1
The e ec o he F ee High School Tui ion law on uppe -seconda y school choice in
Japan
Pe Ma iel
1
Depa men o Quan i a i e Me hods
Uni e si y o he Basque Coun y (UPV/EHU)
A da. Lehendaka i Agui e, 83
E48015 Bilbao, Spain
E-mail: pe [email p o ec ed]
Tel.: +34.94.601.3848
Nobuhi o Sanko
G adua e School o Business Adminis a ion
Kobe Uni e si y
2-1 Rokkodai-cho, Nada-ku, Kobe-shi 657-8501 Japan
E-mail: [email protected]
Tel.: +81.78.803.6987
Ainhoa Vega-Bayo
Depa amen o Economic Analysis
Uni e si y o he Basque Coun y (UPV/EHU)
A da. Lehendaka i Agui e, 83
E48015 Bilbao, Spain,
E-mail: ainhoa. eg[email p o ec ed]s
Tel.: +34.94.601.7076
Funding: The au ho s acknowledge inancial suppo om he FEDER/Minis y o Science,
Inno a ion and Uni e si ies (ECO2017-82111-R, PID2020-113650RB-I00) and he Basque
Go e nmen h ough g an s IT1359-19 (UPV/EHU Econome ics Resea ch G oup) and
IT1336-19 (Bilbao Resea ch Team in Economics).
1
Co esponding au ho
This is he accep manusc ip o he ollowing a icle ha appea ed in inal o m in S udies
in Educa ional E alua ion 70 : (2021) ID 101065 which has been published in inal o m
h ps://doi.o g/10.1016/j.s ueduc.2021.101065. Copy igh © 2021 The Au ho (s)
Published by Else ie unde CC BY-NC-ND licence ( h ps://c ea i ecommons.o g/
licenses/by-nc-nd/4.0/)
2
The e ec o he F ee High School Tui ion law on uppe -seconda y school choice in
Japan
Japan is one o he mos homogeneous coun ies in he wo ld. Japanese is he uni o m,
uni e sally used language in he coun y, and he educa ion sys em is based on high-quali y
public and p i a e schools. The p i a e educa ion sec o has some cu iosi ies. I is gene ally
used by highe classes looking o p es ige, as is usual in o he coun ies, e en hough quali y
and selec i i y indices ypically indica e ha high-quali y educa ion is also o e ed in public
schools (James e al., 1988). P es igious p i a e uppe -seconda y schools a e gene ally loca ed
in u ban a eas. In p e ec u es in u ban a eas, he highes - anked public and p i a e uppe -
seconda y schools a e compe i i e, whe eas, in p e ec u es in subu ban a eas, he e a e
gene ally many public uppe -seconda y schools ha a e anked highe han he highes - anked
p i a e uppe -seconda y schools.
The uppe -seconda y school choice in Japan seems o be a decisi e poin a ec ing
s uden s’ u u e job oppo uni ies and social s a us. The quali y o he chosen uppe -seconda y
school o a la ge ex en de e mines s uden s’ p ospec i e oppo uni ies, and ha is why his
decision can be seen as he s a o a long- e m compe i ion o a p o essional ca ee .
Acco ding o a ional choice heo y, indi iduals a e conscious decision make s, and hei
choices a e in luenced by a cos –bene i analysis. This analysis o he uppe -seconda y school
choice decision in Japan was based on s uden s’ socio-economic backg ound, educa ional
aspi a ions, and p e ious academic achie emen . These aspec s can cause class-speci ic
decisions ha may con ibu e o educa ional as well as social inequali ies.
The aim o his s udy is o conduc a de ailed analysis o he i s o hese h ee aspec s
o he uppe -seconda y school choice decision, ha is, o s uden s’ socio-economic
backg ound. The i s speci ic goal o his s udy is o analyse he ac o s ha lead amilies o
choose a speci ic uppe -seconda y school in Japan. The second speci ic goal is o de e mine
3
he impac o he new law called “F ee High School Tui ion”, which was in oduced in 2010.
We use da a on he school and amily cha ac e is ics included in he PISA (OECD, 2009,
2012) ques ionnai es and apply a disc e e choice model o analyse school choices be o e and
a e he implemen a ion o his law and he way in which hese choices a y depending on
s uden s’ socio-economic backg ound and an u ban o a non-u ban loca ion. These PISA da a
ca ego ise Japanese schools acco ding o wo dimensions: hei o ien a ion
(academic/ oca ional) and hei unding (public/p i a e). We analyse he ela ionship be ween
hose wo school dimensions and he socio-economic cha ac e is ics o he s uden s’ amily.
PISA da a ha e been widely used in ecen yea s o s udy simila and ela ed opics.
Fo example, school choice and school e iciency we e in es iga ed by Mon es and Rubalcaba
(2014), who assessed he ex en o which choice and choice equi y mechanisms in luenced
school e iciency. Ano he ecen analysis based on PISA da a was pe o med by Gi o d
(2019), who showed ha he sha e o s uden s whose schools admi ed s uden s based on hei
home add ess educed by app oxima ely 20% in Japan be ween 2000 and 2015. A ypical
s udy based on s uden s’ achie emen was ha o Sakella iou (2017), who analysed he
p i a e e sus public school pe o mance in ma hema ics o 40 coun ies and showed, in he
case o Japan, a public school ad an age.
The pape is o ganised as ollows. Sec ion 2 desc ibes he educa ion sys em in Japan,
Sec ion 3 p esen s he da a and me hods, and Sec ion 4 discusses he empi ical esul s.
Finally, Sec ion 5 p esen s he main conclusions.
The Educa ion Sys em in Japan
Japan has a single- ack comp ehensi e school sys em ha is simila o he US model
(En ich, 2015). The highly compe i i e and ela i ely igid educa ion sys em in Japan is
compulso y a he i s wo s ages, which a e elemen a y school (shōgakkō, o 6- o 12-yea -
olds) and lowe -seconda y school (chūgakkō, o 12- o 15-yea -olds). A e his s age,
4
s uden s can con inue in uppe -seconda y school (kō ōgakkō, o 15- o 18-yea -olds) o
a end a college o echnology (kō ō-senmon-gakkō). The uppe -seconda y school can be
ei he oca ional (senmongakka) o academic ( u sūka). A e wa ds, uppe -seconda y school
g adua es ad ance o uni e si ies o specialised aining colleges, unde ake oca ional
aining, o en e he labou ma ke di ec ly. The pe cen age o hose who en ol in uni e si ies
di e s signi ican ly be ween academic and oca ional uppe -seconda y school s uden s.
Colleges o echnology we e es ablished in 1962, and, unlike uni e si ies o junio colleges,
hey accep hose who ha e comple ed lowe -seconda y schooling. S uden s o hese
ins i u ions a e g an ed he i le o associa e (jun-gakushi) and may apply o admission o he
uppe di ision o uni e si y. Gene ally, hese ins i u ions ocus on eaching specialised
subjec s, wi h he aim o helping s uden s o de elop he abili ies ha a e equi ed o
oca ional li e (Minis y o Educa ion, Cul u e, Spo s, Science and Technology (MEXT),
2021a).
