Planning oral narrative tasks : optimizing strategic planning condition through strategy instruction

This article presents the results of a master thesis, which aimed at investigating the impact of strategic planning instruction on the speech performance of 6 L2 Brazilian learners. The participants, Letras-Inglês students, performed three now-and-there picture-cued narrative tasks under three different conditions: (1) no planning, (2) planning before instruction, and (3) planning after instruction. In addition, the participants filled in post-task questionnaires after the performance of each task, aiming at understanding their opinion on the conditions and tasks. Quantitative and qualitative analyses were conducted in order to examine participants’ oral production and perception, respectively. In general, there was no statistical evidence supporting the impact of instruction on participants’ oral planned performance; however, some statistical results approached significance, which may suggest some positive effects. Qualitative analyses provided positive evidence of the impact of strategic planning instruction on participant perception and their use of strategies during planning time. Moreover, the results of this study can contribute to the fields of Second Language Acquisition and Language Pegadogy.


Introduction
Within a task-based perspective, providing learners with the opportunity to plan their oral task prior to the actual task performance is a pre-task condition called strategic planning (Ellis, 2005).The impact of strategic planning on second language (henceforth L2 1 ) oral performance has been 1 In this study, L2 will be adopted as a general term, defined as "[…] a cover term for any language other than the first language learned by a given learner or group of learners, irrespective of the type of learning environment" (Sharwood-Smith, 1994, p. 7).
researched for the last two decades, and results have shown positive evidence regarding the benefit of the pre-task condition to the improvement of different speech dimensions such as fluency, complexity, and accuracy (Skehan, 1998).Furthermore, strategic planning is viewed as a promising construct because, in addition to being used for research and theorybuilding purposes, it can be manipulated pedagogically.Thus, it opens discussion for the possibility of an interface between theory, research and practice (Ellis, 2005).
Aiming at understanding the impact of strategic planning on overall L2 oral performance, studies have focused on exploring: the role of different types of tasks and task complexity (Foster, 1996;Foster & Skehan, 1996, for instance), the impact of the amount of time available for planning (Mehnert, 1998, for instance), the relationship between strategic planning and working memory capacity (Guará-Tavares, 2008;2011;2016), the relationship between strategic planning and other performance conditions (D'Ely, 2006), the strategies used by learners while they plan (Ortega, 1999(Ortega, , 2005;;Guará-Tavares, 2008, 2016), the benefits of planning collaboratively (Xhafaj, 2014;Xhafaj, Muck, & D'Ely, 2011), and the role of familiarity with task topics and strategic planning (Gavin, 2014;D'Ely, 2011, respectively), to cite but a few.All of these factors seem to cause an impact on learner oral performance, favoring different speech dimensions.
According to Skehan (1998), oral performance has been considered a multifacetated phenomenon, and it has been divided into three different dimensions: fluency (the capacity to produce speech in real time), accuracy (the ability to perform in the target language forms), and complexity (the use of more elaborated and complex language structure).Fluency is the most affected dimension when students are provided with some time to plan their tasks (as can be seen, for instance, in Foster & Skehan, 1996;Mehnert, 1998;Ortega, 1999), while there seems to be a trade-off effect between complexity and accuracy, being the latter the less affected speech dimension.
Notwithstanding the positive evidence, studies have shown mixed results in relation to the dimensions that are affected when the opportunity for planning is given.Some studies, for instance, have found that strategic planning has an impact on fluency and complexity (Yuan & Ellis, 2003, for instance), while some have shown no impact on any dimension whatsoever (D'Ely, 2006, for instance).These mixed results may be explained under some grounds.To start, learners' attention resources are limited (Schmidt, 2001); therefore, learners may select the dimensions they will direct their attention to while planning.Moreover, this selection may be guided by the type of task the participant will perform (Skehan & Foster, 1997).Finally, students may not be familiar with the planning condition, and they may not take advantage of this time to plan their tasks properly, which, consequently, would not affect their oral performance positively (D'Ely, 2006;Ellis, 2009;Mehnert, 1998).This may occur because students do not know the strategies they can use during the planning time, or simply because they are not strategic planners.Such fact may also be explained by the complex nature of planning, a problem solving activity involving other minor activities as highlights Guará-Tavares (2008): When planning an oral task, learners need to activate task-relevant information, maintain them activated and accessible until this information can be integrated to subsequent information in a coherent way; learners also need to sustain, maintain, and switch attention from the various components of the task (e.g., from meaning to form and vice-versa), suppress irrelevant L2 and L1 information, and monitor (p.180).
Bearing in mind (i) the existence of trade-offs, (ii) the complexity of the process of planning as regards problem solving when planning and retrieving pre-planned ideas on-line, and (iii) the learners' lack of familiarity with the planning time, it is possible to state that strategic planning per se may not be enough.That is, giving learners the opportunity to plan produces positive impact on their speech performance, but it does not seem to improve all the speech dimensions.
Considering the ideas above and the lack of research concerning the use of instruction to enhance strategic planning as a pre-task condition, this article intends to present the results of a master thesis, which aimed at investigating the extent to which instruction sessions on how to plan cause an impact on learners' accurate planned oral performance in L2.The goals of the instructional sessions were threefold: a) to raise learner's awareness on strategic planning; b) to make them familiar with strategic planning; and c) to assist learners in becoming strategic while planning.Moreover, the study also aimed at establishing learners' perception on the impact of the strategic planning instructional sessions on their oral performance.
In order to achieve these objectives, the following research questions guided the study: a) Does strategic planning per se produce an impact on students' accurate oral performance on there-and-then narrative tasks 2 ?b) Does strategic planning after an instructional period produce an impact on the students' accurate oral performance on there-and-then narrative tasks?c) Can strategic planning instructional sessions assist learners to become more strategic when they plan?d) What are the students' views on the instructional sessions and the strategic planning process?

