Effect of an Online Language Course on Students` Grammar Achievement and Its Relationship with L2 Motivational Self System

: More and more courses are now offering online English classes. However, to date most studies have been limited to the investigation of the implementation of quizzes to improve students` grammar ability, while the use of video tutorials has been less explored. Thus, the current study aimed to investigate how video grammar presentations, in addition to quizzes, could help students improve their grammar achievement. 170 Indonesian students were enrolled in a grammar class as part of TOEFL ITP program conducted on an online language course. This course utilised videos to teach grammar points explicitly and deductively by using students` native language (Indonesian language). The effectiveness of the class was measured based on the results obtained from the pre-test and three progress tests. Each of the progress tests was conducted after a two-week treatment with around five grammar topics each. It was found that students` grammar achievement appeared to improve statistically significantly after receiving the first two-week treatment, yet students` scores were not found to be statistically significant between the progress tests. Thus far, it confirmed that online language programs with video-based grammar instruction as well as grammar exercises could lead students to increased grammar achievement. In addition, taking into account the importance of motivation as a powerful predictor of students` success, this study also investigated the relationship between students` motivation and their grammar achievement. Overall, no association between students` motivation and students` final progress test was found in this study.


INTRODUCTION
In Indonesia, more and more language courses are offering English classes, thereby demonstrating the importance and increasing interest in English as a foreign language (EFL) (Zein et al., 2020).One of the main purposes, among others (e.g., work requirement), to learn English is to study abroad with a scholarship.Most of the scholarship providers in Indonesia require students to take the TOEFL ITP test; thus, a large number of Indonesian students sit for this test each year.Despite the importance, most Indonesians still find the preparation courses for the test expensive, and some of them who are living in remote areas also have a lack of access to these courses.
Thus, the current study aimed to help Indonesian students learn English materials that could better prepare them to sit for the TOEFL ITP test.To achieve this objective, an experimental study was conducted at a free online course in which students were instructed with English grammar (or structure) as one of the tested sections in this TOEFL test.This study was also expected to provide practical evidence for teachers and course designers to make informed decisions to adjust their teaching strategies and materials within the contexts of online language learning programs.
Furthermore, considering the belief in the importance of motivation as a powerful factor to enhance students` achievement, the current study also aimed to investigate the relationship between Dörnyei's (2009) Second Language Motivational Self System (L2MSS) and students` grammar achievement.Thus, the results could provide evidence-based information for teachers to understand the learning motivation of their students within the course.The findings can also contribute to the major pedagogical implication in which it could help teachers in forming the ideal L2 self-motivational support in order to improve students` grammar achievement particularly for online language program contexts.

Computer-based and Online Grammar Instruction
Research has shown that online and computer-based instruction could help students improve their grammar ability.Exposing students` to computer-based grammar instruction, Kılıçkaya (2015) found that Turkish college students` achievement of adverbial clauses improved statistically significantly after receiving a two-week (18 hours) treatment.The teaching materials were mostly in the form of grammar exercises (e.g., fill-in the blanks, multiple choices, matching) with immediate feedback; audio files, graphics and animation were also used to engage students` interest.
Similarly, Torlaković and Deugo (2004) conducted a study for university learners in Canada and found that computer-based grammar teaching could significantly improve students` performance and confidence in using adverbial clauses.The teaching input was also mostly in the form of grammar tasks by which students learnt at their own pace.Students received intermediate feedback during their learning processes.The instruction lasted for two weeks with approximately six hours in total.
Using Moodle, an open-source learning management system (LMS), AbuSeileek (2009) also reported the effectiveness of online grammar instructions either deductively or inductively to improve Saudi Arabian university students` grammar ability over a month.The target structures in this study were simple, compound, and complex sentences; the teaching activities included grammar practices, role plays, and group discussions.With a similar finding, Scida and Jones (2016) also found that online video grammar tutorials and grammar practices could help improve students` grammar achievement; however, it is worth noting that the reported effect size was small.Thus far the literature showed that online language programs could indeed help students improve their grammar ability.It is also shown that most online language courses have focused mostly upon using grammar quizzes to help students` learning, whereas the use of video grammar presentations appears to be less explored.Hence, the current study utilised video tutorials in addition to quizzes to investigate how students` grammar achievement can be significantly improved.

