Purpose – The purpose of this study is to test the impact of entrepreneurship education on business, engineering and arts and science students using the theory of planned behaviour. The study adopted a pre-test–post-test (time 1, t 1 and time 2, t 2 ) to measure the change of attitudes and intentions over a period of six months. The participants who took entrepreneurship as a compulsory or elective course within their curriculum are 450 in total. To measure attitude, the subjective norm and perceived behavioural control, the study adopted a measure proposed by Kolvereid (1996b). For the intention to become self-employed, the study adopted a three-item measure of career intention, proposed by Kolvereid (1996b), which captures the intention of an individual to start a business. The results showed that the post-programme mean values of subjective norm, attitude towards self-employment, perceived behavioural control and intention towards self-employment increased in relation to the pre-programme ones. But the mean difference value in all four variables is higher for business students when compared to the other two student groups. Also, t -tests indicated no significant differences between respondents and “incomplete” non-respondents (students who filled the t 1 questionnaire but failed to respond at t 2 ). Design/methodology/approach – The study adopted a pre-test–post-test (time 1, t 1 and time 2, t 2 ) to measure the change of attitudes and intentions over a period of six months – one semester. A convenient sample technique has been used. The participants who took entrepreneurship as a compulsory or elective course within their curriculum are 450 in total – 100 (Bharathidasan University, Trichy) from business, 100 (National Institute of Technology, Trichy) from engineering and 250 (Bharathidasan University, Trichy) from art and science). The total 250 arts and science students were selected from four reputed art and science colleges in India where entrepreneurship course is offered either as compulsory or elective course, and due share of 60 was given to each college where the total number of students in the final year was 1,000 to 1,500 in each college. The 100 engineering students were selected from one reputed engineering college where the total number of final year students was 750. Finally, 100 business students were selected from two reputed business schools where the number of final year students was 600. All the students from arts and science and engineering were soon-to-graduate undergraduates and business students were soon-to-graduate postgraduates. It was clearly explained to the surveyed students that the questionnaires were for research purposes only, participation was voluntary and their views would not affect their grades. Both time 1 (t 1 ) and time 2 (t 2 ) questionnaires were reviewed by three academics and five non-participating students to ensure clarity of wording and face validity of the constructs. Findings – The overall response rate was 55.3 per cent. The mean and standard deviation of variables, attitude towards self-employment, subjective norm, perceived behavioural control and intention towards self-employment for the samples of business students, engineering students and arts and science students are presented in . To test the hypothesis, the present study used the following tests: Correlation (Tables III-V) and regression (Table VI) to test the relationship between attitudes and intention at t 1 and t 2 . To test the effect of the programme on the change of attitudes and intentions, the current study used one-way ANOVA on the difference scores (for sample of business, engineering and arts and science) with the group membership (programme) as the independent variable. The “difference scores” method is preferable to split-plot repeated measures ANOVA for pre-test–post-test designs, because it gives equivalent results in a simpler and less confusing way (Girden, 1992). No significant violations of the assumptions for t -test, repeated measured ANOVA and regression were identified. Specifically, the common problem of multicollinearity was not evident for all the three majors of students, as the correlations between independent variables were moderate and the tolerance values were all higher than 0.70 for business group, 0.72 for engineering group and 0.73 for arts and science group. Research limitations/implications – The study aimed to address the attitudes and intentions among business, engineering students and art and science students, but not actual behaviour, and therefore, the study echoes the suggestion that longitudinal studies following the subjects for years after graduation are the only way to prove with accuracy the intention–behaviour link (Kolvereid, 1996b). The study is a comparative study on the effect of entrepreneurship education through the Azjen’s theory of planned behaviour on the scores of variable attitudes towards self-employment, subjective norms, perceived behavioural control and intention towards self-employment and has not made any attempt to find out the causes for such scores. Originality/value – Using the theory of planned behaviour, the study tested the impact of entrepreneurship education on business, engineering and arts and science students. The study adopted a pre-test–post-test (time 1, t 1 and time 2, t 2 ) to measure the change of attitudes and intentions over a period of six months. The participants who took entrepreneurship as a compulsory or elective course within their curriculum are in total 450. To measure attitude, the subjective norm and perceived behavioural control, the study adopted a measure proposed by Kolvereid (1996b). For the intention of becoming self-employed, the study adopted a three-item measure of career intention, proposed by Kolvereid (1996b), which captures the intention of an individual to start a business. The results showed that the post-programme mean values of subjective norm, attitude towards self-employment, perceived behavioural control and intention towards self-employment increased in relation to the pre-programme ones. But the mean difference value in all four variables is higher for business students when compared to the other two student groups.
