Analyzing Students' Behavior in a Beginner's Programming Course
description
Transcript of Analyzing Students' Behavior in a Beginner's Programming Course
Analyzing Students' Behavior in a Beginner's
Programming CourseMarija Brkić, Higher Teaching Assistant
Maja Matetić, Associate Professor
Department of Informatics, University of RijekaRadmile Matejčić 2, 51000 Rijeka, Croatia
http://www.inf.uniri.hr
Why are we doing this?
• Task 6 of the strategy of the University of Rijeka for the period 2008-2013:
• Pass rate increase for 2nd year enrollment to 75%
• The course pass rate in the academic year 2012/2013 was 61%.
• The course pass rate in the academic year 2011/2012 was 63%.
• We are facing a falling pass rate!!!
Course info
• Programming 1
• mandatory course
• 1st year of undergraduate study of Informatics
• C++
• procedural programming
• 82 students in our case study
• LMS Moodle, supplemental instruction classes
ActivityActivity ScoresScores
Online quizzesOnline quizzes 2525
Self-evaluationSelf-evaluation 44
LabsLabs 99
GlossaryGlossary 22
ExamExam 2525
ActivityActivity 66
Visualization as a pre-processing tool
2012/2013
35%
21%
35%
9%
final exam
remedial exam
fail
dropout rate2011/2012
36%
24%
36%
4%
final exam
remedial exam
fail
dropout rate
Grade distribution
2012/2013
10
7 7
5
0
2
4
6
8
10
12
A B C D
Avoiding examination
2012/2013
0
5
10
15
20
25
30
35
1st quizz 2nd quizz exam
# o
f a
bs
en
t s
tud
en
ts
regular date
extra date
Repeating the course2012/2013
79,3%
13,4%
20,7%
7,3%
novices
students repeating the coursewho failed again
students repeating the coursewho passed
2011/2012
84,2%
7,9%
7,9%
15,8%
novices
students repeating the coursewho failed again
students repeating the coursewho passed
Additional activity I
2012/2013
0
10
20
30
40
50
60
70
1st 2nd 3rd 4th 5th 6th
Labs
# o
f s
tud
en
ts
success
total
Additional activity II
Relationship with the final grade
Pre-processing
• Missing values for one part of activities have been replaced with minimum values
• Examples with the remaining missing values have been filtered out
• Additional attribute has been generated (Labs)
Data mining techniques
• Association
• Classification
• Clustering
• Outlier detection
Association rulesNo Antecedent Confidence Lift
1 Exam 1 1,761
2 Self-evaluation, Exam 1 1,761
3 Labs, Exam 1 1,761
4 Quizzes, Exam 1 1,761
5 Self-evaluation, Labs, Exam 1 1,761
6 Self-evaluation, Quizzes, Exam 1 1,761
7 Labs, Quizzes, Exam 1 1,761
8 Self-evaluation, Labs, Quizzes, Exam 1 1,761
9 Labs, Quizzes, Glossary 0,95 1,677
10 Self-evaluation, Labs, Quizzes, Glossary 0,95 1,677
11 Quizzes, Glossary 0,91 1,601
12 Self-evaluation, Quizzes, Glossary 0,91 1,601
13 Labs, Quizzes 0,88 1,55
14 Self-evaluation, Labs, Quizzes 0,88 1,55
15 Quizzes 0,83 1,467
16 Self-evaluation, Quizzes 0,83 1,467
Classification rules
Clustering
Activity6
Exam25
Self-evaluation
4
Glossary2
Quizzes25
Labs9
Total71
Items
Cluster 0 2.765 20.618 3.806 0.379 21.018 7.059 55.644 17
Cluster 1 0.765 14.412 3.361 0.296 18.021 4.412 41.266 17
Cluster 2 1.200 5.548 3.212 0.140 15.794 3.000 28.843 20
Cluster 3 0.071 1.093 2.788 0.069 9.907 1.607 15.392 14
Cluster 4 0.000 0.000 0.886 0.006 1.750 0.364 2.941 13
Outlier detection
Activity Exam Self-evaluation Glossary Quizzes Labs Total
Cluster 0 6.0 16.5 3.86 0.28 19.0 6.0 51.64
Cluster 0 1.0 25.0 4.0 1.2 23.75 6.0 60.95
Cluster 0 0.0 23.0 3.86 1.28 17.5 3.0 48.64
Cluster 0 5.0 20.0 3.86 0.0 23.75 9.0 61.61
Cluster 2 0.0 15.5 3.14 0.0 13.5 1.5 33.64
Cluster 2 0.0 19.0 3.71 0.0 14.0 6.0 42.71
Cluster 2 5.0 11.0 3.93 1.04 21.8 3.0 45.77
Cluster 2 2.0 13.0 3.68 1.12 21.0 7.5 48.3
Cluster 3 0.0 7.5 2.0 0.0 8.25 1.5 19.25
Cluster 4 4.0 1.5 3.5 0.08 18.0 0.0 27.08
Student comments on newly introduced activities
• Official evaluation
– I liked the labs because they force us to work on new materials continuously
– I liked the labs because they encourage us to exercise regularly
– professors gave us a lot of materials and organized everything perfectly – from labs to supplemental instruction
• Class evaluation
– the labs made us work continuously
– it is a good idea for getting scores, though the evaluation system should be less harsh and give partial credits
– excellent idea set to practice perfectly
– labs helped a lot for continuous engagement
Future work
• Time analysis (self-evaluation)
• Log analysis (forum, laboratory exercises, etc.)
• Classification issues
Conclusion
• We are actually not facing a falling pass rate!!!
Thank You for your attention!