Millennial Career-identities: Reevaluating Social Identification and Intergenerational Relations

ABSTRACT This article extends generational literature by investigating the self-perceptions of U.K. based Millennials on how they formed their career-identities within intra- and inter-generational interactions (n = 36). Thematic analysis of interviews indicated that participants engaged in personal identification more than generational identification, interacting with both contemporaneous and older generations to co-produce and role model their career-identities. These findings challenge normative readings of sociological perspectives on generations which emphasize divergence and weakening ties between generations. Rather, findings suggest that an integrative framework provides a more appropriate theoretical lens for conceptualizing generational career-identities, incorporating kinship perspectives on interactive relationships within and across generations.


Introduction
Identity includes the self-concept of personal traits, values, attitudes and standards of behavior, along with social roles and social identification (Vignoles et al., 2011). It also encompasses past, present and future (possible) selves (Markus & Nurius, 1986). At work, career-identity comprises 'the career aspirations, values and beliefs that inform our self-concept which enable us to answer the question "who am I"' (Lysova et al., 2015, p. 40). Career-identity therefore transcends vocational identity and encompasses our career attitudes and behaviors. Empirical investigation of workplace identities often attempts to explore the ways in which personal and social identities are formed and mutually influenced (Watson, 2008), relating to 'who am I?' and 'who should I be?' Given the plethora of literature on generational differences in career attitudes, this paper seeks to explore the impact of generation, as a social identity, on the process by which personal career-identities are formed.
Notwithstanding doubts on the strength of empirical evidence for the concept of generations (Aboim & Vasconcelos, 2014;Constanza & Finkelstein, 2015;Rudolph & Zacher, 2017), workplace differences between older and younger generations continue to be widely discussed (Lyons et al., 2019). Younger Western generations include the Millennials, or Generation Y, and despite divergent opinions on the demarcation of cohorts (Parry & Urwin, 2011), those born between 1980 and 2000 are generally described as Millennials. Millennials are thought to be the most distinctive from earlier generations in their career attitudes (Kultalahti & Viitala, 2014). More than their elders, they are said to have relatively strong preferences for intrinsic values and enjoyable work (Maxwell & Broadbridge, 2016), desire more autonomy (Luscombe et al., 2013), and have highly protean (Broadbridge et al., 2007) and boundaryless career attitudes (Laird et al., 2015). They prioritize careerism by changing organizations frequently to gain faster promotion (De Hauw & De Vos, 2010) and, being less risk averse, are more likely to leave an organization if dissatisfied (Twenge et al., 2010). Overall, therefore, Millennials are thought to be more intrinsic, agentic and adaptable.
Accordingly, literature on generations in the workplace is ensconced in discourses surrounding social change and difference (Foster, 2013). In achieving social change, peer-to-peer sameness and peer-to-elder difference is assumed (e.g., Mannheim, 1952). However, there remains a lack of agreement and explanation for how and why workplace generational differences develop (Lyons et al., 2019). Constanza and Finkelstein (2015, p. 3) are unequivocal in their assertion that there is 'no sufficient explanation for why differences should even exist,' particularly when so many empirical studies have offered mixed results. As a rebuttal, Lyons et al. (2015) draw our attention to theoretical explanations offered by Alwin and McCammon (2007), Eyerman and Turner (1998), and Urick (2014), and of course Mannheim (1952). Nonetheless, it is generally acknowledged that generational research remains largely descriptive without a substantive focus on theoretical development (Cadiz et al., 2015;Lyons et al., 2015).
Of particular concern is the use of abductive reasoning, or inference-to-thebest-explanation, to adopt theoretical positions without subsequent empirical exploration. While cohort differences in career attitudes have been hypothesized, and in some cases successfully tested, the premise behind why or how there may be generational differences has been assumed rather than fully investigated. Thus, while many scholars routinely utilize a cohort perspective (Lyons & Kuron, 2014), the processual link between cohorts' sharing of formative periods and their consequent development of similar career attitudes has not been explicitly explored (Constanza & Finkelstein, 2015;Parry & Urwin, 2011). Equally, Mannheim's (1952) theory of generations and the sociological or psychological processes pertaining to peer-to-peer collective consciousness and peer-to-elder divergence also remain largely unexplored (Aboim & Vasconcelos, 2014;Parry & Urwin, 2011). And while many scholars highlight the age-period-confound (e.g., Rudolph et al., 2018), its interrelationships have not been disentangled to establish the root causes of any cohort differences (Constanza & Finkelstein, 2015). Equally, the social forces perspective which encompasses an integrative framework to examine the causes and processes of generational change and identity has not been fully investigated (Lyons & Kuron, 2014), nor has Rudolph and Zacher (2017) advocated lifespan developmental perspective. Resultantly, in such a nascent field, 'the development, testing, and refinement of theory are critical next steps' for generational researchers (Lyons et al., 2019, p. 2).
Therefore, this article answers calls to make a theoretical contribution around generational differences in career-identities (e.g., Constanza & Finkelstein, 2015;Lyons & Schweitzer, 2017), and to offer more qualitative research in the field (e.g., Lyons & Kuron, 2014). Specifically, it investigates the processes by which workplace generational differences develop (responding to Constanza & Finkelstein, 2015), and how different generations may influence and interact with each other to form differentiated career-identities (responding to Cadiz et al., 2015).

