How Career Identity Shapes the Meaning of Work for Referred Employees

Sociological explanations for why referred employees typically have longer tenures than nonreferred employees tend to be either that referred employees enter their jobs possessing a clearer sense of employer expectations or that they often receive support from their referrers while on the job. However, through analysis of work-history interviews conducted with salespersons in Toronto, Canada, I find that the significance of each of these factors for a person’s tenure depends on their career plans. For individuals with clear career plans, information mattered but support was less important. Conversely, for individuals with unclear career plans, support mattered but information was less important. I find that this divergence was based on the fact that individuals who had clearer career plans cared more about the fit they had with the tasks they performed jobs which they were referred into while those with unclear plans tended to be more concerned about their overall fit with the job’s culture. I delineate these this difference in job satisfaction by demonstrating how the combination of information and support respondents had at any given job led them to either support, interrogate, or re-route their career plans differently based on the initial clarity of these plans. Based on these findings, I argue that the role that referrals play in shaping turnover intentions should be nested within individuals’ career identities. Doing so prevents researchers from seeing turnover intentions as being solely based on expectations at the time of hire or connections made, strengthened, or weakened on-the-job and, instead, necessitates a more grounded view of turnover decisions.


INTRODUCTION
A key contributing factor to the overall unemployment rate in Canada is high turnover in frontline jobs such as those in call centers, retail stores, and fast food restaurants (Crémieux and Van Audenrode, 1996;Amine, 2013). Individuals occupying these jobs often report feelings of overwork, burnout, and job insecurity, all of which are reported in their reasons for leaving their jobs (Sallaz, 2015). The high turnover rates of employees in these jobs are of special concern given that they are disproportionately occupied by recent immigrants, visible minorities, women, and youth (Creese and Wiebe, 2009;Uppal and LaRochelle-Côté, 2014).
Given the reasons that frontline workers give for leaving their jobs, one might be surprised to find that referrals have a major impact in increasing their job tenure. Despite the precariousness of much frontline work, research has shown that, similar to employees in other jobs, frontline workers who are referred to their jobs tend to have lower turnover rates (Burks et al., 2015). Explanations for this tend to be either that referred employees enter into their jobs possessing a clearer sense of employer expectations (Sicilian, 1995) or that they often receive support from their referrers while on the job (Chapman and Mayers, 2015). These factors vary in regard to their effects on turnover across both companies and industries (Burks et al., 2015), however, leaving the reasons why referred employees stay in their jobs an empirical question.
Given the potential for contingencies specific to particular workplaces that could be shaping this relationship (see Leschziner, 2015), combined with the role that individual employees' prior experiences and human capital investments could be playing in shaping these outcomes (see Holtom et al., 2008) more work is needed to determine the role that expectations and peer support play in shaping the employment tenures of referred employees. Gauging the impact of individuals' sense of career or work identity in shaping this process is a promising route to take (see Ibarra, 1999;Leschziner, 2015). This is because, by isolating referral's impact outside of the context of individuals' broader ideas about themselves as workers, I argue that neither expectation-based or peer support-based explanations provide full accounts of individuals' motivations to either stay in or leave their jobs (see Williams, 2018). Indeed, the presence of inconsistent findings in regard to both of these two causal pathways attests to the omission of important variables. I assert that this omission is a product of many social scientists' tendency to focus either on dispositional or contextual variables (see Patterson, 2014)-the former represented by information and expectations, and the latter by ongoing peer support.
To specify the role that referrals play in shaping tenure in frontline jobs, I have interviewed 25 referred individuals who work in sales. Interviews focused on delineating dispositional and contextual elements that might precipitate turnover by examining these individuals' motivations for entering into this area of work, entering into the particular jobs they currently occupy, and assessing motivations that they had for leaving previous jobs as well as any turnover intentions they may now have regarding their current job. The role of their referrer was also assessed, in terms of the information they provided them or any additional benefits they felt being referred offered them.
Through analysis of these interviews, I found that the role that referrals played in shaping turnover was mediated by respondents' broader career identities and expectations of specific jobs. These expectations were not, however, conceived simply as expectations about the tasks or pay of certain jobs. Rather, these expectations were part of respondents' larger sense of identity and, in particular, their career identity (see Ibarra, 1999). Respondents tended to make turnover decisions based on how they saw any given job as fitting into what they imagined (see Emirbayer and Mische, 1998) their careers to be and where they saw themselves going. Being referred often was a reason why these individuals stayed, but was typically articulated as such only due to a feeling of that they shared of owing their referrers-they did not want to jeopardize either their relationships with their referrers or their referrers' jobs. Importantly, the strength of an individual's relationship with their referrer tended to correlate with that individual's turnover choices, leading to the conclusion that while weak ties may be more beneficial for individuals to obtain employment on average (Granovetter, 1973; see Burt, 1992;Lin, 1999), strong ties keep individuals on the job. The strength of an individual's ties is not static, however, as a key reason for individuals deciding to leave a job was when a once strong connection became weakened. Such weakening of ties often led them to question their commitment to their employer now that their relationships were weaker within this organization.
Connected to the role that relationships with individuals' referrers play in shaping turnover intentions, I demonstrate how career plans or career identity (see Ibarra, 1999) shaped how information and peer support led to respondents staying in or leaving their jobs by supporting, interrogating, or re-routing their career plans. This was due to respondents experiencing any given job as being part of a career plan, forcing them to question their career aspirations, or as happily but unexpectedly taking their careers in a new direction. While being referred was an important dimension of how respondents valued any given job, this valuation was marked heavily by this broader, more existential sense of their awareness of their particular career phase (see Ibarra, 1999). As such, I argue that the role that referrals play in shaping turnover intentions should be nested within individuals' career identities. Doing so prevents researchers from seeing turnover intentions as being solely dispositional or situational, and, instead, as being connected to individuals' broader self-concepts (Miles, 2014;see Leschziner, 2015) and career identities (see Ibarra, 1999). It also enables analysis of individuals' career decisions in ways which make sense of seemingly impulsive or sudden job exits as individuals may anchor themselves to different images of ideal work, and in turn, come to make different career decisions despite possessing similar employment backgrounds.
I introduce the concept of spontaneous conflict resolution to explain what I observed as being individuals' attempts to reframe their work as meaningful in light of unmet expectationsattempts which sometimes helped individuals find reasons to stay in their jobs but, more often than not, led them to seek out opportunities which could better enable them to satisfy their career needs given perceived shortcomings in their current workplaces. I further discuss how the perception of such conflicts varied based on both the strength and the quality of respondents' initial expectations about the job. For those who expected particular tasks, conflict resolution tended to be performed when the tasks aligned but other facets of the job were not as expected. This would lead these respondents to disattend to these latter issues in light of appealing job tasks. Conversely, for those who expected a particular workplace culture, this resolution tended to involve amplification of other components of the job. In turn, this would lead these respondents to disattend to their initial drivers in light of appeals to workplace culture. Thus, as will be shown, based on whether they initially possessed a task-or culture-orientation respondents would narrow the parameters of their expectations due to them going unmet in two opposite ways.

