This document provides an overview of learning analytics and discusses how learning analytics can be used to gain insights from educational data to improve teaching and learning. It discusses key concepts in learning analytics like educational data mining, the types of data that are collected, common analysis methods used, and how results can be visualized and applied. The document also presents examples of learning analytics research on analyzing discussion forum data to identify patterns in how students engage with each other's posts and the relationships between different listening behaviors and student contributions.

1. A Vision for Learning Analytics
Alyssa Friend Wise
Associate Professor, Learning Sciences & Educational Technology
Director, Learning Analytics Research Network at NYU (NYU-LEARN)
alyssa.wise@nyu.edu
@alywise
Learning
Theory
Computational
Power
Actionable
Insight
Big
Data
+ +
=>

2. THE DEVELOPMENT
OF METHODS FOR
MAKING DATA-BASED
DISCOVERIES WITH
THE UNIQUE KINDS
OF INFORMATION
THAT COMES FROM
EDUCATIONAL
SETTINGS
BAKER (2010)
EDUCATIONAL
DATA MINING

3. THE COLLECTION AND
ANALYSIS OF DATA
TRACES RELATED TO
LEARNING IN ORDER TO
INFORM AND IMPROVE
LEARNING PROCESSES
AND/OR THEIR
OUTCOMES
SIEMENS ET AL., (2011)
LEARNING
ANALYTICS

4. DATA
THEORY
COMPUTATION
DISPLAY
USE

5. USE
INSIGHT
+
ACTION

6. HOW CAN ANALYTICS INFORM TEACHING
AND LEARNING DECISIONS?
K E Y Q -
W H AT A R E T H E
G O A L S O F T H E
E D U C AT I O N A L
A C T I V I T Y ?
I S I T G O I N G A S
W E E X P E C T / WA N T ?
W H AT S H O U L D W E
D O D I F F E R E N T LY ?
I N S T R U C TO R S -
I D E N T I F Y +
H E L P S T U D E N T S
W I T H S P E C I F I C
N E E D S
S E E C L A S S - W I D E
PAT T E R N S + C H A N G E
M AT E R I A L S /
S T R AT E G I E S
S T U D E N T S -
C A N B E T T E R U N D E R S TA N D + TA K E
C O N T R O L O F H O W T H E Y S T U DY

7. DATA
USE

8. DEMOGRAPHICS ( W H O T H E Y A R E )
E X : G E N D E R , PA R E N TA L B A C KG R O U N D
PERFORMANCE ( H O W T H E Y ’ V E D O N E P R I O R )
E X : S TA N D A R D I Z E D T E S T, R A N K I N G , G R A D E S
ACTIVITY ( T H I N G S T H E Y D O )
E X : LO G - F I L E S , S E L F - R E P O R T, P H Y S T R A C E S
ARTIFACT ( T H I N G S T H E Y C R E AT E )
E X : P R O B L E M A N S W E R S , W R I T T E N T E X T
ASSOCIATION ( C O N N E C T I O N S B E T. T H I N G S )
E X : S T U . I N T E R A C T I O N S , R E S O U R C E C O - U S E
DATA
HOPPE IN
SUTHERS ET
AL. (2015)

9. DATA
THEORY
USE

10. DATA MINING
Image Credit: Scott Clark via Flickr (CC BY 2.0), adapted

11. DATA GEOLOGY
Image Credit: APS Museum via Flickr (CC BY 2.0), adapted
( S H A F F E R , 2 0 1 3 )

12. DATA ARCHEOLOGY
Image Credit: Pedro Szekely via Flickr (CC BY 2.0), adapted
( W I S E , 2 0 1 4 )

13. “FLAT ONTOLOGY” BEHAVIORAL
PATTERNS HAVE LITTLE USE IN
EDUCATION (RIEMANN IN WISE & SCHWARZ, 2017)
EXAMPLE: DOES MORE TIME /
CLICKS SHOW A STUDENT IS:
- ENGAGED?
- STRUGGLING?
THEORY CAN GIVE GUIDANCE
ABOUT HOW TO DISTINGUISH
DATA
IS A
PROXY

