A longitudinal study of the entrepreneurial intentions of university students.

Audet, Josee

ABSTRACT

This paper presents a longitudinal study of entrepreneurial intentions among university students enrolled in a business administration program. Data was collected twice: first during the last semester in school and then 18 months later, once students had graduated and started to work on a full-time basis. Results confirm that the perceptions of the desirability and feasibility of launching a business significantly explain the formation of an intention to go into business on a long term horizon. However, the model does not hold as well when the time frame is shorter. Indeed, both perceptions fail to explain to a significant degree the intention to go into business. When work satisfaction is added to the model, both this variable and perception of feasibility become significant predictors of short term intentions, perception of desirability remaining non significant. Our results also tend to indicate that entrepreneurial intentions and perceptions vary over time. As the temporal stability of intention is a condition for an intention-based model to accurately predict behavior, the link between entrepreneurial intentions and actual venture creation may prove difficult to establish. Data collected in the following phase of the study should provide more information about the temporal stability of variables.

INTRODUCTION

If there is one question that has not been answered satisfactorily, it is that of whether it is possible to identify the people within a society who will eventually go on to start their own businesses. This question is of considerable interest. Let us imagine for a moment that it is possible to identify the individuals in a given group who have what it takes to succeed in business, or whose profile is consistent with the composite profile of the successful entrepreneur. Financial institutions would find it considerably easier to decide who qualifies for a loan and who does not, simply by administering a test to determine whether the applicant meets the "successful entrepreneur" criteria. Similarly, "future entrepreneurs" could be identified as soon as they entered the educational system, and directed towards an academy of entrepreneurship whose mission would be to provide suitable training, in the same way that the former communist countries used to recruit and train their most promising athletes. Clearly, these examples are completely unrealistic, but they nevertheless illustrate society's insatiable curiosity about new venture creation and its principal actor, the entrepreneur. Indeed, this quest for the "Holy Grail" has fuelled research in the field of entrepreneurship for many years.

Several theoretical approaches have been developed to explain why some people eventually become entrepreneurs. Among these, a relatively new stream of research has emerged, based on entrepreneurial intentions. More specifically, the intention to start a business is thought to be the best predictor of actual venture creation, such intention being formed by perceptions of the desirability and feasibility of going into business. This research perspective looks promising as the few empirical studies that have verified the link between perceptions and intentions have yielded convincing results. However, we do not know yet how entrepreneurial perceptions and intentions evolve over time. As a significant amount of time may elapse between the moment the intention to start a business is formed and the moment the potential entrepreneur initiate activities leading to venture creation, temporal stability is crucial for the model to hold. The objective of this research is to fill this important gap, through a longitudinal study of entrepreneurial intentions and perceptions.

The first part of the paper describes and discusses the theoretical approach selected for this study, namely that intention is a valid predictor of planned behavior such as starting a business. The next section situates the theoretical framework within the research context, i.e. a group of university students on the point of selecting their future careers. Research objectives are then exposed; they are firstly to test whether the students' perceptions of the desirability and feasibility of going into business have an impact on their intention to become entrepreneurs and secondly, to verify if those perceptions and intentions remain stable over time. The paper also describes the research methodology used and presents the results from the first two phases of this longitudinal study, along with some comments. It ends with an overview of the forthcoming phase of the research.

THEORETICAL FOUNDATIONS OF THE ENTREPRENEURIAL INTENTION APPROACH

The two main theory-driven models used by researchers having adopted the entrepreneurial intentions approach to study the venture creation phenomenon are those of Ajzen (1991), borrowed from social psychology, and of Shapero and Sokol (1982). This latter was developed specifically for the field of entrepreneurship.

According to Ajzen's Theory of Planned Behavior ("TPB"), any behavior that requires a certain amount of planning, as is unquestionably the act of venture creation, can be predicted by the intention to adopt that behavior (1991). Thus, it would be possible to predict whether or not an individual will eventually launch a business by studying his or her intention to do so. In the TPB, three variables precede the formation of intention, which itself predicts behavior: 1) the subject's attitude toward a given behavior, 2) subjective norms, i.e. the subject's perception of other people's opinions of the proposed behavior, and 3) the subject's perception of his or her control over the behavior.

