revealed the contribution of web-based grammar practice to their grammatical competence and writing skill.
D. Data Processing
This study conducts a correlational research. Correlational study is conducted to determine to what degree two variables are related Creswell, 2008.
In this study, the variables are between web-based grammar practice and grammatical competence, and between web-based grammar practice and writing
skill. Therefore, the purpose of this study is to determine the degree of the relationship between the two variables; in this case, to what degree web-based
grammar practicecorrelates to students‟ grammatical competence and writing
skill. All of variables are measured and converted into numerical data or scores using testing technique, which is believed to be valid and reliable. Since the
variables are measured using testing technique, the data are converted into scores; it means that the variables are continuous.
In measuring the degree of association or relationship, this study employs correlation statistical test using SPSS 21, which is Pearson Product Moment.
Before investigating the relationship between the two variables using Pearson Product Moment test, several assumptions underlying this test are checked first.
There are three assumptions, which are scale assumption, normality assumption, and linearity assumption.
1. Scale Assumption Scale assumption means that the data should be interval or ratio data. This
correlational test required two sets of continuous data. The data are in the form of numbers, and thus, they are already in continuous scale.
2 Assumption of Normal Distribution Normality assumption means that the data should be normally distributed.
To check whether the data are normally distributed or not, tests of normality and the normal Q-Q plot are needed. The Shapiro-Wilk Test of Normality was
selected since the samples were less than 50. If the significance value of the Saphiro-Wilk Test Sig is greater than
α = 0.05, the data are normally distributed. Q-Q Plot shows the percentiles of a standard normal distribution against the
corresponding percentiles of the observed data. The data are normally distributed if the resulting plot is roughly straight line with a positive slope.
3 Linearity assumption To check the linearity, the data should be plotted using scatter plot. The
scatter plot should be more or less in a line. Scatterplot shows a clear direction of the relationship between the two variables. The illustration of the features of the
correlation is presented in scatter diagrams or scatterplot. A scatter diagram plots each individual case on a graph which shows the points at which two variables
intersect. Those points show the position of one variable to another variable.
Figure3.1 Positive Correlation
Rougly straight line falls directly on a line with an upward incline, or with positive directtion, shows positive correlation, in which high values of web-based
grammar practice are associated with high values of grammatical competence and writing skill. In this case, it means that the higher web-based grammar practice
variable is, the higher grammatical competence and writing skill variable will be. Roughly straight line falls directly on a downward incline, or with negative
direction, shows negative correlation, in which high values of web-based grammar practice are associated with low values of grammatical competence and writing
skill. It means that high web-based grammar practice variable is not followed by high grammatical competence and wrtiing skill variable.
Figure 3.2 Negative Correlation Meanwhile, no correlation means that the values of web-based grammar practice
are not at all predictive of the values of the grammatical competence and writing skill.
Figure 3.3 No Correlation
After checking the three assumptions, the correlation coefficient r can be calculated using Pearson Product Moment test in SPSS 21. Since it is expected
that web-based grammar practice have positive contribution to students‟
grammatical competence and writing skill, then one- tailed test is used. Pearson‟ r
varies between -1 and +1. This sign determines whether the correlation is positive or negative. Cohen and Holliday 1982 suggest the strength of relationship as
follows: PLAGIAT MERUPAKAN TINDAKAN TIDAK TERPUJI
Table 3.2 Correlation Coefficient r
Pearson‟s r Indication
0.19 Very low
0.20-0.39 Low
0.40-0.69 Modest
0.70-0.89 High
0.90-1 Very high
The closer r is to 1 whether positive or negative, the stronger the relationship of the two variables. The nearer r is to zero or the further it is from
+1 or -1, the weaker the relationship. In this study, it is expected and predicted that the direction of the relationship between the two variables is positive. After
finding the correlation coefficient, then the next step is hypothesis test. Hypothesis test is conducted to check whether the high correlational coefficient is
due to chance or due to the significant relationship between the variable. Inferential statistically, the positive relationship between the variables can be seen
from the value of Sig. α H : ρ=0; H
1
: ρ≠0. The null hypothesis is formulated to state that there is no correlation between the two variables, while the alternative
hypothesis stated that there is a significant correlation between the two variables.
E. Data Analysis and Verification