CC BY-NC-ND
67
Patterns of School Choice: Two Districts in St. Petersburg
D. Alexandrov, K. Tenisheva, S. Savelyeva
Received in Daniel Alexandrov
January 2018 Candidate of Sciences in Biology, Head of the Laboratory of Sociology in Education and Science, National Research University Higher School of Economics. Address: 19 Pechatnikov St, 190008 St. Petersburg, Russian Federation. Email: alexandrov@hse.ru Ksenia Tenisheva
Senior Lecturer, National Research University Higher School of Economics. Address: 55/2 Sedova St, 190008 St. Petersburg, Russian Federation. Email: tenishewa.soc@gmail.com Svetlana Savelyeva Deputy Head of the Laboratory of Sociology in Education and Science, National Research University Higher School of Economics. Address: 55/2 Sedova St, 190008 St. Petersburg, Russian Federation. Email: ssavelieva@hse.ru
Abstract. Parental choice of primary school is analyzed using the example of local education systems in two districts of Saint Petersburg. The empirical basis of the research is provided by the results of a parent survey conducted in 34 schools (1,055 respondents). The following data is sorted and compared successively: whether parents make education-
al choices at all, whether they consider alternative options, what school characteristics they believe to be important, what sources of information they use, and what actions they take. The study explores how characteristics of choice are related to parental education and socioeconomic status as well as to the probability of selecting a school of a specific status.
Insight is provided not only into how the desire of parents to analyze all possible school options and sources of information correlates with their educational and socioeconomic backgrounds in general but also how parental choice is affected by neighborhood structural characteristics (school diversity, proportion of higher-status schools). Districts with broader structural opportunities and larger middle classes feature a variety of choice strategies, which is not observed in districts with limited structural opportunities even if they are socioeconomically heterogeneous.
Keywords: school differentiation, parental choice of primary school, neighborhood context, structural opportunities.
DOI: 10.17323/1814-9545-2018-4-199-229
The article is focused on the patterns of parental choice of primary school in different structural contexts within a megalopolis. Original empirical evidence from two districts is used to show that parents with similar educational backgrounds pursue different educational choice strategies depending on local structural contexts. insisted from Research on the reproduction of social inequality through educa-Russian by tion conducted in Germany, Great Britain and France shows that po-I. Zhuchkova. larization of educational institutions results from choice differences
across social classes [Broccolichi, van Zanten 2000; Ball 1993; Ball et al. 2002; Kristen 2003]. Even being entitled to exercise the freedom of choice, parents tend to make educational decisions that match their social class or socioeconomic status. More advantaged parents send their children to more effective schools to boost their chances for success. As a result, social advantage is reproduced through education: it is not only the level of education selected by students and their families but also the quality of educational institutions, starting from the first grade or even preschool, that is critical.
The role of school in social inequality reproduction became particularly visible in the 1980s, when a number of countries began to reform their education systems and integrate market mechanisms. Institutional changes affected school choice, which became free for parents. With all the diversity of national educational contexts, the major consequences of introducing free school choice were the same for most countries. Instead of improving education quality, as had been expected, competition resulted in school differentiation. Market rules benefit schools that are already better off and make it worse for those attended by working-class children [Lauder, Hughes 1999; Reay, Ball 1997].
In a situation like that, school choice becomes the "middle-class strategy" [Ball 1993]. Access to educational services is unequal in favor of the middle class. What is more, middle-class parents regulate the education market according to their needs and goals [Ball, Bowe, Gewirtz 1996; Ball, 2003]. However, such opportunities are fraught with risk and may require heavier investments than ever before, as market conditions decrease the probability of replicating the social status of parents. Although the middle class controls the education market, the market as such is so open and disorganized that it disturbs the order and long-term planning which middle-class parents want so badly [Ball et al. 2002].
Vague choice criteria are an important feature of modern education markets. Which kindergarten or school is better? There are no evaluation tools that could be applied universally, so middle-class parents spend a lot of time choosing the "right" school. Not infrequently, they pay attention to student performance or ethnic composition [Hastings, Kane, Staiger 2005; Saporito, Lareau 1999], but they do not have a clear idea of what exactly should be selected and what criteria should be considered. Uncertainty is giving rise to a new type of moral panic around schooling and school choice [Ball at al. 2002], which is especially typical of large cities with their saturated education markets, fierce competition and high levels of school diversity and accessibility.
Middle-class parents living in large cities think in such a strategic way as to preventively move to better neighborhoods with the same socioeconomic status as theirs, which is called moving to opportunity [Leventhal, Brooks-Gunn 2003]. In this case, parents are not involved
in the school choice process as they made this choice in advance by choosing where to live [Gabay-Egozi 2015]. Some researchers, understanding how such concentration of the middle class in neighborhoods contributed to inequality reproduction, suggest breaking the link between the school market and the housing market [Benson, Bridge, Wilson 2015].
Working-class parents are involved in school choice much less actively than middle-class families. In a number of cases it happens due to the lack of information and cultural and social resources in the family to make optimal decisions. Disadvantaged (migrant or low-income) families often refuse to make a choice, being convinced that all public schools are the same and simply sending their children to the nearest one [Broccolichi, van Zanten 2000; Kristen 2003]. However, even if they do choose an institution, they tend to be guided by the criteria that are of little interest to middle-class parents, attaching most importance to school proximity and road safety [Warrington 2005]. As a result, working-class children turn out to be disadvantaged even though their parents enjoy the freedom of school choice [Ball 1993; Reay, Ball 1997].