14. DATA
THEORY
COMPUTATION
USE

15. S I N H A E T A L . ( 2 0 1 4 )
GOOD DATA
FEATURES
NEED TO BE
ENGINEERED
Raw
Clicks
Aggregate
Features
Critical
Concept
Info
Process.
Play
SeekFwd
ScrollFwd
RateFast
Skipping Disengaged Low
Play
Pause
SeekBw
SeekFwd
Checkback Searching
for specific
info
Med
Play
Pause
SeekBw
Rewatch Reviewing
content
High
Play
Pause
SeekBw
ScrollBw
Clarify
Idea
Tussling
with
content
Very High
F U Z Z Y PAT T E R N M ATC H I N G O F
M O O C V I D E O WATC H I N G DATA

16. STRUCTURE DISCOVERY
(CLUSTER, SNA)
PREDICTIVE MODELS
TEMPORAL ANALYSIS
NATURAL LANGUAGE
PROCESSING (NLP)
VISUAL ANALYTICS
SOME
COMMON
METHODS
USED IN
LEARNING
ANALYTICS

17. DATA
THEORY
COMPUTATION
DISPLAY
USE

18. Resource 1
Resource 2
Resource 3
Resource 4
Resource 5
Resource 6
Resource 7
Resource 8
Resource 9
Resource 10
Resource 11
Resource 12
Resource 13
Resource 14
Resource 15
Resource 16
Resource 17
Resource 18
Resource 19
Resource 20
Resource 21
Resource 22
A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
My Learning Dashboard - Student Resource Activity

19. DATA
THEORY
COMPUTATION
DISPLAY
USE

20. E T H I C S
E T H I C S
I
M
P
A
C
T
I
M
P
A
C
T
DATA
THEORY
COMPUTATION
DISPLAY
USE

21. D A T A
T H E O R Y
C O M P U T A T I O N
D I S P L A Y
U S E
Detecting different ways learners
attend to others’ posts in online
discussions and how they relate
to the contributions they make
themselves as an element of
dialogic participation

22. T H E O R Y
D A T A
U S E
D I S P L A Y
Online
Listening
Theory
C O M P U T A T I O N
• Goal of online discussions is for
learners to build understanding
through dialoging with others
• Drawing on (neo-piagetian)
theories of social constructivism
at a basic level this involves
Wise et al., 2012a; Wise et al., 2013
Externalizing one’s ideas by contributing
posts to an online discussion
Taking in the externalizations of others
by accessing existing posts

23. T H E O R Y
D A T A
U S E
D I S P L A Y
Listening
Not
Lurkers
C O M P U T A T I O N
Lurker
A person
w/ nefarious
connotations
who looks at
posts as a stand
alone activity
and doesn’t
contribute
themselves
Listening
A behavior
that is normal /
productive,
part of a larger
engagement in
discussion,
and can inform
future discussion
contributions
Wise et al., 2013; Muller et al., 2010; Preece et al., 2004

24. T H E O R Y
D A T A
U S E
D I S P L A Y
Not “if”
but “how”
lllllisten
C O M P U T A T I O N
• Prior research assumed students
generally attend to others’ posts
• Yet evidence of limited attention
to previous discussion posts
• Our data-mining work shows that
learners make widely varying
choices in the decision space
• Frequency /length of log-in
• Which posts attended to, in what
order, for how long
• Revisit as often as wanted
• Reply when + where desired
Wise et al., 2013; 2012a, Brooks et al., 2013; Palmer, et
al. 2008; Dennen, 2008; Hewitt, 2003;

25. U S E
D I S P L A Y
C O M P U T A T I O N
• Online discussion forums don’t
naturally collect listening data
• Need to know what specific post a
student is looking at
• Modifications made to open-source
forum “Phorum”
• Collect ~1000 + records per student
Collecting
Listening
Data
D A T A
T H E O R Y