TURNOVER IN THE SOCIOLOGY OF WORK AND OCCUPATIONS: VOLUNTARY AND EXPEDIENT
To study the ways in which individuals stay in and leave their jobs, the term turnover is frequently used (see Burks et al., 2015). Turnover refers to the amount of job-leaving or "churn" that occurs in an organization and is categorized according to whether it occurs voluntarily by employees or involuntarily at the hands of managers within the organizations (Burks et al., 2015).
As turnover rates remain stable across industries and continue to constitute one of the highest cross-industry costs (Radford, 2013), and as turnover negatively impacts both individual employees' levels of commitment to their workplaces (Mueller and Price, 1989) and organizations' overall performance (Glebbeek and Bax, 2004), this topic has been increasingly addressed by both academics and industry leaders (Holtom et al., 2008;Burks et al., 2015). The issue of voluntary turnover, in particular, has received an extraordinary level of attention (Saunders and Maxwell, 2003;see Baldry et al., 2007). As jobs become more flexible (Saunders and Maxwell, 2003), and as currently employed individuals continue to actively seek out alternative job positions (Pissarides and Wadsworth, 1994), efforts to understand the forces that work to prompt such behavior have become important both for organizations wanting to increase revenues by reducing the amount of hiring and training that they need to do of churning employees, as well as for social scientists interested in studying changes in labor market trends.
Researchers have found that many factors influence voluntary turnover, ranging from macro-level contextual trends to more micro-level, individual factors (Mobley, 1977;Mobley et al., 1979;Holtom et al., 2008;Hamori et al., 2012). These factors have often been utilized in complex models to map out the decisionmaking process that individuals engage in once they begin to think about leaving their jobs. For example, factors such as current labor market conditions (Mobley et al., 1979) and the nature of the job itself (i.e., the level of routinization or autonomy present in the job; Mobley et al., 1979;Schervish, 1983), the culture of the organization (Abelson, 1993;Arthur, 1994;Coff, 2011), and the tenures of employees currently working at an organization (O' Reilly et al., 1989) are frequently conceptualized as distal causes of turnover due to these factors either prompting or inhibiting employees' initial thoughts about leaving their jobs (see Mueller and Price, 1989). Factors such as employees' own perceptions of stress as well as their performance at the job are conceived as being more proximate causes of turnover, as these perceptions may result in employees quickly and more thoroughly beginning to think about leaving their jobs (Holtom et al., 2008).
Despite differences in motivations for leaving jobs across industries, as well as differences in distal causes of turnover, referred employees tend to have lower turnover rates across all industries (Sicilian, 1995;Chapman and Mayers, 2015). This finding is quite intriguing to researchers who study voluntary turnover, as it questions many of the existing theories and models explaining the turnover process (see Burks et al., 2015). While many explanations and models exist which explain the role that referrals might play in shaping this process, two core factors have been found by scholars to explain this relationship: the information about the jobs that referred employees have and the peer support that they receive once on the job (Fernandez et al., 2000;Burks et al., 2015).

