Advertisement

Data and Methods

How to Empirically Grasp the Implications of Shadow Education on Educational and Social Differentials in Japan?
Chapter
  • 633 Downloads

Abstract

The field of shadow education research remains largely empirically unexamined. Existing empirical studies often remain very limited in their explanatory power due to the use of inadequate data sources and limited methodical approaches. In addressing the question which data and methods are best suited to achieve empirically founded comprehensive findings when empirically analyzing the implications of shadow education for social inequalities, first, existing data sources are evaluated. Second, by understanding the mixing of methods as beneficial instead of mutually exclusive, two surveys are introduced, which complement each other and allow both quantitative and qualitative analyses. Whereas the Hyōgo High School Students (HHSS) surveys of the years 1997 and 2011 allow for quantitative analyses of the demanding side of shadow education and across student cohorts and times, the Juku Student and Teacher Survey (JSTS) of the year 2013 allows for qualitative analyses from the angle of providers of shadow education as well. Third, the importance to use alternative estimates besides odds ratios, etc. to achieve comparable findings, is stressed. Finally, it is shown in which of the four dimensions outlined in the Shadow-Education-Inequality-Impact (SEII) Frame (Access, Effects, Continuity, and Change) the introduced data can be used.

Keywords

Shadow education Juku Private tutoring Supplementary education Social inequality SEII Frame PISA study HHSS 1997 HHSS 2011 JSTS 2013 Japan 

4.1 Problematic

To achieve empirically founded comprehensive findings, a suitable data is of essence. In the process of evaluating existing data sources that have been used to analyze the phenomenon of shadow education in the Japanese or any other context, inevitably the question whether these data are actually fit for empirical analyses concerning the implications of shadow education for social inequalities occurred. For this work, thus, the question which data would be best suited to realize reliable findings concerning the issue of social inequality and shadow education in the Japanese context is central. In addition, fitting methods to achieve the intended findings based on the chosen data proves essential. These are hurdles all researchers have think more about when researching this topic in any context, whether local, regional, national, or international. Reliable findings bearing the potential to increase our knowledge on social inequality in educational attainment in all its forms can only be produced, if much care is put into the problematic of fitting theory and data.

As I outlined in the Shadow-Education-Inequality-Impact (SEII) Frame (Chap.  1, Fig.  1.2), to achieve reliable and convincing findings on the subject, the used data should provide the opportunity to analyze determinants for the access to shadow education and its effects for individual educational pathways in present Japan. To further understand the relative impact of shadow education on social inequalities, analyses over time focusing on the possible determinants for the continuity and change of this industry are necessary. On first glance, international large-scale assessment studies such as PISA and TIMSS seem to provide the best suited data for analyzing shadow education in different national contexts including Japan. Until now, however, only few international educationalists drew on these large, representative student samples with the aim to analyze the determinants or effects of shadow education (or rather out-of-school time lessons; e.g., Baker et al. 2001, Darby E. Southgate 2009, Ojima and von Below 2010, Entrich 2013, Park 2013, Darby E. Southgate 2013, Entrich 2014a, b, Matsuoka 2015). In fact, as is the case with all studies, a number of limitations is apparent, which are not only of methodological nature. Besides cultural patterns, the questionnaires’ item styles, and sample variations between countries (Hamano 2011: 3–4), we find definition and translation inaccuracy regarding the items used to capture shadow education (particularly for Japan), whereas the focus of these studies on certain student cohorts in the middle of their school careers (e.g., either 4th or 8th graders in TIMSS or 15-year-olds in PISA) does not allow to draw overall conclusions on the impact of shadow education for final educational success or participation across students’ whole school life courses (see also Entrich 2014a: 83–84, 96–97). Hence, to analyze shadow education and its implications for social inequalities in Japan, nation-specific datasets are better suited.

After reviewing several promising Japanese datasets such as the Social Stratification and Mobility (SSM) surveys,1 I decided to use the Hyōgo High School Students (HHSS)2 surveys of the years 1997 and 2011 as the main data basis for my quantitative analyses, because this dataset allows for broad analyses of the phenomenon of shadow education in Japan, as I will explain further in the subsequent part of this chapter. However, since even the HHSS data did not include all the necessary information to capture the implications of shadow education participation for educational and social inequalities, I designed and carried out an additional survey in 2013, called Juku Student and Teacher Survey (JSTS). In contrast to the HHSS data, the JSTS includes quantitative and qualitative data, allowing for generalizations based on “hard” data while “explaining the participants’ perspectives and developing an understanding of the meanings they attach to the phenomena of interest” (Fairbrother 2014: 76). As the above quotation by Todd D. Jick shall illustrate, I view the mixing of methods as beneficial instead of being mutually exclusive.

In this chapter, the purposes, specifics, and advantages of both surveys will be shortly introduced followed by an introduction of the variables used in the later carried-out analyses. Following this, I will discuss the specifics of the later applied methods and estimations. Finally, a brief summary is provided.

4.2 HHSS: Hyōgo High School Students Survey

The HHSS survey is a cooperative research project of several universities across Japan and was conducted under the guidance of professors Fumiaki Ojima (Dōshisha University) and Sōhei Aramaki (Kyūshū University).3 Targeting students at the end of their school life course (end of 12th grade), this survey provides a great amount of valuable data regarding the school life of high school students, their social backgrounds, and their expectations about life after school. Of particular interest for this work are students’ experiences with shadow education, which have been surveyed in detail. Furthermore, several items were conducted in retrospective, thus allowing for calculations across students’ whole school life courses.

The first HHSS survey has been carried out in 1981 on the largest of the four main Japanese islands, Honshu, in the prefecture Hyōgo, west central Japan. Hyōgo prefecture is not only part of the Kansai area, Japan’s second largest economical and second most populated area following the conurbation Tōkyō; it also consists of urban and rural parts ranging from the Sea of Japan in the North to the Inland Sea in the South, where the capital Kōbe is located. Kōbe is the sixth largest city in Japan with a population of 1.5 million people. Along with Ōsaka and Kyōto, Kōbe is part of the Kyōto-Ōsaka-Kōbe metropolitan center of the Kansai region. Consequently, Hyōgo prefecture reflects the average Japanese population and is thus a good area to conduct data on the Japanese schooling system as well. Here, differently ranked high schools across the prefecture have been chosen in order to reflect the current school life situation of high school students in this prefecture and to show the diversity not just between students but schools as well. Basically similar surveys were repeated in 1997 and 2011 using comparable questions on the core items, such as social background. Thus, cross-temporal comparisons are possible.

As shown in Fig. 4.1, the first survey included 18 schools (2.782 students), the second survey was carried out at 15 schools (2.397 students), whereas the third survey included 17 schools (3.826 students). Even though each survey includes different schools, ten schools remained constant in all three surveys. However, in spite of its high relevance, the prevalence of shadow education and its impact on social inequality formation has not been a prominent research subject for a long time and was rather neglected in earlier research. Therefore, the first HHSS survey includes no questions on shadow education participation at all. This changed in the second round of the survey, where questions concerning students’ participation in shadow education of several types during primary and middle school were included, before students’ experiences with shadow education during their high school years were added in the 2011 survey as well. Hence, certain cross-temporal comparisons are possible and may shed light on changes in the access to shadow education against the background of major societal and educational changes from the early 1990s to the late 2000s. In addition, students were asked about their future plans as well as their experiences in retrospective, e.g., their post-high school graduation plans. This provides us with valuable data regarding the whole school life course of students of the late baby boomer generation attending the school system in the 1990s in comparison to students of the post-baby boomer generation who attended school in the 2000 years.
Fig. 4.1

HHSS 1981/1997/2011 – sample overviews (Ojima 2001; Ojima and Aramaki 2013)

In the following, the variables used in the upcoming analyses of part two are shortly introduced. The focus lies primarily on data of the 2011 HHSS survey, complemented by data of the 1997 survey, where applicable. Based on the theoretical discussion presented in Chap.  3, of particular importance for the later carried-out analyses are students’ experiences with shadow education, their family background, educational aspirations, academic achievement, and the institutional and structural schooling context as reflected in school ranking. Accordingly, these variables are briefly introduced in the following.

4.2.1 Shadow Education Participation

Based on the discussion by Stevenson and Baker (1992), Bray (2010) defined shadow education as academic, supplementary, and private (see Chap.  1). The HHSS data provides us with information on the three major types of shadow education in Japan, which comply with this definition of Bray (2010): gakushū juku (academic juku, including yobikō), katei kyōshi (private home tutors), and tsūshin tensaku (correspondence courses). In the 2011 survey, students were asked whether they have participated in these three types of supplementary education during their primary, middle, and high school years or not. Students’ experiences with shadow education were separately encoded as dummy variables (1 = yes; 0 = no) for each type and for total participation experience.

