THE INFLUENCE OF PEER COMMENT ON FACEBOOK AND GENDER TOWARD WRITING ABILITY TO THE ELEVENTH GRADE STUDENTS OF SMA NEGERI

Abstrak Writing is one of the four English skills which considered the most difficult skill for the students to study. The difficulty might be in organizing the ideas, structure, and mechanic. Meanwhile the existence of the personality types or attitude of the students is also considered as the factor the success of writing. In the light of the above issues, the researcher wants to investigate the influence of peer comment on facebook and gender toward writing ability. The population of this study was the eleventh grade students of SMA Negeri 5 Palembang academic year 2011/2012 consisting 174 students. 40 students were represented as an experimental class and the other 40 of them were as the control class. By using Stratified Random Sampling, in carrying out the research, the writer used a factorial design. The data distributions were considered normal. It was proved by statistic calculation of normality test which shows that p-output (0.60 and 0.70) were higher than α (.05). Then based on the calculation of Levene statistics, it was found that p-value is higher than α (.05). Based on the result of statistic calculation using independent t-test, it was found that all four hypotheses were accepted since the p-output was .000 which less than α (.05).Furthermore, this article also discusses the significant value of the technique which is the influence of peer comment on facebook and traditional technique.


Introduction
The facebook is one of the familiar and top social networking sites nowadays.There are two primary ways for companies to market on the facebook.They are the Facebook Groups and the Facebook Pages.The former are focused more on bringing together groups of people who share common interests.They are communities of friends with similar interests, and are also becoming extremely popular.They promote brand presence, share news/events with their community and stimulate relevant discussions.A prime example of an organization leveraging Facebook Groups is Contikithrough their group they share brochures, the trip/video of the week, contests, and polls to their 37,000+ members.The latter are places where users can interact around and with a particular brand or product.It can share our business and products with facebook users.The example of this page is Harley Davidson Motor Company page.
The phenomenon of using the facebook has already been widespread in Indonesia especially for students.Almost every student has an active facebook.Some schools have hot spot area so they can access it easily.Sometimes they can access it by their mobiles.Because of that we use facebook media to make the students not only have fun in facebook, but also they can use it in a positive way.Especially in facebook group the students can write something in their mind.On the other hand, other students can comment directly on that story.It can make improvement for them in appreciating especially a fairy tale.
There are a lot of fairy tales that can be found in many sources, such as in the textbooks and magazines.According to Murcdoch (2002: 17-19), literature learning process nowadays still applies the old paradigm, structuralism approach.Literature subject is taught in educational institution and it is taught traditionally and conventionally and only focuses on intrinsic values Alderson (1995).The students only analyze the elements of fiction that have already been found in the text.Consequently, literature subject does not contribute to increase students' literary appreciation.
According to Barlian (2003:5), teachers do not know how to teach literature in a proper way.Similarly, Rudy (2007:7) quotes from some expert's opinion that literature teaching cannot run well since teachers are lack of teaching strategy.She has also observed the phenomenon.Consequently, literary works become such a boring subject to learn.The students complain that the way their lecturers present the story in the classroom is not interesting.The lectures only provide them with the elements of fiction to be appreciated.For these reasons, a strategy is required to help students appreciate the story better.
Writing is one of the basic skills which one acquires in teaching and learning activities in the classroom.The skill of expressing one's thoughts and communicating ideas and views to others is developed here.Writing is the primary basis upon which the student work, student learning, and student intellect will be judged in college, in the workplace, and in the community.Other source accessed on February 21 st , 2012 http://wiki.answers.com/Q/What is the importance of writing state that writing is one of the ways that we translate our thoughts for other people.Some people are better at expressing themselves in writing than the others, and we thus get a better translation when we read what they have to say rather than hearing them speak.Writing assists the students with other language tasks as well as writing helps the student learn how to form language, how to spell, how to put together a plot.The students learn how to make a logical argument, or how to persuade, mainly through writing.Speaking can help us learn those things as well, but it is easier to self-examine and evaluate how to improve when we have something concrete in front of us, and can revise it.
Based on the syllabus, writing in the classroom activities taught by the teacher 3 (three) times in one semester.English taught by the teacher twice a week with time allocation of 2 x 45 minutes for every meeting.The teachers teach it with whilst activity just explain and demonstrate how to make short story and ask the students to make short story based on RencanaPelaksanaanPembelajaran (RPP) prepared by the teacher before the new of academic year starts.
An alternative technique in making writing more interesting which has already been done by the writer is using the web-log in the Internet.The writer plans to do a research on the use of the facebook as the media in teaching writing.No other researchers have done this kind of study, at least at the Graduate Program of PGRI University Palembang.
Nowadays, people live in technologically advanced society and everything is connected to the Internet.The Internet world is competing in gaining the sophisticated technology in every part of life.Now the Internet becomes more familiar in the field of education including in language teaching.According to Kauchak, Eggen and Charter (2002:389) in Rudy (2007:6), the computers are interconnected through an Internet which enables people to communicate and share information worldwide.In relation to language teaching and learning, one of the familiar technologies that are very well-known now on the Internet is by using the web-log (Peachey, 2008:20).
Teaching is communication process in which a body of knowledge is delivered from an instructor to students (Gagne and Briggs, 1985:10).With the proliferation of computer software's and hardware's, especially by using the facebook.There, the teachers and the students can interact and have teaching and learning process.This fact shows that the use of the Internet especially facebook and related information technologies will foster an alternative teaching and learning environment.
The Internet has become a part of daily lives.It helps people search everything entailed.It can also be a powerful teaching tool, providing access to information and allowinglinks to other learners.Similarly, Richard and Renandya (2002:362) list some experts in making use of the Internet in teaching language.
Based on the explanation above, the researcher is interested in investigating students' writing ability through peer comment on facebook.

