Towards a broader look at inequalities in education: a response to the comments

© Educational Research Institute knowledge on inequalities in education. The comments by Michał Sitek are not limited to the discussion of my proposals, but they form a separate article with a clearly outlined hypothesis that the author consistently tries to prove, referring to numerous studies. The aim of my article was to convey a much more modest message. I was mostly interested in the fact that the current state of research does not allow us to conclude that after the school reform in Poland in 1999, inequality in education decreased. I tested three hypotheses, the first of which concerned the impact of social background on students’ performance, the second – the relationship between social origin and the choice between different types of upper secondary schools, and the third on the consequences of the gradual differentiation of lower secondary schools introduced by the reform. In all three cases, available data did not allow me to reject the hypothesis that no changes occurred in educational inequalities during the 15 years of the functioning of new schools in Poland. In the article I made no attempt to explain why this is the case. Referring to research, I tried to convince the reader that inequalities Ignorance is bliss

knowledge on inequalities in education. The comments by Michał Sitek are not limited to the discussion of my proposals, but they form a separate article with a clearly outlined hypothesis that the author consistently tries to prove, referring to numerous studies.
The aim of my article was to convey a much more modest message. I was mostly interested in the fact that the current state of research does not allow us to conclude that after the school reform in Poland in 1999, inequality in education decreased. I tested three hypotheses, the first of which concerned the impact of social background on students' performance, the second -the relationship between social origin and the choice between different types of upper secondary schools, and the third on the consequences of the gradual differentiation of lower secondary schools introduced by the reform. In all three cases, available data did not allow me to reject the hypothesis that no changes occurred in educational inequalities during the 15 years of the functioning of new schools in Poland.
In the article I made no attempt to explain why this is the case. Referring to research, I tried to convince the reader that inequalities

Ignorance is bliss
O ne cannot remain indifferent to the problem of inequality in education. And not just because so much attention in public policy is devoted to the issues of equal opportunities. A more important reason is an elementary sense of justice, deeply rooted in human nature. It makes us sensitive to different kinds of exclusion, such as barriers to learning opportunities in the case of children from disadvantaged backgrounds. However, writing about inequality is a bit risky. It interferes with the picture of the world in which we tend to believe in order to feel comfortable.
Interrupting the feeling of bliss, however, also has its positive side. It supports discourse, which may result in a better understanding of inequality in education. That is how I received the comments of Michał Sitek to my article "Education reform and inequality: fifteen years of new lower secondary schools in Poland" (Edukacja, 144(5), 2017,). The author of the polemic considers many issues which I have not considered, conducting supplementary analyses, many having scientific originality and enriching in education appear to be stable also in other countries, despite reforms aimed at ensuring equal opportunities for young people from different social backgrounds. At the end, I formulated a conclusion that sources of educational inequalities should be sought outside school -in social stratification. This conclusion is not new, but I am convinced, that this is still promising way to understand the nature of inequality in education.

