A Comparative Study of the Effectiveness of Two Methods of Teaching Resumptive Pronouns in Writing : TBLT and Blended Learning

Resumptive pronouns (RPs) are one of the most challenging grammatical points for EFL learners because this structure is different in their L1. We aimed to examine whether blended learning/TBLT are useful to teach RPs. We examined the extent to which such methods improve performance on the posttest. Forty learnerstook part in the study who were assigned to 2 groups: one group was taught via TBLT, and the other via blended learning. Before piloting the study, the participants were given an OPT to check their homogeneity. Besides,they were given a researcher-made test on RPs to check their knowledge, the result of which indicated that the participants did not have sufficient knowledge about this point. Finally, the participants were given a researcher-made test as the posttest to check the effect of the treatment and the extent to which it was helpful for the correct use of RPs. Findings indicated that TBLT was more fruitful. Findings of the present study have pedagogical as well as practical implications.


Introduction
After the introduction of computers into classrooms, computer-assisted language learning (CALL) has been widely used in various fields to facilitate the teaching and learning of different aspects of L2. (Bahrani, 2011;Pena-Sanchez & Hicks, 2006;Stracke, 2005Stracke, , 2007)).
Task-based language teaching (TBLT) must be sited within the period of the end of the 20th century.It is not a remote or unique language teaching methodology.It can be fully understood if it is been differentiated by preceding methods and analyze them within the conventional communicative methodology.The arrival of TBLT is connected to what became famous as the Bangalore Project (Prabhu, 1987) which started in 1979 and completed in 1984.The word task here refers to the special kind of activities carried out in the classroom which are characterized by the importance assigned to meaning, process, and the content the teaching procedure.
According to Cook (1993), relative clauses are "subordinate clauses that modify nouns within nouns phrases in the main clause above them" (p.138).In the sentence The man who spoke to him went out, the relative clause who spoke to him modifies the noun man in the subject NP the man.The word man, which is so modified, is the head of the relative clause.The relative clause often marks the element that is related to the main clause with a relative pronoun such as who.There are four types of relative clause: relative elements, relative pronouns, relative complementizers, and resumptive pronouns (RPs).
Based on Karimi (2001), there is a particular difference between the wh-movement languages such as English and languages like Chinese and Persian, that is, RC as a RP, is usually placed where the operator or wh-word has moved in English.Thus, one can conclude that, in English, there is no RP in a simple restrictive RC, whereas Persian allows the existence of the RP in the RC (except in subject RCs).That is, in Persian, the reformed noun in the CP is not transferred to Spec-CP and remains in situ.

Review of Literature
L1 writing and L2 learning affect L2 writing mostly, and Kassen (1995) explains the relationship between these three: From the first area, we have learnt to view writing not as pure writing down, but as a complicated interplay of cognitive processes by which writers discover and create meaning (Emig, 1971;Flowers & Hayes, 1981).Based on previous research, L2 learning involves the expression of communicative intent as mediated by various competencies, including grammatical, sociolinguistic, discursive, and strategies (Canale & Swain, 1980;Hatch & Brown, 1995).Different models of language learning like the models based on input (Krashen, 1982;Krashen & Terrell 1983), based on output (Swain, 1985), and based on interaction (Burmfit, 1984) have some essential components, which are meaningful, purposeful, and contextualized language use.The use of technology has also long been introduced to complement traditional writing classes (Chang, 2005;Goldberg, Russell, & Cook, 2003;Shang, 2007).Montali and Lewandowski (1996) found that poor readers not only felt more successful with bimodal presentation, but were more successful in terms of comprehending content.In addition, technology has been used as the complement in traditional writing classes for a long time (Chang, Chang, Chen, & Liou, 2008;Fidaoui, Bahous, & Bacha, 2010;Goldberg, Russell, & Cook, 2003;Liou, Wang, & Hung-Yeh;Shang, 2007).Chang et al. (2008) showed that online assistant facilitates L2 learner-writer collocation use.