It is not only parents' socioeconomic status but also their race and religious beliefs that affect school choice. The most important thing, however, is that socioeconomic inequality at the level of neighborhoods results in choice being determined by local contexts, namely the ethnic and sociodemographic composition of individual neighborhoods and differences in the organization of local education systems. Population density plays an important role in shaping local education markets: as the population decreases, more places become available at schools, while as neighborhoods get more populated, schools become overcrowded, hence less accessible. The distribution of students among schools depends on the availability and ratio of schools of different statuses and their location. Based on these characteristics, researchers identify types of competitive environments in local education markets [Lubienski, Gulosino, Weitzel 2009; Taylor 2001].
Until very recently, the physical and social contexts of school choice had hardly been addressed by researchers [Lubienski, Gulosino, Weitzel 2009]. However, modern studies are increasingly built around analyzing local educational markets or even micromarkets and microsystems. At this level of analysis, it is easier to understand context-conditioned processes, including the differentiation of schools across neighborhoods [Taylor 2001].
Russian research findings also reveal class differences in school choice strategies. Better-educated and wealthier parents attach more importance to teacher competence than school proximity [Sobkin, Ivanova, Skobeltsina 2011; Roshchina 2013]. Low-educated parents tend to follow the "package investment strategy", delegating concerns about their children's academic achievement to the education system, while highly educated parents make "targeted investments"
at every stage of education, deciding what classes or activities will contribute the most to their children's future success1. However, there have been no comparative studies that would explore the choice of school by Russian parents with due regard for their structural opportunities.
The Changing Rules of School Choice
"What country of opportunities are you talking about??? I don't understand why not allow everyone to choose the schools they want? Very few parents struggle for strong schools today, very few! So why not allow overloaded schools to run admission tests??? Let them enroll students who are capable and willing to learn! Not everyone can afford a gymnasium or a lyceum! And failure in such tests does not mean infringement of anyone's rights! Decisions should be made by schools only! And if we have to queue overnight, we'll do it right by the school building because it's closer and easier to understand what's going on in there."2
The relevance and social significance of the problem of school choice today are comparable to those of introducing the Unified State Exam (USE) several years ago. However, while the USE integration involved a universal set of rules and procedures applicable to the whole country, the rules on admission to schools still vary year after year in some regions.
The 1990s brought complete freedom of choice to education, enabling parents to choose from a variety of learning formats and schools of different statuses offering different sets of educational services [Cherednichenko 1999]. That was when education markets began to emerge. Meanwhile, schools were inheriting their statuses, as well as reputations, from the Soviet era, using them as the main signals to translate to the market.
The education market grew and evolved for two decades, followed by critical national policy decisions on the rules of school choice designed to improve the opportunities for children from low-resource
1 http://ria.ru/ratings analytics/20120514/647531719.html
2 http://www.shkola-spb.ru/
The problem of school choice is very acute nowadays. Parents who consciously build their children's educational strategies are ready to spend nights camping in queues to get their children enrolled in the school of choice. Such situations can be observed from time to time in various cities.
Address-based school enrollment frustrates highly educated parents who want quality education for their children:
families. The Federal Law No. 273 On Education in the Russian Federation of December 29, 2012 introduced the notion of "school zone".
The law allowed every federal subject to interpret the term in their own way. Whole administrative districts were declared school zones in St. Petersburg in 20133, so every family had the right to choose any school within their administrative district in 2014. However, this practice only lasted for one year. As early as 2014, some schools had already begun to be officially linked to specific addresses, the residents of which were the only ones entitled to attend those schools. An experiment carried out in St. Petersburg revealed the heart of the problem of school choice: while some parents advocate for free school choice, all of them want to have the right to attend the school nearest to their home.
Linking specific schools to restricted neighborhoods aims at reducing the segregation of schools: the larger the territory within which parents are allowed to make free choice, the higher the inequality among educational institutions, which inevitably results in reduced educational chances for lower-class groups. However, when free choice is limited by school zones, it harms parents who want their children to attend higher-status schools but have none near their home. The decision of the government of St. Petersburg to link schools to residential districts is an attempt to come to a compromise and balance the interests of all market participants.
The issue is especially acute in megalopolises that have more advanced and differentiated education systems, i. e. more educational institutions, higher school density and status diversity.
Data Collection In order to analyze the process of parental school choice, two sur-and Sampling veys of primary school students' parents were conducted, one in the schools of Vasileostrovsky District in 2013 and the other in the schools of the left-bank part of Nevsky District in 2015. Schools were sampled randomly. The resulting sample included 581 parents in 21 schools (of the total of 30) in Vasileostrovsky District and 474 parents in 13 schools (of the total of 19) in the left-bank part of Nevsky District. The surveys were carried out by students and researchers of the National Research University Higher School of Economics, who did short interviews with parents outside the school, recording their answers in questionnaires4.
3 St. Petersburg Law No. 461-83 On Education in Saint Petersburg of July 17, 2013.
4 Data was collected and analyzed within the framework of the projects supported by the HSE Program for Basic Research in 2013-2015. The analytical part of the study was also supported by the Russian Humanities Research Foundation (Project No. 16-03-00802 "Differentiation of Schools and Educational Choice: Schools and Parents", 2016-2018).
The questionnaire tool offered both open- and closed-ended questions on various aspects of school choice: how many options were considered, how much time it took to make a decision, how exactly the choice was made, how different schools were compared, what the choice criteria were, which sources of information were used, etc. Next, parents were asked whether they considered changing school before the middle grades as well as about their plans concerning middle and high school education and their college ambitions. A separate module of questions was devoted to sociodemographic characteristics, namely parental education and parental socio-occupational status (SOS)5.
Characteristics of Differentiation in local education markets in general and in school the District Cases choice in particular is largely affected by the historical, socioeconomic, geographic and residential contexts of specific neighborhoods.
