Investigating the effect of location , specifically shopping centres , on franchisees ’ entrepreneurial orientation : A cluster analysis

Franchisees play an important role in franchise systems, as they are responsible for managing their business and ensuring its overall performance. Since earlier research confirmed the relationship between entrepreneurial orientation (EO) and performance, this study explores the effect of location on the five dimensions of EO for franchisees. This study assessed franchisees, their risk taking, innovation, pro-activeness, competitive aggression and autonomy, and investigated location clusters related to this. Location was investigated from a geographical and operational stance. A face-to-face survey was conducted among 104 franchisees operating in four major cities in the retail sector in South Africa. Two clusters were identified: ‘operating inside of shopping centres’ (50%) and ‘operating outside of shopping centres’ (50%). Contradictory to what was found in the literature, the results indicate that a franchisee is likely to be more entrepreneurially orientated when operating outside of a shopping mall in the Johannesburg region. The strongest and weakest EO dimensions in terms of cluster construction were identified. Furthermore we found that, since the EO dimensions are highly correlated, it is likely that a franchisee will experience high levels of all of the dimensions simultaneously and therefore share in the positively associated benefits of EO, such as growth and performance.


Introduction
The quest for South Africa to create employment, stimulate growth and foster young and growing businesses has been a difficult journey with mixed results (Barnard, Kritzinger & Krüger, 2011: 112).Franchising has become a thriving source of new businesses and the concept of franchising has grown phenomenally worldwide (Emerson, 2014: 456).The United States of America (USA), for example, has over 3 000 active franchise systems, with 901 093 franchisees employing approximately 18 million people, which in turn generates an economic output of over $2.1 trillion.This is a significant contribution to the American GDP and accounts for about 40.9% of the retail sector (Dant, Grünhagen & Windsperger, 2011: 253).
A great deal of research has focused on the relationship between the franchisor and the franchisee (Labrinidis & Jagadish, 2012;Norton, 1988;Sharma, 2013;Weaven, Grace, Dant, & Brown, 2014).Dada and Watson (2010: 3) describe franchising as a system faced with opposing forces of standardisation and uniformity versus innovation and adaptation.However, Kauffman and Dant (1998: 5) infer that there is room for the franchisee to be innovative and to show some degree of EO, which would separate the franchisee from a typical manager.According to Kickul and Gundry (2002: 85), the drive to find answers in the entrepreneurial domain has grown significantly and has also changed direction from looking at entrepreneurial traits, to activities, and now to behaviours.It is this investigation of behaviours that has led to explore the effect of location on the five dimensions of EO for franchisees.Dada and Watson (2013: 790) revealed that EO is significantly and positively related to the performance outcomes of franchise systems, both from a financial and non-financial perspective.This notion is well established in the literature by scholars such as Zahra and Covin (1995) and Wiklund and Sheperd (2005).According to the authors, franchisees with high levels of EO are thus likely to be more productive, which can lead to a more profitable franchise system.
Many scholars focus their efforts on the EO of franchisees (Kauffman & Dant, 1998;Dada & Watson, 2012;Dada & Watson, 2010;Grünhagen, Wollan, Dada & Watson, 2014;Ketchen, Short & Combs, 2011).However, earlier work on EO focused only on one or two of the EO dimensions (Lumpkin & Dess, 1996: 429;Dada & Watson, 2010, 2012, 2013;Hughes & Morgan, 2007: 651) and not on all five dimensions of EO as introduced by Lumpkin and Dess (1996).The five dimensions of EO are: innovation; risktaking; pro-activeness; competitive aggression; and autonomy (Lumpkin & Dess, 1996;Frese, Lumpkin, Rauch & Wiklund 2009: 761).In addition, Ghosh and Craig (1991: 466) emphasise that, in highly competitive businesses such as franchise systems where the product or service offering is quite similar, franchises can obtain significant competitive advantage by making accurate locational decisions.Noor and Sarker (2015: 102) concur and established that location, brand image, service quality, price, hygiene factor, product quality, interior advertisement and sales promotion all affect the preference of consumers toward franchise organisations.Limited research investigates the effect of location on the EO of franchisees, specifically those within shopping malls to those outside of shopping malls (operational location).In this paper, the terms 'shopping centres' and 'shopping malls' are used interchangeably.According to Muller (2008: 41), there has been a boom in the number of and the size of shopping malls, which has yielded thousands of stores within these retail space giants.Prinsloo (2016: 1) agrees and reports: "There a strong drive from international brands that are very interested in South Africa.There are a number of international retailers that have identified South Africa as an important and popular retail destination.South Africa has about 1 785 shopping centres and is the sixth largest in the world as far as number of shopping centres are concerned.We are just behind the USA, Japan, Canada, the UK and China."Since there are various types of stores, as well as different activities and initiatives, this paper specifically focuses on franchisees operating food franchise systems inside and outside of shopping malls across four major cities namely Johannesburg, Pretoria, Durban and Cape Town within South Africa (geographical location).
The core research problem is to determine whether franchisees within different geographical locations differ in terms of their EO.We specifically want to understand whether shopping mall franchisees have higher levels of EO than those that operate outside shopping malls.Furthermore, the paper seeks to determine whether the five EO dimensions are correlated for franchisees.Therefore, the research aim is to determine whether franchisees can be grouped into clusters that are differentiated by three variables namely: (1) the EO dimensions; (2) the city that the franchisee operates in; and (3) whether the franchisee operates inside or outside a shopping mall.The paper additionally explores which of the five EO dimensions have the most prominent role in the construction of the clusters.This is done firstly by providing theoretical support for the concepts above, followed by the development of hypotheses.
The contribution of the research lies in illuminating the significance of location to the EO of franchisees.The decision where to locate a business has consequences for the level of EO of the franchisee.The study does not conclude that there are no benefits to a franchisee being within a shopping mall, but rather that the franchise system is more likely to be associated with higher levels of EO if they operate outside of shopping malls.The importance of the variables included in the construction of the cluster is highlighted, whereby risk-taking is identified as the variable with the most prominent role in the formation of the clusters.In practice this means that franchisees who have more freedom to take risks tend to have a higher level of EO overall.