4.2.1.1 Total Enrolment

As illustrated in Fig. 4.2, a high percentage of the 12th grade students who participated in the 2011 survey received shadow education lessons from 2000 to 2005, when they were in primary school (61%), and from 2006 to 2008, when they attended middle school (83%), but to a lesser extent from 2009 to 2011, during their high school years (38%). While correspondence courses were used more often in earlier grades (21%), the investment rates for private tutors (9%) and academic juku (73%) were highest in middle school, where the peak of out-of-school lesson attendance is reached.
Fig. 4.2

High school students’ experience with different types of shadow education during their primary, middle, and high school years, multiple responses possible, in % (HHSS 2011)

However, since multiple responses were possible, we also find students who participated in more than one type of shadow education concurrently during their school life, but this only applies for students who aimed for a university. In total, 9.1% of all students participated in more than one type of shadow education during primary school, 13.5% during middle school, and 4.7% in high school. Most multiple users supplemented attendance at a juku with correspondence courses (7.8% in primary, 9.1% in middle, and 4.1% in high school). In contrast, students who participated in all three types of shadow education at the same time are almost nonexistent (<1%).

Although it seems reasonable to assume that the growth of the Japanese juku-industry is already beyond its peak, the participation in shadow education has only decreased partially. When comparing the student cohorts of 1997 and 2011, who were enrolled at the same high schools (Fig. 4.3), we find a slight but overall decrease of the student population’s enrolment in shadow education in primary and middle school.
Fig. 4.3

High school students’ experiences with different types of shadow education in the 1990s and 2000s, in % (HHSS 1997/2011)

Of the 1997 cohort who attended primary school from 1986 to 1991, in total 66% participated in shadow education, whereas 88% of the same student cohort were enrolled in shadow education during their middle school years (1992–1994). In comparison, of the 2011 student cohort which attended primary school from 2000 to 2005, in total 60% participated in shadow education, whereas still 82% of the same student cohort reported to have participated in shadow education during their middle school years (2006–2008).

In general, our data report a decrease in enrolment rates in all fields of shadow education except for the utilization of correspondence courses during primary school, where a slight increase of 3% is found. Despite the low fertility rate and the carried-out reforms in education (see Chap.  10), the still high enrolment ratios during middle school might reflect the high concerns of families regarding admission to high ranked schools and thus social status maintenance. According to Takehiko Kariya, one of the leading Japanese sociologists in education, due to low birth rates simply more parents nowadays have the resources and the will to pay for their child’s additional learning support.4 Also, this indicates an unchanged difficulty level of certain entrance examinations and that the Japanese entrance examination hell is far from being overcome. According to our data, students in Hyōgo prefecture show generally similarly high enrolment ratios in shadow education in the 1990s and 2000s when compared to national data of the MEXT (see Chap.  7, Figs.  7.2 and  7.5).

4.2.1.2 Length of Enrolment

In addition to these numbers, we find considerable differences in the intensity of shadow education enrolment as measured in terms of length of enrolment. Table 4.1 gives an overview regarding students’ long-term and short-term investments in shadow education according to period of participation and type of shadow education. In total, 88% of all students of the 2011 cohort have shadow education experience, and most of these students made long-term investments, either over their whole school life course (from primary to high school, 33%) or until entrance to high school was achieved (from primary to middle school, 31%). Considerable differences in investment periods are apparent according to type of supplementary lessons though: whereas almost two thirds of the students who have participated in juku-classes at some point during their school life course made long-term investments, students who decided for lessons given by private tutors or correspondence courses mostly followed short-term investment strategies.
Table 4.1

Long-term and short-term investments in shadow education, in % (HHSS 2011)

 

Period of participation

Total

Academic juku

Private tutor

Corr. courses

Long-term

High, middle, and primary school

32.6

21.9

2.1

13.2

High and middle school

9.0

13.9

9.0

5.8

Middle and primary school

31.4

28.4

6.4

18.9

Total

73.0

64.2

17.5

37.9

Short-term

Only high school

0.8

2.2

9.9

6.7

Only middle school

21.1

26.8

55.4

15.4

Only primary school

4.1

5.6

15.9

35.3

High and primary school

1.0

1.3

1.4

4.9

Total

27.0

35.9

82.6

62.3

Valid responses

3308

2982

435

1112

Percentage of all valid responses (N = 3748)

88.0

79.6

11.6

29.7

4.2.1.3 Study Purposes

The generally found high participation ratios, however, do not account for students’ actual study goals, meaning the educational goal they want to achieve. To measure students’ study purposes when pursuing shadow education lessons during high school, I used students’ higher education entrance intentions (i.e., the chosen entrance methods to enter a university or college). In general, students can either enter the higher education sector or the job market following high school graduation. If the decision for staying in the education system is made, different entrance methods can be chosen. In the HHSS 2011 survey, students who responded that they would pursue an academic career following high school graduation (see also educational aspirations) were additionally asked how they intend to enter higher education.

Figure 4.4 shows the distribution of high school students according to used type of shadow education during high school (2009–2011) and planned entrance method to higher education at the end of 12th grade. Accordingly, the primary study focus of high school students using shadow education was preparation for entrance examinations (83.3%). Particularly students who participated in juku-classes and correspondence courses intended to enter higher education by taking the central entrance examination tests, the sentā shiken. In contrast, students receiving lessons from a private tutor at home more often seem to wish for entrance to higher education without taking the sentā. This kind of shadow education is only used by a very small proportion of students, though.
Fig. 4.4

High school students’ experience with shadow education, according to the type and planned university entrance method, multiple responses possible, in % (HHSS 2011)

In general we see that students consider an investment in shadow education a probate strategy to enter higher education by succeeding in entrance exams. As an adequate alternative strategy, entrance by recommendations or through admission office (AO) has become prevalent also. Recent ministerial data showed that the percentage of students entering higher education by alternative entrance methods has increased since the 1990s for entrance through AO, but not for other alternative university entrance methods (M. o. E. MEXT, Culture, Sports, Science and Technology 2012: 1, 4). However, the total number of students taking the sentā has remained stable since the 1990s (see Chap.  10, Fig.  10.1). Hence, students who intend to enter higher education continue to participate in shadow education to prepare for a university and college entrance exams at the end of high school and are very likely to take the central entrance exam (sentā shiken). Consequently, the main study focus of students attending juku or other types of shadow education at the end of high school is to succeed in the central university entrance exam. These students’ study focus emphasizes the traditional ways of school “transition” (shingaku). Still, a considerable proportion of students also continues to receive remedial and other supplementary lessons without focus on entrance exams. These students are thus understood as pursuing a general “supplementation” strategy (hoshū). In addition, we find a high percentage of students who do not confine themselves to using only one of the above study goals and thus follow a “comprehensive” approach (sōgō). Chapter  6 further discusses these main study purposes of students pursuing shadow education.

Based on the 2011 HHSS data, all high school students who stated that they intend to enter a university or college in the old-fashioned way by taking the central entrance exam are classified as having a shingaku study goal. Students who received supplementary lessons without intentions to enter higher education at all or via entrance exams were classified as pursuing a hoshū study goal. Students pursuing shadow education considering both strategies were classified as following an sōgō study purpose (Fig. 4.5).
Fig. 4.5

High school students’ experience with shadow education, according to study goal, in % (HHSS 2011)

According to our data, most students participate in shadow education during high school to prepare for university entrance exams (shingaku, 22%). Only about 7% of all students receive shadow education without having the intention to enter a university or college the traditional way (entrance exam) but pursue either alternative entrance methods or attempt to enter the job market after graduation (hoshū). An additional 9% of all high school students purchase lessons in the shadow with the intention to enter higher education through entrance exam or alternative methods (sōgō).

4.2.2 Social Origin

To measure the social origin of students as the most important explanatory variable in an analysis that is concerned with the identification of a possible increase of inequalities as an outcome of investments in shadow education, several different variables will be considered, which reflect certain dimensions of social origin fit for our analyses.

