Students' collaboration skills have recently been brought under the spotlight, especially after the latest PISA-measurement (OECD, 2017). Now that we have the knowledge from PISA, what do we need that knowledge for? Let us consider what happens when students collaborate face-to-face. In addition to individual competencies, the affective environment that surrounds students is an important factor in social learning and students' possibilities to be engaged (e.g. Reyes, Brackett, Rivers, White, & Salovey, 2012). The affective environment draws from a broader cultural-geographical environment including micro, meso, exo and macro systems (Bronfenbrenner, 1993). In Bronfenbrenner's model, these ecological systems include proximal processes (transfer of energy between the individual and the persons, objects, and symbols); competencies (demonstrations of acquisition, development or ability); and dysfunctioning (difficulties in maintaining control and integration of behavior). Here, we take a look at what kind of student initiated actions (i.e. proximal processes) can be identified in the student interaction around mathematical problem solving? How do the 12 collaborative problem-solving competencies measured in PISA (OECD, 2017) connect to those actions, and does possessing the competencies always lead to fruitful collaborative outcomes or not (dysfunctioning)? Drawing from a large database of the Social Unit project, a 45-minute student interaction is closely examined. As an example, a PISA-measured competence is demonstrated in the following arguments: Student 1: My God, what are you doing? Student 2: What? It's a square. This can be interpreted as a competence A1: Discovering perspectives and abilities of team members. However, the students' actions seem to reveal more than just a sign of a particular competence. The socio-affective interaction in the meso system may also influence students' ability to fulfil the potential of their competencies. The analysis of students' actions show that 1) the collaboration in question proceeded in stages: it was clearly more intensive at times and included everybody, and 2) many collaborative skills were displayed, yet, the collaboration fell apart frequently because of socio-affective factors and students' understanding of the teacher's expectations. As a conclusion, we claim that while measured at the micro-meso level, information about collaboration competencies might be more useful for decision making at the exo level. Such information might be relevant to the skills level of a larger group of people (such as a nation), but that information might not be easily applied within a specific group, e.g. for instruction or feedback purposes, if further aspects of socio-affective meso system are ignored.

Bronfenbrenner, U. (1993). The ecology of cognitive development: Research models and fugitive findings. In. R. H. Wozniak & K. W. Fischer (Eds.), Development in context: Acting and thinking in specific environments. The Jean Piaget symposium series. (pp. 3- 44). Hillsdale, NJ, England: Lawrence Erlbaum Associates. OECD (2017), PISA 2015 Results (Volume V): Collaborative Problem Solving, OECD Publishing, Paris. http://dx.doi.org/10.1787/9789264285521-en Reyes, M. R., Brackett, M. A., Rivers, S. E., White, M., & Salovey, P. (2012). Classroom emotional climate, student engagement, and academic achievement. Journal of educational psychology, 104(3), 700.