Participants
The participants were 40 EFL Iranian learners, randomly selected from different English institutes.Their age was between 15 and 19 years old.According to their class term in the institutes and also according to their scores from their previous term, their English proficiency level was intermediate.

Materials
The first test was the Oxford Placement Test (OPT) to check the homogeneity of the participants, and the second one was a researcher-made test.We calculated the reliability (0.82) and validity of the test prior to the study by asking university professors to judge its validity, which was approved by all.For the task-based group, the material was the same as the other group, but the difference was that, here, the teacher taught the participants the planning task.For the blended learning, the material was partially given through a Weblog (http://learningenglishonline.mihanblog.com) and partially in the class.The participants were to hand in their writings by e-mail.Also, the topics were given through the Weblog.After conducting the study, the researcher-made test was again given to the participants with some minor differences, making it almost impossible for them to use their prior knowledge.

Task-Based Method
For this method, the participants, first, had the researcher-made test.Then, we told them how to do planning before writing a paragraph, and the relative elements and specially the RPs were taught to the participants-there were also some other points for teaching about the paragraph, such as the order of the paragraph, and about how to organize the paragraph in a proper way, that is, using the topic sentence and some supporting sentences and at the end including a concluding sentence.At the end of the term, they all had the researcher-made test again, the aim of which was to analyze and check the effect of the teaching method.

Blended Learning
In this method, we asked the participants to answer the researcher-made test.After that, they were to check the first teaching material on the Weblog.They had a two-day deadline to send their writings to our e-mail.They had the second material presented in the class, and again they had a two-day deadline to send their homework to us.The third part of the lesson was again presented in the class, and a two-day deadline was the same for this time.The last lesson was given through the Weblog, and the participants had the same deadline to send their writings.After the term was finished, the participants were given the researcher-made test again.

Data Analysis
The data from the pretest and the posttest were subjected to statistical analyses to explore the probable effect of the treatment, applying matched t test.There were two paired t tests and three ANOVAs-one ANOVA for the placement test and two for the experiment.

4.1Results of the Pretest
The purpose of the pretest was to make certain that the BLG and TBG were homogeneous at the outset of the study.One-way between-groups ANOVA was used to help achieve the purpose of the pretest (see Table 1).In the BLG, the mean score on the pretest (M = 12.65) is substantially less than the mean score on the posttest (M = 15.70).However, to check the statistical (in)significance of this difference, one needs to consult the Sig.
(2-tailed) column in Table 3.It can be seen that in the TBG on the pretest, the mean score of the learners (M = 13.05) is enormously less than that of the posttest (M = 17.70).To make certain that the difference was, indeed, statistically significant, Table 5 should be consulted:

Discussion
The main result of this study was the impact of task-based instruction on the learning of RPs for the participants.The effect of this method was more than the other experimental group.This is mainly because this method focused on tasks as the building blocks of its teaching/learning exercises and strategies (Nunan, 1989).
This study is in tune with recent studies such as Long (1988) in which he found that tasks focus on specific aspects of language within larger communication frameworks.This way, there would be no decontextualization for impeding the effective learning and the use of particularly focused part of the language in tasks.Krashan (2002) mentions a range of studies (e.g., Day & Shapson, 1991;Harley, 1989;Lyster, 1994;Salaberry, 2000;Van Patten & Sanz, 1995) which claimed that TBLT instruction was far more effective than the usual traditional instruction.
According to Flyman-Mattson (1999), one of the special characteristics of a task is that it focuses on different particular aspects of a language, no matter the feature is grammatical or it is related to vocabulary or discourse functions.She believes that tasks virtually embrace all aspects of a language without over emphasizing one on the other one.Thus, one can make use of tasks for every all of the skills (i.e., speaking, listening, reading, and writing.)