Seconda y schools in Japan do no usually include bo h s ages o seconda y educa ion
(lowe seconda y and uppe seconda y), so he as majo i y o s uden s ha e o change
schools be ween hese wo s ages. Acco ding o he Minis y o Educa ion, Cul u e, Spo s,
Science and Technology (MEXT, 2012), only 3% o s uden s in 2012 did no change schools
be ween hese s ages. Nea ly 1.2 million s uden s in ha yea s a ed hei educa ion a he
uppe -seconda y le el, bu ewe han 5,000 s uden s con inued o s udy a a seconda y school
ha had bo h lowe - and uppe -seconda y di isions, and ewe han 30,000 s uden s en olled
in an uppe -seconda y school ha was ope a ed by he same educa ional body as hei lowe -
seconda y school o ha had a close ela ionship wi h hei lowe -seconda y school.
Admission o he uppe -seconda y school sys em plays an impo an ole. A e ha ing
inished hei compulso y educa ion, s uden s a e admi ed o high- o low-p es ige uppe -
seconda y schools depending on hei sco es in compulso y en ance examina ions. Schools a
5
he uppe -seconda y le el a e alued by hei epu a ion, ha is, acco ding o he pe cen age
o hei s uden s who pass he di icul en ance examina ions o he mos p es igious
uni e si ies in he coun y (S e enson & Bake , 1992). Repu able uppe -seconda y schools
ecei e many applica ions; he e o e, s uden s applying o hese schools need highe exam
sco es. Thus, schools a e anked by he minimum exam sco es o be admi ed, and s uden s
apply o schools ha i hei academic achie emen s. The sys em he e o e ep esen s a ype
o hie a chical academic anking sys em (Ka iya & Rosenbaum, 1999). Howe e , in ecen
yea s he e ha e been ce ain changes o he admission p ocess o uppe seconda y schools.
Fo example, En ich (2019, p 275) ecen ly w o e: "F om 1997 onwa ds, s uden s’ indi idual
mo i a ion and ap i udes we e gi en mo e weigh o high school admission h ough he
e alua ion o p ac ical o echnical examina ions, essays, and in e iews, and by s onge
conside ing ex a-cu icula ac i i ies and ecommenda ion le e s. Addi ionally, he
ca chmen a eas o high school admission we e expanded, whe e o e s uden s can choose
om a la ge ange o high schools now and ake mo e han one en ance examina ion in he
same yea ."
We in es iga e how he uppe -seconda y school choice is ela ed o he socio-
economic cha ac e is ics o he amily and how his ela ionship changed in 2010 wi h he
implemen a ion o he new “F ee High School Tui ion” law. This di ec app oach would no
be alid o he elemen a y school choice because elemen a y schools and lowe -seconda y
schools a e compulso y and mos s uden s a e alloca ed o public schools wi hin hei school
dis ic . The e o e, some pa en s a e willing o pay mo e o housing in speci ic school
dis ic s. Ku oda (2018), o example, showed ha pa en s exhibi ed highe willingness o pay
housing en in a be e public elemen a y school dis ic . Howe e , his has been shown only
o ma ied couples who ha e child en a ending o expec ing o a end elemen a y schools.
Simila ly, Ushijima and Yoshida (2009) and Yoshida e al. (2008) analysed he e ec o
6
school quali y on land p ices. They concluded ha he quali y o elemen a y schools
in luences land p ices in he school dis ic bu ha his e ec a ies depending on he ime
pe iod analysed and he ype o dis ic . A possible ela ionship be ween he ype o uppe -
seconda y school and s uden s’ socio-economic cha ac e is ics has been s udied indi ec ly in
he ollowing wo s eams o esea ch.
The i s s eam is based on s uden s’ es sco es. In his li e a u e, he a iances o
sco es a e decomposed in o be ween-school and in-school a iances. I s uden s we e
alloca ed o schools based on hei achie emen le el, hen a high be ween-school a iance in
sco es would be expec ed. This was shown, o example, by Knipp a h (2010), in an analysis
o he ma hema ics and science sco es o PISA 2000, 2003, and 2006, and by Taki (2011), in
a s udy ocusing on he ma hema ics sco es o PISA 2003. Al hough hese s udies showed ha
he s uden s in he same uppe -seconda y schools ha e simila sco es, hey did no di ec ly
model school choice. Ne e heless, Knipp a h (2010) ound co ela ions be ween s uden s’
pe o mance in ma hema ics and science and hei socio-economic s a us a he high school
le el and concluded ha “The PISA s udies also showed ha Japanese s uden s a e alloca ed
o high schools acco ding o hei achie emen le el and hei economic, social, and cul u al
backg ound” (Knipp a h, 2010, p. 403).
The second s eam o esea ch ocuses on he quali y o schools. De ining he quali y
o a school is no a s aigh o wa d ask, which is why he de ini ions di e among s udies.
Yamamo o and B in on (2010) analysed he 1995 Social S a i ica ion and Mobili y Su ey
and de ined he quali y o uppe -seconda y schools using esponden s’ epo s on he
p opo ion o classma es who p oceeded o junio college o uni e si y. They concluded ha
he op- anked seconda y schools we e chosen by amilies wi h a highe socio-economic
s a us. Nakanishi (2011) applied a di e en app oach and de ined he quali y o an uppe -
seconda y school using s uden s’ sel - epo ed achie emen a lowe -seconda y schools,
7
concluding ha he achie emen , and he e o e he quali y o he school, is highly ela ed o
he a he ’s socio-economic cha ac e is ics. Recen ly, En ich (2019), using he Hyōgo High
School S uden s (HHSS) su ey (Ojima & A amaki, 2018), no only ela ed he school
anking o classical school choice d i e s, like pa en s’ socio-economic s a us, s uden s’
academic pe o mance, and ins i u ional cons ain s, bu also showed ha a signi ican sha e
in he school choice decision is made by indi idual s uden s hemsel es. This was achie ed
by showing ha s uden s’ conc e e u u e plans signi ican ly a ec ed hei decision making a
he ansi ion o high school.
The e is also a as li e a u e ocusing on he impo ance o acking, ha is, he ea ly
de e mina ion o whe he s uden s will ollow an academic ack o a oca ional ack, in he
Japanese sys em. Taki (2010, p. 247) concluded ha “Japan is a coun y ha ing he dis inc
cha ac e is ic whe ein almos all he ele ance be ween SES [socio-economic s a us] and he
academic pe o mance is being con e ed in o in e -school dispa i y by he high school
en ance exam”. Simila ly, Hallinan (1994) and Oakes (1994) obse ed an impac o acking
on achie emen , a i udes, and educa ional a ainmen . They also concluded ha acking is
ela ed o s uden s’ economic, social, and cul u al backg ound; he school en i onmen ; and
he class oom clima e, among o he a iables. This is an impo an ac because, subsequen ly,
s uden s om low- anked schools a e unlikely o en e compe i i e uni e si ies (Ono, 2001).