The cohort perspective
The cohort perspective is the most widely cited to help account for generational differences in the workplace (Lyons & Kuron, 2014). Ryder (1965) defines cohorts as individuals born within the same time interval who share common experiences and critical events in their formative years, resulting in a similar character as distinctive from earlier cohorts. For Millennials, such formative experiences might include parental redundancies, economic uncertainty, rapid advances in technology and telecommunications, greater and faster developments in economic expansion, and globalization (Cogin, 2012). Although contextual or societal changes may influence all ages, they arguably affect those in their developmental years more since younger cohorts grow up without strong frames of reference to earlier social norms (Campbell et al., 2015). This in part mitigates the age-period-cohort confound, since cohorts likely reference these shared primary developmental years when responding to later life stages (Lyons et al., 2015). Resultantly, people of the same cohort may develop similar belief and value systems, leading to predictable career attitudes (Kultalahti & Viitala, 2014).
The cohort perspective, however, comes with staunch criticism. First, it has not been explicitly tested (Constanza & Finkelstein, 2015;Parry & Urwin, 2011). Critically, age and cohort effects cannot be distinguished in crosssectional studies, while cohort and period effects cannot be differentiated in cross-temporal studies (Rudolph et al., 2018). Second, the idea that one's early years are the most formative can be challenged (De Hauw & De Vos, 2010;Rudolph et al., 2018). Third, notwithstanding disagreements in demarcating cohort spans (Lyons & Kuron, 2014), arbitrary cutoff dates also disguise the process of gradual generational change (Twenge et al., 2010). Fourth, it lacks sufficient accounting for heterogeneity wherein individuals sharing chronological and spatial proximity hold different work values according to, for example, gender and ethnicity (Parry & Urwin, 2011). Accordingly, Purhonen (2016, p. 96) argues the cohort perspective leads to 'a simplified and exaggerated view of generations.' Fundamentally, the approach is overly deterministic (Aboim & Vasconcelos, 2014) and fails to fully account for generational identity as a social construct.