Theoretical
foundation and hypotheses development Frese et al. (2009: 763) note that more than 100 EO studies have been conducted, which has led to a wide acceptance of the conceptual meaning and relevance of the concept.Entrepreneurial orientation (EO) is defined by Lumpkin and Dess (1996: 429) as the organisational processes, methods, styles, practices, and decision-making activities employed by entrepreneurs that lead to new ventures.Prior research on EO clearly divides the concept of EO into two approaches: (1) a Uni-dimensional approach, and (2) a Multidimensional approach.The uni-dimensional approach is supported by Miller (1983) and Covin and Slevin (1989), who suggest that the dimensions underlying EO work concurrently, such that a firm is entrepreneurial to the extent that it scores high on these dimensions collectively (De Clercq, Dimov & Thongpapanl, 2013: 510).The multidimensional approach is supported by Lumpkin and Dess (1996), who indicate that the EO dimensions operates individually and should be treated as separate constructs.In this paper we adopt a multi-dimensional approach, whereby individual EO dimensions are measured on franchisees.We do, however, explore the uni-dimensional approach by testing the correlation between the individual EO dimensions.
Businesses are embracing EO in order to drive their organisations towards success.Indeed, technology, social media and more effective communication create a business environment that is more competitive and fast-paced.As a method of responding to this competitive business environment, there is an increasing interest in firm-level entrepreneurial efforts.This is perhaps even more so the case for franchise systems, where competitiveness for similar products and services are fiercer.Dant et al. (2011: 254) suggest several important factors when studying franchising, including: diffusion of innovation; knowledgetransfer mechanisms; entrepreneurial orientation; turf issues; interdependencies; and autonomy.For most of the studies conducted on franchise systems only three of the five EO dimensions were measured, namely innovation, risk-taking, and pro-activeness (Maritz & Nieman 2006;Bolton & Thompson, 2003;Dada & Watson, 2012, 2013;Grünhagen et al., 2014;Chien, 2014).However, Wiklund and Shepherd (2005: 75) suggest that EO comprises five dimensions namely, innovation, risk-taking, pro-activeness, competitive aggression, and autonomy.These five dimensions of EO are supported in the literature by Lumpkin and Dess (1996); Frese et al. (2009: 761), as well as Richard, Barnett, Dwyer and Chadwick (2004).In this paper, we review all five of the EO dimensions in a franchise context.