Mo eo e , s uden s wi h a highe socio-economic s a us usually ha e easie access o
shadow educa ion lessons ha help hem o imp o e hei academic pe o mance, which in
u n leads o admission o compe i i e uni e si ies. Ma suoka (2015), o example, indica ed
ha , in 2007, he pe cen age o s uden s a ending c am schools (juku) o lea n academic
subjec s inc eased hea ily in he pe iod when s uden s ook high school en ance
examina ions, eaching 50.9% o eigh h-g ade s and 65.4% o nin h-g ade s. Fu he mo e,
mo e ecen in o ma ion has shown ha he impo ance o shadow educa ion is no
8
dec easing. Acco ding o e-S a (2018), 79.8% o nin h-g ade s in public schools a end c am
schools ha each academic subjec s, and hei es ima ed a e age annual expendi u e is JPY
393,000. One o he mos impo an and comp ehensi e s udies on his opic is En ich
(2018), who conduc ed an analysis o he impac o shadow educa ion on social inequali y
o ma ion in Japan based on se e al de ailed empi ical analyses. This wo k examined he
easons o he high Japanese en olmen a es in c am schools and p i a e u o ing, oge he
wi h hei causes and hei implica ions o social inequali y.
The e is also an ex ensi e li e a u e ocusing speci ically on he choice o high school.
Fujiha a (2012), o example, de eloped wo hypo heses ela ed o his opic. The i s is he
ela i e isk a e sion hypo hesis (B een & Gold ho pe, 1997), and he second is he
downwa d educa ional mobili y a e sion hypo hesis (Kikkawa, 2006). These hypo heses we e
es ed wi h da a on second-yea s uden s a Japanese senio high schools. As expec ed, he
esul s showed ha a he s’ occupa ion and pa en al educa ion had di ec e ec s on he ank
o high school a ended.
An impo an conclusion wi h espec o he aim o ou wo k was d awn by Ka iya
(2016), who in es iga ed he ope a ion o mechanisms ha ha e p oduced social inequali y in
educa ion o e ecen decades. Speci ically, his wo k analysed h ee da a se s collec ed in
h ee di e en decades ocusing on he ela ionship be ween he hie a chy o Japanese senio
high schools and inequali y. The esul s p o ided empi ical e idence o an inc easing impac
o s uden s’ amily backg ound on hei academic g ades and he ank posi ions o he high
schools in which he s uden s a e en olled. Mo eo e , he esul s showed a dec easing and
indi ec e ec o he amily backg ound on he s uden s’ alloca ion o high schools o e ecen
decades. This indi ec e ec has occu ed p ima ily h ough i s in luences on s uden s’
academic achie emen . Speci ically, he “selec ion o s uden s in o di e en u u e SES s a a
9
akes place isibly h ough educa ional di e en ia ion a he uppe seconda y educa ional
le el” (Ka iya, 2016, p. 151).
Finally, Sakai (2010) analysed he ca ee consciousness o s uden s a an u ban
comme cial high school and hei possible mo i a ions o a end high- o middle-le el
uni e si ies. His esul s helped o p o ide an unde s anding o how he economic ecession o
he 1990s in luenced schools’ policy on ca ee guidance. Speci ically, he concluded ha he
s uden s’ consciousness lacked a conc e e u u e pe spec i e and ha helping hem o
o e come his issue was consis en wi h he managemen s a egy o he school.
Uppe -seconda y school is no compulso y, bu he as majo i y o all lowe -
seconda y school g adua es con inue hei s udies a ei he public o p i a e uppe -seconda y
schools. Acco ding o MEXT (2021b), he pe cen age o p i a e uppe seconda y s uden s in
2010 was 29.8%. Nei he public no p i a e schools a e ee, bu he cos o public uppe -
seconda y schools is lowe han ha o p i a e uppe -seconda y schools. In Ap il 2010, he
Japanese Go e nmen made public uppe -seconda y schools ui ion ee. A he same ime,
s uden s a p i a e uppe -seconda y schools s a ed o ecei e an amoun equi alen o he
ui ion ee a public uppe -seconda y schools as a subsidy. Howe e , he household
expendi u e on educa ion pe s uden (including p i a e school ui ion a e he subsidy and
o he expenses o school educa ion and ex acu icula school ac i i ies) o p i a e school
s uden s is wo o h ee imes he expendi u e o public school s uden s. Since Ap il 2014, a
household income limi has been applied o de e mine eligibili y o ecei e he subsidy.
Modi ica ions a e con inuously being made o his sys em.
The goal o he ui ion- ee high school p og amme was s a ed as ollows:
“Minimizing he inancial bu den on households o ensu e ha all mo i a ed high school
s uden s can eel secu e abou ecei ing educa ion, i espec i e o he inancial si ua ions o
hei amilies, is an issue ha needed o be ackled u gen ly” (MEXT, 2009). The jus i ica ion
16
Table 2
Summa y S a is ics o he Explana o y Va iables
Va iable
2009
2012
Ci y and la ge
ci y
Town and small
own
Ci y and la ge
ci y
Town and small
own
Numbe o s uden s
3,394
1,495
3,608
1,368
Dummy a iables
P opo ion
P opo ion
Fa he no wo king
0.07
0.08
0.07
0.07
Mo he no wo king
0.66
0.59
0.65
0.60
Two-pa en amily
0.91
0.91
0.92
0.91
Siblings
0.88
0.91
0.87
0.89
G andpa en s li ing wi h he amily
0.27
0.40
0.25
0.34
Values
Disc e e a iables
1
2
3
4
5
6
2009
Educ. le el mo he
Ci y and la ge ci y
0%
2%
12%
33%
25%
28%
Town and small own
0%
4%
16%
38%
22%
19%
Educ. le el a he
Ci y and la ge ci y
0%
4%
13%
26%
7%
50%
Town and small own
0%
7%
18%
36%
7%
32%
2012
Educ. le el mo he
Ci y and la ge ci y
0%
2%
10%
33%
26%
29%
Town and small own
0%
4%
16%
40%
22%
18%
Educ. le el a he
Ci y and la ge ci y
0%
5%
12%
27%
8%
48%
Town and small own
0%
7%
19%
35%
8%
31%
Ci y and la ge ci y
Town and small own
Con inuous a iables
Mean
S.D.
Min.
Max.
Mean
S.D.
Min.
Max.
