As the field of learning analytics continues to mature, there is a corresponding evolution and sophistication of the associated analytical methods and techniques. In this regard social network analysis (SNA) has emerged as one of the cornerstones of learning analytics methodologies. However, despite the noted importance of social networks for facilitating the learning process, it remains unclear how and to what extent such network measures are associated with specific learning outcomes. Motivated by Simmel’s theory of social interactions and building on the argument that social centrality does not always imply benefits, this study aimed to further contribute to the understanding of the association between students’ social centrality and their academic performance. The study reveals that learning analytics research drawing on SNA should incorporate both – descriptive and statistical methods to provide a more comprehensive and holistic understanding of a students’ network position. In so doing researchers can undertake more nuanced and contextually salient inferences about learning in network settings. Specifically, we show how differences in the factors framing students’ interactions within two instances of a MOOC affect the association between the three social network centrality measures (i.e., degree, closeness, and betweenness) and the final course outcome.

1. Ties that matter
Effects of the network context on the association between
social centrality and academic performance
17 December 2015
PhD SeminarSrećko Joksimović, DraganGašević
s.joksimovic@ed.ac.uk
@s_joksimovic
www.de.ed.ac.uk/people/srecko-joksimovic

2. Social network
analysis
Slide 2 out of 18
 Structural environment as opportunity or constraint
 Structure (e.g., social, economic, political) as lasting
patterns of relations among actors
 Actions are viewed as interdependent
 Ties as channels for flow of resources

3. Centrality
measures
Slide 3 out of 18
 Eigenvalue centrality
 Betweenness
centrality
 Degree centrality
 Closeness centrality

4. Strength of
ties
Slide 4 out of 18http://www.informationweek.com/why-your-weak-relationships-pack-strength/d/d-id/1107476?
Connections through
strong ties
Connections through
weak ties
“The argument asserts that
our acquaintances (weak ties) are less likely
to be socially involved with one another
than are our close friends (strong ties)” (Granovetter, 1983, p.1).

5. Structural
holes
Slide 5 out of 18http://rzhengac.github.io/Comp4641Main_tutorial.html
Structural
hole
Node A’s position implies
structural advantage relative to
node D.

6. SNA in
educational
research
 Structural centrality measures as predictors of:
 Cognitive learning outcomes
 Final grade
 Higher sense of belonging to a group
 Course satisfaction
 Comprehension of learning materials
 etc.
Slide 6 out of 18

7. Motive
Slide 7 out of 18
Russo and Koesten (2005)
prestige (in-degree) Cognitive learning
outcomecentrality (out-degree)
degree centrality
Course
grade
Cho et al. (2007)
closeness centrality
betweenness centrality
Jiang et al. (2014)
degree centrality
GPAcloseness centrality
betweenness centrality
eccentrality
Gašević et al. (2013)
degree centrality
Course
grade
closeness centrality
betweenness centrality
degree centrality
Course
grade
closeness centrality
betweenness centrality
Positive, statistically
significant association
Note:
No statistically
significant association

8. Theoretical
approach
Slide 8 out of 18https://cvcedhlab.hypotheses.org/author/mduering
 Centrality does (not) necessarily imply less constraints
and more benefit (Krachardt, 1999)
 Importance of contextual factors
 Triads as the fundamental unit of analysis
 Simmel’s theory of social
interactions
 No inherent motivation to form a
clique

9. Study
objective
Slide 9 out of 18
Network structural
properties Learning outcome
Social dynamical
processes?
Research questions:
1. Differences in the underlying processes that determine
network formation?
2. Propensity for forming Simmelian ties?
3. The impact at the association between social centrality
and academic performance?
Tie dynamics:
• Homophily/heterophily
• Reciprocity
• Triadic closure
• etc.

10. Method
(Data)
 Platform: Coursera
 Courses: CodeYourself! (English), ¡A Programar!
(Spanish)
 Certificate: 50% for the coursework; 75% - distinction
Slide 10 out of 18
59,531
26,568
1,430
25,255
13,808
1,818
0
10000
20000
30000
40000
50000
60000
70000
Enrolled Engaged Engaged with
forum
Course participants
Codeyourself Aprogramar
0
500
1000
1500
2000
Codeyourself Aprogramar
Obtained certificate
Normal Disctinction

11. Method
(Analysis)
Slide 11 out of 18
Discussion forum
extract
Weighted,
directed graph
SNA
Descriptive
network analysis
Statistical
network analysis
 Centrality measures
 Exponential random graph models
 Homophily
 Achievement
 Domestic/Guest
 Gender
 Reciprocity
 Popularity
 Expansiveness
 Simmelian cliquesCourse outcome
 Obtained certificate
 Normal
 Distinction
 None
Multinomial logistic
regression Association?
Interpretation

