Entrepreneurial Intention: Theory of Planned Behaviour and the Moderation Effect of Start-Up Experience Entrepreneurial Intention: Theory of Planned Behaviour and the Moderation Effect of Start -Up Experience

Entrepreneurial activity is considered to be an intentionally planned behaviour. Consequently, entrepreneurial intention (EI) may be evaluated via theory of planned behaviour (TPB). According to Ajzen’s TPB, EI is explained by three antecedents: attitude towards entrepreneurial behaviour, perceived social norms and perceived behaviour control in other words, self‐efficacy. Although this model is widely tested empirically, new research regarding moderation effects may be valuable [1]. Moreover, [2] argues that personal factors such as previous start‐up experience are relevant concerning the model. Accordingly, in this study, moderation effect of start‐up experience is added in TPB model from a convenience sample of 528 undergraduate business administration students from the three most economically developed cities in Turkey. Hypotheses are tested by means of hierarchical multiple regression analysis. Coefficients are estimated using ordinary least squares. In order to test the moderator effect, significance values of the interaction term is assessed. According to the results, all of the relations within the model are significant. Ajzen’s TPB holds for the Turkish case. Moreover, for the stu ‐ dents with a past start‐up experience, the effect of both self‐efficacy and personal atti ‐ tude towards entrepreneurial behaviour on entrepreneurial intensity increases. This is a promising result for the future studies.


Introduction
Entrepreneurship is considered to be the discovery/creation, evaluation and exploitation process of opportunity [3], and it requires the preparedness to realize and/or create that opportunity [4,5]. Within this perspective, entrepreneurial intention is considered to be the best predictor of this behaviour in comparison with other factors such as, demographic and trait variables. This is because entrepreneurship is taken to be an intentionally planned behaviour likewise all other strategic decisions [6]. Accordingly, within the entrepreneurship literature, cognitive research gain considerable popularity and most of this attention is given on intention models [6].
Theory of reasoned action, which is a widely used intention-based theory, explains intention via a formulation considering subjective norms of significant others and personal attitude towards the related behaviour as the antecedents of intention. Theory of planned behaviour (TPB) is an extended form of theory of reasoned action, with the addition of a new variable, perceived behavioural control (PBC) [7]. TPB is used extensively within different study areas besides the entrepreneurship literature, and it still provides a rich potential for the area [8].
In addition to the theoretical importance of TPB, it has also practical importance. According to the model, these perception-based intentions and beliefs may be learnable not inborn [4]. Besides, it is a widely accepted fact that personal differences such as past experience, knowledge, etc. may affect the evolution of the intention [9]. As a result, studying intention may provide policy makers the opportunity to realize the cognitive frameworks of individuals who are considering to engage in entrepreneurial behaviour, which is more beneficial for policy-making purposes compared to those who have already started up a business [10]. This is important because entrepreneurial behaviour acts as an important and locomotive force of innovation within an economy [11].
This paper is following the entrepreneurial intention model of Ajzen's TPB. There are many studies in the literature following this path; however, it is argued that a lot of work is needed for figuring out the factors effecting entrepreneurial intention [12]. Moreover, although this model is widely tested empirically, it is argued that new research regarding moderation effects may be valuable [1]. Additionally, [2] asserts that personal factors such as cognitive short-cuts, self-related concepts and previous start-up experience are relevant concerning the relation between entrepreneurial intention and perceived behavioural control plus personal attitude towards entrepreneurship. One of the most important factors that are proposed to add the TPB is past experience [13,14]. Accordingly, in this study, moderation effect of startup experience is added in TPB model.
Consequently, the aim of this paper is to test the effect of previous start-up experience on the entrepreneurial intention within a revised model of TPB. Recent studies on TPB take prior actions into consideration [15]. However, studies considering for moderation effects [1] and past experience [16] are still asked for. Because it is claimed that the effect of past behaviour is not a direct one, accordingly, it is suggested to add past behaviour as an indirect factor for intention instead of a direct link [17][18][19][20][21]. This addition is important both theoretically and practically. Theoretically, both examining a widely tested model-TPB-within an eastern cultural context is important for the generalizability of the theory.
