The logic of attraction: exploring the institutional complexity of job preferences

Purpose – Applying the institutional logics perspective to applicant attraction, this study investigates the level of uniformity among preferences for consulting job attributes associated with the institutional logics of the corporation, the profession and the family, and tests for the influence of anticipatory socialization differences. Design/methodology/approach – The study uses a discrete choice experiment with 232 business students. A hierarchical Bayes approach to conjoint analysis uncovers part-worth heterogeneity and allows for subsequent cluster and regression analysis of the choice data. Findings – The findings identify a dominant job-oriented preference type and a minor career-oriented preference type. Anticipatory socialization through personal prior work experience and the occupation of friends decreases adherence to the logic of profession and increases the relevance of the family logic. The parents ’ occupation has only a minimal influence on preferences. Practical implications – The study provides attribute-based recommendations on how professional service firms can effectively address the complex expectations of potential applicants in their job ads for an entry position and underlines the role of intra-generational reference groups as important anticipatory socializers. Originality/value – By testing individual socialization effects at the pre-hire stage and beyond the organizational level, the study fills a void in both the recruitment and the institutional literature. socialization as explanatory factor for heterogeneous job preferences, the study adds new insights into the micro-foundations of institutional agency as we empirically test specific forms of previous exposure to values (anticipatory socialization) for adherence to those values or logics in decision-making. Drawing on institutional research, this study shows that in the mature field of PSFs with strong internalized rules of the game (Alvesson and K € 2004) actors at the margins of the field, who are deeply embedded in other social contexts such as university or family, question the rules when they get to know them (Greenwood and Suddaby, 2006) In previous exposure to the logic of profession does not necessarily strengthen the level of adherence to this logic but can result in selective defection due to for example saturation, or more appealing alternatives. To participants who have already experienced PSFs, either directly or indirectly, the family logic becomes more important. These students chose the family logic for the work climate attribute and for the training and development attribute. Friends and colleagues familiar with PSFs also influence students to assign more importance to the family logic for the salary and bonus attribute. These results are consistent with the observation that previous interns tend to appreciate a more balanced work Soulez, 2014). The influence of work experience may result from the “ reality shock ” newcomers experience after organizational entry, which leads to critical reflection of work practices 1980). The results suggest that the anticipatory socialization perspective advances knowledge on individual differences, a renewed focus in applicant attraction research and S € u ß , 2017) and in upcoming differentiated approaches to human resource management et 2019). In particular, the results


Introduction
Much of applicant attraction research builds on the idea that people want to join organizations which match their own interests and personality. Since organizations also want to recruit people who fit, this implies a homogeneity of applicant characteristics (Schneider, 1987). However, business students of the Y generational cohort were found to have more heterogeneous job preferences than the Generation Y stereotype proposes (Ng et al., 2010;Guillot-Soulez and Soulez, 2014). The underlying mechanisms producing differences in job The logic of attraction preferences and in individual perceptions of (mis-)fit between people and organizations remain puzzling to date (Acikgoz, 2019). This is especially burdensome for research on employee attraction in professional service firms (PSFs), such as consulting companies, because of the "war for talent" for the brightest graduates that knowledge-intensive organizations face (Malhotra et al., 2010). Understanding applicants' preference structures and their drivers therefore needs to be a priority. Institutional logics provide individuals with legitimate repertoires of behavior, vocabulary and ways of assessing other actors, e.g. future employers (Friedland and Alford, 1991;Thornton et al., 2012). The institutional logics perspective and its application to recruitment studies (Bullinger and Treisch, 2015;Hansen and Schnittka, 2018) suggest that the young business consultants' expectations of PSFs as employers are linked to broader value systems in society, specifically to the logic of the profession (PSFs bringing together experts at their jobs), the logic of the corporation (PSFs as a place for a lasting career), and the family logic (PSFs taking care of their employees' social and personal needs) (Thornton et al., 2012). With multiple institutional logics providing different prescriptions, PSFs and their (potential) employees face institutional complexity (Greenwood et al., 2011). Institutional logics guide individual decision-making and could therefore explain heterogeneous reactions of potential applicants during the attraction phase. That individuals draw on different (combinations of) institutional logics "on the ground" (McPherson and Sauder, 2013), i.e. in practical situations, is due to socialization and familiarity with a logic, which constitutes a recent focus in institutional theory (Brandl and Bullinger, 2017;Glaser et al., 2016).
