Introduction
The signing of the first memorandum of agreement between the Greek government and the tripartite international decision group known as the “troika” (European Commission, 2010), signalized the end of the Greek metapolitefsi, the prosperous period that followed the restoration of democracy in 1974. The “memorandum” contained the first economic adjustment programme for this country. The global financial crisis, which made the memorandum necessary, hit Greece in an anthropological level. Resentment, desperation, and social unrest are permanent elements in the public discourse (Antonakakis and Collins, 2014; Drydakis, 2015; Giannakopoulos and Anagnostopoulos, 2016; Ifanti et al., 2013; Tsekeris et al., 2015). In this socio-economic context we explore the inequalities in access to higher education in Greece with a view to contributing to an international discussion (see Meyer et al. 2013; Mountford-Zimdars and Sabbagh 2013). The map of Greece is the canvas upon which patterns of percentages of access to higher education have been drawn.
The main reason for adopting a geographical approach in this study is that social geography has always played a significant role in the History, culture, and economy of Greece (Caraveli and Tsionas, 2012). Anthropologists like Dubish (1995) and Seremetakis (1991) have discussed the culture of some of these communities. Sociologists like Mouzelis (1978) have described how Greeks left their communities and moved to towns and how this move has shaped the character of modern Greece. Laying at the imaginary border between East and West, Greece is full of geographical antitheses. Eighty percent of its mainland is mountainous, something that has created the conditions for the formation of small and remote communities. Small towns in the mainland are the economical centers that provide the basic services to the surrounding rural areas. Seaside towns are spread across the country’s thirteen thousand kilometers coastline. In the seas that lay to the west, east and south, there are more than one thousand islands, many of which are secluded with only a few residents while others are vacation destinations for the upper social classes. More than half of the country’s population of 11 million lives in the two major urban centers.
The key theoretical idea in the current study is that inequalities in access to higher education are connected with differences in social capital across regions. Social capital is a broad notion in sociology. It has been connected to individuals’ overall positioning in the social space (Bourdieu, 2008), individuals’ resources for action (Coleman, 1988), the “networks, norms, and social trust that facilitate coordination and cooperation within societies” (Putnam, 2000), and the “information, trust, and norms of reciprocity inhering in one’s social networks” (Fucuyama 1997, 378). Social capital has a strong spatial element, as the Moving to Opportunity experiment in five large American cities has shown (see Chetty, Hendren, and Katz 2016). In our study, the spatial accumulations that constitute social capital are thought to be the neighbourhoods from which the examinees are eligible to attend their local schools.
The catchment area characteristics that we examine in our study, include the size and use of the houses, and the “economic inactivity” of the residents. The first two are proxy measures of neighbourhoods’ social capital and degree of urbanicity. The third is a measure of social relationships in the neighbourhoods among members of non-typical extended families. A fourth measure is the educational level of the residents. How do however these neighbourhood characteristics affect educational achievement? In his aptly titled paper “why it takes a village”, Ainsworth (2002) has listed a number of ways. He has included mechanisms like “collective socialisation”, “social control”, “social capital”, and “differential occupational opportunity”. Later in our study we do discuss how one of these mechanisms affects parents’ educational choices in a special mountainous community. The statistical verification of such relationships, however, is beyond the scope of the present study. Our research questions can be stated as follows: (a) how large are the geographical disparities in access to higher education in Greece and (b) which factors at catchment area level explain the variation in the examination scores?
Literature review
The neighbourhoods as intermediary levels in social progression
James Coleman (1966), the author of one of most cited reports in the field of educational inequality, theorised that social phenomena at the macro level progress through processes that take place at the micro level. In the Foundations of Sociological Theory, Coleman (1990) introduced a figure that explained his theory, known by its outline as “Coleman’s boat”. According to this theory, a macro level social phenomenon can set off particular actions between individuals at the micro sociological level. Individual behavior can in turn bring horizontal changes in the choices of other individuals at the micro level through complex social and psychological processes (the hull of the “boat”). These changes result in subsequent macro-level phenomena.
Our study is compatible with a slightly different sociological model. We propose that neighbourhoods mediate between the “macro” and the “micro” level. This idea has been expressed by American urban sociologists like Sampson’s (2012) who used a vast data-base from his two decades work in Chicago and described patterns of social inequalities among different areas of this city. Although he did not fully explain the mechanisms that created neighborhood effects, he showed that these effects were strong. Another American sociologist communicated the same idea twenty five years before Sampson. In The Truly Disadvantaged Wilson (1987) coined the terms “social isolation” and “accumulation of poverty” to describe the social networks in the neighbourhoods of the urban poor in great American cities.