26. U S E
D I S P L A Y
C O M P U T A T I O N
• Turn comprehensive point-in-time
entries into individual activity records
• Coding events and calculating duration
• Clicks can initiate or complete actions
• Conceptual event categories
• Scan vs. read (6.5 wps threshold)
• Post and review (self-oriented)
• Divide activity into sessions using
abandonment criteria (time of inactivity)
• Determine time window for variable
calculation (study-dependent – could be
overall or based on learning activity,
fixed time, individual student session)
Data
Processing
Needs
D A T A
T H E O R Y

27. U S E
D I S P L A Y
• mySQL queries merging log + post tables
produces list of all actions, time-date
stamp, ID of user performing action, post
acted on, ID of user who created post
• Excel VBA macros clean data, separate
users, calculate initial event duration,
divide sessions of use, estimate session-
ending events (avg. speed (user| event) x
post length), recode reads and reviews
• Macros also calculate variables at desired
unit of analysis –(e.g. # of sessions, av.
session length, % sessions with posts, %
other's posts viewed (or read), % real
reads, av. real read length, # of reviews
(self/other) , av. review length)
T H E O R Y
Data
Processing
Techniques
C O M P U T A T I O N
D A T A
Wise & Chiu, 2014; Wise et al. 2014; 2013; 2012a

28. U S E
D I S P L A Y
• Cluster Analysis used to identify and
characterize learner subpopulations
• Ward’s hierarchical technique w/
squared Euclidean distance used to
determine cluster # + membership
• Multi-dimensional cluster contrasts
• HLM Mixed Modeling used to examine
listening-speaking relationships + take
into account potential interdependencies
of students collaborating in groups
• Overall relationships, Group Differences
Intervention Effects, Time Effects
• Statistical Discourse Analysis used to
detect distinct segments of discussion
and build explanatory models of these
• Multiple time-lag variables included
T H E O R Y
Modelling
Techniques
C O M P U T A T I O N
D A T A
Wise & Hsiao, in preparation; Wise et al. 2014; 2013;
Wise & Chiu, 2014; Wise & Chiu, 2011

29. Roles
Listening
Study
Explanatory
Model # of
sessions
% of
sessions w/
posts
Average # of views per
session
% of
others’
posts
viewed
% of
others’
posts read
% of total
views that
were reads
Average length of
timereading post
# of
reviews
# of posts
made
Average
length of
posts
Average time to create
posts
1st
Week
Group
size
% of women
Mean age
SD age
Mean course grade
Woman
Age
Course
grade
Goal orientation
Synthesizer
Wrapper
Role week
already occurred
Week Group Group-week Student Student-week Dependent variables
(no sig. var.)
–0.29 **
+0.13 **
+0.20 ***
–0.14 **
–0.25 ***
+0.38 ***
+0.53 **
–0.17 **
–0.14 *
–0.14 **
+0.25 ***
+0.16 *
+0.36 **
+0.22 ***
+0.26 ***
+0.34 *
+0.48 ***
+0.60 ***
–0.13 **
+0.18 ***
+0.35 ***
–0.27 ***
+0.56 **
+0.31 ***
–0.84 ***
–0.39 **
+0.86 ***
+1.16 ***
–0.96 ***
–0.66 ***
+0.27 ***
+0.84***
+0.30 ***
+0.21 **

30. U S E
D I S P L A Y
Microanalytic case studies using dynamic
discussion maps; temporal analysis efforts
T H E O R Y
Additional
Data
Analysis
C O M P U T A T I O N
D A T A
Wise et al. 2012a; 2012c; 2012d
Date Time Session Action Duration
(min)
Length
(words)
Post
6/3/2011 23:46 1 Read 12.43 413 447
6/3/2011 23:52 1 Read 1.73 60 455
6/4/2011 00:08 1 View 0.23 117 459
6/4/2011 00:09 1 Read 4.51 204 460
6/4/2011 23:49 2 Post 3.18 121 477

31. T H E O R Y
D A T A
U S E
D I S P L A Y
Studies +
Findings
C O M P U T A T I O N
Research conducted on large UG courses in
business + ed psych (online + blended offerings)
Wise & Chiu, 2014; Wise et al. 2014; 2013; 2012a; 2012c;
Common
Patterns
Characteristic Behaviors
Disregardful
Minimal attention to others’ posts (few posts
viewed; short time viewing). Brief and
relatively infrequent sessions of activity.
Coverage
View a large proportion of others’ posts, but
spends little time attending to them (often
only scanning). Short but frequent sessions
of activity, focusing primarily on new posts.
*May be socially-oriented or content-driven.
Focused
View limited number of others’ posts, but
spends substantial time attending to them.
Few extended sessions of activity.
Thorough
Views a large proportion of other’s posts +
spends substantial time attending to many of
them. Long overall time spent listening;
considerable reviews of others’ posts.