4.2.2.1 Economic, Social, and Cultural Status (ESCS)

Following Bourdieu (1983), students’ social origin is composed of different forms of parental human capital, which was defined as economic, social, and cultural. To reflect students’ economic, social, and cultural status (ESCS), we created a variable including parents’ highest occupational status5 and highest education level,6 as well as household possessions7 and number of books8 as the cultural component. Figure 4.6 shows the proportion of students participating in different types of shadow education across different social strata, ranging from the lowest quartile (low ESCS or underclass) over the two mediocre quartiles (middle ESCS or middle class) to the highest quartile (high ESCS or upper class).
Fig. 4.6

High school students’ experiences with different types of shadow education during their primary, middle, and high school years, according to ESCS, multiple responses possible, in % (HHSS 2011)

The displayed distribution of students according to the four quarters of socioeconomic and cultural background (ESCS) shows significant differences in the participation rates in different types of shadow education from primary to high school. With decreasing ESCS background, the participation in shadow education seems to decrease as well. Low ESCS students (bottom quarter) are generally participating in all kinds of shadow education to a lesser extent. Particularly during high school (2009–2011), most low strata students stopped their investment in shadow education.

Our data further suggests that the length of enrolment is depending on the ESCS of a student. Low ESCS strata seem to generally pursue short-term investment strategies, whereas high ESCS strata can afford to make long-term investments more often, in particular in the cases of juku-classes and correspondence courses. This supports the argument that low ESCS students who pursue shadow education follow purposeful and goal-oriented investment strategies. Because low ESCS students do not possess the resources to make long-term investments, short-term investments simply have to pay off (see Appendix, Fig. 4.14). Furthermore, significant differences in the proportion of students which pursue certain study goals when participating in shadow education in high school are found across ESCS groups. Almost one third of all high ESCS students pursue shingaku-oriented juku (focus on entrance exam preparation), whereas only 7% of low ESCS students invest in these kinds of lessons. The same general discrepancy is found for correspondence courses and private tutors. Besides these general differences in enrolment rates, similarities are also detected. If low ESCS strata find ways to participate in juku-classes during high school, they will mostly attend shingaku programs. Correspondence courses are more often pursued using shingaku programs but also sōgō programs (focus on entrance exam preparation and remediation). In contrast to each other, high ESCS strata use private tutors mostly with shingaku focus; low ESCS strata mostly receive such lessons with hoshū purposes (focus on remediation) (see Appendix, Fig. 4.15).

4.2.2.2 Highest Parental Education Level

Bukodi and Goldthorpe (2013) stressed that social origin needs to be “decomposed” to get a better understanding of how certain components of social origin might affect students’ educational attainment. The authors proposed using components such as class, status, and education of parents separately when analyzing educational inequalities, especially when making comparisons over time to assess whether stability or change occurred in the impacts of social origins on educational attainment. Since the education level strongly determines social status in Japan (Kikkawa 2006), the parental education level is supposed to play a crucial role for social reproduction and thus is the most important component of the ESCS in the Japanese context. Also, this measure of social origin is very well suited for comparisons across time, since there are no variations within its definition and measurement.

In both HHSS surveys (1997 and 2011), parents’ highest education level was measured separately for fathers and mothers using a five-degree scale ranging from the highest education level to the lowest. Category one thus includes all parents holding a university (daigaku) degree of at least four years study, including parents with a master or doctor degree also. Following this, parents who have obtained a degree from a junior college (tanki daigaku) are placed. These institutions differ from universities particularly in terms of length of study, which is only about 2–3 years. Another level of tertiary training is achieved at the technical colleges, which show a stronger focus on vocational training. The fourth category consists of those parents who did not enter tertiary education but completed high school. The lowest education level is the middle school diploma. When comparing the enrolment ratios for parents from the 1997 and 2011 cohorts (see Appendix, Table 4.8), our data reflects the earlier verified ongoing educational expansion (see Chap.  2, Fig.  2.2). In particular, mothers of the 2011 cohort are more often enrolled in tertiary education. Only few parents remain with a middle school diploma. Most parents have achieved a high school diploma or higher.

To represent parents’ highest education level in each household, an additional variable for both parents’ education level was computed reflecting the highest level of education achieved by both parents. Accordingly, in 2011 in almost 44% of all households, at least one parent holds a university diploma, whereas almost 21% have at least a college diploma, leaving about 36% without tertiary education training. In comparison, in 1997 a considerably higher percentage of households are found where no parent holds a college or university diploma (53%).

This categorization of parental education level allows the construction of three educational strata, from which students originate: (1) advantaged educational background (parents possess a university degree), (2) mediocre educational background (parents possess a college degree), and (3) disadvantaged educational background (parents have no tertiary education training). Similar to the found differences in participation rates according to ESCS quartile, differences according to educational stratum are to be expected. However, as Fig. 4.7 shows, even though such differences are apparent, they are not extreme.
Fig. 4.7

Students’ experiences with shadow education, according to educational stratum, in % (HHSS 1997/2011)

Students who went to middle school in the 1990s (1992–1994) show no particular differences in enrolment in shadow education at all (advantaged educational background, 89%; disadvantaged educational background, 88%). In contrast, a smaller proportion of students from disadvantaged educational background who received shadow lessons during their middle school years in the 2000s (2006–2008) is detected (78%). Nevertheless, enrolment ratios during middle school remain considerably high across all educational strata.

4.2.2.3 Number of Siblings

To reflect households’ varying sizes and distribution of income, the number of siblings9 will be considered as another social origin category in some of my analyses. Previous research indicates that students without siblings are more likely to attend out-of-school classes since their parents can concentrate their resources on only one child (e.g., Tomura, et al. 2011). The number of siblings should thus affect whether and in what dimension investments in shadow education are affordable. This variable is encoded as possessing the following values: 1 = no siblings, 2 = one sibling, and 3 = two or more siblings. These categories seem adequate, since only 7.5% of the students had more than two siblings in 2011. According to our data (see Appendix, Fig. 4.16), there seem to be almost no differences in the proportion of students with or without siblings using shadow education during middle school. However, during primary and high school, students with more siblings are less often enrolled in extracurricular lessons in the shadow. It has to be mentioned here that 10% of the whole 2011 HHSS student sample reported to have no siblings at all, whereas most students had one sibling (58%).

4.2.2.4 Parents’ and Students’ Educational Aspirations

Following Rational Choice Theory (RCT) and Shadow Education Investment Theory (SEIT) approaches (see Chap.  3), particularly educational aspirations are decisive for the making of educational decisions and thus the formation of educational and social inequality. However, existing research refrained from measuring aspirations separately and focused their analyses on constant social origin variables such as highest parental education level, occupation status, cultural resources, etc. (e.g., Erikson and Jonsson 1996, Jonsson and Erikson 2000, Fujihara 2011, Breen, et al. 2014). Hence, besides “decomposing social origins” into classic measures such as education level, class, and status as proposed by Bukodi and Goldthorpe (2013), I argue that especially in the Japanese case, educational aspirations need to be included in analyses concerned with educational choices or investments, as they determine the reproduction of social inequality.

To measure educational aspirations, I use the future pathway goals (shinro kibō) parents have for their children and students’ own career ambitions, respectively. To collect such data, in the 1997 and 2011 HHSS survey, students were asked about their post-high school graduation plans. Students could choose between (1) start working (including working at home) immediately after graduation (shūshoku), (2) four-year university (daigaku), (3) junior college (tanki daigaku), (4) technical college (senmongakkō), (5) some other institutions, or (6) undecided. Using the same scale, students were asked whether they know about their parents’ aspirations for them. If students were not sure or did not know their parents’ aspirations, their answers were included in the “undecided” category. The final aspirations of parents and students were categorized on a five-degree scale from highest to lowest educational aspiration, omitting aspirations for other institutions (see Appendix, Table 4.9).