Moreover, revising the model may also be valuable in order to strengthen the explanation capability of the theory. For this issue, we have rested on the past experience, which is considered to be one of the most important explanatory factors for behavioural intention. Practically, adding the past experience may shed light to practitioners and politicians for new opportunity exploitation activities, such as empowering behavioural experiences of entrepreneurs via initiatives, etc. These issues stand as the theoretical and practical importance of the paper.
Within this study, first TPB will be explained in brief, and then the effect of past behaviour within the TPB model will be discussed. Thereafter, the study held in Turkey, on 528 undergraduate students will be presented. Terminally, conclusion and practical and theoretical implications of the study is initiated.

Theory of planned behaviour
Theory of planned behaviour (TPB) that "predicts and explain behaviour in specific contexts" [7] is a frequently used theory in different disciplines [22]. This is also true for entrepreneurship research since to become an entrepreneur is considered to be a conscious activity and intention is taken to be a cognitive state [22]. Moreover, it is argued that entrepreneurial decision is a complex one and need intentional cognitive process [18]. As a result, instead of personality traits or demographic studies, cognition contains more and significant information regarding the entrepreneurial behaviour, since it is a "closer antecedent" for behaviour [6,23]. The outcomes of the previous work also suggested that theory of planned behaviour is an applicable theory for entrepreneurial research [12]. Consequently, intention-oriented researches within the entrepreneurship literature are gaining popularity.
For infrequent behaviours/unstable contexts (which is true for entrepreneurial context), the explanatory power of intention regarding behaviour is increased [16,[24][25][26]. Because, it is argued that strategic entrepreneurship is taken to be an intentionally planned behaviour, and this is true for even necessity motivated and unexpected entrepreneurship [16]. Therefore, studying the decision-making process for entrepreneurial behaviour via theory of planned behaviour (TPB) seems reasonable [12].
There are two known attitudes for individuals that are intuitive and rational [27], and the main assumption for Ajzen's intention-behaviour relation is human behaviour is rational [28]. Intention means how much a given behaviour is tended to be tried plus how much effort is made for this behaviour [7] and the stronger the intention the chance of the behaviour to be realized is increase [28]. Intention provides a link between the beliefs of an individual and corresponding behaviour [29]. It is considered to be typical for entrepreneurial context, even though the new venture start-up process may evolve suddenly due to an opportunity realization [16].
According to TPB, there are two major sources of intention: desirability (motivation to act for the intended behaviour) and feasibility of the given behaviour [16]. To be more precise, perceived behavioural control (PBC) stands for feasibility; subjective norms and personal attitude towards entrepreneurial behaviour together define the desirability part of the entrepreneurial intention. These are three kinds of conceptually independent [30] beliefs which are behavioural, normative and control beliefs, respectively [31]. Although this model is a generic one that holds across cultures and contexts, the relative importance of the factors may change [7]. In generic form, intention may increase when there is a positive attitude towards the behaviour, subjective norms regarding the behaviour is favourable and individual has a belief that he/she can accomplish the behaviour effectively [32].
Subjective norms (SN) represents the perception of significant others about a given behaviour. The main assumption for adding this factor to the model is the argument that human behaviour is adopted according to other people's attitude towards given behaviour [24]. Although the effect is taken to be effectual across cases and cultures, the significant others differ for different individuals [4]. For instance, for individuals holding a job, the co-workers or other work-related networks are important. On the other hand for students, family and friends may be important. The effect of subjective norms within the model is questioned due to insignificant and non-systematic previous results regarding it. However, it is argued that, when intention is measured appropriately, there appears a strong relation between norms and intention [25]. By definition, regarding this factor, the belief of others is weighted by individuals' readiness and willingness to act according to these beliefs [17]. As a result, it is argued that especially within collectivistic cultures such as Turkey, subjective norms play a positive and important role for explaining the intention [2,6,12]. Accordingly following hypotheses are proposed: H1: Subjective norm positively affects entrepreneurial intention.
