|
|
|
| 2018年度 異文化言語教育評価論 |
Chapter 9
Coming to Terms with Assessment Innovation:
Conclusions and Recommendations
9.5 Where to from Here?
n For the reasons the authors reached, it may
be proposed that interactional evidence required for interact is best
secured through offering students opportunities to engage in a range of peer to
peer interaction throughout the school year.
Ø That is because students can select
evidence of their best performances for summative grading purpose.
n The preceding arguments presents a
convincing case for the usefulness and validity of interact.
n However, there are no easy answers to
resolve how a model of assessment that is built seamlessly within a teaching
and learning program can be used for high-stakes purposes where issues of
reliability and consistency of measurement and accountability are required.
n It is necessary for those who are
responsible for decision-making to recognize the impracticality and challenges
of interact. However, it may mean that New Zealand should return to the
one-time test model.
n Indeed, speaking assessments that focus on
interaction between two interlocutors are now well-established (e.g., Cambridge
English language assessment), indicating that paired/group one-time tests are
sufficient, fit for purposes, and adequately representative of the construct to
measure spoken communicative proficiency.
n In order to ensure the success of interact,
a shift in understanding among stakeholders is required.
n Although impracticality was seen as
comparative disadvantage of interact, the teacher survey comments showed
some evidence to suggest that the benefits of collecting on going evidence.
Ø A suggestion for collecting on-going
evidence is to pass ownership of recording their interaction to the students by
using their mobile phones.
9.6 Recommendations
n The authors advanced the following
recommendations for on-going classroom practice.
1.
Acknowledge
that the ultimate goal is measurement of automaticity with regard to potential
real-world interactions with target language speakers.
2.
Promote
maximal opportunities for the development of automaticity. Students need to be
exposed to as many opportunities as possible to interact with others in
authentic situation.
3.
Provide
feedback on students’ interactions that will help them to enhance their
performances across all dimensions of a spoken communicative proficiency
construct.
4.
Move
towards an understanding of “performance-based” assessment. Students are
assessed as they perform actual or simulated real-world tasks and measured in
the process of performing the targeted linguistic acts.
5.
Measure
performances not only in terms of task completion but also in terms of a clearly
articulated construct. The interest is not the demonstration of sophisticated
grammar or lexis, but rather, linguistic acts to the task.
6.
Interpret
“real-world task” in terms of situational and interactional authenticity. Tasks
that aim to replicate situational authenticity require a dimension of
interaction that moves beyond the rote-learnt and artificial scenarios.
7.
Give
students more ownership of what they want to talk about. Students are able to
engage with the task in ways that enable them to demonstrate the full extent of
their proficiency. Task must be seen relevant by candidates.
8.
Be
mindful of the potentially negative impact of interlocutor variables. It may be
beneficial to allow students to select their own partners.
9.
Be
realistic. Automaticity is not to be equated with perfection. Automaticity will
be determined by appropriate use of language and the ability to sustain the
interaction.
9.7 Conclusion
n The evidence gathered from this project
indicates that, at least as far as the participants are concerned, interact
is working relatively well.
n Teachers perceive interact to be a
significantly more useful assessment than converse.
n They consider interact to be a valid
form of assessment which reflects New Zealand’s revised school curriculum and
its emphasis.
n For students, there is a wide range of
perceptions (positive and negative).
n Open-ended comments reveal perspective that
would suggest that interact would likely be preferred over converse
when it is subject to some modification.
Discussion point
How can we measure appropriate use of language in a
speaking test? Propose a concrete idea with some reasons.