Finnish students’ high achievement and small between-school differences in the OECD PISA studies have been seen to indicate the equality and equity of the Finnish education system. Undermining this, Finland stood out among the Nordic countries in TIMSS 2011 (Trends in International Mathematics and Science Study) due to its considerably higher between-class differences (Yang Hansen, Gustafson, & Rosén, 2014). The results were confirmed in a longitudinal large-scale study compris-ing 10 000 lower secondary students in 129 schools and 460 classes in the Helsinki metropolitan area, showing that there not only were considerable differences between classes already at the beginning of grade 7 but that these differences grew during the three years of lower secondary education.

To complement the metropolitan study, the present study looks at the role the way the municipalities allocate students to different schools and the schools to different classes, plays in the creation of between-school and between-class differences in the Finnish comprehensive school. To do this, we draw on a large-scale national assessment study, including a questionnaire for school principals regarding class formation. We will focus first on the allocation of students to different types of classes, and secondly on the differences between the students’ cognitive competence, learning motivation and choice of track in the Finnish two-track (academic and vocational) upper secondary education as well as students’ home background according to the type of class they study in.

The allocation of students into schools and classes by achievement is a common practice. There are clear differences both in the form of the selection of students into the different tracks and in when in student’ educational path the selection takes place, however. In some countries, students are selected into schools or tracks with differing curricula early in their education careers (e.g., Austria, Belgium, Germany, the Netherlands) whereas in others (e.g., the Nordic countries, France, Poland, Japan, USA), all students follow the same curriculum at least through compulsory education (Eurydice 2014; OECD 2012). The first type of tracking represent open ability-based differentiation, but also the latter can and often does include differentiation based on ability or other student characteristics, allocating students into different schools or into different tracks or streams within schools. The latter may compass permanent grouping covering all subjects such as music classes in the Finnish primary school or be subjects-specific as in the advanced placement (AP) courses common to many American schools (Chmielewski, 2014).

Ability-based tracking and streaming has been a hotly debated topic for a long time with arguments often resting on a perceived trade-off between equity and efficiency (Entwistle & Alexander, 1992; Hanushek & Woessman, 2006; OECD, 2012, 2013). The central argument for tracking is that homogeneous classrooms permit a focused curriculum and appropriately paced instruction, leading to better learning for all. Arguments for ungrouped classrooms centre on the concern that lower ability students will be disadvantaged by less ideal learning environments regarding teacher proficiency, curricular content, and peer effect. However, some have found no evidence of ability grouping (Figlio & Page, 2002; Lefgren, 2004) and some have even found de-tracking to affect negatively the attainment of just the students the heterogeneous groups are supposed to support (Loveless, 2009). One interpretation for the former has been that the positive effects of achievement-specific instruction of tracking to overcome the negative (or lacking positive) peer effects for students in lower-achievement tracks (e.g., Zimmer, 2003; Duflo et al., 2011). Loveless (2009) attributes the negative impact of de-tracking to a diminishing incentive to try one’s best, increasing especially minority students’ risk to succumb to peer pressure for under-achievement.

The study draws on a large-scale national assessment of 7 500 ninth graders’ (49 % boys) in spring 2017 with a random sample of 82 schools across Finland. The assessment comprised all ninth grade classes and students in the participating schools. Theoretically, the assessment is based on the notion of learning to learn (Csapó, 2007; Hoskins and Fredriksson, 2008), relying on a rich battery of cognitive tasks and attitudinal questionnaires. The learning to learn skills scale comprises four tasks on reasoning, two tasks on mathematical reasoning and two tasks on reading comprehension. The reasoning tasks included three tasks from the Ross and Ross test of higher cognitive processes (1979), and a modified version of the Piagetian task of Control of variables (Hautamäki, 1984; Inhelder & Piaget, 1958; Shayer, 1979). The tasks of mathematical reasoning included one tasks each of Demetriou’s and colleagues (1991) and of Stenberg and colleagues (2001). In addition, the study included a separately implemented test on ninth-grade curricular mathematic. The attitudinal scales used in this study comprise agency beliefs (e.g. Chapman, Skinner, & Baltes, 1990), academic self-concept (e.g. March, 1990), and goal orientation (e.g. Elliot & Dweck, 1988). In the analyses, they were grouped form four dimensions relevant form the point of view of the school: attitudes supporting learning, attitudes detrimental for learning, trust in personal ability, and academic self-concept. In addition to the student-level measurements, the presentation draws on a concurrent questionnaire to school principals regarding the allocation of students to different classes in the beginning of grade 7, and on the character of each of the ninth grade classes participating in the study (type of student selection). The principals’ answers to questions regarding the formation of classes at the beginning of grade seven have been classified using both the quantitative and the open answer data from the questionnaire. Data on students’ cognitive competence and motivational attitudes have been analysed using basic statistical methods. Between-school and between-class differences have been analysed both independently using ANOVA and by using variance component and multi-level analyses (SPSS24 and Mplus). Despite the nested structure of the data, the former has been adopted on side of the latter to reflect the special Finnish form of school choice (Seppänen et. al, 2015) where the object of choice is not primarily the school but a class offering a special curricular emphasis (i.e., music, mathematics, languages).