Information: Human Capital on Arrival
Sociologists who do research on work and occupations often focus on relatively objective measures when exploring individuals' employment trajectories (Sorenson, 2000), performance (Burks et al., 2015), and compensation (Peterson and Morgan, 1995). The umbrella term human capital captures the diversity of metrics used by these sociologists, as these metrics may all be conceptualized as relatively stable investments that individuals bring with them into particular jobs or occupations (Schultz, 1961;Becker, 1964;Burton-Jones and Spender, 2011). Made famous by Schultz (1961) and Becker (1964), human capital refers to the host of resources that an individual has to offer a waiting world of prospective employers. Factors such as education and previous employment are oft-used metrics when discussing the returns individuals receive on human capital investments (Lepak and Snell, 1999). Similarly, the markets in which such investments are received are also studied in terms of how they foster exchanges between these experiences and job opportunities, salaries, benefits, and more personal on-the-job treatment by both employers and fellow employees (Hamori et al., 2012).
Indeed, as evidenced by the surge in Human Resource Management firms, Talent Management agencies, and other third-party organizations designed to manage organizations' recruitment strategies, organizational leaders have increasingly been attempting to hire individuals whom they believe will "fit" into their organizations (O' Reilly et al., 1989;Holtom et al., 2008;Hamori et al., 2012). Fit, in this sense, is defined as the quality of match that a prospective employee is assumed to have in a given organization based on criteria such as educational background, personality type, work style, leadership ability, future goals, and work skills (Kristof-Brown and Billsberry, 2013). Literature on this topic has indeed established that people do fit into their jobs differently and this fit is indicative of how productive individuals will be on the job, how they will relate with co-workers on the job, and how long they will stay on the job (Kristof-Brown and Billsberry, 2013).
The prospective nature of employees' sense of fit-employees have not entered into their jobs yet, so their fits are being assessed based on criteria that they possess prior to job entryforegrounds the role that referrals play in this process as a form of capital that employees bring into employing organizations. For instance, the better match theory holds that referred employees, through the process of being informed about the job, have a better sense of what the job entails than do those non-referred (Sicilian, 1995;Rivera, 2015). Simply by having a better sense of what the job entails, referred employees will more likely enjoy and be good at the job they are entering into (Wanous, 1980).
Referrers are also likely to refer individuals whom they believe will be good at the job. This can be due to either them believing they possess the required technical or soft skills or that they are adaptive enough to fit the company's needs (Fernandez et al., 2000). Due to the referrer's own experience with the company to which they are referring the prospective employee, such assessments will likely be good [although there is variation in quality of assessments made by referrers based on referrers own performance and fit within their companies (Fernandez et al., 2000)]. As the referred employee will serve as a reflection of the referrer's judgment, referrers seldom refer individuals that they do not believe will succeed at the job (Burks et al., 2015).

Peer Support and Social Capital: Human Capital Developed on the Job
While the information that referred employees have about the job on first-entry may be important, some researchers have sought to emphasize the active role that referrers play in shaping both employees' job abilities and job satisfaction rates once on the job. The social enrichment theory holds that referred individuals will have stronger social connections and peer support within the organization due to their referrers being present in the organization (Fernandez et al., 2000). Social enrichment's claims are borne out in the rich sociological literature on social capital, which refers to investments individuals can make in themselves by way of increased connections with individuals who may be able to help access hard to reach employment opportunities or other kinds of opportunities which involve individuals capitalizing on "structural holes" (Burt, 1992) where individuals can directly talk to key decision-makers or other individuals' whose words carry more weight (Erickson, 1996;Lin, 1999). In short, investments in social capital yield expected returns and can predictably be accumulated (Burt, 1992;Erickson, 1996;Portes, 1998;Lin, 2001). Together, the social enrichment and social capital perspectives help make sense of the fact that, aside from the human capital individuals bring into any given organization, the extent to which these skills, intentions, and capacities impact turnover decisions is impacted by the nature of the relationships one has on-the-job.
This line of work is inspired by Granovetter's (1973) work on weak ties, namely that individuals will often find employment through individuals that they hold more acquaintance-like relationships with than from strong ties-those individuals with whom they are closely attached, such as long-term friends or family members. From this perspective, referrers may importantly shape referred employees' trajectories at a workplace by providing them support and networking on the job (Fernandez et al., 2000). Indeed, Moynihan and Pandey (2008) found that individuals with strong network connections inside their workplaces tended to demonstrate lower turnover retentions.
Critically, however, these kinds of internal referralsparticularly from strong ties such as family members-can lead to bias in hiring decisions and cultural homogeneity in the workplace (see Padgett and Morris, 2005). Internal referral systems can lead to nepotism in hiring, a process by which employees hire those not only like them, but closely related to them due primarily not to their suitability to the particular job but based on those similarities and connections in themselves (see Bellow, 2003). For this reason, though internal referrals lead to high levels of retention (Miller, 2018), organizations need to ensure that workplace diversity is maintained to prevent unfair hiring practices from becoming established as institutional logics of the organization (see DiMaggio and Powell, 1983).

Information and Peer Support: Unfolding Human Capital
Human capital upon entry and peer support may serve as distinct pathways by which employees stay on the job, but they also overlap. This makes sense intuitively, as individuals always have some degree of human capital when entering a job, but also typically engage in some kind of contact with their referrer or other employees once on the job. For example, referred employees who enter into the job with a high level of personorganization fit may also have close relationships with their referrers.
To gauge the way that prospective employees' ideas about the job prior to entry influence their tenure and performance once on the job, initiatives have been drafted by many companies to make their job descriptions more transparent for potential employees in the hope that more employees will enter into the job with "realistic" expectations. This has been echoed by state officials who believe that discrepancies between individuals' ideas about jobs and those jobs' actual characteristics are a key contributor to unemployment rates, and supported by personnel psychologists who have found that individuals' expectations of what their jobs entail influence how well they perform and how long they stay at any given job (Reilly et al., 1981).
Shifting the focus to on-the-job behavior, other companies and scholars have focused on how to increase peer-to-peer relationships to foster the same sort of social bonding among recently acquainted employees that they have witnessed between referrers and referred employees (Mueller and Price, 1989). Individuals in these firms believe that a key contribution to turnover rates in their firms is a lack of reinforcement of their firm's core values in its employee base (Mueller and Price, 1989). In this view, person-organization fit can be continually developed and maintained in employees, and this serves as a feasible way for firms to manage their employees and create the workforces they desire.

