INSTRUMENT Construction of the Instrument

26 from teachers so they had the same opportunity to be used as sample. It also means that each of the students as the member of the population had equal chance of being included in the sample. It has been mentioned before that the writer took 60 students as samples or 50 of the total number of the population. To take those 60 students as sample, the writer used a proportional random sampling because she took the number of students from each class proportionally. After deciding the number of samples of each class, she started to do the random sampling. First, she gave numbers to the students 1-30. Second, she wrote the numbers on small pieces of paper. These small pieces of paper were rolled and put into a box. After being mixed, the pieces were dropped out of the box one by one. The same procedure was done four times until 60 students were obtained.

3.2 INSTRUMENT

All research studies involve data collection. In collecting the data, research instrument is very important in which the validity and reliability of the instrument will influence the validity and reliability of the data obtained. In general, instruments used to collect the data can be divided into two types, a test and non-test Arikunto, 1992:122. A test is a written set of questions to which an individual responds in order to determine whether he or she passes. According to Gay, as quoted by Hapsari, 1998:45, more inclusive definition of a test is a mean of measuring the knowledge, skill, feeling, intelligence, or aptitude of an individual or group. Excluded from the test is said to be a non-test. 27 In collecting the required data for the analysis in this study, the writer used a test as the instrument to measure the knowledge of the respondent in grammatical structure at the level of question tags. Based on the test, she identified, classified and analyzed the numeral scores of the result of the test. To collect the data, the writer used only one type of the test that was multiple-choice test. She conducted this kind of test because multiple-choice type test is easy to be scored and also it does not take much time to score because it just has one correct answer. Harris, 1967:7, points out that because of the highly structured nature of these multiple-choice tests, the writer can get directly at many skills and learning she wishes to measure. This statement affirms that besides being easily arranged, a multiple- choice test will show the occurrence and frequency of the specific errors right away. Moreover, based on the nature of this study which was designed to find out some specific difficulties of question tags, the writer assumed that multiple-choice test would be the most appropriate type of the test.

3.3 Construction of the Instrument

As has been previously mentioned that the writer chose multiple-choice test, to be the type was applied in the research. The test consisted of 111 items with four options A, B, C and D. Each item represents a certain mastery of question tags in which the students might have difficulty. The range of vocabulary in this test included words that might have never been practiced by the students. In 28 the instrument the materials were arranged suitable for the third grade students of SLTP Institute Indonesia in the academic year of 20052006.

3.4 Preparation of Test