Sex, age, place of birth, ethnicity, parents’ education, father’s occupation and marital status have significant^ and independent direct effects on educational attainment of children (Table 3.3). The ‘R square change’ at the bottom of Table 3.3 shows that parents’ education contributes almost half (14 per cent out of the 30 per cent) of the proportion explained in the model. Thus in this study, education of parents appears to be the most powerful predictor of children’s educational attainment.
The second highest contribution was age, followed by socio-environmental factors, mainly place of birth and ethnicity. Marital status also has a significant effect on educational attainment; on the other hand, differences in the number of siblings, religion of mother, region and migration statuses appear to be insignificant influences on children’s educational attainment.
Model E of Table 3.3 tells us that everything else being equal, males could a chance to obtain a higher (32 per cent of one year of schooling) level of educational attainment than females. This is equivalent to roughly 3. 8 months longer in school than females. For both males and females, a one-year increase in their ages will increase the probability of remaining in school by 28 per cent of a year of schooling. For those who were married, the average years of schooling were around ten months (83 per cent of one year of schooling) shorter than for those who were single.
Number of siblings has no significant effect on the years of schooling. This is far from Blake’s (1989:298) findings in the United States where sibling size was the second most important factor after father’s education in affecting the total
S ig n ifica n t (*)refers to the significant o f Model F-test in a regression model. If the T value was greater than 1.96 or less than -1.96 the parameter is significantly different from zero at the 5 per cent level o f confidence (Gray, 1994:121).
years of education. In Indonesia, especially in the areas of the study, although parents with lower occupation tended to have more children (Table 3.2), nevertheless, the government’s heavy subsidy^ of educational provision, especially through Inpres schemes, may have mitigated the negative effect of sib-size on children’s educational attainment.
Urban-bom children also have a higher probability (17 months longer or 143 per cent of one year of schooling) in their schooling than the rural-bom. Different demand for and supply of education between urban areas and rural areas generally is seen as one of the factors in the rural-urban differential in educational attainment. In most villages of rural areas of Indonesia and in areas surrounding the three cities in the study in particular, the number and variety of schools were limited, most of the villages only having primary schools. Therefore, among those who were bom in rural areas, only those who were bom in relatively affluent families could continue their education.
Compared to other ethnic groups, Javanese youth have the highest probability of continuing their education (around 80 per cent of one year of schooling or around 9 months above the Babesuma ethnic group). Educational attainment of Babesuma^ was lower than that of Javanese and ‘Other’ ethnic groups. Low probability of continuation of education among the youth of the Babesuma ethnic group may be due to cultural factors such as a stronger preference for traditional education (pesantren5) and possibly a lower expectation
3 According to the World Bank (1982), in Indonesia, in the 1970s, for example, the proportion o f GNP devoted to education (ranging from 2 to 5 per cent) was similar to that in most developed countries. In 1992, in public university, the tuition fees paid by students only covered 15 per cent o f the public university cost (Boediono, McMahon and Adams, 1992:33).
4 Babesuma refers to Bantenese, Betawinese, Sundanese and Maduranese: indigenous ethnic groups in Java apart from the Javanese ethnic group.
5 A strong preference towards pesantren will result in low years o f schooling for some reasons: first,
enrollment in pesantren was not regarded as schooling by the respondents. Second, pesantren
graduates were unable to further their education in ‘modem’ education (sekolah umum). To solve this
problem, in the early 1980s the government introduced the 'madrasah ’ type o f school that combines
on the social as well as economic rate of return of modem education, since they have experienced a strong resistance toward modem (colonial and secular) education in the first half of the century. Moreover, modem education seems to be far from the needs of the traditional economy of most members of this ethnic group.
It is also interesting to note that the probability of continuing education was higher among children of the Javanese than among children of ‘other’ ethnic groups. One possible explanation was that most parents of the ‘other’ ethnic group, mostly Chinese and Arabs, were engaged in trades, which do not require high formal educational qualification, while high formal educational qualification seem to be a necessity for children of the Javanese ethnic group, most of whom were attracted to service sector activities.
Among all variables, parents’ education had the strongest effect on the educational attainment of children: one-year difference in schooling of parents affects the chance for the continuation of education among children by around a half-year of schooling, or 4. 5 months 0.37 per cent, in standardized coefficient. So assuming everything else was equal, children of parents (both father and mother) with university qualification could obtain roughly six and a half years longer in school than children of parents with primary education qualification.
