In this paper we propose that the unpopularity of science in many industrialised countries is largely due to the gap between the subculture of science, on the one hand, and students' self-image, on the other. We conducted a study based on the self-to-prototype matching theory, testing whether the perceived mismatch between the typical representative of the science culture (the science prototype) and students' self-image is linked to not choosing science as a major. Fifty-four Dutch ninth-grade students currently choosing their subject majors (so-called profiles) completed a Dutch version of a questionnaire previously designed by Hannover and Kessels, which measures students' perceptions of typical peers favouring different school subjects (prototypes for physics, biology, economics, languages) and students' self-image. Students chose a profile to the extent that they conceived of themselves as similar to the typical peer who likes the key subject of that profile. Fifty per cent of variance was explained when using an aggregated science versus humanities distance score and predicting whether a student had chosen a science-related or a humanities-related profile. A comparison of Dutch students' description of the physics prototype with the German data from Hannover and Kessels revealed similar prototypes in both countries. The traits ascribed to the physics prototype were in line with science-related values and the culture of science as described by Merton and Traweek, for example. The relevance of the perceived fit of the culture of science to students' selves for academic choices is discussed. (Contains 5 tables and 2 figures.)