Experiences with Active and Collaborative Learning

1
Experiences with Active and Collaborative Learning

• PACISE 2005
• Bloomsburg University
• Tom Briggs
• thb_at_ship.edu

2
Introduction

• Active and Collaborative Learning
• Students interact directly with material
• 70 increase in long-term retention
• Reduced drop out rate

3
Active Learning in CS

• Commonly reserved for intro courses
• Used for solving a problem or writing code
• CS is a contact sport
• What about upper division courses?

4
Felder-Silverman Learning Styles

• Four groups of learning styles
• Active and Reflective
• Sensing and Intuitive
• Visual and Verbal
• Sequential and Global
• Identifies preferred learning style

5
Active vs. Reflective

• Active Learners
• Prefer concrete knowledge
• Hands-On
• Interactive
• Reflective Learners
• Prefer abstract concepts and theory

6
Sensing and Intuition

• Sensing Learners
• Learning facts and concepts
• Intuitive Learners
• Prefer possibilities, applications and
  relationships

7
Visual and Verbal

• Visual Learners
• Prefer visual representation of material they can
  see
• Charts, graphs, figures
• Verbal Learners
• Prefer words, either spoken or written

8
Sequential and Global

• Sequential Learners
• Follow material in a step-by-step sequence
• Global learners
• Need material in the context of its domain
• Need to understand the relationships between new
  and old material

9
Computer Science Students
10
CS Students Preferences

• The demographics
• 83 of CS students are visual learners
• 55 were active
• Implications to in-class time

11
Simple Active Technique

• Active does not imply difficult
• Ask students to interact with material
• Solve a problem
• Sketch a proof
• Trace a section of code
• Ask students to break into groups
• Set a time limit to complete task
• Call on a few groups to share solutions
• Ask students to judge goodness of solutions

12
What does this do?

• Breaks the sequential flow of a lecture
• Students interact with material and their peers
• try out their conceptual understanding of the
  material
• Get immediate feedback
• Students get a break from information assault
• Provides time for students to cognitively process
  knowledge

13
Active Learning in CS

• Most literature addresses intro course
• Usually describes using code review / peer
  programming
• Advanced Courses ?
• Frequently taught as
• abstract facts / theory courses
• straight lectures

14
Operating Systems

• Background
• Introduction to operating system concepts
• Juniors and Seniors
• Taught two sections (¼ 30 each)
• Four credit course

15
Worksheets

• Frequently made use of in-class worksheets
• Example
• Computing system utilization with I/O
• Described the context of the equations (global)
• Relationship of equations (intuitive)
• Guided exploration, concrete examples (active)
• Worksheet Lecture slides (visual)

16
Hypothesis Testing

• Abstract material difficult for students
• Ad-hoc, instructor lead demonstrations
• Example mmap( ) system call
• Student question provoked discussion (open/fopen)
• Lecture / slides put aside
• Lead class to develop hypothesis
  (global/intuitive)
• Created code to test hypothesis (visual)
• Students helped prof look up system calls
• Students ran program (truss) (active)

17
System Calls

• In previous example, truss command

int main(int argc, char argv) int x FILE
fp int in open("test.c",S_IREAD)
read(in,x,sizeof(in)) close(in) fp
fopen("test.c","r") fread(x,sizeof(int),1,fp)
fclose(fp)
gcc test.c o test truss ./test
open("test.c", O_RDONLYO_NOCTTY)
3 read(3, "inc", 4)
4 close(3)
0 brk(0)
0x86da000 brk(0x86fb000)
0x86fb000 brk(0)
0x86fb000 open("test.c", O_RDONLY)
3 fstat64(3, st_modeS_IFREG0600,
st_size292, ...) 0 mmap2(NULL, 32768,
PROT_READPROT_WRITE, MAP_PRIVATEMAP_AN, -1, 0)
0xb75f0000 read(3, "include \ninclude
0 munmap(0xb75f0000, 32768)
0
18
Student Perception

• Threading and context switches
• Class Lecture
• Read discussed theory (sensing)
• Discussed different OS implementations
  (intuition)
• Students challenged which is faster
• Small group discussions (active)
• Lead to develop hypothesis to test
  (active/intuition)
• Out-of-class assignment
• implement test, collect results, submit graphs
  (visual)
• small groups compare (varied) results (active)
• group presents one set of results

19
Evaluation Synthesis

• Active environment
• Peer review of work
• Challenge pre-conceived beliefs
• Evaluate goodness of results
• Hypothesis testing leads to new results
• Evaluation Synthesis
• Highest levels of Blooms taxonomy

20
Computer Organization

• Assembly Programming, CPU Architecture, ILP,
  Memory, IO
• Sophomores with CS1 CS2 experience
• Taught two sections (¼ 20 each)
• Four credit course

