THE USE OF CLUSTERING TECHNIQUE TO IMPROVE THE STUDENTS’ SKILL IN WRITING DESCRIPTIVE PARAGRAPH

Writing process involves thinking and creative skills. To stimulate the students’ thoughts to express their ideas, clustering technique is effective brainstorming activity to help the students explore their ideas and write consistently. Thus, the objective of this study is to find out whether or not there is any significant difference between the students who were taught in writing descriptive paragraph through Clustering Technique and those who were not. The population of this study was all of the eighth grade students of SMP Srijaya Negara Palembang in academic year 2016/2017. There were three classes that consist of 113 students. The samples of this study were VIII C as a control group and VIII B as an experimental group. This study was conducted by using quasi experimental method with nonequivalent control group design. Then, the data were analyzed by using paired sample t-test and independent sample t-test. The result of independent sample t-test shows that the value of sig. (2-tailed) was 0.00 < 0.05 in the level of significant 0.05. While, the value of t-obtained was 5.213 higher than the critical value of t-table was 1.665 with degree of freedom 76. It can be concluded that the null hypothesis was rejected and the alternative hypothesis was accepted. It means that there was any significant difference between the students who were taught in writing descriptive paragraph through Clustering Technique and those who were not.


Introduction
Writing has always been an important component of study for language leaners and today it has become increasingly important.It is the area in which leaners are expected to be offered adequate time to develop their writing skill.Rao (2007) pointed out that writing enhances people's thought processes and the organizing ideas, developing the ability to summarize and criticize.In addition, writing is really seen as a process of discovery as the writers try to find their way while they are struggling to think, compose, and put their ideas together (Ismail, 2011).
Unfortunately, writing is the most difficult language skills.According to Miranda (2003), writing is more difficult than other language skills even if the language used is their first language.In writing, students need to pay attention carefully to use English in writing because the students are expected to set down the words, phrases, sentences, and paragraphs in the appropriate structure, composition, and construction in writing.For example, the students are unable to express their ideas and organize them in the written form.These problems can be influenced by several factors derived from internal and external factors of students.The internal factors that can affect the students' writing are IQ, cognitive ability, talent, and motivation.And the external factors that can affect the students' writing are curriculum, material, approach, method, strategy, and technique.So that teacher should be creative in choosing the strategies for teaching.
To solve the problems above, teachers should choose an interesting topic.An interesting topic may motivate students feel free to write what is in their thoughts to be expressed in the written form.Clustering Technique is one of alternative to stimulate the students' thoughts in writing activity.According to Oshima and Hogue (1999), clustering technique is another brainstorming activity that can be used to generate ideas in writing.This technique is also known as diagramming or mapping.Moreover, the clustering technique can motivate the students to write and to stimulus their ideas.This technique is an effective tool for the prewriting stage of the writing process.In clustering technique, there are lines, boxes, arrows, and circles to show relationships among the ideas and details.It is added by Pharr and Santi (2005) in clustering the students write a topic in the center of a piece of paper, then write ideas suggested by the topic around it, connecting these to the topic with lines and with their subtopics.So, this technique can help the students to stimulate their ideas and to organize them before they develop into paragraphs.
This research focused on descriptive text which is referring to the syllabus of school -based curriculum (KTSP) this kind of text is taught at the eighth grade of junior high school.Based on basic competency of writing skill, the students could describe things, places, and a person in detail in written form.The fact, writing descriptive paragraph is still complexity.They do not have any ideas when they are asked to describe them and also emerged choosing the topic, arranging paragraph and using of vocabulary.Thus, clustering technique is introduced to the students to stimulate them to discover and become engaged in a topic.In order to achieve the purposes mentioned above, this research focuses on the following question is there any significant difference between the students who are taught in writing descriptive paragraph through clustering technique and those who are not.

Literature Review Writing Skill
Barly (1995) stated that writing is an active process to organize and formulate the ideas on the paper.It is a complex activity with the control language both of the sentences level (grammatical, structure, vocabulary, punctuation, spelling, and later information) and beyond the sentence rank (organizing and integrating information into cohesive and coherent paragraph or text).
Writing is a process that involves several steps.According to Harmer (2007), there are five steps in writing process; as follows: 1. Prewriting Prewriting is a way to get started to write which includes activities of finding topics, generating ideas, planning content and organization.This study focused on this initial stage of writing.This activity requires the student to gather, obtain, or develop ideas for the writing successfully.
2. Drafting Drafting is a kind of practice exercise for the final draft.A draft is based on the planned ideas made in prewriting activity.In drafting, the writer elaborates our ideas in complete sentences and paragraphs.If the first draft is not a perfect so the writer need to do revision.