2009
Cul u al possessions
-0.25
0.93
-1.39
1.24
-0.30
0.90
-1.39
1.23
Family weal h
-0.43
0.70
-2.88
2.81
-0.43
0.71
-2.56
2.06
Home educ. esou ces
-0.32
1.02
-4.53
1.60
-0.51
1.07
-4.53
1.60
Highes pa en al occupa ional
s a us
52.78
14.27
23.00
80.00
50.2
15.20
23.00
80.00
2012
Cul u al possessions
-0.38
0.96
-1.51
1.27
-0.57
0.94
-1.51
1.27
Family weal h
-0.24
0.63
-3.30
2.92
-0.16
0.62
-2.50
1.80
Home educ. esou ces
-0.44
0.81
-3.93
1.12
-0.60
0.81
-3.93
1.12
Highes pa en al occupa ional
s a us
52.30
20.07
11.56
88.70
46.32
19.95
11.56
88.70
17
Ano he index a iable coded by PISA 2012 is Home educa ional esou ces, which is
cons uc ed using answe s o ques ions conce ning, o ins ance, whe he s uden s ha e a desk
and a quie place o s udy, a compu e o schoolwo k, educa ional so wa e, books o help
wi h hei schoolwo k, echnical e e ence books, and dic iona ies. As can be seen in Table 2,
o ou s udy, his index anged be ween -4.53, 1.60 and -3.93, 1.12 o 2009 and 2012,
espec i ely. Finally, he highes pa en al occupa ional s a us shows he highe In e na ional
Socio-Economic Index o occupa ional s a us sco e o ei he pa en o , in he case o single-
pa en households, o he only a ailable pa en . As s a ed abo e, Bukodi and Gold ho pe
(2013) decomposed he social o igins in o pa en al class, pa en al s a us, and pa en al
educa ion. In ou case, he pa en al class can be ep esen ed by Family weal h, he pa en al
s a us by Highes pa en al occupa ional s a us, and he pa en al educa ion by Educ. le el
mo he and Educ. le el a he .
The es ima ion o a mul inomial logi model o a school choice using PISA da a could
seem o be a non-s anda d app oach, as PISA s udies usually ake in o accoun he complex
sample s uc u e h ough he use o mul ile el models and ela ed echniques. Ne e heless,
hese s udies a e usually ocused on he collec ed s uden s’ achie emen s, which may depend
on he schools’ and s uden s’ cha ac e is ics, and he complex sample s uc u e mus be aken
in o accoun . In ou case, he explained a iable (school choice) was de e mined p io o he
da a collec ion; he e o e, he me hodology does no need his mul ile el app oach.
Empi ical Resul s
We es ima e mul inomial logi models by maximum likelihood using he 2009 and
2012 samples o he wo ci y/ own a eas o analyse he easons ha d i e Japanese pa en s’
and s uden s’ decision o choose a ce ain ype o uppe -seconda y school. The dependen
a iable in ou model is a ca ego ical a iable, Type o school, wi h ou di e en alues, as
de ined in Table 1. The explana o y a iables a e hose included in Table 2. We checked o
18
possible mul icollinea i y among hese a iables bu ound e y low in e dependency. Tables
A1 and A2 in he Appendix p esen he es ima ion ou come o he model wi h 36 pa ame e s
ob ained o he wo a eas and wo yea s. To summa ise his ou come, we p esen he
likelihood a io es s o all he explana o y a iables in Table 3. Wi h ou dependen
ca ego ies in ou dependen a iable, he e a e h ee se s o pa ame e s associa ed wi h each
explana o y a iable. The join hypo hesis ha an explana o y a iable does no a ec he
dependen a iable (school ype) he e o e in ol es he se o h ee coe icien s co esponding
o each speci ic a iable being equal o ze o.
Table 3
Likelihood Ra io Tes s o he Explana o y Va iables
2009
2012
Ci y and la ge
ci y
Town and small
own
Ci y and la ge ci y
Town and small
own
𝜒V
s a is ic
p-
alue
𝜒V
s a is ic
p-
alue
𝜒V
s a is ic
p-
alue
𝜒V
s a is ic
p-
alue
Fa he no wo king
10.43
0.02
2.34
0.51
0.19
0.98
10.04
0.02
Mo he no wo king
12.91
0.01
0.56
0.91
10.20
0.02
3.25
0.35
Two-pa en amily
2.42
0.49
8.06
0.05
1.47
0.69
6.74
0.08
Siblings
15.30
0.00
1.74
0.63
13.01
0.01
6.63
0.09
G andpa en s li ing wi h
he amily
7.12
0.07
5.39
0.15
3.04
0.39
4.85
0.18
Cul u al possessions
24.00
<0.01
6.42
0.09
24.52
0.00
18.98
0.00
Educ. le el mo he
58.51
<0.01
20.67
<0.01
46.51
0.00
12.09
0.01
Educ. le el a he
119.12
<0.01
34.51
<0.01
78.29
0.00
19.64
0.00
Family weal h
11.72
0.01
11.57
0.01
3.26
0.35
1.57
0.67
Home educ. esou ces
0.46
0.93
7.40
0.06
3.26
0.35
5.67
0.13
Highes pa en al
occupa ional s a us
11.45
0.01
4.30
0.23
24.82
0.00
3.51
0.32
The speci ic e ec s o each a iable on he school choice a e discussed below, bu
p obably he mos impo an esul in Table 3, ela ed di ec ly o ou goal o analyse he impac
o he new law, can be obse ed o Family weal h. Al hough i s e ec is signi ican a he 5%
le el in bo h ci y/ own a eas in 2009, p io o he passing o he law, i is no longe signi ican
in 2012. This is in line wi h he conclusion eached by Ho i and Shimizu ani (2018) ha he
ui ion- ee p og amme enhanced he high school en olmen a e o lowe -income households;
19
ha is, i ga e hem incen i es o send hei child en o high school hanks o he exemp ion
om he ui ion paymen .
Ano he impo an pa o he in e p e a ion o ou esul s is based on changes in he
p obabili y de ined in equa ion (1) o choosing a speci ic ype o school depending on
di e en alues o ou explana o y a iables (
𝑥$
). Figu e 1 shows he change in he
p obabili y o choosing a ce ain ype o school in he wo analysed yea s and a eas i one o
he explana o y a iables changes i s alue. In his compa ison, he emaining explana o y
a iables a e se o a ep esen a i e alue, aking in o accoun he desc ip i e s a is ics
p esen ed in Table 2, by se ing hem o a “benchma k amily”. Rep esen a i e alues o he
dummy a iables a e de e mined based on he majo i y ule, ha is, he a he is wo king, he
mo he is no wo king, i is a wo-pa en amily wi h one o mo e siblings, and he e a e no
g andpa en s li ing wi h he amily. Mo eo e , he disc e e a iable o a he s’ and mo he s’
educa ion le el is se o a median alue (5), while he alues o he con inuous explana o y
a iables (Family weal h, Cul u al possessions, Home educ. esou ces, and Highes pa en al
occupa ional s a us) a e se o hei co esponding mean alue. All he changes p esen ed in
Figu e 1 a e calcula ed as disc e e changes (i.e. 0/1) o he dummy a iables, as he uni
change o he median alue o a he s’ and mo he s’ educa ional le el (i.e. om 5 o 6), and
as a s anda d de ia ion change o he con inuous a iables (i.e. om (mean - s . de ./2) o
(mean + s . de ./2)). Mo e de ails can be ound in he s udy by Long and F eese (2005).
Figu e 1 allows o se e al compa isons. Gi en ou goal o analyse he impac o he
new law, he mos ele an esul s o us a e he di e ences be ween 2009 and 2012. We can
obse e ha , o example, a change om ze o o one in he Fa he no wo king a iable in he
ci y a ea in 2009, ha is, a change om a amily wi h a wo king a he o a amily wi h a non-
wo king a he , dec eases he p obabili y o amilies choosing he p i a e and academic
(ou come 3) ype o school by abou 0.08, bu in 2012 he e ec o his a iable is close o
20
ze o. On he o he hand, his a iable inc eases he p obabili y o choosing a public and
academic (ou come 1) ype o school in he ci y a ea in 2009 by abou 0.06, bu his change in
2012 is nega i e and close o ze o.
