Career shock ambivalence, univalence or indifference? Novel insights on the drivers and outcomes of gig workers’ responses to the disappearance of an online labor platform

Although scholars have long agreed that careers have become much more dynamic and volatile in the past decades, career research has not focused sufficiently on the role of significant disruptions in people’s careers (Akkermans et al., 2021). However, such disruptions are everywhere and happen to everyone, whether they be major global events such as the COVID-19 pandemic or an economic crisis or smaller and more personal events like a layoff, an unexpected promotion or a major private life event (e.g. having a baby or going through divorce). Research on such disruptive events has picked up momentum in recent years (see, e.g. Baruch and Sullivan (2022)), but is still in need of significant development to understand how such events may impact contemporary career development and success (Seibert et al., 2024).

Therefore, in this study, we examine the drivers and consequences of workers’ evaluation of a career shock. A career shock is “a disruptive and extraordinary event that is, at least to some degree, caused by factors outside the focal individual's control and that triggers a deliberate thought process concerning one's career” (Akkermans et al., 2018, p. 4). Research shows that career shocks can be positively and negatively valenced (Akkermans et al., 2018; Seibert et al., 2013), with existing empirical studies differentiating discrete career-related events based on whether these events are positive or negative. Examples of positive career shocks are considered an unexpected pay raise or promotion, while examples of negative career shocks are the sudden departure of a mentor, a layoff or an organizational bankruptcy (Akkermans et al., 2021; Seibert et al., 2013).

Typically, it will indeed be possible to accurately classify a career shock as either positive or negative. However, this may not always be the case. Specifically, some workers may view a disruptive career-related event to be simultaneously positive and negative (or neither positive nor negative). Insights from social psychology indeed show that for some people, a dualistic view on events as either positive or negative may not always apply (Baek, 2010). Instead, for those people, career shocks trigger ambivalence, which refers to a state where an individual holds both positive and negative feelings toward an event (Priester and Petty, 2001; Thompson and Zanna, 1995) and contrasts with the state of univalence (either positive or negative evaluation) and indifference (neither positive nor negative evaluation).

This paper adopts an approach to career shocks that highlights career shock valence and, crucially, the diversity of responses that workers may have to career shocks beyond univalent positive or negative reactions. To the best of our knowledge, little is known about the conditions under which workers evaluate a disruptive career-related event as negative or positive (univalent), both positive and negative (ambivalent) or neither negative nor positive (indifferent) – and how this relates to career-related behaviors. Studying career shock ambivalence, univalence and indifferences – and their antecedents and outcomes – is important for several reasons. First, it helps to show that workers’ attitudinal responses to career shocks are more complex than the binary positive versus negative responses that are traditionally assumed in the careers literature. This will assist future studies to better capture workers’ diverse, bipolar attitudinal reactions to unexpected, disruptive career-related events. Second, a study of career shock ambivalence, univalence and indifference helps to predict behavioral outcomes more accurately. In fact, it could help explain why individuals delay and fail to make changes in their career or engage in inconsistent career behaviors that contradict each other (Piderit, 2000; Akkermans et al., 2020; Verbruggen and De Vos, 2020). Finally, a study of the antecedents of career shock ambivalence will help to explain why workers develop contradictory experiences of unexpected, disruptive career events – or, on the contrary, are indifferent. This should help relevant stakeholders (e.g. managers, mentors and career coaches) to support workers in seeing the upside of a career shock and prevent situations in which workers feel paralyzed in making career changes needed to overcome a career shock.

Accordingly, our study aims to examine the antecedents and behavioral outcomes of univalence, ambivalence and indifference in workers’ responses to career shocks. To do so, we draw on unique data obtained from gig workers who experienced the sudden exit of an online labor platform (OLP) from the Dutch market. Gig workers are freelancers who rely on OLPs such as Uber, Deliveroo and Upwork to access short-term, paid work (Duggan et al., 2020; Meijerink and Keegan, 2019). Research shows that OLPs – after their initial rapid international expansion – have recently started withdrawing from various national markets as they jostle for dominant positions. This entails streamlining operations based on expansion strategies in what are often seen as winner-takes-all markets (Veen et al., 2025; Altenried, 2024) and can lead to sudden withdrawal from certain markets where the platform has been a prominent player.

The sudden disappearance of an OLP provides a unique research opportunity, and its study enables us to make the following contributions to the literature on career shock valence. First, the disappearance of an OLP is likely to be an unexpected and disruptive event as OLPs cease their operations literally within days. Deliveroo shut down its app one day after informing workers about its exit from the Australian market (Guardian, 2022) and announced its departure from Germany less than a week before terminating its operations (Altenried, 2024). The sudden disappearance of an OLP is a unique and distinct event, allowing us to examine the degree to which individual workers evaluate the associated career shock positively, negatively or in another way. Accordingly, our study contributes by shedding light on the patterns in ambivalence, univalence and indifference in career shock valence associated with a single career-related event for a specific group of workers.

Second, the OLP context is a useful empirical setting for explaining the drivers of patterns in career shock valence. The ambivalence literature has shown that a person’s evaluation of an event, object or actor (e.g. an employer) varies depending on which attributes (e.g. income, development opportunities or relationships with co-workers and supervisor) are considered important. This means that multiple attributes towards an event precede an individual’s ambivalent evaluation of that event (Baek, 2010). It is likely that gig workers render multiple OLPs’ attributes salient, given that gig workers are a highly heterogeneous group (Schor et al., 2020). On the one hand, research shows that a significant portion of gig workers is highly dependent on platform work for their income (Schor et al., 2020; Kuhn and Maleki, 2017). In such cases, the sudden exit of an OLP cannot be easily absorbed by a worker and may result in a negative career shock. On the other hand, as freelancers in precarious work settings, gig workers may acquire the core self-evaluation and resilience needed to absorb unexpected career events such as a sudden loss of income (Granger et al., 2022; Seibert et al., 2013). Platform work may be seen as a stepping stone toward a career as an entrepreneur (Newlands, 2024). In such cases, the disappearance of an OLP may also be seen as a last necessary “push” toward a new career trajectory, simultaneously functioning as a positive career shock. Accordingly, we examine whether the economic dependence of platform workers is positively related to a negative career shock and whether core self-evaluation relates positively to a positive career shock. In doing so, we contribute by showing the conditions under which gig workers experience the disappearance of an OLP as a negative shock (univalence, because of high economic dependence), a positive shock (univalence, because of high core self-evaluation), as both a positive and negative shock (ambivalence, because of high core self-evaluation and high economic dependence) or as neither a positive nor negative shock (indifference, because of low core self-evaluation and low economic dependence).