Literature Review
Peers are a good source of feedback, the helpfulness of their comments on peer writing can be very limited.In addition, they tend to be inexperienced in writing and generating constructive comments.Therefore, it is critical for the successful use of peer reviewing that students come to generate helpful reviews for their peer writers in the absence of strong expert input.Peer reviewing of writing is hard for the instructor to monitor.As the class size increases, the amount of peer comments increases in an exponential way.
When we peer-review another writer's work, we evaluate it, criticize it, suggest improvements, and then communicate all of that to the writer.As a first time peer-review, we might be a bit uneasy about criticizing someone else's work.For example, how do we well somebody his essay is boring.Read the discussion and steps that follow; we'll find advice and guidelines on doing peer reviews and communicating peer-review comments.
When we are peer-reviewing other people's writing, remember above all that we should consider all aspects of that writing, not just in fact, at least of all the grammar, spelling and punctuation.If we are new to peer-reviewing, we may forget to review the draft for certain aspects.The following outline may serve as guidelines or points to consider.
1. Make sure that the review is comprehensive.Consider all aspects of the draft we're reviewing, not just the grammar, punctuation, and spelling.In peer response, the students are given plenty of opportunities to brainstorm ideas in pair groups, to give feedback on each other's writing and to check and edit for each other.Besides pairing students up in writing is an ideal way to promote learning effectiveness, it not only gives teachers more quality time to work with students but also provides students with plenty of opportunities to brainstorm ideas and to learn from each other (Teo, 2006:9).
Facebook is one of the top social networking sites, quickly gaining Myspace in terms of total users and traffic.Facebook presents a unique marketing opportunity for businesses through the creation of Facebook Business Pages.As more people explore social media, social networking sites have become some of the key online sources they use to learn more about products, organization, artists and world events.Marketing on facebook has a viral effect information spreads through networks exponentially.
There are two primary ways for companies to market on facebook.The first is through Facebook Groups and the second is through Facebook Pages.Groups are more focused on bringing together groups of people share common interests, whereas pages are places where users can interact around and with a particular brand or product.Finocchiaro (1989:136) states that there are four kinds of teaching media.They are: 1. Audiovisual materials 2. Audio material 3. Visual material 4. Realia Media and materials are software which conveys message or educational information usually prepared to be used with equipments, while hardware is a medium to convey the massage.It means software is as instructional media such as television, computers, etc.In this research, facebook is included as a part of audiovisual material in teaching media.
Research in EFL writing class suggests teacher's feedback not emphasize on surface level (grammar, punctuation, and spelling) for its harmful (Nafsiah, 2007:5).A study by Robb, Ross and significant effect to improvement of the student's writing by the end of the source (Gray, 2004:8).
In contrast, students have different view regarding the feedback on their writing.Studies by Ancker (2000:11) and Diab (2006:13) disclose the discrepancies between teacher beliefs are important, but student's beliefs that are more important than other features.Students feel that need to know what's wrong and what are right as well as the way to avoid the same mistakes in the future.
To avoid such miscommunication, it is better for both parties to negotiate, in the beginning of lesson, the type of feedback they would like to have.