Mannheim and generational identity
Therefore, some academics call for a return to Mannheim (1952) who remains the most influential theorizer on generations to date (Purhonen, 2016). Mannheim combined both realist and relativist understandings of generation, though he ultimately emphasized intellectual agency over social structure (Aboim & Vasconcelos, 2014). He described a tripartite construction of generation: location in history; actuality or common collective viewpoint; and generation units as self-conscious subunits of people interacting in society (Mannheim, 1952). The first two elements involve external time, but the third involves internal time (Aboim & Vasconcelos, 2014). Critically, it is not enough for a cohort to experience temporal and spatial proximity -to become a generation there also has to be social proximity and conscious interaction to form a shared sense of identity (Aboim & Vasconcelos, 2014;Mannheim, 1952). Additionally, given society is stratified by class and race, for example, different people can respond to the same critical events in their formative years in different ways. This gives rise to differentiated generation units (Mannheim, 1952). Accordingly, generations are socially constructed (Nakai, 2015) wherein individuals have collective memories and worldviews as well as common consciousness (Cogin, 2012;Lyons & Kuron, 2014;Parry & Urwin, 2011). Hence, generational identity can be defined as an individual's awareness of being similar to or different from a certain generation unit (Lyons et al., 2015;Urick, 2014).
Consequently, generation can be viewed as a social identity (Lyons et al., 2019). While Mannheim's theory does not directly lend itself to testable hypotheses, what could be empirically explored therefore is 'the development of a shared consciousness among members of a generational group' (Rudolph & Zacher, 2017, p. 114). Given people of similar generations are purported to share similar values, the chance of peer-to-peer identification is likely, thus identity theories can be employed to help understand the social construction of shared values and behavior (Urick, 2012). Social identity theory (Tajfel & Turner, 1979 is concerned with self-perceived group membership and collective identity. It provides a theoretical framework to investigate individual perceptions of generational in-groups and out-groups (Joshi et al., 2010;Lyons & Kuron, 2014;Lyons et al., 2015;Lyons & Schweitzer, 2017;Urick, 2012). Motives for social identification include creating a positive self-image, furthering self-knowledge, and achieving self-distinctiveness (Ashforth, 2001, as cited in Ashforth et al., 2008. Accordingly, generational differences can be operationalized as the extent to which people positively identify themselves with one generation, and dis-identify with others (Lyons et al., 2019;Wang & Peng, 2015).
More than this, identification leads to individuals adopting attitudes and behaviors congruent with their chosen group identity (Ashforth et al., 2008;Ashforth & Mael, 1989;Reicher et al., 2012). Mannheim (1952) proposed only a key group of younger people interact with each other to socially construct a new generational identity, forming a prototype which their peers then objectify. Subsequently, peer-to-peer socialization occurs, involving the acceptance and internalization of a generation's perspective. Thus, social identification with a prototype 'both describes and prescribes one's attributes' as a group member (Hogg et al., 1995, pp. 259-60). Individuals can choose to act congruently with that generational prototype, thereby 'enacting societal expectations' (Urick, 2014, p. 401). Reicher et al. (2012) note, however, that SIT alone does not fully theorize on such social interaction. Rather, self-categorization theory (SCT, Turner, 1987) is required to conceptualize mutual social influence. Turner (1987) proposes that we not only internalize group perspectives, but are also more receptive to ongoing influences from those whom we perceive to be in the same social group. Accordingly, while recent efforts have employed SIT to investigate work-related generational identities (e.g., Lyons et al., 2019;S.T. Lyons & Schweitzer, 2017;Weeks et al., 2017), they only investigate how employees identify with or stereotype certain generations. They do not employ SCT to explain how the contents of a person's career-identity, for instance, are influenced by that generational identity.
We also know that role modeling plays an important part in career construction (Sealy & Singh, 2010). Role models have three main functions as inspirations, representations of the possible, and behavioral exemplars (Morgenroth et al., 2015). Sealy and Singh (2010) draw on links with social identity theories to explain the processes underlying role modeling, highlighting the importance of similarities between a subject and their role model. Indeed, individuals use personal identification to seek out role models with similar personal attributes, for example, age (Ashforth et al., 2016). Congruent with identification theories, role models can be based on real or perceived social relationships between the subject and role model (Gibson, 2004). Likewise, Grote and Hall (2013) propose that social referents can come from known people as well as those who are distant, including abstract social categories. Gibson's approach to role models also distinguishes between positive and negative role models, i.e. the 'cognitive process in which individuals actively observe, adapt, and reject attributes of multiple role models' (Gibson, 2004, p. 136). Altogether, this agentic element helps explain how Millennials might model themselves on close or distant peers.

The kinship perspective
Nevertheless, Mannheim (1952) also discussed the transmission of cultural heritage from older to younger generations. Initially, he theorized, this may happen consciously or unconsciously. As younger adults experience contexts different from older generations', however, they are likely to problematize some of these inherited worldviews. Resultantly, there is a conscious break away from the attitudes of previous generations. Such an understanding may go some way toward conceptualizing the influence of period. Ultimately, Mannheim's view of generations expresses a weakening of kinship ties and increasing conflict of opinion between generations (Aboim & Vasconcelos, 2014;Joshi et al., 2011). Joshi et al. (2011), however, remind us that younger generations might reject -or accept -inherited worldviews. The authors are particularly concerned with reconciling chronology with genealogy. The latter informs a kinship perspective which describes the exchange mechanism whereby value systems are passed down through generations, rather than across peers or global cohorts (Joshi et al., 2011). As such, values can be transmitted in the workplace from senior to junior members, however, kinship can also describe how values are transmitted from parent to child (Joshi et al., 2010), accounting for the socialization of children into a plethora of gender, class, cultural and other identities. Likewise, Alwin and McCammon (2007, p. 221) remind us that generation is 'first and foremost a kinship term' and look to relationships between generations, including parental socialization. Indeed, parental support has been positively associated with personal career-identity evaluation (Stringer & Kerpelman, 2014). Similarly, Whiston and Keller (2004) point to a large body of extant literature that discusses parental influences on children's vocational aspirations.
Nevertheless, value transmission between generations is an agentic process (Eyerman & Turner, 1998). Intergenerational interactions can range from transmittive, i.e. cooperative and reciprocal, to resistive, i.e. conflictual and competitive (Joshi et al., 2010). Ultimately then, the attempted transference of ideas from older to younger generations results in three potential outcomes: the duplication; mutation; or deletion of inherited values (Joshi et al., 2011). Given much of the literature on Millennial career attitudes focuses on generational differences, it is the mutation and deletion of inherited values that is often presupposed.