Participants
Following other studies on oral production (D'Ely, 2006;Guará-Tavares, 2016;Ortega, 2005 andRossi, 2006 to cite just a few), the six participants in the study had an intermediate level of proficiency in English, which was controlled through prior observation classes and informal conversations with the students and their English teacher.All of the participants were Brazilian undergraduate students from a Letras-Inglês program of a university in the state of Paraná, aged from 19 to 23.They were fictionally named as: Carla, Daiane, Hugo, Mauro, Márcia and Tatiane.Furthermore, the decision of choosing teachers-to-be as participants was twofold: firstly, it was observed a difficulty in developing oral production in this specific program, in which students graduate having deficient oral proficiency in the language they will teach3 , and secondly, studies in the area using this type of population are rare.

Narrative tasks and their conditions
The participants performed three narrative tasks.The use of narrative tasks seems to be popular for research purposes (D'Ely, 2006;Skehan & Foster, 1997;Yung & Ellis, 2003).Also, the tasks were monologic, that is, students did not need a partner to perform them.This choice was made due to the purpose of the study, because if the participants had the opportunity to listen to their peers' story, this could be an intervening factor, affecting the results.Moreover, monologic tasks are more cognitively demanding (Kawauchi, 2005) since they elicit linguistically denser talk (Bygate, 1999), fact that meets the study's purpose.The narrative tasks were picture-cued with no written story.In order to control for task effect, the tasks shared a similar topic: relationships; however their stories were different.Each task was performed under a condition: (1) no planning time, (2) planning time before instruction; (3) planning time after instruction.In every condition, in order to increase the task complexity, the participants had fifty seconds to see and understand the story, having no access to the story afterwards.In the first condition, the participants narrated the story right after the fifty seconds, and in the second and third conditions, the participants were provided with ten minutes to plan their stories, which is the regular amount of time provided in most studies.During the planning time, participants received a sheet of paper as a draft for planning purposes; however, they were not allowed to use it during the task performance, otherwise they could read the information in the draft, which is not the purpose of an oral task.At the end of the ten minutes, the participants narrated the story.Participants' narratives were recorded individually, and later they were transcribed and used for quantitative analysis mostly.