Finally, existing research has also shown that OLPs bring unique opportunities and challenges for gig workers’ careers (Zwettler et al., 2024; Duggan et al., 2022; Kost et al., 2020; Newlands, 2024; Akkermans et al., 2025). Most of these studies have examined gig workers’ careers while they are working via an OLP. As a result, we still know little about gig workers’ careers after they stop working via an OLP (for an exception, see Altenried, 2024). We advance knowledge by examining whether positive and negative career shocks relate to career crafting and career inaction. Career crafting and career inaction are opposite (or bipolar) career-related behaviors: career crafting concerns proactive behaviors that individuals perform to self-manage their career (Tims and Akkermans, 2020), which are more likely to occur when workers experience a positive career shock. On the contrary, career inaction concerns failure to act sufficiently over some period of time on a desire to make a change in one’s career (Verbruggen and De Vos, 2020) and is more likely to be driven by negative career shocks. By studying these bipolar career behaviors as outcomes, we contribute by showing that ambivalence in career shock valence also drives behavioral ambivalence in terms of workers’ simultaneously being inactive and proactive in making changes to their careers.

A career shock is a disruptive event that triggers career reflection (Akkermans et al., 2018). Hence, by definition, career shocks consist of both a discrete external event and an internal reflection (Akkermans et al., 2021). Importantly, a career shock should not be confounded with its potential impact on people's career paths (Akkermans et al., 2021). This means that a career shock can, but does not necessarily lead to actual changes in people’s careers. For example, a mentor leaving (event) may lead someone to consider leaving themselves (reflection), which constitutes a career shock. However, they may still decide to stay in the end (no change).

Career shocks can be positively and negatively valenced. For the purpose of this study, we define career shock valence as the negative and/or positive charge of unexpected events that trigger career reflection. As noted by Akkermans et al. (2018), whether a career shock is positive or negative depends on the individual’s experience of the external event, rather than the degree to which the underlying event has positive or negative consequences for the individual’s career or the organization that employs them. Although positively valenced shocks are more likely to lead to positive outcomes and negative shocks to negative outcomes, this is not always the case (Akkermans et al., 2020). For instance, an unexpected promotion (positive shock) can lead to increased career success (positive outcome) but also to stress and burnout (negative outcome) if a person was not ready for it yet. As such, career shock valence reflects the affect/emotions associated with the unexpected career-related event, rather than the career outcomes that this event brings about.

Existing studies classify individuals’ affective/emotional responses to distinct career-related events as either positive or negative. For instance, in one of the first empirical studies of career shocks, Seibert et al. (2013) assumed that people would uniformly and negatively respond to career shocks like “organization change” and “mentor departure” and positively to shocks such as “visible job success” and “quick raise or promotion.” Along similar lines, Holtom et al. (2012) – who examined negative shocks only – conceptualized responses to incidents like “disagreement with a supervisor” or “being encouraged to leave” as uniformly negative. Later studies follow the tradition to classify responses to distinct career shocks as either positive or negative, based on events such as “finding a job faster than expected after graduation” or “encounters with influential people” as positive career shocks and “a layoff,” “organizational bankruptcy” or “being overwhelmed by a first job” as negative career shocks (Blokker et al., 2019; Rummel et al., 2021).

Classifying career shocks into positive and negative has merit, as they trigger different behaviors and outcomes (Akkermans et al., 2020), thus enabling more accurate measurement and theorizing (Akkermans et al., 2021). However, research on career shocks has thus far overlooked the possibility that workers may view a distinct career-related event as simultaneously positive and negative (or neither positive nor negative). According to social psychologists, a state in which individuals hold both positive and negative feelings toward an object, person, or event can occur and manifests as ambivalence (Priester and Petty, 2001; Thompson and Zanna, 1995). Such bipolar, ambivalent feelings contrast to (1) the state of univalence, in which individuals hold either a positive or negative feeling (but not both) toward an object, person or event and (2) the state of indifference, which manifests when an individual experiences neither a positive nor a negative feeling. Applied to career shocks, what extant studies have done so far is conceptualize workers’ emotional/affective responses to career shocks as states of univalence (either positive or negative), while overlooking states of career shock ambivalence and indifference. To the best of our knowledge, little is known about the degree to which career shock univalence, ambivalence and indifference manifest, under what conditions they manifest, and what implications they have for career-related behavioral outcomes. This is problematic, as it (1) obscures the complexity in career shock evaluations, (2) presents missed opportunities for explaining bipolarity in career behaviors of individual workers and (3) overlooks ways for how ambivalent/indifferent responses can be prevented and univalent, positive career shock evaluations fostered. Accordingly, we examine the drivers and outcomes of career shock (uni-/ambi)valence.

To better understand these drivers and outcomes, we further draw on the ambivalence literature, which has shown that a person’s evaluation of an event, object, or actor varies depending on which attribute(s) is considered more seriously (Baek, 2010). Taking the employer as an example of an object of study, workers may perceive different attributes of that employer – such as compensation and benefits, development opportunities or relationships with co-workers and supervisor – more or less relevant. Understanding the saliency of multiple attributes is relevant as this antecedes an individual’s ambivalent evaluation (Baek, 2010). If workers deem only one attribute of an object, event or person to be relevant, they are more likely to develop univalent assessments of that object, event or person. Workers who experience a career shock as either positive or negative are thus likely to emphasize only one attribute of the career shock, while those who valence a career shock as both positive and negative likely consider multiple, bipolar career shock attributes relevant. As shown by the review study of Shuman et al. (2013), valence depends on differences in needs, goals, values and/or norms of individuals, which implies the need to study different antecedents – i.e. at least one that explains positive valence and at least one that explains negative valence – to understand why univalence, ambivalence or indifference in career shock experiences manifest. As discussed next, we propose that the gig economy represents a unique context that puts forward two such relevant antecedents (i.e. economic dependent and core self-evaluation), which helps to better understand the patterns in career shock univalence, ambivalence and indifference.

