To gain insight into the complex entrepreneurship process of venture creation (Gartner 1985), which involves the entrepreneur, environment and the opportunity, the cognitive factors inspiring the entrepreneur to establish a new venture are being increasingly explored (Autio et al. 2001; Forbes 1999; Krueger, Reilly & Carsrud 2000; Liñán & Chen 2006; Peterman & Kennedy 2003; Tkachev & Kolvereid 1999). Entrepreneurial cognitions are described (Mitchell et al. 2002:9) as ‘the knowledge structures that people use to make assessments, judgments, or decisions involving opportunity evaluation, venture creation, and growth’. Thus, entrepreneurial intentions are defined as an individual's conscious awareness and conviction to start a new venture in the near future (Thompson 2009).

The theory of planned behaviour (TPB) (Ajzen 1991) postulates that the greater the intention to initiate a behaviour, the higher the probability of its implementation. The TPB applied to entrepreneurial intentions of individuals hypothesises that a combination of three antecedents, personal attitudes, subjective norms and perceived behavioural control, act as initiators of the intention to create a new venture. Personal attitudes are evaluations of a specific behaviour including what individuals deem likeable, attractive or advantageous and can be both affective and evaluative (Ajzen 1991). Subjective norms denote normative beliefs and compliance thereto, as well as the perceived social pressures from significant others to perform or avoid certain behaviours (Ajzen 1991). Significant others include reference people, such as family, friends and colleagues. Perceived behavioural control refers to a person's subjective assessment of the ease or difficulty of performing a task or behaviour and the level of control over the behaviour (Ajzen 2002).

Grounded in the TBP (Ajzen 1991) and the model of entrepreneurial event (Shapero & Sokol 1982), the Shapero-Krueger model of entrepreneurial intention was developed (Krueger et al. 2000). Likewise, Liñán, Rodríguez-Cohard and Rueda-Cantuche (2005) combined the models of Shapero and Sokol (1982) and Ajzen (2006) and the theory of social capital (Grootaert & Van Bastelaer 2001; Uphoff 2000; Woolcock & Narayan 2000) to create an entrepreneurial intent model with three antecedents: personal attraction towards entrepreneurship, perceived social norms and perceived feasibility (self-efficacy). Whilst Thompson (2009) was involved in the clarification of the entrepreneurial intent constructs, Liñán and Chen (2006; 2009) refined and validated the entrepreneurial intentions questionnaire (EIQ) which consisted of the following constructs: entrepreneurial intent (EI), personal attitude (PA), perceived behaviour control (PBC) and subjective norm (SN). They tested the EIQ on students from two culturally diverse samples from developed countries, namely Spain and Taiwan, and found it valid. Subsequently Liñán partnered with different authors to further test the EIQ (Liñán, Nabi & Krueger 2013; Liñán, Rodríguez-Cohard & Rueda-Cantuche 2011; Liñán, Urbano & Guerrero 2011). Several other authors used different constructs to develop an EI measurement instrument (Hammami & Affes 2013; Parente & Feola 2013; Yang 2013). It can be concluded that a standardised EIQ does not yet exist.

The Liñán and Chen (2009) EIQ was used to test entrepreneurial intent of final year students in rural SA (Malebana 2012) and it was decided to use the same EIQ to test the entrepreneurial intent of the South African and Polish students. However, it was critical to first determine the applicability of the constructs and factors in the Liñán and Chen EIQ to the students from the two developing countries, SA and Poland. Swanepoel and Nieuwenhuizen (2014) applied factor analysis to each one of the seven construct scales to test for unidimensionality. Only two of the seven constructs were unidimensional, whilst for three constructs one or more items had to be discarded and a further two constructs subdivided into several factors. These redefined EI scales were used to differentiate between the EI of South African and Polish students and are further explained in the section on methodology.