Perceived behavioural control (PBC) refers to the "perceived ease or difficulty of performing the behaviour" [7]. PBC is a perception instead of an actual control and can be operationalized via self-efficacy [7]. Self-efficacy (SE) is considered to be an appropriate measure for PBC since both deals with the "perceived ability to perform a behaviour" [7,31,32]. To put it another way, both PBC and SE copes with the perception not the actual skills or abilities [33]. Self-efficacy does not only improve goal setting but also provide persistency for the pre-set goals as a result strengthen the intention [34]. In other words, self-efficacy positively affects various stages of entrepreneurial behaviour [33]. Other factors such as role models effect intention if they affect the self-efficacy [16]. It is argued that, the greater the self-efficacy, the entrepreneurial intention will be stronger. Contrary to subjective norms, self-efficacy is considered to be an important factor for entrepreneurial intention universally [2]. Besides, it appears to be the most powerful antecedent of intention within the literature and, moreover, under the conditions intention explains little regarding the entrepreneurial behaviour, perceived behavioural control and self-efficacy accordingly has an influence for separately predicting the behaviour [25]. H2: Self-efficacy positively affects entrepreneurial intention.
Personal attitude refers to the people's "favourable or unfavourable evaluation of the behaviour in question" [7,35]. In other words, personal attitude explains the "personal" desirability of any given behaviour, in comparison with the subjective norms that refers to the desirability of significant others [16]. According to the model, personal attitude is not an inherent position for an individual but can be learned [17]. As a general rule, the favourable the personal attitude towards a behaviour, the stronger the intention to perform that behaviour.
Insofar, the generic model of TPB is explained and related hypotheses are developed within the entrepreneurial context. Now, the effect of previous start-up experience will be added as a moderator to the generic model.

Effect of previous experience
Learning is an important factor for cognitive theory including Ajzen's framework [24]. Accordingly, it is a very common argument that the past behaviour is a strong predictor for the future behaviour. To be more specific, past behaviour is told to explain additional 13% of future behaviour variance [26] since it serves as the most important "human capital variable" [36]. However, the formulation of how this may be true within an intention-based model is controversial [32]. Past behaviour somehow change the intention [37] and this relation is told to be not linear and unidirectional [38]. There appears to be a two way relation, that means attitudes and intentions influence behaviours and behaviours also influence the intentions back [38].
Effect of past on future behaviour is controversial within the Ajzen's theory of planned behaviour (TPB) also. It is argued that although TPB is a well-formulated model to explain intention, one major weakness of the model is it does not consider past behaviour [26]. However, [7] argues that, the past behaviour is inherently considered within the model since the information gained via past experience is processed on the antecedents of intention. As a result, past behaviour has no helpful value for future behaviour while considering the direct effect of past experience. However, the significant effect of past behaviour for entrepreneurial intention models contradicts this perspective [37].
To put it briefly, regarding the previous experience-intention relation, there are two main approaches [8]: one, which follows Ajzen's path and argues that past behaviour is irrelevant for the intention model since it is already absorbed via the antecedent factors of intention. Second approach, that is started to be adopted widely argues that past behaviour may be added to the model and improve the explanatory power of the model [8]. Accordingly, the past knowledge is checked for the inclusion within the entrepreneurial intention and behaviour framework. This is because it is argued that two kinds of knowledge may be gained from past experience, which are tacit and explicit knowledge [8]. Although tacit knowledge is embedded in the beliefs (i.e. self-efficacy/control beliefs, personal attitude towards entrepreneurial behaviour/ behavioural beliefs and subjective norms/normative beliefs), explicit knowledge may provide more variance that cannot be explained via these beliefs [8]. And if the situation is not stable, due to reasons such as experience, the intention model may not work well.