Figu e 1
Change in P obabili y wi h Respec o he Benchma k Family
2009
2012
Ci y
and
la ge
ci y
Town
and
small
own
1 – Public and Academic 3 – P i a e and Academic
2 – Public and Voca ional 4 – P i a e and Voca ional
To in e p e he mos ele an esul s in Figu e 1, we will ocus i s on he ci y a ea.
Compa ing all o he changes and aking in o accoun ha he e a e ze o/one changes, uni
changes, and s anda d de ia ion changes in he mix, we can see ha he bigges changes in
p obabili y come om he a iable Siblings, o bo h 2009 and 2012. Tha is, ha ing mo e
han one child inc eases he p obabili y o choosing he public and academic ype o school
(ou come 1) by abou 0.08 bu dec eases he p obabili y o choosing he p i a e and academic
(ou come 3) ype o school by abou 0.10 in bo h yea s.
Change in P edic ed P obabili y o SCTYPEREC
-.11 -.07 -.04 0 .04 .07 .11
1 2 3 4
1 2 3 4
1 2 3 4
1 2 3 4
1 2 3 4
1 2 3 4
1 2 3 4
1 2 3 4
1 2 3 4
1 2 3 4
1 2 3 4
Fa he no wo king-0/1
Mo he no wo king-0/1
Two-pa en amily-0/1
Siblings-0/1
G andpa en s-0/1
Educ. le el mo he
Educ. le el a he
Family weal h-s d
Cul u al possessions-s d
Home educ. esou ces-s d
Pa en al occup. s a us-s d
Change in P edic ed P obabili y o SCTYPEREC
-.11 -.07 -.04 0 .04 .07 .11
1 2 3 4
1 2 3 4
1 2 3 4
1 2 3 4
1 2 3 4
1 2 3 4
1 2 3 4
1 2 3 4
1 2 3 4
1 2 3 4
1 2 3 4
Fa he no wo king-0/1
Mo he no wo king-0/1
Two-pa en amily-0/1
Siblings-0/1
G andpa en s-0/1
Educ. le el mo he
Educ. le el a he
Family weal h-s d
Cul u al Possessions-s d
Home educ. esou ces-s d
Pa en al occup. s a us-s d
Change in P edic ed P obabili y o SCTYPEREC
-.11 -.07 -.03 .01 .05 .09 .13
1 2 3 4
1 2 3 4
1 2 3 4
1 2 3 4
1 2 3 4
1 2 3 4
1 2 3 4
1 2 3 4
1 2 3 4
1 2 3 4
1 2 3 4
Fa he no wo king-0/1
Mo he no wo king-0/1
Two-pa en amily-0/1
Siblings-0/1
G andpa en s-0/1
Educ. le el mo he
Educ. le el a he
Family weal h-s d
Cul u al possessions-s d
Home educ. esou ces-s d
Pa en al occup. s a us-s d
Change in P edic ed P obabili y o SCTYPEREC
-.11 -.07 -.04 0 .04 .07 .11
1 2 3 4
1 2 3 4
1 2 3 4
1 2 3 4
1 2 3 4
1 2 3 4
1 2 3 4
1 2 3 4
1 2 3 4
1 2 3 4
1 2 3 4
Fa he no wo king-0/1
Mo he no wo king-0/1
Two-pa en amily-0/1
Siblings-0/1
G andpa en s-0/1
Educ. le el mo he
Educ. le el a he
Family weal h-s d
Cul u al Possessions-s d
Home educ. esou ces-s d
Pa en al occup. s a us-s d
21
Mo eo e , he pa o Figu e 1 de o ed o he ci y a ea indica es ha , app oxima ely,
he g oup o a iables wi h he second mos impo an e ec in 2009 (behind he la ges
Siblings e ec ) can be conside ed o be Fa he no wo king and, as expec ed om he
li e a u e e iew, Educ. le el mo he and Educ. le el a he .
I is also impo an o no e ha he e ec s o he a iables Fa he no wo king and
Mo he no wo king a e di e en . Acco ding o Table 2, in he ci y a ea, he mo he is no
wo king in 66% and 65% o he amilies in 2009 and 2012, espec i ely, and he e ec o his
a iable in he ci y a ea in Figu e 1 emains ela i ely s able be ween 2009 and 2012. On he
o he hand, he a he is wo king in 93% o he amilies in he wo analysed yea s (Table 2),
and, as men ioned ea lie , i s e ec in Figu e 1 changes d as ically be ween 2009 and 2012.
This shows close dependence o he school choice on he amily budge , which, acco ding o
his esul , seems o be linked o a he s’ bu no mo he s’ employmen s a us in he ci y a ea.
The e o e, Fa he no wo king seems o be an indica o o a low-income household. On he
o he hand, Mo he no wo king seems o ep esen a b oade amily cha ac e is ic han he
wo k s a us pe se – o example, a su icien ly high amily income ha allows his mo he ’s
employmen s a us. A non-wo king mo he , he e o e, is likely o indica e a speci ic amily
class. I also highligh s he ac ha he change in Mo he no wo king is simila o a s anda d
de ia ion inc ease in Cul u al possessions, bo h implying a highe p obabili y o choosing an
academic bu a lowe p obabili y o choosing a oca ional ype o school.
As expec ed om he li e a u e, he educa ion le el o he a he and mo he also ha e
an impo an e ec on he ype o school chosen. A uni change in hese a iables ( om 5 o
6) inc eases he p obabili y o choosing he p i a e and academic (ou come 3) ype o school
in he ci y a ea and in he wo yea s by app oxima ely 0.04–0.07, bu i dec eases he
p obabili y o choosing he public and oca ional (ou come 2) ype o school by abou 0.03–
0.04.
22
A ypically obse ed e ec in he li e a u e is ha highe weal h leads o a highe
p obabili y o choosing a p i a e and academic school. This esul , ep esen ed by he e ec
o he a iable Family weal h, is obse ed only in 2009. A s anda d de ia ion inc ease aises
he p obabili y o choosing he p i a e and academic (ou come 3) ype o school by
app oxima ely 0.03 in he ci y a ea, bu i dec eases he p obabili y o choosing he public and
academic (ou come 1) ype o school by app oxima ely he same amoun .
The compa ison o he changes in p obabili ies be ween 2009 and 2012 is expec ed o
shed ligh on he e ec o he “F ee High School Tui ion” law implemen ed in 2010. The
di e ences in he e ec s o he explana o y a iables be ween 2009 and 2012 s ay he same
o some a iables and a y o o he s. Con inuing o ocus on his in he ci y a ea, he la ges
Siblings e ec s ays e y simila in 2009 and in 2012. This is a e y impo an conclusion
because he co esponding pa o he amily budge de o ed o ui ion cos s ob iously
inc eases wi h he numbe o child en and he ac ha his e ec emains cons an in he wo
yea s seems o sugges ha he new law has a mo e mi iga ed e ec on amilies wi h mo e
han one child. Howe e , his s able e ec o Siblings is no obse ed o o he a iables
ela ed o amily budge .