DECISION-MAKING AND COGNITION
While originating in scholarship on work and occupations, the two core pathways that I discussed above-human capital on entry and peer support-tie into recent debates in sociology regarding the role that prior experiences play in shaping individuals' decision-making processes. Broadly framed in terms of two schools-what I term the dispositional school and the situational school-sociologists have typically explained the decisions individuals make in terms of the properties they possess and the situations they enter into (Williams, 2017a). While much work has been done recently to move toward a synthesis of dispositional and situation accounts (see Lamont et al., 2014), similar to the study of turnover, scholars have argued that more work is needed to move from abstract models to concrete processes of decision-making (see Emirbayer and Mische, 1998;Vaughan, 1998;Baldry et al., 2007;Leschziner, 2015). The study of turnover among employees is uniquely suited to contributing to this discussion due to the amount of contextual factors that influence decisions about work, coupled with the fact that qualitative interview methods enable the examination of these individuals' own evaluations, thoughts, and reflections about their work situations. The inclusion of referral status as my main conditioning factor also provides an opportunity to tease out the role of dispositional or antecedent and situational or contextual factors in shaping individuals' decisions to stay in or leave their jobs. Referred employees bear the unique quality of possessing organization-specific human capital-strong social connections and advanced knowledge upon entering an organization relative non-referred employees (Fernandez et al., 2000;see Burks et al., 2015)-that is not clearly visible to either fellow employees or to potential recruiters or headhunters at other organizations (Coff, 2011). While high performing employees at an organization may come to be seen as "star" or "core" employees (Lepak et al., 2011, p. 334), employees with tight bonds to one or more co-workers, due to being referred into the organization by these individuals, may or may not make these connections visible to others. In this sense, referred employees have invisible bonds to others that disincentive job-leaving, while not signaling to competing firms strong performance and, hence, attractiveness (see Coff, 2011). Put another way, while factors such as performance might simultaneously promote both job tenure and job-exit, it may be the case that referral status uniquely promotes job tenure through invisibly anchoring an employee to co-workers.

Dispositions and Action
Sociologists have long been torn between how to best conceive of individual thought and action: Currents have flowed between crediting much of this to social structure or to some idea of personal choice or agency not fully reducible to any context outside of individuals, be it psychological in nature or even more ephemeral (see Fromm, 1969;Smelser, 2014;Williams, 2017a,c). While proponents of both views share the same basic assumptions about thought and action-namely that thought and action can be conceived of as the interplay between (a) internalized mental structures that individuals collect or build up as they live their lives which enable the perception of their surroundings (Mead, 1934;Williams, 2017c) and (b) or some sense of emerging personality that, while informed by individual experiences, is not fully reducible to these experiences (Archer, 2007)-they differ in regards to which of these factors they emphasize in their explanations.
This distinction is quite important as it shapes both the kinds of methodologies that scholars use as well as the implications that they draw from both their own research and the work of others (Williams, 2017d). For example, Archer (2007, p. 2), a thinker who has repeatedly announced her commitment to studying individuals' "reflexivity" and decision-making capabilities, critiqued many of the conclusions drawn from Bourdieu's (1984) contemporary sociological classic Distinction. Asserting that Bourdieu's conclusions rested on the idea that children essentially inherit their parents' class positions and, in turn, life chances, Archer (2010) argued that not only was such a seamless transfer of class position unlikely in Bourdieu's own study sample, but was highly improbable in other contexts. Using interview data drawn from British youth, she found that the processes by which parents pass on their class positions to their children-primarily through parenting and educational choices made by parents in the young lives of their children-do not result in any clear transfer of class position to children. This is largely due to the fact that individuals in contemporary Britain require very specific credentials to engage in most varieties of paid work, and therefore need to actively choose which kinds of occupations they hope to enter into (Archer, 2010). In short, Archer (2010) questions the extent to which class reproduction can be seen as an automatic process, arguing that, at the very least in modern British society, reflexivity mediates this process due to the fact that individuals must actively seek out credentials and skills in order to be hired into any career regardless of what their parents have or have not done.
Archer's critique of more heavily structural accounts extends to criticisms of the recent use of the sociological dual process model (SDPM). According to the SDPM, individuals think and act automatically and unconsciously, or deliberately and consciously (Lizardo, 2017). The core of this model is called the dual process model, a model which assumes that cognitions occur either automatic and unconsciously relying on what are called Type 1/System 1 processes or consciously and deliberately and relying on what are called Type 2/System 2 processes (Evans, 2008;Vaisey, 2009). This model gained footing in sociology in response to criticisms of sociological research as privileging situational aspects of individual action over more dispositional and ingrained elements (Vaisey, 2009). However, the dual process model is fragmented in that many versions of it have been formulated based on assumptions about the nature of autonomous or unconscious and controlled or conscious processing (Chaiken, 1980;Smith and DeCoster, 2000;Thompson, 2009;Evans and Stanovich, 2013;Ackerman and Thompson, 2017;De Neys, 2018;Evans, 2018;Pennycook, 2018). In short, a wave of sociologists believed that the discipline had been privileging the contexts of action at the loss of more deep-seated, often moral components of action (see Patterson, 2014;Cerulo, 2018). This wave believed that sociologists were increasingly abandoning the role of intra-individual, habitual action, and used the SDPM to support their claims about the role of habit in human behavior.
While the SDPM accounts for the role that dispositions may play in shaping action, the overall model of action that it has supported has recently come under criticism from a range of psychologists (e.g., Evans, 2012) and sociologists (e.g., Leschziner, 2015;Williams, 2017b;Cerulo, 2018). The primary point of contention for such critics is the way the SDPM conceives of thought and action as bifurcated between fully automatic and fully deliberative action, and as such does not account for the ways in which individuals might automatically come to perceive their surroundings differently or fail to interpret stimuli as they normally would (Williams, 2017b). For example, in her study of elite chefs, Leschziner (2015) found that many chefs would come up with new recipes without deliberative intent. Chefs would often be involved in the creation of particular dishes, when suddenly they would think to add a particular novel ingredient to a dish. The addition of such a novel ingredient would hit them without conscious effort, making this experience far from purposive. Thus, in such a scenario the motivation for change in a person's actions-cooking, in this case-is not a conscious attempt to override her existing cognitive patterns, but a non-conscious impulsion triggered by a feeling she has when confronted with a particular stimulus. A chef that Leschziner (2015) mentions who thought to pair pink peppercorns with panna cotta and dried apricots did so not out of conscious intent, but out of a sudden, almost impulsive sense that the addition of pink peppercorns would just work. By framing thought and action as automatic or deliberate, the sociologists using the dualprocess model have no conceptual tools to understand this kind of action and, thus, spontaneous creativity of this sort cannot be made sense of very easily.
Such a view of spontaneous creativity was found by Ibarra (2004) in her work on how professionals made career moves. Introducing the concept of working identity, she argued that individuals' career identities should be seen as provisional and always in progress as they tend to change as new skills are acquired, new roles are assumed, new role models become available, and occupational environments change. While vocational preferences and habituated ways of understanding one's competencies still impacts the kinds of work an individual seeks and how they feel about that work, Ibarra (2004) found that individuals were much more pliable in their work orientations than were often assumed in many vocational theories. She argued that individuals should experiment with their working identities, as taking on new roles could inspire individuals to re-frame their existing talents and transform them into what she called convertible competencies: knowledge and skills that could be used in domains other than those in which they were originally formed. For example, a graduate student may feel that her research skills are only valuable in the academy, but such skills can relatively quickly be translated to skills that are required in the private sector such as market research and trend analysis.
As will be shown below, when applied to the case of turnover, the notions of spontaneous creativity and working identity in light of occupational demands can be extended into individuals' turnover decisions. I introduce the concept of spontaneous conflict resolution to explain what I observed as individuals attempted to reframe their work as meaningful when their expectations about their jobs went unmet either in terms of the tasks they were expected to perform or the kind of relations they had with others on the job. Only by seeing these individuals' turnover decisions as nested within their overall career identity does the emphasis on tasks or relations become analytically sensible-for individuals who were more task-oriented, unmet expectations about other components of the job could be neutralized while for those less focused on tasks and more on workplace culture and relationships tasks mattered far less. These two different valuations were related to the extent to which an individual's career identity was well-developed, as those with clearer career goals more highly valued tasks and those with vaguer career identities valued other, more cultural aspects of their jobs.

