Promoting Interaction in Large Classes with Computer-Mediated Feedback Richard Anderson, Ruth Anderson, Tammy VanDeGrift, Steven Wolfman, Ken Yasuhara.

Promoting Interaction in Large Classes with Computer-Mediated

Feedback

Richard Anderson, Ruth Anderson, Tammy VanDeGrift,

Steven Wolfman, Ken Yasuhara

http://www.cs.washington.edu/research/edtech/

t

Classroom Presenter

Tablet PC-based presentation system– integrates writing on computer-projected slides– separates instructor’s view of presentation

from class view– basis for classroom technology research

t

Context

• University level

• Focus on large classes (> 50 students)

• Computer Science and Informatics

t

Modern Pedagogy vs. Modern Practice

active learning

participatory

interactive

student-directed

lecture

instructor-dominated

passive

disconnected

In the context of current university practice, how can a technological intervention promote interaction in the

classroom?

t

Design Process

1. Discover what inhibits interaction

2. Understand what makes a good design

3. Design intervention

4. Evaluate

t

Inhibiting Factors

Through participant observation, pilot studies, and literature search, identified:– Student apprehension

– Feedback lag

– Single-speaker paradigm

t

Design Goals

• Address inhibiting factors

• Support student-initiated interaction

• Scale to large classes

• Impose low cognitive load

• Exploit existing classroom structures

t

Slides as a Mediating Artifact

t

Slides as a Mediating Artifact

• In the classroom:– facilitates communication– structures discussion

• Outside the classroom:– used as memory aid– used as study guide

• Across terms– reifies of course knowledge

t

Designed System:Classroom Feedback System (CFS)

t• A class’ full name includes its package.

– for example, java.util.ArrayList or java.lang.String

• Often it is more convenient to use the class name without the package, e.g., ArrayList, String

• The import statement tells the compiler where to find class definitions that don't have a complete package name and aren't in the current package– Classes can be imported individually, or all classes in a

package can be imported– java.lang.* is imported automatically by the compiler– is not like #include in C/C++

import statement

• A class’ full name includes its package.» for example, java.util.ArrayList or java.lang.String

• Often it is more convenient to use the class name without the package, e.g., ArrayList, String

• The import statement tells the compiler where to find class definitions that don't have a complete package name and aren't in the current package» Classes can be imported individually, or all classes in a

package can be imported

» java.lang.* is imported automatically by the compiler

» is not like #include in C/C++

import statement

t

Summer 2002 study:Example slide from lecture on Java packages

t

t

t

Evaluation

Intro. programming course, summer 2002:– 150 students total– 12 with laptops– 9 week course, 3 weeks with CFS

Data: observations, surveys, focus groups, interview w/instructor, electronic logs

t

CFS increased classroom interaction

Spoken interaction pre-CFS

Spoken interaction with CFS

Total interactions with CFS

Total w/out “Got it”

# per class 2.4 2.6 15.9 7.9

p-value .91 .04 .14

t

Contributions

• Slide context as medium for interaction

• Designed system

• Evidence of available student feedback

• Successful “round-trip” interactions

• Novel interaction patterns for computer-mediated communication [Anderson et al., CHI 2003]

t

Future Work

• Broader study/deployment

• Support for instructor-planned interaction

• Archival use of feedback

• Support complex feedback

• Scale to more participation

t

Related Work

• ActiveClass [Griswold, CSCL 2003]

• WILD [Roschelle and Pea, CSCL 2002]

• ClassTalk [Dufresne et al., 2000]

• Active learning [Bonwell and Eison, 1991]

• “CATs” [Angelo and Cross, 1993]

t

Acknowledgments

• UW CSE Education & Educational Technology Research Group

• MSR Learning Sciences & Technologies

• Students and instructors from the study

http://www.cs.washington.edu/research/edtech/