The Effects of On-line and Pre-task Planning on Descriptive Writing of Iranian EFL Learners

The aim of this paper was to investigate the effects of pre-task planning (PTP) and on-line planning (OLP) on descriptive writing of EFL learners. It is discussed extensively that planning factor influences task performance of language learners. Thirty seven learners of English as a foreign language, aged between 19 and 24, were recruited and randomly assigned into two groups with pre-task and on-line planning conditions. The participants in PTP group were given 10 minutes to plan their performance before the main task performance, while the participants in OLP group had to begin writing immediately but they could take as much as time they liked. The collected data was then coded and analyzed using independent samples test statistical procedure. The results of data analysis showed planning time had no effect on complexity and accuracy of participants’ performance but it influenced positively the fluency of PTP group.


Introduction
As early as 1970s, the communicative language teaching (CLT) approach became popular among second language acquisition (SLA) researchers and second language teachers (Skehan, 2003).Task-based language teaching (TBLT) is a realization of communicative language teaching.It is indeed the strong version of CLT, as tasks provide the foundation for an entire language program (Ellis, 2004).Foster and Skehan (1999) state that there are some pre-mid-and post task activities that can be utilized to help language learner pay a balanced attention to both form and meaning simultaneously and improve the quality of learner language.Planning is one of the task condition factors that affects second language production and has been of both theoretical importance to second language acquisition (SLA) researchers and of practical importance to language teachers (Ellis, 2005).Due to the importance of task planning factor in learners' task performance there have been plenty of studies that have focused on the interaction of planning and task performance of language learners (Ellis, 1987;Foster and Skehan, 1996;Ortega, 1999;Robinson, 1995;Skehan andFoster, 1997, 1999;Yuan and Ellis, 2003;Mochizuki and Ortega, 2008).The studies mentioned above mainly concern oral task performance and only a few studies have concentrated on written task performance (e.g.Ellis and Yuan, 2004).This fact highlights that further studies are needed to explore written performance of learners and task planning in a more extensive area and on different task types.Consequently this research focused on finding the effects of pre-task and on-line planning on descriptive writing of language learners, since no study has touched this writing task type before.

Literature Review
Variety of design factors (such as planning) and how they influence the language produced by learners regarding accuracy, complexity and fluency have been main focus of studies of many researchers (Ellis, 2009;Foster and Skehan, 1999;Housen and Kuiken, 2009;Wigglesworth and Storch, 2009).To make the significance of planning in the field of SLA understood Ellis (2005) states that even the language that seems to be effortless and naturally occurring involves planning.He believes that "planning is essentially a problem solving activity; it involves deciding what linguistic devices need to be selected in order to affect the audience in the desired way" (Ellis, 2005: 3).Planning whether pre-task (before the task performance), on-line (during the task performance) rehearsal (doing the task or a similar task before the main performance of the task) provides chance to attend to language as form (Ellis, 2005).Planning conditions can also be classified differently besides pre-task and on-line planning categorization.They can be divided into guided planning, in which learners are guided in the planning phase about what and how to plan, and unguided planning, in which learners receive no guidance or advice in the planning phase (as investigated in studies by Hulstijn and Hulstijn, 1984;Foster and Skehan, 1996;Foster and Skehan, 1999;Sangarun, 2005).Besides these, Whalen and Menard (1995) discussed planning that takes place in discourse level or in the aspects of text constructions that learners attend to.In addition to clarifying types of planning in SLA research, the theories and models that inform processes at work in planning procedure should be taken into account.The theoretical structures and models that are directly or indirectly related to study of task planning in second language acquisition are Tarone's (1983) account of stylistic variation, Levelt's (1998) models of speech production and Kellog's (1996) model of writing (Cited in Ellis and Yuan, 2004: 62), and Robinson's (2001) and Skehan's (1998) cognitive models of L2 performance and language learning.The mentioned theories and models explicitly or implicitly draw on three central constructs involved in psycholinguistic accounts of language processing -attention and noticing, a limited working memory capacity, and focus-on-form.Since 1980s SLA researchers have investigated the effects of planning on language performance and language acquisition (e.g.Ochs, 1979;Wigglesworth, 1997;Foster and Skehan, 1999;Ellis and Yuan, 2004;Kawauchi, 2005).These studies have had different foci and have addressed the issue of task planning from different perspectives but most of them have investigated oral production of learners and have concentrated on exploring the effects of planning on accuracy, complexity, and fluency of language production.Findings of majority of these studies have shown clear effects for planning on complexity and fluency of learners' language (Foster and Skehan 1996;Skehan and Foster, 1997;Wigglesworth, 1997;Mehnert, 1998;Ortega, 1999;Foster and Skehan 1999;Yuan and Ellis, 2003;Ellis and Yuan, 2004;Kawauchi, 2005) but findings in terms of accuracy have not been homogenous.Ellis (2004) believes these mixed findings are due to learners' difference in orientation towards accuracy and their proficiency level, and also due to different task types and particular grammatical features used in the studies.Some of the studies have reported positive effects on accuracy (e.g.Mehnert, 1998;and Kawauchi, 2005) but it was not supported in studies by Yuan and Ellis, (2003) and Ellis and Yuan (2004).