MATERIALS AND METHODS
Data for this project was collected over a 14-month period spanning from Summer 2016 to Fall 2017. Respondents were recruited using a popular classified advertisement website. An ad was placed in the "volunteers" section of the website, with the description asking that any individuals currently employed in sales would be suitable for interview. The study received ethics approval from the University of Toronto Research and Ethics Board (protocol #3306). All subjects gave written informed consent in accordance with the University of Toronto Ethics Committee.
Thirty-two respondents were recruited, all having current employment in sales jobs, and a wide variety of work and educational histories. This sample was reduced to twenty-five, as seven respondents had never been referred to a job. Of the 25 had been referred to a job at some point in their careers, 12 respondents were referred to their current jobs. Of the 25 respondents included, 14 respondents identified as male, and 11 as female. Five of the referred respondents occupied managerial roles, with the remaining 20 occupying managerial roles.
Interviews took place in the author's office 1 and were semistructured and in-depth, focusing on the core themes of how individuals find meaning at work, conceptualize their own career trajectories, and view the impact of being referred on their work lives and career identities. Interviews began with a series of questions about the respondent's work history, moved onto current employment, and ended with a discussion about expectations about the future. Interviews ranged from 1 to 3 h, with the average length being ∼2 h. Several probes were used to unpack what work meant to each respondent, as well as to flesh out the impact and interpretation of different life events for respondents on their career trajectories. The use of probes evolved after each interview, with interviews becoming more conversational both as they each progressed and overall from interview to interview. I found that giving more information about myself and my experiences throughout the interview greatly increased my rapport with the respondents and got them to expand on their own answers to my questions more often.
This conversational approach is also a product of my choice of analytic strategy: narrative analysis (Hollway and Jefferson, 1997). By seeing interviews as whole narratives composed of sub-narratives, plots, and narrative arcs, this approach enables researchers to focus on the decision-making paths that respondents take rather than focus on particular responses individually (Hollway and Jefferson, 1997). Given my focus on how respondents conceptualize their careers, their life histories, and their futures, foregrounding narrative themes, and sensemaking strategies allowed me to map out the different strategies respondents use. Specifically, it allowed me to see when and which key moments surfaced in the interviews, how these moments were explained, and how the introduction of these moments shaped the rest of the responses given by respondents. In turn, this method made visible the reflexive modes that respondents had by making explicit both the kinds of events that respondents experienced in regard to their work, how they interpreted them in the past, and how they interpret them now both consciously-in the form of the explanations they explicitly gave me-and tacitly-in the form of the interview itself in terms of which experiences they emphasized or left out (see Hollway and Jefferson, 1997;Randall and Koppenhaver, 2004).
Importantly, however, the data that I have collected is limited to the Canadian labor market and mostly to low-to midlevel income earners (in the $20,000-$50,000 CAD income range). While this limits the generalizability of my findings, by examining the turnover intentions of workers in occupations with mostly higher than average turnover rates-such as call center representatives and tradeshow salespersons-I am able to tap into the motivations of a group of individuals that are likely to be highly reflective about this topic due to having already had experience voluntarily exiting multiple jobs. Indeed, the majority of my sample have had more than three jobs in the last 5 years, attesting to the fact that these individuals frequently negotiate unmet expectations about their work.

