Students’ ability-based tracking or streaming into schools and classes is a common practice but its forms differ by countries. Students can be allocated into different schools or tracks with differing curricula relatively early, as is the case in Austria, many German states, and the Netherlands. Or, as in France, the Nordic countries, and the USA, all students follow the same curriculum through compulsory education or at least grades 1–9 (Eurydice, 2014). The first can be seen to represent open ability-based differentiation but also the latter can – and often does – include differentiation within schools. This may mean permanent grouping covering all subjects or to be subject-specific as in the Success for All programme for early reading instruction (Slavin et al., 1996) or the US advanced placement (AP) courses (e.g., Loveless, 2009).

Arguments for and against ability-based tracking or streaming rest mainly on a perceived trade-off between equity and efficiency (e.g., Entwistle & Alexander, 1992; Hanushek & Woessman, 2006). The central argument for tracking rests on a homogeneous classroom permitting a better-fitted curriculum and pace of instruction based on each student’s actual performance level, leading to better learning for all (i.e., Vygotsky’s Zone of Proximal Development). Arguments against tracking centre on concern for lower-ability students being disadvantaged by less rigorous curricular content, the lack of positive peer effect, and potential lower teacher proficiency. Evidence for the effects is not unanimous, however, with some studies having found no negative evidence of the impact of ability grouping (e.g., Figlio & Page, 2002; Lefgren, 2004) while Loveless (2009) found de-tracking to affect negatively the attainment of high-performing minority AP-students.

Finnish students’ high achievement and small between-school differences in the OECD PISA studies have been seen to indicate the equity of the Finnish education system. In TIMSS 2011, however, Hansen and her colleagues (2014) found Finland to stand out among the Nordic countries with its disproportionally large between-class differences. These were confirmed in a large-scale study of 10 000 Helsinki metropolitan region students, with considerable between-class differences that increased during the three years of lower secondary studies (Kupiainen, 2016). Reasons for the differences remained open due to lacking information on class-formation.

Despite the officially non-streamed Finnish comprehensive school for grades 1–9 and a strong neighbourhood-school-principle, municipalities can provide parents and students choice of school and/or of a class with extra lesson hours in some curricular subject. These so called ‘classes with an emphasis’ have been shown to cater disproportionally to higher-achieving and more strongly mastery-oriented students from higher-SES families, and have been accordingly criticized for threatening the equality of Finnish basic education (e.g., Seppänen et al., 2015; but see also Kantasalmi & Kupiainen, 2021).  

The first results of the current study revealed that 36% of the 4,075 Helsinki ninth graders who participated in the assessment the data is based on, studied in a class with aptitude-based student selection (academic or non-academic emphasis) and 5 p% in small classes comprising solely of SEN students. Students in the selective classes clearly outperformed students in the regular and SEN classes in all cognitive tasks, had stronger mastery attitudes, and were set to continue their studies in the academic track of upper secondary education more often that the other students.  In our presentation, we will deepen the analyses by the means of propensity score matching, investigating the state and development of between-school and between-class differences in the assessed cognitive competences, learning motivation, and school achievement and their development during the three years of lower secondary education, as well as their final GPA and choice of academic vs. vocational track in the two-track upper secondary education.

The study draws on a longitudinal assessment of learning to learn, implemented in Helsinki schools in 2011–2018. The study began with a random sample of 744 students at grade 1, and ended with a full participation of 4,075 students at the end of grade 9 (50 % boys). The assessment comprised a rich array of cognitive tasks and motivational/attitudinal questionnaires under the theoretical notion of learning to learn (see Hautamäki & Kupiainen, 2014). For both time points (beginning of grade and end of grade 9) the cognitive component comprises three tasks or dimensions: one task in mathematical reasoning (Demetriou et al., 1991), one in reading comprehension (Kinch & van Dijk, 1978; Lehto et al. 2001), and one in verbal reasoning (Ross & Ross, 1984). In addition, the students were given in grade 9 a sample of free PISA-tasks from 2003 and 2006, which will be used in the analyses as a proxy for more curricular content. The self-reported attitudinal component covers agency beliefs (Chapman et al., 1990), goal orientation (Elliot & Dweck, 1988), and academic self-concept (Marsh, 1990). As an additional measure for task effort, we will use the log data provided by the digital assessment. Students’ final 9th grade GPA and choice of upper secondary school and track was attained from the National Joint Application Register. Basic descriptives on students’ cognitive competence and motivational attitudes will be analysed using SPSS24, and multilevel modelling for the nested data using MPlus with between-school and between-class differences analysed by intraclass correlations. The actual focus of the presentation, however, will be on the impact of the different types of selective classes on the development of their students’ cognitive competence, motivation, and school attainment across the three years of lower secondary education, investigated by propensity score matching (PSM, e.g., Baser, 2006), using R (e.g., Randolph & Falbe, 2014). In addition to background variables (gender, parents’ education, 6th grade GPA) the basis for the PSM will be chosen from the cognitive and motivational measurements of the assessment implemented at the beginning of grade 7. Regarding the classes formed already in the primary grades, we hope to be able to extend the PSM to rest on base data from the earlier measurement points of grades 1 and 4.