Instructional sessions
In order to optimize the participants' planning time, instructional sessions were designed based on the six most reported strategies by Guará-Tavares' (2008;2016) participants: writing/outlining/summarizing, rehearsal, lexical search, elaboration, organizational planning and monitoring, and also a communication strategy -paraphrasing -which was added after the pilot phase of the study.A total of four instructional sessions were administered.The first session occurred one week after the participants performed the first two tasks, and its main objectives were to raise students' awareness in relation to the benefits that strategic planning could have in their oral performance and to present strategies they could use while they plan.The strategies were divided in the other three sessions, when the participants had the chance to practice each of the strategies by doing activities.At the end of each session, the participants were asked to reflect upon the strategies they practiced in order to understand whether they would be useful to them.

Post-task questionnaires
Right after performing each task, participants were asked to respond a questionnaire, whose aim was to unveil their perception on the task performance, planning time (for tasks 2 and 3), task difficulty, and also the process they engaged during planning time and task performance.The answers of the questionnaires were used for qualitative analysis.

Researchers' diary
During the instructional sessions and the data collection, a diary was kept in order to take notes of aspects that could assist data analysis.

Analysis
As already mentioned, the data was analyzed both quantitatively and qualitatively.Regarding the qualitative analysis, the narrative transcriptions were analyzed in terms of percentage of error free clauses, and four statistical treatments were adopted.First, a descriptive analysis was conducted in order to provide an overall picture of the group's performance in the three conditions.Second, in order to understand individual performance and also explain and discuss the results, the gain scores of each participant in relation to each pair comparison was identified.Third, Friedman's test was run to check for significance among the three task conditions.And forth, in order to see whether there was significance between pair task conditions (Task 1 -Task 2; Task 1 -Task 3; Task 2 -Task 3), Wilcoxon signed-rank tests were applied.The probability level of p <.05 was used to determine statistical significance; however, considering the last statistical treatment, as there was the need to run three pairs of comparisons, the probability level of p was adjusted to .0167, using Bonferroni correction (i.e., dividing .05by the number of conditions) (Larson-Hall, 2010).In what concerns the qualitative analysis, the answers from the questionnaires were tabulated and compared in order to grasp participants' perception.

Descriptive Statistical Results
The descriptive statistics are presented in Table 1.The table provides the minimum and maximum scores (which are the lowest and highest scores of the participants in the group), the mean performance of the group (i.e. the sum of the participants' scores and the division by the total number of participants), and also the standard deviation (which is the variation in the participants' scores).Considering the mean number for each task, it is possible to see that the group produced more errorfree clauses in Task 3, that is, about 50% of their clauses did not contain any grammatical error.This was expected since this task was performed under the more enhanced condition -the participants received instructional sessions on how to plan and, after that, they had the opportunity to plan this task.In a speculative manner, this result can tentatively suggest that instruction on planning led participants to focus their attention on accuracy, and, therefore, they produced more accurate outcomes.On the other hand, the outcome of Tasks 1 and 2 presented similar mean values, which may suggest that only strategic planning itself does not seem to present any impact on accuracy, considering the researched sample.
The standard deviation was high in every task, which indicates that there was a great variation in the scores of the participants in the tasks.This difference in the participants' performance scores may be due to a number of reasons: (1) different levels of proficiency in English, which was not controlled by placement tests in the study; (2) the impact of trade-off effects among different dimensions of speech performances (Foster & Skehan, 1996).The participants may have improved in terms of fluency and/or complexity at the expense of accuracy4 ; and (3) the participants' anxiety5 in performing the task, which was reported by some of the participants.The participant that got 7.2% of error-free sentences in Task 2, for instance, was trembling while performing Task 1.In fact, some studies point out that anxiety may negatively affect learners' performance (Ellis, 1998).
The individual scores of the participants, listed in Table 2, show a confusing scenario if task pair comparisons are made.From Task 1 to Task 2, and from Task 1 to Task 3, only two participants (Carla and Marcia) produced more accurate outcome in Task 2 compared to Task 1.The other four participants (Daiane, Hugo, Mauro, and Tatiane) produced more accurate outcome in Task 1 when they did not have time to plan.The same phenomenon occurred between Task 1 and 3. Three participants produced more accurate outcome in Task 3 compared to Task 1.The situation seems to be more consistent between Task 2 and Task 3. Five participants produced more accurate outcomes in Task 3. The participant who did not present a more accurate performance in Task 3 compared to Task 2 was Carla; however, the difference is only of 5.8%, which is low.In short, the results seem to favor the performance in Task 3, taking into consideration that the group presented more error-free clauses in this task, and the participants seemed to have improved from Task 2 to Task 3. Thus, receiving instructional sessions and having the opportunity to plan a task seems to have led the group to produce more accurate outcomes.