The gig economy is an economic system where online labor platforms (OLPs) such as Uber, Deliveroo or Fiverr match supply and demand for freelance labor. OLPs selectively combine principles of the market and hierarchy to coordinate economic activities (Frenken et al., 2020; Meijerink et al., 2021; Vallas and Schor, 2020; Robinson et al., in press). They externalize their gig workforce by classifying workers as independent contractors who are nominally free to compete on price or service (Keegan and Meijerink, 2023). Gig workers bear significant economic risks because OLPs neither pay for workers’ idle time nor compensate for any sudden loss of income. OLPs simultaneously subvert these free-market principles by restricting gig workers’ autonomy using algorithm-based human resource management (HRM) activities such as performance appraisal (e.g. online rating schemes), workforce planning (e.g. shift systems) and remuneration policies dictated by platforms (e.g. pricing algorithms) (Meijerink et al., 2021; Veen et al., 2020; Wood et al., 2019). The use of such HRM activities is an indicator of hierarchical control that characterizes employment relationships (Frenken et al., 2020; Meijerink and Keegan, 2019; Robinson et al., in press). As a result, platform-based gig work is conceptualized as a blend of independent contracting (i.e. freelancing) and employment (i.e. hierarchical control by OLPs). Accordingly, gig workers have been termed “dependent contractors,” that is, workers who are externalized by OLPs but simultaneously dependent on OLPs’ discretionary power and control (Kuhn and Maleki, 2017). As such, there are various work-related attributes that gig workers may deem more or less relevant, meaning that gig workers may develop ambivalent responses to work-related events (such as career shocks).

The “dependent contractor” status has attracted scrutiny from various societal stakeholders. OLPs have been criticized by labor unions, journalists, politicians and labor activists for creating “false self-employment” (Meijerink et al., 2021; Veen et al., 2025). OLPs aim to prevent the reclassification of gig workers as employees because this would significantly increase labor costs. In countries like Australia, the Netherlands and Spain, governmental agencies have successfully challenged OLPs’ dependent contractor model. It is therefore not surprising that an OLP such as Deliveroo exited these countries in 2019 (Germany), 2021 (Spain) and 2022 (Australia and the Netherlands). Their use of the dependent contractor model enables OLPs to exit local markets swiftly due to their limited obligations vis-à-vis their gig workforces (Meijerink and Keegan, 2025; Altenried, 2024). Gig workers are not protected from being “laid off” by an OLP, as the latter does not employ them. At the same time, the idea that gig workers cannot be “laid off” in the context of platform work is complicated due to the ambiguous employment situation created by platforms using a dependent contractor model. Formally, platforms deny they employ gig workers, and so from that perspective, laying off gig workers is not obvious. However, from the workers’ perspective, the dependence of contractors on platforms (Schor et al., 2020) raises the possibility that a rapid market exit by a platform is, to a large extent, similar to a situation where workers are laid off. Such ambiguities and complexities are rife in the context of gig work and require a nuanced perspective that does not rely on notions of work and employment built on clear distinctions between those employed and those not employed. That said, limited obligations enable OLPs to exit markets literally within days. For instance, Deliveroo shut down its app one day after announcing its exit from the Australian market to workers (Guardian, 2022). Similarly, when leaving the German market, Deliveroo announced its departure less than a week before closing access to its app (Altenried, 2024).

The sudden and swift disappearance of an OLP likely constitutes a career shock. Gig workers may perceive the disappearance of an online labor platform as a negative career shock, a positive one, or a combination of both (see also Tamontseva et al. (2025)). On the one hand, it is likely that they experience the disappearance of an OLP as jarring because gig workers can incur a loss of income (Altenreid, 2024), part of their identity (Van Fossen et al., 2024) or access to a “good” bad job (Wood et al., 2019). On the other hand, the disappearance of an OLP may also be seen as beneficial by workers. OLPs aim to lock in workers to prevent them from moving to rival platforms and to drive down prices (Meijerink et al., 2021; Veen et al., 2020). The sudden disappearance of an OLP may be perceived as a welcome event for workers that frees them from the shackles of an (exploitative) situation perpetuated by an OLP. Additionally, workers may see gig work as a stepping stone toward a career path as an entrepreneur (Newlands, 2024) or more traditional forms of employment. In these cases, the exit of an OLP may be perceived as a desirable nudge to (finally) embark on an entrepreneurial career path or to search for more standard employment.

The above implies that the sudden disappearance of an OLP likely constitutes a career shock to workers, to which they may develop bipolar/ambivalent responses. To better understand why workers develop positive and/or negative evaluations, we examine economic dependence and core self-evaluations as relevant (bipolar) antecedents. The choice for these two variables follows from the dependent contractor status of gig workers, which may explain career shock ambivalence. That is, economic dependence reflects the situation where gig workers bear significant economic risks (i.e. they are dependent parties in dependent contractor models), while core self-evaluation reflects the situation where gig workers need to develop resilience, creativity and an entrepreneurial attitude (i.e. they are contractors in dependent contractor models and carry responsibility as an equal party to a contract). We expect that economic dependence is positively correlated with workers perceiving Deliveroo’s disappearance as a negative career shock due to the dependent nature of the work situation, and core self-evaluation to be positively correlated with perceiving it as a positive career shock, as it offers new options and pathways for entrepreneurial contractors. As such, besides aligning with the dependent contractor model, a study of these two independent variables also helps to shed light on whether gig workers respond to a career shock in a univalent positive way (because of high core self-evaluation and low economic dependence), in a univalent negative way (because of low core self-evaluation and high economic dependence), in an ambivalent way (because of high core self-evaluation and high economic dependence) or in a state of indifference (because of low core self-evaluation and low economic dependence) (see Figure 1).