Nonetheless, for all their research on the inner city areas, American urban sociologists did not explain the mechanisms of the neighbourhood effect. Our study suffers from the same limitation, due to the longtime shortage of educational statistics in Greece. Even if, however, detailed background information on examinees’ socioeconomic background was available for this Mediterranean country, researchers “would not be sure about the causal mechanisms which produce them [the neighbourhood effects], their relative importance in shaping an individual’s life chances, the circumstances or conditions under which the neighbourhood effects are most important”, van Ham and Manley (2012, 2787) claim. Similarly, Rees, Power, and Taylor (2007) have warned against the “simplistic conceptualizations” of correlations among geographic regions when social phenomena are studied. “There is an obvious need”, they write, “to avoid the accusation of ‘spatial fetishism’, where spatially-specific outcomes are delineated in ways which obscure the reality of the social processes which underpin them” (Rees et al. 2007, 269). The current study simply maps the patterns of geographical inequalities in access to higher education in Greece, without making any claim about causal relationships between spatial level variables.
The education inflation in Greece
Before examining the distribution of the scores in the university entrance examinations in Greece, we should give some background information on why a university degree is so sought after in this country. A university degree in Greece has always been associated with high social status (Fragoudaki, 1985; Tsoukalas, 1986), reflecting its value in terms of personal prestige, job security, and salary level (Gouvias 1998). “The expansion of the capitalist activities favored the rapid expansion of large comprador strata, composed of merchants, agents, lawyers, intermediaries, civil servants and speculators of all kinds”, writes Tsoukalas (1981, 116). Research on returns to education tells the same story. Tsakloglou and Antoninis (1999) found considerable differences in the living standards between university graduates and high school graduates in Greece, even when factors such as age, region, and employment are controlled for.
From a statistical point of view, around 64% of secondary education graduates in Greece succeed each year in entering tertiary education. The average percentage in the OECD countries is 60%. Tertiary education entry rates for Greece are 42% for Type A institutions (theory based studies) and 26% for Type B institutions (practical, technical or occupational skills). For the 19 countries of the Euro Area, the corresponding rates are 55% and 12% (OECD, 2011). Greece however sends more students on a per capita basis to universities abroad than any other country with similar economic growth (Antoninis and Tsakloglou, 2001). According to an OECD policy brief, Greece is ranked among the top countries in student migration with 55,074 Greek students study abroad (OECD 2004, 3). In addition to these students, some 25,000 more are awarded degrees from foreign universities without leaving the motherland. Half of the 32 private post-secondary non-tertiary education colleges that operate in Greece as “centers of further studies”, are franchised institutions of foreign universities. This means that through a number of economic strategies, known in the literature as being the driving forces behind the internationalisation of access to tertiary education (Altbach and Knight 2007), the Sixteenth Article of the Constitution of Greece that disallows the operation of non-state universities is circumvented.
Macro level policies affecting the distribution of the examination scores
The university entrance examinations in Greece are known as the “Pan-Hellenics”. They are carried out every year from May to June under the auspices of the Ministry of Education and are the only requirement for entering one of the country’s 20 universities and 14 technological institutions. Their structure has changed seven times in the last 50 years. The changes concern the number and the grouping of the examined subjects, the weight of the particular scores in the calculation of the final score, the weight of teachers’ grades, the duration in terms school years, the age of the examinees and, in many cases, all the above characteristics combined. Though national examinations in Greece have been accused of forcing students into a “distorted” competition, they have managed to retain high levels of social acceptance as being “objective” (Antoninis and Tsakloglou 2001, cited in Giamouridis and Bagley 2006).
Three partially implemented policies can illuminate key ideologies in Greece regarding the university entrance examinations: the inclusion of “social abilities” (Arsenis, 2016; Kassotakis, 2003), the standardisation of the examination scores, and the requirement of a minimum mark. The inclusion of “social abilities” was a plan to democratise access by taking into account candidates’ interests and talents. The standardisation of the examination scores through item banking was an effort to introduce accountability in education by comparing schools on the basis of examination results. The minimum required mark, commonly known in Greece as the “basis of 10” -the middle point of the 20-point grading scale- was an unsophisticated but populist threshold for entering higher education. The above-mentioned policies have had their equivalents in other educational systems. A personal statement of abilities and interests, for example, has been an integral part in the application process for entering British universities. In Italy, school comparisons based on examination results have taken the place of the regional progressive educational policies (Barzanò and Grimaldi, 2013). Threshold test scores for entering higher education can be found nowadays in countries like China and Japan. The “basis of 10” policy in Greece needs to be discussed, due to the effect that it has had on the distribution of the examination scores.