Friedman's Test
In order to assure that the difference in Table 1 is statistically significant, a statistical test called Friedman was employed.This test was chosen considering that the same group of students performed three different tasks under three different conditions and that their scores differed from each other, not providing normal distribution for the data.The test compares the mean score of the tasks and reports whether there was a significant difference between them.As presented in Table 3, the comparison between the three tasks was not statistically significant (p = 0.311).In other words, the performance improvement of the group in relation to accuracy could have been due to chance alone.This may be further explained by the small number of participants and the different performances of the participants in the tasks.Even though the result of Friedman's Test did not present statistical significance when comparing the three tasks together-which means that no significance would be expected in the results from Wilcoxon Signed-Ranks Test (Field, 2009), the test was run in order to scrutinize whether the results from the comparison between each task pair approached significance.Unlike Friedman, the Wilcoxon Signed-Ranks Test reports the significance of each task pair comparison.Its only problem, though, is that, because it is comparing more than one pair of tasks, its number of significance is divided by three.Therefore, the pvalue of each comparison is not 0.05 anymore, but 0.0167.
Table 4 showed that no pair task comparison presents a significant difference, given that the pvalue for this test should be equal or smaller than 0.0167 in order to be statistically significant.However, it is worth noticing that the results from the comparison between Tasks 3 and 2 approached significance at p = 0.075.This result is based on the positive ranks, which represent the participants that gained in accuracy from Task 2 to Task 3.Each participant represents approximately 16.7% of the group; therefore, a negative impact of a unique participant may have diminished the chances of having a statistically significant impact.Probably if the number of participants was higher, significance would be detected.Nevertheless, almost the whole group improved in accuracy from Task 2 to Task 3, which did not occur in the other pair tasks.This fact might partially indicate that the optimization of performance condition through instructional sessions plays a role and positively impacts participants' accurate performance.For a more consistent claim about the positive role of the instructional sessions to have been achieved, the comparison between Tasks 3 and 1 should have presented a higher level of significance as well.However, this lack of consistency between Tasks 3 and 1 may be explained by the role of familiarity.The participants were not familiar with narrative tasks when they performed Task 1, which may have influenced their performance.
These results are partially consistent with those of Sangarun (2005), and D'Ely (2011), who provided some type of instruction to their participants-either as metacognitive guidance, or teacher-led planning.The instructional sessions led the participants' attention to focus on meaning and form, which may have assisted them in planning their tasks strategically.In consequence, the processing in the conceptualizer and/or formulator may have been reduced, freeing up their attentional resources and allowing the participants to monitor their language structures on-line; thus producing accurate language.
To sum up, the quantitative analysis of this study did not show any significant impact on the accurate oral performance of the group regarding the task condition imposed.That is, providing the opportunity for strategic planning (Task 2) and teaching students how to plan (Task 3) did not affect students' accurate performance significantly compared to Task 1, in which students did not have time to plan.However, comparing Task 2 and Task Acta Scientiarum.Language and Culture Maringá, v. 39, n. 2, p. 203-212, Apr.-June, 2017 3, significance was approached, which might indicate at some level that the optimization of strategic planning through instructional sessions plays a role and causes an impact on participants' accurate performance.
The Qualitative Analysis of Data The impact of instructional sessions on strategic planning The analysis of question 4 from Post-task questionnaire 3, in which the participants were asked to report their opinion on the instructional sessions and whether or not they assisted them in doing Task 3, reveals that the sessions were perceived as positive by all the participants.They reported that the instructional sessions assisted them in learning new strategies that, in turn, were useful while they were planning their tasks.For instance, Carla stated that the instructional sessions showed her other ways to plan what she was going to say6 .Consonantly, Daiane said 'I used the strategies when was planning what to say'.Besides that, the instructional sessions provided the participants with the opportunity to reflect upon the strategies they were using and to try new strategies that could be more beneficial when planning their performance, which was the case of Mauro.In Task 2, he said he wrote down the entire story he would narrate and ended up forgetting parts of the story.This can be perceived in his voice when he says: Mauro: (The instructional sessions) helped me because in the previous activities I wrote the entire story and I ended up not remembering everything.With 'organizational planning' (the strategy) I believe I improved a lot.
In addition, it is possible to notice, in the participants' drafts of Task 2, that they all wrote down the entire story, and they complained having trouble remembering the story on-line and being nervous because of that.For a narrative task in which the participants had to tell a relatively long story, 'writing the entire story' as a strategy does not seem beneficial, considering that our attentional resources are limited (Schmidt, 2001).When performing the task on-line, which requires a lot of attention (Levelt, 1989), the participants could not direct enough attention to remember everything they wrote, as emphasized by Tatiane and Mauro.
Tatiane: I planned my story and when I had to tell it I got confused.
Mauro: I wrote one thing, but when I told the story I added other things, eliminated others, in sum, it was different from the draft.
Still regarding strategies, the participants had to report the strategies they used while they were planning Tasks 2 and 3.In Post-task questionnaire 2, the most cited strategy was 'writing'.Some participants reported using organizational planning (Tatiane: I tried to create a story compatible with the images), lexical search (Daiane: I reflected about vocabulary choice), and paraphrasing (Carla: I substituted the words I didn't remember).In posttask questionnaire 3, the participants mentioned using the seven strategies presented to them in the instructional sessions.
Even though, the participants reported the use of some strategies in Task 2, it is possible to notice in their answers that they were not strategic at using them.They did not use them as metacognitive strategies, because they had not reflected upon them before the instructional sessions.Hugo, for instance, reported using 'organizational planning' in Task 2; he also reported that the 'organizational planning' he was presented to and practiced in the instructional sessions was the most useful strategy to him.The same applies to 'writing'; all the participants used 'writing' in Task 3, but not as in Task 2. They used it as outlining and summarizing; therefore, the strategy 'writing/outlining/summarizing' in this study is separated into two qualitatively different strategies: 'writing' and 'outlining/summarizing'.
Another way to unveil the impact of the instructional sessions is by analyzing the participants' evaluation of their own performance in each task.The analysis result of the post-task questionnaires indicates that the participants in general felt more comfortable with their performance in Task 3 compared to the two previous tasks, as it is possible to see in Table 5. Almost all the participants seemed to perceive an improvement in their last oral text compared to the previous ones.Daiane's opinion was the only one that did not corroborate the others', and it is worth mentioning that this participant, at first, did not want to do this task, but encouraged by her peers changed her mind.Therefore, she was probably not motivated and engaged in the task.According to Ellis (2009), the learner's attitude toward the task is a variable that contributes to his/her perception and performance of the task.
In short, it is possible to see that the instructional sessions assisted learners in becoming more strategic when they plan.The instructional sessions assisted the participants in reflecting upon the strategies they can use.They used the strategies they learned as reported by them in the Post-task questionnaires.In addition, the participants seem to have become more confident about their oral production.In the next section, the analysis of the participants' perception of the process they underwent is presented.