Using university students to test EI is appropriate as the findings can contribute to education policies and more specifically to entrepreneurship education policies and holds implications for public decision-makers who develop support programmes for entrepreneurship. Several authors have used university students to test EI (Audet 2004; Gird & Bagraim 2008; Liñán & Chen 2009; Liñán et al. 2013; Liñán, Rodríguez-Cohard & Rueda-Cantuche 2011; Liñán & Santos 2007; Wu & Wu 2008), specifically the effect of entrepreneurship education on EI (Malebana 2012; Peterman & Kennedy 2003; Saeed et al. 2013; Souitaris, Zerbinati & Al-Laham 2007).

Whereas many researchers focus on the extent of EI in a specific group in a country, several researchers have conducted country comparisons (Table 3). The largest comparison was completed by Iakovleva, Kolvereid and Stephan (2011) who compared EI in 14 countries, whilst Engle et al. (2010) compared 12 countries and Heuer and Liñán (2013) compared 10 countries. South Africa has not been included in any of the country comparative studies whilst Poland was included in one (Moriano et al. 2012). Nevertheless, the EI of South African students has been researched by several authors (Farrington, Venter & Neethling 2012; Gird & Bagraim 2008; Malebana 2012; Urban 2006; 2012).

The results of a 12-country study (Engle et al. 2010) suggested that Ajzen's model of planned behaviour, as operationaliaed in that study, successfully predicts EI in each of the 12 countries. The countries included both developed and developing countries, namely Germany (Germanic Europe), France and Spain (Latin Europe), Ghana (sub-Sahara Africa), Russia (eastern Europe), Sweden and Finland (Nordic Europe), Egypt (Middle East), China (Confucian Asia), Bangladesh (southern Asia), Costa Rica (Latin America) and the United States (Anglo), with the regional cluster in parenthesis. As foreseen by Ajzen, the significant contributing model elements differed by country as did the percent of the variance explained by the model, although one model element, social norms, was a significant predictor of EI in each country.

Research into the determinants of EI of students in China and the USA revealed significant differences between the two groups pertaining to the effect of work experience and entrepreneur parents on entrepreneurial intention (Wang, Lu & Millington 2011).

Regional differences exist not only between countries but also within a country. A study by Liñán, Urbano and Guerrero (2011) determined the environmental cognitive elements that may explain regional differences in start-up intentions of potential entrepreneurs from two Spanish universities. In their study they differentiated between the valuation of entrepreneurship in a person's closer and wider environments. Items that measured the closer environment involved the valuation of entrepreneurship by direct personal contacts including family, friends and colleagues. This construct was named ‘closer valuations’. The wider environment items included culture and country with the construct ‘social valuation’ (Liñán et al. 2011). In this study it was determined that the social valuation of entrepreneurship was higher in the more developed region of Spain with a positive effect on the perceived SNs and behavioural control of potential entrepreneurs. In the other region the ‘closer valuation’ was more important and it affected the attitude towards entrepreneurship and SNs of potential entrepreneurs. The conclusion was that more positive entrepreneurial values should be promoted (Liñán et al. 2011). Other studies on regional differences within a country were conducted by Kibler (2013) in Finland and Lin et al. (2013) in Shri Lanka.

A study that focused on the differences between developing and developed countries with regard to the entrepreneurial intentions of undergraduate students (Iakovleva et al. 2011) found that the TPB is fully replicable in both country groups with equal predictive power. The authors found that entrepreneurial intentions as well as attitudes, SNs and PBC are higher in developing countries than in developed countries. The authors suggested exploration of the robustness of the TBP on other types of students from different countries for future research. In addition they advised that it would be informative to determine how cultural antecedents and formal national institutions might influence attitudes, SNs and planned behavioural control. In compliance with these suggestions, this study will investigate the robustness of the TPB on master's in business management students from two developing countries, SA and Poland. It follows from Table 3 that no research has yet been conducted comparing these two developing countries, nor comparing SA with any other country, with regard to EI.

The research question is: Does EI and its antecedents differ between South African and Polish master's in business management students from the University of Johannesburg (UJ) and the Krakow School of Business (KSB)?

