45
2. Data validity and reliability
An instrument can be said to be a good one if it is valid and reliable. Before the instruments are used, they have to be tried out. It is intended to
find out the validity and reliability of the instruments.
a. Validity
An instrument can be valid if it can reflect what is being measured Arikunto, 2002:145. Cooper and Schindler 2003:231 say that validity
refers to the extent to which a test measures what we actually wish to measure.
In this study, to verify the item validity of instrument, each item of the test is correlated with the total score by using Point Biserial Correlation formula.
The item of the test is considered valid if the result of the correlation coefficients r
xy
is as many as the r table of product moment. The number of students joining the try out is 28, with the significance level
α = 0.05 the r table is 0.374
.
The item of the test is considered not valid if the correlation coefficient is lower than r table.
The criterion is as follows. r
o
r table = valid r
o
r table = invalid Budiyono, 2000:69
The results of the try out: 1
From 45 items of the test of vocabulary mastery, 25 items are valid and 20 items are not valid.
46
2 From 50 items of the test of reading comprehension, 20 items are valid and
30 items are not valid.
b. Reliability
Nunan 1992:14 says that reliability refers to the consistency and of the results obtained from a piece of research. Sugiyono 2006,states that
reliability refers to the level of internal consistency or stability of the test over time.
To verify the reliability of the test, the writer uses the Alpha Cronbach formula.
Based on the analysis using the formula above, the result of reliability coefficientof the test of vocabulary mastery and the test of reading
comprehension are 0.817 and 0.853. it can be said that the instrument are reliable. The computation of reliability test of the two variables can be seen
at Appendices 16 and 18.
E. Data Analysis Technique
As stated above, this research tends to know the contribution of the independent variable to the dependent variable. In order to achieve that
purpose, it must be known from the relationship of the variables. The research tests the hypothesis using Product Moment and Multiple Linear regression
formula. According to Borg and Gall in SuharsimiArikunto, 2002:251, Product Moment is used to describe the strength of relationship between two
variables, while Multiple Linear regression is used to describe the strength between one independent variable and one dependent variable. In this