There are many antecedents that can explain negative career shock perceptions among workers (see Akkermans et al., 2018, 2021 for an overview and discussion). We consider economic dependence to be one of those highly relevant antecedents as it sits right at the heart of the “dependent contractor” status that characterizes gig workers (Kuhn and Maleki, 2017) and because research has shown that gig workers’ (negative) work-related experiences are highly contingent on their economic dependence on gig work (Fisher et al., 2025; Schor et al., 2020). Economic dependence refers to the degree to which an individual’s financial situation and living conditions are contingent on the income derived from their work as a gig worker (Kuhn and Maleki, 2017; Schor et al., 2020). As noted by Schor et al. (2020), while economic dependence is relevant in various workplaces, its importance is heightened in the context of gig work given the heterogeneity among workers. While some perform gig work to earn money in addition to a main source of income, other gig workers have to work full-time to make ends meet (Keith et al., 2019; Schor et al., 2020).

We expect that workers who are highly dependent on gig work for their income are more likely to experience the disappearance of an OLP as a negative career shock. This is consistent with conservation of resources (COR) theory, which predicts that individuals are strongly motivated to avoid resource loss, as this creates stress (Halbesleben et al., 2014). We follow Hobfoll (1989), who refers to resources as “objects, personal characteristics, conditions or energies that are valued by the individual” (p. 516). In the case of gig work, access to paid freelance labor is a key resource provided by OLPs. Indeed, Granger et al. (2022) showed that gig workers experienced the COVID-19 pandemic as a shocking event that caused resource loss due to a substantial reduction in paid gig work. From a COR theoretical perspective, the exit of an OLP amounts to a significant resource loss (i.e. opportunity to generate income), which creates stress for workers, thereby likely creating perceptions of a negative career shock. This particularly holds for gig workers who are strongly economically dependent on gig work. First, the share of income generated from gig work is higher for economically dependent gig workers. Gig workers who are less economically dependent have significant supplementary sources of income or their living conditions are less dependent on platform-generated income (Kuhn and Maleki, 2017; Schor et al., 2020). Second, those who are economically dependent on gig work spend more hours per week working via an OLP. From a COR theoretical perspective, economically dependent gig workers thus invest a personal resource – that is, time. According to COR theory, individuals with limited resources (e.g. time spent) are more likely to experience high levels of stress and, thus, a depletion of energy to reinvest resources (Hobfoll, 2001; Halbesleben et al., 2014). Moreover, economically dependent gig workers that exhibit long working hours are more likely to experience technology overload and job insecurity, which have shown to be positively related to burnout – and thus resource loss – among gig workers (Bunjak et al., 2021; Hajiheydari and Delgosha, 2024). As such, rather than seeing the disappearance of an OLP as uplifting, economically dependent gig workers are more likely to view it as a negative event that equates to the (further) loss of resources. Taken together, we therefore propose the following hypothesis:

Core self-evaluation refers to a higher-order trait composed of four traits of an individual: (1) self-esteem, the overall value that a person place onto themselves; (2) generalized self-efficacy, the self-evaluation of how well the individual is able to perform across a variety of situations, (3) emotional stability, the tendency to have a positive approach to explaining events and focus on positive aspects of the self and (4) locus of control, one’s beliefs about the cause of events in one’s life with locus being internal when individuals believe that events are contingent on their own actions (Judge et al., 2003; Chang et al., 2012).

We expect that workers with high levels of core self-evaluation are more likely to experience the disappearance of an OLP as a positive career shock. This is consistent with COR theory, which predicts that individuals with ample personal resources – such as high levels of core self-evaluation – are better able to absorb upsetting events (Hobfoll, 2001). First, individuals with high core self-evaluations are more likely to thrive outside the gig economy, as their generalized self-efficacy enables them to perform well across a variety of situations. Indeed, research has shown that core self-evaluation is positively related to an individual’s perceived employability and ability to obtain a (new) job (Onyishi et al., 2015) as well as career competencies (Tims and Akkermans, 2017) and career adaptability (Chen et al., 2023). Given that individuals with high generalized self-efficacy are more positive about their employability and adaptability, we expect them to be more likely to view the disappearance of an OLP as a positive event. This is consistent with insights from marketing research, which shows that individuals' evaluations of events (i.e. service encounters) are dependent on their self-efficacy (McKee et al., 2006). Specifically, research has shown that self-efficacy is positively related to the use of active coping strategies that enable consumers to find positive resolutions after service failures (Tsarenko and Strizhakova, 2013). Along similar lines, we expect that workers – as consumers of OLPs’ intermediation services – are more likely to view the sudden disappearance of an OLP as a positive shock when they have high levels of core self-evaluation.

Second, the emotional stability dimension of core self-evaluation likely explains why workers view the sudden disappearance of an OLP as a positive event. Emotionally stable individuals are more likely to view events as uplifting, are better able to cope with unexpected events and are more likely to appraise stressful situations as opportunities (Gallagher, 1990). Accordingly, we expect that those who are emotionally stable are more likely to develop positive career thoughts. On this basis, we expect that workers who score high on emotional stability are more likely to experience the disappearance of an OLP as a positive shock.

Finally, depending on their internal locus of control, workers are more likely to view career shocks as positive. Workers with an internal locus of control are more likely to attribute change to themselves and are more likely to have a growth mindset. Individuals with a growth mindset believe that they can bring about change and improvements themselves (Yeager and Dweck, 2020). Indeed, research has shown that a growth mindset is associated with positive outcomes such as enjoyment, resilience and engagement (Frondozo et al., 2022; Guo et al., 2023). According to Seibert et al. (2013), these very attributes, such as resilience, increase the likelihood of being able to absorb career shocks and see the bright side in sudden career-related events. We therefore propose the following hypothesis:

To understand how gig workers respond to career shocks, we examine career crafting and career inaction as behavioral responses. Career crafting and career inaction are bipolar, career-related behaviors, and thus relevant to study, as they help to better understand the consequences of ambivalence in career shocks experiences of workers. Career crafting refers to “proactive behaviors that individuals perform to self-manage their career and that are aimed at attaining optimal person–career fit” (Tims and Akkermans, 2020, p. 176). It includes cases where workers proactively reflect and think about their career aspirations and what they value in their career, as well as when workers proactively update career goals and share them with others (Tims and Akkermans, 2020). Career crafting can be beneficial to individuals by increasing career success (Janssen et al., 2021) and career sustainability (Ge et al., 2023). More broadly speaking, proactive career behaviors, including career crafting, are central to modern careers and relate to various beneficial outcomes, such as employability and success (Jiang et al., 2023).