32. T H E O R Y
D A T A
U S E
D I S P L A Y
Studies +
Findings
C O M P U T A T I O N
Wise & Chiu, 2014; Wise et al. 2014; 2013; 2012a; 2012c;
Dimension Characteristic Behaviors
Breadth The quantity of unique posts one views -
important in terms of the diversity of ideas
a learner is exposed to.
Depth The length of time spent reading posts –
important to allow for deeper consideration
of others’ ideas.
Temporal
Contiguity
The degree to which learners disperse or
concentrate their participation - important
for integration / evolution over time.
Revisitation The extent to which students return to posts
made by themselves and others – important
in metacognition and self-regulation.
Research conducted on large UG courses in
business + ed psych (online + blended offerings)

33. T H E O R Y
D A T A
U S E
D I S P L A Y
Studies +
Findings
C O M P U T A T I O N
Wise & Chiu, 2014; Wise et al. 2014; 2013; 2012a; 2012c;
Breadth Depth
Depth
Breadth
Low High
Low Disregardful Coverage
High Focused Thorough

34. T H E O R Y
D A T A
U S E
D I S P L A Y
Studies +
Findings
C O M P U T A T I O N
Relationships
Listening Depth
• A greater % of real reads predicts richer
argumentation ( reasoning/qualifiers)
Listening Reflectivity
• Reviewing others’ posts multiple times
predicts greater responsiveness
• Reviewing one’s own posts multiple
times predicts richer argumentation
(# of claims made)
Listening Breadth
• Reading a greater % of posts and
viewing a greater % of posts than those
read predicts richer argumentation
(evidence used to support claims)
Wise & Chiu, 2014; Wise et al. 2014; 2013; 2012a; 2012c;

35. T H E O R Y
D A T A
U S E
D I S P L A Y
Studies +
Findings
C O M P U T A T I O N
Interventions
Assigning Students Roles to Play
• Synthesizer / Wrapper roles increased the
% of posts read during in-role weeks (and
sometimes the % of real reads), but effect
weakly sustained post-role
• Roles often contributed extensive
summaries that acted as pivotal posts
spurring whole discussion to more
advanced knowledge construction phase
Open-Ended vs Distinct Alternative Tasks
• Open-ended business challenge led to a
higher % of real reads and average time
spent reviewing one’s own posts
• Follow-on argumentation effects possible
Wise & Hsiao, in prep; Wise & Chiu, 2014; Wise & Chiu, 2011;

36. D A T A
D I S P L A Y
U S E
Embedded
& Extracted
Analytics
C O M P U T A T I O N
Metric Me
(Week X)
Class
(Week X)
Participation
range
6 days 5 days
# of sessions 3 11
Average session
length
48 min 39 min
% of sessions
with posts
67% 49%
# of posts made 4 7
Average post
length
386 words 216 words
% of posts read 68% 79%
#of self reviews 2 5
#of peer reviews 12 8
T H E O R Y

37. D A T A
D I S P L A Y
U S E
Analytics
Integration
C O M P U T A T I O N
Clear guidelines and discussion of:
• purpose of engaging in online
discussion
• instructor’s expectations for a
productive process of engaging in
online discussions
• how the learning analytics provide
indicators of this process
articulating one’s ideas, being exposed to the ideas
of others, negotiating differences in perspective
attending deeply to a spectrum of others’ ideas,
and contributing comments that are responsive
and rationaled,
percent of posts read introduced as a metric that
has clear meaning in the context of the activity
T H E O R Y

