Open-endedness in CALL presentation

Open-endedness in CALL
Richard Watson Todd
King Mongkut’s University of Technology Thonburi
©2006 Richard Watson Todd
Open-endedness in teaching



Open-ended = range of possible appropriate
responses
Closed-ended = only one correct answer
Open-endedness can be applied to:
– teacher questions
– classroom activities
– methods of assessment

Traditionally closed-endedness has
dominated
©2006 Richard Watson Todd
Advantages of open-endedness


Little evidence that open-endedness leads to
language learning
Open-endedness:
– promotes thinking skills
– promotes learner autonomy

As teachers, we should probably try to use
open-ended approaches where feasible
©2006 Richard Watson Todd
CALL


Common addition to classroom learning
Variety of types of CALL:
–
–
–
–
Computer-mediated communication (CMC)
The Internet
Generic software programs (e.g. Word)
Multimedia CALL software
©2006 Richard Watson Todd
Multimedia software




Heavily promoted multi-million dollar
industry
Closed-ended, relying on multiple-choice,
matching, gap-fill
Very weak when compared to teaching
Need for open-endedness in multimedia
software
©2006 Richard Watson Todd
Making CALL more open-ended


Problems with open-endedness in
computing e.g. giving feedback
4 approaches in this paper:
– CMC (e-mail lists and bulletin boards)
– Supplementing closed-ended software with
user-user interaction
– AI chatterbots
– Website for analysing texts of the user’s choice
©2006 Richard Watson Todd
Approach 1: CMC


Main approach to open-endedness in CALL
The study:
– compare use of e-mail lists and bulletin boards
– students from classes in Thailand, China, Japan,
Australia
– student contributions collected and analysed for
length, complexity and functions
©2006 Richard Watson Todd
Approach 1: CMC

Findings
– Both e-mail and boards are open-ended
– Student contributions have similar levels of
syntactic complexity
– Student contributions for e-mails are much
longer
– Student e-mails contain a wider range of
functions

E-mail lists may be preferable to bulletin
boards
©2006 Richard Watson Todd
Approach 2: Supplementing closedended software

The program
– Noughts-and-crosses (tic-tac-toe) game
– Multiple-choice questions
– 3 versions
• 1. One player against the computer
• 2. Two players at one computer (talk to each other)
• 3. Two players over a network (with chat facilities)
©2006 Richard Watson Todd
©2006 Richard Watson Todd
Approach 2: Supplementing closedended software

The study:
– 12 students used all versions
– Program use was recorded
– Questionnaire about preferences distributed
©2006 Richard Watson Todd
Approach 2: Supplementing closedended software

Findings
– Version 2: Two players at one computer (talk to
each other)
• More interesting than play with computer
• It’s very fun because I can play the game with my
friend and I can see everything
– Version 3: Two players over a network (with
chat facilities)
• I think this version work, I like it
• OK! It’s very fun too because I can play the game
©2006
RichardIWatson
with my friend
and
canToddchat too
Approach 2: Supplementing closedended software

Findings concerning interaction
– Version 2: Two players at one computer (talk to
each other)
• Dominated by reading aloud and use of first
language
– Version 3: Two players over a network (with
chat facilities)
• Syntactically very limited
• e.g. “Who are you?” “Hello” “OK” “he he he” “Ah”
©2006 Richard Watson Todd
Approach 3: AI chatterbots




Chatterbots simulate a conversational
partner
Not genuinely intelligent, but simulate
intelligence
ELIZA
FRANK, updated psychiatrist
©2006 Richard Watson Todd
Approach 3: AI chatterbots

The study:
– 5 students’ conversations with 3 chatterbots are
recorded
– Students were interviewed for reactions
©2006 Richard Watson Todd
©2006 Richard Watson Todd
Approach 3: AI chatterbots

Findings
– Some students engage in open-ended discourse
– Others give up quickly
– One student:
• “I really like these programs. I don’t like Internet
chat rooms because if you want to talk about your
problems, no one wants to listen and you can’t speak
strongly. These programs listen to anything you
say.”
©2006 Richard Watson Todd
Approach 4: Website for analysing
texts of the user’s choice

Multimedia software reading:
– Texts provided
– Closed-ended exercises (e.g. multiple-choice)
– Program controls content

Website analysing texts of user’s choice
– Texts chosen by users
– Closed-ended exercise (identify keywords)
– User controls content
©2006 Richard Watson Todd
Approach 4: Website for analysing
texts of the user’s choice

How to use the website
–
–
–
–
–
User pastes a text into the website
Basic statistics provided
Exercise on identifying keywords
User identifies words to learn
Gap-fill test for chosen words generated
©2006 Richard Watson Todd
©2006 Richard Watson Todd
©2006 Richard Watson Todd
©2006 Richard Watson Todd
©2006 Richard Watson Todd
Approach 4: Website for analysing
texts of the user’s choice

The study:
– 24 students used the website
– Student use was monitored
– Students completed questionnaires
©2006 Richard Watson Todd
Approach 4: Website for analysing
texts of the user’s choice

Findings
– On average, 3.35 keywords identified per text
– Problems with students choosing unusual words
(e.g. labview) for vocab test
– 23 students said they would use the website
again
– Quotations
• “It good program. I think it can help me very much
to learn the reading story.”
• “This website is good for people who want to study
©2006 Richard Watson Todd
English by yourself.”
Conclusions



For all 4 approaches, users react positively
to open-ended CALL
But only evidence that open-ended CALL is
preferable comes from Approach 2:
Supplementing closed-ended software
Nevertheless, CALL software developers
should consider including open-endedness
©2006 Richard Watson Todd
Directions for CALL software

Integrating CMC into software
– Benefits of open-ended CMC
– Need for interlocutors
– Most applicable for networked software to be
used by classes
©2006 Richard Watson Todd
Directions for CALL software

Integrating chatterbots into software
– Need to include Spellcheck/Grammarcheck
– Need to identify specific content/functions

Allowing users to select input
– Need to expand analysing algorithms
– Most applicable for networked software

Time for CALL software to progress to the
next stage. Open-endedness provides a
©2006 Richard Watson Todd
direction for progress.