21
Differences from OS

• Students lack extensive background
• Most had CS1 CS2
• Discrete Math, some Prob. Stat.
• Sophomores
• Relied on more structured / guided activities
• Fewer expectations of independent thinking
• Computer Organization
• Use of simulators and counters
• Focus on architecture
• Closer to familiar hardware

22
Worksheets

• 15 worksheets
• Guided students through various activities
• Deriving and using Amdahls Law
• Observing and computing speed-ups
• Researching processor specifications
• Building assembly programs
• Use simulators and counters to observe machines

23
Simulators

• Simulators
• software to simulate a physical system
• SimpleScalar tool chain
• Free MIPS R4000 32-bit simulator
• GNU C compiler and binutils (cross compiler)
• Different execution models
• Simple, no ILP
• Pipeline, no cache
• Pipeline and cache, in-order execution
• Pipeline, cache, and speculative execution (ROB)

24
Counters

• Counters
• Machine status registers
• Pentium (RDMSR/WRMSR)
• UltraSPARC v8, v9 CPU control masks
• Software configures events
• Track execution of real program on real hardware
• Stochastic element

25
Worksheets
Section from a worksheet (white space / student
response fields omitted)
26
Inconsistent Results

• Simulators little variance
• Single thread of execution
• Not simulating entire system
• Counters high variance
• Context switching and interrupts
• Process affected by external events
• Inconsistent / surprising results challenge
  students expectations

27
Active Learning

• Hypothesis testing (e.g. best cache) (active)
• Data collection forecasting (visual/active)
• System comparison (active/global/intuitive)
• Tracing execution on simulator (visual)
• Assembly programming / registers (visual)

28
Conclusions

• Active learning
• Effective in upper division theory courses
• Engaged and challenged students
• Appealed to a range of learning styles
• Did not require significant preparation overhead

29
Hypothesis Test Explain

• Instructor guides students to
• Understand problem
• Develop hypothesis
• Identify tests to prove/disprove hypothesis
• Execute tests and collect results
• Explain results
• Support hypothesis
• Develop new hypothesis to explain inconsistent
  results
• Students
• Exposed to the science in Computer Science
• Cognitive process challenged and reinforced

30
Results

• First offering for courses
• Initial exam scores were generally good
• Student feedback on end-of-term surveys
• Overall very positive / higher than department
  college averages
• Over 60 listed worksheets and class
  discussions as most positive aspects of course
• Comments

31
Take Home Points

• Active environments
• Challenge students knowledge
• Move students higher in Blooms taxonomy
• Improve student comprehension and retention of
  material
• Provide another vehicle to assess student
  comprehension
• Do not require sophisticated or overwhelming
  class preparation
• Promote faculty role as leader/guide

32
Bibliography

• (see paper for in-text citations)
• 1 Owen Astrachan, Concrete teaching hooks and
  props as instructional technology, ITiCSE '98
  Proceedings of the 6th annual conference on the
  teaching of computing and the 3rd annual
  conference on Integrating technology into
  computer science education, ACM Press, 1998, pp.
  21-24.
• 2 R. M. Felder and R. Brent, Learning by doing,
  Chem. Engr. Education 37 (2003), no. 4, 282-283.
• 3 Scott Grissom and Mark J. Van Gorp, A
  practical approach to integrating active and
  collaborative learning into the introductory
  computer science curriculum, Proceedings of the
  seventh annual consortium on Computing in small
  colleges midwestern conference, Consortium for
  Computing Sciences in Colleges, 2000, pp. 95100.
• 4 Lewis E. Hitchner, Judith Gersting, Peter B.
  Henderson, Philip Machanick, and Yale N. Patt,
  Programming early considered harmful, SIGCSE '01
  Proceedings of the thirty-second SIGCSE technical
  symposium on Computer Science Education, ACM
  Press, 2001, pp. 402403.
• 5 SimpleScalar LLC, Simplescalar 3.0.
• 6 Jeffrey J. McConnell, Active learning and its
  use in computer science, ITiCSE '96 Proceedings
  of the 1st conference on Integrating technology
  into computer science education, ACM Press, 1996,
  pp. 5254.
• 7 K. Silverman R. Felder, Index of learning
  stylels, World Wide Web., February 2005.
• 8 Lynda Thomas, Mark Ratclie, John Woodbury,
  and Emma Jarman, Learning styles and performance
  in the introductory programming sequence, SIGCSE
  '02 Proceedings of the 33rd SIGCSE technical
  symposium on Computer science education, ACM
  Press, 2002, pp. 3337.
• 9 Henry M. Walker, Collaborative learning a
  case study for cs1 at grinnell college and
  austin, SIGCSE '97 Proceedings of the
  twenty-eighth SIGCSE technical symposium on
  Computer science education, ACM Press, 1997, pp.
  209213.