Moreover, one's cognition depends on his own experience providing him/her information regarding the relevant behaviour and its consequences [39], because, individuals do not realize the worth of all opportunities around them [40]. In other words, different people discover and realize different opportunities, due to the possession of different prior knowledge and experience [40]. Besides, this experience provides the related behavioural skills via affording relevant setting for essential training [32]. Because, direct experience affords different information compared to the indirect one [30]. Related to entrepreneurial behaviour and value creation context, this relation is significant [6]. In addition, these cognitions are more stable compared to the cognitions formed via indirect sources of information [39]. These results support the idea that the previous experience amplifies the relation between entrepreneurial intention and its antecedents, namely self-efficacy and personal attitude towards entrepreneurial behaviour.
It is argued that the relation path that is proposed to be added to the TPB model is not a direct one [6]. This is because the direct effect is absorbed through the antecedents of intention. Prior start-up experience does not affect personality or entrepreneurial potential of the individual, but it changes the "individuals' perceptions about the opportunities available" [15]. As a result, it may be argued that prior experience provides knowledge that amplifies the relation between both personal attitude and self-efficacy to entrepreneurial intention. In other words, with the experience, even if the attitude and self-efficacy levels are same, the cognitive process works different to shape the relation regarding these factors with entrepreneurial intention positively.
For instance, although individual's attitude towards entrepreneurship may be low, due to past experience he/she may knows that if he/she does not realize the opportunity, someone else will do it [8]. As a result, without increase in the attitude, intention may increase due to this "explicit" information gained through past experience. Similar situation may also be hold for the SE. for instance, individual may not be "confident" about his/her control regarding the situation but due to past experience he/she may be aware of resources to cope with the situation [8]. In other words, without an increase in SE, with the knowledge and experience gained through the past behaviour, the intention level may increase. Moreover, with past experience, individuals expend minimal effort for intention although this is not due to an increase in the antecedents [20].
To sum up, past start-up experience may provide individuals this explicit knowledge that is not reflected through the antecedents, and the effect of this knowledge and experience on intention is not a direct one, but a moderator effect exists. As a result, following hypotheses are suggested: H4: With the moderation of start-up experience, the relation between self-efficacy and entrepreneurial intention increases.
H5: With the moderation of start-up experience, the relation between attitude towards entrepreneurial behaviour and entrepreneurial intention increases.

The study
In this study, the effect of past start-up experience is added to the Ajzen's theory of planned behaviour (TPB), as a moderator factor between entrepreneurial intention and personal attitude towards entrepreneurial behaviour plus self-efficacy. Table 1 figure out the proposed model and relations of this study. Now, the empirical analysis of the proposed hypotheses will be realized.

Measures
All of the factors except past start-up experience is measured by seven-point scales that ranged from 1 (strongly disagree) to 7 (strongly agree) in relation to entrepreneurial behaviour. The scales used revealed adequate reliability among undergraduate students in the past studies. For avoiding the response bias, some items for each measure are negatively worded [14].
For the entrepreneurial intention and personal attitude, Ref. [12]'s scales are applied. For entrepreneurial intention six items, for personal attitude five items that ask for the level of agreement is used. It is argued that there is a lack for reliable measures for entrepreneurial intention [41], and this study argues to be statistically robust and theoretically sound for applying different cultures. Related to intention there are three kinds of measures: desire ("I want to…"), self-prediction ("How likely it is…") and behavioural intention ("I intend to…"), and the last one is told to have better results for behavioural prediction [12]. Moreover, for the entrepreneurial intention concept, it is appropriate to measure via a reflective measure not a formative one [41], which is true for this scale. Accordingly, we also use the behavioural intention measure. For attitude, an aggregate scale is used consistent with Ref. [7].
For subjective norms, identification of the appropriate significant others is important [17] and within this study, two questions are formulated that are about the decision of family and friends. These two are considered to be the significant others for undergraduate students. Accordingly, following two items to agree/disagree are asked "My family/friends would see it as very positive if I would start my own business". For self-efficacy, Ref. [42]'s scale, which is widely assessed, is used.  For past start up experience, a direct "yes or no" question asking for "whether they initiate a start-up experience before" is used. This question is considered to be a dummy variable as a moderator in the classic TPB model.