According to Shapero and Sokol (1982), the decision to change direction significantly in life, for example by launching a business, is precipitated by an event or a break in the established routine. The person's choice will then depend on three elements, namely (1) his or her perception of the desirability of the proposed behavior (a combination of the first two variables in the preceding model), (2) his or her propensity to act (i.e. to act in accordance with his or her intentions), and (3) his or her perception of the behavior's feasibility, which is similar, conceptually speaking, to the third variable in the preceding model (see Figure 1).

[FIGURE 1 OMITTED]

The beauty of these two theoretical models lies in their simplicity. Both main constructs (perception of desirability and perception of feasibility) are in fact the product of the combined effects of several other variables studied in connection with the venture creation phenomenon. For example, the attraction of the idea of starting a business is probably dependent on the entrepreneurial models an individual has in his or her immediate environment, the prestige and respect ascribed to entrepreneurship as a career choice by the people around the individual, the individual's need for achievement and independence, the opportunities available in the environment, and so on.

But what is a theory worth if it cannot be verified? Generally speaking, the empirical results have been convincing. A meta-analysis of the findings of more than a hundred studies revealed an average correlation of .65 for attitudes and intentions, and .46 for intentions and behavior (Kim & Hunter, 1993). In a similar meta-analysis of the findings of 185 studies carried out up to 1997, TPB accounted for 39% of the variance in intention and for 27% of the variance in behavior (Armitage & Conner, 2001). The behaviors observed for these analyses included contraceptive use, engaging in physical exercise, giving blood, breast-feeding a child, exercising voting rights and so on--in other words, behaviors somewhat removed from the act of launching a business. However, some interesting findings were also obtained for a behavior similar to that of starting a business, namely growing a business (Orser, Hogarth-Scott & Wright, 1998). In this latter study, analysis of data collected from 139 small business owner-managers suggested that the respondents' intention to grow their business was a key factor in actual growth at the end of a four-year period.

The two models described above were also tested on the intention to start a business, and here too the results were significant. Krueger, Reilly and Carsrud tested and compared both models (2000): the model of Shapero and Sokol was found to be slightly superior to that of Azjen (adjusted [R.SUP.2] of .408 compared with .35) and in both cases the results were statistically significant (p<.0001). Reitan (1996) obtained results that were even more convincing with a model combining both the above and including situational variables: this new model explained 63% of the variance in the intention to start a business. However, these models appear to have less explanatory power when short-term intentions are used as the dependent variable: Reitan's model only explained .30 of the variance in the intention to go into business in the next two years, compared to .22 for Autio, Keeley, Klofsten & Ulfstedt (1997) using a one year intention variable.

It is important to mention that all these studies were aimed at explaining the formation of intention; they did not address the relationship between intention and action. Yet, this is the most important connection. As pointed out by Chrisman, "although the relationship between intentions and subsequent venture creation has considerable face validity and logic, it is not established that intentions always, or even usually, lead to entrepreneurial activity"(1999, p. 47). The few studies where this link was tested produced mixed results. In the five-year period of Katz's study (1989), only 10 out the 33 respondents who had the intention of seeking self-employment at the beginning of the study actually did so at some point. Two other studies showed a higher incidence of business start-ups (Carter, Gartner et Reynolds 1996; Chrisman, 1999). However, in both these cases respondents were in fact nascent entrepreneurs at the beginning of the study, as they had already initiated activities associated with starting a new firm (seek outside assistance, look for facilities, develop a business plan, etc.). This means that they had more than a strong intention to start a venture: they were engaged in the process of making the transition from intention to actual behavior. If it is not possible to verify the existence of a link between the intention to start a business and the realization of that intention, the models described above will remain only partially valid. We tend to think that it may in fact be difficult to establish a statistically significant association between intention and action for two reasons, namely the level of control an individual has over the act of launching a business, and the temporal stability of the intention to launch a business.

As pointed out by Ajzen (1991), if perception of control is to be used to predict behavior, there must be a good fit between that perception and reality. For example, if an individual believes he or she is perfectly capable of going into business and has mastered the process when in fact he or she lacks the knowledge, skills or resources required to succeed, the intention will probably not be realized. This gap between perception and reality may lie in the individual's evaluation of his or her own skills and competencies, or be external to the individual, in an area over which the individual has little control (e.g. access to a bank loan). It may also be a result of the individual's lack of knowledge about what is required (time, effort, resources, etc.) for the behavior to occur. In either cases, such inaccurate perceptions of the feasibility to start a business will weaken the predictive value of an intention-based model.