An integrative approach
An alternative approach is to consider integrative frameworks. Rudolph and Zacher (2017), for instance, advocate the lifespan developmental perspective which embraces multi-disciplinary frameworks to investigate historical context, socio-cultural context, and agency-structure interactions. This includes researching generational identity/identification as an alternative to the deterministic operationalization of cohorts. For their part, S. Lyons and Kuron (2014, p. 141) propose the social forces perspective which integrates ageperiod-cohort effects (as 'complementary, rather than opposing influences' on generational differences), social identity/identification, and the multifaceted interactions between social groups. Accordingly, Cadiz et al. (2015, p. 359) note the benefits of using a social forces perspective to better understand how knowledge and culture are transferred between historical periods, and investigate the differences as well as 'reciprocal influences and exchanges within and between generations.' Consequently, an integrative perspective underpins the current study to analyze individual Millennial perceptions of generational groups, interactions, and career-identity. Following a pragmatist approach, the research question is scoped to explore peer-to-peer sameness and peer-to-elder difference whilst remaining open to alternative perspectives: How do Millennials respond to peers and elders when developing their careeridentities?

Design
The study focus was on how Millennials directly or indirectly interacted with other social actors to form their own career-identities, including career attitudes commonly discussed in generational career literature such as intrinsic values, agency and adaptability. The project was granted approval by the Research Ethics Committee at Glasgow Caledonian University and consisted of 36 semi-structured interviews with Millennials. The interviews were conducted, audio-recorded, and transcribed by the researcher, averaging 1 hour 20 minutes.

Sample
While the concept of cohorts is questionable, for operational and exploratory reasons, a purposive sample of Millennials was sought from graduates with birth years ranging from 1980to 1993(Md = 1988 who had progressed to the workplace. To increase the variableness of the sample, participants were recruited via public announcements on LinkedIn from a variety of job sectors in the U.K., i.e. education (n = 8); finance (n = 5); IT (n = 5); public services (n = 4); Higher Education (n = 4); marketing (n = 3); hospitality (n = 3); business (n = 2); and transport (n = 2). Equal numbers of genders were sought: 18 participants self-identified as women (W1-W18) and 18 as men (M1-M18).

Data collection
Open descriptive questions were utilized to elicit spontaneous contextual answers rooted in participants' experiences. The opening questions were worded as tours (e.g., Rubin & Rubin, 2005) with participants guiding the researcher round their personal experience, pointing out significant moments along the way. First, participants were asked to narrate their own career stories, wherein the researcher could discern participants' own career attitudes, contexts, turning points, and the sequencing of cause-and-effects. Second, participants were asked, "Can you describe for me the times when you became aware of your peers' career attitudes?" and "Can you describe for me the times when you became aware of older people's career attitudes?" Third, follow-up questions were utilized to gain more detail and to confirm causal processes, for example, "How do you think their career attitudes may or may not have impacted your own?'

