The Relationship between Writing Self-efficacy and Writing Performance of Iranian EFL Students

A strong sense of self-confidence in the writing task is referred to as writing self-efficacy. There is a discrepancy between the scholars’ views regarding the relationship between writing self-efficacy and writing performance. Therefore, this study aimed to investigate if there is any significant relationship between the students' writing selfefficacy and their writing performance. Also, there was an attempt to see if there is a gender difference in the students' writing self-efficacy and their writing performance. Convenient sampling was used to select 59 medical students (28 males and 31 females) from among medical university students. The instruments used in this study were Writing Selfefficacy Questionnaire and students’ writing compositions. To analyse the data for both objectives, Pearson productmoment correlation was used. The results of this study showed that there was not any significant relationship between the students’ writing self-efficacy and their writing performance. It was also found out that there was not any significant relationship between our male and female participants’ self-efficacy and their writing performance.


Writing self-efficacy and writing performance
Self-efficacy is considered as a person's belief in the ability to perform certain behaviors (Bandura, 2000).This states that individuals who judge themselves as capable (efficacious) of performing certain tasks or activities are more likely to be capable of doing the particular task.One of the essential factors in this self-efficacy theory is the notion of selfreferent.In fact, it is believed that there are many forms of self-referents among which the individual's opinion about their own capability in dealing with the constant changes with the realities of life is very important (Bandura, 2003).Accordingly, academic achievement of a person can be influenced by the behavior which is itself affected by selfefficacy.Many other studies supported the effect of self-efficacy on academic achievement including writing achievement (e.g.Lane & Lane, 2001;Lane, Lane & Kyprianou, 2004;Pajares, 2000;Pajares & Johnson, 1996;Pajares & Valiente, 2001).
In writing, however, a strong sense of self-confidence in the writing task is called writing self-efficacy.In other words, individuals may feel better to write when they have self-belief or self-confidence in their ability to write.They may also be more assertive and face with the difficulties with more perseverance when doing a writing task.In the self-efficacy discussions, three degrees of efficacy are mentioned: high, mid and low.Those who have a high confidence in the writing ability are considered as people with high self-efficacy or having a positive sense of self when it comes to writing.The definition for the other types are in the same direction.So, the students with high self-efficacy consider the hard writing task as a challenge to fulfill and try their best to accomplish the task by making productive use of their cognitive strategies (Lavelle, 2006).Conversely, non-self-regulated students don't get involved in learning process and as a result they might be subjected to any kind of sophomoric knowledge rather than deep knowledge which is needed for high academic achievement (Zimmerman, 1986).
The relationship between self-efficacy and students' writing was examined by researchers in the 1980s for the first time.
Shell and his colleagues are among the pioneers who examined the relationship between writing self-efficacy and writing achievement.In one of their studies (Shell, Murphy, & Bruning, 1989), they investigated the relationship between undergraduate students' writing and reading self-efficacy and their writing and reading achievement.They Flourishing Creativity & Literacy constructed a scale for writing self-efficacy which measured the students' skills in different writing tasks (e.g.writing a letter and an essay) and students' skills in writing components (e.g.spelling and parts of speech).Their writing achievement was obtained from the students' scores on an essay and the scores were given holistically.The results showed that the writing self-efficacy (both task efficacy and component efficacy) had a significant predictive power on writing achievement.The same result has also been reported in other studies from the same era (e.g.Pajares & Johnson, 1996;Pajares & Valiente, 1997;Pajares & Valiente, 1999;McCarthy, Meier, & Rinderer, 1985).
Moreover, Pajares and Valiente (2001) also investigated elementary-school students' writing self-efficacy and found that the students' self-predictions significantly predicted their writing performance; furthermore, they found that their self-efficacy beliefs directly influenced their anxiety about the task of writing, their feelings about its perceived usefulness, and their evaluation of essay writing specifically.White and Bruning (2005) did a research on the relationship of writing beliefs, including writing self-efficacy, and writing performance of the post-secondary learners.They administered a quantitative study through three tests that indicated the learners' implicit beliefs about the writing hardness, the attitudes toward writing they held, and their writing self-efficacy.The researchers deduced that the learners' beliefs about writing had a great impact on the quality of their writing.As a result, they persuaded writing instructors to pay attention to the self-efficacy perceptions of their learners so that "integrated models of writing" (p.186) can be designed to meet their individual writing demands.
In a more recent study, Jones (2007) investigated the relationship between 118 freshman learners' writing self-efficacy and their internal sense of control over their writing performance.The participants were from many nations including Asians, African-Americans, Latinos, Whites, and "Other" men and women.The results revealed that there was a significant difference between the learners showing weak writing skills with those having stronger writing skills."Selfbeliefs had a far greater effect on course grade than previous writing achievement for the weaker students than the stronger students" (Jones, 2007, p.18).
Based on what was discussed earlier, although self-efficacy has been known as having a great impact on the performance of the writer, there are some other studies which state that the students' high self-efficacy in writing may not surely indicate better writing performance (see Alverez & Adelman, 1986;Sawyer, Graham, & Harris, 1992).One of the most recent studies that confirms this point was done by Hashemnejad, Zoghi and Amini (2014) on one hundred and twenty Iranian EFL learners.In another study conducted by Al-Mekhlafi (2011) on forty-four female EFL teachertrainees at Sohar University in Oman, the results revealed that there was no relationship between the participants' writing self-efficacy and their writing performance.So, as it can be seen here, there is a discrepancy between the results of the above-mentioned studies regarding the relationship between writing self-efficacy and writing performance.Therefore, the researchers of this study thought there is a need for more investigation in this regard.In light of this research, the researchers formulated the following research questions to test the relationship between the two variables, writing self-efficacy and writing performance.1) Is there any significant relationship between the students' writing self-efficacy and their writing performance?
2) Is there a gender difference in the students' writing self-efficacy and their writing performance?

