Presentation at San Diego State University on April 12, 2013.
Educational researchers have found that students from under-represented minority families and other disadvantaged demographic backgrounds have lower achievement in online (or hybrid) courses compared to face-to-face course sections (Slate, Manuel, & Brinson Jr, 2002; Xu & Jaggars, 2013). However, these studies assume that "online course" is a homogeneous entity, and that student participation is uniform. The content and activity of the course is an opaque "black box", which leads to conclusions that are speculative at best and quite possibly further marginalize the very populations they intend to advocate for.
The emerging field of Learning Analytics promises to break open this black box understand how students use online course materials and the relationship between this use and student achievement. In this presentation, we will explore the countours of Learning Analytics, look at current applications of analytics, and discuss research applying a Learning Analytics research method to students from at-risk backgrounds. The findings of this research challenge stereotypes of these students as technologically unsophisticated and identify concrete learning activities that can support their success.
The Achievement Gap in Online Courses through a Learning Analytics Lens
1. The Achievement Gap in
Online Courses through a
Learning Analytics Lens
John Whitmer, Ed.D.
Academic Technology Services
California State University, Office of the Chancellor
San Diego State University
April 12, 2013
2. Motivating Questions …
1. Do our current uses of academic technologies
(such as online learning) decrease or
exacerbate the achievement gap?
2. If we don’t believe our current uses serve
these students, how do we know? What can
we do about it?
3. Agenda
1. Context: CSU Achievement Gap &
Conceptual Framework
2. Recent Conventional Research in
Online Courses
3. Research using Learning Analytics &
Course Redesign
4. Next Steps & Discussion
5. Increasing Access to Higher Education in the U.S.
Percent of Percent of
Percent
1976 Enrollment 2010 Enrollment
Increase
(1976) (2010)
All Students 10,986 21,016 91%
White 9,076 83% 12,723 61% 40%
Asian/Pacific
198 2% 1,282 6% 548%
Islander
URM Students 1,493 14% 5,977 28% 300%
American
76 1% 196 1% 158%
Indian
Black 1,033 9% 3,039 14% 194%
Hispanic 384 3% 2,741 13% 614%
Table adapted from data in NCES Digest of Educational Statistics (2011)
6. Increased Access to Higher Education
Enrollment Increase by Race/Ethnicity in Higher Education (1976-2010)
700%
600%
500%
400%
300%
200%
100%
0%
All Students White Asian/Pacific URM Students American Indian Black Hispanic
Islander
Table adapted from data in NCES Digest of Educational Statistics (2011)
9. No Significant Difference:
Framing Academic Technology
[academic technologies]
are “mere vehicles that
deliver instruction but do
not influence student
achievement any more
than the truck that
delivers our groceries
causes changes in our Image courtesy bsabarnowl @ Flickr
nutrition”. (Clark, 1983)
Index of studies: http://www.nosignificantdifference.org/
11. Adaptability to Online Learning:
Differences Across Types of Students and
Academic Subject Areas (2013: Xu, Jaggers)
Compares persistence and
grade between online and f2f
courses
Compares same student
Washington Community and
Technical Colleges (2 yr)
Studied 500,000 course
enrollments (10% online),
40,000 individuals
Data 2004-2009
13. Major Finding by Population
Overall Course Result Select Populations Result
Online GPA Average 2.77 Black -0.394
F2F GPA Average 2.98 Males -0.288
Entire Population -0.215 Academic Preparedness
Effect by Subject -0.267 (F2F GPA<3.0 First Term) -0.314
Age < 25 -0.300
Cohort Effect
(Courses w/+75% online
at-risk students v. less
than 25%) -0.359
Table adapted from Xiu & Jaggers, 2013
14. Major Finding by Subject
Overall Course Result Select Subjects Results
Online GPA Average 2.77 English -0.394
F2F GPA Average 2.98 Applied Knowledge -0.322
Entire Population -0.215 Social Science -0.308
Effect by Subject -0.267
Table adapted from Xiu & Jaggers, 2013
17. What’s your experience with
Online or Hybrid Course Design?
Who has designed a fully online or hybrid
course? (raise hands)
Of those who have, how many think it’s harder to
create online/hybrid materials than to create face
to face activities? (keep hands raised)
18. Peering into the blue box
Online Course
Faculty
Course Faculty Course &
Design Development Student
Support Training
(Reassigned
time, incentives, etc
.) Student
Specialist
Support use of
(Instructional online
designers, etc.) materials
19. Do we need
*students* to adapt …
or
Do we need to change how *we*
approach creating & evaluating
our technology-enhanced
instructional materials?