The principals list a variety of factors guiding class formation in lower secondary schools. Main deviations from random allocation are classes offering a special curricular emphasis (15 % of students) and smaller classes reserved for student with need for support (5 % of students). Half of the classes with a special emphasis are based on literacy subjects (e.g., language, mathematics) and the other half on other subjects (e.g., music, sports. Acceptance to these classes is based on a test or other indication of ability or interest. Students in the classes with special emphasis outperformed the students in the other classes in all cognitive tasks (independent effect of class ƞ2 = .17–.25, p < .001), and their learning-supportive attitudes were stronger than were those of students in other classes (independent effect of class ƞ2 = .12–.15, p < .001). The students in classes with an emphasis on some literacy subject outperformed those in classes with a none-literacy emphasis even if the overall better-performing girls were clearly overrepresented in the latter type of classes. As expected, the performance of students in the smaller classes was lower than that of those in other classes, confirming the Finnish phenomenon of a contrary correlation between class-size and achievement. The differences are somewhat bigger in the grades of all literacy subjects in students’ final report cards (independent effect of class ƞ2 = .21–.27, p < .001). The difference in the mean achievement (independent effect of class ƞ2 = .23, p < .001) can be seen to guide students’ choice of upper secondary track with attendance of a selective class reducing radically students’ probability of choosing the vocational track (46 % for students in ‘normal’ classes, 24 % for those in selective classes, and 85 % for students in the smallest classes).

Chapman, M., Skinner, E.A., & Baltes, P.B. (1990). Interpreting correlations between children's perceived control and cognitive performance: Control, agency or means-ends beliefs. Developmental Psychology, 26 (2), 246-253. Chmielewski, A. K. (2014). An international comparison of achievement inequality in within-and between-school tracking systems. American Journal of Education, 120(3), 293–324. Csapó, B. (2007). Research into learning to learn through the assessment of quality and organization of learning outcomes. The Curriculum Journal, 18(2), 195-210. doi: 10.1080/09585170701446044 Demetriou, A., Platsidou, M., Efklides, A., Metallidou, Y. and Shayer, M. (1991) ‘The development of quantitative-relational abilities from childhood to adolescence: Structure, scaling, and individual differences’, Learning and Instruction, 1:19–43. Elliot, A.J. & Dweck, C.S. (1988). Goals: an approach to motivation and achievement. Journal of Personality and Social Psychology, 54 (1), 5-12. Entwistle, D., & Alexander, K. (1992). Summer setback: race, poverty, school composition and educational stratification in the United States. American Sociological Review, 57, 72–84. 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. Inhelder, B. and Piaget, J. (1958) ‘The early growth of logic in the child’, London: Routledge and Kegan Paul. Kintsch, W. and van Dijk, T.A. (1978) Toward a model of text comprehension and production. Psychological review. Volume 85, Number 5, September 1987, 363-394. Loveless, T. (2009). Tracking and detracking: High achievers in Massachusetts middle schools. Thomas B. Fordham Institute. Marsh, H.W. (1990b) ‘The structure of academic self-concept: The Marsh-Shavelson model’, Journal of Educational Psychology, 82: 623-636. Ross, J.D. and Ross, C.M. (1979) ‘Ross test of Higher Cognitive Processes’, Novato, California: Academic Therapy Publications. Seppänen, P., Kalalahti, M., Rinne, R., & Simola, H. (Toim.). (2015). Lohkoutuva peruskoulu. Perheiden kouluvalinnat, yhteiskuntaluokat ja koulutuspolitiikka. Jyväskylä: Suomen kasvatustieteellinen seura. Kasvatusalan tutkimuksia 68. Sternberg, R. J., Castejon, J. L., Prieto, M. D., Hautamäki, J. and Grigorenko, E. L. (2001) ‘Confirmatory factor analysis of the Sternberg Triarchic Abilities Test in three international samples. An empirical test of the triarchic theory of intelligence’, European Journal of Psychological Assessment, 17:1-16. Zimmer, R. (2003). A new twist in the educational tracking debate. Economics of Education Review, 22(3), 307-315. Yang 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