A. Narrative Text
Narrative text is a text which contains about story (fiction/ non fiction / tales/ folktales/ fables/ myths/ epic) and its plot consists of climax of the story (complication) then followed by the resolution.
According to Utami (2009:229), narrative text has purposes that to tell a story to amuse, entertain, teach a lesson or moral, explain something or make a comment. In

B. Fairy Tales
A fairy tale is a type of short story that typically features folkloric fantasy characters, such as fairies, goblins, elves, trolls, dwarves, giants, mermaids or gnomes, and usually magic or enchantments.However, only a small number of the stories refer to fairies.The stories may nonetheless be distinguished from other folk narratives such as legends (which generally involve belief in the veracity of the events described) and explicitly moral tales, including beast tables.
In less technical contexts, the term is also used to describe something blessed with unusual happiness, as in "fairy tale ending" (a happy ending) or "fairy tale romance" (though not all fairy tales end happily).Colloquially, a "fairy story" can also mean any fetched story or tall tale.
In cultures where demons and witches are perceived as real, fairy tales may merge into legends, where narrative is perceived both by teller and hearers as being grounded in historical truth.However, unlike legends and epics, they usually do not contain more than superficial references to religion and actual places, people, and events; they take place once upon a time rather than in actual times.
Fairy tales are found in oral and in literary form.The history of the fairy tale is particularly difficult to trace because only the literary forms can survive.Still, the evidence of literary works at least indicates that fairy tales have existed for thousands of years, although not perhaps recognized as a genre; the name "fairy tale" was first ascribed to them by Madame d'Aulnoy in the late 17 th century.Many of today's fairy tales have evolved from centuries-old stories that have appeared, with variations, in multiple cultures around the world.Fairy tales, and works derived from fairy tales, are still written today.
The older fairy tales were intended for an audience of adults, as well as children, but they were associated with children as early as the writings of the precieuses: the Brothers Grimm titled their collection Children's and Household tales, and the link with children has only grown stronger with time.
Folklorists have classified fairy tales in various ways.The Aame-Thompson classification system and the morphological analysis of Vladimir Propp are among the most notable.Other folklorists have interpreted the tales' significance, but no school has been definitely established for the meaning of the tales.

Snow White and Seven Dwarfs
Once upon a time, there lived a little girl named Snow White.She lived with her aunt and uncle because her parents were dead.One day, she heard her uncle and aunt talking about leaving Snow White in the castle because they both wanted to go to the town and they didn't have enough money to take Snow White.
Snow White didn't want her uncle and aunt to do this, so she decided it would be best if she ran away.The next morning she ran away from home.When her aunt and uncle were having breakfast, she ran away into the woods.
Then she saw this little cottage.She knocked but no one answered, so she went inside and fell asleep.Meanwhile, the seven dwarfs were coming home from work.They went inside.They found Snow White sleeping.Then Snow White woke up.She saw the dwarfs.The dwarfs said, "What is your name?" Snow White said, "My name is Snow White."Doc, one of the dwarfs said, "If you wish, you may live here with us." Snow White said, "Oh, can I? Thank you." Then Snow White told the dwarfs the whole story, and Snow White and the seven dwarfs live happily even after.