The “basis of 10” policy was introduced with the law on the “reform of higher education” (2007) and was abolished with the law on the “access to higher education” (2009). In Table 1 we can see how this policy affected the examination scores. The distributions of the examination scores are platykurtic and negatively skewed for a number of years. There is however a notable difference during the years that the “basis of 10” was in effect. In 2008, for example, a score equal to 10 was the 8th percentile. Only 5,477 participants scored below 10. Two years later, when the policy was not in effect, a score of 10 was the 25th percentile and the number of participants with a score under 10 quadrupled (20,868 examinees). In 2012 the middle point of the 20-grade scale was the 32th percentile (26,351 examinees). A possible explanation for the differences in the examination scores can be the leniency in the grading of the examination papers. The questions of the Pan-Hellenic examinations are graded through the traditional pen and paper way. In the years when the “basis of 10” was in effect, graders were possibly reluctant to leave students out of tertiary education. In other words, teachers themselves “corrected” what appeared at that time to be a rather arrogant educational policy.
The late blooming of sociological research in Greece and the economic crisis
Quantitative studies on educational inequality appeared in Greece very late in comparison to the other countries of the historical Western Europe and the United States. In the years before metapolitefsi, the period that started with the restoration of democracy in 1974, the discipline of sociology was considered “Marxist” in Greece, a characterisation that could end any academic career (Kokosalakis 1998). It was therefore only after 1974 that Greek sociologists were able to investigate the relation between socioeconomic status and access to higher education. In an early paper, Eliou (1976) used census data to map geographical disparities in the educational level of the Greek population. Μeimaris and Nikolakopoulos (1978), in what was essentially the first quantitative sociological study in education in Greece, used Correspondence Analysis to model undergraduate students’ social background and the characteristics of the university departments they attended.
During metapolitesfsi the social and educational inequalities decreased (Katsimi et al., 2014). This is partially reflected to the Gini coefficient for Greece, which decreased by 19% during that period (op. cit.). From a sociological perspective, every quantitative study in Greece during metapolitefsi confirmed the relation between socioeconomic status and educational achievement (see Sianou-Kyrgiou, 2006, 2008; Chrysakis, Balourdos, and Capella 2009; Gouvias 1998a, 1998b; Gouvias, Katsis, and Limakopoulou 2012). In her surveys and correlational studies (path analyses), Georgia Kontogiannopoulou-Polydorides (1985 – 1996) confirmed causal relationships between access to higher education and family factors.
The metapolitefsi miracle in Greece was based on Keynesian economics. Katsimi et al. (2014) use the phrase “soft budget constraints”. This is a term coined by János Kornai (1986) to describe the situation in which economic resources flow in a country with the prospect to be repaid through the growth of the local economy. The global financial crisis of 2008 however increased the risk for Greece to not be able to repay its loans (Ozturk and Sozdemir, 2015). This resulted to an increment in interest rates, economic uncertainty and finally a deep economic crisis in this country. One of the measures that successive Greek governments have taken in order to address the lack of economic resources for education is the merging of schools in neighbouring catchment areas. There is therefore an interesting spatial element in the study of educational choices, provisions, and outcomes.
Spatial disparities in educational choices, provisions, and outcomes
Spatial disparities in education provision are usually reported between countries or between large regions within countries (Ballas et al., 2012). Some studies however have actually focused on school catchment area level. In one of the firsts such works, Ainsworth (2002) built a number of regression models in which neighborhood characteristics predicted educational outcomes. Ainsworth (op. cit.) claimed that his study constituted “a step forward to both the urban and the educational literatures” (op. cit., 143). His paper was not influential enough so as to bring a “paradigm shift” in educational research as he had predicted, but a number similar works did appear in Urban Studies, the leading journal in the discipline of urban sociology. In a special issue of this journal (Butler and Hamnett, 2007), Gordon and Monastiriotis (2007) discussed the effect of the geographical context on English secondary school public exams and found that local educational policies affect school-level success rates. In another issue, Webber and Butler (2007) found also that the type of neighborhood in which pupil live is a reliable predictor of GCSE performance.