Revising
Revising means evaluation of the draft that has been made.In revising, the draft may decide to move, delete, or add words, sentences, paragraphs, that will complete and improve the content writing.
4. Editing Editing can be done systematically which means the writer pays attention to the correct grammar, spelling, punctuation, capitalization.Also, this activity is time consuming and need patience to identify the mechanical errors.

Publishing
Available online at: http://jurnal.um-palembang.ac.id/englishcommunity/indexISSN 2549-9009 (print), ISSN 2579-7378 (online) Publishing is an activity of sharing the finished writing with others.The final writing can be published after qualifying preceding processes or stages of writing.
There are five components in scoring writing test based on Brown (2004) as follows: 1. Content Content is about logical development of ideas.The ideas concrete and thoroughly developed or not, whether the essay addressed the issue or not, is there any extraneous material present or not in the text.
2. Organization Organization consists of introduction, body, and conclusion.It's about how the writer makes a good relationship between the titles, introductory paragraph, and the topic, the body of paragraph, generalization, and conclusion.It's about whether all of the components above are support each other and composed orderly by the writer or not.

Vocabulary
Vocabulary is about style and quality of expression.How the writer use precise vocabulary, parallel structures, word choosing, etc.

Grammar
Grammar is about all the rules language application used by the writer.For example, native like fluency English grammar, correctness of using relative clause, prepositions, modals, articles, verb forms and tense sequencing, etc.

Mechanical skill
Mechanical skill is about the punctuation, spelling, and all of the graphic convention of the language.It's about correctly using English writing conventions; all needed capitals, paragraph intended, using coma, fullstop, etc.

The Concept of Descriptive Paragraph
Descriptive paragraph is a part of factual genres.Its social function is to describe a particular person, place or thing.Based on Zemach and Rumisek (2005) stated that descriptive paragraph explains how someone or something looks or feels and a process to explain how something is done.In addition, Oshima and Hogue (2007) analyze that descriptive paragraph is also the ability to describe people, places, or objects accurately is a useful life skill.Furthermore, Mayer (2011:5), "descriptive paragraph has the generic structures namely; identification, description and conclusion".Identification is the part to identify or write the recognition; moreover, it can be a general statement about place, person, or thing that wants to describe.Descriptions or main body of the paragraph are vivid details of place, person, or thing that are described so that the reader can easily imagine or picture the object, or they can feel the object with the five senses.

Clustering Technique
Clustering is a kind of technique which can be used in pre-writing stage.It is alternative brainstorming activitiy before starting to write.Rico (2000:28) also claims that "Clustering is a nonlinear brainstorming process akin to free association.It makes a Design-mind process visible through a nonlinear spilling out of lightning associations that allows patterns to emerge.Through clustering we naturally come up with a multitude of choices from a part of our mind where the experiences of a lifetime mill and mingle".Using this clustering technique indicates that it is effective to stimulate the right brain and explore the image to easier the students generate the ideas in writing process.Baroudy (2008) defines clustering as a non-linear brainstorming process that generates ideas, images, and

Method of the Research
The method of the research was used a quasi-experimental design to find out the difference of students' skill in writing descriptive paragraph through clustering technique and that those who were not.
The design involves an experimental group which is given the treatment by using clustering technique and control group which was not given the treatment at all.The participants of this study were eight grades students of SMP Srijaya Negara Palembang in academic year of 2016/2017.To select the samples, purposive sampling technique was applied because these classes have similar characteristics which are that each class has the equal numbers of students consisting 39 students and the writing level of most students can be classsified as low level.In conducting his study, samples were divided into two groups: experimental group (VIII B) consisting 39 students and 39 students for control group (VIII C) Before giving the treatment, pretest was given to the experimental and control group.The treatment that using cluster technique was conducted in experimental group and no treatment for control group, those activities were implemented for ten meetings including pretest and posttest.Writing test was employed to collect the data from the subjects of this study.The test was relevant to their curricula and appropriate with standard competence and basic competence.To validate the test, content validity was established to measure the writing test by observation and judgment from the experts to assess the test items have the high degree of content validity.In judging the students' writing score, analytical scoring rubric adapted from Brown (2007) which has five components that are content, organization, grammar, vocabulary, and mechanics, was applied to examine the students writing desriptive paragraph.Reliability of the test was applied to judge the accurateness and consistency of the writing test score.In this case, the raters are provided to examine whether the writing score is reliable or not by using inter-rater reliability that was analyzed by Spearman Brown.The result of reliability coefficient of the test for control group was 0.795, while experimental group was 0.911.The result of reliability coefficient of tests indicate preferably higher than 0.70, it means that is considered reliable.
For analyzing the data, paired sample t-test was used to obtain the information dealing with the results of pretest and posttest to find out the significance differences of the tests before and after given the treatment through clustering techniques and those who were not.The results of post test between experimental and control groups were analyzed by the independent sample t-test in order to find out whether or not there is a significant difference between the students who are taught to write a descriptive paragraph through Clustering Technique and those who are not.

