What is the source of social capital? The association between social network position and social presence in communities of inquiry

What is the source of social capital?The association between social network position

and social presence in communities of inquiry

Vitomir Kovanovic1 Srecko Joksimovic1

vitomir [email protected] [email protected]

Dragan Gasevic2 Marek Hatala1

[email protected] [email protected]

1School of Interactive Arts and Technology 2School of Computing ScienceSimon Fraser University Athabasca University

Burnaby, Canada Edmonton, Canada

July 4, 2014,London, UK

IntroductionBackground

MethodsResults

Discussion and ConclusionsConclusionsReferences

Objective

Main question

Why different students have different socialcapital and social network positions?

Can we use some of the online learning theories to providemore insights into the underlying factors that contribute tothe observed differences?

V. Kovanovic et al. What is the source of social capital? 1 / 23


MethodsResults


Objective

Main question

Why different students have different socialcapital and social network positions?

Can we use some of the online learning theories to providemore insights into the underlying factors that contribute tothe observed differences?



MethodsResults


Final goal

For instructors

Provide instructors with information on student’s learning progresswithin a learning community so that appropriate instructionalinterventions can be planned and implemented.

For students

Provide learners with the real time feedback of their own progress, andprogress of their peers so that they can self-regulate their learningactivities more successfully.

For researchers

Use data to better operationalize current Community of Inquiry modelof online learning.



MethodsResults


Asynchronous online discussionsSocial network analysisCommunity of Inquiry (CoI) modelProposed approach

Asynchronous online discussions -“gold mine of information” Henri [9]

• Frequently used in both blended andfully online learning [11],

• Their use produced large amount ofdata about learning processes [4],

• Particularly important insocial-constructivist pedagogies [1].

• Frequently used for constructingstudents’ social networks.



MethodsResults



Social network analysis

• Social capital: value resulting from occupying a particularlyadvantageous position within a social network [2]

• Many studies indicated importance of students’ social capital onmany important aspects of learning and educational experience:

• Academic performance,• Retention,• Persistance,• Program satisfaction,• Sense of community,• . . .



MethodsResults




However,

• Typically isolated studies focusing on a single aspect of particularinterest,

• Typically not explaining what might be the cause of observeddifferences in network positions,

• Lack of well-established learning theories which explicitly addresssocial network position.

Can we leverage existing comprehensive models of online learning toprovide insight into the nature of the observed differences in socialnetworks?



MethodsResults




However,

• Typically isolated studies focusing on a single aspect of particularinterest,

• Typically not explaining what might be the cause of observeddifferences in network positions,

• Lack of well-established learning theories which explicitly addresssocial network position.

Can we leverage existing comprehensive models of online learning toprovide insight into the nature of the observed differences in socialnetworks?



MethodsResults



Community of Inquiry (CoI) model

Conceptual model outlying the important constructs that defineworthwhile educational experience in online education setting.

• Social presence: relationships andsocial climate in a community.

• Cognitive presence: phases ofcognitive engagement and knowledgeconstruction.

• Teaching presence: instructionalrole during social learning.

CoI model is:

• Extensively researched and validated,

• Adopts content analysis forassessment of presences.



MethodsResults



Social presence

Social presence

“Ability of participants in a community of inquiry to project themselvessocially and emotionally, as “real” people (i.e., their full personality),through the medium of communication being used.” [7, p. 89]

Three different dimensions of communication:

1 Affectivity and expression of emotions: defined as “the abilityand confidence to express feelings related to the educationalexperience.” [7, p. 99]

2 Interactivity and open communication: defined as ““reciprocaland respectful exchanges of messages” [7, p. 100].

3 Cohesiveness: Activities that “build and sustain a sense of groupcommitment” [7, p. 101]



MethodsResults



Social presence

Social presence

“Ability of participants in a community of inquiry to project themselvessocially and emotionally, as “real” people (i.e., their full personality),through the medium of communication being used.” [7, p. 89]

Three different dimensions of communication:

1 Affectivity and expression of emotions: defined as “the abilityand confidence to express feelings related to the educationalexperience.” [7, p. 99]

2 Interactivity and open communication: defined as ““reciprocaland respectful exchanges of messages” [7, p. 100].

3 Cohesiveness: Activities that “build and sustain a sense of groupcommitment” [7, p. 101]



MethodsResults



Social presence coding scheme

• Content analysis scheme for analysis of discussion messages,• Use of whole message as unit of analysis,• Look for particular indicators of different sociocognitive processes,

Social presence categories and indicators as defined by Rourke et al. [12].

Category Code Indicator

Affective A1 Expression of emotionsA2 Use of humorA3 Self-disclosure

Interactive or OpenCommunication

I1 Continuing a threadI2 Quoting from others’ messagesI3 Referring explicitly to others’ messagesI4 Asking questionsI5 Complementing, expressing appreciationI6 Expressing agreement

Cohesive C1 VocativesC2 Addresses or refers to the group using inclusive pronounsC3 Phatics, salutations



MethodsResults



Proposed approach

General idea

Investigate the relationship between students’ social capitaland social climate in the course.