Since all participating students were in 12th grade, the percentage of students still undecided regarding their future career was marginal (2.6%). The majority of all parents and students aim for university entrance (>50%). However, there is also a considerable percentage of students who did not know or were not sure about their parents’ expectations. In particular, the father’s wish for his child’s future career is often unknown by the child (23.9%). This suggests that parents’ influence on students’ aspirations for life after school was not very strong in these cases, and thus students themselves might have greater say in general decision-making, as I will further discuss in Chap.  5. As for parents’ aspirations, we assume that these were already present when their children attended primary school (2000–2005) and have remained relatively stable. I thus use the reported final aspirations throughout my analyses. In contrast, students’ aspirations are formed over the courses of their school lives. Most students remain insecure regarding their future career until they reach high school (Table 4.2). Still, we find considerable proportions of students who knew where they wanted to go after graduating from high school at the early stages of their educational lives, ranging from approximately 11% in primary school to 35% in middle school.
Table 4.2

Educational aspirations of students and planned university entrance method, in % (HHSS 2011)

 

Primary school

Middle school

High school

Planned entrance method

(2000–2005)

(2006–2009)

(2009–2011)

Exam

Othersb

University

8.5

24.0

59.3

40.4

19.0

 Definitely top university

13.8

11.7

2.2

 If possible top university

21.3

15.0

6.3

 Average university is enough

24.1

13.7

10.5

Collegea

1.0

4.3

17.8

5.3

14.6

Job

1.4

6.3

19.7

No decision yet

89.1

65.4

0.9

Valid responses

3365

3365

3388

3358

aThis category consists of junior and technical college

bThis category includes admission by recommendation (by school or publicly offered), admission office, and other possibilities

In addition to general aspirations, we have data on students’ concrete goals concerning their entrance to university. Almost 60% of students have university aspirations, and the majority aims at first-class universities (35%). However, only a few students think they can enter top universities without taking the entrance examination (8.5%). Careful preparation is thus essential, and an investment in shadow education promises a higher pass probability in these exams.

In accordance with theories on social reproduction, students’ educational aspirations vary considerably across social strata (Fig. 4.8).
Fig. 4.8

High school students’ educational aspirations in primary, middle, and high school, according to ESCS, in % (HHSS 2011)

Not only does a higher percentage of high ESCS students pursue the highest educational track following high school graduation (university), these students also tend to make this decision earlier than their peers from less advantageous background. During middle school, 43% of students from high ESCS backgrounds have already decided to enter university following high school, whereas only 13% of low ESCS students have made the same decision at this point. In contrast, low ESCS students favor entering the labor market to a higher degree compared to other social strata (low ESCS, 10% in middle school and 23% high school; high ESCS, 2% in middle school and 11% in high school; similar results are found for educational strata; see Appendix, Table 4.10). Whether ambitious students from disadvantaged ESCS backgrounds can actually increase their chances of entering university by benefitting from gaining access to shadow education remains to be scrutinized.

Taking into account differences in educational preferences over time, Table 4.3 shows students’ educational aspirations in the 1990s compared to the 2000s. In particular, the percentage of high school students aiming for university increased by more than 8% from 1997 to 2011. However, future educational opportunities are already predominantly predestined at the end of high school, since the rank of the high school a student attends largely determines whether certain educational goals can be achieved. Therefore, educational aspirations at the end of middle school play a major role for future opportunities and thus students’ chances to achieve a high education level. In addition, at this stage in students’ educational life course, students and their families are often still uncertain which future pathway they should pick. This uncertainty concerning future prospects slightly increased from 1994 to 2008 (by more than 3%), thus leaving 67% of all middle school students insecure about their future educational pathway.
Table 4.3

Students’ educational aspirations in the 1990s and 2000s, in % (HHSS 1997/2011)

  

University

Junior college

Technical college

Other

Still uncertain

Job

Valid responses

Primary school

1991

 7.3

 0.9

 0.3

0.2

90.0

 1.3

1734

2005

 8.3

 0.1

 0.9

0.1

89.1

 1.6

1967

Middle school

1994

20.1

 2.8

 2.4

0.4

63.4

10.8

1734

2008

21.1

 0.4

 3.4

0.1

67.0

 7.9

1967

High school

1997

47.1

11.1

12.3

3.2

 1.7

24.7

1773

2011

55.4

 4.1

14.8

0.5

 2.5

22.7

2055

Only students of those schools (n = 10) were considered, which took part in both surveys

To increase their chances of gaining entrance to high-ranked high schools, shadow education is pursued by most middle school students. However, from 1994 to 2008 the percentage of students participating in shadow education decreased partly (Table 4.4). In particular, students which are uncertain which pathway to follow have considerably less often received lessons in the shadow in 2008 (65%) compared to 1994 (88%).
Table 4.4

Percentage of students who participated in shadow education during middle school, according to educational aspirations, 1992–1994 and 2006–2008 (HHSS 1997/2011)

 

University

Collegea

Still uncertain

Job

Valid responses

1992–1994

93.5

83.1

87.5

81.1

1637

2006–2008

87.5

77.6

65.0

79.8

1958

Only students of those schools (n = 10) were considered, which took part in both surveys

aThis category consists of junior and technical college

It seems that while students with high aspirations generally continue to make investments in shadow education, students without clear goals more often question whether such an investment is necessary.

4.2.3 High School Stratification

4.2.3.1 High School Ranking

In Japan, all high school graduates have the opportunity to access higher education. However, two major high school tracks can be divided: a general academic (futsūka) and a vocational track (senmongakka). But, as stated earlier, the Japanese academic high school system is also highly stratified through the prestige high schools have gathered according to the percentage of students that attain entrance into high-ranked universities (Stevenson and Baker 1992: 1641, Ojima and von Below 2010: 277). Following Shirakawa (2013), the academic track was subcategorized using the expected advancement ratio of students to higher education, students’ socioeconomic background, and their academic standing in middle school to create a variable to depict the ranking of Japanese high schools. Following this, the 17 high schools of this study were classified into four different ranks: Academic A, B, and C as well as vocational high schools. In accordance with the expected outcomes of educational expansion, a higher percentage of the 2011 cohort attended Academic A high schools (1997, 17%; 2011, 26%), whereas fewer students entered vocational high schools (1997, 24%; 2011, 21%; see Appendix, Table 4.11). Taking into account this high school stratification, we are able to differentiate students’ participation ratios in shadow education. Figure 4.9 shows the percentage of students who participated in shadow education at the time they were in primary, middle, and high school according to the rank of high school they finally attended.
Fig. 4.9

High school students’ experiences with different types of shadow education in primary, middle, and high school, according to high school ranking, multiple responses possible, in % (HHSS 2011)

During middle school, most parents seem to somehow manage the high costs for the participation in shadow education, as it appears to have become an accepted part of children’s school life. In high school, however, a high diversity can be found. Clearly students that attend Academic A high schools have more experience with shadow education in general and continue to invest far more in these lessons after middle school (80.4%) than students in Academic C (17.9%) or vocational high schools (11.6%), where the majority quits their investment. This indicates that the additional financial, time and effort investments are no longer considered worthwhile for students in Academic C high schools or vocational schools, since career aspirations make such an investment unnecessary.

4.2.3.2 Type of High School’s Administration

In addition to this ranking, we need to consider whether students attend private or public schools, since students’ participation ratios in shadow education vary between those school types (MEXT 2014). Hence, a dummy variable for administrative type of school (1 = public, 2 = private) is included. According to the 2011 HHSS survey data, on average twice as many public high school students (44%) participated in shadow education lessons during their high school years (2009–2011) than students attending private institutions (22%).

4.2.3.3 Location of High School

Furthermore, previous research has shown that the location of a school has effects on whether or not students participate in shadow education, because there might be less availability in rural areas (Stevenson and Baker 1992: 1647–1650). Also, differences in the effects of shadow education were verified according to residential area (Entrich 2014a). Based on our data, we classified three areas: metropolitan (high-density areas with more than one million residents and more than 4000 residents per square kilometer), urban (semi-dense areas with more than 100,000 but less than one million residents and more than 900 residents per square kilometer), and rural (less than 100,000 residents and less than 900 residents per square kilometer). As shown in Fig. 4.10, participation in shadow education is highest in urban areas, not metropolitan areas. In fact, more students of rurally located high schools pursue supplementary lessons than students living in metropolitan regions. For the latter students, especially correspondence courses are more valued, as these lessons are easily accessible in rural areas where only few juku exist.
Fig. 4.10

High school students’ experiences with shadow education in high school, according to location, in % (HHSS 2011)

4.2.4 Post-High School Status

After entrance to high school is achieved, students and their families need to make up their minds regarding future options and decide which pathway to follow. Besides leaving the schooling system by entering the job market after high school graduation, all high school graduates have the opportunity to enter higher education. As data of the national Ministry of Education show, approximately 57% of the 18-year-old cohort proceed to higher education in present Japan (MEXT 2017). However, due to the still existing entrance examination system, a hierarchical ranking of universities exists. To calculate whether students achieve a higher education status as a result of shadow education investments, calculations need to focus on transition to university differentiating the university level into different ranks. However, in the HHSS survey, we did not directly measure whether students actually entered a certain university. Since our data was surveyed at the end of high school, we only measured whether students aspire to enter a certain kind of university or leave the schooling system. As described above (see educational aspirations), students were first asked to give details about their post-high school career plans, ranging from entering university to joining the workforce. Additionally, all students aspiring for higher education were asked whether they intend to enter a top university or not. Of course, these categories only reflect what students actually aspire, but not, how certain it is that they will actually achieve their goal. Because of that, a further item was included in the questionnaire asking whether students are sure that they will realize their set goal. On a three-degree scale, the respondents could choose between: (1) I will definitely realize my goal; (2) I might realize my goal, if I am capable enough; and (3) I do not attach to much value to the realization of my goal. According to our data, at the end of 12th grade, most students (82%) are sure that they will achieve their chosen post-high school goal (see Appendix, Table 4.12). Particularly students who intend to enter top universities following high school graduation are generally sure that nothing stands between them and the realization of their goal (approximately 90%). For this analysis, only the latter students are considered as possessing “realistic future goals,” since for the rest of the cohort, a high probability remains that they will not enter their chosen school. These “realistic future goals” of students are used as a proxy for students’ post-high school education status, i.e., whether students enter a top university (1), an average university (2), a college of different sort (3), or the job market (4). This assumption seems justified considering that families have to decide very early whether a certain kind of university is pursued and how to prepare for the entrance, as I will further discuss in Chap.  5.

