Table 4.26 Results of MGT question “c” by education level – SI test: chi-square
Table 4.27 Results of MGT question “c” by education level – JM test: chi-square
4.4.4.4 Summary of education variable
In summary, it appears that education level does indeed have a statistically significant impact on language usage and attitudes in Jambi, as seen in the ANOVA results from both the questionnaire and MGT. In the
individual questionnaire questions analyzed, though, not all questions showed statistically significant chi- square results. But all questions did follow the trend that the higher the education level, the less Jambi
Malay usage and the lower the attitudes towards it.
Though the matched guise test composite results showed significant differences among the education levels, the individual questions examined with the chi-square test were less clear. Only the status
question revealed a significant impact of education level on answers given, with a predictable pattern. This pattern—the higher the education level, the lower the number of positive answers towards Jambi Malay—
was similarly seen in the questions for Indonesian, however, which was not as expected. Thus, Jambi
Malay language attitudes according to education level as determined by the MGT are difficult to tell with precision.
4.4.4.5 Results: College
graduates
4.4.4.5.1 Questionnaire Data were also gathered from twelve informants in Mudung Laut who had already finished university level
education. Since college graduates are not very representative of the whole population, their results were excluded from the main education analysis above, but will be examined here for comparative purposes. It
must be stressed that results from only twelve informants, compared with much higher numbers of those with lower education levels, may not be very reliable.
In figure 4.9, one can see the placement of college graduates’ scores along the continuum of scores according to education level for the questionnaire +C = completed college. With a mean questionnaire
score of 1.308, compared to the next education level finished SMA 1.354 it is clear that the questionnaire scores were substantially lower for the college graduates. The ANOVA results in table 4.28 show that the
college scores were statistically significantly lower F statistic 23.06, p .0001, especially when paired with those with no education. For the questionnaire, then, these results strengthen the idea presented above
that the higher the education level, the lower the scores—and in effect, the lower the usage of Jambi Malay and attitudes towards it.
Figure 4.9 Average questionnaire scores by education level including college
Table 4.28 Questionnaire results by education level including college test: one-way between subjects ANOVA
4.4.4.5.2 Matched guise test In the matched guise test results, the trend was the same for attitudes towards Jambi Malay see figure
4.10. The college graduates had the lowest average score of all 1.463, at the bottom of a steady decline in attitudes towards Jambi Malay as the education level rises. And like the questionnaire, the differences
between college graduates’ scores and the other categories were statistically significant F statistic 4.39, p. 0018 according to the ANOVA test. This was especially the case between those with no education and
those who finished college see table 4.29.
Figure 4.10 Average MGT scores by education level including college – JM
Table 4.29 MGT results by education level including college – JM test: one-way between subjects ANOVA
The MGT results for Indonesian among college graduates were similar to the results for Indonesian among those with less education, presented in the previous section. It was expected at the outset
that the higher the education level, the higher the MGT scores, meaning the more positive attitudes towards Indonesian. However, as seen in figure 4.11, the college graduates’ average score was not the highest
1.567. In fact, they were the second to the lowest, after those who finished SMA 1.553. The highest scores were found among those with no education. And, although according to the ANOVA test the
difference in scores between education levels were statistically significant p .0009, there was no significant contrast between the college graduates’ scores and others, specifically see table 4.30.
Figure 4.11 Average MGT scores by education level including college – SI
Table 4.30 MGT results by education level including college – SI test: one-way between subjects ANOVA
4.4.4.5.3 Summary of college graduates College graduates’ questionnaire results and the results of the MGT for Jambi Malay were similar to the
main education level analysis no education through finished SMA. The same overall trend was seen: the higher the education level, the less usage of Jambi Malay and the less positive the attitudes towards it.
It was also found that the MGT results of the college graduates for Indonesian echoed results of the main education analysis, in that the average scores of those with lower education levels were
surprisingly high, and the scores of those with higher education were surprisingly low. Although the data from college graduates, with just twelve informants, is minimal in comparison, it lends more credence to
the findings of the main education analysis since the results are not contradictory.
4.4.5 Results: Sex variable