Determining the entrepreneurial orientation (EO) of franchisees
Franchising literature has focused heavily on franchise system relationships (Strutton, Pelton & Lumpkin, 1993;Spinelli & Birley, 1996;Chiou, Hsieh & Yang, 2004;Antia, Zheng & Frazier, 2013), specifically the effect of EO on the franchise relationship (Dada andWatson, 2012, 2013).Grünhagen et al. (2014: 829) emphasise that disagreement exists regarding the extent to which franchisors really want their franchisees to be entrepreneurial.While several studies concluded that EO enhances firms' performance (Lumpkin & Dess, 1996Frese et al., 2009: 762;Wiklund, 2006: 37;Lindsay & McStay, 2004;Chien, 2014), Wiklund and Shepherd (2005:71) found that not all entrepreneurial efforts result in enhanced firm performance.While some of the inconsistency in results may be due to differences in research methods, there is enough evidence to prove that EO may sometimes contribute to improved performance.Recently, human resource (HR) operational autonomy was added as a research avenue to the EO-performance link in franchise systems (Grünhagen et al., 2014), as well as franchisor resources, EO and performance in a coupleowned franchise outlet (Chien, 2014).Although some research has been conducted on EO and the franchise system (Lindsay & McStay, 2004;Maritz & Nieman, 2006), specifically measuring the EO of franchisors (Dada & Watson, 2012, 2013;Grünhagen et al., 2014), limited research is available regarding EO of franchisees.In this paper we answer the call for research regarding all five EO dimensions, measured on franchisees.

Innovativeness and franchisees
Innovativeness captures a bias towards embracing and supporting creativity and experimentation, technological leadership, novelty, and research and development of products, services and processes (Hughes and Morgan, 2007: 652).This includes developing new products, services, operational techniques, technologies, and practices for improved efficiency.Stanworth, Healeas, Prudy, Watson and Stanworth (2003) state that while franchisors' demonstrate innovative behaviours in their decision to expand their firms through the use of franchising, franchisees desire entrepreneurial activity to enable local market adaptations and generate innovations and become an important source of innovative behaviours (Ketchen et al., 2011: 585).Prior studies have shown that franchisees, as a result of their daily customer interaction, are a major source of innovative ideas in the franchise system (Cox & Mason, 2007;Dada & Watson, 2010).Flint-Hartle and de Bruin (2011: 70) agree and found that almost all of the franchisees in their study described innovative ways of doing business alongside franchisor innovative system processes.

Risk taking and franchisees
Risk-taking focuses on the element of uncertainty and encompasses a firm's tolerance for and rewarding of uncertain projects, together with its reliance on novel procedures and methods (Miller, 1983).The risks associated with entrepreneurial ventures may include moving into new and unfamiliar markets, committing substantial resources to ventures with vague outcomes, and also incurring significant debt while pursuing these opportunities.Dada and Watson (2012: 4) found that both franchisors and franchisees have moderately high levels of risk-taking tendency.Ketchen et al. (2011) argue that the franchisor's risk taking is displayed in their attempt to build a franchise system while franchisees run the risk of introducing the franchisor's concept into new and untried markets, and also risk resources devoted to the development of the local markets (Kaufmann & Dant, 1998: 6).
Pro-activeness and franchisees Lechner and Gudmundsson (2014: 41) state that proactiveness anticipates competitive moves and maintains first-mover advantage; an important factor for differentiation.A firm that is proactive is characterised by a tendency to seek opportunities, foresight, introducing new products or services ahead of the competition, and acting in anticipation of future demand to create, change and shape the environment (Dada & Watson, 2010: 4-6).Furthermore, the decision to franchise itself could be seen as evidence of pro-activeness, given that franchising is a technique to bring together resources so as to rapidly create large chains and gain first-mover advantage (Michael, 2003: 76).Dada and Watson's study (2011) also suggested elements of proactiveness in the franchisees' actions in an attempt to be the leaders in their local marketplace.