Using the created variable reflecting students’ post-high school status, Table 4.5 verifies that students who will enter high-ranked universities pursue mostly lessons concentrating on the preparation of entrance examinations (shingaku, 78%) during high school. Also, lessons where students prepare for entrance examinations and, at the same time, aim for alternative university entrance methods by improving their grades, etc. (sōgō, 63%) are considerably often used. In contrast, students focusing on the improvement of school grades or with remedial needs (hoshū) also often aim for college (26%) or sometimes the job market (13%). Hence, only those students who are certain to enter higher education following high school graduation will enroll in shingaku or sōgō lessons. The generally few students pursuing shadow education during high school knowing that they will enter the job market after graduation are likely to only enroll in hoshū lessons due to certain study difficulties. For these and other students who chose hoshū and sōgō lessons, catching up in certain academic subjects might be the major reason for participation. The impact of shadow education on transition to higher education is thus very likely varying according to the pursued study goal, as I will investigate in Chap.  6.
Table 4.5

Students’ post-high school status in relation to their participation in shadow education in high school, according to type and purpose, in % (HHSS 2011)

Type and study purpose of shadow education in high school

Post-high school status

Top university

Average university

Collegea

Job

Total

All types

Shingaku

77.5

20.8

1.7

525

Sōgō

62.5

29.1

8.4

275

Hoshū

32.8

28.4

25.5

13.2

204

Academic juku

Shingaku

78.7

19.7

1.5

456

Sōgō

66.2

27.1

6.7

225

Hoshū

32.0

29.9

25.9

12.2

147

Private tutor

Shingaku

69.6

21.7

8.7

23

Sōgō

53.3

33.3

13.3

15

Hoshū

34.5

27.6

31.0

6.9

29

Cor. courses

Shingaku

75.2

24.8

133

Sōgō

49.3

41.8

9.0

67

Hoshū

30.6

33.3

16.7

19.4

36

Valid responses

1053

609

508

470

2640

aThis category consists of junior and technical college

4.2.5 Academic Achievement

Because of the merit-based class consciousness in Japan (Rohlen 1983: 311), the academic achievement of students is believed to be decisive for educational success in general. The transition from one school level to the next is meritocratic in nature and thus highly determined by academic performance. Accordingly, generally high-achieving students are believed to automatically enter higher-ranked schools, since they are able to pass more difficult entrance examinations or will get a recommendation for a high-ranked school. Also, RCT research stresses that the academic achievement level of students is critical in making educational decisions. This includes the decision for shadow education, since students’ participation was found to depend on academic standing in school, i.e., students grade point average (GPA; e.g., Seiyama and Noguchi 1984, Stevenson and Baker 1992), and thus students’ academic standing needs to be included as an influential determinant to predict whether shadow education is pursued or not. Previous research has stressed that different types of shadow education are chosen according to academic standing and purpose of study (e.g., Rohlen 1980, Komiyama 1993, 2000, Roesgaard 2006). Also, SEIT argues that the primary outcome of investments in shadow education is the improvement of academic achievement level, i.e., school grades or performance in entrance exams. However, empirical evidence is scarce. In my upcoming analyses, students’ academic standing will thus be included.

The HHSS data includes the participating 12th graders’ GPA during high and middle school. This GPA is generally coded on a five-degree scale ranging from 5 (= high academic standing) to 1 (= low academic standing). Using the same scale, we are able to differentiate students into five main achievement groups, according to their participation in shadow education in their respective school levels (see Appendix, Fig. 4.17). Of our sample, 91% of all students with a high GPA in middle school responded to have participated in shadow education during this period. With lower academic achievement level, participation rates in middle school decrease. A similar clear tendency is not found for students’ high school years. Apparently, students’ academic achievement level in high school is strongly related to the rank of high school. Accordingly, participation rates vary much stronger between high schools than general academic achievement level.

As is general usage in social sciences, I will treat this ordinal variable as if it was metric in nature in my upcoming analyses.

4.2.6 Gender

Since Japanese parents traditionally have different ideas about the future of boys and girls and generally tend to invest more in the education of boys (Schultz Lee 2010: 1582), a dummy variable concerning gender (1 = male, 2 = female) is generally included in my upcoming analyses.

4.3 JSTS: Juku Teacher and Student Survey

In addition to the Hyōgo High School Students survey, I designed the Juku Student and Teacher Survey (JSTS) to complement the existing data by targeting a different population: jukusei, students enrolled at private schools in the shadow education sector, as well as the operators of this industry, jukuchō (juku principals), and their teachers (jukukōshi). Based on a multi-sequence-multimethod design (Teddlie and Tashakkori 2006), this survey was carried out by the author during several fieldwork periods from 2012 to 2014 following a two-stage random sampling method:

First, juku were chosen based on the following premises:
  1. 1.

    Specialization: Juku of different types were chosen.

     
  2. 2.

    Success: Following the yūshōreppai (survival of the fittest) principle, only juku were targeted, which are at least 30 to more than 50 years in business, thus managing to stay competitive and successful due to specific reasons.

     
  3. 3.

    Size: Juku of different sizes and organizations were chosen, thus including small- or middle-sized juku (kojin/chūshō juku) and major corporations operating as chain juku (ōte juku).

     
  4. 4.

    Location: Juku in Japan’s metropolitan centers Kansai (Kyōto, Ōsaka) and Tōkyō (Setagaya) and in two less populated prefectural cities outside these metropolitan areas (Shiga prefecture: Kusatsu; Fukushima prefecture: Iwaki) were chosen.

     
Before collecting the data in the form of a questionnaire-based survey early in 2013, contacts to juku were established on the basis of personal recommendations and introductions. After winning general interest of several juku for my research, meetings were set. In these meetings, I usually first sat down for an interview with the head of the school (jukuchō), before taking a tour through the school, visiting the classrooms, taking a look at the teaching approaches and methods, and getting in contact with teachers and students. The exploratory semi-structured interview with the jukuchō followed several guiding questions to deliver basic information about the juku, its teachers, students, supply, and general development, while leaving room for the interviewee to talk about issues they feel are important. Following this first meeting, generally a second visit was paid to provide the school with the needed questionnaires for jukusei (students attending a juku) and jukukōshi (teachers at a juku), including jukuchō.
Second, a questionnaire survey following a simultaneous design, i.e., simultaneous collection of quantitative and qualitative data (Gürtler and Huber 2012: 39), was carried out from January to March 2013. The eight juku that took part in the survey represent five kojin/chūshō juku (JukuA to JukuE) and three chain juku (JukuF to JukuH) with numerous branches. In total, 20 juku schools participated, of which 500 jukusei and 102 jukukōshi filled in the questionnaire (Table 4.6). JukuF and JukuG are two different chain juku belonging to the same parent company, a local big player operating in the prefectures Kyōto, Ōsaka, Hyōgo, Shiga, and Nara. JukuH is another large joint stock company operating nationwide. Hence, at these two ōte juku, the number of students and teaching staff was too large to include everybody in the sample. After setting the targets, we decided on a suitable number of questionnaires to be given out. It was, however, not always possible to get as much participants as originally envisaged due to certain circumstances, such as the juku’s time schedule, individual timetables, and seasonal variation in study periods.
Table 4.6

