the Royal Society, Interface /
the Royal Society, 11 (98), 20140378. https://doi.org/10.1098/rsif.2014.0378
Pea, R. D. (1993). Practices of distributed intelligence and designs for education. In G. Salomon
(Ed.), Distributed cognitions: Psychological and educational considerations (pp. 47–87).
Cambridge, UK: Cambridge University Press.
Resnick, L. B., Säljö, R., Pontecorvo, C., & Burge, B. (Eds.). (1997). Discourse, tools and
reasoning: Essays on situated cognition. Springer Berlin Heidelberg.
Rosé, C. P., & Ferschke, O. (2016). Technology support for discussion based learning: From
computer supported collaborative learning to the future of massive open online courses.
International Journal of Artificial Intelligence in Education, 26(2), 660–678.
https://doi.org/10.1007/s40593-016-0107-y
Russo, T. C., & Koesten, J. (2005). Prestige, centrality, and learning: A social network analysis of
an online class. Communication Education, 54(3), 254–261.
https://doi.org/10.1080/03634520500356394
Scardamalia, M., & Bereiter, C. (2003). Knowledge building environments: Extending the limits
of the possible in education and knowledge work. In A. DiStefano, K. E. Rudestam, & R.
Silverman (Eds.), Encyclopedia of distributed learning (pp. 269–272). Thousand Oaks,
CA: Sage Publications.

PRESTIGE IN ONLINE DISCUSSIONS
24
Scardamalia, M., Bereiter, C., & Lamon, M. (1994). The CSILE project: Trying to bring the
classroom into World 3. In K. McGilly (Ed.), Classroom lessons: Integrating cognitive
theory and classroom practice (pp. 201–228). Cambridge, MA: The MIT Press.
Scott, J. (2012). Social Network Analysis (3rd ed.). Thousand Oaks, CA: SAGE Publications.
https://doi.org/10.5040/9781849668187
Sfard, A. (1998). On two metaphors for learning and the dangers of choosing just one.
Educational Researcher, 27(2), 4–13.
Shea, P., Hayes, S., Smith, S. U., Vickers, J., Bidjerano, T., Gozza-Cohen, M., … Tseng, C.-H.
(2013). Online learner self-regulation: Learning presence viewed through quantitative
content- and social network analysis. The International Review of Research in Open and
Distributed Learning, 14(3), 427–461.
Stahl, G. (2006). Supporting group cognition in an online math community: A cognitive tool for
small-group referencing in text chat. Journal of Educational Computing Research, 35(2),
103–122. https://doi.org/10.2190/q435-7611-2561-720p
So, H.-J. (2009). When groups decide to use asynchronous online discussions: Collaborative
learning and social presence under a voluntary participation structure. Journal of
Computer Assisted Learning, 25(2), 143–160. https://doi.org/10.1111/j.1365-
2729.2008.00293.x
Thomas, S. L. (2000). Ties that bind: A social network approach to understanding student
integration and persistence. The Journal of Higher Education, 71(5), 591–615.
Torfason, M. T., & Kitts, J. A. (2011). Prominence. In G. A. Barnett (Ed.), Encyclopedia of social
networks (pp. 720–721). Thousand Oaks, CA: SAGE Publications, Inc.
van den Heuvel, M. P., & Sporns, O. (2011). Rich-club organization of the human connectome.
The Journal of Neuroscience: The Official Journal of the Society for Neuroscience,
31(44), 15775–15786. https://doi.org/10.1523/JNEUROSCI.3539-11.2011
Vaquero, L. M., & Cebrian, M. (2013). The rich club phenomenon in the classroom. Scientific
Reports, 3, 1174.
Wang, C.-H. (2005). Questioning skills facilitate online synchronous discussions. Journal of
Computer Assisted Learning, 21(4), 303–313. https://doi.org/10.1111/j.1365-
2729.2005.00138.x
Warren, K. J., & Rada, R. (1998). Sustaining computer-mediated communication in university
courses. Journal of Computer Assisted Learning, 14(1), 71–80.
https://doi.org/10.1046/j.1365-2729.1998.1410071.x
Wegener, B. (1992). Concepts and measurement of prestige. Annual Review of Sociology, 18(1),
253–280. https://doi.org/10.1146/annurev.so.18.080192.001345
Wise, A. F., & Chiu, M. M. (2011). Analyzing temporal patterns of knowledge construction in a
role-based online discussion. International Journal of Computer-Supported Collaborative
Learning, 6(3), 445–470. https://doi.org/10.1007/s11412-011-9120-1

PRESTIGE IN ONLINE DISCUSSIONS
25
Wise, A. F., & Cui, Y. (2018). Learning communities in the crowd: Characteristics of content
related interactions and social relationships in MOOC discussion forums. Computers &
Education, 122, 221–242. https://doi.org/10.1016/j.compedu.2018.03.021
Wise, A. F., Hausknecht, S. N., & Zhao, Y. (2014). Attending to others’ posts in asynchronous
discussions: Learners’ online “listening” and its relationship to speaking. International
Journal of Computer-Supported Collaborative Learning, 9(2), 185–209.
https://doi.org/10.1007/s11412-014-9192-9
Wise, A. F., Zhao, Y., Hausknecht, S., & Chiu, M. M. (2014). Temporal considerations in
analyzing and designing online discussions in education: Examining duration, sequence,
pace, and salience. In Assessment and evaluation of time factors in online teaching and
learning (pp. 198–231). IGI Global. https://doi.org/10.4018/978-1-4666-4651-3.ch008
Zhou, S., & Mondragon, R. J. (2004). The rich-club phenomenon in the Internet topology. IEEE
Communications Letters, 8(3), 180–182.
Zingaro, D., & Oztok, M. (2012). Interaction in an asynchronous online course: A synthesis of
quantitative predictors. Journal of Asynchronous Learning Networks, 16(4), 71–82.
Retrieved from https://eric.ed.gov/?id=EJ982683