Competitive aggressiveness and franchisees
Competitive aggression is the intensity of a firm's efforts to outperform competitors, ambitious market share goalsetting, or aggressive actions such as price cutting (Lumpkin & Dess, 1996).Cox andMason (2007: 1070) reveal that franchisees could act autonomously in response to local conditions by setting prices competitively and implementing local marketing campaigns.Vroom and Gimeno (2007: 901) agree and state that in some situations, firms may delegate competitive decisions to franchisees, while their competitors delegate these same decisions to franchisors.Conlin (2002) found that within a given franchise chain, competition between franchised units is fiercer than it is between company-owned units, because franchisees do not consider how their pricing behaviour affects demand for other units.

Autonomy and franchisees
On the one side, franchisors transfer autonomy to franchisees in expectation of innovative results in the franchise system (Kuratko & Hodgetts, 2004) and on the other side; franchisors want to protect their franchises system from change.Maritz and Nieman (2006) concur and suggest that franchisees do exhibit autonomy in certain situations, despite the constraints of the franchise system (Lindsay & McStay, 2004).The results of the study by Dada and Watson (2012: 15) reveal that franchisees, as stewards of the franchise system, welcomes greater autonomy and that they enable a true entrepreneurial partnership between the franchisor and the franchisee.Although Cochet and Garg (2008: 137) point out that franchisors need to ensure that where franchisees are granted autonomy there are strong incentives in place to ensure appropriate behaviour.Frese et al. (2009) confirm that the five salient dimensions of EO usually show high interrelationships with each other.This notion is supported by Bhuian, Menguc, and Bell, (2005); Richard et al. (2004); Stetz, Howell, Stewart, Blair and Fottler, (2000) and Tan and Tan, (2005).In this paper we contribute to the franchising literature by determining whether this fact holds true for franchisees and not S. Afr.J.Bus.Manage.2017,48(3)necessarily for franchisors.Thus the following hypothesis is stated: H1: All five EO dimensions are correlated with each other for franchisees.
Location, specifically shopping centres, and entrepreneurial orientation (EO) of franchisees The choice of enterprise location has been a subject of research interest since the seminal work on "Theory of Location of Industry" (Weber, 1910).Furthermore, research about the location decision has been a strategic focus since the 1930s and 1940s (Reynolds & Wood, 2010).According to Hsu (2013), location plays a vital role in the success of certain types of franchises.The choice of location also influences the success of entrepreneurs, where entrepreneurs who choose to locate their businesses close to their home area tend to outperform their counterparts who are located further from their home area (Dahl & Sonerson, 2012).Gikonyo, Berndt and Wadawi (2014: 444) identified location as the most vital part to success in franchised restaurants.The most important benefit that can be gained from a prudent location decision is improved performance (Huff, 1963 andBarnard et al., 2011: 124).As established in the literature, EO is significantly and positively related to performance outcomes (Rauch, Wiklund, Lumpkin & Frese, 2009: 761;Tajudin, Aziz, Mahmood & Abdullah, 2014: 224;Engelen, Gupta, Strenger & Brettel, 2015: 1069).If location decisions are linked to improved performance and performance is linked to high levels of EO, there ought to be a positive relationship between the choice of location and levels of EO.However, this could not be established in the literature.Taatila (2013: 72) and Eggers, Hansen and Davis (2012: 219) found that strategies with a focus on convenient locations for entrepreneurs are not related to levels of EO.Alegre and Chiva (2013: 499) furthermore found that, by looking at a firm's performance through EO, control variables such location had a non-significant influence on the levels of EO.Access to markets is another vital consideration with regards to location decisions for entrepreneurs (Kimelberg & Williams;2013).The development of shopping centres created a highly competitive market in communities by being the focal point for retail activities in a specific community (Lee, Johnson, Gahring & Lee, 2008).Handa and Grover (2012) suggest that the accessibility to customers shopping centres provide make them attractive locations for entrepreneurs.The close proximity to competitors inside a shopping centre increases competitiveness amongst businesses (Rogerson, 2011: 320;Teller & Alexander, 2014: 28).Therefore, competitive aggressiveness appears to be the only dimension which indicated that competitiveness might suggest a slight increase in the levels of EO.
After conducting an extensive literature review by using the following keywords: "Entrepreneurial orientation AND location"; "Entrepreneurial orientation AND shopping centres"; "Entrepreneurial orientation AND location decision"; "Entrepreneurial orientation AND location choice"; "Entrepreneurial orientation AND operational location"; and "Entrepreneurial orientation AND risk taking"; "Entrepreneurial orientation AND innovation"; "Entrepreneurial orientation AND pro-activeness"; "Entrepreneurial orientation AND competitive aggressiveness"; and "Entrepreneurial orientation AND autonomy", no significant positive relation could be found between location decisions and levels of EO, except for a limited number of studies (Rogerson, 2011: 320;Teller & Alexander, 2014: 28) which showed that there might be link between competitive aggressiveness and a location inside a shopping centre.As far as can be established, no previous studies have investigated the relationship between the levels of entrepreneurial orientation of franchisees, particularly inside shopping malls.Based on the above, the following hypothesis is stated: H2: The choice of location, specifically being inside or outside of a shopping centre, does not have an influence on the levels of EO for franchisees.