The two districts selected for analysis represent contrasting cases in terms of their socioeconomic contexts, spatial accessibility and location of schools in them. Vasileostrovsky District (VD), which occupies the territory of Vasilyevsky Island (it also includes two smaller islands) is characterized by the high population density and spatial accessibility of virtually all schools. The island has an area of only 10.9 km2, stretching up to 4.2 km north to south and up to 6.6 km west to east. The district has good transportation and is part of the city's historic center. For instance, the eastern part of the island is home to such sightseeing attractions as the Spit of Vasilyevsky Island, the Kunstkamera Museum, The Twelve Collegia edifice headquartering St. Petersburg State University, and others. At the same time, the district is isolated from the rest of the city, being connected to the mainland by bridges, which are raised nightly to allow the passage of sea vessels along the Neva River. These characteristics make VD a unique locality. Some of its spatial characteristics are displayed on the map, which also shows the location of all the schools in the district (Fig. 1).
Nevsky District (ND) differs significantly from VD in the historical, geographic and demographic contexts. Despite being rather stretched out (20 km from north to south, 8 km from west to east) and having an impressive area of 61.79 km2, it is sparsely populated. Residential housing occupies only 1.6 percent of the territory. The district is split in two by the Neva River, over which transportation is rather difficult: only three bridges connect the two parts of ND, turning them into isolated "ecosystems" (Fig. 2).
5 ISEI-08 (International Socio-Economic Index of Occupational Status) was used as an indicator of socio-occupational status. The index is based on the detailed international classification of occupations ISC0-08 and shows the social prestige of various occupations and the relevant levels of education required.
Fig. 1. Vasileostrovsky District Schools Plotted on the Map
(dark circles indicate schools covered by the surveys).
The left-bank part of Nevsky District involved in the study is very stretched along the Neva River with residential neighborhoods alternating with industrial parks, which creates structural limitations for school choice. The district is bordered by the river on one side and by the Nevsky Overpass (Sortirovochny Bridge) on the other. Only two main roads run along the Neva River connecting the district with the city center (Alexander Nevsky Square) and Kolpinsky District, and they are congested most of the time. Transport accessibility of the district and transportation within it thus cannot be considered satisfactory.
There are similarities as well as differences between the geographical contexts of the two districts. Both VD and the left-bank part of ND are fairly isolated from other districts by topographical barriers, which hamper student mobility between schools of neighboring districts. VD, however, is more compact, and nearly all families have more than one school available within walking distance. The left-bank part of ND is very stretched and divided into sectors, which may affect student mobility within the district.
The socioeconomic contexts of the districts are inextricably associated with the processes of their historical and real estate development. VD was among the first districts in the city to be involved in real estate development. It has a lot of housing which was built before the first third of the 20th century and which is now dilapidated or used as communal apartments. At the same time, the district has quarters developed in the 1960s-1970s as well as luxury infill apartment buildings. Such a diversity of residential housing renders the district attractive to all social classes.
Nevsky District was developed during the Soviet period and constructed by the proletariat for the proletariat. It still has housing with-
Fig. 2. Nevsky District Schools Plotted on the Map
(dark circles indicate schools covered by the surveys).
out shower facilities—it was suggested that residents would go to public saunas. Although the population changed over years, in fact the composition stayed the same, consisting mainly of the working class. Numerous industrial zones virtually turn the residential quarters into isolated "pockets" with a small number of schools in each. As a result, parents have to choose from those few institutions near their homes.
Differences in the cost of housing (purchase and rent) help to see the gap in the socio-occupational standings of residents between the two districts: the average price per square meter is 103,027 rubles in VD, as compared to 77,094 rubles in ND; average monthly rent for a one-bedroom apartment is 35,574 rubles in VD, as compared to 25,028 rubles in ND6.
The contextual differences described above influence the structure of local education markets, the VD market being more differentiated and the ND market being more homogeneous, as shown below.
6 Estimated using the 2017 statistics from the Byulleten Nedvizhimosti (Real Estate Bulletin) website.
Figure 3. The Number of Schools (Adjusted for Population) and the Proportion of Higher-Status Schools in the City Districts.
Vasileostrovsky Nevsky
Admiralteysky
Vyborgsky
Kalininsky
Kirovsky
Kolpinsky
Krasnogvardeysky
Krasnoselsky
Kronshtadtsky
Kurortny
Moskovsky
Petrogradsky
Petrodvortsovy
Primorsky
Pushkinsky
Frunzensky
Tsentralny
Number of Schools per 10,000 District Population
1.34
1.06
1.07
0.92
1.31
1.12
1.05 1.05 1.12
177
1.58
1.02
1.40
1.09
0.94 1.09 1.08
Proportion of Higher-Status Schools
1.91
0.64
0.39
0.31
0.48
0.39
0.52
0.24
0.59
0.33
0.25
0.34
0.53
0.27
0.42
0.33
0.45
0.60
0
Characteristics of Most city districts have approximately one school per 10,000 of the the Local Educa- population, the highest school per capita rates being observed in tion Markets Tsentralny and Admiralteysky Districts. VD has a 30 percent higher rate than ND, and it also features the highest proportion of higher-status schools (lyceums, gymnasiums, specialized schools) in the city-64 percent (Fig. 37), as compared to only 39 percent in ND, which is close to the average city rate.
VD and ND differ in accessibility of educational institutions, their characteristics and, as a consequence, the aspects of school choice. Assumedly, choices in ND may be limited due to the specific spatial organization of the district, while VD parents will be more concerned about school choice—not because of the high limitations but because they have choice opportunities.
The contexts described and the differences in local education markets between the two districts frame the basic conditions of parental school choice. The important factors affecting it include population density, the number of educational institutions in the district, and
7 The data was provided to the Laboratory of Sociology in Education and Science by St. Petersburg Center for Assessing the Quality of Education and Information Technology.
