2.1.6. Data Collection Procedures
This experiment investigates the following questions:
· Are linguistic and cultural knowledge prerequisites
to learning translation? More specifically,
o Could translation scores be explained by prior scores in
language and cultural disciplines?
o Is there a positive relationship between prior scores in
language and cultural disciplines and subsequent translation scores?
o What is the strength of this relationship? In other words: Is
it systematic?
To answer these research questions, three sets of data were
collected. The first represented the dependent variable: translation
competence. The 30 highest and the 30 lowest grades in the le
Arabic-English-Arabic translation exam were recorded. These scores represent
students' performance in only one test: that of the third year. For the sake of
validity, the mean of each student's third and second years' scores was
calculated. Then only the 22 highest and the 22 lowest means were kept for
analysis. (See appendix C for all sets of scores concerning both groups).
The second set of data concerns the first independent
variable, namely students' linguistic competence as measured on the
Baccalaureate exams. It is represented by the mean of each student's score in
Arabic and in English exams. The scores were not used separately because this
study is not concerned with the effect of each language competence apart. It is
rather interested in overall linguistic competence. This is why individual
means were obtained from each pair of English and Arabic scores.
The third set of data represents the second independent
variable, namely students' cultural knowledge. As stated earlier, it is the
mean obtained from three scores: History and Geography, Philosophy, and the
general Baccalaureate mean. It is assumed that the obtained scores would
measure the students' achievement in the academic disciplines studied during
the third year of secondary school, with a specific importance given to the
mentioned disciplines. (See Appendix C)
2.1.7. Statistical Analysis
Data collected within the ex post facto experiment
were quantitative data. This, obviously, called for quantitative analysis. Two
different types of analysis were used to answer each of the research questions.
Statistically speaking, these questions read as follows:
a. Is there a significant difference between group A's and group
B's language means?
b. Is there a significant difference between group A's and group
B's culture means?
c. Is there any correlation between scores in translation and
prior scores in language and culture?
d. What is the strength of this relationship? In other words: Is
it systematic?
Indeed, a T-test is used to investigate questions (a)
and (b), and Pearson product-moment correlation coefficient, questions
(c) and (d).
2.1.7.1. Means Comparison: Research Questions (a) and
(b)
To answer these research questions, the following hypotheses
were
set:
Hia: Group A's language mean is significantly higher than that of
group B.
Hlb: Group A's Culture mean is significantly
higher than that of group B.
A T-test was used to compare each independent variable pair of
data. Therefore, the null hypotheses (Ho a) and (Ho b) were stated as
follows:
Hoa: Group A's language mean is not significantly higher than
that of
group B.
Hob: Group A's Culture mean is not significantly higher than
that of group
The next step in the analysis procedure was the choice of a
probability of error level (alpha level). The decision was set at a
conservative level i.e. a < 0.01. This decision implies that the present
study tolerated no more than a probability of 1% that the differences between
means be chance occurrence. The type of significance level this study adopts
was directional (as opposed to non-directional). This means that the
study expected one specific mean to be higher than the other. Thus the study
predicted the direction of the difference.
| 