In the same line of thought, if the intention to launch a business is likely to change over time and as a result of circumstances, an intention formulated at time 1 will not predict action taken at time 2. Indeed, the longer the time between formation of the intention and its expected realization, the greater the chance that some new circumstance will occur to change the initial intention and invalidate the prediction. This is recognized by Ajzen who stated that the temporal stability of intentions is a condition for accurate behavioral prediction (1991). Using the TPB model, Sheeran, Orbell & Trafimow (1999) tested the hypothesis that the temporal stability of intentions moderates the relationship between intentions and behavior. Their results support this hypothesis: almost 20% more variance was accounted for in the behavior of respondents with stable intentions, as opposed to those with unstable intentions. Furthermore, the results indicate that when a person has a positive intention to perform a behavior, the stability of that intention is vital to successful performance. The intention to go into business has never been proven to be stable, indeed, some findings suggest that it may in fact be unstable (Audet, 2000; Katz, 1989). In Audet's study, students' perceptions of the desirability and feasibility of going into business were measured before and after they had taken a course on entrepreneurship, and some significant differences were found. Given that perceptions form the basis of intentions, it is reasonable to suppose that a change in perception will also trigger a change in intention. In Katz's study, only two of the 33 persons who indicated an intention to seek self-employment in the first year of the study reported the same intention one year later. The only other mention of entrepreneurial intentions came from a single respondent in the fourth year of the study.

RESEARCH QUESTIONS

In light of the foregoing, we felt it was essential to build on current knowledge of the venture creation phenomenon by testing a model based on the intention to go into business. More specifically, answers to the following questions were sought:

* Are entrepreneurial intentions predicted by perceptions of the desirability and feasibility of starting a business ?

* Do these perceptions remain stable over time ?

* Does the intention to start a business remain stable over time?

* What factors influence the temporal stability of the intention to start a business?

The theoretical model selected is that developed by Shapero and Sokol (1982), as reviewed by Krueger, Reilly & Carsrud (2000), with one change, namely the omission of the "propensity to act" variable.

METHODOLOGICAL FRAMEWORK

The project is scheduled to cover a three-year period, since the research questions require longitudinal observation of the variables. The research population is composed of university students from a business administration degree program. The interest of this group lies in the fact that, in terms of skills and knowledge, they appear to be excellent candidates for venture creation. Moreover, when the project started they were near the end of their studies and were expected to have the time and energy available in the near future to plan a business project. As research has shown that the entrepreneurial aspirations of students are highly sensitive to the image of entrepreneurship as a career path projected by the university community (Autio, Keeley, Klofsten & Ulfstedt, 1997), it is important to understand their entrepreneurial path and perhaps help and encourage them in the process.

The research specification provided for an initial data collection process at the beginning of the first year, i.e. during the respondents' last term at university ("Phase I"). The purpose of this process was to measure the variables of the theoretical model at time 1 ([t.sub.1]) and to test the first part of the model (antecedents of intention). A second data collection process was carried out in the second year, to measure the same variables at [t.sub.2], and identify and explain any variations ("Phase II"). The third and last data collection process will take place in the third year, for the same purposes ("Phase III").

Phase I

The sample consisted of 107 third-year undergraduate business students from Concordia University in Montreal, Canada. Data were gathered using a questionnaire distributed in the classroom and over the Internet. To avoid bias in the responses, the students were assured that participation in the study would not affect their grades. Perception variables were measured with a single item on a 1 to 5 scale. The following questions were asked: "How appealing is the idea of one day starting your own business?" (perception of desirability) and "How confident are you in your skills and abilities to start a business?" (perception of feasibility). Intentions were measured as a % probability that the respondent would go into business in the three-year period following graduation (short term) and at some point in their life (long term).

Phase II

The research population for Phase II was the same as for Phase I. However, the occupation of respondents changed: in the 18 months separating Phase I and II, they graduated from university and obtained full-time employment. The sample was much smaller than in the previous phase; of the 107 initial respondents, only 54 were found and agreed to participate to the study. This low retention rate is disappointing but not all that surprising since young adults are known to be very mobile. It is not a major concern as this smaller sample is nevertheless representative of the larger sample, there being no significant difference between these two groups when looking at the means of the four variables measured at [t.sub.1]. Data was collected through telephone interviews. All four variables were measured at [t.sub.2] and in addition open-ended questions were asked to explain differences in answers provided at [t.sub.1] and [t.sub.2]. Finally, respondents were asked about their current work situation to investigate a possible relationship between work satisfaction and short-term intentions to start a business, as will be explained below.