For understanding applicant attraction, sociodemographic differences and social influences, in particular socialization and work experience, have also been largely ignored and have just recently been rediscovered (Kulkarni and Nithyanand, 2013;Evertz and S€ uß, 2017;Torsello, 2019). Socialization is the process through which individuals acquire knowledge and acceptable behaviors to become members of a group, thereby "preparing [them] for the types of role [they] will be expected to play" (Ryder, 1965, p. 852). Preparing students for the role of business consultants, anticipatory socialization as the learning and development of expectations before starting the job is especially important. Depending on who influences their expectations, soon-to-be graduates might prioritize the logic of the profession, corporation or family when assessing the attractiveness of job advertisements.
The study aims at understanding why some business students are attracted to certain job attributes of PSFs while others are not. Developing an exploratory cluster and regression analysis with a Bayesian approach to conjoint analysis, this exploratory study addresses two research questions. First, what worth do business students assign to the institutional logics of the profession, the corporation and the family encoded in the attributes of PSF job ads and how homogeneous are these preferences? And second, how do primary socialization by family members and secondary socialization by friends and colleagues and the students' own work experience influence job pursuit intentions by increasing or decreasing adherence to these logics?
Theoretical background Professional service firms and institutional logics Professional service firms (PSFs) invest a lot of effort in recruiting university graduates by seeking to anticipate the preferences of potential future business consultants and to address them in their job advertisements and employer branding material (Erlenkaemper et al., 2006). As professional service firms themselves combine professional principles with the bureaucratic efficiency of a formal organization (Swart and Kinnie, 2013), job candidates are likely to draw on the logic of the profession and the logic of the corporation when assessing the attractiveness of a PSF as an employer. Professions and corporations are two out of the six key institutional logics of Western societies, besides family, religion, state and market, that "shape how reasoning takes place and how rationality is perceived and experienced" (Thornton et al., 2012, p. 2). The logic of the profession centers around personal expertise and membership in a profession as basis for identification, positioning the client as the most important stakeholder (Anderson-Gough et al., 2000). The logic of the corporation is common in any form of business organization and depends on hierarchy, roles and culture within a specific company (Thornton et al., 2012). With studies stressing the growing importance of work-life balance due to the preference of Generation Y employees entering the workforce (Guillot-Soulez and Soulez, 2014), the family logic as the firms' responsibility to take care of employees (providing a familiar atmosphere and work-life balance) might also gain importance.
Institutional logics present a concept for understanding the interrelationships between individuals, organizations and institutions in societies (Thornton et al., 2012), as both individuals and organizations draw on these prevalent principles, practices and symbols (Friedland and Alford, 1991). In their strategic human resources management (HRM) choices regarding organizational culture, performance and reward systems, training and development activities. PSFs draw on multiple institutional logics (Swart and Kinnie, 2013), just as potential applicants' expectations are influenced by institutional logics (Dejordy and Creed, 2016). Comparable with organizational values espoused in corporate communication (Bourne et al., 2017) yet less explicit, potential applicants can encounter and assess institutional logics as they are encoded in job ad descriptions of the organization, positions and roles (Thornton et al., 2012, p. 40). Typically, job ads feature a description of the company (organizational culture), market the offerings (compensation and benefits, employee advancement) and outline the requirements (job design, work climate) for a particular function (Backhaus and Tikoo, 2004;Elving et al., 2013). These job attributes can each reflect either professional, corporate or family values, depending on the employer and its recruitment messages. The tensions and incompatibilities in handling multiple logics turn institutional environments from pluralist into complex (Greenwood et al., 2011). As research investigating how organizational members interpret complex institutional logics and translate them into practical action is burgeoning (e.g. McPherson and Sauder, 2013;Pache and Santos, 2013), individuals are increasingly recognized for their active role in reproducing and changing institutions (Abdelnour and Hasselbladh, 2017).
Individual agency should not be equated with autonomous, rational decision-making (Emirbayer and Mische, 1998, p. 970). Individual actions and associated choices are iterational; although they are directed towards the future, they are driven by the routinizations of past experiences in particular socio-cultural settings instead of expected outcomes (alone). In decision-making situations, individuals might draw on one or combinations of multiple logics, depending on previous experiences (Pache and Santos, 2013). Previous exposure to institutional logics determines which logics are likely to be activated. Accordingly, without relevant socialization experience, a logic is not likely to be available to novices. With experience, a logic is available and understandable but not obligatory for familiar individuals. Identified individuals actively commit to a logic as part of their identity (Pache and Santos, 2013). As this strand of research treats the questions of socialization superficially, it would benefit from an explicit focus on socialization and social influences.