Why does parents’ education strongly affects children’s educational attainment in Indonesia? First, parents of the respondents were mostly bom in the situation where the ‘contradiction’ between secular and religious education was rampant. So, parents with high educational qualifications were more likely to be bom in families that were more receptive toward ‘modernization’ or ‘Westernization’; of which educational institutions were seen as one of the agents. So, children of parents with high educational qualification were more likely to
have a social milieu that supported high achievement in education. Second, parents with high educational qualification were also a prime source of cultural and emotional capital for children’s value attitudes, intellectual and verbal development. Parents’ education seems to be the most powerful source for ‘cultural reproduction’, whereby children with parents holding high educational qualifications were equipped with a culture that was likely to be matched with school culture.
Unlike parents’ education, father’s occupation only has a small effect on the probability of the continuation of children’s education (R square change was below one per cent). Assuming everything else was equal, compared to children of farmers, children of fathers with professional occupations could spend 14 months (1.19 x 12 months) longer in school. The pattern was clear that children of fathers with higher occupations have a higher probability of continuing their education than children continue of parents with lower occupations.
Table 3.3
Regression on the probability of educational attainment among school leavers 15 to 29 years in three cities of Java, 1994. (The omitted category was coded 0.)___________________________________
Independent Ca- Regression Coefficient Standard-
variables ses Model A Model B Model C Model D Model E ized
Sex Female 1503 Male 1497 .48*** .50*** .53*** .51*** .32* .04* (.14) (-13) (.1 2) (.1 2) (.1 2) (-0 1) Age 3000 .30*** .29*** .24*** .24*** .28*** 24*** (.0 2) (.0 2) (.0 1) (.0 1) (-0 2) (-0 1) Number of siblings 3000 - . 0 2 . 0 1 .04 .03 .03 . 0 1 (.0 2) (.0 2) (.0 2) (■0 2) (-0 2) (.0 1) Place of birth Rural 325 Urban 2675 2.56*** 1.6 6*** 1.44*** 1.43*** .1 0*** (•2 2) (.2 0) (.2 2) (-23) (.0 2) Region Jakarta 1750 Semarang 500 -.6 8** -.09 - . 1 2 -.11 - . 0 1 (.2 1) (.2 0) (.2 0) (.2 0) (-0 2) Surabaya 750 . 0 1 .32 .32 .37 .04 (.18) (.17) (.17) (.17) (.0 2) Ethnicity Babesuma 846 Javanese 1462 1.47*** .83*** .8 6*** ..81*** .1 0*** (.2 1) (.18) (.17) (.17) (-0 2) Others 690 9 7*** .39* .38* .34 .03 (.18) (.19) (.19) (.19) (.0 1) Religion of mother Non-Moslem 348 Moslem 2652 -1.34*** -.52* -.54 -.44 -.03 (•25) (.23) (.23) (-23) (0 1) Parents’ educ. 3000 .53*** .5\*** .50*** .37*** (.0 2) (-0 2) (.0 2) (.0 2) Father’s occup. Professional 109 1.19** 1.19** .09** (.41) (.41) (.0 1) Clerical 383 1.07*** 1.03*** .08** (.29) (-2 0) (.0 2) Trades 719 .82** .80** .08** (.26) (.26) (.0 2) Services 398 .42 .42 .06 (-32) (.32) (.0 2) Production 1106 .41 .41 .05 (.25) (.25) (-25) Farmers 219 M arital status Single 2095 Married 905 -.83*** -.09*** (.15) (0 1) Migration status Non-migrant 2464 Migrant 536 .16 . 0 1 (.2 2) (-0 2) Constant 4.9805 3.1648 .6541 .7195 .1191 Model R2 .070 .140 .285 .291 .298 F 75.97 54.21 119.71 81.84 74.58 R2 change .070 .069 .145 .005 .007 F change 75.97*** 40.38*** 609.54*** 4.64*** 14.65***
Source: ‘The Dynamics of Youth Education and Employment’, 1994.
Herein after numbers in parentheses refer to the standard errors, * refers to significant at 0.05, ** =significant at 0.005 and ***= significant at 0.0005.