PRESENTATION OF RESULTS AND DISCUSSION

Phase I

When questioned, the respondents estimated the probability of launching their own business in the next three years at 25 %, which is fairly low (see Table 1 next page for descriptive statistics). Indeed, only nine students out of 107 estimated this probability as being 75% or more. Thus, very few students appeared to be considering an entrepreneurial career in the short term. In the longer term, the figures were more encouraging, with an average probability of 61%. Nearly half the students (43%) estimated the probability of starting a business at some time in the future at 75% or more. Thus, many students appear to have a significant interest in the idea of becoming an entrepreneur, but for most it is not something they feel they will do in the immediate future. These results are in line with those obtained from a sample of students enrolled in another Quebec university (Filion, L'Heureux, Kadji-Youaleu & Bellavance, 2002) and from university students in Russia and Norway (Kolvereid, 1996; Tkachev & Kolvereid, 1999).

Perception of desirability and perception of feasibility variables explained 49% of the variation in the long-term intention to start a business, compared with just 32% in the case of short-term intentions. The predictive power of both variables is fairly high (see Table 2 next page). Not too surprisingly, these results confirm previous findings (Autio, Keeley, Klofsten & Ulfstedt, 1997; Krueger, Reilly & Carsrud, 2000; Reitan, 1996) to the effect that perceptions of the desirability and feasibility of going into business explain the formation of entrepreneurial intention to a significant degree. Perception of desirability appears to be a stronger predictor than perception of feasibility, for both long term and short term intentions, this departing from Krueger's results but confirming Reitan's.

The model is more robust when the time frame is longer as opposed to shorter. Autio, Keeley, Klofsten & Ulfstedt argue that adopting a model using a strict short-term intention has both merits and weaknesses. On the positive side, its power to predict venture creation is probably higher than models using long-term intentions. However, the occurrence of a strong intent to go into business within a short time frame is relatively rare. Reitan explained his weaker results with a model using short-term intentions by saying that certain contextual variables excluded from his model (e.g. type of job) probably had a greater impact in the short term than the variables included in the model. For this reason and in light of Phase I results, it was decided to collect data about the work situation of respondents in Phase II; more precisely, the respondents' level of satisfaction with their current jobs was measured on a 1 to 5 scale.

Phase II

A regression analysis confirmed the Phase I results for long-term intentions (see Table 3 below) but not for short-term intentions. Indeed, the two perception variables failed to contribute significantly in the prediction of short-term intentions.

In an attempt to verify the explanation set forward by Reitan (1996), the variable "work satisfaction" was added to the regression model (see Table 4 below). This yielded disturbing results: both perception of feasibility and work satisfaction proved statistically significant in explaining short-term intentions, but not perception of desirability. As might have been expected, correlation between work satisfaction and entrepreneurial intentions was negative ([r.sup.2] of -.28), suggesting that some respondents reconsidered their short-term intention to go into business because they enjoyed their current jobs and were satisfied with their employment situation. Interestingly, work satisfaction had no impact on the intention to start a business in the distant future: when this variable was added to the model, its p value was non significant at .35. We are thus lead to believe that respondents' level of satisfaction with their current jobs explains the probability that they will start a business in the near future better than their perception of the desirability of going into business.

There may be an alternate explanation for the fact that perceptions of the desirability and feasibility of going into business fail to explain the intention to start a business in the near future. Sutton argues that one explanation for poor predictability by intention-based models is violation of the principle of compatibility (1998). This principle states that to maximize predictive power both the predictor and the criterion should be measured at the same level of specificity in terms of action, target, time and context. This principle is violated when measures of perceptions and intentions are not set in the same time frame. For example, in this study respondents were asked about their perception of the feasibility and desirability of starting their own business at some point in their life to predict their intention to go into business in the three-year period following their graduation. Results would have probably been different, and more convincing, if respondents had been asked about their perceptions toward starting a business in the next three years. First, their level of work satisfaction would have influenced their perception of the desirability of starting a business in the short term and made this variable more significant in explaining short-term intentions. Second, their perception of the feasibility of starting a business in the near future would have been a more accurate predictor of short-term intentions. To verify this hypothesis, perceptions of desirability and feasibility will be operationalized within the same time frame as intentions in Phase III of the research.