Socialization and sociodemographic factors influencing job pursuit intentions As individuals are socialized into the rules and norms of society (Berger and Luckmann, 1966), they have a basic understanding of all institutional logics (Friedland and Alford, 1991).

The logic of attraction
Preparing for various roles, also relating to career and occupations, socialization conveys a profound sense of meaning about "what is important, what is to be valued, what is to be lived for" (Arnett, 1995, p. 618). Socialization occurs in different contexts and through different social influences that might vary in their impact on which institutional logics individuals activate. In primary socialization, parents pass on cultural knowledge to their children; in secondary socialization, organizations, such as schools or employers, signal appropriate cultural behavior to individuals. Work-related socialization knows three main forms. Trough organizational socialization, newcomers learn to understand the culture of a certain employer.
Professional socialization serves to understand the culture of a profession, such as a being a doctor, lawyer or consultant (Van Maanen and Schein, 1979). Anticipatory socialization refers to the expectations individuals develop towards their first job and "encompasses all of the learning that takes place prior to the graduate's first day on the job" (Garavan and Morley, 1997, p. 119). Realistic expectations and value congruence between organizations and potential applicants are the desired outcomes of successful socialization at the anticipatory stage (Feldman, 1976). In order to attract professionally committed graduates and to ensure their future socialization to the work culture of both the organization and the profession, PSFs seek to recruit already somewhat socialized "veteran" candidates over "neophytes" (Carr et al., 2006), especially former interns of the industry (Narayanan et al., 2010). While numerous empirical studies analyze the role of business students' anticipatory socialization with regard to relevant post-entry outcomes, such as professional skepticism for the accounting profession (Farag and Elias, 2016), we lack understanding of its role during the recruitment phase, particularly for the decision to apply.
Our study aims to account for the influence of anticipatory socialization on the activation of institutional logics by students deciding on attractive job attributes of PSFs. In order to identify measurable sociodemographic factors that reflect differentiated adherence to institutional logics, we draw on career and recruitment studies, both interested in understanding graduates' choices for an entry position. Given the limited availability of articles within the research area of HRM, we supplement our theoretical arguments with work-related literature from neighboring fields such as communication, childhood and nursing studies.
At the anticipatory stage, young adults prepare through socialization for work by family and educational institutions and consolidate their expectations through socialization by work on the basis of first professional experiences (Cohen-Scali, 2003). Parents play a pivotal role during anticipatory socialization (Jablin, 2001;Sang et al., 2009). Empirical research on organizations, however, tends to ignore the role of the family (Dyer, 2003). Communication research shows that the family as socialization agent in contrast to television conveys not only extrinsic but also intrinsic work values for developing a professional identity (Hoffner et al., 2008). Psychological and sociological research traditionally acknowledges the role of the family (for an overview see Whiston and Keller, 2004). Most of these studies are related to specific occupations, such as teachers, lawyers and nurses, and confirm the phenomenon of "occupational inheritance," the influence especially of the fathers' occupation on their children's first job (e.g. Jodl et al., 2001;Gubler et al., 2017). The job information parents pass on to their children in family communication about work might also include many unfavorable aspects, i.e. addressing work dissatisfaction or gender issues (Medved et al., 2006). Observing their parents struggle with economic turmoil, workplace restructuring and automatization is considered to be a key reason why the generation Y workforce finds it difficult to commit to an employer and rather pursues rapid career advancement and opportunities for personal development ( € Ozçelik, 2015). Friends also influence occupational choice through socialization. Important reference persons are those in the same educational institution. During the transition from home to university, it is vital for first-year students to forge new relationships with fellow students (Mudhovozi, 2012), a finding that might be an indication that friends generally also constitute an important reference group for the transition from university to first employment. This research also shows that socialization by friend and peers rests on a different socialization mechanism than socialization by family. General adolescence research suggests that friends shape individual behavior through modeling (resulting in preferences) (Biddle et al., 1980), while parent influence is mainly transmitted through norms (resulting in standards). Therefore, newcomers rely on their peers from their social network for comparison, for example in terms of job search performance (Cheung and Dougherty, 2005), and girls favor gender-atypical jobs more when the most popular girl in the group also favors them (Alm and B€ ackman, 2015). Career studies suggest that the discussion of career and occupational norms during leisure time activities with friends during adolescence and later with peers from college is an important micro-mechanism of anticipatory socialization (Jablin, 2001;Weidman, 1984). For example, in the socially intense setting of top business schools characterized by long joint study sessions and regular comparison among peers, there is a strong normative influence to apply at popular employers (Higgins, 2014). The relevance of socializing with friends and peers for job choice might also be reflected in the workplace motivation generation Y students draw from belonging to a team of friendly and collaborative colleagues (Calk and Patrick, 2017). However, as this finding varies greatly among respondents, it requires further empirical study. So, despite its theoretical importance, there is a lack of research on peer influence on anticipatory socialization.