Research Question
What are the effects of pre-task and on-line planning on accuracy, complexity, and fluency of Iranian EFL learners' descriptive writing?

Research hypothesis
There is significant difference between the performance of pre-task and on-line planning groups in terms of accuracy, complexity, and fluency.

Participants
In this study thirty seven participants took part.They all were sophomore students of English language and literature in Tabriz University.They were randomly distributed in two groups based on their scores in their writing course that they all had recently passed in a way that the average score of groups were the same.Azerbaijani Turkish was the dominant mother language among the participants and they aged between 19 and 24 years old.The number of participants in OLP group was 20 of which 13 were females and 7 were males.There were 17 participants in PTP group, 13 females and 4 males.

Materials
The task material employed in this study was the topic "Describe one of the nationwide ceremonies or festivals in your country".The selection of this topic as task material was based on criteria and definitions mentioned for descriptive essay by Langan (2000) and Meyers (2006).And it was also based on the fact that participants had shared background knowledge about the topic proposed as there are some nationwide ceremonies common to all people in Iran (e.g.Norouz Eve).3.5 Variables of the study a) Independent variables: In this research planning condition was the independent factor which was manipulated by length of planning time and by the time of occurrence (before or during task performance).b) Dependent variable: Written performance at three levels of accuracy, complexity and fluency.

Accuracy, complexity, and fluency measurements
In order to find whether there were any identifiable differences in the texts written by the learners in OLP and PTP groups, the texts were analyzed to determine the number of T-units, independent clauses and dependent clauses.The term T-unit was coined by Hunt (1965).He defined T-unit as the "shortest grammatically allowable sentences into which (writing can be split) or a minimally terminable unit".Dependent clause is defined as the clause that must be used with another clause to form a complete grammatical construction and it depends on the other clause.And independent clause is defined as the clause that can stand on its own (Richards and Schmidt, 2002).a) Accuracy measurement: The accuracy of written performance of the participants was measured by calculating the percentage of error free T-units to the total number of T-units (the same measurement used by Rahimpour (1997, 2008), Errasti, 2003;Larsen-freeman, 2006;Storch, 2009;Storch and Wigglesworth, 2009;Rahimpour and Nariman, 2010;Rahimpour and Hosseini, 2010).b) Complexity measurement: Complexity of the texts produced by the participants in this research was measured by calculating the percentage of dependent clauses to total number of clauses (following Storch and Wigglesworth, 2009).c) Fluency measurement: Fluency was measured by the number of the T-units per text.This was the same procedure followed by Storch and Wigglesworth (2009).Sentence fragments were not taken as T-units (Ishikawa, 2006;Foster and Skehan, 1996) and also backchannel cues such as hmm and yeah and discourse boundaries markers such as ok, thanks, or good were not counted as T-units following young (1995).

Procedure
In PTP group students were given 10 minutes to plan their performance and 17 minutes to commit the task.They were required to produce at least 200 words to reduce opportunities of on-line planning.Following Ellis and Yuan, (2004) no detailed guidance was provided but participants were told to plan their writings in terms of content, organization, and language.At first papers on which only the topic was written were distributed among students and they started taking notes to get prepared to do the main writing task.The participants in this group were asked not to write the complete text in those papers as they would be taken away after 10 minutes.They were asked to plan what they would write during the main writing task on the papers given first.After 10 minutes the papers given first were collected and second papers on which task topic and instructions were written were distributed.Then they were given 17 minute to write the essays.On the other hand, the participants in OLP group were given papers on which topic and instructions were written and they were asked to start writing immediately to eliminate the possibility of PTP.No detailed guidance was given but the students were asked to plan their writing in terms of content, organization, and language.They were told that there would be no time limitation for them.So the participants in this planning condition were not provided with pre-task planning time but ample time to plan their writings during their main performance.The start and finish times of OLP group were recorded to make sure that they spent more than 17 minutes (i.e. the time allocated to PTP group) to write their essays).