JSTS 2013 – sample overview and return ratios

 

kojin/chūshō juku

ōte juku

Total

JukuA

JukuB

JukuC

JukuD

JukuE

JukuF

JukuG

JukuH

Juku

Official classification

Shingaku

Doriru

Sōgō

Hoshū

Shingaku

Shingaku

Sōgō

Yobikō

Number of branches

1

1

1

1

1

>50

>70

>100

>200

Sample

1

1

1

1

1

4

2

9

20

Jukukōshi

Total number

6

2

15

4

10

>100

>200

>100

>500

Targeted population

6

2

10

4

10

80

20

20

152

Sample

5

1

2

4

9

61

14

9

102

Return ratio

83%

50%

20%

100%

90%

76%

70%

45%

67%

Jukusei

Total number

120

50

350

90

80

>1000

>1000

>10000

>10000

Targeted population

80

50

80

80

40

200

100

300

930

Sample

45

38

0

62

31

100

0

224

500

Return ratio

56%

76%

0%

78%

78%

50%

0%

75%

54%

As shown in Table 4.6, return rates of jukusei vary from 50 to 78%, whereas return rates of jukukōshi vary considerably as well. However, apart from JukuC and JukuG, where the jukuchō did not want to be responsible for the missed study time of their jukusei, at all schools, a suitable student sample was achieved. The samples of jukukōshi are generally smaller, since their responses are meant to provide general information on the juku and their connection to regular schools and to complement these data with open items to get qualitative statements of the operators’ side about reactions to recent changes in the regular schooling system and decreasing student populations. In addition to the questionnaire survey, personal conversations with students, parents, juku-operators, and researchers in this field served to get a greater understanding about functions and implications of the Japanese juku-industry and its relation to formal education. Follow-up research was carried out in August and September 2013 as well as in June 2014.

Since the quantitative data of the student survey will be used to predict why families choose a certain juku-type over another, certain variables need introduction. Particularly important for our analyses prove, again, social origin and school ranking variables. In addition, the JSTS provides us with the opportunity to include concrete reasons for choosing a certain juku.

4.3.1 Social Origin

To measure the social origin of students, gender and educational stratum are considered. Similar to the variables concerning gender and educational stratum in the HHSS survey, gender will be included as a dummy variable (1 = male, 2 = female), while educational stratum as the major indicator for students’ family background was measured in terms of fathers’ and mothers’ highest completed education level. Both parents’ highest education level was independently measured using a five-degree scale ranging from the highest education level to the lowest: (1) master’s degree or a doctorate, (2) university degree, (3) technical or junior college degree, (4) high school diploma, and (5) middle school diploma. As shown in Table 4.7, the conducted sample of jukusei originates from more advantaged educational backgrounds for the most part (>70%).
Table 4.7

Highest parental education level, distribution in % (JSTS 2013: Student Survey)

 

Graduate school

University

Collegea

High school

Middle school

Valid responses

Father

8.2

64.3

6.9

19.5

1.1

437

Mother

2.2

42.8

28.4

25.3

1.3

451

Parentsb

9.1

67.0

9.1

14.1

0.7

460

aThis category consists of junior and technical college

bAt least one of both parents holds a degree in this category

In general, fathers have more often completed graduate school or university, whereas mothers more often hold a college degree. To represent parents’ highest education level in each household, an additional variable for both parents’ education level was computed reflecting the highest level of education achieved by at least one of both parents. Accordingly, the participating students’ parental education level is generally considerably high. Only approximately 15% of all households have no tertiary education degree. In contrast, in 76% of all sampled cases, at least one of both parents holds a university degree.

4.3.2 School Stratification

To reflect the structural frame of the Japanese schooling system, variables concerning the stratification of schools need to be controlled for. To measure this stratification of schools, data on the reputation of a school to express the ranking of schools and their administration type (private/public) were collected. In the JSTS Student Survey questionnaire, students were asked whether the school they attend has a generally high reputation (1), an average reputation (2), or a low reputation (3) and whether they attend public schools (1) or private schools (2). According to my data, most students attend public schools (78.7%) and rate the reputation of their school as average (64.5%).

4.3.3 Unmet Demands

To find out why families choose a certain type of juku over another type in present Japan, particularly students’ reasons to enter their juku and their view on school classes have been surveyed in detail. The following item seems best suited to predict the likelihood of entering a certain juku-type to meet new demands: “Thinking about your reasons to attend this juku, please answer to the different categories below. Please choose from [that is true] to [that is not the case at all]. Make your mark where you think it is most applicable” (Q22, Student Survey). Using a five-degree Likert scale, students could choose from that is true (1) to that is not the case at all (5). The four most applying response categories will be used as metric predictors for juku attendance and are displayed in Fig. 4.11.
Fig. 4.11

Students’ motives to attend juku, in % (JSTS 2013: Student Survey)

According to these data, the main motives to attend juku are students’ intention to perform well in school, i.e., get good grades and prepare for tests and exams. In addition, students often rate the teaching methods and the general learning environment at juku higher than at their school. Particularly important are further lesson contents outside the regular school curriculum and high ambitions, i.e., the transition to a high-ranked school. Only few students attend juku just because they like studying; most of them need help with their homework or want to meet or make new friends. Whether their parents’ are actually glad that they attend such a school is of less concern to students. Some students also find it important to receive individual support from the teacher at juku. Many of these motives thus imply that students do not feel that regular school actually meets all their demands, whereas several categories point to shortcomings that might have been created through the new reform course (see Chap.  10).

In addition to students’ motives to attend a juku, their general view on the quality of regular school’s classroom instruction was measured using the four items displayed in Fig. 4.12. Accordingly, only few students actually think that school is a waste of time (16%). However, a much higher percentage of students is not sure whether school has actually prepared them sufficiently for their future life (43%) or rate class in school as not very good (55%). Also, almost one third feels that they receive not much attention when they need help during class. For my upcoming analysis in Chap.  8, the latter two variables concerning help during class (“when I need help during class, the teacher helps me”) and instruction quality (“in general, class in school was not very good”) will be included as dummy variables (1 = I agree; 2 = I disagree).
Fig. 4.12

Students’ view on regular school’s instruction quality, in % (JSTS 2013: Student Survey)

4.4 Methods

To achieve reliable and convincing findings, the methodological approaches used in each of the four subsequent empirical chapters are of great importance. To methodically approach the first dimension outlined in the SEII Frame (Chap.  1, Fig.  1.2), the Access to shadow education, binary logistic regression analyses are carried out to predict the likelihood of actually deciding for and taking part in shadow lessons based on major influential variables, such as social origin and educational aspirations of parents and students. This kind of regression analysis calculates the probability that the event occurs that students actually take part in lessons in the shadow in reference to those who do not. By applying Rational Choice Theory as a theoretical approach, these analyses are used to measure whether families actually decided to make an investment in shadow education at a certain point of time in the student’s school life course or not.

To further estimate the impact of shadow education investments on educational success in Chap.  6, thus concentrating on the Effects dimension of the SEII Frame, multinomial logistic regression analyses are applied, allowing for similar regressions with the difference that nonbinary but nominally scaled variables, in this case the ranking of high schools and students’ post-high school status, are used as independent variables reflecting educational success. Hence, in this analysis, more than two outcomes are possible, ranging from high educational returns in terms admission to an advantageous or prestigious school level to low educational returns without comparable returns.

In Chap.  7, the focus shifts to the relativity of Access to and Effects of shadow education across time, applying the Continuity dimension of the SEII Frame. Similar to Chaps.  5 and  6, access to and effects of shadow education are predicted using binary and multinomial regressions; only this time these analyses are meant to compare different cohorts and models across time to clarify the extent of possible inequality increase or decrease through shadow education. Hence, quantitative data of the HHSS 1997 and 2011 surveys are used.

In Chap.  8, particularly the relativity of Access to shadow education across time is questioned emphasizing the Change dimension of the SEII Frame based on quantitative and qualitative data alike. To achieve comprehensive findings, I will approach this final issue based on mixed methods, using a so-called within-method triangulation model (Jick 1979: 602–611, Gürtler and Huber 2012). Hence, in addition to quantitative descriptive and multinomial logistic regression analyses in the first part of the results section, qualitative content analyses are used to help explain the quantitative findings (Creswell et al. 2003), i.e., explain whether the found effects are an outcome of certain developments in the Japanese education system.