Data analysis
Interviews were thematically analyzed using guidelines from Corbin and Strauss (2015) to code processual elements in the form of actionsinteractions. Lincoln and Guba (1985) criteria for trustworthiness were applied: credibility; confirmability; transferability; and dependability. Credibility and confirmability were achieved through prolonged engagement (e.g., exploration of participants' career stories), follow-up questions to confirm the researcher's understanding, transcription, constant comparison, and memo-writing. Memos were developed into coding tables, providing an audit trail (see Table 1 for an example). Codes were phrased as gerunds to denote participants' active responses to interactions. First-level codes often became Inspiration is a function of positive role models (e.g., Morgenroth et al., 2015) Positive role modeling Career role modeling ' . . . they're the negative role model that I'm seeing' Role models (in vivo code). Role models can be negative (e.g., Gibson, 2004) Negative role modeling the properties of second-order sub-codes, which themselves were categorized under a final third-order code. Transferability and dependability were achieved through theoretical and negative case sampling. As the memo-writing progressed, new research participants were sought, asked questions for confirmation, and prompted for alternative explanations. This resulted in thick descriptions and exploration, for example, of both personal and generational identification with role models. All coding and analysis was grounded directly in the data and not derived from prior theory. Nevertheless, Corbin and Strauss (2015) acknowledge the role of literature in assisting the researcher to make comparisons between findings and extant theory. After a broad outline of first and second order codes was generated from the first 23 interviews, the researcher engaged in theoretical sensitivity. Various theories were drawn on as outlined above, but the researcher did not force an interpretation of one or the other. Theoretical saturation was achieved after 30 interviews, with the remaining six interviews analyzed to confirm no new codes or properties were evident.

Findings
The final third-order code encompasses the over-arching theme from the interviews: connecting with other career-identities (making links to the careeridentity of others as a guide to the development of one's own career-identity and self-image). This theme incorporated three sub-codes, or processes, by which participants came to develop their career attitudes within inter-and intra-generational interactions. These are discussed in turn below: coproducing career attitudes; career role modeling; and (re)confirming individual career-identity.

Co-producing career attitudes
The first process, co-producing career attitudes, involved interacting with other people and actively or passively acquiring their attitudes in the development of one's career-identity. A direct interaction was required between the participant and another social actor. Participants engaged in coproduction as a way of increasing awareness and reducing uncertainty when attempting to answer, "who should I be?" The intended outcome was therefore to have a career-identity socially validated by significant others of different generations. From the sample, this was done by: inheriting from parents; being directed by parents; being mentored by elders; and co-creating with peers. First, inheriting from parents occurred when participants consciously or unconsciously adopted their parents' career attitudes. Rather than intentionally copying their parents, the process was subconscious. For example: you realize you've done exactly what he's done. It's like a modern version . . . (M10) . . . you do generally eventually turn into your parents, you can see it happening sometimes, it probably has <happened> from the family unit make-up point of view . . . (M18). Therefore, rather than consciously breaking from previous generations and demonstrating a weakening of kinship ties (cf. Mannheim, 1952), Millennial participants sometimes inherited the career attitudes of older generations in line with Joshi et al.'s (2010) kinship perspective. Findings therefore highlight the predominant influence of social identities stemming from familial rather than peer cultures, and the potential similarities between younger and older members of comparable socio-cultural contexts, as in Rudolph and Zacher (2017) lifespan perspective.
Second, interviewees could be directed by parents and explicitly encouraged to have certain career attitudes. M6 discussed how his father actively encouraged him to be intrinsic: Yeah he's certainly on the same sort of ilk as well as my mum on that, he didn't want me to sort of just get a trade and go and just join and be another cog in the machine really, just making sure that I'm doing something that I'm happy with . . . it was more of a gentle encouragement of "go into further education, take time to see what you want to do rather than pigeonhole yourself early on" (M6) M6 explained his father had not attended university, but had continued in the same trade since young adulthood. By encouraging his son to do different, therefore, he was encouraging him to learn from the previous generation's (his) negative experience. In such cases, rather than transmitting their own worldview (cf. Joshi et al., 2010;Mannheim, 1952), older generations themselves provided the impetus for generational change by asking younger generations to do things differently.
However, some parents offered contradictory advice and individuals had to rationalize which advice to follow. M4 described shifting from his father's to his mother's career attitudes: And then as I got older my mum's attitudes towards work actually probably informed me a bit more because she would always say "do something that you enjoy". You know what I mean, you work to live, you don't live to work. So when I was younger my dad's attitudes pushed me but as I got older I realized my mum was probably doing the more savvy thing (M4) By this time, M4 had engaged in some work experience he did not enjoy, making him reassess his career-identity. He consequently matched his new personal experience with the alternative views of his mother. This reflects that personal context was an important evidence-base used by participants to help them decide whether to mutate or delete transmitted values from older generations. This corresponds with Mannheim's (1952) and Joshi et al.'s (2010) perspectives on intergenerational transmissions reevaluated in new contexts.
Third, some participants felt they were being mentored by elders at work to change their career attitudes. This parallels Joshi et al.'s (2010) notion of intergenerational interaction in the workplace. Taking advice from mentors opened up individuals to ideas they had not considered before. W2 explained: . . . the chairman of X used to mentor me a little bit and he actually said make sure you don't stay in a job for too long so I've always kind of kept that at the back of my mind. (W2) Again, rather than Millennials being innately psychologically and physically adaptable as inferred from a cohort perspective (cf. Ryder, 1965), it was actually the older generation's perspective on contemporary contexts which drove them to instill such attitudes in younger generations.
Fourth, participants could also interact with people their own age, thereupon co-creating with peers. They discussed and evaluated their careers, like W5 who interpreted her preferred career-identity through the insight of her husband: W5's situation demonstrates how individuals are receptive to ongoing influences from those of the same social group (e.g., Turner, 1987), a consequence of tendencies to socialize closely with others of a similar age.