The L2 Motivational Self-system (L2MSS)
In the theory which is put forward by Dörnyei (2009), context and selves are the highlighted components of the L2 motivational support systems.Ideal L2 self represents the individual desires of the l2 learners to become a fluent user of a target language (Higgins, 1987).This factor is highly motivated by internal factor of the L2 learners to be able to reach the expected proficiency level in the L2 language (Dörnyei, 2009).In the concept of L2 motivational selves support system, ideal L2 self has the biggest influence in the successful acquisition of L2 language of the learners (Dörnyei, 2009;Thompson & Vasquez, 2015).This is also consistent with the Self-Discrepancy Theory, suggesting that the L2 learners should have a clear ideal goal and aspiration in order to possess a considerably high motivation in the process of learning L2 language (Higgins, 1998).In that sense, the learners with high ideal L2 self will put maximum effort and initiate the learning activities even without the influences of external factors (Csizér & Dörnyei 2005).Different from L2 ideal self which is strongly affected by internal driving force, the ought-to L2 self is more likely affected by external factors.This motivational framework is mostly driven from the expectation of social environment as the way to prevent the future consequences of not learning the L2 language (Dörnyei, 2009).In response to anti-ought-to self, learners with this kind of motivation develop active resistance towards external pressures or influences in learning a second language (Thompson & Vasques, 2015).That is, the learners are motivated to study L2 language in an attempt to break the social expectations in society (Alharbi, 2017;Lanvers, 2016;Thompson, 2017).For instance, the learners choose to have a degree in language even if it is not lucrative among society, or they choose to study languages which are not commonly studied within the majority of society.

Previous Research on L2MSS
The prominent study on L2MSS carried out by Taguchi et al., (2009) found that L2MSS was not the only factors influencing the L2 learning effort of the Japanese students.Cultural differences and other factors seem to contribute more to the participant's attitudes towards L2 language in this context (Norton, 2000;Dornyei & Ushioda, 2011).An empirical study done by Papi (2010) submitted different findings in which the study found that the L2MSS variables significantly influenced the anxiety of Iranian L2 learners.More specifically, the study coined that ideal L2 self reduces the L2 language anxiety of these participants as opposed to Ought-to-L2 self which increases the anxiety of the participants.This result is consistent with Dornyei's and Ushioda's (2011) argumentation in which ideal L2 self is highly correlated with the hope and expectation of the L2 learners, whilst Ought-to-L2 Self makes the learners have anxious feeling of learning the target language.
In relation to L2 motivational supports and its relationship to language proficiency, Moskovsky et al., (2016) conducted an investigation inviting 360 L2 learners in Saudi Arabia.The study found that L2MSS components were not consistently correlated with the L2 learner's proficiency and achievements or scores.The study further elaborated that, the L2 self-motivational supports does not always predict the scores, proficiency, and behaviour of the students in their L2 language.Furthermore, Lamb (2012) examined the influences of this motivational support to the overall attainment of L2 learners within the Indonesian educational context.This study reported that the L2 motivational systems do not have a strong effect on the proficiency of the learners.However, ideal l2 self was found to have a marginal effect on the participant L2 achievement.In other words, ideal selves have less contribution in stimulating the actualization of learning attitudes (Lamb, 2012).
Another study by Thompson and Liu (2017) found that only ideal L2 self is correlated to language ability, and other two (ought-to and anti ought-to) do not necessarily correspond to the improvement of language ability.Meanwhile, a more specific study on anti ought-to l2 self done by Huensch and Thompson (2017), anti ought-to self is considered to develop pronunciation skills of the participants.From all these mentioned research, there is still inconclusive ideal on the association of this motivational variable to the L2 learner's proficiency and achievement.Contrasting views among researchers in this field further motive this study to be conducted.
To reiterate, the study currently investigated the following research questions: 1. Are there any statistically significant differences in students` achievements after attending an online grammar instruction?2. Does students` grammar achievement correlate with the three L2 motivational systems (ideal L2 self, ought-to L2 self and anti ought-to self)?