21. Learner Analytics
“ ... measurement, collection, analysis and
reporting of data about learners and their
contexts, for purposes of understanding and
optimizing learning and the environments in
which it occurs.” (Siemens, 2011)
22. Case Study: Intro to Religious Studies
• Undergraduate, introductory, high demand
• Redesigned to hybrid delivery format
54 F’s
through “academy eLearning program”
• Enrollment: 373 students
(54% increase on largest section)
• Highest LMS (Vista) usage
entire campus Fall 2010
(>250k hits)
• Bimodal outcomes:
• 10% increase on final exam
• 7% & 11% increase in DWF
• Why? Can’t tell with aggregated data
23. LMS Use Variables Student Characteristic
Variables
Administrative Activities Enrollment Status
(calendar, announcements) First in Family to Attend
Assessment Activities College
(quiz, homework, Gender
assignments, grade center) HS GPA
Content Activities Major-College
(web hits, PDF,
content pages) Pell Eligible
Engagement Activities URM and Pell-Eligibility
(discussion, mail) Interaction
Under-Represented
Minority
URM and Gender
Interaction
25. Predict the trend
LMS use and final grade is _______ compared to
student characteristics and final grade:
a) 50% smaller
b) 25% smaller
c) the same
d) 200% larger
e) 400% larger
26. Predict the trend
LMS use and final grade is _______ compared to
student characteristics and final grade:
a) 50% smaller
b) 25% smaller
c) the same
d) 200% larger
e) 400% larger
27. Correlation LMS Use w/Final Grade
Scatterplot of
Assessment Activity
Hits vs. Course
Grade
28. Combined Variables Regression Final Grade by
LMS Use & Student Characteristic Variables
LMS Student
Use Characteristic
Variables Variables
25%
(r2=0.25)
> +10%
(r2=0.35)
Explanation of change Explanation of change
in final grade in final grade
33. COURSE REDESIGN
Flagship: Program in Course Redesign, led by Carol Twig
(1999-2004)
– Pew funded 30 grants, $8.8M budget to redesign
courses for improved outcomes & lower costs
(institutionalization)
Result: 25 of 30 courses reported increased learning
outcomes, 5 no change (not worse!)
– 17 reduction DWF (10-20%)
– Cost reduction 20-77%, $3M annual savings
Adopted (with modifications by CSU, SDSU, Chico State)
– Chico State evaluations:
http://www.csuchico.edu/academy/outcomes/index.shtml
34. Sample Improvements
DWF (drop-failure-withdrawal) rates at Drexel were consistently reduced 10-
12 percent in the redesigned course.
At OSU, withdrawals were reduced by 3 percent, failures by 4 percent and
incompletes by 1 percent. As a result, 248 more students successfully
completed the course compared to the traditional course.
At TCC, students in redesigned sections had a 68.4 percent success rate
compared to 60.7 percent for traditional sections. Success rates were higher
for all groups of students regardless of ethnicity, gender, disability, or
original placement. The overall success rate for all composition students was
62 percent for the 2002-2003 year compared to 56 percent for the 1999-2000
year prior to redesign.
In the traditional course at USM, faculty-taught sections typically retained about
75 percent of students while adjunct- and TA-taught sections retained 85
percent. In the redesign, the retention rate was 87 percent. The rate of D and F
grades dropped from 37 percent in the traditional course to 27 percent in
the redesigned course. DFW rates dropped from 26 percent in the traditional
course to 22 percent in the redesign.
Source: Program in Course Redesign Round III: Lessons Learned
(http://www.thencat.org/PCR/R3Lessons.html)
35. Underserved Student Experiences
More comfort, higher
participation in online
forums
Appreciate ability to
anonymously rewind /
repeat / review materials
English language Source: Twigg, C. (2005). Increasing Success for
Underserved Students: Redesigning Introductory
Courses. URL:
learners decreased http://www.thencat.org/Monographs/IncSuccess.htm
social anxiety
36. Proven PCR Techniques for
Underserved Students
Interactive online tutorials (not PPT decks)
Continuous assessment / feedback
Increased student interaction
Individualized, on-demand support
Undergraduate learning assistants
Structural supports that encourage student
engagement / progress
38. Discussion
Do you think there is an achievement gap at
SDSU for under-served students in online /
hybrid / tech enhanced courses? What evidence
do you have for those beliefs?
What is SDSU doing about any existing
achievement gap w/academic technology?
What supports are in place or could be
developed?