The influence of one's own work experience and that of colleagues on anticipatory socialization has not gained much attention in empirical studies either. This seems largely attributed to the fact that during school and tertiary education, young people gain much of their work experience from low-end, low-skill part-time jobs (Levine and Hoffner, 2006). From the surprisingly limited empirical studies on the career consequences of student internships, however, we can assume that favorable work experience during an internship positively affects the intention to start a career in the respective industry (Chen and Shen, 2012; Rothman and Sisman, 2016). Drawing on organizational socialization literature, one can further assume that colleagues function as important socialization agents (Saks and Gruman, 2012). At least regarding graduate entrepreneurship, prior personal work experience plays a role in career decisions (Henderson and Robertson, 2000).

Method
In this study, 239 business students close to finishing their studies were asked to decide which of two job ads from consulting companies they would find more attractive and therefore would apply for a job. Consulting is a typical entry job for university graduates, with leading consulting firms ranking among students as popular "employers of choice" and competing with each other to attract new talents (Rampl, 2014;Rampl et al., 2016). New labor market entrants constitute a specific type of job seeker due to their inexperience with the realities behind job attributes and the resulting difficulties in assessing job opportunities (Collins and Stevens, 2002). This is why students are a key target of organizations' orchestrated recruitment activities and are being dominantly addressed in applicant attraction research (Berthon et al., 2005). Because students of the Gen Y cohort are now entering the labor market (Muskat and Reitsamer, 2020), current applicant attraction studies particularly focus on enrolled students belonging to the generation Y (Alonso-Ameida and Llach, 2019; Baum and € Uberschaer, 2018;Ronda et al., 2020). Theoretically, the choice of the sample is relevant as the concept of anticipatory socialization applies to the individual job expectations accumulated before the first job (Garavan and Morley, 1997). Therefore, to understand the role of anticipatory socialization in the context of application attraction to The logic of attraction consulting firms, the study needs to concentrate on the pre-employment, "education-to-work transition" career stage (Guillot-Soulez and Soulez, 2014).
In order to answer the first research question of which job attributes and institutional logics potential employees prefer and how homogenous these preferences are, we implemented a two-stage design. In the first stage, we draw on a decision situation in which different logics might become relevant. Based on the discussion of logics in PSFs (Swart and Kinnie, 2013;Thornton et al., 2012), we created job vignettes thatsimilar to real job advertisementsdescribe certain attributes of consulting jobs in PSFs (company description and job design, salary and bonus, work climate and training and development) either according to the institutional logic of professions, corporations or family (see Table 1).
Drawing on these vignettes, we conducted a choice-based conjoint analysis (Luce and Tukey, 1964). This is often also referred to as a discrete choice experiment, which mimics the decision-making process in real life situations. These choice data were used to find out which of the job attributes and which institutional logics business students prefer. Our model is based on Rossi et al. (2005) and only included the covariates of gender and age. The choicebased conjoint analysis resulted in part-worth utilities which show the potential employees' preferences. The choice-based conjoint analysis employed the software R-Package bayesm (Rossi et al., 2005). It was run with 232 respondents remaining after seven respondents were eliminated due to missing values and implausible answers in the socio-demographic variables.
In the second stage, we used the preference results of the choice-based conjoint analysis to conduct a cluster analysis in order to find out whether there are archetypical preference types Attributes Levels Company description and job design (1) Wea competitive companyare looking for process-oriented, wellstructured and dedicated employees for varied tasks. (Institutional logic of the corporation) (2) Wea client-oriented companyare looking for a professional, competent and highly qualified expert for demanding consulting responsibilities.  who favor different job attributes and institutional logics. The results of the cluster analysis show how homogenous the potential employees' preferences are. The cluster analysis is based on the rescaled part-worth estimates for the three levels of the four attributes used in the choice-based conjoint analysis. The classification was performed using Ward's method of hierarchical agglomerative clustering (Murtagh and Legendre, 2014) and the software R-Package pvclust (Suzuki and Shimodaira, 2006). The number of classes was identified by using p-values for clusters, which state how strongly the clusters are supported by the data.