On the contrary, career inaction refers to an individual’s “failure to act sufficiently over some period of time on a desire to make a change in one's career” (Verbruggen and De Vos, 2020, p. 378). Career inaction occurs when a person wants to make a change in their career yet does not act sufficiently on their desire to change over a prolonged period of time (Verbruggen and De Vos, 2020) and relates to unfavorable career outcomes, such as lower well-being and career satisfaction and more career regret (D'Huyvetter et al., 2025). We examine these bipolar career-related behaviors to better understand the consequences of univalence, ambivalence and indifference in career shock experiences. Specifically, we expect that univalent career shock experiences result in univalent career outcomes (i.e. either high career crafting or high career inaction), that ambivalent career shock experiences result in ambivalent career outcomes (i.e. high career crafting as well as high career inaction), and that indifferences in career shock experiences result in indifference in career outcomes (low career crafting and low career inaction).

We expect that workers who experience the disappearance of an OLP as a positive career shock are more likely to engage in career crafting. This is consistent with COR theory, which argues that individuals with generous resource pools are better equipped to reinvest and build additional resources (Hobfoll, 2001). A positive career shock signals to workers that a career-related event provides valuable resources to them in the form of, often unexpected, career opportunities (e.g. the chance to transition into entrepreneurship or explore a new career trajectory). In line with COR theory (Halbesleben et al., 2014), employees who experience an unexpected yet resourceful career event are more motivated to increase their resource base through career crafting (e.g. proactively building or strengthening relationships with others to inform them about their career aspirations or seeking opportunities to increase their person-career fit).

On the contrary, workers who experience the disappearance of an OLP as a negative career shock are more likely to report career inaction. Career inaction is often accompanied by negative emotions and stress (Rogiers et al., 2022) and can impact people’s well-being, health and performance (D'Huyvetter et al., 2025). In line with COR theory (Halbesleben et al., 2014), we expect that career inaction is more likely to occur when workers incur a negative career shock. That is, a negative career shock is a resource-draining event that triggers negative emotions such as anxiety, anger or apathy. This prevents gig workers from spending sufficient time on making career-related changes and thus experience career inaction (i.e. not act sufficiently on their desire to change). We therefore propose the following hypotheses:

To test our hypotheses, we used survey data obtained from a sample of former Deliveroo workers in the Netherlands. Deliveroo is a food delivery platform that matches restaurants, consumers and workers (so-called “riders”). Deliveroo was a suitable setting for testing our hypotheses for at least two reasons. First, its riders are “dependent contractors” because Deliveroo selectively combines indicators of independent contracting and employment for managing its workforce (Meijerink et al., 2021; Veen et al., 2025). Second, Deliveroo ceased its activities in the Dutch food delivery market in December 2022. This disappearance can be considered an extraordinary event for multiple reasons: (1) Deliveroo was the first gig-based food delivery platform to leave the Dutch market; (2) its disappearance left approximately 2,500 riders without a gig-based job and (3) Deliveroo negotiated a layoff agreement with the Dutch labor union FNV. The latter is unprecedented, as labor unions traditionally do not represent non-employees such as gig workers. In addition, the departure of Deliveroo was broadly covered by the Dutch media and debated by spectators such as labor unions and politicians. As such, the departure of Deliveroo served as a unique case for testing our hypotheses regarding gig workers’ uni-/ambivalent experiences of an OLP’s disappearance.

At the start of the research, there was neither a publicly available list of Deliveroo riders in the Netherlands nor a complete overview of the workforce comprising gig workers that Deliveroo was willing to provide or disclose. It was, therefore, impossible to invite all Deliveroo riders in the Netherlands to participate in our survey. We nevertheless managed to access a significant number of potential respondents with the support of an organization that provided administrative services to 896 Dutch Deliveroo riders. To ensure minimal dependence on this organization, we were provided with a database containing the email addresses of these Deliveroo riders. This enabled us to directly invite these individuals, evaluate the response rate and send direct reminders to the participants ourselves. We invited all 896 riders by email to complete an online survey in February 2023. The two-month time gap between Deliveroo’s departure and the distribution of our survey was chosen for two reasons. First, career shocks are disruptive events that trigger career reflection (Akkermans et al., 2018). The two-month time lag allowed workers to reflect on Deliveroo’s sudden appearance, yet it is not too long for them to forget about their evaluation of this sudden career-related event. Second, from a more ethical perspective, we did not want to pounce on workers right after they were laid off and probe them on that experience, which is why we sent out the survey two months after Deliveroo’s exit from the Dutch market. After receiving the survey in February 2023, the respondents had three weeks to complete it and received two reminders by email. To safeguard confidentiality, their responses were directly sent to the research team. The responses did not contain any missing values because our online survey included a forced-entry feature ^[1].

In total, we received 227 completed surveys, which equals a 25% response rate. Of the respondents, 87% identified themselves as male and 10% did not have a permit to work in the Netherlands. The respondents were 28.5 years old on average (SD = 7.5 years) and had worked via Deliveroo for 2.9 years on average (SD = 1.5 years).

Whenever possible, we relied on validated scales to measure our variables and used a five-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree) as answer options. Two researchers translated the items from English to Dutch and back again to ensure the correct representation of translated items, following best practice principles of back translation (Klotz et al., 2023). We pilot-tested the survey with five former Deliveroo riders to ensure the content validity and understandability of our survey items. The pilot test resulted in minor changes to the item wording that were included in the final version of the survey instrument.

Economic dependence was measured using a four-item “job dependency” scale developed by Clark (2005), which measures workers’ perceptions of the degree to which their financial situation and living conditions are contingent on the income from a particular job. We adjusted the items such that they reflect workers' economic dependence on the Deliveroo platform. Example items are “My income from my work at Deliveroo was important to me” and “To maintain the standard of living I desire for myself, I had to keep working with Deliveroo.” The four-item measure had a Cronbach’s alpha of 0.83.