More specificially,

We looked at the relationship between social networkcentrality measures and social presence, one of the threemain components of Community of Inquiry model of online

learning.



MethodsResults



Proposed approach

General idea

Investigate the relationship between students’ social capitaland social climate in the course.

More specificially,

We looked at the relationship between social networkcentrality measures and social presence, one of the threemain components of Community of Inquiry model of online

learning.



MethodsResults



Proposed approach

Measure three dimensions of social presence for each student and seehow they relate to their network centrality measures.

• Are three dimensions of social presence statistically significantpredictors of network centrality measures?

• What is the relative importance of different dimensions of socialpresence?



MethodsResults



Proposed approach

Measure three dimensions of social presence for each student and seehow they relate to their network centrality measures.

• Are three dimensions of social presence statistically significantpredictors of network centrality measures?

• What is the relative importance of different dimensions of socialpresence?



MethodsResults


Data setSNA centrality measuresMessage codingStatistical Analysis

Data set

• Six offerings of graduate level course in software engineering atdistance learning university.

• Total of 1747 messages by 81 students.

Course offering statistics.

Student count Message count Graph density

Winter 2008 15 212 0.52Fall 2008 22 633 0.69Summer 2009 10 243 0.84Fall 2009 7 63 0.58Winter 2010 14 359 0.84Winter 2011 13 237 0.77

Average 13 291 0.71Total 81 1747



MethodsResults



Social network centrality measures

• Directed social network graph based on post-reply activity.

• Extracted popular centrality measures.

Descriptive statistics of social network metrics.

Mean SD Min Max

Betweenness 9.04 14.51 0.00 74.20In-degree 19.84 8.62 4.00 42.00Out-degree 19.86 9.37 3.00 44.00In-closeness 0.09 0.04 0.04 0.17Out-closeness 0.08 0.04 0.03 0.18



MethodsResults



Social presence message coding

• Each message manually coded by two coders on the levels ofindicators (Percentage agreement = [84.1 − 98.9]%).

Social presence indicators.

Category Code Indicator Count Percentage Percent Agreement

Affective A1 Expression of emotions 288 16.5% 84.4A2 Use of humor 44 2.5% 93.1A3 Self-disclosure 322 18.4% 84.1

Interactive I1 Continuing a thread 1664 95.2% 98.9I2 Quoting from others messages 65 3.7% 95.4I3 Referring explicitly to other’s messages 91 5.2% 92.7I4 Asking questions 800 45.8% 89.4I5 Complementing, expressing appreciation 1391 79.6% 90.7I6 Expressing agreement 243 13.9% 96.6

Cohesive C1 Vocatives 1433 82.0% 91.8C2 Addresses or refers to the group using

inclusive pronouns144 8.2% 88.8

C3 Phatics, salutations 1281 73.3% 96.1



MethodsResults



Social presence message coding

• Some indicators way too frequent.• Limiting discriminatory power of the whole category.

• We removed indicators occurring in more than 75% of the messages.

Social presence categories.

Category Count Percentage Percent Agreement

Affective 530 30.3% 80.8Interactive (Excluded I1 and I5) 1030 59.0% 86.2Cohesive (Excluded C1) 1326 75.9% 93.4



MethodsResults



Statistical Analysis

Multiple regression analysis:

• DV: Social network centrality metrics.

• IVs: CoI Social presence codes.

• Backward stepwise model selection using AIC criterion [8].

• Holm-Bonferroni correction [10]:• Guaranteed to keep family-wise error rate (FWER) α at the desired

level (i.e., α = 0.05).• Significantly more powerful than classical Bonferroni correction [5].



MethodsResults



Holm-Bonferroni correction procedure

For family of N tests and desired α significance:

• Sort all N observed p-values from smallest to largest.• Cutoff for the smallest p-value: α/N.• Cutoff for next p-value: α/(N− 1).• . . .• Cutoff for largest p-value: α.

Important rule

If any of the tests gets rejected, all the subsequent tests are also rejectedautomatically.

Current study

In our study with 5 tests, cutoff p-values areα = [0.01, 0.0125, 0.0167, 0.0250, 0.05]



MethodsResults






Important rule


Current study




MethodsResults






Important rule


Current study




MethodsResults


Results

Regression results for selected centrality measures after stepwise model selectionusing AIC criterion.

Betweenness In-degree Out-degree In-closeness Out-closeness

β SE p β SE p β SE p β SE p β SE p

Affective 0.27 0.12 0.024 0.18 0.054 0.001 0.23 0.059 <0.001Interactive 0.38 0.12 0.002 0.65 0.064 <0.001 0.65 0.07 <0.001 0.27 0.11 0.015 0.37 0.15 0.017Cohesive 0.2 0.061 0.001 0.14 0.066 0.041 -0.23 0.15 0.137

F (3, 77) 19.6 <0.001 159 <0.001 130 <0.001 6.24 0.015 3.03 0.054Adjusted R2 0.32 0.86 0.83 0.061 0.048



MethodsResults


Main findings

• All but one regression models were significant, one marginallysignificant.

• Interactive dimension of social presence is the most stronglyassociated with all of the network centrality measures.