Conclusion
It worth mentioning that whereas blended learning had a great effect on the learning of the RPs, the effect of task-based method was even more noticeable in the posttest.The results imply that although the application of technology (i.e., the Internet and computer) resulted in a better writing performance, the interaction of L2 learners and teachers and the usage of tasks and brainstorming must be taken into account, too.In other words, it is the combination of traditional methodology and using tasks that results in better and more fruitful results.
To conclude, the blended online writing instruction was found to be the effective method for teaching problematic grammatical points to the learners.This finding is in line with other studies on blended learning in different disciplines.Sitzmann et al. (2006) conducted a meta-analysis of 96 experimental studies on online and classroom instruction between 1996 and 2005.They concluded that blended learning optimized the instructional advantages of online learning and classroom instruction.Zhao et al. (2005) found that there was no difference in the overall effectiveness between online and face-to-face learning; they also noted that courses applying blended learning resulted in better learning outcomes than distance or face-to-face education alone.
In the same tune with this study, Kılıçkaya and Krajka (2010) studied the impact of online learning on English vocabulary knowledge of Turkish learners.The learners of the online class were compared to the learners of traditional classes.The results showed that the learners who received online teaching performed better than the other group.The results of the present study are in line with those of this study.Kirkgöz (2011) investigated the speaking ability of learners and teachers of English using face-to-face instruction as well as technology, that is, video.At the end of the experiments, the students' oral communication skills changed significantly.The results of the present study are in line with the results of Kirkoz.
According to Nunan (1987) and Williams (1998), in spite of the positive feedback from the students in successive years in the mentioned studies, it should be noted that computer can never replace the effect of teacher's presence, who is responsible for developing appropriate CALL programs and caring about students' progresses.Based on Richard (1997), teachers, by selecting learning activities, preparing students for new learning, presenting learning activities, asking questions, conducting drills, and checking students' understanding provide opportunities to practice new items, monitor learners' learning, give feedback to learners, and review and reteach when deemed necessary.

Implications
The application of blended online learning, from a pedagogical point of view, presents helpful insights to L2 teachers, learners, and syllabus designers.The results of this study can contribute to a better understanding of the effect of TBLT and blended learning on the learning of some problematic grammatical points.The results will be useful for L2 teachers to prepare the most suitable teaching methods for the learners according to their specific conditions.
The results have practical implications for materials developers to redesign the curriculum to include TBLT and blended learning classes or introduce online CALL materials into the curriculum, according to compatibility of the lesson and the learners because every method has its own benefits and limitations.
This study is also useful for the policymakers because it provided a description of two upper-intermediate level writing classes.Other intensive English programs can determine similarities with their writing classes at the same proficiency level and apply the successful methods mentioned in the present study to achieve better outcomes.

Limitations
Like any other research, this study is not without limitations.First, motivating some students to participate in the virtual class was very difficult.Sometimes, they had problems receiving the materials through the Weblog, and we had to e-mail them individually, and sometimes the e-mails failed.This caused the process of data collection to last longer than we expected.The other limitation was the delay some participants had for e-mailing their homework.This caused some problems in collecting the data.The next limitation was that we did not differentiate the participants' gender.The other limitation was that only RPs were studied.It may be more fruitful if we had analyzed the effect of the blended learning method and TBLT on other grammatical points, as well.
Still, the other limitation was the placement test.Because for the blended learning group, the placement was done online, the process was somehow time-consuming to gather the data, as some participants did not cooperate properly.

Table 1 .
Descriptive statistics for comparing the BLG and TBG pretest scores Results of the First Research QuestionThe first research question was whether blended learning affected Iranian L2 learners' learning of RPs in their writing.A paired samples t test was run to capture any possible difference between the BLG participants' performance on the pretest and the posttest (see Table2).

Table 2 .
Descriptive statistics for comparing the pretest and posttest scores of the BLG

Table 3 .
Results of the paired samples t test for comparing the pretest and posttest scores of the BLG

Table 4 .
Descriptive statistics for comparing pretest and posttest scores of the TBG

Table 5 .
Results of the paired samples t test for comparing the pretest and posttest scores of the TBG