Career role modeling
The second process participants described was career role modeling: actively selecting another person as a source of motivation and inspiration, and as an exemplar of who it is desirable or possible to be. In contrast to co-producing career attitudes, no direct interaction was required between participant and social actor. Through this process, participants were toying with possible selves (e.g., Markus & Nurius, 1986;Morgenroth et al., 2015), answering questions such as, "who can I be?" and "who do I want to be?' It involved looking at both positive and negative exemplars from older and peer generations, but was tempered by the relevance participants assigned to social actors.
Consistent with the function of role models as being inspirational (e.g., Morgenroth et al., 2015), M7 talked about feeling "inspired" by his father's work ethic. W11 also felt inspired by her mother's achievements, however, this was tempered by other observations of her father. She explained: It kind of inspired me to achieve. Now when I'm obviously just starting my career, I'd love to be able to say in ten years' time that I've been on holidays in amazing places and I'd love to be able to have a nice car and do that, and so I think that's really massively down to her and her whole attitude to succeeding financially. But also seeing how stressed they were a lot of the time, I think my kind of priority is to do something that's comfortable but also something that you enjoy, and I do think a lot of the time, not so much my mum, but my dad, I don't think he enjoyed what he did, so I think that . . . you know I'd like to be comfortable but also do something that I enjoy and be happy.
Interviewees therefore individualized their career-identities, sourcing multiple representations of the possible from positive and negative role models (e.g., Gibson, 2004), in these cases from older generations. Career role models could also be sourced out-with the family as distant social referents (e.g., Gibson, 2004;Grote & Hall, 2013). M18 described seeing older men at his industry's annual dinner, and how he consciously wanted to be different: . . . so when I'm a 50-something year old white male I'm not going to be the same as them but they're the negative role model that I'm seeing so I'm thinking I'm not going to act like them when I'm older, that's how the generations change I think because you see people doing stuff before you and you're like "I'm not going to be like that" (M18) For M18, therefore, generational change was indeed initiated from the younger generations, expressing a weakening of kinship ties in line with Mannheim (1952).
Affecting the influence of career role models, however, was the relevance individuals imparted to them. Some could not source appropriate role models from their immediate family. While literature suggests Millennials are more intrinsic than older generations (cf. Maxwell & Broadbridge, 2016), W16 and M17 felt slightly less intrinsic than their same-gender parents. In both cases, their parents had been teachers -something W16 and M17 perceived as intrinsic jobs. Therefore, although they were inspired by their parents' intrinsic motivation, they felt their own jobs (in business) were not quite comparable. M15 also recounted the difficulty in sourcing a realistic role model because both his parents stopped working when he was young. He consequently found it difficult to know what he wanted from a career: . . . (I) didn't really focus on what I wanted to do as a career, so maybe that was because I didn't see a career example to say "well maybe I could have a career doing this possibly".
In such cases, participants appeared to prioritize personal over generational identification in assessing the relevance of role models (see Ashforth et al., 2016;Sealy & Singh, 2010). Likewise, W18 described how peers (friends) could not always provide appropriate role models: . . . one of them is about to retire next year and he's 30. Like this is my point, I can't keep up with that, the effort's not comparable . . . Therefore I've had to make this decision for my own career worth and my own self-worth to say "well I'm not comparable with them". I can't behave on that kind of level. I can't keep up with that . . . (W18) Processes aligned with SIT/SCT were therefore utilized for personal identification to underpin the role modeling process, in contrast to the assumption that generational identification is a central influence on contemporary careers (cf. Urick, 2014).

