Explanations for the extensive variation in rates of entrepreneurship across countries and gender have been biased by the dominant paradigms of the field, the limits of the resulting research designs, and ultimately by the lack of a comprehensive multilevel theoretical framework for predicting and understanding the decision to start a business. I propose an alternative view of entrepreneurship based on contemporary social theory; in particular on Bourdieu's theory of practice. The resulting practice theory model provides a set of theoretical propositions and hypotheses that fit well with existing evidence from the field of entrepreneurship and from other studies of gender and work. Results from a series of 2-level random coefficient models investigating the relative impacts of micro- and macro-level correlates of nascent entrepreneurship strongly support a practice theory view of entrepreneurship. The capital structures -- economic, cultural, social and symbolic -- that define social positions explain most of the gender variation and a large extent of the variation across countries. Country characteristics also explain a good portion of the country variation. Perceptions, social ties, and national gender culture are significant correlates of nascent entrepreneurship, net of all other micro- and macro-level factors. The variation in rates of nascent entrepreneurship is more extensive across countries than across gender. Micro-macro interactions are also explored.