The effect of family socioeconomic status on students' educational outcomes is well-documented in the literature, particularly since the release of the Coleman Report (Coleman et al., 1966). This phenomenon has been categorized in rational choice theory as primary effects when it influences students' performance, or secondary effects when it affects their educational choices (Boudon, 1973; Jackson, 2013). It is also referred to as social reproduction theory when scholars study students who remain trapped within the same social class as their parents due to their socioeconomic status (Bourdieu & Passeron, 1977; Rodriguez-Hernandez, 2020). In other words, researchers typically examine the extent to which socioeconomic and cultural status can explain educational outcomes (Broer et al., 2019).

However, while socioeconomic status is one of the variables with the greatest impact on educational outcomes (OECD, 2019), it does not fully account for all educational choices and performance. Some students perform better (or worse) than expected based on their socioeconomic status: these are the students we refer to as Deviant Cases (Boudon & Lazarsfeld, 1966).

Deviant Case Analysis (DCA) originated in empirical sociology as a tool to test various variables derived from the same underlying properties. It is particularly used in indicator correction but also finds use in the analysis of cases which defy statistical regularities. Our analysis focuses on the latter, as we take into consideration students who, despite their disadvantaged backgrounds, exceed expectations.

On this consideration our first two aims are: (1) to understand the determinants of being an upward deviant student and (2) to examine the factors that increase the likelihood of belonging to this category.

A previous study on the Italian case, where we analyzed PISA 2018 test performances net of socioeconomic status (Bonanni & Moreschini, 2024), revealed that certain ascriptive variables, individual choices, and structural factors are significant determinants of being or not being a deviant student. In addition to these variables, the role of students' educational and occupational expectations, as well as those of their parents, is crucial. Many other aspects of this social phenomenon remain to be explored: one of these is the school effect, specifically whether attending one school rather than another influences the likelihood of being an upward deviant case.

The literature includes many studies analyzing the effect of school on educational outcomes (Hanushek, 1997; 2016; Gamoran & Long, 2006; Woessmann, 2016; Barret et al., 2019; Agasisti et al., 2021), and building on these, our third aim (3) is to understand how much the school explains the phenomenon of deviant cases.

In summary, the aims of this study are threefold: (1) to identify some of the determinants of being an upward deviant case, (2) to determine the likelihood of being one, and (3) to examine how much of this phenomenon is due to the school that students attend.

This analysis will be conducted in four European countries, varying in terms of selectivity (Benadusi & Giancola, 2014): Finland, France, Germany, and Italy. We hypothesize that the variables outlined in the methodology will play a crucial role in determining the belonging to the upward deviant cases category, and that, depending on the level of selectivity in the education system, the school attended will significantly impact the probability of being a deviant case.

Our hypotheses will be tested through mono-, bi-, and multivariate analyses using the OECD PISA 2022 database. Initially, deviant cases will be identified for the four countries under analysis. These will be operationalized as the residuals from the regression between individuals' socioeconomic and cultural status (ESCS) and their performance in reading tests (one regression per country). Subsequently, we will examine, through linear regression for each country, the determinants of being a deviant case, including the secondary school track (excluding Finland, which has a common core system), student gender, migration background, and the experience of having repeated at least one school year. The model will then include (as a nested model) the school-level ESCS and the individual’s standard deviation from the school’s means (as a proxy for the individual status position related the ESCS of the school) as structural measures. In the second stage, binomial logistic models will be conducted using the same approach and variables to study the probability of being a deviant case across all four countries. The fourth quartile of the residuals of the regression between the ESCS and the reading performance to PISA test will be used as dependent variable (this variable represents the overperforming students versus all the others that we will assume as reference category). This will allow us to study deviant cases in a high-stakes context. Finally, for the same category, two multilevel binomial logistic regression models will be conducted for each country to examine how much the school attended affects the probability of being a deviant case, or more precisely, how much of this variance is explained by the school level (a total of 8 models). Four of these will include only the school level for each country, while four will include all the previously mentioned variables. We will use the Intraclass Correlation Coefficient (ICC) to observe the difference in explained variance between the null-model and the full-model, to determine how much the probability of being a deviant case depends on the school. The models, their coefficients, and the variance they reproduce cannot, however, be compared across countries due to structural reasons (differences in educational systems). However, this is only a partial limitation, as it allows us to theoretically compare the effects due to the context and the presence or absence of certain factors related to educational systems.