All of the scales have high reliability scores within the past studies. According to Ref. [43], 0.80 reliability level is required. This is also true for this study since all constructs' reliability scores are higher than 0.80 ( Table 2).

Sample
Data are collected from a convenience sample of 528 third and fourth year undergraduate business administration students (232 male, 296 female) from the three most economically developed cities in Turkey: Istanbul, Ankara and İzmir. Using a student sample is not an issue for this study because, the study deals with the entrepreneurial intention of "potential entrepreneurs" [33], which is consistent with our sample.
It is argued that university graduates between 25 and 34 ages are the closest group toward entrepreneurial behaviour, and third and fourth year students which are soon-to-graduate are close to this age group since they are close to their career choice [6,12,16]. This is because they may see entrepreneurial career as a smart option compared to wage employment [11]. Moreover, when the time between intention and behaviour is close, the relation may be more reliable [44] and our sample suits for this issue. This sample is convenient with the previous-related literature [6].

Analysis
Hypothesis are tested by means of hierarchical multiple regression analysis using PASW Statistics 18. Coefficients are estimated using ordinary least squares (OLS). In order to test the moderator effect of start-up experience, significance values of the interaction term is assessed.
Correlation values for the model are presented in the Table 3.
Correlations between the constructs are all statistically significant at the p < 0.01. All of the correlation values between independent variables are ranging between 0.320 and 0.460 that corresponds to low-moderate correlation levels. This level indicates to ignore the presence of multicollinearity [45]. Moreover, VIF and tolerance values are checked in order to look for multicollinearity. For all of the factors, these two values indicate for point out for overlooking multicollinearity.
The model's regression results are given in the Table 4. According to the Entrepreneurship -Practice-Oriented Perspectives R 2 is 0.457. Previous studies find that the generic TPB model explains 30-45% of the variance for intention [46] which is consistent with our results.
In order to test for the correlation between the error terms, the Durbin-Watson value is checked. According to Ref. [46], the Durbin Watson value needs to be between 1.5 and 2.5, for independency of the observations. It is 1.864 for this study.

Results
Findings of regression analysis supported all of our five primary hypotheses. Both the model of Ajzen's theory of planned behaviour (TPB) and the moderation effect of past experience are supported by our sample. In other words, TPB holds for the Turkish case. Besides, the moderator effect of past start-up experience between entrepreneurial intention and both selfefficacy plus personal attitude towards entrepreneurial behaviour holds.
In the model, the most influential factor for entrepreneurial intention is personal attitude towards entrepreneurial behaviour and self-efficacy along with subjective norms follows it.  The effect of personal attitude towards entrepreneurial behaviour and self-efficacy on entrepreneurial intention is valid for most of the past studies.
According to past research, subjective norms are not hold in general and provide complex results. In other words, some of researches the effect is so small, and for some others, it is insignificant. For this study, subjective norms have a significant effect on entrepreneurial intention. This may be due to the collectivistic nature of Turkish culture. Because, within the collectivistic cultures, the perceptions of significant others are important in general. For the entrepreneurial context, it is also the case [2,6,12].
Moreover, the relation between entrepreneurial intention and the attitude towards entrepreneurial behaviour is moderated by start-up experience. Additionally, start-up experience has a moderator effect on the relation between entrepreneurial intention and self-efficacy. In other words, for the students with a past start-up experience, the effect of both self-efficacy and personal attitude towards entrepreneurial behaviour on entrepreneurial intensity increases. We may interpret this result with the existence of explicit knowledge gained through past start-up experience which provide more variance that cannot be explained via antecedents of entrepreneurial intention [8]. As far as our research, this moderation effect is not studied within the relevant literature. This is a promising result for the future studies.