As revealed by the early analyses, 36% of the assessed students studied in selective classes, based on a parent-made choice in primary school or an interest- or aptitude-test based selection at the transition from primary to lower secondary school. Half of the selective classes (18% of students) had an academic emphasis (e.g., foreign language, STEM) and the other half a non-academic emphasis (e.g., music, art, PE). In addition, 5% of the students studied in small SEN classes. Hence, only 60% of students studied in mainstream classes. Girls were overrepresented in the selective classes based on a foreign language, music or art, while boys were overrepresented in classes with an emphasis on a STEM subject or PE. Students in selective classes had, on average, higher educated parents than students in the other class types. Students in selective classes outperformed others in all cognitive tasks (ƞ2=.113; p<.001), and they were more mastery-oriented (ƞ2=.022; p<.001). Students in classes with an academic emphasis outperformed the students of non-academic classes despite the overrepresentation of girls in the latter. Yet, selective classes explained less than half of the overall between-class differences in reasoning (ƞ2=.026; p<.001). Between-class differences were slightly bigger (ƞ2=.029; p<.001) in GPA, which is critical for the choice of and acceptance to the academic track of upper secondary education. The academic track was chosen by 87% of students in selective classes versus 60% in regular classes and 20% in SEN classes. In the multi-level regression model, school-level accounted for 8.7% and class-level 20.2% of variation in the PISA tasks – a clear difference from the between-school differences reported in the original 2003 and 2006 OECD studies for Finland. In the presentation, the above results will be explored further with PSM to unravel the added value the different selective classes provide their students.

Baser, O. (2006). Too much ado about propensity score models? Comparing methods of propensity score matching. Value in Health, 9(6), 377-385. Entwistle, D., & Alexander, K. (1992). Summer setback: race, poverty, school composition and educational stratification in the United States. American Sociological Review, 57, 72–84. Eurydice (2014). The structure of the European education systems 2014/15: schematic diagrams. Eurydice – Facts and Figures. http://eacea.ec.europa.eu/education/eurydice/facts_and_figures_en.php#diagrams Figlio, D. N., & Page, M. E. (2002). School choice and the distributional effects of ability tracking: does separation increase inequality? Journal of Urban Economics, 51(3), 497-514. Hansen, K., Gustafsson, J.E., & Rosén, M. (2014). School performance differences and policy variations in Finland, Norway and Sweden. In Northern Lights on TIMSS and PIRLS. Differences and similarities in the Nordic countries. TemaNord 2014:528. Hanushek, E.A. & Woessman, L. (2006). Does Educational tracking affect performance and inequality? Differences-in-differences evidence across countries. The Economic Journal, 116 (March), C63–C76. Hautamäki, J. & Kupiainen, S. (2014) Learning to learn in Finland. Theory and policy, research and practice. In Ruth Deakin Crick, Cristina Stringher & Kai Ren (Eds.) Learning to Learn. International perspectives from theory and practice. Routledge. Kantasalmi, K., & Kupiainen, S. (2021). Classes with selective intake in Finnish Comprehensive School: A Problem of Societally Equal Opportunity for Schooling or a Boost for Learning and Equity in Pedagogical arrangements? International Journal of Educational Research, 109, 101857. Kupiainen, S. (2016). Luokkien väliset erot. In R. Hotulainen, A. Rimpelä, … & T. Wallenius, Osaaminen ja hyvinvointi yläkoulusta toiselle asteelle: tutkimus metropolialueen nuorista, Opettajankoulutuslaitos, Tutkimuksia 398, 67–95, Helsinki: Helsingin yliopisto. Lefgren, L. (2004). Educational peer effects and the Chicago public schools. Journal of Urban Economics, 56(2), 169-191. Loveless, T. (2009). Tracking and detracking: High achievers in Massachusetts middle schools. Thomas B. Fordham Institute. Randolph, J. J., & Falbe, K. (2014). A step-by-step guide to propensity score matching in R. Practical Assessment, Research & Evaluation, 19. Seppänen, P., Kalalahti, M., Rinne, R., & Simola, H. (eds.). (2015). Lohkoutuva peruskoulu. Perheiden kouluvalinnat, yhteiskuntaluokat ja koulutuspolitiikka. Suomen kasvatustieteellinen seura. Kasvatusalan tutkimuksia 68. Slavin, R.E., Madden, N.A., Dolan, L.J., Wasik, B.A., Ross, S., Smith, L. & Dianda, M. (1996). Success for All: A Summary of Research. Journal of Education for Students Placed at Risk, 1:1, 41–76.