Participants
The target population of the current study was Indonesian students enrolled in the TOEFL ITP course for the structure and written expression class at Edinmelb course (kelasdaring.edinmelb.com),an Indonesian private online course.Although the class was free and open to any Indonesian learners.We received a total of 177 students who registered for the class and initially agreed to participate in this research.However, fifty students (17 males & 33 females) completed the course by receiving the approximately six-week treatment and forty-two (15 males and 16 males) submitted the online questionnaire.The average age of the students was 24 ranging from 18 to 40.For ethical reasons, the data of the participants who were under 18 years of age were not included as research data.

Instructional Techniques
This online course utilised videos to teach grammar points explicitly and deductively by using students` native language (Indonesian language).Each video presentation lasted around 10 to 20 minutes and was accompanied by an untimed 10-question exercise.Solutions to each of the exercises were provided also in the form of a video.Thus, students first watched a video presentation on a grammar point, and they completed a related exercise to check their understanding.After completing an exercise, they could access a video that discussed the answers to the exercise; a discussion board was also provided for students to discuss each exercise.
To ensure that students follow the above-mentioned learning process (e.g., watching a video explanation before doing an exercise), a feature of gamification was employed.This gamification feature was provided by Moodle, an LMS that the course was using.This feature allows access restrictions to some activities (e.g., quizzes) prior to the completion of another required activity.Thus, all of these materials (e.g., videos, exercises, discussion boards) were accessed by the students on Moodle.(The sample appearance of the site can be seen in Appendix A.)

Treatment
The first two weeks, after the pre-test, students learnt simple sentences that were divided into five sections (or five video explanations): sentence and verb, preposition, appositive, present participle, and past participle.After they learnt these five grammar points, students took Progress test 1 to check their progress.For the next two weeks, they then learnt compound sentences, adverbial, nominal, and adjectival clauses.Next, they took Progress test 2.For another two weeks again, students were presented with materials on reduced adverbial and adjectival clauses and inversion (e.g., place and negative expressions, conditionals, comparisons).Lastly, they then took Progress test 3.The class duration in total was approximately 2 months; each of the progress tests was open for access within around one week.(The syllabus can be seen in Appendix B.).

Materials
Most of the teaching materials along with the exercises were adapted from Longman Preparation Course for the TOEFL Test by Phillips (2003).The pre-test along with the three progress tests were authentic TOEFL ITP tests for the structure and written expression; thus, each test contained 40 multiple-choice items with a duration of 25 minutes.These authentic tests were adopted from TOEFL ITP Practice Tests and Official Guide to the TOEFL ITP Test, both of which are published by the Educational Testing Service (ETS, 2010 and 2013 respectively).(For the purpose of this assignment, all of these sources were unfortunately used without any prior permission.).

Instruments
Test validity and reliability.All of the tests (the pre-test and three progress tests) that were used within the study were official tests published by the ETS; thus, in that regard reliability and validity then can be ensured.The duration of each of the tests (25 minutes) was also set for this study.L2 Motivational Questionnaire.In this study we measure the learner's motivational factors using a-6 point Likert scale motivational questionnaire ranging from strongly disagree (1) to strongly agree (6) (Appendix C).The first part of the questionnaire included the participant's demographic and educational information which include name, age, institution and educational background.The second part consisted of 30 items which measure the motivational self-system of the participants.The 30 items are divided into three motivational systems which are ideal l2 self (10 items), ought-to L2 self (10 items) and anti ought-to L2 self (10 items).The items for ideal L2 self and ought-to L2 self were adopted from Dörnyei and Taguchi (2010), and the other ten items for anti ought-to self were taken from the interview data of Thomson and Vasques (2015).Since the 30 items are originally written in English, we translate the items into Bahasa as our participants are all Indonesian.To ensure the reliability of the items, we conducted internal reliability.The results of the internal consistency are ideal L2 self (α=.76), oughtto L2 self (α=.87) and anti ought-to L2 self (α=.82).Hence, these scores confirm the internal reliability of the items (Dörnyei, 2001).