Table 1. Socio-occupational status and College Education of VD and ND Parents: Mean Values with Confidence Intervals.*
VD ND
Average maternal SOS 51.3 (±15.7) 49.9 (±13.4)
Average paternal SOS 51.4 (±14.2) 50.4 (±13.6)
Percentage of college-educated mothers** 63.8 (60.3-67.8) 59.1 (54.2-63.7)
* The international index ISEI-08 used to assess socio-occupational status takes values from 10 to 90, the lowest values corresponding to unprestigious, low-paying unskilled labor jobs (e. g. cleaning lady), and the highest ones to prestigious, high-paying occupations that require a college degree (e. g. surgeon or lawyer). In this sample, SOS values range from 10 to 79. ** Confidence intervals for college education were calculated by bootstrapping.
their characteristics. School choice contexts differ greatly at the level of city districts, which coincide with "school zones".
School Choice in There are no established social classes in Russia today that would St. Petersburg use their advantages or lose within the education market. Rather, we are talking about socio-occupational status and different educational backgrounds. At least, these parameters can be measured and included into the analysis model.
Despite the differences in socioeconomic development between the city districts described above, the data collected indicates that parental socio-occupational status and educational background of parents do not differ statistically significantly between VD and ND (Ta-ble 1).
However, the districts differ considerably in their socioeconomic composition, as evidenced by the paternal and maternal SOS density functions (Fig. 4, 5).
ND has a population of heterogeneous composition. Both pictures show two peaks, particularly prominent in the paternal SOS density function graph (Fig. 5). The first one indicates highly-qualified managers and professional engineers, and the second one denotes specialists, i. e. vocational teachers, junior managers in the construction and welfare sectors. The proportion of the lower middle class is very small, which results in a gap in the middle of the distribution.
Vasileostrovsky District has a high-SOS stratum of both fathers and mothers—specialists such as engineers, geophysicists, physicians, dentists, judges, etc.—which is barely distinguishable in ND.
The majority of parents in both districts—67.3 percent in VD and 71.9 percent in ND—are firmly convinced that their children will obtain a college degree in the future. In VD, 10.2 percent of parents intend to move their children to another school later, as compared to 2.4 per-
Figure 3. The Number of Schools (Adjusted for Population) and the Proportion of Higher-Status Schools in the City Districts.
Vasileostrovsky Nevsky
Admiralteysky
Vyborgsky
Kalininsky
Kirovsky
Kolpinsky
Krasnogvardeysky
Krasnoselsky
Kronshtadtsky
Kurortny
Moskovsky
Petrogradsky
Petrodvortsovy
Primorsky
Pushkinsky
Frunzensky
Tsentralny
Number of Schools per 10,000 District Population
1.34
1.06
1.07
0.92
1.31
1.12
1.05 1.05 1.12
177
1.58
1.02
1.40
1.09
0.94 1.09 1.08
Proportion of Higher-Status Schools
1.91
0.64
0.39
0.31
0.48
0.39
0.52
0.24
0.59
0.33
0.25
0.34
0.53
0.27
0.42
0.33
0.45
0.60
0
cent in ND. College aspirations and school changing intentions are positively related with mother's higher education.
Only 42 percent of VD families did not consider other school options apart from the school attended by their children. The proportion is considerably higher in ND (59%). In most cases, parents chose from two schools only (29% in both districts).
The response "Only one" (no other school options were considered) cannot be interpreted unambiguously. As seen from the interviews8, different motivations and family backgrounds may be behind it. Parents could have chosen that specific school because they or their friends had attended it, so they consider it trustworthy enough and do not have to engage in the complicated process of school selection. Alternatively, such a response may indicate that other nearby schools were so bad that they were not even considered as options. From this point on, families that did not make a choice are treated as a homogeneous group in being compared to those who made a choice, but it
8 After completing the formalized questionnaires, some of the parents (N=40) gave extended interviews on school choice. In particular, they were asked to specify how they interpreted certain items of the formalized questionnaire.
Figure 6. The Distribution of Parents' Responses to How Many School Options They Considered.
Vasileostrovsky District Nevsky District
College-educated No college degree
should be kept in mind that this category includes rather diverse families with different educational decisions.
Figure 6 shows the distribution of responses on the number of school options considered in VD and ND, adjusted for mother's education. In both districts, parents without college degrees were more likely to consider only one school option, while college-educated parents chose from two or more institutions.
Children of 42 percent of ND families attend schools other than the nearest one. This question was not asked in VD, but it can be safely assumed that the proportion of such families is higher, since the journey to school takes on average 1.5 minutes longer in VD than in ND.
School Characteristics Perceived as Important by Parents
Parents consider a number of factors when choosing a school, balancing their choice criteria with their family's needs and opportunities. The respondents were asked about the school characteristics that had mattered the most to them. In VD, parents were supposed to rate all of the characteristics proposed, while ND parents were asked to select and rank only three. Seven response options were proposed: "school status (gymnasium, lyceum, specialized school)", "availability of extracurricular activities", "high USE performance", "ethnic composition", "cultural background of classmates", "neat and well-equipped facilities" and "security guards and student safety"; they could also select "other" and provide an answer of their own. The first survey in VD offered two more options, "proximity to home" and "good teachers and administrators", which were later excluded as they were selected by nearly all parents and thus did not allow for discriminating among different categories. The resulting response distribution is shown in Figure 7. Response differences between the districts are reflected in Table 2.
Figure 7. The Distribution of Responses to Which School Characteristics Were Considered When Making a Choice.
