The total number of the second semester students in Sanata Dharma University was 150 students who ranged from 17-20 years old. This number of
students were divided into six classes of Writing II. However, since it would not be possible to involve the entire classes of the entire population of the students,
the researcher applied random cluster sampling to select the subject of the study. Random cluster sampling is occurred when a researcher chose a class from a list
of classes randomly and used all the members in that class as the sample in his or her research Ary et al, 1990: 175. It was chosen one class randomly, which is
class D, as the representation of the second semester students in Sanata Dharma University. There were 28 students in this class. Of these 28 students, 18 were
female and 10 were male. In this class, there were three students who were retaking writing II. Since the subjects of this study were the second semester
students, the three of them could not be included in the subjects. Therefore, the total number of the subjects would be 25 students.
C. Setting of the Study
The study was conducted at the English Language Education Study Program of Sanata Dharma University. It was conducted during the period of the even
semester in the academic year 2006. The researcher collected one of the assignments from the students of Writing II class, especially class D.
D. The Nature and Source of Data
In order to gather the data about collocational errors in Writing II, the researcher collected Writing II assignments from the students. In these
assignments, the students had freedom to choose one of the topics given. The assignments were compositions about news report. The students were asked to
write a news report by choosing one of the topics given. The topics were Kidnap Attempt, Tsunami Disaster, Bomb Terror and Plane Crash. The students were not
told that their use of collocations would be studied. Had they been told, they might have underused or overused such word combinations.
Writing II assignments were the documents to be analyzed in this research. Document analysis is the analysis of the written or visual content of a document
Frankel and Wallen, 1993: 389. When using document analysis, it must be established the authenticity of the documents itself, as well as the validity of its
content. The assignments were regarded as authentic texts since they were the students’ productions, which were written by their own hands. Besides, the
assignments were considered valid because the assignments were the first draft in which the lecturer had not yet revised. As the compositions were written as part of
their weekly writing exercise, the language they produced was linguistically natural and there were so many grammatical errors. However, since this study
focused on the collocational errors, the grammatical errors in the students’ writings were ignored. It would be the word combinations in the students’
writings as the subject to analyze in this study.
E. Research Instruments
In this study, it was not used any instrument to gather the data since the data was gathered by collecting the students’ writings. In analyzing the collocational
errors, the researcher used Simple Concordance Program SCP. A Simple PLAGIAT MERUPAKAN TINDAKAN TIDAK TERPUJI
Concordance Program SCP is designed to help the researcher look through computer readable text files for the occurrence of key words. When each word is
found, SCP will display the context in which the word is used, and also gives a reference to where that word occurs in the text. In the SCP, there is an option that
can be used to analyze the text, which is KWIC Key Word- in Context option. As seen from the concordance lines, the main adva ntage of the KWIC is that it makes
it easy for the investigator to scan rapidly to identify typical combination collocational patterning. SCP will also show the vocabulary used by the author
and the frequency with which words are used. These can be displayed in alphabetic order, as in a dictionary, or according to their frequency of use. SCP
will also produce a word frequency profile and some statistical properties of the texts. In this study the researcher used SCP version 4.05, which was designed by
Alan Reed 1997-2001.
F. Data Collection