do this, the researcher used the dependent t-test. Figure 3.2 shows the formula for the dependent t-test.
̅ √
Figure 3.2 The Dependent t-Test Formula
where: t
: t ratio ̅
: average difference : different scores squared, then summed
: different scores summed then squared N
: number of pairs
F. Research Procedure
The researcher began the study by stating the null hypothesis and alternative hypothesis. Having decided the two hypotheses, the researcher went on
to choosing the subjects of the research, who were one class of first semester students of the English Extension Course of the 20112012 academic year. As the
reader has read in Theoretical Description in Chapter II, the researcher did not randomly assign the subjects into two groups as the research required only one
group of subjects, following the one group pretest-posttest design by Cohen, Manion and Morrison 2000.
Hereafter, the research subjects would learn how to guess the meanings of the unknown words using the context in which the words were found. The
researcher would introduce them to Clarke and Nati on’s procedure for guessing
from context. The subjects would then do some tasks, both in groups and individually, in which they had to apply
Clarke and Nation’s procedure for guessing from context. This step would last for two meetings.
After two meetings, they were posttested using the same material. Their scores on the pretest and posttest would then be compared and analyzed using the
t-test in order to see if they had improved their guessing from context skills. The result of the analysis would answer the first question of the thesis. The researcher
would distribute the questionnaire to the group to see how the subjects view Clarke and Nation’s procedure for guessing from context as a vocabulary learning
technique. To answer the last question of the study, the researcher would also analyze
and examine the research subjects’ posttest results to see in which stage of the
procedure the students’ mistakes occurred the most.