Vasileostrovsky District
Good teachers
Proximity to home
Cultural background of classmates
School status
Security guards and student safety
Neat and well-equipped facilities
Ethnic composition
High USE performance
Availability of extracurricular activities
Other
58.0
60.5 62.0 64.5 63.5 62.5 63.5 64.0
Nevsky District
0 0
66.0 0 0
63.5
57.0
59.5
58.4
59.0
56.5
62.6
60.4
College-educated No college degree
Table 2. Differences in School Characteristics Perceived as Important by Parents (%).
Total ND VD
School status 35.3 36.7 33.3
Extracurricular activities 25.4 19.8 32.7
USE performance 20.3 12.7 30.2
Ethnic composition 10 5.3 16.2
Cultural background 17.6 15.8 20
Facilities 28.9 25.2 33.5
Safety 21.2 13.8 30.8
School status, neat and well-equipped facilities and availability of extracurricular activities were found to be the most significant criteria of parental school choice. Cultural background of classmates is also considered as quite important. Ethnic composition is a much greater concern for parents in ND than in VD. Items on school proximity and the quality of teachers and administrators were only offered in the VD survey, and the overwhelming majority of parents marked these characteristics as important. Below, we zero in on each of the school characteristics as well as the categories of parents who attach the most importance to them.
School status (gymnasium, lyceum, specialized school). School status orientation is strongly related to parental education—this criterion was selected more often by college-educated mothers in both districts. For fathers, however, this relationship is only observed in VD. School status plays an important role for parents who chose the school other than the nearest one to their home. VD parents who paid attention to school status consider their choice to be final and have no intention of moving their children to another institution.
Some parents obviously understood "school status" in a way that was different from what was implied by the questionnaire. Despite the bracketed explanation that formal status (gymnasium, lyceum, specialized school) was meant, they assessed the informal status of schools, i. e. their prestige and popularity in the local community. That is why a good proportion of parents who marked school status as an important criterion had actually sent their children to regular schools.
High USE performance. In both districts, this school characteristic was selected as important by parents who considered two or more school options. It is considered equally significant by mothers of all educational backgrounds. In VD, this parameter was selected more often by college-educated fathers than by fathers with no college degree. School effectiveness is valued more by those VD parents who have college aspirations for their children.
Availability of extracurricular activities. This characteristic is equally important to parents who chose from two or more school options as well as those who did not consider any alternatives. The gap between parents who attach importance to this criterion and those who do not is only observed in ND, where extracurricular activities are valued more by high-school-educated mothers and fathers, while college-educated mothers are not likely to consider this parameter significant. In VD, attaching importance to extracurricular activities is related to parents' college aspirations: those who find extracurricular participation important are more likely to expect their children to obtain a college education (less likely to opt for "Unlikely"). Only parents whose children attended regular schools had been interested in ex-tracurriculars, while those who sent their kids to gymnasiums and lyceums had barely taken this factor into account.
Ethnic composition. This criterion is equally important to mothers of all educational backgrounds. VD fathers' attitudes vary across educational levels, ethnic composition being more likely to be considered an important school characteristic in families where fathers had vocational education. In VD, this factor also plays a greater role for parents who consider the possibility of changing school. This finding is based on the responses of parents from a number of schools, and there are no signs that their children attended schools with high proportions of ethnic minority students.
Cultural background of classmates. This school characteristic is valuable most of all for better-educated parents with fairly high ambitions. In VD, it was selected by college-educated fathers and parents who had considered two or more school options before making a choice. In ND, cultural background of classmates is significant for college-educated parents (significance level=0.10). ND parents who attach importance to the level of classmates' cultural development expect their children to obtain higher education.
Neat and well-equipped facilities, security guards and student safety. The survey results do not allow for identifying the specific categories of parents considering or not considering these two characteristics. In ND, neat and well-equipped facilities play a more significant role for parents who have no intentions of changing school. In VD, the proportion of parents with college aspirations was higher among those who valued school facilities than among those who attached no importance to the factor. In VD, student safety is more important to parents who chose from a few school options than to those who only considered one.
The items on "school proximity" and "good teachers and administrators" were only proposed during the first survey conducted in VD. School-home distance is an important factor for parents who did not consider alternative options. Interestingly enough, college aspirations are found more often in parents attaching importance to school location. Good teachers and administrators represent a characteristic that is more likely to be considered by parents who intend to provide their children with a college education; it is rarely taken into account by parents with high school and vocational education. In addition, it is valued by parents who were involved in the school-choice process and disregarded by those who were not.
Parents in ND were asked to rank three school characteristics that they had considered when making school choice by their importance. School status was the first choice of most respondents (24.7%), followed by high USE performance (15.6%). Other factors were much less likely to be selected as the first choice. The second most important factor was availability of extracurricular activities for most parents, followed by neat and well-equipped facilities, security guards and student safety, and high USE performance. As for the third most impor-
Figure 8. Differences in Choosing the Top Priority School Characteristic (First Choice) between College- and Lower-Educated Parents (%).
Other
33.3
Security guards and student safety
Neat and well-equipped facilities
Cultural background of classmates Ethnic composition
8.5
5.7
2.5
Availability of extracurricular activities
School status
High USE performance
College-educated parents No college degree
tant school characteristic, the most popular answers were school facilities, followed by student safety and extracurricular activities.
College-educated mothers are significantly more likely to rank school status as their top priority criterion, their second choice being high USE performance, while non-college-educated parents value availability of extracurricular activities and student safety the most. No difference in the frequency of selecting specific school characteristics as the third choice was found among parents with different educational backgrounds (Fig. 8).
Some patterns can be traced in choice combinations, too. Parents who valued school status the most were very likely to rank high USE performance or availability of extracurricular activities as the second most important factor. Those who attached the most importance to extracurriculars were also concerned about the cultural background of classmates or ethnic composition. Ethnic composition was also the next thing considered by parents whose primary choice criterion was school effectiveness. Cultural background of classmates is related to other characteristics of school effectiveness and composition, while neither neat and well-equipped facilities nor security guards and student safety are related to any other aspect of school effectiveness.