From the discussion the following propositions emerged:

For the cross-sectional survey design, two samples, one group from each university, were drawn from master's students in business management: MCom students at UJ, SA, and MBA students at the KSB, Poland. It should be noted that these students are not entrepreneurship students and were not subject to entrepreneurship education. In June 2013, a total of 182 questionnaires were distributed and all were completed and returned.

A comparison of the profiles of the UJ MCom and the KSB MBA students appears in Table 4. The gender distribution for the two universities was similar, approximating a 40/60 female and/or male split. The age distribution differed in the older age categories with nearly twice the number of UJ MCom students (20.5%) in the 40–49 age bracket as opposed to KSB MBA students (11.2%). Substantially more of the UJ MCom students (90.4%) were employed at the time of completing the questionnaire than the KSB MBA students (73.0%). However, nearly all of the students have been employed at some stage, UJ MCom students 96.0% and KSB MBA students 90.9%.

With regard to entrepreneurial experience (Tables 4a and 4b), nearly three times more UJ MCom students (22.0%) compared to KSB MBA students (8.0%) were running a business. Similarly, more than three times as many UJ MCom students (47.6%) had tried to start a business before, compared to the KSB MBA students (15.9%). This, in spite of the fact that approximately the same percentage of KSB MBA students (54.5%) has a family member running a business as the UJ MCom students (56.6%). In addition, the KSB MBA students had more friends running a business (88.6% to 72.3%) and knew more other entrepreneurs (96.6% to 87.7%).

The heads of the departments at UJ and KSB granted permission for the study to be conducted. The data collection took place in a classroom setting. The purpose of the study was verbally explained to the students who were informed that participation was voluntary and confidentiality and anonymity was guaranteed.

The Liñán and Chen (2009) validated EIQ adapted by Malebana (2012) was used. The latter added questions to assess the effect of university education and governmental support programmes on EI. The list of governmental support programmes was deleted from the EIQ as it did not provide meaningful information.

Section A of the EIQ collected demographic information. Section B assessed ‘experience and entrepreneurial knowledge’ (7 questions). Section C focused on ‘entrepreneurial intent’ (9 questions), Section D on ‘attitude towards becoming an entrepreneur’ also known as ‘personal attitude’ (6 questions), Section E on ‘perceived behavioural control’ (9 questions), Section F on ‘entrepreneurial support’ (5 questions), Section G on ‘social capital’ (SC) (15 questions), Section H on ‘entrepreneurial self-efficacy’ (ESE) (24 questions) and Section I on ‘entrepreneurial competencies’ (4 questions). A nominal scale was used for questions in sections A and B, whilst questions in sections C, D, E, F, G, H and I used a five-point Likert-type scale (for C, D, E, F & G: 1 = strongly disagree, 5 = strongly agree; for H & I: 1 = very little confidence, 5 = very confident). The final instrument consisted of 82 questions.

The research analysis was completed in two phases. In phase one the applicability of the EIQ scales to samples from the two developing countries was assessed (Swanepoel & Nieuwenhuizen 2014). EI (C) is the independent variable with the antecedents (D, E, F, G, H & I) the dependent variables. For each of the seven scales (C to I) of the EIQ, the properties (construct validity and internal consistencies) were evaluated using unrestricted exploratory factor analysis (EFA), using SPSS 20 (SPSS 2012). For a detailed discussion of the seven factor analysis see Swanepoel and Nieuwenhuizen (2014). The factors that emerged were mostly different from the existing EIQ, with the exception of the first two factors, namely entrepreneurial intent and attitude to becoming an entrepreneur (personal attitude). The means instead of the interquartile range was used as the data contained few outliers (mostly fewer than four). The results appear in Table 5. The reliability was acceptable for all the factors with a Cronbach's alpha coefficient (α) greater than 0.7 (Nunnaly 1978) (Table 5), except G1.4.

The items relevant to each factor appear in Table 6.

Independent samples t-tests for equality of means were conducted on the newly created factors (included in Table 5 and Table 6) to assess the differences between the two independent groups, master's students from UJ and the KSB.