38. D A T A
D I S P L A Y
U S E
Analytics
Integration
C O M P U T A T I O N
Discussion Participation Guidelines
Attending to Others Posts
Broad Listening: Try to read as many
posts as possible to consider everyone’s
ideas in the discussion. This can help
you examine and support your own ideas
more deeply. However, when time is
limited it is better to view a portion in
depth, then everything superficially.
*The visual interface shows posts that
you have viewed in blue and new
ones in red to help you track this.
T H E O R Y
Learning Analytics Guidelines
Attending to Others’ Posts
% of
posts
read
The proportion of posts
you read (not scanned)
at least once.
It is good to read as many posts as
possible to consider everyone’s
ideas in the discussion However,
when time is limited it is better to
view a portion in depth, then
everything superficially.
Metric Me
(Week X)
Class
(Week X)
Participation range 6 days 5 days
# of sessions 3 11
Average session
length
48 min 39 min
% of sessions with
posts
67% 49%
# of posts made 4 7
Average post
length
386 words 216 words
% of posts read 68% 79%
#of self reviews 2 5
#of peer reviews 12 8

39. D A T A
D I S P L A Y
U S E
Analytic
Agency
(Goals)
C O M P U T A T I O N
T H E O R Y
Discussion guidelines present metrics as a
starting point for consideration, not as
absolute arbiters of engagement
Goal-setting is an explicit and structured
part of the learning activity as students set
weekly goals for engaging in the online
discussions in an online reflection journal
(in the LMS)
S A M P L E S T U D E N T G O A L S
“I aim to read all (most) posts [in the discussion], and actively
participate in two threads in addition to any I create”
“Well, since I didn't hit last week’s goal really I [still] need to
do that, also keep the length [of my posts] down and get more
interactive with the other kids.”
“As a goal for the next discussion, I will try to synthesize ideas
from different thread areas”

40. D A T A
D I S P L A Y
U S E
Analytic
Reference
Frame
C O M P U T A T I O N
T H E O R Y
Establish a rhythm for reflection
• Weekly cycle of reviewing the analytics
• Evaluate progress towards the goals
• Assess when the goals themselves need
to be updated or revised
Provide a dedicated space
• Online reflective journal (private wiki)
• Examine trajectory over time
S A M P L E S T U D E N T R E F L E C T I O N
“I found that I wanted the challenge of trying to
up the % of overall posts that I reviewed each
week. This also meant slowing down my reading
since the data would not record a quick read of
the information. The overall result was that I
think I learned more and was able to get a
broader sense of opinion concerning the
readings.”

41. D A T A
D I S P L A Y
U S E
Analytic
Agency
(Reflection)
C O M P U T A T I O N
T H E O R Y
• Continually reminding students of
theoretical patterns
• Prompting reflection on individual
progress and goals
• Value and danger of comparisons to peers
S A M P L E S T U D E N T R E F E R E N C E U S E
“I was surprised to see that most of classmates
checked the forum more than I did…”
“Since all my numbers are below the average so that
makes me feel, ‘Oh my gosh, I’m kind of jumping out
of this class’ or something. It is kind of a little bit –
sometimes depressing.”
“Compared to the previous week, [my] number of
reviews of others’ posts has been hugely increased …
and I did spend more time to read and understand
others’ posts.”

42. D A T A
D I S P L A Y
U S E
Analytic
Implementation
Design
C O M P U T A T I O N
T H E O R Y
Wise, 2014 ; Wise et al., 2016; Wise & Vytasek, 2017
Sensitizing issues for the design of analytics use
that open a space for inquiry and conversation
Aligned Design

43. Integration (technological and pedagogical) made analytics a
coherent part of the learning process
Students embraced agency in setting (often recurring) personal
goals and evaluating their progress, no “big brother” issues
Individual, peer, and instructor reference frames were important for
making sense of the data; reactions were both cognitive and
emotional
Reflection on data a powerful starting place
Concrete and proximal goal-setting is harder
Change happens slowly, isn’t always intentional, requires support!
PILOT STUDY
FINDINGS