Descriptive statistics for Phase II variables do not appear to differ much from those at [t.sub.1] for the 54 respondents who participated to both phases of data collection (see Table 1). A t-test for equality of means and a paired samples test revealed no significant difference between [t.sub.1] and [t.sub.2] means for all four variables. For each of the two perception variables, approximately half the respondents provided the same answer at [t.sub.2] as at [t.sub.1] and the change in the probability of starting a business was greater than 20% for one third of the sample. Correlation coefficients for measures of the same variable at [t.sub.1] and [t.sub.2] were relatively high for perception of desirability ([r.sup.2]=.63), perception of feasibility ([r.sup.2]=.55) and long term intention ([r.sup.2]=.57) and slightly lower for short term intention ([r.sup.2]=.35), suggesting less temporal stability.

When looking at differences in frequencies between [t.sub.1] and [t.sub.2], changes in perceptions were found across all score categories (from 1 to 5). The trend appears to be that the perceptions of respondents in the lower categories at [t.sub.1] shifted to higher categories at [t.sub.2]. This indicates that, at [t.sub.2], more respondents felt they had the skills and abilities to start a business and regarded it as a desirable career option. However, it must be borne in mind that this trend does not show when we look at the differences in means. Such apparent changes in perceptions were associated with changes in both short and long-term intentions of starting a business at [t.sub.2]. Fewer respondents rated the probability of starting a business in the short term at 50% and above, but more expressed such an intention in the long run.

According to theory, respondents with a strong positive orientation qualify as "potential entrepreneurs" in the sense that, if the theory holds, they are the respondents with the highest probability of starting a business. These would-be entrepreneurs are also the ones that entrepreneurship scholars have been trying to identify a priori for years. If one takes a closer look at the answers provided at [t.sub.1] by students with a supposedly high entrepreneurial potential (% probability of starting a business > or = 75%), it can be seen that their short-term intentions changed drastically. Indeed, at [t.sub.2] only one respondent out of four still intended to start a business within the three-year period following graduation (see Table 5). The other two respondents who stated a high short-term probability of starting a business at [t.sub.2] had indicated probabilities of 65% and 5% (!) at [t.sub.1]. On a long term horizon, intentions are slightly more stable. Of the 21 respondents who, at [t.sub.1], had strong intentions of starting a business at some point in their life, only six had changed their minds to a significant degree at [t.sub.2] (>25% change in probability). On the other hand, among those who did not have a strong intent at [t.sub.1] but had one at [t.sub.2], 5 out of 6 respondents expressed a change in intention greater than 25%. In fact, the average increase in the probability of starting a business between [t.sub.1] and [t.sub.2] for these respondents was 37.5%, which is a change of rather large magnitude. Furthermore, two of these six respondents had expressed a very low probability of starting a business at [t.sub.1] (respectively 5% and 20%). This means that these "potential entrepreneurs" at [t.sub.2] would have never been identified as such at [t.sub.1] if an intention-based model had been used. There is thus a risk of missing real aspiring entrepreneurs when using such a model. Katz goes as far as to say that "studying those who intend to enter business ownership would result in missing over 98% of those who become business owners." (1989, p.530)

In light of the above, it is difficult to evaluate with much certainty the temporal stability of the intention to start a business and of the perceptions that form such intention. What is the range within which an intention can change over time without being labeled "unstable" ? Is a 10% or 20% change in the probability of starting a business indicative of temporal unstability ? Would a change from a probability of going into business of 85% to 75% affect the likelihood that the behavior will occur as much as a change from 75 to 85% ? The answer to these questions is highly consequential as respondents with stable strong positive intentions are considerably more likely to perform the predicted behavior than those with positive but unstable intentions (Sheeran, Orbell & Trafimow, 1999). There is an urgent need to clarify the notion of temporal stability as applied to the intention of starting a business.

Identifying the factors and events that are most likely to impact on the intentions of respondents may help us understand how these intentions change over time. Among the reasons given by respondents to explain changes in their intentions or perceptions were:

Positive change

* Being more mature, having more work experience, this having an effect on their self-confidence;

* Money: some realized that their earning potential was limited when working as an employee and they want more, much more.

