The purpose of this paper is to describe, illuminate and analyze individual learning strategies within the context of the automation/electrical engineering programme in vocational education in Swedish upper-secondary school system. The study was carried out during 2002-2004 in the subject of control technique. The aim of the study was to investigate an describe students´ learning and the learning environment brought about when computer-based learning software is employed in the education, and whereby labs take place through virtual simulations. Our main questions are:

• What learning strategies do the students develop in that sort of learning environment?
• How and what do they learn?
• What resources do they use in learning?
• What motives are there for their actions and behaviour in the classroom?

The study has been performed and monitored from a combination of a socio-cultural perspective (Säljö 2000; Vygotsky 1934/1986; Wertsch 1998), and activity theory (Leontiev 1978; Engeström 1987).In the first analysis, we describe the learning strategies of four students. These descriptions are based on video recordings of their interaction with the learning software, and interviews based on replayed selections from the video recordings. We also describe the learning software and the assigned tasks, as well as the pedagogical support offered. Research on computer-based simulation for educational purposes emphasises the importance of adequate instructional support built into the software; otherwise it must be supplemented by teacher support. Our study presumes that students be able to work independently with the learning software. This involves programming and testing functions in the virtual simulation models, in order to draw conclusions about how to control processes using different programming languages. However, our results show that only the single most competent of the students in the study masters the simulations in practice; all the other students perceive it as remarkably difficult. One of the decisive factors is whether or not the student perceives the virtual simulations as real. The comprehension of the students' actions in the classes is enhanced by the next analytical step, in which classroom interaction is put into focus. We employ activity theory as our chief analytical tool, and the results provide a different and complementary description of students' behaviour. It might seem that what takes place in the classroom is one activity, but if one takes the students motive for participating in it as one's point of departure, the result is that there are two parallel activities taking place alongside one another; one with respect to learning control techniques, and another for assessing the students' performance. These activities appear to coincide only for three of the students in the study, namely those whose grades are higher than 'pass'. Students who risk failing are offered special examination opportunities in order to pass, but it is not clear how they learn the substance of the control techniques. In the final analytical step, control techniques are pictured as a learning process at the crossroads of professional and school perspectives. We describe students' experience of on-the-job training as well as their perceptions of what control techniques require in terms of professional skills and knowledge. Our results highlight the problem of adapting and 'translating' what is taught in school to the circumstances of professional practice. In an abstract subject such as control technique, it is primarily a matter of understanding symbols and the functions they represent, and furthermore to see the symbols as tools and components in an authentic process. Those skills are necessary in order to perceive the virtual simulation practice as real with respect to professional practitioners.