Students' perception on the process
In some questions of the post-task questionnaires and interview, the participants were encouraged to provide their opinion on the process they were exposed to, and from the analysis of the participants' answers, it seems that having the opportunity to plan and being instructed on how to use the planning condition more strategically were positive for all of them.
Regarding Task 1, which was used by means of control, the participants reported that it was the first time they had to do a now-and-then narrative task, and some of them classified the task as being fun and challenging, but also scaring.In general, the speaking activities they were used to doing in the classes were related to answering questions or giving their opinion about some topic, and not actually performing tasks.During the class observation period, we could observe some speaking activities.The teacher gave some time to the students to plan their oral performance.All of the students wrote what they would say, and when performing the activity, they ended up reading aloud what they had written, which would not be considered an effective speaking activity.
In the second trial, the participants had the opportunity to plan their tasks and most of them viewed this condition as positive.The participants claimed that it was the first time they were exposed to this condition -at least a pre-task condition in which they did not have access to the notes they wrote while planning -, and that they could use this planning time to organize what they would say.Mauro was the only participant that stated that having time to plan was not beneficial for him, because he did not implement what he had previously planned on-line.This may indicate that he did not know how to use the planning time strategically (D'Ely, 2006;Ellis 2009), leading him to forget what was previously planned.Regarding the opportunity to plan Task 3, which occurred after the instructional sessions, all the participants claimed that they felt more comfortable to plan because they could apply the strategies they learned in the instructional sessions.
In relation to the interviews, a link between the experience of learning how to plan and real life situation was brought by the participants.They claimed that this process assisted them in becoming more strategic while speaking in general, as can be perceived in the voices of Carla and Marcia Carla: I've learned better how to organize my way of speaking, what to say when I have to tell someone something.
Marcia: The experience we underwent was very important, considering that we didn't know the speaking strategies.Little by little, it is very rewarding to apply these strategies and be able to English it up.
The strategies they learned in the instructional sessions seem to assist them in becoming aware of their own speech production, which may cause an impact on their speaking skills.
In conclusion, the process to which the participants were exposed seems to have been positive and beneficial for them.The next section addresses and answers the research questions of this study, and brings a general discussion of the results.