C. Gender
Gender is a range of characteristics used to distinguish between males and females, particularly in the cases of men and women and the masculine and feminine attributes assigned to them.Depending on the context, the discriminating characteristics vary from sex to social role to gender identity.Sexologist John Money introduced the terminological distinction between biological sex and gender as a role in 1995.Before his work, it was uncommon to use the word "gender" to refer to anything but grammatical categories.However, Money's meaning of the word did not become widespread until the 1970s, when feminist theory embraced the distinction between biological sex and the social construct of gender.Today, the distinction is strictly followed in some contexts, like feminist literature, and in documents written by organizations such as the World Health Organization (WHO), but in most contexts, even in some areas of social sciences, the meaning gender has expanded to include "sex" or even to replace the latter word.Although this gradual change in the meaning of gender can be traced to the 1980s, a small acceleration of the process in the scientific literature was observed when the Food and Drug Administration started to use "gender" instead of "sex" in 1993."Gender" is now commonly used even to refer to the physiology of non-human animals, without any implication of social gender roles.
In the English literature, the tracheotomy between biological sex, psychological gender, and social sex role first appeared in a feminist paper on transsexualism in 1978.Some cultures have specific gender related social roles that can be considered distinct from male and female, such as the hijra of India and Pakistan.
While the social sciences sometimes approach gender as a social construct, and gender studies particularly do, research in the natural sciences investigates whether biological differences in males and females influence the development of gender in humans; both inform debate about how far biological differences influence gender identity formation.

Method of Research
The method used in research was factorial group design.This method is the modification of pre-testpost-test control group design which permits the investigation of additional independent variable (moderator variable).In this design, there were two groups, the first group was as the experimental group, and the other one was control group.This design can be shown as follow: The symbols above are interpreted as follows: • R refers to Random

Sample
Sample is part of whole population of investigation (Arikunto, 1989:104).To determine the sample of investigation, Arikunto (1989:107) says if the subject of investigation is less than 100, all subjects of investigation must be taken.If the subject is more than 100, the sample is 100%, 15%, 20%, 25%.The sample of the Eleventh Grade Students of SMA Negeri 5 Palembang can be seen in table 2.
The sample of investigation was 80 students.The 80 sample students were taken randomly from six classes by using interview, questionnaire, and test.The writer selected the students' names which were chosen as sample.

Technique for Collecting the Data Interview
The writer interviewed two teachers of English.The writer asked them about the English textbook they used, what kinds of teaching media they used in the process of teaching and learning.Media they used in the process of teaching and learning writing items, the schedule of English lesson, how many times they taught writing and motivation of the students in learning English.

Questionnaire
In collecting the data, the writer used the questionnaire yes-no questions for the model of Facebook.This questionnaire was adapted from Rudy (2005:21).

Test
Test was used because it was considered as the most reliable way to get some information.According to Tenbrink (1977:6), testing tends to be most objective technique for obtaining information and is likely to lead the most reliable information.
In this model, the students typed the story on the comment box that they could find on facebook.Kinds of writing that they were assigned on their comment box were narrative text.
The results of the students' analysis are then scored with holistic scale.Alderson (1995:108) explains "when examiners use this type of scale, they are asked not to pay much attention to any particular of the candidates' production, but rather to make a judgment of its overall effectiveness.

Reliability
The degree, to which a test is consistent, or reliable, can be estimated by calculating a reliability coefficient.In this investigation, to estimate the reliability of test, the inter-ratter reliability is used.Interratterreliability is estimation based on the correlation of scores between two or more raters who rate the same item, scale, or instrument (Garson, 1998).Moreover, (Brown, 2004:21) says that Inter-ratter reliability occurs when two or more scores yield inconsistent scores of the same test, possibly for lack of attention to scoring criteria, inexperienced, inattention, or even preconceived biases.
To minimize the subjectivity in scoring process, there were two raters get involved in giving scores.The first one was the researcher as a "Rater A" and the second one was his collogue who has the experience in handling teaching and testing writing as a "Rater B".The correlation coefficient between two raters was calculated by Cohen's Kappa using SPSS computation.Cohen's Kappa for inter-rater reliability can be used to assess inter-rater reliability if there are just two raters (Garson, 1998:31).He also adds that "am inter-rater reliability analyzing using the Kappa statistic was performed to determine consistency among raters.The output and interpretation of the reliability of the pretest was figured out in the following table 6.
The result of the inter-rater analysis were Kappa = .631with p < .000and 95% CL (.504, .848).As a rule of thumb values of Kappa from 0.40 to 0.59 are considered moderate, 0.60 to 0.79 substantial, and 0.80 outstanding (Landis & Koch in Garson, 1998).Most statisticians prefer for Kappa values to be least 0.6 and most often higher than 0.7 before claiming a good level of agreement.
From the above output and interpretation, it can be concluded than consequent the test is considered as a reliability test.

