5.4.1 Age variable
From the discussion on questionnaire findings according to the age variable, we can say that age does indeed have a strong impact on the amount of JM spoken, and, to a lesser degree, the self-claimed attitudes
towards it. Considering the questionnaire results, then, null hypothesis 1.1 that there is no difference in scores according to age can be rejected, and alternative hypothesis 1.1 can be accepted. The young
displaying the lowest JM language usage may be reflective of the increase in recent years of those receiving higher education compared with earlier generations, since education level strongly influences language use,
as was discussed in section 4.4.4. The same may be true for the more negative attitudes towards JM showed by the young, at least where societal advancement is concerned, since a significantly higher number of the
young said that even though they are happy with Jambi Malay they would not choose to attend or send their children to a Jambi Malay-medium school.
Attitudes indirectly accessed through the MGT, however, are not impacted significantly by age.
Based on the MGT results, null hypothesis 1.1 is not able to be rejected. In fact, some surprising trends are instead observed: the Old have fairly positive attitudes towards Indonesian and less positive attitudes
towards Jambi Malay, while it is the young who have the most negative attitudes towards Indonesian. The contrasting findings between the MGT and the questionnaire for the age variable may be a
result of built-in problems with the measurement tools themselves, as will be discussed later, or the fact that there were 20 fewer MGTs completed than questionnaires.
Whether or not these factors skewed the results, the case may be that overall, age is less important in predicting language attitudes than in predicting language usage. There may be more unity than disunity
among the Jambi Malays, without age distinction, when it comes to attitudes towards their language. A desire for increase in status andor economic advancement, which appears to be to some degree higher in
the young, may be the main factor that has the potential to disturb the unity.
5.4.2 Education variable
The impact of the education variable on language use and attitudes seems stronger, and is statistically significant in both the questionnaire and the MGT. Based on the results of the education variable from the
questionnaire and the MGT, null hypothesis 1.2 that there is no difference in scores among varying education levels can be rejected and alternative hypothesis 1.2 that there is a significant difference
between varying education levels in scores can be accepted. More specifically, it is apparent that the higher the education level, the less usage of Jambi Malay, and the less positive the attitudes towards it. This
is true in both the questionnaire and MGT for Jambi Malay, and is continued in the case of the small number of college graduates as well.
Although this trend regarding usage and attitudes towards Jambi Malay is fairly clear, there was an interesting issue encountered in the education level analysis for Indonesian. The MGT results for
Indonesian are the opposite of what was expected. Instead of following the trend of the higher the education level, the more positive the attitudes towards Indonesian, those who finished high school and
those who finished college have the lowest scores, while those with no education have the highest scores.
As with the age variable, this issue could have to do with a fault in the MGT. The informants tended to answer the same way for both Indonesian and Jambi Malay; therefore a problem inherent in the
MGT could be a cause. An alternate explanation could be that the educated indeed have more negative attitudes towards Indonesian, as indirectly discovered in the MGT. Possibly, it is linked to a matter of
necessity as a driving force. For example, the uneducated may need to have positive attitudes towards Indonesian; they need it for upward mobility. Those who are educated mostly likely have attained fluency
in it, and the need for positive attitudes towards it is less, given their already relatively high social status. Labov 1972 discussed a phenomenon similar to this and labeled it “hypercorrection” cf. Francis 1983.
He found that lower middle social classes were more inclined to approximate the prestige variety, and avoid linguistic behaviors that would mark the nonstandard language. In the case of the uneducated in this
study, hypercorrection extended to attitudes may indicate a desire for increased socioeconomic status.
5.4.3 Sex variable