Validity Reliability The Difficulty of an Item

30 the lecturing technique for the whole material by the teacher. After getting the test result, the writer scores the students’ work. 4 Comparing both difference between pre-test and post-test result of both experimental and control group by applying t-test formula. 5 Making conclusion of the result.

3.8 Method of Collecting Data

3.8.1 Try-Out

The quality of data, whether it is good or bad, is based on the instrument to collect the data. A good instrument must fulfil validity and reliability. So before the test was done, the writer had tried out the instrument. The try-out was held on January 12, 2011. It was administered to the students in same level as population. 20 students were taken as the subject of the try-out. They were given 45 minutes in doing the try-out test. After administering the try-out, the data was analyzed to find out the validity and reliability of the test.

3.8.1.1 Validity

Dealing with validity, Heaton 1979: 152 proposes that validity of a test is the extent to which it measures what is to measure. Every test, whether it is a short, informal classroom test or public examination, should be valid as the constructor makes it. The result is consulted to critical value for r-product moment. If the obtained coefficient of correlation is higher than the critical value for research- 31 product moment, it means that a paragraph is valid at 5 alpha level of significance. To calculate the validity of a test, the writer used this formula: } }{ { 2 2 2 2 ∑ − ∑ ∑ − ∑ ∑ ∑ − ∑ = y y n x x n y x xy n r xy In which, r xy = the correlation of the two scores on two halves of the test n = the number of the students x = the score of each component y = the sum of all paragraphs’ scores ∑ xy = the sum of multiple score with the total score ∑ 2 x = the sum of the square score in each component of writing scoring ∑ 2 y = the sum of all paragraphs’ score square Arikunto 2006: 170 In this writing, there are five aspects that would be analyzed. They are grammar, vocabulary, spelling, content and fluency.

3.8.1.2 Reliability

Heaton 1979:154 states, reliability is necessary characteristic of any good test: for it to be valid at all, a test must be reliable as a measuring instrument. Reliability is a consistency of its measurement whether it was administered in different occasion. In this study, the reliability of the test was measured by comparing the obtained value with research-value product moment. That is, if the obtained value is higher than the table r-value, it can be said that the test is reliable. 32 To calculate the reliability of the test, the writer used this formula: In which index reliablity number of item item variance total variance Arikunto 2006: 196

3.8.1.3 The Difficulty of an Item

The index difficulty of an item shows how easy or difficult a particular item proves in a test Heaton 1975: 172. It is usually express as the fraction or percentages of students who answer the item correctly Arikunto 2006: 211 states that difficulty level is a number that shows the easy and difficulties of an item. A good test is a test which is not too easy and difficult. The calculation as follow: Where, P = difficulty index B = number of students getting item correct JS = total number of respondent Arikunto 2006: 211 33 He classifies the difficulty level of an item as follows: Item with P 0.00 – 0.30 is a difficult item Item with P 0.31 – 0.70 is a medium item Item with P 0.71 – 1.00 is an easy item

3.8.1.4 Discriminating Power