Methodology
In this quantitative research study, simple random sampling was employed to food franchisees in shopping malls, sampling a minimum of four franchisees per mall in four different cities.A total of 24 shopping malls were identified, which contribute to an average of six shopping malls per city.In addition, this study targeted a sample of 15 franchise systems within each of the four cities that are not within a shopping mall.The target sample size to test the hypotheses was 120 respondents but 16 were wasteful therefor the sample size was 104. Figure 1 illustrates that 55 franchisees were located inside shopping malls and 49 outside shopping malls.A non-probability convenient sampling technique was employed, since every food franchisee in the shopping mall had a chance to have been selected to participate in the survey.

Food franchisees in the sample
Often, researchers have to accept a sampling frame that includes people or cases beyond those in whom they are interested (Cooper & Schindler, 2008: 384).Fortunately, this can be easily resolved by a sample that is drawn from the larger population with a screening procedure in place to eliminate those who are not members of the group the researcher wishes to study.In this paper, respondents who were not franchisees were excluded from the sample.The researchers understood the diligence necessary when designing a sample and executing the sampling process.In Figure 1, the dark grey areas show the selected sample from each of the four cities, between 13-16 inside shopping malls and between 10-13 outside shopping malls.

Hypotheses testing
Cluster Analysis is used as an exploratory method to segment the data into natural groups.According to Anderberg (1973: 372), cluster analysis is a collective term covering a wide variety of techniques to delineate natural groups or clusters in data sets.Bratchell (1989: 106) defines a cluster as a group of similar objects and further defines cluster analysis as a term applied to a number of techniques that seek to divide a set of objects into several groups or clusters so that the objects within the group are more similar to each other than objects in different groups.There are three basic requirements in order to use cluster analysis: firstly, there must be a measure of similarity or dissimilarity between the objects in the data set; secondly, there must be a technique or algorithm for forming clusters; and lastly, a means to decide when a group of objects indeed represent a cluster.In this paper, a two-step clustering analysis, using SPSS was used to determine if clusters are present in the data set.Bratchell (1989: 106) suggests that in order to perform cluster analysis, variables must be selected according to relevant criteria as set out by the researcher.For the purpose of this paper the five dimensions of EO is used as variables, as well as the city of operation (geographical location) and the location of the franchisee in terms of being inside or outside a shopping mall (operational location).The five dimensions of EO are considered to be continuous variables, while geographical and operational locations of the franchisee are categorical variables.

Reliability and validity
Cronbach's Alpha indicates the degree to which a set of variables measure a single one-dimensional latent construct.
The alpha value ranges from zero to one, with a minimum reliable coefficient in the region of 0.6 (Cooper & Schindler, 2008: 293).A Cronbach's Alpha value was calculated for the each of the five dimensions of EO and a value of 0.8 or higher was considered to indicate reliability.The five dimensions were presented by grouping similar and related questions relating to that factor together.The Cronbach Alpha value for innovation is 0.96, for risk taking and proactiveness 0.94, for competitive aggression 0.99, and for autonomy 0.96.All five EO dimensions showed internal consistency between the grouped questions.This implies that the questions regarding the five EO dimensions are reliable.

Demographics of the sample
In terms of the number of years operating the franchise inside a shopping mall, the results were relatively evenly spread between one and 12 years.The franchises that were outside shopping malls have mostly been in operation between three and six years, with 61.23% of respondents falling within this category.The average years of those franchises within shopping malls are 7.25 years, with a standard deviation of 3.34 years.The average years of those franchises outside shopping malls are 5.61 years, with a standard deviation of 2.66 years.