Findings The Results of Pretest and Posttest for Experimental Group
In experimental group, the lowest score of pretest was 25, there were 19 students in category very poor.While the highest score obtained by students was 93 with standard deviation 18.67, there were 2 students in excellent category.In post test, the lowest score is 57, there were 11 students in fair category and no student in very poor category and the highest score is 100 with standard deviation 12.40 that obtained by 14 students in excellent category.These results is displayed in figure 2. The distribution of score in pretest and posttest control group showed that the lowest score in pretest was 25, there were 25 students in very poor category, and the highest score was 90, there was one student in excellent category.While the lowest score of postest was 32 that there were six students in poor category and the highest score was 100 in excellent category was four students.The result of pretest and post test for control group is displayed in figure 3.  Based on the results of paired sample t-test, the data indicated that the use of clustering technique was effective to improve the students' scores in writing descriptive paragraph.This is proven that t-obtained 16.386 was higher than t-table 1.685.

Statistical Analysis of Independent sample t-test
In order to find out whether or not there was a significant difference in achievement between the experimental group and control group, independent sample t-test was applied to compare the results of the posttest in the experimental   The result of independent sample t-test showed that the value of t-obtained was 5.213 and the value of t-table was 1.665.At the sig (2-tailed) was 0.00 < 0.05.The result of analysis of the research shows the value of t-obtain is higher (5.213) than the value of t-table (1.665).The table also showed that the p-output of students' skill in writing descriptive paragraph in both groups was lower than the level of significant 0.05.It means that there was a difference in both groups.It claims that the null hypothesis (Ho) was rejected and alternative hypothesis (Ha) was accepted.It means that there was significant difference between the students who were taught to write a descriptive paragraph through Clustering Technique and those who were not.

Discussion
The computation result of the scores gained by the students in the posttest indicate the use of clustering technique was effective to improve the students skill in writing descritive paragraph.By implementing this technique, the students could explore their ideas relevant to the topics of paragraph and organize the ideas to transform them into the writing product.This is in line with Pica's (1986, as cited by Baroudy, 2008) proposition that clustering technique is very helpful to unleash ideas, images, and feelings related to the main topic in writing.In addition, clustering technique has assisted them to recognize clear association among the ideas which have been generated and it is another brainstorming activity before writing stage (Oshima and Hogue, 1999).
Clustering technique is considered very effective to implement it to assist the students use the keywords that may use in their writing.By giving the keywords, that the students can organize their ideas in determing the generic structure which they wrote into descriptive paragraph and it is also helped the students to explore their imagination to connected with topic in writing descriptive paragraph.
The students who were taught by using Clustering Technique showed their progress in drawing up idea into their writing while given and after the treatment.Before given the treatment, they did not know how to draw up idea, how to deliver in a good manner, and how to make the conclusions.But, with the treatment using Clustering Technique, they could think more critically than before.By learning the problem given, they could enhance their knowledge and drew it up into their writing.Moreover, they were treated to ask and deliver their opinions and arguments with their group.They learned how to make a good teamwork and how to increase their independent work through that good teamwork.According to the statement, it could be said that clustering is a good way to develop idea before starting the writing paragraph.

Conclusion
Clustering technique is proven to be effective in elevating students' score in writing descriptive paragraph.It is helpful for the students to stimulate their ideas and organize the paragraph.Having conducted the research, it can be found that the students have a fun

Figure 3 .
Figure 3.The results of pretest and post test for control group : http://jurnal.

Table 1 .
Statistical Analysis of Paired sample ttestThe result of paired sample t-test in mean between pretest and post test of writing descriptive paragraph in experimental group was 28.205 with standard deviation 10.761 at the significance level 0.05 and degree of freedom 38, the critical value of t-table for the two tailed test 1.685, t-obtained 16.386 was higher than the critical value of t-table 1.685.It is displayed in table 1.The result of paired sample t-test in mean between pretest and post test of writing descriptive paragraph in control group was 14.949 with standard deviation 11.932 at the significance level 0.05 and degree of freedom 38, the critical value of t-table for the two tailed test 1.685, t-obtained 7.824 was higher than the critical value of t-table 1.685.It is displayed in table 1.The results of paired sample t-test

Table 2 .
The results of Independent sample t-test