126 Mohamad Fahmi
in this group is 21.9 per cent, whereas the proportion of WORKSMA in
public, private non-religious and private Islamic schools are 6.4, 12.2, and
12.9 per cent respectively.
Students from public and private Christian schools have also superior
academic achievements. The mean of final exam score in lower secondary
education is the highest for the public schools group. The average of this
score among the public school group is 37.96, whereas the final exam scores
in private non-religious, private Islamic and private Christian schools are
35.28, 36.75, and 35.88 respectively. The private Christian school group has
the highest final exam score in upper secondary education (SCORESMA)
or 33.31, whereas the upper secondary exam scores of public, private non-
religious and private Islamic schools are 31.26, 29.65, and 28.31 respectively.
The mean of PRIFAIL in public schools is also the lowest among these
groups. Hence, public school students have higher average pass rate in
primary education than students from private schools.
5. EMPIRICAL RESULT
5.1 Upper Secondary School Choice Estimation
Table 5.5 presents the multinomial logit model of choice among the four
upper secondary school categories with 2,101 observations. I follow some
previous studies about school choice effect (Bedi and Garg 2000; Le and
Miller 2003; Newhouse and Beegle 2005) to assess whether selection bias
is present in the probit model of the higher education participation using
the multinomial logit estimation. The model’s independent variables
include: personal characteristics and family background (MALE, ISLAM,
LANGINDO, URBAN, WORKSMP, and SIBLING), parents’ education
(FATHSHHE and MOTHSHHE), control variable for ability (PRIFAIL
and NEMSMP), and junior secondary school choice (PRIVATENRJH,
PRIVATEISSH, PRIVATECPOJH).
The positive coefficient on MALE in all estimation indicates that
female student has higher odds to attend public lower secondary school
as compared to male student. As expected, the positive and significant
coefficient on ISLAM shows that a Muslim student is more likely to attend
private Islamic school and private secular than public school. On the other
hand, a non-Muslim student prefers to study at private Christian school
than public school. Students who lived in urban area have higher odds to