A number of similar studies can be reported here. Brattbakk and Wessel (2013) explored the influence of neighborhood on educational achievement in Oslo and found small but statistically significant effects. Johnston et al. (2006) compared the cultural segregation across local education authorities in England between school and neighborhoods and found that segregation is larger between schools than between neighbourhoods. In one of the first studies to consider the spatial dimension on educational outcomes, Bradford (1990) discussed the importance of residential environment on students’ attainment and calculated its effect. Kauppinen (2008) investigated how the neighborhood effect in Helsinki is associated with the socioeconomic composition of the school. Finally, Maloutas (2007) investigated educational inequality in different types of residential areas in Athens and discussed the middle class strategies of school choice.
Housing and economic inactivity in school catchment area
Apart from the educational level of the parents, two other important factors in explaining patterns of educational inequality in Greece can be “housing” (the size and the use of the house) and “economic inactivity”. The size of a house can be used as a proxy measure of social status (see Foley 2003). Emmanuel (2004) found positive correlation between dwelling size and the socioeconomic status of the households in the city of Athens. Maloutas and Ramos-Lobato (2015) discussed how housing in Athens and Dortmund (Germany) is connected with educational inequalities. According to Eurostat (2016) the mean dwelling size in Greece for all degrees of urbanicity is 88.6 m2, slightly over the mean dwelling size of 2001 for this country (Dol and Haffner, 2010, 51). According to the Hellenic Statistics Authority (2014), the mean Greek house usually consists of 3 rooms and the mean dwelling size per person is around 34,6 m2. According to the same source, the percentage of houses with size ≥100 m2 in Greece is 27.2%. There are 1,734,612 such houses in this country, out of the total 6,371,901 houses (27%). Second home ownership and the use of the dwellings as holiday homes or rentals, a widespread strategy in many countries, has been associated with the urbanity of the neighborhood (Paris, 2011).
Another factor in our study is “economic inactivity”. Economically inactive are people who are neither “employed” nor “unemployed”. In most cases they are grandparents, uncles, aunts, cousins and other relatives, who may not live in the same dwelling but often use a number of “housing strategies” to live nearby, as Tosi (1995) explains. The relatives in rural and urban areas alike create non-typical extended families and contribute to the raising of the children and the school-family relations. According to the Hellenic Statistical Authority (2016), around 40,69% of people in Greece over the age of 15 are economically inactive. The typical form of extended families in Greece are only 3.9% of the population of families (Iacovou and Skew, 2011). The percentage of the non-typical extended families that we just described, however, is unknown.
Inequalities across “Kallikratian” municipalities
Figure 1 depicts patterns of inequality in access to higher education in Greece. The map displays 325 administrative areas, known as “Kallikratian municipalities” i.e. the administrative regions that were created in Greece under the Kallikratis programme. Officially known in Greece as the “new architecture of self-governing entities and deconcentrated administration” (law 3852/2010), “Kallikratis” was seen as a step to combat political clientism at local level but also to transfer responsibilities for austerity policies at the same geographical level (Souliotis, 2013).
Each of the Kallikratean municipalities encompass cities, towns, and small settlements. Different shades of grey in Figure 1 represent different percentages of students who in the examinations of 2013 achieved a score greater than or equal to the third quintile of the distribution of scores (QU3 = 13.4). A score in the two upper fifths of this distribution can secure a place in a good university department. Lighter shades represent higher percentages. The colour white indicates lack of statistical information. The map suggests that more than half of the participants in some parts of major urban areas gain a place at prestigious institutions.
Darker shades of gray represent smaller percentages of scores in the two upper fifths of the distribution. In cartography a specific colour can “promote particular connoted narratives”, according to Fotiadis (2009, 33). A number of “black” regions are scattered on the periphery of our map. These are the mountainous parts of Xanthi at northeastern Greece, a number of mountainous municipalities at the western and northwestern parts of Greece, a number of mountainous regions at Peloponnese and Grete, and some small islands. More than 50% of the examinees in the “black” areas scored below the first quintile (QU1=7.5). The grades in the first fifth of the distribution are not sufficient for entrance even at the country’s less prestigious institutions. Apart from answering our first research question, Figure 1 conveys what Fotiadis (2009, 44) has defined as “cartographic homogenization impulse”. Municipalities are presented in our map as homogenous unified spaces. The unmapped mosaic of inequalities, however, should not disappear from our view and attention.
