O n f e r e n c e

Session
Group A 
Group B 
Group C 
Group D 
1 
Intervention Text chat 
+ 
Task A 
Text chat 
+ 
Task B 
Face-to-
face 
+ 
Task A 
Face-to-
face 
+ 
Task B 
2 
Follow-up 
Face-to-
face 
+ 
Task A 
Face-to-
face 
+ 
Task B 
Face-to-
face 
+ 
Task A 
Face-to-
face 
+ 
Task B 
3 
Intervention Face-to-
face 
+ 
Task B 
Face-to-
face 
+ 
Task A 
Text chat 
+ 
Task B 
Text chat 
+ 
Task A 
4 
Follow-up 
Face-to-
face 
+ 
Task B 
Face-to-
face 
+ 
Task A 
Face-to-
face 
+ 
Task B 
Face-to-
face 
+ 
Task A 
Table 1: Counterbalancing of task conditions and tasks 
Analysis and results 
Quality of the language produced: Analysis of the quality of the language produced by 
the students during the follow-up face-to-face tasks is in progress. It is common in 
research on task-based language learning to investigate the impact of task variables on 
the CAF of the language produced during tasks. The literature on task-based language 
teaching, however, provides little detail on how these have been measured. The main 
body of this talk will therefore comprise a discussion of the main questions raised 
regarding the measurement of CAF, namely: 

Which measure of the various CAF measures proposed are most appropriate?

How much data should the measures be based on?

When is a pronunciation error a lexical error?

What constitutes a morpho-syntactic error? 

---

-172- 
2014 CALL Conference 
LINGUAPOLIS
www.antwerpcall.be 

When is a pause a pause?

Self-efficacy: The self-efficacy data was submitted to a three-way mixed ANCOVA with 
chat condition as a within-participants factor, the order of presentation of the chat 
conditions and the order of presentation of the tasks as between-participant factors, and 
scores on the c-test as a covariate. No main effects or interactions were found. 
Conclusions 
In contrast with previous research on text-chat mediated task-based language learning 
which found that participating in text chat reduced students’ levels of foreign language 
learning anxiety (Satar & Ozdener, 2008), in this st
udy no difference in students’ levels 
of self-efficacy between the text chat and face-to-face conditions. The absence of an 
effect in this study might be attributed the proficiency of the students - in this study the 
students were advanced learners of English studying abroad interacting with fellow non-
native speakers, while in Satar & Ozdener’s study the participants were high school 
students 
–
and differences in the measures used.
It remains to be seen what the impact of text chat is on oral production. Before this 
analysis can be completed, a valid and reliable approach to coding the data for CAF must 
be established.