This study contributes to existing literature by exploring a relatively understudied aspect of educational outcomes: deviant cases, or students who perform better than expected given their socioeconomic status. By focusing on upward deviant cases across four European countries with varying levels of educational selectivity—Finland, France, Germany, and Italy—we aim to shed light on the determinants of this phenomenon and the role of schools in shaping such outcomes. The use of PISA 2022 data and a wide range methodological approach — including residual analysis, logistic regression, and multilevel models — can provide insights into individual, structural, and contextual factors influencing the likelihood of being a deviant case. Our findings will enhance understanding of how ascriptive characteristics, school characteristics (such as the track), and structural variables (school-level ESCS mean) interplay with school-level effects, offering new perspectives on the potential for educational systems to foster upward mobility. Lastly, this research emphasizes the importance of addressing structural inequalities within education systems and highlights the pivotal role schools can play in mitigating or amplifying these disparities. By identifying key factors that contribute to exceptional educational outcomes for disadvantaged students, the study offers valuable implications for policymakers seeking to promote equity and excellence in education.

- Agasisti, T., Avvisati, F., Borgonovi, F., et al. (2021). What school factors are associated with the success of socio-economically disadvantaged students? An empirical investigation using PISA data. Social Indicators Research, 157, 749–781. https://doi.org/10.1007/s11205-021-02668-w - Barrett, P., Treves, A., Shmis, T., Ambasz, D., & Ustinova, M. (2019). The impact of school infrastructure on learning: A synthesis of the evidence. Washington, DC: World Bank. http://hdl.handle.net/10986/30920 License: CC BY 3.0 IGO. - Benadusi, L., & Giancola, O. (2014). Secondary school systems: ‘Comprehensive’ versus ‘selective’. A comparison in terms of equity. Scuola democratica, Learning for Democracy, 2, 461-482. https://doi.org/10.12828/77426 - Bonanni, M., & Moreschini, I. (2024). Deviant cases from expected performance: The role of expectations beyond socio-economic and cultural status. Italian Journal of Sociology of Education, 16(3), 55-78. https://doi.org/10.14658/PUPJ-IJSE-2024-3-3 - Boudon, R. (1973). L’inégalité des chances. Paris: Colin. - Boudon, R., & Lazarsfeld, P. F. (1966). Méthodes de la sociologie: II. L’analyse empirique de la causalité. La Haye: Mouton & Co. - Bourdieu, P., & Passeron, J. C. (1977). Reproduction in education, society and culture. London: Sage. - Broer, M., Bai, Y., & Fonseca, F. (2019). A review of the literature on socioeconomic status and educational achievement. In Socioeconomic inequality and educational outcomes (Vol. 5, pp. 1-22). Cham: Springer. - Coleman, J. S., Campbell, E. Q., Hobson, C. J., McPartland, J., Mood, A. M., Weinfeld, F. D., & York, R. L. (1966). Equality of educational opportunity. Government Printing Office. - Gamoran, A., & Long, D. A. (2006). Equality of educational opportunity: A 40-year perspective. Sociology of Education, 79(1), 1-27. - Hanushek, E. A. (1997). Assessing the effects of school resources on student performance: An update. Educational Evaluation and Policy Analysis, 19(2), 141-164. https://doi.org/10.3102/01623737019002141 - Hanushek, E. A. (2016). What matters for student achievement. Education Next, 16(2), 18-26. - Jackson, M. (Ed.). (2013). Determined to succeed? Sociology: Performance versus choice in educational attainment. Stanford: Stanford University Press. - OECD. (2019). PISA 2018 results (volume III): What school life means for students’ lives. Paris: PISA OECD Publishing. - Rodríguez-Hernández, C. F., Cascallar, E., & Kyndt, E. (2020). Socio-economic status and academic performance in higher education: A systematic review. Educational Research Review, 29, 100305. https://doi.org/10.1016/j.edurev.2019.100305 - Woessmann, L. (2016). The importance of school systems: Evidence from international differences in student achievement. Journal of Economic Perspectives, 30(3), 3-32. http://www.jstor.org/stable/43855699