Regional policies at the villages of the historical Muslim community
The lowest scores in the Pan-Hellenic examinations of 2013 were reported in a number of villages on the Rodope mountain range at Thrace (northeastern Greece). These villages are known as “Pomakohoria”, a rather defamatory name deriving from the Bulgarian word “pomachamedanci” that means “Islamicised”. Most of the people in these villages represent what Tsitselikis (2012) has characterised as the “historical Muslim minority” of Greece. They are descendants of the Thracian Muslims (mostly Sunnis) who were exempt from the compulsory population exchange between Greece and Turkey in accordance with the Treaty of Lausanne (see Martin 2007). The community consists of three separate ethno-linguistic groups. A large Turkish group, which amounts to half of the community’s population, a smaller Pomak group (the indigenous people) that amounts to 30% of the population, and a Roma group that amounts for the remaining 20%. For the last 20 years, Turkish has practically become the linqua franca in the region (Lytra, 2007).
Apart from the language, the people at “Pomakohoria” share the same topography and the same living conditions. They are also engaged to similar livestock activities. Most importantly, however, they share the same educational opportunities. According to the Treaty of Lausanne, Muslim students in the region can opt to attend one of the 228 local primary minority schools and the 2 secondary minority schools in the city of Xanthi. The minority schools, however, implement both a Greek and a Turkish language curriculum. To these two languages, one has to add the foreign languages that are offered in Greek schools, and the mother tongue that is spoken at home. Plainly, many of the students of the minority schools are linguistically confused, according to the head of the local educational authority (Papadopoulos, 2013).
In a macro level policy, an additional 0.5% of the places in the Greek university departments is annually reserved for the Muslim who either live in Thrace or are of Thracian origin (Ziaka, 2009). Between 1996 and 2009, three thousand students have made use of this affirmative action (Zachos and Chava, 2015). Some members of the Muslim community, however, who identify themselves with the Turkish macro level policies (see Katsikas 2012)), resist the aforementioned policy and affect individual educational choices through local minority associations and unofficial social gatherings. The political implications from the clash of the official macro level policies with the local tactics are beyond the scope of the current study. Educational choices of minority families, however, are examples of how a social mechanism at the “meso” level comes between the policies of the macro level and the choices at the micro level. The isolating terrains of Rodope Mountains, the poverty that has been accumulated in the region, and, most importantly, the historical and social process that have formulated an identity different from the dominant Greek Orthodox one, are the building blocks of the assumed “meso” level.
Discussion
The current work has confirmed previous quantitative findings in Greece and has offered evidence for the existence of geographical patterns in access to higher education in this country. The mapping of educational inequalities can broaden the scope of educational research and strengthen evidence-based policies. As Larsen (2009, 138) writes, “a map does not just chart, it unlocks and formulates meaning, it forms bridges … between disparate ideas that we did not know were previously connected”.
Maps however are “narratives with a purpose, stories with an agenda”, according to Short (2003, 24) and the narrative of the current study is that access to higher education in Greece differs among upland, lowland and coastal areas, and also between the affluent urban neighborhoods and the inner city zones. Characteristics of these areas, like the size and the use of the houses and the existence of non-typical extended families, play a significant role in the construction of educational outcomes.
The importance of geographical approaches to educational phenomena can be the subject matter of an interesting academic discussion. The boundaries between a regional approach to educational inequalities and the broader sociological explanations can be vague. Human geography is a branch of the social sciences and comprises sub-branches, like “social geography”, “cultural geography”, and “children’s geography”. These sub-branches are neither far from each other nor very different from canonical sociological research. As a matter of fact, the development of human geography as a tool for understanding society has followed the development of the mainstream sociological thought (Graham, 1990; Nayak and Jeffrey, 2011). In recent years for example, human geography, similarly to sociology, has moved away from reliance on Marxism and has embraced critical and postmodern epistemologies (Best, 2009; Reynolds, 2003; Thomassen and Minca, 2002).
From what we have reported so far, we have come to the conclusion that future educational research could investigate regional and neighbourhood effects in tandem with the school effects. Neighbourhood studies, however, are in “crossroads”, according to van Ham and Manley (2012). The challenges that van Ham and Manley discuss in their paper (op. cit.) are also present in the current work. According to the authors (van Ham & Manley, 2012, 2787), the number and complexity of the explanatory variables in exploring the neighbourhood effect should increase. Longitudinal studies should address the intergenerational neighbourhood effects and also the neighbourhood selection bias. Neighbourhoods should be “operationalised”, not only in terms of their administrative borders but also in relation to the particular borders of each one of the explanatory variables (ibid). The very idea of studying educational outcomes from a spatial perspective needs to be evaluated. Our work has hopefully made a contribution to this direction.
2017-1-14-1484387338