The nega i e e ec on he p obabili y o choosing he p i a e and academic ype o
school (ou come 3) in he ci y a ea o Fa he no wo king (a low-income household indica o )
obse ed in 2009 diminishes in 2012. This is con i med by he ac ha i s e ec is signi ican
a he 5% le el in 2009 bu no signi ican a he 5% le el in 2012 (Table 3). In e es ingly
enough, he same esul is ob ained in Table 3 o Family weal h, a ac ha is also e lec ed
g aphically in Figu e 1 as i s e ec diminishes in 2012 in he ci y a ea. The e o e, acco ding
o Table 3 and Figu e 1, he e ec o he wo a iables di ec ly o indi ec ly ela ed o he
amily budge (Fa he no wo king and Family weal h) dec eased signi ican ly in 2012. This
seems o be a di ec e ec o he new law’s implemen a ion in 2010.
23
The esul s a e sligh ly di e en o he own a ea. The di e ence in he school o e as
well as in he social class composi ion wi h espec o he ci y a ea becomes appa en in he
desc ip i e s a is ics in Tables 1 and 2. Rega ding he school o e , he la ges di e ences a e
ep esen ed by a g ea e p opo ion o public and oca ional and a lowe p opo ion o p i a e
and academic ypes o schools in he own a ea. When i comes o he socio-economic
cha ac e is ics, he la ges di e ences a e ep esen ed by a g ea e sha e o g andpa en s li ing
wi h he amily and a lowe educa ion le el o bo h pa en s in he own a ea.
Focusing on Figu e 1, he bigges changes in p obabili y in he own a ea come om
he a iables Fa he no wo king and Two-pa en amily. The e ec o Fa he no wo king in
he own a ea, wi h a lowe o e o p i a e schools, seems o be di e en om ha in he ci y
a ea, wi h a much la ge sha e o p i a e schools (Table 1). In his case, i s e ec inc eases he
p obabili y o choosing he oca ional ype and dec eases he p obabili y o choosing he
academic ype o school in 2012. This e ec is no signi ican a he 5% le el in 2009 (Table
3). This kind o amily does no seem o be a ec ed by he new law as he decision seems o
be mo e be ween academic and oca ional schools han be ween public and p i a e schools.
This is p obably because o a lowe o e o p i a e schools locally; conside ing p i a e
schools in ci ies nea by would be linked o highe ime and a el cos s. Mo eo e , he
si ua ion o a non-wo king a he in his mo e u al a ea could be associa ed wi h a desi e o
as e inco po a ion o he o sp ing in o he labou ma ke .
In spi e o he la ge e ec in Figu e 1, he e ec o he Two-pa en amily is only
ma ginally signi ican in 2009 and no signi ican a he 5% le el in 2012 (Table 3). Howe e ,
simila o he ci y a ea, he e ec s o Educ. le el mo he and Educ. le el a he a e bo h
signi ican a he 5% le el (Table 3). Again, i appea s ha , gi en he lowe o e o p i a e
schools, he p e e ence o pa en s wi h highe educa ion is mo e o he academic han o he
24
oca ional ype o schools, unlike he p e e ence o p i a e o e public schools obse ed in
he ci y a ea.
On he o he hand, an impo an esul is ep esen ed by he same e ec o Family
weal h as ob ained in he ci y a ea. I s e ec on he p obabili y o choosing a p i a e and
academic school in 2009 is posi i e, whe eas i is nega i e o he public academic ype o
school. This e ec p ac ically disappea s (Figu e 1) and becomes no signi ican a he 5%
le el in 2012 (Table 3). These a e he same esul s ha we e ob ained o he ci y a ea.
The e o e, he di ec e ec o he new law can be obse ed bo h in he ci y and in he own
a ea.
Figu e 2 shows he p obabili y o choosing a speci ic ype o school o speci ic alues
o he a iables Educ. le el a he (as an example o a a iable wi h simila e ec s in he wo
yea s conside ed) and Family weal h (as an example o a a iable wi h di e en e ec s in he
wo yea s). The emaining a iables in Figu e 2 a e se a he benchma k amily’s alues, as
p e iously speci ied in Figu e 1.
Focusing on he g aphs showing he e ec o he a he ’s educa ion, he bigges
changes in he p obabili y o choosing a speci ic ype o school belong, in he ci y a ea, o he
p i a e–academic and public– oca ional ypes o schools. Fo he lowes le els o he a he ’s
educa ion, he p obabili y o choosing he public and oca ional ype o school is highe han
he p obabili y o choosing he p i a e and academic ype o school. This si ua ion is e e sed
o he highe le els o educa ion. These e ec s a e almos he same in 2009 and 2012. The
same esul s o he own a ea a e ep esen ed by he downwa d end o he p obabili y o
choosing he public and oca ional ype o school. The expec ed upwa d end o he p i a e
and academic ypes o school is no obse ed due o hei limi ed o e in he own a ea. The
mos impo an esul is ha he ends o all ou p obabili ies a e e y simila in he wo
a eas and in he wo analysed yea s.
25
Figu e 2
E ec o he Fa he ’s Educa ion and he Family Weal h on he P obabili y o School Choice
Fa he ’s educa ion
2009
2012
Ci y
and
la ge
ci y
Town
and
small
own
Family weal h
2009
2012
Ci y
and
la ge
ci y
Town
and
small
own
The e ec o amily weal h, howe e , changes in 2012 compa ed wi h 2009 in he wo
a eas (lowe pa o Figu e 2). In 2009, ha ing g ea e amily weal h inc eases he p obabili y
o choosing p i a e schools bu dec eases he p obabili y o choosing public schools bo h in
0.1 .2 .3 .4 .5
1 2 3 4 5 6
Educ. le el a he
Pub Academic Pub P ac ical
P i Academic P i P ac ical
P obabili y
0.1 .2 .3 .4 .5
1 2 3 4 5 6
Educ. le el a he
Pub Academic Pub P ac ical
P i Academic P i P ac ical
P obabili y
0.2 .4 .6
1 2 3 4 5 6
Educ. le el a he
Pub Academic Pub P ac ical
P i Academic P i P ac ical
P obabili y
0.2 .4 .6 .8
1 2 3 4 5 6
Educ. le el a he
Pub Academic Pub P ac ical
P i Academic P i P ac ical
P obabili y
0.2 .4 .6
-2 0 2
Family weal h
Pub Academic Pub P ac ical
P i Academic P i P ac ical
P obabili y
0.1 .2 .3 .4 .5
-2 0 2
Family weal h
Pub Academic Pub P ac ical
P i Academic P i P ac ical
P obabili y
0.2 .4 .6 .8
-2 0 2
Family weal h
Pub Academic Pub P ac ical
P i Academic P i P ac ical
P obabili y
0.2 .4 .6 .8
-2 0 2
Family weal h
Pub Academic Pub P ac ical
P i Academic P i P ac ical
P obabili y
32
Ishida, K., Nakamu o, M., & Takenaka, A. (2016). The academic achie emen o immig an
child en in Japan: An empi ical analysis o he assimila ion hypo hesis. Educa ional
S udies in Japan, 10, 93–107. doi:10.7571/esjkyoiku.10.93.