Results
Table 1 provides the descriptive statistics for accuracy, complexity, and fluency of L2 learners' descriptive writing in pre-task and on-line planning conditions.Insert Table 1 Here As the table shows mean of OLP group (X=0.73482) was higher than that in PTP group (X=0.72048) in the case of accuracy.In terms of complexity mean of PTP group (X= 0.32278) was greater than that in OLP group (X= 0.26442).And finally mean of fluency of PTP group (X= 15.8824) was higher than mean fluency of OLP group (X= 11.8000).Table 2 demonstrates the results of independent samples t-tests for accuracy, complexity, and fluency of L2 learners' descriptive writing in pre-task and on-line planning conditions.Insert Table 2 Here As the table shows, since the significance level of independent samples test was not higher than the alpha level of 0.05 in the case of accuracy and complexity it was concluded that the there was no significant difference between the performance of OLP and PTP groups.Conversely, regarding fluency, the significance level of independent samples test (0.008) was lower than the alpha level of 0.05.It means that planning had influenced only the fluency of descriptive writings of language learners.Thus our the research hypotheses that "there is significant difference between the performance of pre-task and on-line planning groups in terms of accuracy, complexity, and fluency' was confirmed only in the case of fluency of participants written performance and not in terms of its accuracy and complexity.Figure 1 shows the means of accuracy of L2 learners' descriptive writings in pre-task versus On-line planning conditions.As mentioned before, the mean of accuracy of OLP (X=0.73482) was larger than the mean of PTP group (X=0.72048),although it was not statistically confirmed.Insert Figure 1 Here Figure 2 presents means of complexity of L2 learners' descriptive writings in pre-task versus on-line planning conditions.As can be seen, mean of complexity of PTP group (X= 0.32278) was greater than that in OLP group (X= 0.26442), although this difference in mean value was not significant.Insert Figure 2 Here Figure 3 demonstrates means of fluency of L2 learners' descriptive writings in PTP and OLP groups.As it is shown the means of two groups were significantly different (PTP group,X= 15.8824 and OLP group,X= 11.8000) and this difference was statistically confirmed.Insert Figure 3 Here

Discussion
The results showed no significant difference between accuracy and complexity of pre-task planning and on-line planning groups.However, there was statistically significant difference between fluency of descriptive writings of two groups and PTP group outperformed OLP group.The results of this study in terms of accuracy showed that participants have used their time to focus on content (Sangarun, 2005) and also this result can be accredited to lack of guided planning (Mochizuki and Ortega, 2008).The finding in terms of complexity revealed that pre-task planning group used the time to focus on propositional content and identifying the main and on-line planning group spent their time finding suitable lexical terms and presumably to encode temporal and modal meanings (Ellis and Yuan, 2004).Besides, this result can be due lack of instruction in operationalizing this experiment (Foster and Skehan, 1996;and Skehan and Foster, 2005) and also the emphasis in Iranian context on formal grammar teaching (Rahimpour and Nariman, 2010).In the case of fluency the greater fluency in PTP group is because pre-task planning facilitates process and text planning for content and organization and it may help in increasing L2 writers' confidence in their ability to write clearly and effectively and, for this affective reason, it may reduce their need to engage in extensive monitoring (Ellis and Yuan (2004) based on Kellog's model of writing).

Conclusion
The findings of this study showed that planning influenced the fluency of the descriptive writing of language learners but it did not have any influences on accuracy and complexity of their language production.In the case of fluency pre-task planning group outperformed on-line planning group and regarding accuracy there was not a statistically significant difference between the performances of two groups and also the syntactic complexity of two groups was not different.These findings can be considered as partial support to Skehan (1998) who claimed that trade-off effects are likely to occur between different aspects of language production as a result of human's limited attentional resources (i.e.learners are not able to pay a balanced attention to different aspects of language simultaneously) .At the same time these findings can be in odds with Robinson's (2001) Cognition Model in which he claimed attentional resources are not limited and learners like native speakers have the capacity to attend to more than one aspect of language simultaneously.

Implications
The findings of this study can contribute to task-based teaching methodology since planning is considered as one of the task implementation factors that can be manipulated by giving chance or not providing time for planning, offering different types of planning to the learners in task performance, and providing learners with various lengths of planning time and planning effects can be observed in the performance of language learners (Ellis, 2009).These findings may also add to the present literature in SLA theory, language testing, syllabus design, and material development.

Figure 1 .
Figure 1.Mean differences in terms of accuracy of OLP and PTP groups

Table 1 .
Descriptive statistics for learners' descriptive writings Accuracy

Table 2 .
The summary of data analysis to compare the accuracy, complexity, and fluency of pre-task planning group and on-line planning group