44. DISTINCT D
PROFILES

45. D A T A
T H E O R Y
C O M P U T A T I O N
D I S P L A Y
U S E
Investigation of the interaction
practices in large-scale learning
environments based on analysis
of the artifacts left behind by
students’ and instructors’ activity

46. D A T A
D I S P L A Y
U S E
Looking to
Learn
(in all the
wrong places)
C O M P U T A T I O N
MOOCs offer exciting possibilities for
learning but suffer from high drop-out,
low support + limited interaction
Discussion forums can foster social,
interactive and responsive learning
but often have problems of overload,
disorganization + chaos
Messages focused on learning course
material are mixed with social +
logistical (and don’t get replies).
Sub-forums + votes don’t solve this…
can we use NLP and ML to identify
content-related discussion threads?
T H E O R Y

47. Courses
Course Name Domain Length Pre-Reqs Assessment Usage
StatMed’13 Statistics 9 wks N/A
Quizzes,
homework,
final
Training Set
StatMed’14 Statistics 9 wks N/A
Quizzes,
homework,
final
Test Set:
Cross-Offering
StatLearn Statistics 9 wks
Intro stats,
L algebra,
Quizzes
Test Set:
Cross-Course
PSY Psychology 12 wks N/A Quizzes, final
Test Set:
Cross-Domain
(Near)
YBW Physiology 9 wks N/A Quizzes, final
Test Set:
Cross-Domain
(Far)

48. D A T A
D I S P L A Y
T H E O R Y
Modelling
Questions
1. Do content-related threads in a statistics
MOOC discussion forum have distinct
linguistic features ?
2. Can these be used to create a model to
reliably identify them?
3. Does the model generalize to
• another offering (same MOOC)?
• a different statistics MOOC?
• MOOCs on other topics?C O M P U T A T I O N
U S E

49. D I S P L A Y
T H E O R Y
Natural
Language
Processing
[in Lightside RW]
U S E
• Use linguistic features to predict if post
is about learning content
• Unit of analysis = Thread
• Initial representation by Starter post
• Hand-coding by research assistants
• Detailed coding guide + training
• Good interrater reliability (a > 0.75)
• Bag-of-words feature extraction
• Unigrams and bigrams only, parts of
speech unhelpful, stop words *IN*
C O M P U T A T I O N
D A T A

50. Courses
Course
Name
Total # of
Posts
# of Threads
(Starting
Posts)
# of Starting
Posts Coded
Content-Related
Starting Posts
StatMed’13 3320 844 837 [844] 47%
StatMed’14 1218 310 304 [310] 54%
StatLearn 3030 626 298 [300] 51%
PSY 2307 438 438 [438] 28%
YBW 2467 825 299[300] 40%

51. D I S P L A Y
T H E O R Y
Predictive
Modeling
[in R]
U S E
• 2236 extracted features used to train
a binary L2 regularized logistic
regression model
• Confusion matrix and data
restructuring for model
optimization
• Evaluation via 10-fold cross
validation + 4 test sets
C O M P U T A T I O N
D A T A

52. Note: Views and votes did not improve model and
were poor predictors on their own
D I S P L A Y
T H E O R Y
Model
Results
U S E
C O M P U T A T I O N
D A T A
StatMed
’13
StatMed
’14
Stat
Learn
PSY YBW
Accuracy 0.80 0.81 0.80 0.80 0.73
Kappa 0.61 0.62 0.60 0.52 0.42
Recall 0.79 0.85 0.90 0.72 0.60
Precision 0.79 0.81 0.76 0.62 0.68
0.0
0.5
1.0
Recall Precision
0.0
0.5
1.0
Accuracy Kappa

53. D I S P L A Y
T H E O R Y
Improving
Results
U S E
Dynamic Interrelated Post and
Thread Categorization (DIPTiC)
• Apply classifier to both thread starter
and all replies (after separately
verifying performance on replies)
• Establish cutoff threshold percent for
content replies in content thread
• Compare starter- and reply-based
classifications, manual triage on
mismatches
• Accuracy improvement on sample
data estimated as .81 -> .88
C O M P U T A T I O N
D A T A