Technique for Analyzing the Data
Three techniques were applied to analyze the obtained data, they were: (1) the statistical on measuring homogeneity and normality of the data, (2) independent t-test, and (3) Two ways ANOVA, and all the data analyses were measured by using the statistical Package for Social Science (SPSS) 17.

Statistical Analysis on Measuring Homogeneity and Normality.
This test was used to find out whether or not the samples taken fromthe same population was normal and whether or not the sample had the same variance.It can be said that both score (peer comment on facebook and traditional technique) are categorized into normal since p-output is higher than α (.05), here the writer uses Kolmogorov-Smirnov statistic calculation.The sample is considered homogeneous whenever the p-output is higher than α (.05) and the writer uses Levene Statistic calculation.

Independent t-test
The independent t-test was applied to find out the significant difference in means between control and experimental group.

Two-way ANOVA
To measure the interaction effect from students' writing average scores using peer comment on facebook toward gender, two-way ANOVA found in SPSS 17 is used.
The two ways ANOVA enables to test the difference between two or more samples and every sample consist of two or more variance simultaneously (Riduan, 2004:170).

Findings Data Description The Pretest Score in the Experimental Group
After the pretest was administered to experimental group, it was found that there was one student who got the lowest score, namely 60 (2.5%).The score among 62 up to 69 were got by thirty six students (90%), the score 70 were got by three students (7.5%).
Based on the descriptive statistic, it was found that the mean score was 66.05, the median was 66.00, the mode of the score was 65, standard deviation was 2.287, the variance was 5.228, the range of the score was 10, the minimum score was 60, and the maximum score was 70.

The Pretest Score in Control Group
Then, when pretest was done in control group, it was found that there were five students who got the lowest score, namely 60 (12.5%).The scores among 62 up to 69 were got by thirty four students (85.5%), and the scores 70 was got by 1 student (2.5%).
Based on the descriptive statistic, it was found that the mean score was 64.76, the median was 65.00, the mode of the score was 63, standard deviation was 2.167, the variance was 7.885, the range of the score was 10, the minimum score was 60, and the maximum score was 70.The following table (table 10) shows the descriptive statistic of the data.

The Posttest Score in Experimental Group
We can see from the result of the posttest score that it was 2.5% (one student) got the lowest score i.e., 75.Besides, the score 76 (5%) were got by two person, the scores among 80 up to 89 were got by 33 students (82.5%), the scores 90 (7.5%) were got by three students, and the highest score i.e., 91 (2.5%) was got by one student.
Based on the descriptive statistic, it was found that the mean score was 84.28, the median was 85.00, the mode of the score was 86, standard deviation was 3.816, the variance was 14.563, the range of the score was 16, the minimum score was 75, and the maximum score was 91.

The Posttest Score in Control Group
We can see from the table the result of the posttest score in control group were that 7.5% (three students) got the lowest score i.e., 62.In addition the score from 63 to 69 (90%) were got by thirty six students and 2.5% (one student) was got the highest score i.e., 70.
Based on the descriptive statistic, it was found that the mean score was 84.28, the median was 85.00, the mode of the score was 86, standard deviation was 3.816, the variance was 14.563, the range of the score was 16, the minimum score was 75, and the maximum score was 91.

The Result of Gender in Experimental Group
Based on the result of students' questionnaire, it was found out that the gender was only male and female.

The Result of Female Students in Experimental Group
The following table explained that the score 51 (20%) were got by four students, the score 53 (10%) were got by two students, the score 54 (20%) were got by four students, score 55 (5.0%) was got by one student, score 58 (20%) by four students, score 59 (15%) were got by three students, score 60 (5.0%) by one student, and the score 63 (5.0%) was got by one student.