Core self-evaluation was measured using the scale validated by Judge et al. (2003). The original scale includes 12 items that each measure one of the four components of core self-evaluation, that is, self-esteem, generalized self-efficacy, emotional stability and locus of control. A sample item is: “When I try, I generally succeed.” Half of the items in the original scale are negatively worded (e.g. “Sometimes I feel depressed”). The confirmatory factor analysis revealed that these six items did not significantly load onto the latent core self-evaluation construct and these were, therefore, left out ^[2]. Each of the four components of core self-evaluation was measured by the same amount of positively and negatively worded items. As such, all relevant components of the core self-evaluation construct were measured despite the fact that the negatively worded items were left out. The scale that included the remaining six items that were positively worded had a good reliability (α = 0.82).

Positive career shock was measured using a five-item scale ^[3] which measures the degree to which a worker had experienced an unexpected, yet fortunate career-related event in the past six months that was helpful to them. We customized the items to ensure they reflect the departure of Deliveroo as a career shock. An example item is: “The disappearance of Deliveroo was a positive event that caused me to rethink what I can do to achieve my career aspirations.” The five-item measure had a Cronbach’s alpha of 0.88.

Negative career shock was measured using a five-item scale ^[3] that measures the degree to which a worker had experienced an unexpected, yet fortunate, career-related event in the past six months that was difficult for them. Similar to the positive career shock scale, we slightly reworded the items to ensure they reflect the disappearance of Deliveroo from the Netherlands. An example item is: The disappearance of Deliveroo was an adverse event that triggered me to think about how I might attain my career goals in a different way. The Cronbach’s alpha of the five-item scale was 0.90.

Career crafting was measured using the eight-item scale developed by Tims and Akkermans (2020), which measures the degree to which workers engage in proactive career reflection and construction. For each item we asked workers to evaluate their proactive career crafting behaviors since the departure of Deliveroo. Two example items are “I deliberately think about what I would like to achieve in my career” and “I deliberately show to others what I am good at.” The eight-item measure had a Cronbach’s alpha of 0.91.

Career inaction was measured using the validated scale developed by D'Huyvetter et al. (2025). The scale includes nine items that assess the degree to which a worker fails to realize their career desires. Two example items are “I don't manage to realize my career desires” and “I find it difficult to take action to change something in my career.” The nine-item measure had a Cronbach’s alpha of 0.93.

When testing our hypotheses, we controlled for worker tenure, age, level of education, average weekly working hours via Deliveroo and permit to work in the Netherlands. Age (measured in years) was included since workers likely become more economically dependent on (platform) work when they grow older (e.g. having to support their families and paying rent/mortgage). Tenure on the Deliveroo platform was measured in years and was included because economically dependent workers are less likely to leave an OLP (Schor et al., 2020). Level of education was measured using an ordinal scale, with employees indicating the highest level of education they had completed, and was included because education is positively related to core self-evaluation (Zacher, 2014). We asked the respondents to indicate the average of weekly hours they worked via Deliveroo because this correlates with economic dependence on platform work (Schor et al., 2020). Finally, we measured whether workers have a permit to work in the Netherlands using a dichotomous scale. This control variable was included because qualitative research shows that migrant workers (without a working permit) are strongly dependent on platform work for their income (Altenried, 2024).

A measurement model with six latent variables (i.e. economic dependence, core self-evaluation, positive career shock, negative career shock, career crafting and career inaction) was tested to verify whether the measurement items are indeed adequate indicators of their respective underlying construct. This model had an acceptable fit to the data (χ²₍₅₉₆₎ = 1.54; comparative fit index (CFI) = 0.95 and root mean square error of approximation (RMSEA) = 0.05). Alternative models with fewer latent constructs (e.g. all career shock items loading onto one overall career shock factor) fitted significantly worse in comparison to the six-factor model, which was therefore retained.

To test our hypotheses, we employed structural equation modelling (SEM) using Analysis of Moment Structures (AMOS) because it combines factorial analysis with multiple regression analyses. Moreover, it affords estimating relationships involving multiple (in)dependent variables, which is relevant for our current study that seeks to examines multiple antecedents and outcomes of career shocks. For testing our hypotheses and establishing model fit, we used maximum-likelihood estimation and report on the χ² statistics and significance levels. Since χ² significance is affected by the sample size and correlational effect sizes, we also used Hu and Bentler’s (1999) fit criteria, including the CFI and the RMSEA. A well-fitting model was defined as one that had χ²/df below 2.0, an RMSEA below 0.06 and CFI values above 0.95.

The means, standard deviations and correlations among our study variables can be found in Table 1. Most of the proposed variables included in our structural model are significantly related. Moreover, positive and negative career shocks are significantly and positively correlated (r = 0.44, p < 0.01). This shows that gig workers who positively valence the disappearance of Deliveroo are also more likely to negatively valence it at the same time, which suggests that positive and negative career shocks are coexisting (i.e. manifest as career shock ambivalence). Moreover, Table 1 shows that older workers are more likely to experience the disappearance of Deliveroo as a positive career shock (r = 0.24, p < 0.01) and negative career shock (r = 0.27; p < 0.01), indicating that older workers are more likely to be ambivalent in their perceptions of the sudden disappearance of an online labor platform. The number of weekly hours that a gig worker worked on average via Deliveroo is positively correlated with economic dependence (r = 24, p < 0.01) and negative career shock (r = 0.30, p < 0.01). This suggests that Deliveroo riders that were highly dependent on gig work were more likely to work a large number of hours per week via Deliveroo, and that the disappearance of Deliveroo is considered more negative in case gig workers spent a lot of time working via Deliveroo on a weekly basis. Finally, economic dependence and core self-evaluation are negatively related (r = −0.45, p < 0.01), suggesting that workers who have low self-esteem, emotional stability and generalized self-efficacy are more likely to be dependent on gig work.

Finally, in order to assess the number of individuals who experienced the disappearance of Deliveroo in a univalent (positive and negative), ambivalent or indifferent manner, we relabeled their responses into binary categories (e.g. high versus low positive/negative career shock). We took the mean values of the two career shock valence dimensions (see Table 1) as cut-off values. Workers were classified as reporting low (vs high) levels of positive career shock valence when falling below (vs reaching above) the 2.90 threshold. The cut-off value for negative career shock was 2.76. Using these cut-off values, we found that 47 (out of the 227) respondents had ambivalent responses to the sudden disappearance of Deliveroo (i.e. they experienced it as both a high positive and a high negative shock), while 110 respondents were indifferent to the exit of Deliveroo from the Dutch food delivery market (i.e. neither experienced it as a positive shock nor as a negative shock) (see Figure 2). Univalent, positive responses were found in 41 cases, while 29 respondents reported univalent, yet negative, responses to the sudden disappearance of Deliveroo.