To answer the second research question of the impact of socialization on these preferences, we ran a mixture-of-normals model analysis (Rossi et al., 2005). This tests if different forms of exposure to institutional logics (through the respondents' own experience and socialization) make individuals relate to specific institutional logics in a novice, familiar or identified mode. Therefore, we focused on socio-demographic factors (see Table 2) that reveal the influence that exposure to institutional logics might have on the subsequent preference for institutional logics. The results of the mixture-of-normals model analysis show the influence of these socio-demographic factors on the part-worth utilities obtained from the choice-based conjoint analysis.

Results
The fit of the logit model to the data is most commonly assessed with a measure of how well the dependent variable, based on the independent variables, can be predicted and therefore focuses on R 2 measures of predictive power (Rao, 2014, p. 159). The results of the choice-based conjoint analysis conducted with 232 respondents, who each chose between two different job vignettes nine times, has an excellent statistical fit as the likelihood-ratio statistic (or chisquare statistic) (R 2 Chi ¼ 1; 494:67, df 5 10, p ≤ 0.001) indicates that the estimated model is highly significant. McFadden's pseudo R 2 ðR 2 McF ¼ 0:472Þ indicates an excellent model fit. The cluster analysis is based on the rescaled part-worth estimates obtained from the choicebased conjoint analysis. Approximately unbiased p-values over 95% indicate that the two clusters are strongly supported by the data.
The cluster analysis led to a segmentation in two classes (see Table 3) which differ with regard to the part-worth estimates and the relative importance of each job attribute. Partworth utilities represent individual preferences in each class in an objective manner. Partworth estimates portray the impact of each of the 12 levels (see Table 3) and thus portray the potential employees' preference structures. Because the part-worth estimates are converted to a common scale, the factor with the greatest range of part-worth utilities (low to high) is the greatest contribution to the overall utility and hence the most important factor (Hair et al., 2010, p. 302).
This segmentation highlights that two archetypical preference classes exist which value job attributes and institutional logics differently: Class 1 (26.3%), the "career-oriented type," is essentially formed around pay-related job attributes (salary and bonus) as well as promotional aspects (training and development), which are perceived as the most important attributes. The most preferred attribute levels of this type are "job tenure and company performance" and "on-the-job training and personal development," which yield almost equal utility. Thus, an underlying corporate logic as well as family logic play an important role.
Class 2 (73.7%), the "job-oriented type," focuses mainly on work climate and company description and job design which are perceived as the most important job attributes. The most preferred attribute levels of this type are "self-determined work and enthusiasm" and "competitive company and process-oriented employee for varied tasks." Thus, an underlying professional logic as well as corporate logic play an important role.
The heterogeneity of the respondents with regard to their preferences can partly be explained by their exposure to institutional logics through their own experience and The logic of attraction socialization (see Table 4). Therefore, we specified the mixture-of-normals model by including the respondents' own experiences and their peers' occupations as business consultants or in other professional service firms. Thus, the independent variables are the respondents' own experiences as business consultants or in other professional service firms and such experience among people in their social circle (i.e. relatives, friends, colleagues). The dependent variable is the preference for the respective job attribute (see intercepts in Table 4).
The likelihood-ratio statistic (or chi-square statistic) (R Total response time (in minutes) (SD 5 2.47) 10.05 Note(s): a Based on the 232 respondents remaining after elimination of seven respondents due to missing values and implausible answers in the socio-demographic variables  The logic of attraction Subjects with personal experience as a business consultant or in other professional service firms especially shy away from salary and bonuses which are not based on job tenure and company performance (mean 5 À0.30, SD 5 0.25 and mean 5 À0.41, SD 5 0.47). Furthermore, they dislike professional training and expertise (mean 5 À0.16, SD 5 0.25) and instead prefer on-the-job training and personal development (mean 5 0.52, SD 5 0.48). This indicates that personal experience leads subjects to focus more on their own career, especially on salary and bonuses as well as training and development. Furthermore, they prefer worklife balance and a friendly climate (mean 5 0.45, SD 5 0.30).