James, E., Benjamin, G. R., & Mend as, M. (1988). Public policy and p i a e educa ion in
Japan. Sp inge .
Ka iya, T. (2016). 10 unde s anding s uc u al changes in inequali y in Japanese educa ion. In
D. Chia acci & C. Homme ich (Eds.), Social inequali y in pos -g ow h Japan:
T ans o ma ion du ing economic and demog aphic s agna ion (pp. 149–165).
Rou ledge.
Ka iya, T., & Rosenbaum, J. E. (1999). B igh ligh : Unin ended consequences o de acking
policy in Japan. Ame ican Jou nal o Educa ion, 107(3), 210–230.
doi:10.1086/444216.
Kikkawa, T. (2006). Educa ion and social inequali y: Con empo a y educa ional
c eden ialism in Japan [Gaku eki o Kakusa Fubyodo: Seijuku Su u Nihon-Ga a
Gaku eki Shakai]. Tokyo Daigaku Shuppan.
Knipp a h, H. (2010). Wha PISA ells us abou he quali y and inequali y o Japanese
educa ion in ma hema ics and science. In e na ional Jou nal o Science and
Ma hema ics Educa ion, 8(3), 389–408. doi:10.1007/s10763-010-9196-5.
Ku oda, Y. (2018). The e ec o school quali y on housing en s: E idence om Ma sue Ci y
in Japan. Jou nal o he Japanese and In e na ional Economies, 50, 16–25.
doi:10.1016/j.jjie.2018.05.002.
Long, J. S., & F eese, J. (2005). Reg ession models o ca ego ical dependen a iables using
S a a (2nd ed.). S a a P ess.
33
Ma suoka, R. (2015). School socioeconomic composi ional e ec on shadow educa ion
pa icipa ion: E idence om Japan. B i ish Jou nal o Sociology o Educa ion, 36(2),
270–290. doi:10.1080/01425692.2013.820125.
McFadden, D. (1974). Condi ional logi analysis o quali a i e choice beha iou . In P.
Za embka (Ed.), F on ie s in econome ics (pp. 105–142). Academic P ess.
Minis y o Educa ion, Cul u e, Spo s, Science and Technology (MEXT). (2009). Making
public high schools ui ion- ee and high school en ollmen subsidies. Re ie ed om
h ps://wa p.da.ndl.go.jp/in o:ndljp/pid/11402417/www.mex .go.jp/b_menu/hakusho/h
ml/hpab200901/de ail/1305888.h m. Accessed Ap il 27, 2021.
Minis y o Educa ion, Cul u e, Spo s, Science and Technology (MEXT). (2012). School
basic su ey. Re ie ed om h ps://www.e-s a .go.jp/s a -
sea ch/ iles?page=1& oukei=00400001& s a =000001011528&yea =20120. (In
Japanese). Accessed Ap il 27, 2021.
Minis y o Educa ion, Cul u e, Spo s, Science and Technology (MEXT). (2020). Su ey on
issues ela ed o s uden ad iso y including p oblema ic beha iou s and uancy.
Re ie ed om h ps://www.e-s a .go.jp/s a -
sea ch/ iles?page=1&layou =no mal& oukei=00400304& s a =000001112655& esul
_page=1 (in Japanese). Accessed Ap il 27, 2021.
Minis y o Educa ion, Cul u e, Spo s, Science and Technology (MEXT). (2021a). O e iew.
Re ie ed om h ps://www.mex .go.jp/en/policy/educa ion/o e iew/index.h m.
Accessed Ap il 27, 2021.
Minis y o Educa ion, Cul u e, Spo s, Science and Technology (MEXT). (2021b). P i a e
school in Japan. Na ional Ins i u e o Educa ional Policy esea ch. Re ie ed om
h ps://www.nie .go.jp/English/educa ionjapan/. Accessed July 11, 2021.
34
Mon es, O., & Rubalcaba, L. (2014). School choice, equi y and e iciency: In e na ional
e idence om PISA-2012. In es igaciones de Economía de la Educación. In A.
Ga cía & I. Nei a (Eds.), In es igaciones deeconomía de la educación 9 (Volume 9,
Chap e 31, pp. 585–614. Asociación de Economía de la Educación.
Nakanishi, H. (2011). Reexamina ion o Japanese educa ional selec ion sys em: Focusing on
s a i ied con es mobili y. Jou nal o Child S udy, 17, 69–81.
Oakes, J. (1994). Abili y g ouping and acking in schools. In P. Husen (Ed.), The
in e na ional encyclopedia o educa ion, 1, 6–12.
OECD. (2009). PISA 2009 echnical epo . Re ie ed om
h ps://www.oecd.o g/pisa/da a/pisa2009 echnical epo .h m. Accessed Ap il 27,
2021.
OECD. (2012). PISA 2012 echnical epo . Re ie ed om
h p://www.oecd.o g/pisa/da a/pisa2012 echnical epo .h m. Accessed Ap il 27, 2021.
Ojima, F., & A amaki, S. (2018). Kōkōsei- achi no yukue – gakkō pane u chōsa ka a mi a
shin o o seika su no 30- nen [High school s uden s’ pa hways – 30 yea s o li e
cou se and school li e ajec o ies as seen om a school panel su ey].
Ono, H. (2001). Who goes o college? Fea u es o ins i u ional acking in Japanese highe
educa ion. Ame ican Jou nal o Educa ion, 109(2), 161–195. doi:10.1086/444265.
Sakai, A. (2010). Educa ional selec ion, ca ee decisions, and school suppo : The case o an
u ban comme cial high school. In Challenges o Japanese educa ion: Economics,
e o m, and human igh s (pp. 87–106). Teache s College (New Yo k).
Sakella iou, C. (2017). P i a e o public school ad an age? E idence om 40 coun ies using
PISA 2012-Ma hema ics. Applied Economics, 49(29), 2875–2892.
doi:10.1080/00036846.2016.1248361.
35
S e enson, D. L., & Bake , D. P. (1992). Shadow educa ion and alloca ion in o mal
schooling: T ansi ion o uni e si y in Japan. Ame ican Jou nal o Sociology, 97(6),
1639–1657.
Taki, H. (2010). The ela ionship be ween socio-economic s a us and academic achie emen :
Explo ing speci ic cha ac e is ics o educa ion and inequali y in Japan. Sociological
Theo y and Me hods, 25(2), 229–248. doi:doi.o g/10.11218/ojjams.25.229.
Taki, H. (2011). School ack and inequali y in Japanese uppe seconda y schools: An
analysis using PISA da a. Soshio oji, 55(3), 37–52.
doi:doi.o g/10.14959/soshio oji.55.3_37.