* Freedom: some resent the constraints of the workplace (rules, fixed schedule, being supervised, fixed salary, etc.);

* Opportunity recognition: being active in the workplace puts respondents in a better position to identify and recognize business opportunities than when they were students;

* Being your own boss: some long for the satisfaction of working for themselves, of having something left for themselves at the end of the day.

Negative change

* Reality shock: some realized that they knew far less than they thought and that the process of starting a business was not as easy as they had imagined (this is a typical case of mismatch between perceptions of feasibility and reality);

* Corporate orientation: some realized they enjoy employee status and the corporate environment (stability, less risk, good pay, etc).

One fact worth emphasizing is that the respondents had gone through a very major change in their lives between [t.sub.1] and [t.sub.2]: they went from being students to having their first full-time job in their field of expertise. As a result, their income level changed as did their social environment, their status in society, their place of residence (for most), their legal status (for some), and so on. Clearly, changes of this magnitude cannot be expected to occur again in the next several years. It will thus be extremely interesting to see if their intentions continue to change from [t.sub.2] to t3, despite the fact that they have entered a period of relative stability in their life paths.

LOOKING TO THE FUTURE

Data gathered in the first two phases of the project confirmed results from previous studies: perceptions of the desirability and feasibility of starting a business significantly explain the intention to go into business at some point in life. However, the model does not hold as well when the time frame is shorter. Indeed, at [t.sub.2] both perceptions failed to explain to a significant degree the intention to go into business in the next three years. When work satisfaction is added to the regression model, both this variable and perception of feasibility become significant predictors of short term intentions, perception of desirability remaining non significant. The major contribution of this research so far is to provide longitudinal measures of the variables forming an intention-based model of venture creation. Our results tend to indicate that entrepreneurial intentions and perceptions vary over time. As the temporal stability of intentions is a condition for an intention-based model to accurately predict behavior, the link between entrepreneurial intentions and actual venture creation may prove difficult to establish. As pointed out by Niittykangas and Laukkanen, expressed readiness for an entrepreneurial career can probably "serve as a useful indicator and a starting point for officials making decisions about whether or not to proceed further with a prospective self-employer. (...) Importantly, entrepreneurial potential cannot be assumed to be confined to such persons only"(1998, page 61). Data collected in future phases of the project should shed more light on the temporal stability of intentions and on the process by which intentions are transformed into action.

REFERENCES

Ajzen, I. (1991). The Theory of Planned Behavior. Organizational Behavior and Human Decision Processes, 50, 179-211.

Armitage, C. & M. Conner (2001). Efficacy of the Theory of Planned Behavior: A Meta-analytic Review. British Journal of Social Psychology, 40, 471-499.

Audet, J. (2000). Evaluation of Two Approaches to Entrepreneurship Education Using an Intention-based Model of Venture Creation. Academy of Entrepreneurship Journal, 6(1), 57-63.

Autio, E., R. Keeley, M. Klofsten & T. Ulfstedt (1997). Entrepreneurial Intent among Students: Testing an Intent Model in Asia, Scandinavia and in the USA.. Frontiers of Entrepreneurship Research, Wellesley, MA: Babson College.

Carter, N., W.B. Gartner & P.D. Reynolds (1996). Exploring start-up event sequences. Journal of Business Venturing, 11, 151-166.

Chrisman, J.J. (1999). The influence of outsider-generated knowledge resources on venture creation. Journal of Small Business Management, 37(4), 42-58.

Filion, L.J., L'Heureux, D., Kadji-Youaleu, C. & F. Bellavance (2002). L'entrepreneuriat comme carriere potentielle--Une evaluation en milieu universitaire, Chaire d'Entrepreneurship Maclean Hunter, Cahier de recherche HEC 2002-04.

Katz, J.A. (1989) Intentions, Hurdles, and Start-ups: An Analysis of Entrepreneurial Follow-through.. Frontiers of Entrepreneurship Research, Wellesley, MA: Babson College.

Kim, M.S. & J. E. Hunter (1993). Relationships among Attitudes, Behavioral Intentions, and Behaviors: A Meta-analysis of Past Research. Communication Research, 20, 331-364.

