Data from PISA and TIMSS have recently fuelled the debate on the efficacy of Inquiry-Based Science Teaching (IBST). Some analyses of these data show that the effects of different scientific activities related to IBST carried out with students vary according to the frequency of their implementation. Extending this research, the present study focuses on several scientific activities (Conception, Problem, Hypothesis, Model, Manipulation, Discussion-IBST, Interaction), some of which have not been considered previously. It investigates the extent to which these activities are implemented in teachers' practices and their relationships with student learning. The study is based on self-reported practices of primary school teachers in France (98) and data on their students (2250) measuring their scientific knowledge, inquiry skills, views of the nature of science (NOS) and attitudes towards science. Findings show that teachers' practices are diverse, with some teachers implementing scientific activities frequently and others rarely. However, no profile of teaching practices could be associated with high or low levels of student learning. The only scientific activity for which converging positive relationships with student learning were found was modelling. This finding calls for further studies to explore the relationships of the different modelling sub-practices with student learning.