(Re)confirming individual career-identity
Identification with broader social categories was, however, apparent in the way that participants positioned themselves in relation to other generations. In the third process, (re)confirming individual career-identity, participants engaged in comparing themselves to others to confirm and justify their own current career-identity. From the sample, no direct interaction was required between individuals and social actors, and identification had no direct influence on the uptake of particular career attitudes. Ultimately, participants engaged in comparison as a way of achieving self-enhancement, positive self-image, and justification for who they already were, corresponding with some of Ashforth's (2001) purposes of social identification. It ultimately answered the question, "who am I in relation to others?" This involved feeling similar and feeling different, reflecting the basic tenets of SIT and SCT.
At the end of the interviews, after participants had already discussed how they had constructed their career-identities, they were asked whether they felt they belonged to a particular generation. There were no prompts given as to how to define generation, and the interviewer made no mention of Millennials or Generation Y. Participants did perceive some similarity with their peers, and a minority mentioned Millennials or Generation Y, however they often defined their generation on their own terms. M1 described his generation as being a "university generation": I feel there is something different from Mum and Dad's generation, but maybe that is just my feeling of a lot of people <have> moved away to university from X and you know a lot of us haven't gone back . . . so you just feel like your generation is the university generation where everybody has gone to university . . . (M1) However, neither M1 nor any others offered identification as an explanation for why their own career-identities had developed the way they had. M1 merely used his sense of identification to position his own identity alongside others'. M1's quote also shows that in discussing similarities to some, he naturally discussed dissimilarities to others.
Second then, as a necessary counterpart, interviewees also described feeling different. Participants often spoke of differences around changes in context and the modern career path, rather than changes in the innate nature between one generation and the next. M15, for instance, thought of his father operating in a non-graduate context. He felt his own intrinsic values and adaptability were products of his current context, and that his career attitudes could have been similar to his father's if he had been operating in his father's context: Similarly, W3 referred to her increased adaptability as a direct consequence of the current labor market. For her, this explained why there was a difference between her own and previous generations (who had predictable career paths and demonstrated loyalty to companies): <we are> the generation that couldn't find a job, so we've almost had to create our own rules and our own way of dealing with it. And the way to deal with it is we're not loyal to a company . . . you know we seem to have created this culture because of what's been forced on us. (W3) Thus, participants perceived changes in context experienced as an adult as a major contributory factor in generational differences in career-identities. Consequently, the impact of generational identity might be overestimated, necessitating more consideration of the age-period-cohort confound (e.g., Deal, 2007;De Hauw & De Vos, 2010;Lyons & Kuron, 2014) or lifespan perspective (Rudolph & Zacher, 2017).

