Purpose - Entrepreneurial intentions have been extensively studied in student populations, with results suggesting that higher education does not promote formation of entrepreneurial intentions (e.g. Varamaki et al., 2013). However, the gap between intending to start a business and actually doing something to start one remains. The purpose of this paper is to analyze the antecedents of entrepreneurial intentions with higher education students and to analyze the antecedents of actual behaviors related to start-ups with higher education students. The authors use theory of planned behavior (TPB) for both analyzing the intentions and analyzing the actual behaviors. Design/methodology/approach - The authors apply Ajzen's (1991) TPB to entrepreneurial intentions of higher education students and test their relevance as antecedents of actual behaviors. In addition to the basic elements of TPB (attitudes, subjective norm (SN) and perceived behavioral control (PBC)) the authors test the impact of entrepreneurial characteristics (EC) (innovativeness, tolerance of ambiguity, creative problem solving and the ability to organize) of the student for intentions and for actual behaviors related to start-ups. Gender, entrepreneurial role models and basic education (upper secondary school vs vocational) are used as control variables. The data were collected in fall 2012 in seven different universities of applied sciences and consists of 3,754 responses, including 182 from respondents who were starting their own business at the time of the study. Findings - For entrepreneurial intentions, the model explains 47 percent of the variance. The best antecedent of intentions seems to be attitude followed by PBC. EC and SN are significant but their role in the model is quite small. For actual behavior related to starting up a firm, the results are quite different. Gender explains behavior significantly and so does basic education. The most important independent variable is EC followed by PBC. Attitude toward entrepreneurship is not significant in explaining the actual start- up behavior, nor is SN. Research limitations/implications - From an empirical standpoint, the sample was limited to higher education students in one country. This limits the scope of generalization. Further studies, comparing antecedents of intentions and behavior in multiple contexts should be undertaken. Another limitation of the study is that the authors have been unable, due to the small number of students engaged in start- up activities, to examine the differences between study years. Practical implications - The results raise interesting and difficult questions for entrepreneurship educators. Should entrepreneurship education aim at actual behaviors rather than intentions - or neither? What is vocational education doing "right" compared with upper secondary school or, indeed, higher education? Furthermore, if EC are this significant for actual behavior, what should or could be done to promote development of such characteristics in higher education? Originality/value - The paper contrasts entrepreneurial intentions with actual behaviors related to starting up a firm and tests the utility of the TPB antecedents in this context. As the development of entrepreneurial intention has generally been studied in isolation from actual activities, the paper presents an interesting counterpoint to previous research.