General discussion
Research question #1: Does strategic planning per se produce an impact on students' accurate oral performance on there-and-then narrative tasks?
Concerning the results of the quantitative analysis, there is no evidence that strategic planning itself produces an impact on students' accurate oral performance.The results of Wilcoxon Signed-Ranks tests showed no statistical significant differences in comparing the participants' scores from Task 1 and Task 2, which indicates that having time to plan a task does not cause an impact on accuracy.However, this does not mitigate the relevance of strategic planning, since this lack of impact was expected given other results in the area (D'Ely, 2006;Foster & Skehan, 1996;Mehnert, 1998).Such results demonstrated how accuracy tends to be the less impacted dimension (Ellis, 2005), which may occur due to trade-off effects (Foster & Skehan, 1996).The participants may have directed their attention to other speech dimensions: fluency and/or complexity at the expense of accuracy.
The fact of having time to plan but not knowing how to use it strategically may also have contributed to the lack of impact on accuracy (D'Ely, 2006;Ellis, 2009).While planning the task, the participants may have used their attentional resources to focus on conveying the story, and no attention was left to monitor the grammatical structures while telling the story.
Research question # 2: Does strategic planning after an instructional period produce an impact on the students' accurate oral performance on thereand-then narrative tasks?
Concerning the quantitative analysis, the analysis results of Wilcoxon Signed-Ranks Test showed that when comparing Tasks 1 and 3, and Tasks 2 and 3, no statistical significance is reached; however, statistical significance was nearly approached in the comparison of Tasks 2 and 3, a fact that might partially indicate that strategic planning instruction plays a role and positively impacts participants' accurate performance as in D'Ely (2011).It was also expected a more consistent difference between Tasks 1 and 3 to raise a stronger claim regarding the efficiency of the instructional sessions.However, task familiarity seems to have played a positive role on the performance of Task 1, which might have influenced the participants' outcome in this task.
In general, the instructional sessions provided the participants with strategies that allowed them to focus on meaning and form, as in Sangarun (2005), using the planning time more strategically.This may have assisted the participants in reducing the processing in the conceptualizer and/or formulator setting aside more attentional resources to focus on not making mistakes.
Research question #3: Can strategic planning instructional sessions assist learners in becoming more strategic when they plan?
One of the reasons for providing instructional sessions on strategic planning was the assumption that only providing time for planning a task is not enough (D'Ely, 2006;Ellis, 2009), considering that the participants may not know what to do while they plan or that they may even be unfamiliar with the strategic planning condition.The instructional sessions would raise the participants' awareness and open a space for practicing these strategies so they could reflect upon them and fortunately become more strategic when planning their oral performance (Oxford, 1989).
Through qualitative analysis, it was possible to conclude that teaching the participants how to plan can assist them in becoming more strategic when they do so.The instructional sessions provided room for reflection, in which the participants engaged themselves in metacognitive processing.Moreover, the participants became more comfortable with the task (Ortega, 1999;2005) and had the opportunity of getting familiar with and practicing strategies that can be used not only when planning a task, but also for speaking skills in general.These results do not corroborate Kellerman's assertion (1991, as cited in Cohen, 1998) that claims it is not useful to teach students strategies because they already know how to use them from their L1.The participants were already acquainted with some strategies, but considering the experience derived from the instructional sessions they could reflect upon their use and learn how to apply them more strategically.
Research question #4: What are the students' views on the instructional sessions and the strategic planning process?
Through the analysis of the post-task questionnaires and interviews in which the participants were encouraged to share their opinions on the tasks, conditions, and instructional sessions, it was possible to notice that the process as a whole was positive for the participants.They claimed that the instructional sessions assisted them in becoming more strategic while planning their speech, not only for tasks, but for speaking in general.
In brief, the quantitative results of this study did not indicate any statistically significant impact on accurate oral performance.Nevertheless statistical significance was approached when students' performances from Task 2 to Task 3 are compared.Thus, it is possible to say that the instructional sessions on strategic planning seem to present positive effects regarding cognitive variables.There seems to be evidence that teaching the participants how to plan may direct their attention to focus on form while planning, which may free up their attentional resources to monitor grammatical structures on-line.From the qualitative analysis, it was possible to see an impact on affective and metacognitive variables.The participants became more confident and comfortable with planning their speech, engaging themselves in metacognitive processes that gave them the opportunity to reflect on strategies they already used, applying them more strategically, and also learning new strategies.