Correlation of the EO dimensions
Correlation coefficients were analysed for the sample set of data to determine if any correlation patterns exist.Table 1 reflects that the five dimensions are all relatively closely correlated.The two dimensions that have the highest correlation are risk-taking and competitive aggression with a correlation coefficient of 0.93383.The two dimensions that are least correlated are pro-activeness and competitive aggression with a correlation coefficient of 0.77215.Competitive aggression implies that the owner or manager has a deep understanding of his/her competitors and being proactive would imply a level of strategic intent to capitalise on that knowledge.It is therefore interesting that these two dimensions are the least correlated.The dimensions are correlated, hence the null hypothesis can be accepted and it can be concluded that the five dimensions of EO are correlated in a statistically significant way.This is supported in the literature by Frese et al. (2009), Bhuian et al. (2005);Richard, et al. (2004); Stetz et al. (2000) and Tan and Tan (2005).
The implication of this finding is that the dimensions are typically found together and at similar levels for franchisees.This implies that, when a franchisee develops one of the dimensions, it would typically have the EO mind-set to develop the other four.
Since EO is essentially associated with all five dimensions, it is clear that franchisees will likely be proven to have high dimensions of EO or none at all.The same argument holds true: if a franchisee has a low score for one of the dimensions, then it will likely have a low score for the other four as well.

Cluster analysis
A two-step cluster analysis was used to establish if groups of franchisees with a higher mean value for the five EO dimensions could be segmented both into franchisees that operate inside or outside a shopping mall and by city of operation.It was found that two clusters could be formed whereby the cluster with a higher mean value for the five dimensions consisted of franchisees operating outside shopping malls and situated in Johannesburg.The cluster with a lower mean value for the five dimensions of EO consisted of franchisees operating inside shopping malls and situated in Pretoria.
In Table 2 it is evident that cluster analysis produced two clusters with an equal amount of 52 records in each cluster.
The quality of the cluster, as shown in Table 2, is 0.5, which can be considered as good.The results of the cluster analysis are shown in Table 3.The table shows the means of each of the five dimensions for each cluster.Table 3 also indicates dominant values for each of the categorical variables.The following key points are evident from Table 3:  Risk-taking was the most important predictor (value = 1) in terms of cluster construction, followed by innovation (value = 0.85), whilst competitive aggression (value = 0.53) was the least important predictor of the five dimensions. Cluster Two had the highest mean for each of the five dimensions, therefore indicating an overall higher level of EO for that group of franchisees than for Cluster One. The higher level of EO (Cluster Two) consists of those franchisees outside shopping malls situated in Johannesburg and that tend to have been in operation for a shorter period of time.The remaining three cities had relatively smaller values in this cluster with values less than 3%.
 The lower level of EO (Cluster One) consist of those franchisees inside shopping malls situated in Pretoria and tend to have been in operation for longer than those franchisees in Cluster One.

Discussion of the findings
In summary, the research results show that franchisees can be divided into two clusters, which can be differentiated by the means of the dimensions of EO; the city that the franchisee operates in; and the location of operation in terms of being inside or outside a mall.Therefore, the null hypothesis (H2): 'The choice of location, specifically being inside or outside of a shopping centre, does not have an influence on the levels of EO for franchisees', is rejected.
The implication of this finding is that there is a significant difference in EO perspective for franchisees that operate outside a shopping mall compared to those that operate inside a shopping mall.Literature suggests that, although shopping centres create a highly competitive market in communities and is a very attractive choice of location for entrepreneurs (Lee, Johnson, Gahring & Lee;2008;Handa & Grover, 2012), this paper has in fact shown that, according to the cluster analysis, the franchisees that operate outside shopping malls are associated with higher mean values for each of the five dimensions of EO.The cluster analysis also indicate that Johannesburg is the city with higher mean values of the EO dimensions, implying that the city in which a franchisee operates affects its ability to be entrepreneurially orientated.
Franchisees outside shopping malls in Johannesburg are associated with higher innovation, more calculated risks, a more proactive management style, more competitive aggression, and a high degree of leadership autonomy.Furthermore, risk-taking is the variable that played the most prominent role in the construction of the cluster.The group that fell into the "outside of a shopping centre" cluster tend to be less risk adverse than the group "inside of a shopping centre".Competitive aggression played the least prominent role in the construction of the cluster, which is contradictory to the findings of Rogerson (2011: 320) as well as Teller and Alexander (2014: 28) who found that competitive aggression was the only dimension that showed a link between shopping centres and the levels of EO.