Methodology
Based on the above-mentioned research questions, correlational research design was adopted to collect the data.

Participants
The researcher utilized convenient sampling to choose 59 (28 male and 31 female) medical students from medical university students who were required to take writing courses as a compulsory 3-unit course before their graduation from medical university.This sampling is convenient because all 59 students of this study (N=30 from one class and N=29 from another class) were instructed by the same teacher (one of the researchers); hence, the researchers could control the effect that various teaching instructions might have had on the students' writing performance.However, it is worth mentioning that since these students were scattered in two separate classes (due to size limitation policy of the university for the writing classes), the researchers had to make sure that the students in both classes are homogeneous in terms of their writing ability.Consequently, the very first composition the students of both classes wrote at the first session of the semester was marked and analyzed for homogeneity test using an Independent t-test.Significance was considered at the (p<0.05)level.The following Results of the homogeneity test showed a mean (M) of 38.2 and a standard deviation (SD) of 3.3 for Group A and a mean (M) of .387 and a standard deviation (SD) of 5.2 for Group B. With equal variances, the significance level for the difference between homogeneity test mean scores of both groups was 0.507 which is higher than P value of 0.05.It shows that the difference is not significant.Therefore, the researchers concluded that both groups were homogenous in terms of their writing ability.

Instruments
The following instruments were used in order to collect the data in this study.

Writing Self-efficacy Questionnaire
The questionnaire which was used for the writing self-efficacy was designed from Donald O. Prickel research in 1994.He conducted a research on the development and validation of a writing self-efficacy scale for adult writers and its use in correlational analysis.The questionnaire consists of 25 questions in the 5 point Likert scale (Strongly Disagree, Disagree Unsure, Agree, and Strongly agree).Since the Persian version of this questionnaire was not available, the researchers used translation-back-translation process to translate the English version of the questionnaire to Persian.First, the questionnaire was translated to Persian by a competent bilingual translator who was an expert in translating academic questionnaires.Then the Persian version of the questionnaire was again translated into English by another professional translator without reference to the original text and then the two English versions of the questionnaires were compared with one another.It is worth mentioning that although this translation is considered a strict translation in which the translator usually pays attention to the word by word translation of a text rather than underlying concepts that the questions were intended to measure, the researchers consulted 3 professionals to make sure that for each question intended the concepts are actually captured.
After doing the translation-back-translation method, the pilot study was done and the questionnaire was given to 30 university students to see whether they can answer all the questions without any ambiguity; if there was any ambiguity, the question was modified.The pilot study showed a good reliability of the questionnaire.The correlation coefficient was +0.68 which showed a positive correlation between the test items.So the questionnaire was used as an instrument in this study.Persian translation and the English version of the questionnaire are shown in appendices A and B as well.

