Determinants of Engagement in an Online Community of Inquiry

1
Determinants of Engagement in an Online Community
of Inquiry 

• Jim Waters
• College of Information Science and Technology
• Drexel University
• Philadelphia
• james.waters_at_drexel.edu

2
Background

• Problem of maintaining student engagement.
• Online learning creates separation
• Alienation, lack of commitment and antisocial
  behavior ?
• Community of Inquiry ?

3

• Pragmatism Dewey and Addams
• Problematic situation, scientific attitude and
  community as participatory democracy
• Inquiry is controlled or directed transformation
  of an indeterminate situation
• There is a community engaged in inquiry. Inquiry
  is an open-ended process with positive feedback.
• Dewey (1916,1933)

4
Community of Inquiry
Garrison et al 2000
5
Cycles of Inquiry
Garrison et al 2000
6
Building on the Garrison et al Model

• Content Analysis of online Discussion Board
• Graduate Information Systems Students
• Open-ended debate
• Practical and Theoretical questions
• Derived behaviors that incorporated different
  elements of the Garrison model

7
Student Roles
Waters and Gasson 2005
8
Research Questions

• Are there noticeable patterns of interactions
  between participant roles?
• Do patterns of interaction change over time?
• Does the online learning environment support
  critical inquiry ?
• What interactions generate greatest student
  engagement

9
Study

• Post-Hoc analysis of online learning archive
• 10 week graduate IS Management course at a US
  university
• 23 students, experienced professionals
  managers.
• 3 - 4 open-ended questions posted to discussion
  board weekly
• 1063 discussion-board messages
• 951 student responses (analyzed)
• 112 instructor postings (not analyzed).
• Content analysis of postings and responses
• Each student contribution message assigned to
  single response type, reflecting dominant mode of
  behavior.

10
Raw results

• 25,937 individual reads of discussion board
  message (range 331 2179 reads per student)
• 951 student postings (range 1 154 per student)
• Most active period weeks 1 2 (157 posts and
  162 posts)
• Then steady pattern of 70-80 posts per week.

11
Student behavior

• Contributor (61)
• Facilitator (22)
• Fluid patterns of class behavior
• Students adopt different behaviors from week to
  week
• Popularity and volume were unrelated
• Possible connection between facilitation and
  popularity/reference to poster.

12
Detailed Analysis

• Nine typical threads analysed
• Three threads each for weeks 3, 6 and 9
• The most productive debate produced 30 messages
  with a maximum thread depth of 7.
• The least productive produced 14 messages with a
  thread depth of 2.
• The mean number of messages on a discussion was
  22
• Four discussions had a thread depth of greater
  than 3.
• Pattern of responses analysed

13
Are there noticeable patterns of interactions
between participant roles?
14
Ratio of receive to send Contributor 27/106
0.25 Facilitator 25/38 0.65 Complicator
0/17 0.00
15
Do patterns of interaction change over time?
Week 3 (n 63)
Week 6 (n51)
Week 9 (n 60)
16
Does the online learning environment support
critical inquiry ?
Muukkonen et al 1999
Stahl 2006
17
Does the online learning environment support
critical inquiry ?

• Few threads reached a definitive conclusion
• Closer synthesizes and ends debate
• Closer often ignored
• Elements found
• Information Gathering
• Synthesis
• Concrete experience
• Reflective observation.
• Critical evaluation
• Deepening questions
• Generating subordinate questions
• Refining given knowledge
• Generating hypotheses
• Open-ended debate ?
• Not problem centered ?

18
What interactions generate greatest student
engagement

• Analysis of all 951 student messages
• Analysis of Read frequency for different message
  types
• Knowledge-elicitation messages (asking
  questions) generated significantly more (24)
  reads pre message than any other type of message.
• Average reads per message for all messages is
  16.78
• Some participants messages are read more
  frequently than others

19
Who are the most attended to posters ?
20
Why are some posters more engaging ?

• Does frequency of posting messages affect
  popularity?
• Does length of message affect frequency of reads?
  
• Does position of messages affect frequency of
  reads ?
• Does type of participant affect frequency of
  reads ?

21
Is frequency of posting related to popularity?

• Correlation between number of messages and total
  reads of a persons messages is 0.97,
• Weak -0.21 correlation between frequency of
  posting and reads/message.
• Most frequent poster posted 136 messages which
  attracted an average of 15.65 reads per message.
• The average messages per person was 37
• Top three most attended to participants posted
  an above average number but subject 20 did not.
• Two of the least attended to participants posted
  well above average numbers of messages.

22
Does length of message relate to read frequency

• Correlation between length of post and reads for
  that post 0.011
• Grouping messages into very short (lt 101 words),
  Short (101200 words), medium (210300 words) and
  long (gt301 words)
• One-Way ANOVA on frequency of reads gives an f
  value of .373 and a significance level of .773,
  no apparent significant effect

23
Does position of message affect frequency of reads

• Messages posted in the first 2 days of a thread
  are read significantly more frequently (f36.339,
  p 0.000) than later messages.
• Messages posted after the third day are read by
  less than 50 of participants.
• If a message is one of the first 10 posted it is
  much more likely to be read than later messages
  (f22.564, p 0.000).
• However only two of the most attended to
  participants are early posters.

24
Does type of participant affect frequency of reads

• The most attended to participants posted more
  facilitation messages (39 of messages posted)
• The least attended to participants typically
  posted far fewer facilitation messages. (23 of
  messages posted).

25
Conclusions

• Peer Facilitation does work
• Students quickly identify valuable contributors
• Early stages crucial
• Changing Contributor to Facilitator
• Identification of thought leaders
• Asking questions gets responses
• Fluid patterns of behavior within the community
• Volume is not the same as quality

26
Future Work
Limitations

• Small, exploratory study
• Initial framework
• Open to debate

• Influence of prior online learning-experience on
  patterns of behavior
• Larger sample size
• Deeper analysis of content
• Explore vicarious learning contributions more
  fully
• Explore why patterns change
• Compare ill-defined vs. well-bounded questions.

27
Questions?