ANALYSIS
While-in line with previous studies on the impact of referral status on employment and turnover decisions-I found that information and peer support were crucial dimensions of referral that impact respondents' decisions to stay in or leave their jobs, I also found that the reasons they cited were filtered through their broader perceptions about their careers and the role that work played in their lives. Career plans or career identity (see Ibarra, 1999) shaped how information and peer support led to respondents staying in or leaving their jobs by supporting, interrogating, or re-routing their career plans. However, these were also seen to be quite provisional and pliable in that when expectations went unmet individuals often appeared to spontaneously make the most of their situations.
The majority of respondents stated that they stayed in jobs because of the fact that they were referred to them. Reasons given range from a feeling of duty to their referrer, enjoying the work they were doing partly because of working alongside their referrer, and finding the job easier because of information given to them by their referrer. However, respondents varied in terms of how they viewed their jobs. This variation was dependent on where in a three stage process respondents found themselves in while working in any given job: As part of a plan, as making questionable decisions, or as being re-routed.

Supporting: Part of a Plan
Roughly a third of respondents framed their various occupations in terms of how well they fit into a broader career plan or roadmap that they had developed for themselves. How well a job fit into this plan tended to impact how long a respondent would stay in that job, with good fit leading to longer tenure and poor fit leading to shorter tenure.
An exemplar of such a narrative is Raoul, a software consultant at a tech firm who was referred to the company by his sister. He was employed as an elementary school teacher for over 10 years prior to his consulting career. While not seeking out this career on his own, but actively referred to it by his sister who thought he would be a good fit, he claimed in the interview that "the switch simply made sense. Given my expertise in teaching various kinds of contents, and influencing others, I believed I had the skills I needed to sell any software, any time, to any client." Moreover, he Raoul stated that the firm had a "corporate ideology" that matched his skill sets-they were "forward-thinking" and data driven, unlike the schools he had previously taught at which he felt were "bureaucratic" and "stuck in the past." Raoul's narrative shaped how he initially viewed this employment opportunity and augmented his daily interactions both with his referrer and with other co-workers and clients. This was made clear in his rising to the top of sales in his company in just two short years, a feat never accomplished by a consultant in his firm. "My peers were shocked, but I wasn't. I knew I had what it took to not only succeed, but to surpass all expectations and shame my peers. I came. I sold. I conquered [laughs]. And I will continue to do so!" Raoul does not feel the need to counsel his sister for support, having "learned how to sell the product extremely easily" early in his tenure at the software firm. However, he does "find comfort" knowing she is there as someone he can turn to in "times of drama. By that I mean when people come at me for being so good so fast." Thus, while Raoul's referral did not directly shape his experiences there, entering into the job with a relatively clear template of what to expect, what he could offer, and how he could learn to succeed in his role led to early success and long expected tenure.
Raoul's narrative draws a clear parallel to Larry's, a call center manager referred by a former co-worker. His career began in government directly after graduating with an MBA in rural Ontario, Canada. While he found his government work to be "moderately fulfilling, " he did not feel it was as "fun or interesting as it could be." He found that he was more creative and "forwardlooking" than his peers, and that they demonstrated bounded, relatively dull "old-style thinking": thinking and planning work activities in ways which tightly regressed onto existing patterns rather than worked to meet new and anticipated problems. Because of how good of a fit this occupation was for him, he felt no need to rely on his referrer to do his daily tasks.
Raoul and Larry's narratives demonstrate how for respondents who saw their jobs as part of a plan, being referred "made sense." For individuals with relatively strong, clear conceptions of their own career identities, being referred led to the development of desired, well-fitting new careers. Importantly, the role of the referrer for these individuals was not an important part of their career experiences. Due to the tight fit of these jobs for my respondents who saw a given job as part of their overall career plan, even conflicts with those individuals were subsumed in light of these larger career ambitions. For example, Raoul stated that he frequently quarreled with his sister outside of work, and that working together may have "actually hurt our relationship." However, given how much opportunity he felt he had with the tech firm, and how quickly he had risen the ranks there, he felt that "this isn't a huge deal. She'll come around!" Similarly, Archer, a barista at a coffee shop who was referred there by a former class-mate in a community college accounting class that he took several years ago, stated that even though his relations with his referrer had become tense the job was still "satisfactory:" I mean, I wish Gerry wasn't such a dick to me. But I get it-people change. This job is wicked, and I wouldn't leave it on account of our souring relationship. Maybe one day we'll mend things, right? Until then, I'm happy here. Great hours and decent pay, plus I get along great with the rest of the staff. I know I can work my way up, so I won't leave unless that doesn't happen in due time.
Thus, for respondents who fit in well into a job, relations with referrers seemed to decrease in relevance as time passed. What became more salient for respondents was the extent to which their current job fit their current career aspirations. However, for respondents with less clear career aspirations, this process was marked by much more ambivalence.