Writing Test
The students' writing compositions were collected at their final exam in order to be marked for the purpose of this study.The writing topic given to all students was "the effects that smoking can have on the body".The scoring rubric which was applied in this study was developed by Goldburg and Roswell (2000).This rubric was developed for the purpose of advanced composition courses.The components of the writing rubric include content, organization, vocabulary, sentence structure, grammar and mechanics and for each component different point values were given to calculate the total points.
To check the reliability and validity of this rubric, the researchers referred to Knoch (2009) and Gennaro (2009) for the distinctions between holistic and analytic rating scales found in the literature because L2 students usually differ in their control of grammatical, cohesive, rhetorical, sociolinguistic and content components of essays in terms of the strengths and weaknesses relevant to placement decisions into college level composition courses.Based on Alderson's (2005) description of direct holistic assessment of students' writing, this rubric met the criteria because it (1) identified different strengths and weaknesses in the learner's knowledge and use of written language, and (2) enabled a detailed analysis of the student responses to specific elements of the writing task.Moreover, the descriptors for writing ability and the point values reflected in the rating scales of the instrument in this study lend themselves well to a holistic approach by isolating troublesome aspects of writing performance (Knoch, 2009).

Reliability Test
Reliability has been described as the extent to which a measuring procedure produces the same outcomes on repeated trials (Carmines & Zeller, 1979).When a human rater applies rating system in content analysis, Neuendorf (2002) presumes that at least one more rater is necessary in order to make a rating scheme valid.The criteria for choosing a rater for this study is mainly related to the fact that since this study is mainly concerned with writing process in terms of content , organization, word choice, sentence structure, grammar, and mechanics the rater needs to be an experienced English teacher who have already mastered the knowledge of all the above-mentioned criteria himself/herself.Therefore, an authentic rater was chosen to mark the papers besides the researchers.At first, the researchers clarified all the elements of writing rubric for the inter-rater and requested the rater to mark the papers based on the given writing rubric.Then, the inter-rater reliability in this study was calculated using t-test.
As it can be seen in the following Table, the inter-rater reliability test shows that in the final test there was not any significant difference between the mean of the scores the raters had given to students' writings.As it can be seen in Table 2, in the final test the significance level for the difference between the raters' mean score was 0.752 which is higher than the P value of 0.05.Therefore, the mean of the two scores given by two raters was used as the final score of writing for each participant.

Data Collection
The students wrote their writings on A4 papers in the final exam in 45 minutes.Immediately after the students wrote their essays in the final exam, they were given writing self-efficacy questionnaire to be filled out.Students filled out the Likert scale questionnaire in 15-20 minutes.

Data Analysis
For the first objective of this study (to see if there is a relationship between the students' writing self-efficacy and their writing performance) and the second objective (to see if there is a gender difference in the students' writing selfefficacy and their writing performance), Pearson product-moment correlation was used.

Results
As depicted in Table 3, without considering the gender of the participants, no significant relationship was found between the students' self-efficacy and their writing performance as p value was greater than 0.05 (P=.197> 0.05).
Table 3. Results of Pearson's correlation for the students' writing self-efficacy and their writing performance For the relationship between male students' writing self-efficacy and their writing performance, Pearson productmoment correlation coefficient was examined.As it can be seen in Table 4, again the p value was higher than 0.05 (P= .464>0.05); therefore, we can conclude that there is not a significant difference between our male participants' writing self-efficacy and their writing performance.
Table 4.The results of Pearson's correlation for male participants' self-efficacy and their writing performance And finally to see if there is any relationship between female participants' writing self-efficacy and their writing performance, the same procedure was used and the results can be seen in Table 5.
Table 5.The results of Pearson's correlation for female participants' self-efficacy and their writing performance As it can be seen in the above table, the p value is a bit higher than 0.05 (P= .053)which is somehow borderline.As the relationship is also found as a weak one so it cannot be considered significant.