Whether the subjects' parents work as business consultants or in other professional service firms has either only a very small impact on the preferences or standard deviations are very high. However, the work of the subjects' parents has a small impact on the preference for salary and bonuses which are based on qualifications and clients consulted (mean 5 0.22, SD 5 0.25). Nevertheless, the impact of personal experience on this preference is a lot higher (mean 5 À0.30, SD 5 0.25) and leads respondents to dislike these forms of compensation. One explanation could be that it takes a lot more effort to personally increase one's bonuses by consulting more clients in a business consulting company than in other professional services firms. In business consulting, professional fees have to be negotiated, whereas there are more or less predefined fees for services offered by, e.g. lawyers and tax consultants. Within the category of professional services firms, 23.6% of the respondents but only 0.6% of the parents worked as business consultants. This large difference in direct experience with business consultancy could explain these opposite preferences.
The influence of relatives who are not parents is very heterogenic as standard deviations are very high. This indicates that experiences of relatives have a less directional influence as, e.g. personal experience as well as the experience of friends and colleagues.
The occupation of the respondents' friends has a similar influence as their own personal experience. Both lead the subject to dislike salaries based on qualification and bonuses based on consulted clients (mean 5 À0.11, SD 5 0.24) and to dislike professional training and expertise (mean 5 À0.33, SD 5 0.25). The same is true for the preference of work-life balance and a friendly climate (mean 5 0.51, SD 5 0.32).
The colleagues' occupation has a similar, but less intense influence on the aversion to professional training and expertise (mean 5 À0.13, SD 5 0.32) and on the preference for work-life balance and friendly climate (mean 5 0.33, SD 5 0.40). The impact of the preference for salary and bonuses which are based on qualifications and clients consulted is small (mean 5 0.21, SD 5 0.32). The impact of personal experience on this preference is a lot higher and leads respondents to dislike these forms of compensation. However, only 16% of the respondents reported having colleagues who work as business consultants.
Family status has an influence, but standard deviations are mostly relatively high. However, singles especially prefer work-life balance and a friendly climate (mean 5 0.84, SD 5 0.30). Furthermore, they prefer salaries based on qualification and bonuses based on clients consulted (mean 5 0.60, SD 5 0.23). This indicates that singles prefer salaries which they personally can increase by being more successful but still seek work-life balance.
Both personal experience and the friends' and colleagues' occupation have a strong directional influence. This indicates that people who probably are of the same age as friends and colleagues have more influence on a subject's preferences than parents and relatives who most often are much older. It is noteworthy that the direction of the influence is always the same (with one small exception). Overall, personal experience and socialization especially influence the negative attitude towards salaries based on qualification and bonuses based on clients consulted, the preference for work-life balance and a friendly work climate as well as the negative attitude towards professional training and expertise.
It is also remarkable that personal experience, the occupation of friends and colleagues, as well as family status have a strong positive influence on the preference for work-life balance and a friendly work climate, whereas the parents' occupation has a very small negative influence. This indicates that the younger generation might have a decidedly different attitude towards work-life balance.

Discussion
Examining applicant attraction from a sociological perspective, this study explores the institutional uniformities and socialization-based antecedents of decision-making behind job pursuit intentions for an entry position in consulting among business student. It is of interest, whether graduates in a more applicant-driven job market use their sought-after position to develop more heterogeneous preference structures for job attributes (reflecting values) based on non-employer controlled influences, such as their own socialization. This would challenge one of the core assumptions of the fit-paradigm, namely that applicants self-select into organizations by conformity with organizational values (Cable and Judge, 1996), and would confront organizations with the difficulties of managing more diverse applicant pools along with more diverse employee demands in the long run. Using a hierarchical Bayes conjoint analysis, the study makes two main contributions to understanding the heterogeneity of job preferences that are relevant for both the institutional literature and the recruitment literature, and specifically to research on early applicant attraction.