Final considerations
In conclusion, the study, whose objectives were to investigate the impact of strategic planning instruction on learners' accurate planned oral performance and to establish their perception on the whole process, has brought some evidence for the positive effect of teaching how to plan.Findings have contributed to our understanding of the impact that a more enhanced strategic planning condition may play on oral performance at the accuracy level.Nevertheless, it must be borne in mind that much more research should be conducted on the subject in order to scrutinize the constructs dealt with in this study.This study had some limitations such as (1) number of participants; (2) lack of a formal level of proficiency control; (3) lack of a control group; (4) measures for accuracy; and (5) lack of other speech dimensions (fluency, complexity and weighted lexical density, for example).It would be recommended to control for these issues in order to have a more robust picture of the results in future research.
Considering the classroom, the idea of providing learners with strategies they may use during the planning time seems to be suggestive.Nevertheless, it is important for teachers to analyze their teaching and learning contexts previously in order to verify whether or not the inclusion of such intervention would be adequate.Foster (2009) reminds that even though research on strategic planning attempts to build an interface between theory and practice, tasks used in research are more controlled than the ones used in the classroom, for instance.It would be interesting to have a more pedagogical stance toward this issue, in which strategic planning instruction would be conducted within a real classroom.In this way, more direct claims for teaching could be made.

Table 2 .
Individual scores of the participants in the three tasks.

Table 3 .
Friedman's Test -Comparing the three experimental conditions.

Table 5 .
Participants' evaluation on the oral text they produced.