Furthermore, it is worth noting that all completed quantitative regression analyses attempt to test whether shadow education functions as educational opportunity for students with disadvantaged family background and thus helps these students to overcome the gap of educational starting chances. Hence, the access to and effects of shadow education are calculated for different social strata allowing for comparisons between enrolment opportunities and benefits of shadow education for students with advantaged and disadvantaged family background. However, as several colleagues have stressed, neither logit (lnOR) or probit coefficients nor the oft-displayed odds ratios (OR) are suited for comparison (e.g., Allison 1999). In fact, “LnOR and OR reflect effects of the independent variables as well as the size of the unobserved heterogeneity” (Mood 2010: 73), wherefore it is not only invalid to make comparisons between samples, but between groups within samples, models (Allison 1999), and of course, comparisons across time (Mood 2010). Such comparisons are often misleading, as they assume “that the unobserved heterogeneity is the same across the compared samples, groups, or points in time” (Mood 2010: 73). Thus, by comparing lnOR or OR between groups or samples, misinterpretations of one’s calculations including a possible over- or underestimation of certain effects is likely to occur. To solve the problem of comparing coefficients or OR between groups, Williams (2009) proposed to generally apply a more advantageous class of models, so-called heterogeneous choice models (also location-scale or heteroskedastic ordered models). These models allow to use ordinal in addition to binary-encoded dependent variables. Also, “sources of heterogeneity can be better modeled and controlled for, and insights can be gained into the effects of group characteristics on outcomes that would be missed by other methods” (Williams 2009: 531).

Following these suggestions, I use the statistical software STATA that allows the additional post-estimation of Average Marginal Effects (AME) following my regression analyses (Auspurg and Hinz 2011, Williams 2012). I am thus able to realize comparisons between models (e.g., with which probability a top high school or a low-ranking high school is chosen in reference to all other high schools), groups (e.g., students from advantaged versus disadvantaged social origin or with what likelihood a certain kind of shadow education is chosen in comparison to another), and whole samples over time (e.g., comparisons between the student cohorts of the 1997 and 2011 HHSS data surveys). In contrast to OR, which have been preferred by many social scientists for quite some time, AME show no chances for the occurrence of a certain phenomenon; they specify by how many percentage points the average probability of the represented group of one variable is different from the probability in the reference group (if used for binary logistic regressions) or from the probability of all other outcomes (if used in multinomial logistic regressions) (Auspurg and Hinz 2011: 69). For example, we are able to predict by how many percentage points the average probability that students’ participate in shadow education differs from the probability that students do not participate in shadow education. Focusing on social origin, we are then able to compare whether students from advantaged social backgrounds have a higher or lower average probability to participate in shadow education compared to students from disadvantaged social backgrounds. We are also able to predict whether such investments in shadow education have a higher impact on whether students attend a high, a mediocre, or a low-ranking high school or university based on their social origin.

4.5 Summary

This chapter’s aim was to introduce the purposes, specifics, and advantages of two datasets and specifics of the chosen methodical approaches. As discussed above and finally summarized in Fig. 4.13, the presented data provide us with the opportunity to make analyses concerning all dimensions outlined in the SEII Frame: Access, Effects, Continuity, and Change (see also Chap.  1). In the subsequent empirical chapters (Chaps.  5,  6,  7 and  8), the main questions related to shadow education and its implications for educational and social inequality are comprehensively analyzed based on the here introduced data. Whereas Chaps.  5 and  6 draw on the 2011 HHSS survey data to clarify access to and effects of shadow education in present Japan, Chaps.  7 and  8 scrutinize variations in access and effects according to continuity and change of the market. Therefore, in Chap.  7 data of the 1997 HHSS survey is used in addition to the 2011 HHSS data analyzing the persistence of this market from the 1990s onward. In Chap.  8, I draw on the 2013 JSTS data to analyze how the Japanese juku-industry remained successfully in business despite unfavorable circumstances. All carried-out analyses will display average marginal effects instead of reporting the conventional, sometimes misleading logit coefficients and odds ratios. Hence, reliable findings are produced with the intention of achieving comprehensive findings on the discussed issues concerning shadow education and its implications for social inequalities.
Fig. 4.13

Data used to analyze the four dimensions outlined in the Shadow-Education-Inequality-Impact (SEII) Frame

Footnotes

  1. 1.

    The SSM is the most prominent dataset in Japanese social sciences, conducted every 10 years since 1955.

  2. 2.

    Original title: Kōkōsei no shinro to seikatsu ni kansuru chōsa = “A survey concerning the school course and school life of high school students.”

  3. 3.

    Thanks to Professor Ojima, I was directly included in the data evaluation process in 2012 and 2013. For further information about this survey, see Ojima and Aramaki (2013).

  4. 4.

    Personal communication, November 2013.

  5. 5.

    Parents’ occupational status is encoded as (1) leading positions in companies with at least five employees, (2) regular full-time employees (such as sarariiman), (3) part-time employees, (4) self-employed with less than four employees, (5) helping out in the family business, (6) others, and (7) unemployed men or full-time housewives.

  6. 6.

    Parents’ education level is encoded as (1) university degree or higher, (2) junior college, (3) technical college, (4) high school diploma, and (5) middle school diploma (see also the following subsection).

  7. 7.

    The home possessions variable is a score variable consisting of the following 11 variables asking about students’ and their households’ possessions: Q20-1, own room (including a room shared with siblings); Q20-2, own passport; Q20-3, Blu-ray/DVD recorder; Q20-4, digital camera; Q20-5, LCD/Plasma television; Q20-6, air cleaning machine; Q20-7, own computer; Q20-8, own mobile telephone; Q20-9, dishwasher; Q20-10, piano; and Q20-11, water filter machine.

  8. 8.

    The definition of books in this survey excludes comic books, magazines, school textbooks, or reference books.

  9. 9.

    This variable is recoded into a variable with three values: 1 = no siblings, 2 = one sibling, and 3 = two or more siblings. These categories seem adequate, since only 7.5% of the students had more than two siblings in 2011.