First, by looking at job preferences with an institutional lens, the study reveals that potential applicants from the generation Y are no homogeneous group and that their preferences for job attributes are complex, as the favored attributes reflect different value sets that we conceptualize here as institutional logics (Friedland and Alford, 1991). Identifying two preference types for a graduate job in consulting among business students, the level of surface uniformity in preferences is higher than in previous research that identified four preference types (Guillot-Soulez and Soulez, 2014). Overall, the results show that there is a strong collective interpretative scheme among business students focusing on work-related job attributes (work climate and company description, job design) from which a small careeroriented group deviates (26%), focusing on pay and promotion related job attributes. The results add to institutional literature (Pache and Santos, 2013) by uncovering how individuals deal with multiple and potentially conflicting institutional logics in decision making practice. In our choice experiment, both clusters of participants base their job pursuit intention on a combination of job attributes representing all three logics-the logic of the corporation twice, the logics of family and of profession onceto form their job pursuit intentions. Because different logics trigger different demands, this is the first reason why we consider the job preferences complex. Zooming in, we find that for the career-oriented type the four attributes are next to equally important, whereas for the job-oriented type the work climate attribute and the salary and bonus attribute are particularly important. Apart from that, we can show that the two types share literally no preference overlaps, neither in attribute rankings nor in dominant logics per attribute. In other words, participants seek the same logic in different attributes, adding another dimension of institutional complexity (Meyer and H€ ollerer, 2016) and a third reason for the complexity of job preferences. For example, the career-oriented type associates promotion (on-the-job training and personal development) and the job-oriented type pay (fringe benefits and team performance) with the logic of the family. Thus, our results highlight that a relatively narrow typology can disguise institutionally heterogeneous and therefor complex preferences. For recruitment research in the area of applicant attraction which is dominated by organizational psychology (Ehrhart and Ziegert, 2005) and the brand equity perspectives (Theurer et al., 2018), we believe this can underscore the usefulness of social theory for making sense of applicant attraction.

The logic of attraction
What is more, the logic-based preference structure reflects the important role of the corporation as locus of identification in early careers and newcomers' hope to build a caring and reciprocal relationship with the employer (Clarke, 2013). This result challenges previous research stipulating low levels of generation Y commitment to employers ( € Ozçelik, 2015). In fact, it adds further to the critical debate surrounding the "new career" characterized by highly mobile and self-managed development models (Clarke, 2013;King, 2003) by reiterating the relative importance of the corporation. While Guillot-Soulez and Soulez (2014) have already questioned the value of the generation Y concept for early applicant attraction based on the plurality of preference types, we do so on the basis of underlying values. Any specific demands of this new employee generation on the corporation, particularly regarding career development, may get more pronounced once hired, as the demands then can be more legitimately expressed as committed organizational members (Naim and Lenka, 2018).
Second, by highlighting the role of anticipatory socialization for developing job preferences, which to the authors' best knowledge has not been researched in the context of early applicant attraction before, we contribute to understanding their formative milieu (Evertz and S€ uß, 2017). Our study finds that participants who have acquired more personal insights into PSFs have different job preferences than their fellow students. Having friends and colleagues working in PSFs alters the individual preference structure while having parents and close relatives working in PSFs has only marginal influence. Against the background of the prominent position of the family in anticipatory socialization theory, this is an interesting finding. This does not mean that parents and close relatives are per se irrelevant for newcomers' broader career decisions (Super, 1985) but that they may not be so helpful for job choice because their job records typically do not include "new professions" (Schein, 1977), such as consulting (McKenna, 2006). The lack of family influence may be traced back to socialization towards reference groups and not within membership groups (Merton, 1949). Another explanation is that parents lose normative influence to their adult children as family life advances, changing socialization from a unidirectional into a more reciprocal process (Glass et al.,1986).
Examining potential applicants' anticipatory socialization as explanatory factor for heterogeneous job preferences, the study adds new insights into the micro-foundations of institutional agency as we empirically test specific forms of previous exposure to values (anticipatory socialization) for adherence to those values or logics in decision-making. Drawing on institutional research, this study shows that in the mature field of PSFs with strong internalized rules of the game (Alvesson and K€ arreman, 2004) actors at the margins of the field, who are deeply embedded in other social contexts such as university or family, question the rules when they get to know them (Greenwood and Suddaby, 2006) In particular, previous exposure to the logic of profession does not necessarily strengthen the level of adherence to this logic but can result in selective defection due to for example saturation, frustration or more appealing alternatives. To participants who have already experienced PSFs, either directly or indirectly, the family logic becomes more important. These students chose the family logic for the work climate attribute and for the training and development attribute. Friends and colleagues familiar with PSFs also influence students to assign more importance to the family logic for the salary and bonus attribute. These results are consistent with the observation that previous interns tend to appreciate a more balanced work environment (Guillot-Soulez and Soulez, 2014). The influence of work experience may result from the "reality shock" newcomers experience after organizational entry, which leads to critical reflection of work practices (Louis, 1980). The results suggest that the anticipatory socialization perspective advances knowledge on individual differences, a renewed focus in applicant attraction research (Evertz and S€ uß, 2017) and in upcoming differentiated approaches to human resource management (Rofcanin et al., 2019). In particular, the results recall the seminal argument of the "strength of weak ties" (Granovetter, 1973) that shares with institutional theory the assumption that individuals' decisions are neither socially determined nor only based on self-interest (Thornton et al., 2012). Due to their embeddedness, job seekers can draw on weakly tied contacts who may hold different values and understandings than the people in the close and hence coherent social environment. In this regard, it is not unlikely that distant friends and colleagues influence students' job preferences more than their parents.