Data Analysis
A repeated-measures analysis of variance (or a repeated-measures ANOVA) was computed to compare students` pre-test scores and progress tests.Thus, an ANOVA was employed the first research question on students` achievements throughout the course.Although 50 students received the treatment, the ANOVA results showed only 37 participants` data (N=37); this is because not all students completed the three progress tests.
To investigate the association or the relationship between the L2 motivational self system and the students` Progress test 3 scores, Spearman's rho as a non-parametric test was performed.This test is conducted to figure out as to whether there is a strong or weak relationship between two variables (L2 motivational self system and Progress test 3) of the participants in this study.

Research Question 1: Are there any statistically significant differences in students` achievements after attending an online grammar instruction?
Table 1 presents that students appeared to perform better in all of the three progress tests compared to the pre-test as indicated by mean scores.However, it is also apparent that the mean value of Progress test 2 was higher than the two other progress tests.A repeated-measures ANOVA showed that statistically significant differences were indeed found (F(3,108) = 14.092, p < .001)with a medium size effect (η2= .281).Nevertheless, post hoc comparisons indicated that statistical differences were only found between the pre-test and the three progress tests.In other words, no statistical differences were found between any of the three progress tests.Thus, although students` mean scores at Progress test 2 seemed higher than the two others, the difference could be ignored in this case due to the statistically insignificant differences.
The results showed that students appeared to increase their achievement after approximately two weeks attending the online grammar class.However, students' scores can be said to stay the same over the next four weeks.That is, students` scores in their subsequent progress tests (Progress test 1 and 2) did not statistically differ from their scores in their first progress test.

Descriptive statistics
Table 2 portrays general statistical information of the three L2 motivational systems (ideal l2 self, ought-to L2 self, and anti ought-to self).It is apparent that there is a statistical difference among three L2 motivational factors of the students.The table shows that the students seem to be affected the most by ideal l2 selfmotivational system in the process of acquiring their l2 language (English), with the mean score of 5.37, the highest among other two factors.However, the standard deviation (SD=.513) is bigger than ½ of its mean score, indicating that there is variability among students on this specific motivational factor.In other words, the scores are considered to be more heterogeneous.
As for anti ought-to L2 self, the mean score of this particular motivational support is M= 4.48, the second highest after Ideal L2 self-mean score.The standard deviation (SD=1.05) of this motivational item is larger than the mean, so that shows the scores are spread out or heterogeneous.That is to say, the students have relatively distinct and diverse anti-ought-to self-motivation of learning their second language.Lastly, considering ought-to L2 self, this motivational factor is reported to have the lowest mean score (M=3.97)among other two factors.Furthermore, its standard deviation (SD=.85) is also larger than its mean score, suggesting a lot of variability of the scores.Thus, it implies that the students are affected differently by this motivational system.

Scatterplot Analysis
The following scatterplots demonstrate the correlation between Progress test 3 and the L2 motivational systems (Ideal L2 self, ought-to L2 Self, and anti ought-to self) of the students within this course.The explanation of each scatterplot will cover four general patterns which include direction, form, strength and potential outliners.

Figure 1. Ideal L2 Self Scatterplot
Figure 1 shows a non-linear association between the students` Progress test 3 scores and their ideal L2 self.A weak strength indicates that there is a very smallscale relationship between these two variables.Also, it can be seen that the form is nonlinear with few potential outliners.This suggests that association between these two variables (ideal L2 self and Progres test 3 score close to zero relationship.This further can be seen from the linear value (R2 = 0.016).Therefore, this motivational factor is not associated with Progress test 3 of the students based on the presented scatterplot.

Figure 2. Ought-To L2 Self Scatterplot
Figure 2 shows weak and non-linear association between the two variables with few potential outliners in the provided figure 2. Thus, from this scatterplot, it can be concluded that ought-to L2 self-motivational factor and progress self are negatively associated with each other.In addition, the linear value (R2 = 0.019) further tells that there is no association between these two variables (ought-to self and L2 motivational support).Principally, if the linear value is closest to zero, meaning that there is a high potential of no association between variables.