Interrogating: Questioning Motivations
A series of respondents found the jobs they were referred into, however, to sometimes be marked by ambivalence. For these respondents, jobs did not necessarily fit into or out of career plans. Instead, they were awkwardly placed within their plans. As a result, support from their referrers and the perceived quality of their workplace culture were often rather important to their tenure decisions.
This pattern is clearest in the case of Tyra, a customer service representative at a call center who had some time ago become disillusioned with the idea of "work." While she was grateful to her friend for referring her to her current job, instead of viewing this opportunity optimistically, she saw this job as simply "something I had to do. I needed money, but I really didn't want to work!" However, she went on to state that "if Sandy [her referrer] wasn't around, I think I'd go nuts and leave!" Upon further probing, Tyra stated that she "didn't really get along with [her] other co-workers. . . they're too into the job, man!" Tyra felt that her co-workers were "stooges, " in her words, individuals who "lived to work" and provided no opportunity for socializing, and "hanging out." As a consequence, Tyra's tenure at the call center heavily relied on the presence of Sandy, without whom she would be "alone in a sea of trolls." Unlike Tyra, Mick, an information technology (IT) support worker, initially liked the job he was referred into but came to strongly dislike it over time because of his referrer. Mick was referred by a friend and initially enjoyed the job, stating that "it seemed like a great place, the center. Lots of food, friendly coworkers, a live-and-live atmosphere, you know?" However, as his relations with his referrer waned, he came to question his decision. I just don't know why I'm here. I took this job because I thought Todd would be fun to work with. But he's not! I never knew how competitive he was. It's like every day is a battle, a battle for glory or something! Silly, I know, but I get really bad vibes from him now and really, really, really just don't want to be around him.
Mick felt that his relationship with his referrer had changed for the worse. As a result, while he still liked many aspects of his new job, the tense relationship with Todd made it hard to stay around.
Tension was experienced by these respondents, however, as they greatly appreciated the fact that their referrers went out on a limb to help them get their jobs. This is noted by Ezzy, who felt she owed it to her referrer, Tammy, to stay put at the bakery. I know Tammy worked real hard here for many years, and put her neck out to get me this job. I just, I just didn't know how I could go on any longer. It was killing me, knowing that this was a dead end. I just didn't see how it fit.
However, given the fact that she believed she would never assume a managerial position at the bakery, she simply could not imagine herself being happy there. Moreover, she viewed it as detrimental to the career she imagined for herself years ahead: I mean, I'm more interested in being a manager than I am in working at the cash. Hell, I'd even manage a hotel if I had to! I guess it took being in this situation to realize that it was managerial work that was really driving me. I feel bad for Tammy; she didn't know.
This experience thus further worked to enlighten Ezzy to the fact that she deeply wanted to work in managerial work. Due to barriers toward achieving this, her duty to her referrer was trumped by the realization of her potentially thwarted career aspirations of becoming a manager. As a result, she quit this job even though she believed being referred was "a big deal." Thus, among respondents who felt compelled to interrogate their career plans, whether the ongoing support that they expected from their referrers and the quality of the relationships they had on the job or the extent to which they were doing the tasks that they expected to perform appeared to spark this process differently based on their overall career aspirations. The starkest contrasts in this regard are Tyra and Ezzy, the former valuing workplace culture and relationships and the latter valuing managerial tasks.

Re-routing: New Career Paths
For respondents who felt dissatisfied with their jobs, the extent to which this dissatisfaction challenged their career aspirations impacted their next practical steps. Respondents would either seek out jobs which fit their career plans or re-route the roadmap they had developed prior to these experiences. Referrers played little role in shaping this extended process, as once a threshold of dissatisfaction was reached, respondents appeared impelled to re-frame or move on to resolve their vocational conflicts.
Linda is an exemplar of this type of respondent. She was referred into a job as an accountant by a schoolmate during her final semester of "accounting school." She began the job really interested, but soon realized that she needed much more autonomy. "I just realized it [accounting] wasn't doing it for me. I was stuck. I was bored. Bored because I couldn't do anything me way. . . always her [my boss's] way!" Linda noted how the routine tasks of her occupation led her to question her early motivation about being an accountant. Rather than work through this, however, and perhaps try to find ways to experience more of a sense of autonomy at this job, she "had to quit!" This was because she saw no feasible way to take control of her situation: The place was just very regimented. My boss scheduled all of my time, and even time with clients felt extremely regulated. I had no leeway over my schedule, over my interactions with clients, over any of my tasks. I thought this would continue no matter how long I stayed, and while I considered working at other firms, I didn't really think they'd be a whole lot different. So, I looked for other options. I looked for jobs that would be flexible and creative.
Realizing that she was stuck, Linda sought out "creative" jobs. Using her informal networks, she quickly found work at a bridal company doing promotional work. Though this work pays less than her accounting job, she "far prefers it." This new occupation enables her to do a variety of tasks and, although she is managed by others, she "can pick who she works with and leave situations [I] do not feel are suiting me." Her referrer was not mentioned in this narration of work events in any extended detail, signaling that this work prompted her to become reflexive about her blossoming career identity in a rather focused, emergent manner. This became evident as she went on to state that she could not imagine herself doing any other kind of work in the future, as this was "more creative and free than anything else [she'd] ever tried." Linda now wanted to continue her promotional work as it tapped into a desire for creativity and flexibility that she only now realized were important components of her career identity.
Similar to Linda, Robin found that rather jarring experiences at her first chosen occupation ultimately led her into something she "never anticipated." Robin studied fine arts and languages in university, with the intention of "either becoming a painter or an elementary school teacher." A few years into university, she decided that painting "wasn't going to pan out, " so she decided to focus her efforts on getting into teacher's college. Though never making it into teacher's college, she obtained employment at a local community center as an art instructor through a contact she made in an art class: The job was really quite good at first. I absolutely loved the students, and felt I was making a positive impact on their lives. I was really happy that my friend got me this job, and wanted to my best at it. But I didn't stay very long due to drama with my co-workers.
When probed about the nature of the "drama" she experienced, Robin relayed to me that what she thought was a relatively benign interaction with her boss wound up being interpreted as "nasty": I told the program organizer that I was having some issues with one of my students. I was sure I told her nicely and in a respectful manner, but the next day she called me into her office and told me that I was a "problem" case. I was rude, belligerent, and not accommodating with the student apparently. I was shocked.
This experience led Robin to question why she got into teaching in the first place. "It really was a fallback, the more I thought about it. I couldn't be a painter so I settled for this." Now, Robin is studying painting again and attempting to pursue this as her primary occupation. "It'll be hard, that's for sure, but I now know settling isn't worth it." As was the case for Linda, the referrer's role was not focused on, despite probes. This again signals how the emergence of a nascent sense of one's vocational desires can lead an individual into a state of inwardness and reflection not easily amenable to concerns of others-what Archer (2007, p. 93) would term an "autonomous reflexive" who has developed a refined sense of self-awareness and engaged on a determined plan of action. For Linda and Robin, these were new careers-and in Robin's case, the return to a previously discarded career.