Conclusion
Shopping malls remains a popular location choice for many South African entrepreneurs.The number of shopping malls in South Africa has increased, and so the number of franchisees operating within shopping malls.It is important to understand the benefit of operating within a shopping mall, particularly whether they provide any benefit to franchisees.If shopping malls provide more benefits, malls should potentially be promoted at the highest levels, as they could provide greater success for franchisees and hence stimulate job creation and growth in the retail sector.
The theoretical foundation of the literature provided insight into a construct called entrepreneurial orientation (EO).It is well documented that EO has been linked to superior performance of entrepreneurial businesses.The core research problem of this paper has therefore been to understand whether all EO dimensions are correlated for franchisees.In order to achieve this, the construct of EO had to be broken down into its five dimensions, which include innovation, risk-taking, pro-activeness, competitive aggression, and autonomy.The study also aimed to understand whether the choice of location, specifically being inside or outside a shopping centre, has an influence on the levels of EO for franchisees.
The results firstly analysed the correlation of the five EO dimensions with each other, for franchisees.Secondly, the existence of clusters of franchisees that can be differentiated by the city of operation and also the franchisee's location, be it inside or outside a shopping mall, were investigated.Using cluster analysis, the franchisees were grouped into two clusters.The cluster that had higher mean values for the five EO dimensions consisted of franchisees outside shopping malls in Johannesburg.The cluster that had lower mean values for the five EO dimensions consisted of franchisees inside shopping malls in Pretoria.The implication is that there are indeed groupings of franchisees to suggest that EO dimensions differ for franchisees inside and outside shopping malls and also across cities.It was found that the five dimensions are statistically correlated.While shopping malls do provide a physically different environment for franchisees inside a shopping mall compared to those outside a shopping mall, such as sheltered walkways, parking lots, cleaning of common areas, lavatory facilities, and security, the study provides evidence that being situated in a shopping mall does not provide a benefit in terms of EO.The paper does not conclude that there are no benefits to a franchisee being in a shopping mall, but rather that operations outside shopping malls are more likely to be associated with higher levels of EO.Additionally, every city in South Africa may have differences in terms of infrastructure and the primary focus of its respective economies; thus cluster analysis shows more EO for franchisees in Johannesburg.

Limitations and areas for further research
No study is without limitations.There is a limited amount of research within the South African context on the relationship between franchisees and shopping malls; further research is necessary to determine the value this relationship holds for entrepreneurs and franchisees.The study used large shopping malls only, while in South Africa there are various different sizes of shopping malls, from small convenience centres, strip malls to super regional shopping malls.The results could differ for the different sizes of shopping malls.The location of franchisees outside shopping malls was mainly central business district areas, while there are significantly more franchisees in suburban areas.It would be interesting to review the results if franchisees from other business areas are incorporated.
As entrepreneurship is linked to growth and employment (Frese et al., 2009: 762), it is clear that future research in EO would be beneficial to the South African economy.The following recommendations and possibilities for future research are suggested:  The paper only included four cities (Johannesburg, Cape Town, Durban and Pretoria).Future research could include other large cities, such as Bloemfontein, East London, Port Elizabeth and Polokwane, amongst others.
The study could go further by comparing the large cities to some of the smaller growing cities in South Africa, such as Richards Bay, Port Shepstone, Kimberly and Witbank, amongst others. Further research could be done on the EO dimensions of franchise stores in different sizes of shopping malls.South Africa has many different sizes of these centres and there are more being built.It would be beneficial to understand the dimensions of EO in stores in the smaller centres (fewer than 10 stores) compared to the larger super regional shopping malls (over 250 stores). There could lie great value in understanding the dimensions of EO and then linking these to the performance of the franchisees within the shopping malls.It could then provide evidence that EO leads to improved performance of franchisees within shopping malls.In addition to performance, it would be of significant value to South Africa to understand the ability of franchisees within shopping malls to generate and sustain employment.

Figure
Figure 1: Summary of sample design