54. StatMed’13
StatMed’14
StatLearn
YBW
PSY
Course Subject Learning Process Question Words Connectors Existence/Condition Course TasksQuality/Quantity Effort / Action People Appreciation/Greeting
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Content - Related
NonContent - Related
Numberoffeatures Content-Related
Discussion Posts
Non-Content-Related
Discussion Posts
Question Words + Connectors
e.g. “can” “does” “why” “how”
“which “and” “of” “than” “is”
Course Tasks + People
+ Appreciation / Greetings
e.g. “answer” “exam” “course”
“lecture” “thank” “good” “I” “my”
Top Feature Distribution by Category

55. D A T A
D I S P L A Y
T H E O R Y
U S E
Student &
Instructor
Support
C O M P U T A T I O N
Post-Hoc Filtering
• Filter to select only content threads
• Reduce # of threads to review by more
than half and create > 85% hit rate of
those reviewed
Live-Tagging Tool
• Content / non-content label suggested
to learners (manual change possible)
• Support student metacognition,
awareness of contributions

56. D A T A
T H E O R Y
StatMed’14
Instructor
Communities
(SNA)
D I S P L A Y
C O M P U T A T I O N
U1-C U417-C U1-NC U417- NC
# of nodes (% in network)
184
(55%)
75
(22%)
168
(43%)
47
(12%)
# of edges (% in network)
400
(47%)
105
(12%)
315
(44%)
55
(8%)
Avg node degree (SD)
4.35
(11.06)
2.8
(7.56)
3.75
(11.18)
2.34
(6.03)
Avg edge weight (SD)
2.23
(3.21)
1.83
(1.72)
2.11
(2.48)
1.20
(0.44)
U S E
Content Threads
Non-Content Threads

57. D A T A
D I S P L A Y
T H E O R Y
U S E
Identify &
Assist
Instructors
C O M P U T A T I O N
U1
• Responses at all levels
• Coaching and supporting
• Social presence cues
U417
• Responses only to thread starters
• Straight forward answers
• Little social presence
“Think about it again using this hint
and let me know if you have any other
questions.”
“That is correct - Nice! So how would
you use this to solve the question?”
“A bell shape is not necessary. You could
have a 'bimodal' distribution where the
two groups do not follow a bell shape.”

58. D A T A
D I S P L A Y
T H E O R Y
U S E
Identify
Student
Learning
Communities
C O M P U T A T I O N
U225: Congrats [u10]! Yes, it has been hard, but fun, and we
learned an awful lot, right?
U110: Great! Everyone it was a pleasure to work with you.
Thank you….
U10: YES [u225]! And [u110] - the test was scary - I thought of
my discussion board friends often!!
U216: Thanks, thanks so much to [u10], [u152], [u110], [u225]
and everybody who helped us to understand this beautiful
course! And in my case also for writing many posts, I see I have
improved my English skills and my statistics vocabulary!!!
U225: [u10], [u216], [u152], [u110], [u515] and everyone, your
discussions helped me so much. I was always a few days
behind you in homework - glad I was able to catch up in the
last weeks and participate a little bit….
U225: [u10], [u216], [u152], [u110], [u515] and everyone,
your discussions helped me so much. I was always a few days
behind you in homework - glad I was able to catch up in the
last weeks and participate a little bit….
U10: YES [U225]! And [u110] - the test was scary - I thought of
my discussion board friends often!!
Content-Related
Learner Module 1

59. Wise & Cui (in press) Learning communities in the crowd: Characteristics of content
related interactions and social relationships in MOOC forums. Computers & Education.
Wise & Cui (2018). Unpacking the relationship between discussion forum participation and
learning in MOOCs: Content is key. Proceedings of LAK’18. ACM.
Wise, Cui & Jin (2017). Honing in on social learning networks in MOOC forums: Examining
critical network definition decisions. Proceedings of LAK’17). ACM.
Cui, Jin & Wise (2017). Humans and machines together: Improving characterization of
large scale online discussions through dynamic interrelated post and thread categorization
(DIPTiC). In Proceedings of Learning at Scale 2017. Cambridge, MA: ACM.
Wise et al. (2017) Mining for gold: Identifying content-related MOOC discussion threads
across domains through linguistic modeling. Internet and Higher Education, 32, 11-28.
Wise, Cui & Vytasek (2016). Bringing order to chaos in MOOC discussion forums with
content-related thread identification. In Proceedings of LAK’16. ACM.