Parents were allowed to specify important school choice criteria of their own. All in all, 134 "other" responses were provided, Figure 9 showing the distribution of the most popular ones. The most common response was school proximity to home or (in some rare cases)
to a parent's workplace. Parents concerned about school proximity were significantly more likely to choose the school nearest to their homes. The second most common response was teachers. This category includes requesting a particular teacher, mentions of teaching quality, teaching staff competence and "teachers' behavior towards students". This is a pretty powerful factor, as parents who searched for good teachers were significantly less likely to choose the school nearest their home.
Some parents reported having chosen a specific school because it was attended by their older children. First of all, such parents are familiar and apparently satisfied with the school personnel and environment. Second, enrollment priority is given to students whose siblings already attend the school even if the family lives in another school zone. Some parents chose the school that they themselves had attended or that was attended by their friends' children. In this case, the school feels familiar to parents, and they tend to perceive their choice as more informed, even though it may have been over a decade since their graduation. It is probably for the same reason that some parents choose the school where someone whom they know works, hoping for some guarantee of a comfortable learning environment for their children.
Some parents reported having chosen a specific school because a particular subject was taught at a good level there, or because its students demonstrated a high level of knowledge, or because it offered unique learning programs. Foreign languages (English, Chinese) and mathematics were mentioned most often among the particular subjects that parents wanted to be taught at a good level. Parents who specified such school characteristics had made informed choices, caring about specific criteria, collecting information on various schools' offers and searching for the most suitable option. The quality of knowledge, learning programs or teaching of a particular subject was mentioned as a school characteristic significantly more often by college-educated parents. Those who paid attention to such characteristics were significantly less likely to choose the school nearest their home.
Parents described some other characteristics as well, such as teacher tolerance, school discipline ("no smoking in the school building or toilets"; "order, discipline"; "everything is negotiable"; "a Stalinism-hardened principal"). Yet, such responses are unique and thus not shown in Figure 9.
School proximity and requests for a specific teacher, not school, were reported most often among "other" reasons. Parents also mentioned friends' recommendations, education program and knowledge quality, and class composition. However, such responses were rather unpopular.
Figure 9. The Distribution of Responses on the Important Characteristics Specified in the "Other" Category (Proportion of parents who selected the "Other" option).
The school was attended by parents Class composition Education program Quality of teaching Subjects available
The school is attended by older siblings or friends' children
Specific teacher School proximity
36.02
8.07 | 4.97 | 4.97 | 3.73 3.11
27.95
Figure 10. The Distribution of Responses to "What Activities Did You Involve in before Making Your Choice?"
Attended school open days
Analyzed school websites
Browsed online forums
Considered advice from friends
None of the above
Vasileostrovsky District
70.5 76.3 73.7
66.5
Nevsky District
71.5 69.5 74.5
47.5 52.5
College-educated No college degree
Sources of Infor- The respondents were asked a series of questions about what they mation had done before deciding on the school for their children. It was found out that 36.9 percent of ND parents and 33.3 percent of VD parents had taken no action to choose the school.
Overall, college-educated parents actively use different sources of information about schools. They are more likely to consider advice of other parents, analyze online forums and school websites and attend school open days than lower-educated parents (Fig. 10).
In VD, parents who only considered one school option did not collect any information on the institution—they did not consider advice from their friends and family, or analyze online forums and school websites, or attend school open days—as compared to those who
considered two or more school options. No such difference was revealed in ND.
Parents with college aspirations for their children tend to search for information more actively, striving to make informed choices by considering advice from their friends and family and analyzing online forums.
In VD, 57.5 percent of the parents who considered two or more school options sent their children to higher-status schools, as compared to 55.3 percent of the parents who made no choice (the difference is insignificant). Meanwhile, the corresponding percentages in ND are 28.6 and 18.9 percent, indicating a significant gap. Therefore, school status is considered an important factor by parents making their school choices.
In ND, higher-status schools were the nearest options for 46.3 percent of the parents who did not consider any alternatives and 30.2 percent of the parents who compared different schools, while regular schools were the nearest ones for 67.5 and 59.3 percent, respectively.
Both districts feature significant differences between mothers who did not make a choice but sent their children to higher-status schools and those who were engaged in the choice-making process, on the one hand, and mothers who did not make a choice and sent their children to regular schools, on the other, the latter category being characterized by a comparatively low socio-occupational status.
In both districts, parents who considered two or more options and chose higher-status schools were guided by formal school effectiveness characteristics, namely USE performance and status. In addition, VD parents also took the cultural background of classmates into consideration. Virtually the same criteria (school status in both districts and USE performance in VD) were important to parents who sent their children to higher-status schools without considering alternative options. However, they also took some action before making their final decision: VD parents analyzed online forums and ND parents considered advice from friends and family. Parents who chose regular schools from two or more options attached the most importance to neat and well-equipped facilities (this difference is observed in VD). Meanwhile, those who sent their children to regular schools without considering alternative options valued ethnic composition most of all, neglecting good teachers and school proximity (in VD).
Results and Logistic regression analysis allows for comparing the influence of pa-Discussion rental SOS and education on school selection patterns. Two models were constructed, one for each of the two districts, the question as to whether parents considered alternative options (choice) being used as a dependent variable (Table 3).
It is not only the mother's education but also her socio-occupa-tional status that making a school choice is related to in both districts.
Table 3. Logistic Regression Results, Dependent Variable: Making a School Choice.
