Abstract
Students' lack of interest in studying science and in science-related careers is a concern in the UK and worldwide. Yet there is limited data, particularly longitudinal, on the sources and development of science-related aspirations. In response, the ASPIRES (Science Aspirations and Career Choice: Age 10–14) longitudinal study is investigating the development of students' educational and occupational aspirations over time. In the first phase of the project, a questionnaire exploring science-related aspirations and interests was completed by over 9,000 primary school students across England. This survey allowed us to explore possible associations between attitudes and aspirations, links which have not been investigated in previous attitudinal studies of this scope. Overall, students expressed positive attitudes to science, reported positive parental attitudes to science and held very positive images of scientists. Multilevel modelling analyses revealed that aspirations in science were most strongly related to parental attitudes to science, attitudes to school science and self-concept in science, and are also associated with students' gender, ethnicity and cultural capital. However, the images students held of scientists were not as closely related to aspirations. These factors are discussed in more detail within the paper, alongside a consideration of possible school-related effects.
Keywords:
Acknowledgements
The authors wish to acknowledge Simon Rutt from the National Foundation for Educational Research for his role in conducting the statistical analyses reported in this paper. This research was funded by a grant from the Economic and Social Research Council, grant RES-179-25-0008.
Notes
Due to practical considerations, the sample was drawn from schools in England, rather than the whole of the UK (Wales, Scotland and Northern Ireland).
Dummy variables were created for the specific groups within the categorical variables, using males, White British, the skilled category of parental occupation and the medium level of cultural capital as bases for comparison. Missing values on these variables were assigned to an ‘unknown’ category, which was entered into the analysis. Consequently, all students (cases) were able to be included in creating the multi-level models.
Effect sizes of 0.3–0.5 are generally considered medium.
This relationship is reflected schematically because the graph of the data itself in contains 249 lines, making the difference demonstrated by the positive covariance more difficult to discern visually.