• Probable reason is the nature of social networks as a medium forfostering collaborative and productive learning.

• According to Garrison [6], interactive social presence is dominant inthe beginning until students develop trust and sense of community,but it decreases over time, while affective and cohesive increase overtime.

• Practical implication: provide opportunities for focused, on taskinteractions that foster open communication and collaboration.



MethodsResults


Main findings

• All but one regression models were significant, one marginallysignificant.

• Interactive dimension of social presence is the most stronglyassociated with all of the network centrality measures.

• Probable reason is the nature of social networks as a medium forfostering collaborative and productive learning.

• According to Garrison [6], interactive social presence is dominant inthe beginning until students develop trust and sense of community,but it decreases over time, while affective and cohesive increase overtime.

• Practical implication: provide opportunities for focused, on taskinteractions that foster open communication and collaboration.



MethodsResults


Main findings: Degree centrality

• All three categories of social presence were significantly predictive ofIn-Degree and Out-Degree centrality measures.

• Affective and Cohesive are very interesting as they are not directlyaffecting degree centrality.

• Interactive category was most strongly associated with degreenetwork centrality.

• This is expected for In-Degree as activities such as askingquestions, addressing by name or quoting someone’s messageincrease chances of ’provoking’ a response.

• For Out-Degree it more interesting.



MethodsResults


Main findings: Betweenness centrality

• 32% of variability in betweenness centrality scores explained by ourregression model. Effect size: Cohen’s f2 = 0.47 which is consideredto be a large effect size [3].

• Interactive and affective dimensions of social presence weresignificantly predictive of betweenness centrality, with interactivedimension being more strongly associated.

• Probably due to the nature of social networks and the focus oninformation exchange. Also trust and sense of community developslater in the course when student already developed opencommunication.

• As a followup, we want to look at the individual indicators, as theymight contain some answers to our findings.



MethodsResults


Main findings: Closeness centrality

• Only interactive category was significantly predictive ofIn-closeness centrality. Model for Out-closeness was verymarginally significant (p = 0.054).

• Probable reason might be the fact that closeness embeds theinteractive relationships, for which affectivity and cohesiveness arenot much important.



MethodsResults


Limitations and future work

Limitations:

• Data from one course• despite having data from several offerings of the course, there might

be an effect of the particular pedagogical approach.

• Not all student interactions have positive effect and increase socialcapital,

• Other important factors beside social presence.

Future work:

• Replicate on new data set, with larger and more diverse subjects,

• Investigate changes in the distributions of three social presencedimensions over time,

• Maybe look at the levels of indicators instead of categories.



MethodsResults


Summary

• There is an interesting connection between social presence andstudents’ social network positions.

• We can use three dimensions of social presence to predict differentnetwork centrality metrics.

• Interactivity and open communication showed to be the mostsignificant component of social presence.

• Our findings indicate the need for providing student withopportunities for the development of social capital throughcollaboration with other students on focused tasks.

• Educational theories suggest that development of trust and sense ofcommunity follows from on-task interactions. Our data shows somepreliminary support for this.


Thank you

References I

Terry Anderson and Jon Dron. “Three generations of distance education pedagogy”. In:

The International Review of Research in Open and Distance Learning 12.3 (2010),pp. 80–97.

Ronald S. Burt. “Structural Holes versus Network Closure as Social Capital”. In: Social

Capital: Theory and Research. 2001.

Jacob Cohen. “The Analysis of Variance”. In: Statistical power analysis for the behavioral

sciences. 1988, pp. 273–406.

Roisin Donnelly and John Gardner. “Content analysis of computer conferencing

transcripts”. In: Interactive Learning Environments 19.4 (2011), pp. 303–315.

Olive Jean Dunn. “Multiple Comparisons among Means”. In: Journal of the American

Statistical Association 56.293 (1961), pp. 52–64.

D. Randy Garrison. E-Learning in the 21st Century: A Framework for Research andPractice. 2 edition. Routledge, 2011.

D. Randy Garrison, Terry Anderson, and Walter Archer. “Critical Inquiry in a Text-Based

Environment: Computer Conferencing in Higher Education”. In: The Internet and HigherEducation 2.2–3 (1999), pp. 87–105.

References II

Trevor J Hastie, Robert J Tibshirani, and Jerome H Friedman. The elements of statisticallearning: data mining, inference, and prediction. Springer, 2013.

France Henri. “Computer Conferencing and Content Analysis”. In: Collaborative Learning

Through Computer Conferencing. 1992, pp. 117–136.

Sture Holm. “A Simple Sequentially Rejective Multiple Test Procedure”. In: Scandinavian

Journal of Statistics 6.2 (1979), pp. 65–70.

Rocci Luppicini. “Review of computer mediated communication research for education”.

In: Instructional Science 35.2 (2007), pp. 141–185.

Liam Rourke et al. “Assessing Social Presence In Asynchronous Text-based Computer

Conferencing”. In: The Journal of Distance Education 14.2 (1999), pp. 50–71.

What is the source of social capital? The association between social network position and social presence in communities of inquiry

Education

social learning

social presence

source of social capital

social network positions

social climate

dierent social capital

students social networks

learning community