Tsuneyoshi, R. (2013). Junio high school en ance examina ions in me opoli an Tokyo: The
ad an ages and cos s o p i ilege. In G. DeCoke & C. Bjo k (Eds.), Japanese
educa ion in an e a o globaliza ion: Cul u e, poli ics, and equi y (pp. 164–182). New
Yo k: Teache s College P ess, Columbia Uni e si y.
Ushijima, K., & Yoshida, A. (2009). Shougakkou Ni Oke u Kyouiku No Shi su Ha Chika Ni
Eikyou Su uka? – Tokyo o Tokube suku No Chika Da a Wo Mochii a Kenshou.
Jou nal o Applied Regional Science, 14, 37–47.
Yamamo o, Y., & B in on, M. C. (2010). Cul u al capi al in Eas Asian educa ional sys ems:
The case o Japan. Sociology o Educa ion, 83(1), 67–83.
Yoshida, A., Zhang, L., & Ushijima, K. (2008). Gakkou No Shi u To Chika – Adachiku No
Chika Da a Wo Mochii a Kenshou. Ju aku Toshi Keizai, 68, 10–18.
36
Appendix
Table A1
Mul inomial Logi Model Es ima ion: 2009
Ci y and la ge ci y
Town and small own
Coe .
S d e o
Coe .
S d e o
Ca ego y: 1 – Public and academic
Fa he no wo king
-0.491
0.247
**
0.094
0.555
Mo he no wo king
0.359
0.166
**
0.005
0.288
Two-pa en amily
0.375
0.246
1.018
0.436
**
Siblings
0.174
0.250
0.454
0.412
G andpa en s li ing wi h he amily
-0.104
0.181
0.257
0.304
Cul u al possessions
0.181
0.098
*
0.160
0.170
Educ. le el mo he
0.176
0.083
**
0.117
0.145
Educ. le el a he
0.222
0.067
***
-0.036
0.124
Family weal h
-0.262
0.125
**
-0.462
0.220
**
Home educ. esou ces
0.001
0.088
-0.249
0.148
*
Highes pa en al occupa ional s a us
0.009
0.006
0.012
0.010
Cons an
-0.710
0.559
0.145
0.930
Ca ego y: 2 – Public and oca ional
Fa he no wo king
-0.631
0.279
**
0.405
0.569
Mo he no wo king
0.044
0.180
0.080
0.302
Two-pa en amily
0.256
0.266
0.754
0.454
*
Siblings
0.178
0.275
0.438
0.437
G andpa en s li ing wi h he amily
0.010
0.196
0.166
0.317
Cul u al possessions
-0.104
0.108
-0.003
0.179
Educ. le el mo he
0.024
0.091
-0.178
0.152
Educ. le el a he
-0.095
0.073
-0.367
0.130
***
Family weal h
-0.174
0.136
-0.361
0.231
Home educ. esou ces
0.030
0.096
-0.375
0.155
**
Highes pa en al occupa ional s a us
0.002
0.007
0.012
0.011
Cons an
0.836
0.606
2.095
0.969
**
Ca ego y: 3 – P i a e and academic
Fa he no wo king
-0.876
0.280
***
0.024
0.593
Mo he no wo king
0.393
0.175
**
0.099
0.309
Two-pa en amily
0.357
0.269
0.444
0.468
Siblings
-0.308
0.258
0.623
0.460
G andpa en s li ing wi h he amily
-0.312
0.191
-0.104
0.326
Cul u al possessions
0.192
0.103
*
-0.012
0.183
Educ. le el mo he
0.431
0.088
***
0.134
0.156
Educ. le el a he
0.429
0.072
***
-0.099
0.133
Family weal h
-0.062
0.130
-0.114
0.236
Home educ. esou ces
0.026
0.092
-0.324
0.158
**
Highes pa en al occupa ional s a us
0.015
0.007
**
0.003
0.011
Cons an
-3.289
0.599
***
0.067
1.005
37
Ca ego y: 4 – P i a e and oca ional
Base ca ego y
Base ca ego y
Log-likelihood
-3670.0
-1542.7
Numbe o pa ame e s
36
36
Obse a ions
3394
1495
***, **, and *: signi icance a he 1%, 5%, and 10% le els.
38
Table A2
Mul inomial Logi Model Es ima ion: 2012
Ci y and la ge ci y
Town and small own
Coe .
S d e o
Coe .
S d e o
Ca ego y: 1 – Public and academic
Fa he no wo king
-0.089
0.311
-0.673
0.520
Mo he no wo king
0.260
0.171
-0.453
0.369
Two-pa en amily
0.058
0.305
0.966
0.489
**
Siblings
0.335
0.228
0.771
0.418
*
G andpa en s li ing wi h he amily
0.193
0.203
0.130
0.400
Cul u al possessions
0.318
0.100
***
0.092
0.203
Educ. le el mo he
0.092
0.087
0.150
0.171
Educ. le el a he
0.192
0.071
***
-0.092
0.146
Family weal h
-0.079
0.146
0.326
0.307
Home educ. esou ces
-0.199
0.111
*
-0.249
0.230
Highes pa en al occupa ional s a us
0.006
0.004
-0.005
0.009
Cons an
0.105
0.555
1.938
1.088
*
Ca ego y: 2 – Public and oca ional
Fa he no wo king
-0.032
0.328
-0.053
0.528
Mo he no wo king
-0.017
0.182
-0.597
0.378
Two-pa en amily
-0.035
0.319
0.550
0.502
Siblings
0.389
0.247
0.835
0.440
*
G andpa en s li ing wi h he amily
0.338
0.214
0.333
0.408
Cul u al possessions
0.181
0.108
*
-0.205
0.210
Educ. le el mo he
-0.131
0.093
-0.082
0.176
Educ. le el a he
-0.110
0.076
-0.295
0.150
**
Family weal h
-0.172
0.156
0.305
0.316
Home educ. esou ces
-0.171
0.119
-0.054
0.236
Highes pa en al occupa ional s a us
-0.001
0.005
-0.007
0.010
Cons an
1.949
0.587
***
3.250
1.118
***
Ca ego y: 3 – P i a e and academic
Fa he no wo king
-0.036
0.320
-1.411
0.879
Mo he no wo king
0.292
0.175
*
-0.373
0.435
Two-pa en amily
0.200
0.317
1.351
0.773
*
Siblings
-0.027
0.231
0.144
0.506
G andpa en s li ing wi h he amily
0.230
0.208
-0.284
0.484
Cul u al possessions
0.409
0.102
***
0.322
0.239
Educ. le el mo he
0.248
0.089
***
0.126
0.207
Educ. le el a he
0.281
0.073
***
0.121
0.180
Family weal h
-0.011
0.149
0.457
0.372
Home educ. esou ces
-0.192
0.114
*
-0.114
0.277
Highes pa en al occupa ional s a us
0.013
0.005
***
0.005
0.011
Cons an
-1.546
0.576
***
-1.484
1.403
39
Ca ego y: 4 – P i a e and oca ional
Base ca ego y
Base ca ego y
Log-likelihood
-4022.1
-1159.7
Numbe o pa ame e s
36
36
Obse a ions
3608
1368
***, **, and *: signi icance a he 1%, 5%, and 10% le els.