Figure 3. Anti-Ought-To Scatterplot
Figure 3 again indicates no relationship between X (anti ought-to L2 self) and Y (Progress 3) as there no clear pattern or form appeared in the Figure .3. In this vein, the association between these two mentioned variables closes to zero relationship with R2 = 1.31.In that sense, anti ought-to L2 and Progress 3 score do not relate to each other.

Correlation coefficient
To further verify this result, the spearman's rho correlation coefficient between Progress test 3 score was conducted.In regard to Ideal L2 self, it is reported that there was no correlation between the ideal L2 self and Progress test 3 score, with rho=20, n=41, p =.19.It can be seen that the p-value of this particular L2 motivational support indicates no statistical significance between the two variables.Concerning ought-to self, there was also no significant correlation between Progress test 3 and ought-to L2 self, with rho =.-161, n=41, p=.31.Similarly, anti ought-to self was found to have no significant correlation to Progress test 3 score (rho =0.03, 41, p=.81), showing no relationship between the two variables.Moreover, the p-values of ideal l2 self (p=.19), ought-to self (p=.31) and anti-oughtto self (p=.81) are all larger than .05,indicating no statistical significance.This further tells that there is no correlation between students` scores in Progress test 3 and all the three variables: ideal L2 self, ought-to L2 self and anti ought-to self.

DISCUSSIONS
Based on the statistical results, it was shown that students` scores appear to significantly improve after attending the course for two weeks.Thus, this study can be said to be consistent with previous studies (e.g., AbuSeileek, 2009;Kılıçkaya, 2015;Scida & Jones, 2016;Torlaković & Deugo, 2004), all of which showed that online language programs could be useful for students to improve their grammar ability.
This study also supported the findings by AbuSeileek (2009) who showed that a deductive approach to teaching grammar online could lead to an increased grammar achievement particularly for more complex target structures.Like Scida and Jones (2016), the current study also showed that video presentations could help students learn English grammar more effectively; thus, the use of videos for online language courses should be more explored.
Because of the relatively small size of participants, the findings of the current study should be interpreted with caution.This is particularly true due to the fact that this study lacked a control group.Although the reported size effect was medium, whether it was the treatment that significantly affected students` performance cannot be ensured.In other words, students` lower scores in the pretest could be because of their initial unfamiliarity to complete timed quizzes on an LMS.Thus, this could also apply to the interpretations of previous studies (e.g., Scida & Jones, 2016) that lacked control groups and particularly with a small effect size.
Nevertheless, it is also important to note that the current study utilised proficiency tests (or the TOEFL ITP) as the progress tests rather than achievement assessments that are directly related to the grammar points that participants learnt.Had we used achievement tests, we could have expected a higher size effect.
With respect to the questionnaire results, each of the three variables (ideal L2 self, ought-to L2 self and anti ought-to self) was found to have no correlation with the students` scores on Progress test 3.This finding is inconsistent with Thomson and Liu (2017) showing that the ideal L2 self might possibly predict the language proficiency of the L2 learners.The current result supports the study of Lamb (2012) that demonstrates that the ideal L2 is not always found to have statistically significant relationship with the language proficiency of the L2 learners.Moskovsky et al., (2016) also put forward the notion that ought-to L2 self and anti ought-to L2 self are less likely to correspond to language proficiency as opposed to ideal L2 self which is found to have marginal effect to language proficiency.Based on the findings of the current study, it can be concluded that internal factors are not the only factors essential for the L2 learners in the process of learning their L2 language.

CONCLUSION
Upon the first research question proposed in this research, it was found that students` grammar achievement appeared to improve statistically significantly after receiving a two-week treatment.Thus, it confirmed that online language programs with video-based grammar instruction as well as grammar exercises could lead students to increased grammar achievement.
The study has also discussed the association between the L2 motivational self systems (Ideal L2 self, ought-to L2 self and anti ought-to L2 self to Progress test 3 score of the participants.It is found that there is no association between the L2 motivational self systems and Progress test 3 of the students in this study.This is also true for the ideal L2 self which has no association with the language score of the participants.

Table 1 .
Participants' performance on the pre-test and progress test.

Table 2 .
Descriptive statics of the three L2 motivational systems.