Michał
Sitek presents a more pragmatic approach. Probably due to this reason, he is less concerned with what is happening in other countries and focuses his attention on the changes taking place in Polish schools. Referring to the PISA results, he notes that during the 15 years that have elapsed since the reform, the results of Polish students have improved, especially in the group of the weakest students. I understand Michał Sitek's position as follows: even if some indicators of inequality did not change, as I tried to demonstrate in my article, the inequalities from before the reform "stopped being" inequalities, as the chances of all groups of students increased. However, this does not preclude the fact -which the author of the polemic does not deny -that the competition may move to a higher level of students' performance and continue to be strong. The same occurs in the labour market. When the number of people with the highest qualifications actually increases, employers and corporations begin to use protectionist practices, increasing their requirements relating to diplomas, thereby forcing candidates to stay in school for a longer period of time (Collins, 1979). The surprisingly good results of students with disadvantaged social backgrounds may not be of much worth in confrontation with the achievements of those whose parents will be able to assist in further stages of their educational career.
Another issue that differs my position from that of my adversary is the degree of "penetration" into the mechanisms that create inequalities in education. Michał Sitek believes that by limiting the arguments to correlations, I simplified the picture of inequality in Polish schools and he suggests, therefore, to expand this picture by taking into account students' gender, social status, or differences in school resources. Then, using statistical analysis, he demonstrates that such a fragmentation of data leads not to one, but multiple images of inequality, not always consistent with my hypothesis, which states that inequalities have remained stable. The results presented by Michał Sitek in Table A2 serve as an example: between 2009 and 2012, there was a sharp decline in the impact of parental status on the results in reading (with the control of other student characteristics and the differences in school performance), or the results presented in Figure 2, where the impact of origin on the choice of upper secondary school has been decomposed into direct effect, depending on parental education, and the effect related to student performance, which also indirectly depends on parental status. Such results, many of which were presented for the first time, clearly allow for a deeper insight into inequality in Polish education than the simple correlations. Parental status determines the chances for a good education slightly differently in the case of girls and boys, and also creates different opportunities in different types of schools. These are certainly valuable findings, but they provide a fragmented picture of inequality in Polish schools.

Simplicity also has advantages
In my analyses I focused on a single index of inequality, namely the correlation coefficient between a student's background and educational achievements. Such simplicity has its advantages. First of all, it provides a synthetic view of inequality. Helps answer the fundamental question of whether inequalities have decreased, remained stable or increased. It is the starting point in assessing the effectiveness of education reforms, where one of the objectives is to reduce inequalities -as in the case of the reform in Poland. Although correlations seem too abstract for education policy, they exhibit many similarities with other indicators that allow an evaluation of the effectiveness of interventions in favour of creating more equal access to education (Sawiński, 2011). The calculation of the correlation coefficient is the first step that needs to be taken in order to evaluate the effectiveness of the reforms. Usually, further steps include a more detailed analysis of the mechanisms of inequality, such as those proposed by Michał Sitek.
The benefits of such a "two-step" procedure are illustrated by studies of the effects of the school reform carried out in the Netherlands in 1968, the objective of which was, among other things, to reduce the impact of parental status on the choice of secondary school. At the beginning, numerous studies focused on the main question: did the reform help to reduce inequalities in Dutch schools?
The conclusions of such analyses indicated that this was not the case. It was therefore assumed that inequalities did not change and the attention shifted to explain why this was so (Tieben and Wolbers, 2010).

Mechanisms that create inequalities
My adversary rightly points out that in order to explain causal relationships, we should take a closer look at school practice and the mechanisms by which student performance is determined by social origin. He mentions, among other things, issues relating to entrusting school management to local governments or establishing a more rational school network. I fully agree that future analyses should go in this direction.
Relatively little was written about the practical difficulties of introducing new lower secondary schools in Poland. Only recently was an article published, which thoroughly analysed the implementation of the postulate of establishing well-equipped schools with professional staff in place of the numerous poorly performing primary schools existing before the reform (Herczyński and Sobotka, 2017). As it turned out, this crucial postulate of the reform has not been met and rural areas are still dominated by a model in which classes of lower secondary schools are located in primary schools. In such schools, the same teachers teach the same students, so it is hard to state that the reform brought a new quality to Polish education. In-depth interviews show that local government officials found it difficult to convince rural communities about the need to build new schools, to which students from various villages would commute. And the main reason was not the lack of money (as is usually the case), but local antagonisms. Residents were willing to agree to build a new school in their village, but protested when it was to be built elsewhere. In many regions of Poland, protests by local communities effectively inhibited the introduction of new schools. But even in places where this was accomplished, the antagonisms were observed among students who were "better" -due to the fact that it was "their school" as it was located in their village -and "worse" because they had to commute. The first group appeared to be more integrated and more likely to participate in extracurricular activities. Children belonging to the second group were usually isolated during school breaks, and after finishing classes they immediately went to wait for the bus that would take them where they felt safe. Large and well-equipped lower secondary schools were to provide equal opportunities through integration and equal treatment. As it turned out, they reinforced existing divisions (Sobotka, 2016).