References

  1. Allison, P. D. (1999). Comparing logit and probit coefficients across groups. Sociological Methods & Research, 28(2), 186–208.CrossRefGoogle Scholar
  2. Auspurg, K., & Hinz, T. (2011). Gruppenvergleiche bei Regressionen mit binären abhängigen Variablen – Probleme und Fehleinschätzungen am Beispiel von Bildungschancen im Kohortenverlauf. [Group comparisons in regressions with binary dependent variables - Problems and misjudgements using the example of educational opportunities in the cohort course]. Zeitschrift für Soziologie, 40(1), 62–73.CrossRefGoogle Scholar
  3. Baker, D. P., Akiba, M., LeTendre, G. K., & Wiseman, A. W. (2001). Worldwide shadow education: Outside-school learning, institutional quality of schooling, and cross-national mathematics achievement. Educational Evaluation and Policy Analysis, 23(1), 1–17.CrossRefGoogle Scholar
  4. Bourdieu, P. (1983). Ökonomisches Kapital, kulturelles Kapital, soziales Kapital [Economic Capital, Cultural Capital, Social Capital]. In R. Kreckel (Ed.), Soziale Ungleichheiten [Social inequalities] (pp. 185–199). Göttingen: Otto Schwartz Verlag.Google Scholar
  5. Bray, M. (2010). Researching shadow education: Methodological challenges and directions. Asia Pacific Educational Review, 11, 3–13.CrossRefGoogle Scholar
  6. Breen, R., Werfhorst, H. v. d., & Jæger, M. M. (2014). Deciding under doubt: A theory of risk aversion, time discounting preferences, and educational decision-making. European Sociological Review, 30(2), 258–270.CrossRefGoogle Scholar
  7. Bukodi, E., & Goldthorpe, J. H. (2013). Decomposing ‘Social Origins’: The effects of parents’ class, status, and education on the educational attainment of their children. European Sociological Review, 29(5), 1024–1039.CrossRefGoogle Scholar
  8. Creswell, J. W., Clark, V. L. P., Gutmann, M. L., & Hanson, W. E. (2003). Advanced mixed methods research designs. In A. Tashakkori & C. Teddlie (Eds.), Handbook of mixed methods in social and behavioral research (pp. 209–240). Thousand Oaks: Sage.Google Scholar
  9. Entrich, S. R. (2013). Causes and effects of investments in shadow education in Japan – A comparative point of view. In F. Ojima & S. Aramaki (Eds.), Gendai kōkōsei no shinro to seikatsu [Career and life of high school students today] (pp. 157–182). Kyoto: Doshisha Publishing.Google Scholar
  10. Entrich, S. R. (2014a). Effects of investments in out-of-school education in Germany and Japan. Contemporary Japan, 26(1), 71–102.Google Scholar
  11. Entrich, S. R. (2014b). German and Japanese education in the shadow – Do out-of-school lessons really contribute to class reproduction? IAFOR Journal of Education, 2(2), 17–53.CrossRefGoogle Scholar
  12. Erikson, R., & Jonsson, J. O. (1996). Explaining class inequality in education – The Swedish test case. In R. Erikson & J. O. Jonsson (Eds.), Can education be equalized? The Swedish case in comparative perspective (pp. 1–63). Boulder: Westview Press.Google Scholar
  13. Fairbrother, G. P. (2014). Quantitative and qualitative approaches to comparative education. In M. Bray, B. Adamson, & M. Mason (Eds.), Comparative education research. Approaches and methods (pp. 71–93). Hong Kong: Comparative Education Research Center (CERC), Springer.Google Scholar
  14. Fujihara, S. (2011). Breen ando Goldthorpe no aitaiteki risuku kaihi kasetsu no kenshō: Chichioya no kodomo ni taisuru shokugyō kyōiku kitai o mochiita keiryō bunseki [Empirical test of breen and goldthorpe’s relative risk aversion hypothesis in the Japanese society: Data analysis of a father’s occupational and educational expectations for his child]. Shakaigaku Hyōron, 62(1), 18–35.CrossRefGoogle Scholar
  15. Gürtler, L., & Huber, G. L. (2012). Triangulation. Vergleiche und Schlussfolgerungen auf der Ebene der Datenanalyse [Triangulation. Comparisons and conclusions at the data analysis level]. In M. Gläser-Zikuda, T. Seidel, C. Rohlfs, A. Gröschner, & S. Ziegelbauer (Eds.), Mixed Methods in der empirischen Bildungsforschung [Mixed methods in empirical educational research] (pp. 37–50). Waxmann: Münster.Google Scholar
  16. Hamano, T. (2011). The globalization of student assessments and its impact on educational policy. Japan Academic Society for Educational Policy Proceedings, 13, 1–11.Google Scholar
  17. Jick, T. D. (1979). Mixing qualitative and quantitative methods – Triangulation in action. Administrative Science Quarterly, 24, 602–611.CrossRefGoogle Scholar
  18. Jonsson, J. O., & Erikson, R. (2000). Understanding educational inequality: The Swedish experience. L’Année Sociologique, 50, 345–382.Google Scholar
  19. Kikkawa, T. (2006). Gakureki to Kakusa-Fubyōdō: Seijukusuru Nihongata Gakureki Shakai [Education and social inequality: Contemporary educational credentialism in Japan]. Tokyo: Tōkyō Daigaku Shuppankai.Google Scholar
  20. Komiyama, H. (1993). Gakurekishakai to juku - datsu juken kyōsō no susume [Credentialist society and Juku –Stop the entrance examination competition!]. Tokyo: Shinhyōron.Google Scholar
  21. Komiyama, H. (2000). Juku: Gakkō surimuka jidai wo maeni [Juku: Before the era of downsizing schools]. Tokyo: Iwanami Shoten.Google Scholar
  22. Schultz Lee, K. (2010). Parental educational investments and aspirations in Japan. Journal of Family Issues, 31(12), 1579–1603.CrossRefGoogle Scholar
  23. Matsuoka, R. (2015). School socioeconomic compositional effect on shadow education participation: Evidence from Japan. British Journal of Sociology of Education, 36(2), 270–290.CrossRefGoogle Scholar
  24. MEXT, Ministry of Education, Culture, Sports, Science and Technology. (2012). Kōdai setsuzoku tokubetsu bukai (dai 4-kai) haifu shiryō - shiryō 1: AO nyūshi-tō no jisshi jōkyō ni tsuite. [Special group concerned with the connection between high school and university education (4th Meeting). Distribution material – Material 1: Concerning the implementation of the AO entrance]. Retrieved from: http://www.mext.go.jp/b_menu/shingi/chukyo/chukyo12/shiryo/1329266.htm
  25. MEXT, Ministry of Education, Culture, Sports, Science and Technology. (2014). Kodomo no gakushūhi chōsa [Survey concerning children’s learning expenses]. Retrieved from: http://www.mext.go.jp/b_menu/toukei/chousa03/gakushuuhi/kekka/1268105.htm
  26. MEXT, Ministry of Education, Culture, Sports, Science and Technology. (2017). Statistics. Retrieved from: http://www.mext.go.jp/english/statistics/
  27. Mood, C. (2010). Logistic regression. Why we cannot do what we think we can do, and what we can do about it. European Sociological Review, 26(1), 67–82.CrossRefGoogle Scholar
  28. Ojima, F. (Ed.). (2001). Gendai kōkōsei no keiryō shakai-gaku: shinro – seikatsu – sedai. Tokyo: MINERVA.Google Scholar
  29. Ojima, F., & Aramaki, S. (Eds.). (2013). Gendai kōkōsei no shinro to seikatsu. Kyoto: Doshisha University Press.Google Scholar
  30. Ojima, F., & von Below, S. (2010). Family background, school system and academic achievement in Germany and in Japan. In J. Dronkers (Ed.), Quality and inequality of education (pp. 275–297). Dordrecht: Springer.CrossRefGoogle Scholar
  31. Park, H. (2013). Re-evaluating education in Japan and Korea. Demystifying stereotypes. London/New York: Routledge.Google Scholar
  32. Roesgaard, M. H. (2006). Japanese education and the cram school business: Functions, challenges and perspectives of the Juku. Copenhagen: NIAS Press.Google Scholar
  33. Rohlen, T. P. (1980). The Juku phenomenon: An exploratory essay. Journal of Japanese Studies, 6(2), 207–242.CrossRefGoogle Scholar
  34. Rohlen, T. P. (1983). Japan’s high schools. Berkeley/Los Angeles: University of California Press.Google Scholar
  35. Seiyama, K., & Noguchi, Y. (1984). Kōkō shingaku ni okeru gakkōgai kyōiku tōshi no kōka [The effects of outside of school educational investments at the transition to high school]. Kyōiku Shakaigaku Kenkyū, 39, 113–126.Google Scholar
  36. Shirakawa, T. (2013). Gendai no kōtōgakkō ni okeru shinro bunka no jōken – Chii tassei shikō to jiko jitsugen shikō [Conditions for course differentiation in today’s high schools]. In F. Ojima & S. Aramaki (Eds.), Gendai kōkōsei no shinro to seikatsu [Career and life of high school students today] (pp. 33–47). Kyoto: Doshisha University Press.Google Scholar
  37. Southgate, D. E. (2009). Determinants of shadow education – A cross-national analysis. Ohio: Ohio State University.Google Scholar
  38. Southgate, D. E. (2013). Family capital: A determinant of supplementary eduaction in 17 Nations. In J. Aurini, S. Davies, & J. Dierkes (Eds.), Out of the shadows: The global intensification of supplementary education (pp. 245–258). Bingley: Emerald Publishing.CrossRefGoogle Scholar
  39. Stevenson, D. L., & Baker, D. P. (1992). Shadow education and allocation in formal schooling: Transition to university in Japan. American Journal of Sociology, 97, 1639–1657.CrossRefGoogle Scholar
  40. Teddlie, C., & Tashakkori, A. (2006). A general typology of research designs featuring mixed methods. Research in the Schools, 13(1), 12–28.Google Scholar
  41. Tomura, A., Nishimaru, R., & Oda, T. (2011). Kyōiku tōshi no kitei yōin to kōka: Gakkōgai kyōiku to shiritsu chūgaku shingaku o chūshin ni [Determinants and effects of investments in education: focusing on out-of-school education and the transition to private middle schools]. In Y. Satō & F. Ojima (Eds.), Gendai no kaisō shakai: Kakusa to tayōsei [Modern class society: Disparity and diversity] (pp. 267–280). Tokyo: University of Tokyo Press.Google Scholar
  42. Williams, R. (2009). Using heterogeneous choice models to compare logit and probit coefficients across groups. Sociological Methods & Research, 37(4), 531–559.CrossRefGoogle Scholar
  43. Williams, R. (2012). Using the margins command to estimate and interpret adjusted predictions and marginal effects. The Stata Journal, 12(2), 308–331.Google Scholar

Copyright information

© Springer International Publishing AG 2018

Authors and Affiliations

  1. 1.Department for Education, Social Science Educational ResearchUniversity of PotsdamPotsdamGermany

Personalised recommendations