The Result of Male Students in Experimental Group
The following table explained that there were one student who was in 41 (5.0%), two students who were in 43 (10%), two students who were in 45 (5.0%), three students who were in 46 (15%), four students who were in 48 (20%), five students who were in 49 (25%), and three students who were in 50 (15%).

The Result of Gender in Control Group The Result of Female Students in Control Group
The following table explained that there were two students who were in 51 (10%), five students who were in 52 (25%), three students who were in 53 (15%), three students who were in 54 (15%), one student who was in 57 (5%), three students who were in 58 (15%), two students who were in 59 (10%), and one student who was in 61 (5.0%).

The Result of Male Students in Control Group
The following table explained that one student who was in 40 (5.0%), five students who was in 44 (5.0%), two students who were in 45 (10%), four students who were in 46 (20%), two students who were in 47 (10%), four students who were in 48 (20%), three students who were in 49 (15%), and three students were in 50 (15%).

Data Analyses
In order to analyze the data obtained, two statistical analyses were applied.They are: (1) the independent t-test, and (2) the two ways ANOVA on measuring the influence on the students' writing achievement taught using peer comment on facebook or not toward gender (male and female).

The Statistical Analysis on Measuring Normality of the Data
Before conducting an interference statistic, the test of normality and homogeneity were needed.This way used to find out whether or not the samples taken from the same population (distribution of population data) is normal and whether or not that samples has the same variance.
From statistic calculation by using normality test of Kolmogorov-Smirnov, it was found that the significant value of pretest experiment was .060,while teaching writing not using peer comment on facebook was .070.It could be said that both score were categorized normal since p-output was higher than α (.05).Then the result of normality test in Shapiro-Wilk showed that the significant value of pretest experimental group was .142,while teaching writing not using peer comment on facebook was .051.It meant that both scores were also categorized into normal since p-output was higher than α (.05).
The following is the figure of the two normal Q-Q plots of writing by using peer comment on facebook and not using peer comment on facebook.
Based on the normal Q-Q plot writing achievement in peer comment on facebook and traditional technique, the data of writing achievement were in normal category.

Analysis on Measuring Homogeneity of the Data
To know whether the samples were homogeneous or not, the writer used the test of homogeneity of variance.Based on the calculation of Levene Statistic by using SPSS 17, it was found that the p-value was .067.The samples are considered homogeneous whenever the p-value was higher than α (.05).It meant that the samples taken from the experimental and control groups were homogeneous.

The Independent t-test
The independent t-test is the most commonly used method to evaluate the difference in means between the two groups.In this study, the independent t-test was used (1) to measure significant difference between gender's writing ability through peer comment on facebook and not, (2) to measure significant difference between female students' writing ability through peer comment on facebook and not, (3) to measure significant difference between male students' writing ability through peer comment on facebook and not, (4) to measure significant difference between gender students' who taught through peer comment on facebook and not.

Measuring Significant Difference between Gender's Writing Ability through Peer Comment on Facebook and Not.
From the calculation of the t-test, it was found mean score from the gender whom female student through peer comment on facebook was 87.00, standard deviation was 2.176, and standard error of mean was .487.Meanwhile, mean score from the female students who taught not using peer comment on facebook was 66.60, standard deviation was 2.113, and standard error of mean was .472.From the statistics calculation using independent sample test was found out the p-output .000.When poutput is lower than α (.05), it could be predicted that the peer comment on facebook and gender which was applied to develop students' writing ability in a narrative test is effective to be taught in a certain groups.

Measuring Significant Difference between Female Students' Writing Ability Using Peer Comment on Facebook and Not Using Peer Comment on Facebook
From the calculation of the t-test, it was found mean score from the female students through peer comment on facebook was 86.10, standard deviation was 4.778, and standard error of mean was 1.068.Meanwhile, mean score from the female students who taught not using peer comment on facebook was 64.50, standard deviation was 1.573, and standard error of mean was .352.From the statistics calculation using independent sample test was found out the p-output .000.When the p-output higher than α (.05), it could be predicted that the peer comment on facebook which was applied to develop students' writing ability in a narrative text is effective to be taught in a certain groups.