In line with our hypotheses, we fitted a model that includes direct effects of core self-evaluations and economic dependence on positive and negative career shock types, respectively, as well as direct effects of positive and negative career shocks on career crafting and career inaction, respectively (see Figure 3). This model had an acceptable fit (χ²₍₇₆₉₎ = 1.52; CFI = 0.93; RMSEA = 0.05) and was therefore used for testing our hypotheses.

Hypothesis 1 stated that economic dependence would be positively related to workers’ perceptions of a platform’s disappearance as a negative career shock. As Figure 3 shows, there is a significant and positive relationship between economic dependence on platform work and negative career shock (β = 0.13, p < 0.05). Therefore, we can accept Hypothesis 1.

Hypothesis 2 stated that core self-evaluation would be positively related to workers’ perceptions of a platform’s disappearance as a positive career shock. As shown in Figure 3, the relationship between core self-evaluation and positive career shock is significant and positive (β = 0.21, p < 0.01), lending support to Hypothesis 2. Hypothesis 3 stated that positive career shocks would be positively related to career crafting. As shown in Figure 3, positive career shocks are positively related to career crafting (β = 0.12, p < 0.05), which supports our Hypothesis 3. Finally, Hypothesis 4 stated negative career shocks would be positively related to career inaction. As expected, and as shown in Figure 3, the relationship between negative career shocks and career inaction is significant and positive (β = 0.55, p < 0.001), lending support to Hypothesis 4.

This study aimed to examine whether and how platform-based gig workers may have experienced the sudden disappearance of a food delivery platform as a positive and/or negative career shock and tested several antecedents and outcomes of these shocks to better understand workers’ univalent, ambivalent or indifferent responses to career shocks. First, our results confirmed that our participants had varying interpretations of the event, including both positive and negative as well as ambivalent (both positive and negative) and indifferent (neither positive nor negative) responses. Second, our findings support the study’s hypotheses. As expected, gig workers who were more economically dependent on the platform were more likely to experience the platform’s disappearance as a negative career shock. Moreover, those high on core self-evaluations were (also) more likely to experience this event as a positive shock. In terms of outcomes, as hypothesized, when the platform’s disappearance was experienced as a positive career shock, gig workers were more likely to engage in career crafting; when the event was a negative shock, they were more likely to experience career inaction. Our results have several implications for research and practice, which we reflect upon below.

First, our results show that positive and negative career shocks can coexist. We found a positive correlation between both career shock types, indicating that workers who negatively valence the sudden disappearance of an OLP are more likely to positively valence it as well (and vice versa). Moreover, approximately a quarter of the respondents experienced the career shock under study as both positive and negative. These findings directly contribute to the literature on career shocks by showing that the valence of career shocks is not necessarily a mutually exclusive phenomenon (i.e. an “either/or” situation) but, instead, that positive and negative shocks can co-occur for individual workers (Akkermans et al., 2018, 2021). Moreover, this is a first indication that the dependent contractor status of gig workers engenders experiences of the disappearance of an OLP as both a negative and positive career shock. These results indicate that the novel context of gig work may reframe career-related phenomena such as redundancy and layoffs in new and as yet poorly understood ways. Platform work enabled by algorithmic technology creates a grey area for employment questions (Keegan and Meijerink, 2025) and suggests a need to reconsider assumptions in career research linked with definitive ideas on employment, non-employment and their implications for careers and experiences of career shocks.

Second, our research sheds light on why an individual worker may experience a career shock to be both positive and negative by studying economic dependence and core self-evaluation as relevant, bipolar antecedents of career shock valence. As expected, economic dependence on platform work was positively related to negative career shocks, while core self-evaluation was positively related to positive career shocks. This implies that gig workers can develop ambivalent experiences of the sudden disappearance of an OLP when they are highly dependent on platform work for their income and when they have high levels of self-esteem, generalized self-efficacy, emotional stability and internal locus of control (i.e. score high on core self-evaluation). Both antecedents tap into different aspects of the dependent contractor status of gig workers – that is, gig workers are economically dependent parties vis-à-vis the OLP, yet also solo self-employed workers who need high levels of core self-evaluation to develop resilience, creativity and an entrepreneurial attitude (Kuhn and Maleki, 2017). While existing studies have shown that gig workers’ experiences are affected by their dependent contractor status when working via an OLP (Meijerink et al., 2021; Veen et al., 2025), our results imply that post-platform experiences (i.e. workers’ reaction to the exit of an OLP) are equally affected by this “dual” dependent contractor status. For the broader career shocks literature, this implies that our understanding of workers’ ambivalent experiences of other career shocks (e.g. sudden layoff, promotion or mentor departure) can be better understood by examining bipolar antecedents.

Finally, our results imply that ambivalence in career shock perceptions also triggers ambivalence in career behavioral responses. As expected, workers who experience the disappearance of an OLP as a negative career shock are more likely to enter a state of career inaction, while those who (also) experience a career shock as positive are more likely to (also) engage in career crafting. We directly contribute to the nomological net of career shocks (Akkermans et al., 2021), showing that, at least among our sample of platform-based gig workers, positive career shocks are likely to trigger proactive career behavior, whereas negative career shocks can result in a state of inaction that does not allow them to make a desired change. For workers who are ambivalent in their career shock evaluation, this creates a striking, bipolar situation characterized by them simultaneously making proactive (i.e. career crafting) and inactive (i.e. career inaction) career-related changes. This likely occurs provided that individuals in ambivalent states emphasize different elements of an event (Baek, 2010). Applied to careers, this implies that individuals with ambivalent career shock experiences are inactive when making changes to some aspects of their career (e.g. expanding their network) while simultaneously making proactive changes to other aspects (e.g. developing career competences). Gig workers who are ambivalent toward the disappearance of an OLP may, for instance, be very proactive in further developing the knowledge, skills and abilities needed to find a new job (as a result of positive shock experiences that are triggered by high core self-evaluation) while, at the same time, being very inactive in actually searching for a new job (as a result of negative shock experiences that are triggered by high economic dependence).

