Correlation Coefficient The Analysis of the Data

C. The Discussion of the Findings

Even though the reviewed literature has shown that movies could be beneficial for language acquisition and the building of language competence and skills, the findings of the research show the index value of correlation coefficient of 0.177 which is very close to zero. According to Sudiyono, the index correlation value between 0.00 – 0.20 shows very lowweak correlation, so this correlation is usually ignored. 2 The comparison between and the index value of Product Moment table also shows that , which means the null hypothesis is accepted. The scatterplot of the correlation shows a positive relationship; with the regression line tends to the right side. However, the dots are evidently far from the line, which makes the relationship weak. Figure 4.7 To explain this phenomenon, the internal threats to the validity of the questionnaire scores were explored. Even though many respondents’ total scores are relatively high, their scores per item show threats to the validity of the total score. Several findings were found: 1 From the items number 8, 9, 10, 11, and 12, which ask about the kind of input that the respondents usually received by watching movies see Appendix 4, the average scores are relatively high 3.8, 4.5, 3.9, 3.7, and 3.6see Appendix 5. This means that the inputs are rather comprehensible. However, from item number 2, which asks about the number of movies that the respondents watch per month, the average score is 2.7. This shows that the amount of exposure that the respondents receive is actually low, which lessens the probability of an input being present. 2 The kinds of movies that best provide input and build listening skill are the ones with no subtitles. 3 However, from the item number 4, which asks whether the respondents use no subtitles see Appendix 4, the average score is 3. It means that the respondents still have quite a high tendency to watch 2 Anas Sudiyono, Pengantar Statistik Pendidikan, Jakarta: PT RajaGrafindo Persada, 2005, p. 193. 3 Penny Ur, Teaching Listening Comprehension, Cambridge: Cambridge University Press, 1984, p. 66. English movies using subtitles. This, of course, means that the inputs are not optimally received because they read the subtitle rather than listen to the dialogues. Moreover, 60 of respondents choose to use Indonesian subtitle see Figure 4.5. 3 The findings in point number 2 then may directly affect the findings in point number 1. If the respondents use subtitles —and the ones that are in Indonesian, their per ception of ―understandability‖ of the language in the movies may be biased. For more explanation regarding the subtitle use, the respondents’ answers for item number 4 were compared. The item asks “How often do you watch English movies without using subtit le?” None of the respondents answered it with Almost Always, 10 respondents answered Frequently, 11 respondents answered Sometimes, 7 respondents answered Rarely, and 2 answered Almost Never. Table 4.8 The Respondents’ Answers for Item Number 4 Groups of Respondent Answer Respondents 1 Almost Always - 2 Frequently 5, 8, 11, 15, 16, 17, 20, 21, 23, 24 3 Sometimes 2, 3, 4, 7, 10, 14, 19, 25, 27, 28, 29 4 Rarely 1, 6, 9, 12, 18, 22, 26 5 Almost Never 13, 30 Next, the listening scores from group 2 and group 4 were compared. After the sums of listening scores from both groups were calculated, the result showed that group 2, who answered Frequently, has the average listening score of 82.75. Meanwhile, group 4, who answered Rarely, has the average score of 80.5. This result may imply that the respondents who frequently watch non-subtitled movies indeed have better listening skill than the respondents who rarely watch non- subtitled movies.