"A multifaceted approach to investigating pre-task planning effects on paired oral test performance" is a paper written by Ryo Nitta and Fumiyo Nakatsuhara in the year 2014. The paper explores the effect of planning time on oral test task performance using a multifaceted approach. Most of the earlier studies looked at this issue by considering the performance of a group of test-takers as a whole. The problem with such an approach is that the fine differences between test-takers' performances, and the differences within particular test-taker's individual and collaborative performance will be lost. Ryo and Nakatsuhara's approach makes sure that intra-test-taker differences in performances are noticed too.
This study looked at 32 foreign language learners' performance on decision making tasks under planned and unplanned conditions. The study used rating scores, discourse analysis and conversation analysis to understand co-constructed performance apart from a questionnaire to understand test-taker attitudes to planning time. Conversation analysis provided valuable insights and implications for teaching and testing.
In teaching research, planning time is seen beneficial because it cognitively helps the limited attentional capacity/resources of the learner. Planning time activates rule based system. Therefore while performing the task, the learner can use the preactivated rule system, and concentrate on the use of memory based system. It also encourages learners to access explicit analytic knowledge since automatised implicit knowledge is lesser available during live performance. Nature of planning, task type, proficiency level of learners, etc. affects performance. Generally, longer planning time is found to be beneficial to fluency, but lesser useful for accuracy and complexity in teaching research.
Testing Research
In standardised testing, planning time is provided mainly for the sake of fairness. That is, to control the level of cognitive demand imposed by potentially unfamiliar topics and enabling test takers to produce their best performance.
The use of unguided planning for shorter periods in tests has found mixed results in earlier research. The limited effects observed might be due to the high stakes nature of testing context, which focuses test-takers' attention on accuracy. This results in careful online planning. Thus the possible effects of planning might be overridden, says Ellis in his edited book "Planning And Task Performance In A Second Language (2005). Measurement methods used in such research might have influenced results.
Dialogic Tasks
The nature of task used is very important. There are monologic and dialogic tasks. They differ vastly. Monologic tasks do not have an interlocutor. Talking to a microphone in response to a voice prompt or written question is very different from interacting with a live interlocutor in person. They both use very different performance processes. In a monologue, test-takers use their own resources. They solve problems on their own. They construct whole/entire performance. In a dialogic task, the discourse is co-constructed. Language, ideas, vocabulary, constructions, etc. are exchanged. The process is constantly open and is dependent on both (or all) parties involved in the dialogue.
Galaczi in his paper titled Peer–peer interaction in a speaking test: The case of the First Certificate in
English examination" published in Language Assessment Quarterly, 2008 says that there are three patterns of interaction in pair discussions. They are, collaborative, parallel and asymmetric interaction. Collaborative pattern has exchange of roles of listener and speaker. They support each other's topic and develop each other ideas. Parallel pattern involves each one developing their own argument. There is less agreement on each others' ideas. Asymmetric pattern involves unbalanced contributions. One person takes secondary role while the other speaks the most. It is observed that in tests, it was collaborative pattern that received the highest scores, and parallel pattern, the least.
Paired oral tasks are designed to measure interactional competence. But usually in research, we only look at cognitive complexity of different tasks, and linguistic demands of task design without attention to the 'co-construction' aspect of interaction. It is important to know how pre-task planning affects interactive patterns of dialogue because the kind of interaction pattern in a task has important implications for validity of the test.
Multifaceted approach
Performance of particular test-taker could vary at different times within a task. But previous studies have looked at the collective performance of both/all parties involved in the task, assuming uniformity of performance throughout the task. This is not fair to the actual way performance takes place. Interactions involve non-linear processes as this study clearly shows us.
This study is process-oriented (not summative). It studies differences and similarities in performance processes of interactions under different planning conditions. It used two decision making tasks. Tasks were something like this: which item in the picture is important for a happy life? Choose the most important two of them. Rating scales from previous researches were modified with extra bottom points to include the current subjects' performance. Fluency, accuracy, complexity and interaction were considered as targets to be measured.
Results
Score analysis showed that there was a slight upgradation of fluency and complexity with planning time condition.
Discourse analysis showed that there was an improvement in breakdown fluency and longer turn length with planning time condition. Also, planning time reduced the speed of fluency (number of words per minute).
The increase in complexity under planned conditions might be due to presentation of planned language during planning stage. Such increase in complexity was observed only in the beginning of the task. During the advanced stage of the task performance, complexity levels were low. Also, the pattern of interaction was parallel. That is, each one gave their points without building up on each others's points. The utterances were longer, but more like a monologue instead of dialogue. There was no co-construction. When what is planned during planning time is exhausted, the discourse fell into a stagnant period.
But negative planning time condition led to gradual increase in turn lengths, incorporating each other's points collaboratively.
Planned interaction ended clumsily which contrasted with unplanned interactions. Planned interactants only expressed their individual ideas during their turns.
Implications for Teaching and Testing
Use planning time for clear purposes, as planned by the task-designer. For example, It is better not to use planning time for tasks that are aimed at developing interactional competence.
Pre-task planning before a pair task is not advisable as it might change the interactional pattern of a task.
We ought to reconsider the duration of test tasks especially in pair oral tasks. As we have seen in this study, when planning time is provided, interaction became collaborative only after a while into the interaction. Therefore, we must reflect whether the given task performance time is sufficient to reach that threshold point where performance can become collaborative.
There is scope to wonder if provision of planning time in oral pair tasks is wise at all. But further research is necessary before we reach any such conclusion.
The paper is available for download HERE from Sage Publications' website.
No comments:
Post a Comment