Indonesian Higher Education 133
that 54.3 per cent and 40.6 per cent students from private Christian and
public schools continue to post-secondary education. On the other hand,
only 23.4 per cent and 15.9 per cent students from private Islam and secular
school participate in higher education. I use the non-linear decomposition
as a technique to find the strongest determinant of group differences that
contribute to higher education access.
The appraisals of non-linear decomposition estimation for school
choice gap on higher education participation between public, private
secular, private Islam and private Christian schools are in Table 5.9 and
Table 5.10. The independent variables consist of all variables from the
previous probit with selectivity correction, which are earnings foregone
(FOREGONE), school quality (TSRATIO), differences in sex (MALE),
religion (ISLAM), language (LANGINDO), household location (URBAN),
number of siblings (SIBLING), whether working in upper secondary school
or not (WORKSMA), academic performance (ZNEMSMA and SMAFAIL),
parent education background (FATHSHHE and MOTHSHHE), and a
set of dummies of province where individual was educated in upper
secondary school.
The results of the decomposition for private secular schools are reported
in column 2 in Tables 5.9 and 5.10. The gap between public and private
secular group on higher education participation is the largest or 0.249.
There are four determinants that contribute statistically significant to the
gap; personal characteristics (PERSONAL), earning foregone (INCOME),
academics performance (ABILITY) and parents’ education background
(PARENTS). All indicators are both significant at 1 per cent level. The
other two determinants FAMILY and SCHOOL are not significant. Earning
foregone (INCOME) is the highest contributor to the gap between public
and private secular school as it contributes 0.114 or 11.4 percentage points
to the gap. The second largest gap between public and private secular
is contributed by parents’ educational backgrounds. The difference of
parents’ educational background is 0.033 and significant at 1 per cent
level. Both variables on parents’ education determinant, FATHSHHE and
MOTHSHHE, are statistically significant at 1 per cent and the influence
of father’s education is stronger than mother’s education background.
This result support the evidence from the summary statistics in Table 5.3.
In public school, the percentage of father that attends upper secondary
or higher education background is 26.2 per cent whereas private secular
students are almost half of the public school group or 14.9 per cent. The