12. Results
(Network
characteristics)
Slide 12 out of 18
-8 -6 -4 -2 0 2 4 6
Expansiveness
Popularity
Simmelian
Reciprocity
Sel. Mixing (Gender)
Sel. Mixing (Domestic)
Achievement (Normal)
Achievement (None)
Achievement (Distinct)
Edges
Aprogramar Codeyourself
Analysis of the estimates for the two ERG models
***
***
***
***
***
**
***
**
***
***
***
***
***
***
Note: * p<.05; ** p<.01; *** p<.001

13. Results
(centrality vs.
performance)
Slide 13 out of 18
-0.12 -0.1 -0.08 -0.06 -0.04 -0.02 0 0.02 0.04 0.06 0.08
Betweenness (normal)
Betweenness (distinct)
Closeness (normal)
Closeness (distinct)
W. Degree (normal)
W. Degree (distinct)
Aprgoramar Codeyourself
Results of the multinomial regression analysis
Note: * p<.05; ** p<.01; *** p<.001
In order to provide meaningful visualizations, estimates for betweenness centrality were
multiplied by 100 (only for the presentation purposes)
***
**
***
*
**
***
***

14. Conclusions
 Observed networks differ with respect to the
determinants of network formation.
 These discrepancies DO affect the association between
social centrality and academic performance.
 Social centrality within the network characterized with
“super-strong” ties, DOES NOT necessarily imply
benefits.
Slide 14 out of 18

15. Implications &
Further
Research
Implications:
 “Traditional” (descriptive) SNA + statistical network
analysis.
 Account for contextual determinants.
Further Research:
 Examine temporal dynamics?
 SNA + content analysis?
 Language vs. social dynamics?
Slide 15 out of 18

16. References
S. Joksimović,A. Manataki, D. Gašević, S. Dawson,V. Kovanović, and I. F. de Kereki:
“Translating network position into performance: Importance of centrality in different
network configurations”, In Proceedings of the Sixth InternationalConference on Learning
Analytics and Knowledge (LAK 2016), (submitted);
B.V. Carolan, Social Network Analysis Education:Theory, Methods & Applications.Social Network
Analysis Education:Theory, Methods & Applications.SAGE Publications, Inc. SAGE Publications,
Inc., 2014.
S. Goodreau, J. Kitts, and M. Morris, “Birds of a Feather, or Friend of a Friend? Using Exponential
Random Graph Models to InvestigateAdolescent Social Networks*,” Demography, vol. 46, no. 1,
pp. 103–125, 2009.
L. C. Freeman, “Centrality in social networks conceptual clarification,” Soc. Netw., vol. 1, no. 3, pp.
215–239, 1979.
S.Wasserman, Social network analysis: Methods and applications, vol. 8. Cambridge university
press, 1994.
R. S. Burt, STRUCTURAL HOLES. Harvard University Press, 1995.
M. S. Granovetter, “The strength of weak ties,” Am.J. Sociol., pp. 1360–1380, 1973.
D. Krackhardt, “TheTies thatTorture: SimmelianTie Analysis in Organizations,” Res. Sociol.
Organ., vol. 16, pp. 183–210, 1999.
D. Krackhardt, “Super Strong and Sticky,” Power Influ. Organ., p. 21, 1998.
Slide 16 out of 18

17. References
Granovetter, Mark. "The strength of weak ties: A network theory revisited.“ Sociological theory 1.1
pp. 201-233, 1983.
T. C. Russo and J. Koesten, “Prestige, centrality, and learning: A social network analysis of an
online class,” Commun. Educ., vol. 54, no. 3, pp. 254–261, 2005.
H. Cho, G. Gay, B. Davidson, andA. Ingraffea, “Social networks, communication styles, and
learning performance in a CSCL community,” Comput. Educ., vol. 49, no. 2, pp. 309–329, Sep.
2007.
D. Gašević, A. Zouaq, and R. Janzen, “‘ChooseYour Classmates,Your GPA Is at Stake!’:The
Association of Cross-Class SocialTies andAcademic Performance,” Am. Behav. Sci., 2013
S. Jiang, S. M. Fitzhugh, and M.Warschauer, “Social Positioning and Performance in MOOCs,” in
Proceedings of theWorkshops held at Educational Data Mining 2014, co-located with 7th
InternationalConference on Educational Data Mining (EDM 2014), London, United Kingdom, 2014,
vol. 1183, p. 14.
Slide 17 out of 18

18. Ties that matter
Effects of the network context on the association between
social centrality and academic performance
17 December 2015
PhD Seminar
Srećko Joksimović, DraganGašević
s.joksimovic@ed.ac.uk
@s_joksimovic
Q&A
www.de.ed.ac.uk/people/srecko-joksimovic