60. Wise & Hsiao (2018) Self-regulation in online discussions: Aligning data streams to
investigate relationships between speaking, listening, and task conditions. Computers in
Human Behavior.
Marbouti, F. & Wise, A. F. (2016) Starburst: A new graphical interface to support
productive engagement with others’ posts in online discussions. Educational Technology
Research & Development, 64(1), 87-113.
Wise, Hausknecht & Zhao (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.
Wise, Speer & Marbouti & Hsiao (2013). Broadening the notion of participation in online
discussions: Examining patterns in learners’ online listening behaviors. Instructional
Science, 41(2), 323-343.
Wise, Perera, Hsiao, Speer & Marbouti (2012). Microanalytic case studies of individual
participation patterns in an asynchronous online discussion in an undergraduate blended
course. The Internet and Higher Education, 15(2), 108-117.

61. Wise, A. F. (in press). Learning analytics: Using data-informed decision-making to improve
teaching and learning. In Contemporary technologies in education: maximizing student
engagement, motivation, and learning. New York: Palgrave Macmillan.
Wise & Vytasek (2017). Learning analytics implementation design. Handbook of Learning
Analytics. Edmonton, AB: SoLAR.
Wise, Vytasek, Hausknecht & Zhao (2016). Developing learning analytics design knowledge
in the “middle space”: The student tuning model and align design framework for learning
analytics use. Online Learning, 20(2), 1-28.
Wise, Zhao, & Hausknecht (2014). Learning analytics for online discussions: Embedded
and extracted approaches. Journal of Learning Analytics, 1(2), 48-71.
Wise (2014). Designing pedagogical interventions to support student use of learning
analytics. Proceedings of LAK’14. Indianapolis, IN: ACM.
Wise, Zhao, & Hausknecht (2013). Learning analytics for online discussions: A pedagogical
model for intervention with embedded and extracted analytics. Proceedings of
LAK’13. Leuven, Belgium: ACM.

62. Handbook of Learning Analytics
solaresearch.org/hla-17
MORE ABOUT LA:
Sample Issues (all open access online)
2018 - 4(3) – Temporal Learning Analytics (forthcoming)
2017 - 4(2) – The Shape of Educational Data
2016 - 3(2) – Multimodal Learning Analytics
2016 - 3(1) – Ethics & Privacy in Learning Analytics
2015 - 2(2) – Learning Analytics & Learning Theory
2015 - 2(1) – Self-Regulated Learning & Analytics
2014 - 1(1) – Inaugural Issue
Other Selected Overviews
ACM Proceedings of LAK’11-LAK’17
Sclater, N. (2017) Learning Analytics
Explained. London, UK: Routledge
Ferguson, R., et al. (2016). Research
Evidence on the Use of Learning
Analytics: Implications for Education
Policy. Publications Office of the EU.
EDUCAUSE 2016 Horizon Report
LACE Evidence Hub (website)
LASI 2018
June 11-13 New York City March 5-9, 2018
Sydney Australia

63. LASI’18
Learning Analytics Summer Institute
June 11-13, 2018 | New York, NY
solaresearch.org/events/lasi/lasi-2018/
Hosted by Teachers College, Columbia University & the
Learning Analytics Research Network, New York University
• 2.5 day interactive event
• Workshops, tutorials, speakers &
networking
• Topics include:
• Natural language processing of
student work
• Predictive learning analytics,
• Dashboard design
• Data privacy considerations

64. Publications underlying this presentation
accessible via Google Scholar
alyssa.wise@nyu.edu
@alywise
Alyssa Friend Wise
Associate Professor, Learning Sciences & Educational Technology
Director, Learning Analytics Research Network at NYU