The role of central examinations
The focus on the reforms carried out in Poland probably did not allow Michał Sitek to consider the proposals of authors who look at issues of inequality in education from a cross-country perspective. There are many voices in this domain, and the author of the polemic quotes them willingly, adding that I have not paid enough attention to these issues in my article. However, treating these results rather briefly does not allow for a deeper penetration of numerous controversies, of which, as I suppose, the authors of the quoted papers are aware. Therefore, I would like to refer to two findings addressed by Michał Sitek, which are not as unequivocally defined by the authors as it would appear from the short overview of their work.
The first issue concerns the impact of central examinations on equality in education. Michał Sitek stressed that the introduction of central examinations was one of the key changes that accompanied Polish school reform. Then he refers to the article by Thijs Bol and others (Bol, Witschge, Van de Werfhorst and Dronkers, 2014), saying that central examinations could reduce educational inequality in Poland.
As it turns out, however, the article does not claim that the introduction of central examinations leads to the weakening of educational inequalities. On the contrary: the authors clearly point out that the results of the few studies conducted so far do not resolve this issue, as this effect was determined only in some countries, while in others, the opposite effect was found (i.e. that central examinations lead to an increase in inequality) or no connection between these phenomena was established. In discussing their results, the authors write: The cross-level interaction between central examinations and socio-economic status is significantly positive, indicating that in countries with central examinations, there is greater inequality in student performance across different socio-economic backgrounds.
[...] However, we do not attach too much significance to this finding, because in the later models [...] we show that the positive interaction between social class and central examinations is not robust once the school level is included" (Bol et al., 2014(Bol et al., , p. 1560.
It is worth noting that in this paper, central examinations are understood differently than in Poland. It is enough, when standard program requirements are defined in a uniform manner for the entire country and exam results are recognised by schools of upper levels and by the job market (Bol et al., 2014(Bol et al., , p. 1556. According to this definition, the central examinations do not have to be taken in the form of a test or be anonymous. For example, Denmark, where secondary school finishes with an open exam taken in front of a board consisting of a teacher who knows the student and his/her abilities well, was included among the countries that have central examinations. In contrast, the United States, where each year of schooling finishes with an anonymous exam in the form of a test, was included in the group of countries without central examinations. The reason is that tests in the United States are selected at the state level, which does not harmonise the requirements throughout the country. According to the criteria adopted in the article by Thijs Bol's team, Poland, before the introduction of external examinations in 2002, would have also been included in the group of countries with central examinations, because the topics of school leaving exams were determined for all schools by the Ministry of Education, and the results of end-of--school exams were honoured in the enrolment process to schools at the next level. I devoted much attention to this article, as it illustrates the situation in which the results of research do not provide a clear picture of the mechanisms creating inequalities in education. I would also like to believe what Michał Sitek seems to -that uniform examinations will improve the chances of children from less favourable backgrounds, "by drawing the attention of teachers and school principals to the achievements of the weakest students" (Sitek, 2017, p. 175). Although I agree with his argument, we cannot forget that more affluent parents will not remain passive. They have enough funds to pay for extracurricular activities in order to prepare (sometimes even "train") their children to achieve better results in achievement tests. This may mean that the external examinations introduced by the Polish reform are creating a new mechanism of inequality (Sawiński, forthcoming).

