The question of room for manœuvre is present at each level in the field of school-to-work transition. It is a matter of concern for the student who is considering the possible strategies towards the targeted vocational position. It is a matter of concern for guidance counsellors who analyse the relevance and the realism of students' vocational projects. Actually, it is a matter of concern for the persons in charge of education policy and of education institutions, when wondering which actions are to be designed and implemented against inequalities in education.Considering this question of room for manœuvre in school-to-work transition, two attitudes can be distinguished. The first one, following Bourdieu and Passeron, considers that social determinism prevails. In this approach, social origins and context determine the transitional process, and the personal efforts of the agent do not count. On the contrary, the second attitude, following Gary Becker, gives priority to the agent's action, and considers that the transitional process is essentially strategic. The first objective of this paper is to show that determinism and strategy can be referred to neither indifferently nor in any case, but must be referred to in precise different cases. In other words, there is a "topography" of determinism and strategy, each playing in its own influence zone.The basic idea is that determinism can be measured. Chaos theory can help to take this measurement by means of Lyapunov exponents (Kiel et Elliott, 1996). By applying this method to data from surveys on school-to-work transition of University graduates, we show (Tchibozo, 2002a) that, for some diploma specialities and levels, the school-to-work transition process is structurally determinist. In other words, as soon as the holders of these diplomas have similar characteristics (e.g. same age, same gender, same diploma), they experience similar trajectories. For instance, this has been observed in a French survey on Bachelor's degree graduates in Biology. In other cases, we observe rather chaos: given similar initial conditions, graduates experience radically divergent trajectories. For instance, this has been observed in the transitional process of groups of doctorate holders.The second objective of the paper is to stress the consequences of the trajectories regularity or irregularity on the room for manœuvre of agents and of education public policy.For an agent, strategy is always possible, whether the process is determinist or not. But the nature of the strategy depends on the case. When the process is determinist, the strategy consists in placing oneself in the good initial conditions, for instance to obtain one's diploma in the "good" institution at the "good" age, etc., in order to reach the targeted vocational position. On the contrary, when the transitional process is irregular, divergent, chaotic, the agent really has to plan a true strategy, in the full sense of the word: the agent has to take into account the behaviour and reactions of the employers, and especially their functional and meta-functional (Tchibozo, 2002b) requirements; the agent has to be able to acquire information; he/she has to be able to anticipate and to adapt to the evolution of education supply and of the labour market.For public education policy, determinist processes are a field where action is possible. Since the processes are regular, public education policy is justified to set up actions for correcting initial inequality and to set up affirmative action, in order to give students the same chances to reach the same set of vocational positions. On the contrary, insofar as public action is general and does not have to target individual and personal objectives, there is no room for public action when the processes are chaotic, irregular, unforeseeable. Therefore we wonder what responsibility the irregular nature of some processes may have in the partial failure of policies against education inequality experienced in Europe during the last decades.ReferencesKIEL L.D. & ELLIOTT E., eds. (1996). Chaos Theory in the Social Sciences: Foundations and Applications. Ann Arbor: University of Michigan Press.TCHIBOZO G. (2002a). "Measuring the sensitivity of initial conditions in individual procedures for finding first professional employment". Bulletin of Sociological Methodology, n°74, pp. 33-52.TCHIBOZO G. (2002b). " Meta- functional criteria and school-to-work transition ". Journal of Education and Work, vol. 15, n° 3, pp. 339- 352. Tables in vol.15, n°4, pp. 481-484.