Discussion
Considering a purposive sample of 59 EFL students studying at one university in Shiraz/Iran, the findings and conclusions of this study are limited in their generalizability.When considering writing, it is not normally possible to generalize the results to a broad range of learners and each study is limited to a specific group (Jafarian, Soori, Kafipour, 2012).With this caveat in mind, the results of this study demonstrated that there was not any relationship between the participants' writing self-efficacy and their writing performance.This result is contrary to the previous research that indicated that these two variables were strongly related to each other (Zimmerman, 1989;Schunk, 1991;Zimmerman, Bandura & Martinez-Pons, 1992;Webb-Williams, 2006).This result is not also parallel with many studies that Pajares conducted and found a strong relationship between self-efficacy and the subjects' composition (see Pajares, 2000, Pajares & Johnson, 1996, Pajares & Valiente, 2001).Self-efficacy has usually been found to have the strongest predictive power among all the motivational constructs studied over writing performance (Pajares & Johnson, 1996); however, this study doesn't support this.
But why the result of this study is in contrast with some of the most leading studies conducted in the area of writing self-efficacy and writing performance?To answer this question, it is imperative to review the definition of self-efficacy again.As it was mentioned earlier, self-efficacy is considered as one's belief of one's ability to perform certain behaviors and achieve a task (Bandura, 1977).So the result of this study contrasts this assumption that the individuals who think they can fulfill a particular task are more likely to try and finish the task successfully.The only reason behind such finding, according to authors of this study, is students' lack of motivation.This is congruent with findings of some other studies which highlighted the role of motivation in general or internal motivation in specific such as Kafipour, Noordin & Pezeshkian (2011), Yazdi and Kafipour (2014), Kafipour and Naveh (2011), Kafipour, Yazdi & Shokrpour (2011), Soori, Kafipour, Soury (2011).The participants of this study had a strong self-belief in their writing ability; however, they were not motivated enough to produce satisfactory results.This turns us to a study which was conducted by Khojasteh, Shokrpour and Kafipour (2015) on similar population (Shiraz medical students) which reported that students do not have any interest in learning English whatsoever and the only reasons they take these general and specific English courses are that a) they are compulsory courses that needed to be taken by all medical students and b) they just want to pass the course (possibly without an attempt to obtain good marks).So we can see here that just having the belief that you can do something wouldn't necessarily mean that you can successfully accomplish the task if you are not motivated enough to do so.Another possible explanation for such result can also be what Lavelle and Zuercher (1999) called over-estimation of one's self efficacy.It is also possible that our participants' self-efficacy was significantly beyond their actual ability and therefore this led to what we found in our results.As to the second question, this study demonstrated that there was not any significant relationship between our male and female participants' selfefficacy and their writing performance.However, this finding doesn't support the results of previous studies focusing mainly on gender differences in terms of writing self-efficacy and writing performance.For example, Pajares (2002) studied male and female elementary and middle school students and reported that in comparison to boys, girls tended to have higher self-efficacy while boys over-estimated their writing ability.According to Lavelle and Zuercher (1999), students with a self-efficacy significantly beyond their actual ability often overestimate their ability to complete tasks, and this can lead to difficulties.But then, students with a self-efficacy significantly lower than their ability are unlikely to flourish and develop their skills.In terms of gender differences, Wigfield, Eccles and Pintrich (1996) also showed that the girls tended to do better in writing tasks even though their self-efficacy equated that of boys.The result of this study, however, are parallel the findings of Hashemnejad, Zoghi & Amini (2014) who did not find any relationship between not only self-efficacy, but also writing performances of their male and female participants.

Conclusion and implications of the study
The findings of this study did not confirm what has been firmly established in the literature in terms of the relationship between students' writing self-efficacy and their writing performance.One possible reason for such a finding was students' lack of motivation in learning English at all.According to Khojasteh, Shokrpour and Kafipour (2015), lack of integrated motivation to learn English has not given students a reason to learn and a desire to attain the learning goal and one of the main reasons mentioned in that study is teachers, who according to students, did not provide enough friendly atmosphere for students to feel appreciated.Therefore, in order to have more motivated students, it is also important to have motivated and more successful EFL teachers (Hekmatzadeh, Khojasteh & Shokrpour, 2016).
Based on the above mentioned finding, this study doesn't suggest that the students' writing self-efficacy can be ignored throughout the writing course.Schunk (1989) emphasized the fact that the students who believe they will experience much difficulty comprehending material are more willing to hold a low sense of efficacy for learning it, whereas those who feel they are able to handle the information-processing demands should feel more effective.
Many researchers have confirmed that a key element towards an individual's self-regulated learning is self-efficacy.In studies of the relationship between self-efficacy and self-regulated learning, those students who had high self-efficacy showed better state of learning strategies and enhanced levels of self-monitoring of their learning results than those students with less self-efficacy (Kurtz & Borkowski, 1984).Some other researchers showed a positive correlation between self-efficacy and task persistence (Zimmerman & Ringle, 1981) task choice (Bandura & Schunk, 1981), effective study activities and skill acquisition (Schunk, 1984), and academic achievement (McCarthy, Meier, & Rinderer, 1985).

Table 1 .
Table displays the results.The mean of homogeneity scores of the two groups in writing

Table 2 .
Students' writing performance according to the rater's final test scores