And when the inequalities decrease
Looking at the papers cited by Michał Sitek I would like to focus on one more (Breen, Luijkx, Müller and Pollak, 2009). When I was writing my article, I had doubts if I should quote this piece of work or omit it. On one hand, the article by Breen et al. is a milestone in the discussion on educational inequalities that has been going on for over half a century, but on the other hand, it focuses on new issues which require significant attention. In the original version of my article, I included it, but later, due to the comments from reviewers, I decided to focus on the main theses, which did not leave much space for a discussion of other issues. Michał Sitek noted, however, that I acted inelegantly, ignoring the voices saying that "the relationship between social origin and broadly understood educational achievement had weakened" (2017, p. 165). In response to this controversy, I would like to discuss the arguments of why the idea of the permanent nature of educational inequalities is being questioned. Richard Breen, Ruud Luijkx, Walter Müller and Reinhard Pollak (2009) took the effort to explain the peculiar paradox that despite the expansion of education and improvement of living conditions in most countries, studies have consistently shown no signs of a decrease in educational inequalities. The authors of the cited paper formulated the thesis that this was the result of too small research samples, which according to the rules of statistical inference, did not allow the hypothesis of no change in the level of inequality to be rejected. In order to achieve higher sample sizes, they combined the results of various studies in nine European countries. For example, in Hungary there were four studies, in the UK 15, in Germany 30, and in the Netherlands 35. Then the data was harmonised by extracting the same categories of social origin (based on the father's occupation), as well as the same levels of educational attainment, which allowed the results to be compared across different periods and different countries. The data harmonisation and relatively large total sample sizes allowed the authors to draw a statistically valid conclusion that in eight out of the nine countries, educational inequalities were slightly reduced. The ninth country, Hungary, was excluded from the analysis. This was due to the fact that Hungarian studies do not provide a coherent picture of the changes occurring in educational inequality 1 .
The authors achieved their goal, as they managed to show that in the case of a sufficiently long period (the analysis included people born in the years 1908-1964), and sufficiently large samples, the data will reveal a decrease 1 Although it is difficult to argue with the decision of the authors to exclude Hungary, as it is the result of applying uniform criteria for assessing the data quality in all countries, the fact that no effort was made to determine which of the four Hungarian studies was a source of problems seems surprising. Perhaps after rejecting one or two studies raising the most doubts, the criteria would have been met. As I mentioned in my article, in the context of the inequalities in education, Hungary is an important country because of the relatively late integration of schools of different levels into one system, the collectivisation of agriculture, which could have had an impact on the educational opportunities of peasant children, and also because of the far-reaching intervention of the communist regime in the school selection processes. After excluding Hungary, Poland became the only country in the project from Central and Eastern Europe, which limited the possibility of drawing conclusions about the mechanisms of inequality in countries with different political systems. in educational inequality. On this basis, will we be able to claim that sooner or later current educational inequalities in Poland -including those observed in lower secondary schools -will decrease? I think such a conclusion would be too hasty, as Breen, along with other authors, does not guarantee this. In their conclusion, the authors admitted that because the article is of an empirical-descriptive-methodological nature, it says little about the essence of educational inequalities (Breen et al., 2009(Breen et al., , p. 1515. Therefore, we learned that inequalities are decreasing in the long time perspective -but this did not change much in terms of educational practice.

Do actions speak louder than words?
In the 1990s, when the dominant conviction was that the expansion of education, in a logical and necessary way, would improve the chances of students from disadvantaged backgrounds, the book "Persistent inequality ... " by Yossi Shavit and Hans-Peter Blossfeld (1993) elicited shock and forced researchers to reflect. However today, many researchers believe that talking about the ineffectiveness of the reforms is unproductive. Education needs information: what needs to be done, and general reflections do not provide any specifics. Public policy -as they say -is based on evidence, so it is expected that the actions will bring results.
The introduction of many reforms -not only in education -is accompanied by expectations of their effectiveness. I would like my article to be a kind of cold shower for the education reformers who prefer such philosophy. More research and critical discussion is required, which focus on fundamental issues, not practical ones. As long as we do not know enough about inequalities in education, anticipating the effects of educational interventions comes down to pure guesswork.