Kolvereid, L. (1996). Prediction of employment status choice intentions, Entrepreneurship, Theory & Practice, 21(1), 47-57.

Krueger, N. F., M. D. Reilly & A.L. Carsrud (2000). Competing Models of Entrepreneurial Intentions. Journal of Business Venturing,15, 411-432.

Niittykangas, H. & M. Laukkanen (1998). Potential entrepreneurs--Can they be located ?. International Journal of Entrepreneurship, 2, 47-65.

Orser, B.J., S. Hogarth-Scott & P. Wright (1998). On the Growth of Small Enterprises: The Role of Intentions, Gender and Experience. Frontiers of Entrepreneurship Research, Wellesley, MA: Babson College.

Reitan, B. (1996). Entrepreneurial Intentions--A Combined Models Approach. Paper presented at the 9th Nordic Small Business Research Conference, Lillehammer, Norway.

Shapero, A. & L. Sokol (1982).The Social Dimensions of Entrepreneurship. In The Encyclopedia of Entrepreneurship,. C. Kent, D. Sexton & K. Vesper (Eds.), Englewood Cliffs: Prentice-Hall.

Sheeran, P., S.Orbell & D.Trafimow (1999). Does the Temporal Stability of Behavioral Intentions Moderate Intention-behavior and Past Behavior-Future Behavior Relations?. Personnality and Social Psychology Bulletin, 25(6), 721-730.

Sutton S. (1998). Predicting and Explaining Intentions and Behavior: How Well Are We Doing?. Journal of Applied Social Psychology, 28(15), 1317-1338.

Tkachev, A. & L. Kolvereid (1999). Self-employment intentions among Russian students. Entrepreneurship & Regional Development, 11(3), 269-280.

Josee Audet, Universite Laval

Table 1: Mean scores of perception, intention and work satisfaction
variables at [t.sub.1] and [t.sub.2]

 std. std.
Variable [t.sub.1] dev. [t.sub.2] dev.

 n=107 n=54

Perception of desirability 3.51 1.08 3.56 1.11
 (1= low, 5= high)
Perception of feasibility 3.39 0.87 3.57 0.92
 (1= low, 5= high)
Intention to start a business 25.46% 26.76 22.61% 25.40
 <3 yrs (probability as a %)
Intention to start a business 60.73% 27.63 64.07% 22.51
 >3 yrs (probability as a %)
Work satisfaction not available 3.57 0.92
 (1= low, 5= high)

Table 2: Summary of Regression Analyses Results at [t.sub.1]

 Short-term Long-term
 intention at intention at
Variables [t.sub.1] [t.sub.1]

 t value p t value p

Perception of desirability 3.28 .0014 6.29 <.0001
Perception of feasibility 2.89 .0047 2.21 .0292

 Adjusted Adjusted
 [R.sup.2] = .32 [R.sup.2] = .49
 p=<.0001 p=<.0001

Table 3: Summary of Regression Analyses Results at [t.sub.2]

 Short-term Long-term
 intention at intention at
Variables [t.sub.2] [t.sub.2]

 t value p t value p

Perception of desirability 1.68 .0985 4.04 .0002
Perception of feasibility 1.44 .1560 2.89 .0056

 Adjusted Adjusted
 [R.sup.2] = .13 [R.sup.2] = .47
 p=.01 p=<.0001

Table 4: Summary of Regression Analyses Results at [t.sub.2] with work
satisfaction added to the model

 Short-term Long-term
 intention at intention at
Variables [t.sub.2] [t.sub.2]

 t value p t value p

Perception of desirability 1.16 .2541 3.69 .0006
Perception of feasibility 2.39 .0212 3.32 .0018
Work satisfaction -2.52 .0155 -0.94 .3541

 Adjusted Adjusted
 [R.sup.2] = .23 [R.sup.2] = .48
 p=.0021 p=<.0001

Table 5: Variation in probabilities of starting a business among
respondents with a strong short term intent at [t.sub.1] or [t.sub.2]

 Probability Probability Change in
 at [t.sub.1] at [t.sub.2] probability

Respondents with a strong 75% 25% (50%)
intent at [t.sub.1] (n=4)
 75% 20% (55%)
 99% 100% 1%
 100% 0% 100%
Respondents with a strong 5% 95% 90%
intent at [t.sub.2] only 65% 90% 25%
(n=2)