DISCUSSION AND CONCLUSIONS
As demonstrated in my analysis, respondents tended to emphasize the fact that they felt a sense of duty to their referrers for helping them obtain employment. However, the weight placed on being referred varied for respondents, with some respondents remaining committed to their jobs and others feeling the need to move elsewhere.
Key in these divergent outcomes was the role of these referred individuals' initial expectations of their jobs. Those who ended up seeing the jobs they were referred into as part of a clear career plan tended to plan on staying longer. Conversely, those who were less sure up front about how the job would fit into their career plans tended to exit earlier. Moreover, for the former group tasks tended to matter more, while the latter were more concerned with peer support and workplace culture. However, this latter group also often stayed employed as they developed new career identities on the job, demonstrating support for perspectives which assert that career identity and vocational choices are quite provisional and pliable (Ibarra, 1999(Ibarra, , 2004. Thus, expectations appear to be the most salient determinant of job-staying and job-leaving in my sample. It was not necessarily the meeting of held expectations that mattered most, but rather it was having the experience of unmet expectations which most greatly impacted tenure (see Williams, 2018). Put simply, ironically it was more productive for respondents to have no career plans than any at all in terms of staying in the jobs they were referred into. This was found both across respondents and within respondents' own career histories, as while expectations did indeed help those with clearer career goals focus on ensuring that their tasks aligned with skillsets they desired to acquire and/or strengthen, for those with vaguer but still personally important expectations about what a work environment should be like the failure to meet these expectations often led them to cycle through many jobs to not real avail; respondents of this latter sort would relay throughout the interview that they tended to become disillusioned with work entirely, now having reached the conclusion that workplaces offer little and demand much.
However, the possession of expectations upon entering a job did not preclude respondents from developing new expectations while on the job. For respondents such as Robin, experiences at a given job could re-route initial plans rather than derail them.
As such, respondents demonstrated the importance of both dispositional and situational factors-expectations and ongoing peer support-for studying the impact of referrals on employment. While the expectations that individuals brought with them into their particular jobs impacted the extent to which they remained in them, these could either remain constant or change based on both the conditions of the jobs they went into and the quality of their interactions with their referrers. The relative absence of discussion about referrers is also quite telling. Once hired, respondents tended to focus on interactions with co-workers other than their referrers. This could signal that respondents felt referrers had done their part by simply helping them gain employment, and that, once employed, new relationships became central. Alternatively, the relationship with the referrer could become constrained as the newly entered occupational role may reduce the amount of a contacted an individual can have with their referrer. This was indeed the case for many of the respondents who stated that not only did they not work in close contact with their referrers but that their new co-workers provided the kind of support they needed more directly and in line with their actual work tasks.
The confluence of such dispositional and situational factors in the case of referred employees has implications for sociological analysis which extend far beyond this case. Rather than see dispositions as driving individuals to act in relatively stable and patterned ways (e.g., Vaisey, 2009), the divergences noted in my analysis signal that dispositions may be more malleable than they are frequently portrayed in frameworks such as that of the SDPM. Importantly, dispositions become augmented in light of situational contingencies. Dispositions still matter, however, as they impact how situations are initially perceived (see Williams, 2017c). That is, the information a referred employee brings into a job shapes what they expect from that job, but having one's referrer suddenly quit could impact just how much those expectations impact one's decision to stick around. The contingency of one's referrer exiting the job could lead the referee to re-evaluate his or her job now that the commitment to that referrer is gone-since they quit, they are now free to quit as well, having come to the realization that perhaps part of why they remained in the job was not actually congruence between job expectations and on-the-job duties, but rather a tacit commitment to their referrer.
Given this study's focus primarily on high turnover frontline occupations in the Canadian labor market, however, future research is needed to clarify the extent to which referral impacts turnover intentions in lower turnover occupations and in labor markets within other national contexts. Doing so could help further delineate the extent to which referred employees engage in spontaneous conflict resolution in light of unmet expectations in contexts where referrals are rarer, more familial in kind, or more widespread.

AUTHOR CONTRIBUTIONS
The author confirms being the sole contributor of this work and has approved it for publication.