Our findings also have strong practical and policy implications. Our results imply that those who are highly dependent on gig work for their income are more likely to experience the disappearance of an OLP as upsetting than those who are less economically dependent on platform-generated income. Our correlation analysis shows that those who are highly economically dependent spend more hours per week working via a platform (up to the degree of doing a more than full-time job) and have lower levels of core self-evaluation. Taken together, this implies that the most vulnerable gig workers (high economic dependence, high workload and low core self-evaluation) are the most likely to only (univalently) experience the disappearance of an OLP as an adverse event. This finding links to career research emphasizing the potentially polarizing effect of employability development, that is, those who have resources will gain even more whereas those without will lose them (Forrier et al., 2018; Hobfoll et al., 2018). A univalent experience of a career shocks as a negative event may exacerbate this effect as it is mostly likely to occur for the most vulnerable workers, thus creating an even stronger resource drain effect. This finding also contributes to the knowledge on boundary conditions involved in career shocks (Akkermans et al., 2021), showing that more vulnerable workers may be particularly susceptible to negative shocks.

In fact, the most vulnerable gig workers prior to the exit of an OLP are more likely to display high career inaction and little to no proactive career behaviors as a result of a univalent, negative career shock once an OLP departed from a country. Some may argue that Deliveroo left the Netherlands due to reclassification court cases initiated by Dutch labor unions such as FNV. Such court cases play a key role in ensuring that the most vulnerable members of the labor market can (re)assert their right to decent working conditions. Our results also raise questions about whether the opposite – at least in the short/medium run – has been achieved, since those gig workers who are most vulnerable are also most likely to be in states of career inaction and thus experience challenges overcoming the disappearance of Deliveroo. Put differently, the disappearance of Deliveroo made those workers who were already vulnerable even more vulnerable, suggesting that the exit of an OLP, even when it is heralded by social spectators (e.g. politicians, regulators, public policymakers, labor unions and labor activists) as a positive development for workers, may have undesirable effects for those that these spectators lobby for and who are in the worst possible position. In other words, when it comes to the disappearance of platforms and the plight of gig workers, social actors may need to be careful what they wish for.

As with all research, our work needs to be seen in the light of its limitations. First, we cannot confidently attribute causality to the observed relationships, as our data are cross-sectional in nature. At the same time, opposite directions may not necessarily hold conceptually or empirically. For instance, core self-evaluation is a higher-order trait (Judge et al., 2003) that is developed over the years (e.g. generalized self-efficacy) and/or stable over time (such as neuroticism). As such, it is more likely to impact career shocks rather than the other way around. Also, objective states such as economic dependence are more likely to impact subjective thought processes such as career shocks rather than the other way around. Nevertheless, future research based on longitudinal data is necessary to tease out causal interrelations, for instance, to assess whether career crafting and inaction also affect career shocks.

Second, we were not able to assess the generalizability of our results across time and space. We collected our data two months after Deliveroo exited the Netherlands. As a result, we do not know how gig workers responded to this event at a later time. Some workers may have found a new source of income and, in retrospect, may consider the disappearance of the OLP less negatively/more positively than initially reported. Although the nature and significance of our observed relationships may still hold over time, the effect size may diminish over time. Moreover, our results are specific to the gig economy context and may not generalize to non-platform contexts (e.g. traditional employment at an employer). It is likely that other antecedents explain ambivalence in career shock experiences in other contexts and for events other than the departure of an OLP, particularly provided that not all workers are highly dependent on a single employer (e.g. when they have high levels of employability).

That said, we nevertheless believe the unique nature of our data makes a contribution to the career shocks literature by showing the occurrence and drivers of ambivalence in workers’ experiences of career shocks. We call for more research that longitudinally examines the antecedents and (behavioral) outcomes of career shocks, in different contexts and of different career-related events. Moreover, we see great utility in more in-depth, qualitative research that uncovers the underlying cognitive and behavioral processes that outline how career shock ambivalence, indifference and univalence manifest and how these translate into career-related outcomes.

Finally, besides valence (positive and negative shocks), career shocks can have other attributes, such as event novelty, criticality and disruptiveness (Akkermans et al., 2018). Given the limited data points at our disposal, we could not study these additional career shock attributes. It is very likely, however, that such attributes impact the relationships between career shocks and behavioral outcomes. For instance, the positive effect of negative career shocks on career inaction may be weaker when the underlying event is not novel or less critical to a worker. As such, we encourage future studies to examine how different career shock attributes interrelate in explaining behavioral outcomes of career shocks. Moreover, we call for future studies that examine outcome variables other than career inaction and career crafting. Specifically, future studies could help to shed light on how career shock valence relates to other evaluative outcomes such as customer feedback. Customer feedback (e.g. five-star rating schemes and online review systems) plays an important role in the gig economy (Wong et al., 2021; Kuhn and Maleki, 2017). Wong et al. (2021) show how workers may both positively and negatively valence the online feedback coming from customers. It is likely that career shock valence and digital feedback valence interrelate. For instance, workers who experience a career shock as positive may find positive digital feedback more salient and experience it to be more supportive, while negative career shock perceptions may trigger workers to pay attention to negative feedback coming from customers and consider digital feedback to be more destabilizing and reinforcing feelings of insecurity. As such, the interplay between career shock ambivalence and feedback ambivalence is crucial to understanding how gig workers make sense of their careers and how organizations can better support them through tailored communication strategies.

Notwithstanding its limitations, we believe that our study provides important insights on how and why gig workers – given their dependent contractor status – respond to the sudden disappearance of an OLP. Gig workers are more likely to experience such events as both a negative and positive career shock when their core self-evaluation and economic dependence on platform work increases. Moreover, gig workers are more likely to engage in career crafting when experiencing a positive career shock and (also) enter a state of career inaction when unexpected career events are (also) perceived negatively. Taken together, we hope that our study sparks future research that examines career shock ambivalence, univalence and indifference and how unexpected career-related events impact nontraditional workers, particularly those that work in the novel gig economy context.