Discussion
This research has utilized Millennials' individual accounts to explore how they responded to peers and elders when developing their career-identities. Analysis showed that participants formed their career-identities in two main ways: co-producing career attitudes to form a career-identity socially validated by significant others of different generations; and career role modeling of both close and distant multigenerational referents. Analysis also showed how participants used generational identification, not as a way of prescribing to generational prototypes, but as a way of (re)confirming individual careeridentity and its relative positionality to others. This article has therefore responded to the challenge of refining generational theory a little further (e.g., Lyons et al., 2019), by empirically investigating the processes by which generational differences have developed (responding to Constanza & Finkelstein, 2015), and how different generations influence and interact with each other to form their differentiated career-identities (responding to Cadiz et al., 2015). Resultantly, there are four key theoretical contributions.
First, analysis demonstrated that influences on career-identities involved more complex and multigenerational processes than the normative readings of some key generational theories might suggest. Mannheim (1952) emphasized the weakening of kinship ties and the deliberate deletion of inherited values from older generations. Analysis, however, showed that participants naturally looked to people of multi-generations to guide their career-identities, sometimes through role modeling processes. As such, they observed, copied, adapted, and rejected the attributes of multiple individuals (e.g., Gibson, 2004). This somewhat aligns with the kinship perspective of Joshi et al. (2010) where younger generations can choose to accept, mutate, or delete inherited worldviews.
Second, and critically, older generations were sometimes the key drivers behind generational change. They actively encouraged or instructed younger generations to do things differently from previous generations, and to take on characteristics which normative literature suggests are uniquely innate to Millennials, such as higher levels of intrinsic values, agency and adaptability. Mannheim (1952, p. 308) briefly accounted for this, though it remains underrepresented in recent interpretations. He acknowledged that older people are often forerunners of new ideas which are then taken on by younger generations. Nevertheless, he also pronounced that these forerunners are 'isolated in their own generation.' The analysis of interviews does not reflect that parents or mentors were necessarily isolated forerunners. Rather, analysis showed that older generations wanted younger people to learn from their (previous generations') past life experiences or perspectives, and try alternative approaches. Overall, the findings challenge normative readings of Mannheim (1952) that presuppose generational change is primarily instigated by younger generations.
Third, participants did not wholly engage with generational identity as Mannheim (1952) might understand it. He emphasized collective consciousness and behavior, though allowing for differentiated generation units. Many generational scholars seek to explain this collective consciousness by applying social identity theories, and investigating generational identification with peers (e.g., Lyons et al., 2019). However, analysis demonstrated that straightforward in-group and out-group comparisons between peers and elders were not utilized by these participants. Rather, identification was used in two quite separate ways.
On one hand, participants used personal identification (e.g., Ashforth et al., 2016) as the underlying process to evaluate the relevance of role models (e.g., Sealy & Singh, 2010). Since the identification process was so personal, it involved a great amount of variety. For example, participants could see themselves similar to peers as well as dissimilar due to economic disparities. Likewise, they could see themselves similar or dissimilar to older generations due to chosen vocations. Relatedly, Lyons et al. (2019) propose that generational identity can become more salient in certain (workplace) contexts. In the context of career-identity development, however, findings here suggest that personal identification was more salient than generational identification in inspiring career attitudes.
On the other hand, participants did use generational identification, i.e. (re) confirming individual career-identity, but only to position themselves in relation to others. They engaged in social comparison as a way of achieving a positive self-image and justification for who they already were. Thus, contrary to Urick's (2014) proposition, there was no evidence from the sample that generational identities were utilized as prototypes to follow.
Fourth, when describing generational differences, participants often pointed to changes in context as the key driver. The cohort perspective (e.g., Ryder, 1965) forefronts formative experiences as key determinants of adult career attitudes (Parry & Urwin, 2011). In contrast, participants here cited influences from their adult life, such as labor markets and conversations with parents and peers as an adult. This aligns more with the assumptions of the lifespan perspective that no specific age period in human development is inherently more formative than others (e.g., Rudolph et al., 2018). Results therefore partially support contextual hypotheses (e.g., Deal, 2007;De Hauw & De Vos, 2010), along with Mannheim's (1952) concept that previous generation's worldviews are reinterpreted when confronted with new contexts.

Conclusion
Overall, the current study demonstrates that Millennials view personal identification as more salient than generational identification, and interact with both contemporaneous and older generations to co-produce and role model their career-identities. Moreover, both peer and elder reactions to contemporary contexts were influential in shaping Millennial career-identities. These findings, therefore, challenge normative understandings of generations which emphasize difference and weakening kinship ties between generations. Rather, from an empirical base, this article proposes that an integrative framework, such as the social forces (Lyons & Kuron, 2014) or lifespan perspective (Rudolph & Zacher, 2017), is required to understand the complexities of generational and career-identity work.
An integrative framework should involve a reconsideration of age-periodcohort or historical and socio-cultural context effects, complemented by that of intra-and inter-generational interactions, generational identity and agencystructure dynamics. In regards to the latter, an important inclusion is the kinship perspective which provides a nuanced theoretical lens for researching generational change through the duplication, mutation, or deletion of inherited values. An integrative framework should also include theory from other disciplines, and data analysis here indicates role modeling theory is useful. Overall, this article calls for the abandonment of a narrow focus on generational 'differences,' and the active empirical exploration of the processes whereby different generations influence each other to develop their career-identities. Critically, older generations' role in instigating generational change needs to be theorized further.

Limitations
Due to the qualitative and exploratory focus on identity processes, the current study was unable to fully explore all elements of integrative frameworks, such as reciprocal influences of younger generations on older generations' careeridentities. Nor was it able to explore more constructionist perspectives on identification such as classification struggles and discourse (e.g., Aboim & Vasconcelos, 2014;Eyerman & Turner, 1998;Foster, 2013;Purhonen, 2016). While very few participants expressed familiarity about global generations, the functional influence of social schemata (such as 'Millennial' stereotypes) on individual career-identities could also be explored further. Additionally, while the findings here question some elements of Mannheim's vision of generational differences through analyses of individual accounts, sociological research might prove more fruitful in clarifying our thinking on collective consciousness.