Discussion
Theory of planned behaviour (TPB) is an extensively adapted model in many different research areas. Besides, there are a lot of studies that tries to modify TPB via adding new antecedents or moderation/mediator effects. Regarding these factors, past behaviour is argued to be one of the most important ones that contain information about the intention. Moreover, although TPB is widely tested empirically, new research regarding moderation effects may be valuable [1]. Accordingly, within this study, the moderation effect of past experience within the TPB model is evaluated. This issue stands for the theoretical importance of this paper.
The results figure out that the model of theory of planned behaviour is supported via Turkish sample. Moreover, the moderation effect of past start-up experience between entrepreneurial intention and personal attitude towards entrepreneurial behaviour is also supported. For instance, without any increase in individual's attitude towards entrepreneurship, with past experience he/she may identifies that if he/she does not realize any opportunity, he/she may lost it forever since someone else will do it [8]. As a result, without increase in the attitude, intention may increase. This happens due to the experience and "explicit" information gained through past experience.
The similar moderation result is gained for past start-up experience between entrepreneurial intention and self-efficacy. Accordingly, although individuals are not positive about her/his control over any given situation, past start-up experience may provide the awareness of the resources to cope with the situation [8]. Similar to the attitude towards entrepreneurial behaviour, the intention level may increase without any increase in the self-efficacy levels. This also happens due to the knowledge and experience gained through the past behaviour.
As far as our research, with the addition of the past experience as a moderator between the antecedents and the entrepreneurial intention, this study is within the first studies in the entrepreneurial intention literature. These studies are important due to the fact that intention is the most representative antecedent of entrepreneurial behaviour and in order to understand this behaviour specifying the effective factors is essential. As a result, it is hoped that this may be a positive contribution for the relevant literature. These are promising results for the future studies. Additional study is needed in order to sustain the generalizability of the results. More studies within different cultures are also appreciated. Because, our results may be due to cultural factors and as a result for different cultures, they may not hold. Besides, similar studies with non-student samples would also be fruitful in order to check the results for different sample groups.
General limitation for the intention-based models of behaviour is hold within our case also. In other words, intention may not always end up with behaviour and sometimes even it does, there may be a significant time lag [10]. Accordingly, some longitudinal studies are welcomed to check for intention-behaviour relation [32].
Besides the academic implications, there are also some practical implications for this paper.
The first one is regarding the intention-based nature of this study. For public policy makers, the initiatives just affect intention and its antecedents will be helpful for new business formation [16] in order to decrease the perceived barriers for students [11]. This is because, in order to change behaviour, one needs to change the intention first [35]. Besides, the perception of the potential and existing entrepreneurs is more important compared to the reality [47]. As a result, intention-based studies are helpful for the practitioners.
Moreover, according to TPB, both self-efficacy and attitude towards behaviour may be developed and learned via formal education or experience [17]. Accordingly, courses for increasing self-efficacy and emphasizing the advantages of entrepreneurship would be helpful for increasing the entrepreneurial intention [48]. Because it is argued via empirical studies that courses regarding entrepreneurship and skills related to entrepreneurial behaviour encourage undergraduate students for entrepreneurial career [11].
Besides, with the addition of past experience effecting the self-efficacy/entrepreneurial intention relation, practically oriented courses that may be helpful for students to gain entrepreneurial experience may also empower the self-efficacy and entrepreneurial intention [49].
Because it is argued that, within the TPB, the explorative power of direct experience is greater compared to the indirect one [30]. Further, since observing may also increase the self-efficacy [5], internship practices may also be helpful regarding entrepreneurial intention.
Above and beyond, since past start-up experience indirectly effect the entrepreneurial intention, initiating a new business may also be supported via credits and training via public policy makers. Because, supporting potential and existing entrepreneurs will be helpful [4] according to our model since past experience increases the intention.
To sum up, this study following the antecedent studies has both theoretical and practical implications regarding the entrepreneurial intention literature and to be more specific studies based on the theory of planned behaviour model. Hence more following researches are welcomed.