Measuring Significant Difference between Male Students' Writing Ability through Peer Comment on Facebook
From the calculation of the t-test, it was found mean score from the male students through peer comment on facebook was 81.60, standard deviation was 2.981 and standard error of mean was .666.Meanwhile, mean score from the male students who taught not using peer comment on facebook was 64.50.standard deviation was 1.573, and standard error of mean is .352.From the statistics calculation using independent sample test was found out the p-output .000.When the p-output is lower than α (.05), it could be predicted that peer comment on facebook which is applied to develop students' writing ability in a narrative text is effective to be taught in a certain groups.

Measuring Significant Difference between Gender Students' Writing Ability Using Peer Comment on Facebook and Not Using Peer Comment on Facebook
From the calculation of the t-test, it was found mean score from the female students not using peer comment on facebook was 66.60, standard deviation was 2.113, and standard error of mean was .472.Meanwhile, mean score from the male students taught through peer comment on facebook was 81.60, standard deviation was 2.981, and standard error of mean was .666.From the statistics calculation using independent sample test was found out the p-output .000.When the p-output is lower than α (.05), it could be predicted that peer comment on facebook which was applied to develop students' writing ability in a narrative test is effective to be taught in a certain groups.

Measuring the Influence of Peer Comment on Facebook and Gender toward Writing Ability
In these statistical analyses, the writer applied the two-way ANOVA which is considered with the investigation of the interaction between one dependent variable (writing achievement) and two variables (peer comment on facebook and gender).By using two-way ANOVA, the writer wanted to measure the influence of peer comment on facebook and gender (male and female).The following is the description.
To measure the influence between the students' writing narrative text ability through peer comment on facebook and gender (male and female), a two-way ANOVA found in SPSS 17 is applied.
From the calculation analysis, it was found that the p-output (.002) was lower than α (.05).It means that the influence was found between writing narrative text through peer comment on facebook and gender (male and female).

Interpretation
Based on the data analysis, the students' ability in writing a narrative indicated that teaching writing a narrative text through peer comment on facebook gives a significant difference on students' ability in writing than not.Some interpretation was made based on the statistics analysis related to the writer's finding.They are described as follows: First, from statistic calculation by using the normality test of Kolmogorov-Smirnov, it was found that the significant value of peer comment on facebook was .060,while teaching writing using traditional technique was .070.It could be said both scores were categorized into normal since poutput was higher than α (.05).Then, the result of normality test in Shapiro-Wilk showed that the significant value of peer comment on facebook was .142,while teaching writing using traditional technique was .051.It means that both scores are also categorized into normal since p-output was higher than α (.05).Furthermore, based on the calculation of Levene Statistic by using SPSS 17, it was found that the p-value was .067.The samples are considered homogeneous whenever the p-value is higher than α (.05).It means that the samples taken from experimental and control groups are homogeneous.
Second, the highest scores which got b the students were preferred to the female students both in experimental and control group, while the male students both in experimental and control group got the lower score.
Third, t-test analysis of the female students' ability in writing through peer comment on facebook and not.From the statistics calculation using independent sample test is found t the p-output is .000.It means that the p-output is lower than α (.05), and it can be interpreted that teaching writing a narrative text through peer comment on facebook is effectively applied in one group.Or it can be interpreted that there is a significant difference in average scores between female students through peer comment on facebook and not.
Forth, t-test analysis of the female students' ability in writing through peer comment on facebook and the female students not using peer comment on facebook.From the statistic calculation using the independent sample test was found that p-output is .000.It means that the poutput is lower than α (.05), and it can be interpreted that teaching writing a narrative text through peer comment on facebook is effectively applied in one group or it can be interpreted that there is a significant difference in average scores between the female students through peer comment on facebook and not.
Fifth, t-test analysis of the male students' ability in writing through peer comment and not.From the statistics calculation using independent sample test finds the p-output .000.It means that the poutput is lower than α (.05), and it can be interpreted that teaching writing a narrative text through peer comment on facebook is effectively applied in one group or it can be interpreted that there is a significant difference in average scores between the female students through peer comment on facebook and not.
Sixth, t-test of the male students in writing taught not using peer comment on facebook and female students through peer comment on facebook.From the statistic calculation using independent sample test is found out that p-output is .000.It means that the p-output is lower than α (.05), and it can be interpreted that teaching writing a narrative text through peer comment on facebook is effectively applied in one group.Or it can be interpreted that there is a significant difference in average scores between female students taught using peer comment on facebook and not.
Finally, the writer applied the two-way ANOVA analyses.It is aimed at showing the influence of peer comment on facebook and gender (male and female).Based on the Test of Between-Subjects influences analyses, the significant value of gender (male and female) were .002which less than p-value α .05, it means that there is a significant influence between peer comment on facebook and gender (male and female) to the students' writing ability.(Sig.value = .035<α .05).It can be interpreted that there is the influence between students' average scores who taught through peer comment on facebook and gender (male and female).