ND VD
Odds Ratio CI P Odds Ratio ! CI ! p
Intercept 0.37 0.16-0.86 0.022 0.87 ; 0.45-1.69 j 0.687
Maternal SOS 1.00 0.99-1.02 0.612 1.00 Í 0.99-1.01 i 0.917
Mother's education 1.96 1.18-3.29 0.010 1.93 Í 1.21-3.10 i 0.006
N 358 428
AIC 483.712 576.426
Table 4. Logistic Regression Results, Dependent Variable: School Status.
ND VD
Odds Ratio CI p Odds Ratio CI 1 p
Intercept 0.23 0.09-0.59 0.003 0.30 j 0.15-0.59 j <0.001
Maternal SOS 1.00 0.98-1.02 0.673 1.02 i 1.01-1.04 i 0.003
Mother's education 2.54 1.39-4.79 0.003 1.96 i 1.22-3.15 i 0.005
N 358 428
AIC 405.459 554.616
College-educated mothers are more likely to consider two or more options before making a decision.
Two more models for each district were constructed with the status of the selected school (regular or lyceum/gymnasium/specialized) being used as a dependent variable (Table 4). School status is also related to mother's education in both districts: having a college-educated mother boosts the child's chances of attending a higher-status school. VD findings indicate the important role of maternal SOS as well, which is positively related to the probability of selecting a higher-status school (Fig. 11). The graphs also show that SOS has more weight than mother's education in VD. The probability of choosing a higher-status school is 16.4 percent higher among college-educated mothers. Every additional score on the SOS scale increases this probability by 5 percent on average, which results in a 31% gap between the highest- and lowest-SOS parents.
Obviously, the differences described above are explained by different structural opportunities of the two districts. Vasileostrovsky District compares favorably with Nevsky District by school diversity and
Figure 11. Predicted Probability of Selecting a Higher-Status School.
Vasileostrovsky District
i Maternal SOS
Nevsky District
i Maternal SOS
Table 5. Number of Schools of Different Statuses in VD and ND.
School Status VD Percentage ND Percentage
Regular school 9 34.60% 31 57.40%
Gymnasium 3 11.50% 4 7.40%
Gymnasium with enhanced education in foreign languages 1 3.80% 1 1.90%
Gymnasium with enhanced education in physics and mathematics 1 3.80% 1 1.90%
Lyceum 0 0.00% 3 5.60%
Artistic lyceum 1 3.80% 0.00%
School with enhanced education in foreign languages 7 26.80% 10 18.70%
School with enhanced education in mathematics 1 3.80% 1 1.90%
School with enhanced education in chemistry 1 3.80% 0.00%
School with enhanced education in the humanities 1 3.80% 0.00%
Educational center 1 3.80% 1 1.90%
Resource center 0.00% 1 1.90%
Center for culture and education 0.00% 1 1.90%
Figure 12. USE Performance of Figure 13. USE Performance of VD and ND
VD and ND Schools in Russian. Schools in Mathematics.
-VD ND
the proportion of higher-status schools (Table 5). Besides, it is home to two of the six prestigious selective admission schools.
School choice strategies are differentiated little in ND despite the residents being split into two social groups (see Fig. 4 and 5), whereas differentiation is obvious in VD with its extended structural opportunities. It would seem that two peaks in SOS distribution should divide schools into two groups, families of higher SOS creating demand for more prestigious schools, but nothing like this happens in ND. The driving force behind school choice thus seems to be structural opportunities combined with the presence of high-SOS residents. The availability of selective admission schools and the presence of upper-class families in VD create conditions for the so-called conspicuous consumption. As a result, social groups of the same SOS tend to choose prestigious schools in VD and regular ones in ND.
Involvement in school choice is more typical of college-educated parents in both districts. College-educated mothers are more likely to consider two or more school options and to be willing to change school before the middle grades (they are still few, however). They use various sources of information (friends, online resources, school visits) and regard school status as an attractive characteristic of the highest priority, being ready to sacrifice school proximity for a higher school status. USE performance is the second most important factor of school attractiveness for college-educated parents. Attaching importance to school effectiveness is part of long-term educational strategies, as such parents are convinced that their children will obtain college education in the future. Parental education is therefore the fundamental factor determining parents' choice behavior.
School choice strategies have specific characteristics in both dis-
tricts. VD parents are more likely to get involved in the choice-making process, consider more school options, choose more remote schools and use various sources of information. VD also has a higher proportion of parents intending to change school before the middle grades and a higher student mobility rate. ND parents tend to consider other parents' advice and go to school open days more often than VD parents, but the use of specific sources of information is not related to parental education in ND. A specific characteristic of ND is that availability of extracurricular activities is valued by non-college-educated parents.
The school choice strategy pattern shared by a number of countries, which was first illustrated by Steven J. Ball through the example of Great Britain [Ball 1993], is found in this study as well: involvement in school choice is much higher among parents from better educational and socioeconomic backgrounds. In addition, it transpires that the structural opportunities of specific neighborhoods affect the degree of choice stratification. In VD, which represents all social classes and has more high-SOS residents, the middle class can identify itself in the process of social comparison [Festinger 1954] and detach itself using academic choice strategies. This becomes possible due to an extended structure of opportunities, which affects not only school choice as such but also adjustment of choice strategies. In situations of limited supply, the strategy of selecting prestigious schools becomes irrelevant for the middle class, so no further stratification is observed. However, as soon as the menu of prestigious schools is extended, the choice strategy described by Ball becomes important not only to the upper middle class but to other layers of the middle class, too. Choice strategies are always determined by comparison horizon, both at the level of schools and at that of SOS categories of parents pursuing a specific strategy. This is the underlying logic behind conspicuous consumption, and the findings of this study are very much in line with it.