Conclusions
Based on the findings and interpretation in his previous chapter, it can be concluded as follows.First, t-test analysis of female students' writing ability that both taught using peer comment on facebook and who were taught using traditional technique.From the statistics calculation using independent sample test was found the poutput .000.It means that p-output was lower than α (.05).It was concluded that there is a significant difference in average scores between female students who both taught using peer comment on facebook and who were taught using traditional technique.
Second, t-test analysis of female students' writing ability who both taught using peer comment on facebook and male students who were taught using traditional technique.From the statistics calculation using independent sample test was found the p-output .000.It means that the p-output was lower than (.05).It was concluded that there is a significant difference in average scores between female students who both taught using peer comment on facebook and who were taught using traditional technique.
Third, t-test analysis of male students who both taught using peer comment on facebook and who were taught using traditional technique.From the statistics calculation using independent sample test was found the p-output .000.It means that the p-output was lower than α (.05).It was concluded that there is a significant difference in average scores between female students who both taught using peer comment on facebook and who were taught using traditional technique.
Forth, t-test analysis of female students' writing ability who both taught using traditional and male students who were taught using peer comment on facebook.From the statistics calculation using independent sample test was found the p-output .000.It means that the p-output was lower than α (.05).It was concluded that there is a significant difference in average scores between female students who taught using peer comment on facebook and who were taught using traditional technique.
And at last, the calculation using two-way ANOVA (Test of Between Subjects Influences) to measure an influence on the student's writing achievement taught using peer comment on facebook and traditional technique towards students' gender were found that the significant value of the technique (Peer comment on facebook and traditional technique) were .002which less than p-value 0.05, it means that there is significant influence between peer comment on facebook and traditional technique toward the students' gender (male and female) to the students' writing achievement.

Suggestions
From the conclusions above, the writer would like to suggest some points particularly extended to the English teachers, students and other researchers.

For the Teachers
For teacher of English several suggestions can be offered for improving students' writing achievement.First, to improve teaching and learning, writing teachers need to recognize the students' need and interest not only through the eyes of the teacher but also the eyes of the students by involving them in teaching and learning activity.Second, to maintain attention and to aid students in learning writing, teacher should use varieties of writing strategies or skills, and it is suggested that peer comment on facebook is appropriate to teach a narrative writing.Finally, it is good for teachers of English to give the students more practices in writing because repeated practice will help many students develop their writing skills.

For the Students
For students are expected to give more attention to the teacher's explanation and discuss to the teacher if they have a problem in learning writing.In writing a narrative text, especially Persona narrative, students should be able to express their emotion because one of the criteria of narrative writing rubric is voice which meant engage the reader through the writer's feeling or mood about the story that she/he have.

For Other Researchers
For other researchers, the writer suggests them to do further research about the different strategies or skills that appropriate in teaching writing.It is hoped that other researchers can create the newest strategy in teaching writing so that both the teachers of English and students can use it in teaching and learning activity.

For the Institution
The institution is suggested to: 1. Provide learning facilities such as libraries to support or as important source for students or teachers in developing their teaching and learning process continuously.2. Provide language laboratory so that the students are able to practice their English particularly in listening skill.

Table 3 .
Test Specification