Researchers identify the following important effects of school choice: parents who engaged actively in the choice-making process tend to be more satisfied with the school selected [Bosetti 2004], their children being more academically successful and more school-oriented [Shumow, Vandell, Kang 1996]. In a broad sense, school choice strategies depend on family characteristics and contribute to the differentiation of schools and social inequality in general. School choice plays a particularly important role in stratified systems. High social significance of differences in school choice behavior across social classes dictates the need to consider those differences when developing education policies and school choice regulations [Whitty 2001]. In the context of present-day education policies and educational strategy research, it appears vital to switch from nationwide and regional-level samples to a more in-depth and localized analysis. Academic choice strategies in Russia, while being similar to those in Europe, are still contingent on local structural contexts.
References Ball S. J. (2003) Class Strategies and the Education Market: The Middle Classes and Social Advantage. London: Routledge.
Ball S. J. (1993) Education Markets, Choice and Social Class: The Market as a Class Strategy in the UK and the USA. British Journal of Sociology of Education, vol. 14, no 1, pp. 3-19.
Ball S. J., Bowe R., Gewirtz S. (1996) School Choice, Social Class and Distinction: The Realization of Social Advantage in Education. Journal of Education Policy, vol. 11, no 1, pp. 89-112.
Ball S. J., Bowe R., Gewirtz S. (1995) Circuits of Schooling: A Sociological Exploration of Parental Choice of School in Social Class Contexts. The Sociological Review, vol. 43, no 1, pp. 52-78.
Ball S. J., Davis J., David M., Reay D. (2002) "Classification" and "Judgement": Social Class and the "Cognitive Structures" of Choice of Higher Education // British Journal of Sociology of Education, vol. 23, no 1, pp. 51-72.
Benson M., Bridge G., Wilson D. (2015) School Choice in London and Paris— A Comparison of Middle-Class Strategies. Social Policy & Administration, vol. 49, no 1, pp. 24-43.
Bosetti L. (2004) Determinants of School Choice: Understanding How Parents Choose Elementary Schools in Alberta. Journal of Education Policy, vol. 19, no 4, pp. 387-405.
Broccolichi S., van Zanten A. V. (2000) School Competition and Pupil Flight in the Urban Periphery. Journal of Education Policy, vol. 15, no 1, pp. 51-60.
Cherednichenko G. (1999) Shkolnaya reforma 90-kh godov: novovvedeniya i sot-sialnaya selektsiya [The Schooling Reform of the 1990s: Innovations and Social Selection]. Sociological Journal, no 1/2, pp. 18-19.
Festinger L. (1954) A Theory of Social Comparison Processes. Human Relations, vol. 7, no 2, pp.117-140.
Gabay-Egozi L. (2016). School Choice in a Stratified Geography: Class, Geography, Otherness, and Moral Boundaries Education. Journal of Education Policy, vol. 31, no 1, pp. 1-27.
Hastings J. S., Kane T. J., Staiger D. O. (2005) Parental Preferences and School Competition: Evidence from a Public School Choice Program. NBER Working Paper No w11805.
Kristen C. (2003) School Choice and Ethnic School Segregation: Primary School Selection in Germany. Münster: Waxmann Verlag.
Lauder H., Hughes D. (eds) (1999) Trading in Futures: Why Markets in Education Don't Work. Buckingham: Open University Press.
Leventhal T., Brooks-Gunn J. (2003) Moving to Opportunity: An Experimental Study of Neighborhood Effects on Mental Health. American Journal of Public Health, vol. 93, no 9, pp. 1576-1582.
Lubienski C., Gulosino C., Weitzel P. (2009) School Choice and Competitive Incentives: Mapping the Distribution of Educational Opportunities across Local Education Markets. American Journal of Education, vol. 115, no 4, pp. 601-647.
Reay D., Ball S. J. (1997) Spoilt for Choice': The Working Classes and Educational Markets. Oxford Review of Education, vol. 23, no 1, pp. 89-101.
Roshchina Y. (2013) Strategii roditeley shkolnikov i vospitannikov DOU v oblas-ti obrazovaniya detey v2012 g. Monitoring ekonomiki obrazovaniya no 6 (69) [Educational Strategies of Parents of School and Preschool Children. Monitoring of Education Markets and Organizations no 6 (69)], Moscow: National Research University Higher School of Economics.
Saporito S., Lareau A. (1999) School Selection as a Process: The Multiple Dimensions of Race in Framing Educational Choic. Social Problems, vol. 46, no 3, pp. 418-439.
Shumow L., Vandell D., Kang K. (1996) School Choice, Family Characteristics, and Home-School Relations: Contributors to School Achievement? Journal of Educational Psychology, vol. 88, no 3, pp. 451-460.
Sobkin V., Ivanova A. I., Skobeltsina K. (2011) Ot vospitaniya v detskom sadu k shkol'nomu obrazovaniyu: mnenie roditeley [From Kindergarten Education to Secondary School Education: Parent View]. National Psychological Journal, no 1, pp. 111-118.
Taylor C. (2001) Hierarchies and Local Markets: The Geography of the Lived Market Place in Secondary Education Provision. Journal of Education Policy, vol. 16, no 3, pp. 197-214.
Van Zanten A. (2005) New Modes of Reproducing Social Inequality in Education: The Changing Role of Parents, Teachers, Schools and Educational Policies. European Educational Research Journal, vol. 4, no 3, pp. 155-169.
Vincent C., Ball S. J. (2006) Childcare, Choice and Class Practices: Middle Class Parents and Their Children. New York: Routledge.
Warrington M. (2005) Mirage in the Desert? Access to Educational Opportunities in an Area of Social Exclusion. Antipode, vol. 37, no 4, pp. 796-816.
Whitty G. (2001) Education, Social Class and Social Exclusion. Journal of Education Policy, vol. 16, no 4, pp. 287-295.