Practical implications
Applicant attraction in more applicant-driven job markets requires employee-centric recruitment activities grounded in consumer-centric paradigms (Ronda et al., 2018). Therefore, HR practitioners need to understand what potential applicants value most during the initial job search phase and what influences those preferences. Conjoint analysis, the base method of this study, has its roots in marketing and provides recommendations for effectively addressing applicant needs on the basis of the highest utility functions.
Tailoring recruitment messages: At the early recruitment stage, employers have limited control over their target audience in the graduate job market. As the aim is to raise awareness among potential applicants, recruitment messages need to sound attractive to a large number of students. Our results show that preferences differ among job-oriented and career-oriented types. In addition, the study implies that a habituated approach to recruitment (Ertimur and Coskuner-Balli, 2015), recruiting newcomers which are already socialized (e.g. through internships), necessitates different recruiting messages and subsequent realities. Recruitment messages should take into account these differences, which is in accordance with the current trend towards HR differentiation (Rofcanin et al., 2019). Organizations could, for example, use workplace testimonials from employees who represent those three different groups of potential applicants and set up multiple career paths which fit the diverse ambitions.
Designing job ads: Job ads as recruiting instrument grant employers a confined and structured frame to spread recruitment messages. The results from our analyses imply that PSFs make the most of jobs ad when they promote varied jobs for organized and committed employees (company description and job design) and promote a self-determined and enthusiastic work environment (work climate). These two ad attributes appear appealing to the large group of business students representing the job-oriented type (76%) and outweigh pay and promotion aspects. Thus, job ad aspects which allow potential applicants to compare their work values with the values of the company and describe what actual work is like should be at the heart of the job ad. HR practitioners need to make sure that the cultural expectations the job ad generates can be met in practice. Pay and promotion attributes then could be designed to address the preferences of the career-oriented type and as such promote tenure-based salaries plus bonuses along with team-based on-the-job training, mentoring and personal development. Since work-life-policies become more important for experienced graduate applicants, they should be mentioned in the job ad but as addition to the four key attributes, for example in an extra "additional benefits" section. Looking ahead to the future of the consulting field, such an ideal job ad would reflect a PSF employer uniting aspects typically associated with the more structured and process-oriented work environment of inhouse-consulting and tenure-based salaries as well as bonuses.
Managing anticipatory socialization: Our results show that previous experiences with PSFs make a difference in job preferences and let potential applicants put more focus on the social side of a job offering, i.e. on the team and their work-life-balance. That is why organizations need to take the influence of internship experiences and word of mouth serious and HR practitioners should have a genuine interest in enhancing and maintaining positive work conditions for student interns as well as permanent staff. The influence of persons who The logic of attraction have work experience, friends and colleagues, imply that an organization's anticipatory socialization tactics are helpful if they are able to get aspiring members in touch with employees, for example through peer mentoring, buddy programs or campus ambassadors. The opportunity to communicate with employees from intra-generational reference groups seems to be an obvious tactic; however, students may receive different information from more senior contacts outside their generation. HR practitioners may help to facilitate aspiring newcomer-senior communication by organizing casual and confiding networking events. Because our respondents were surprisingly disinterested in the professional side of a PSF job, discussing the role and meaning of expertise should be an integral part of anticipatory socialization tactics.

Limitations and future research
The study did not control for how satisfied the parents, relatives, friends and colleagues were with their occupation. Thus, different experiences of the members of the respective social group and differently told stories could explain high standard deviations. Findings may also vary for other cultural contexts, as socialization in non-Western contexts may be more narrow in terms of demanding obedience to and conformity with cultural standards (Arnett, 1995). Not only personal experience and socialization but personality traits as well could explain personal preferences. The limited role of the profession logic, for development and for salary, indicates that business students lack a clear picture of what business consultants actually do. This finding may vary for other, more traditional service professions such as law and accounting, which require mandatory training, professional certificates and membership in professional bodies (Scholarios et al., 2003). Attribute preferences can hardly be generalized for more experienced job seekers (Berthon et al., 2005), likely because increasing work experience leads to a less fragmented understanding of job attributes (Ito et al., 2013) . There is a recent turn in recruitment research towards reconsidering the complexity of values underlying job attributes (Bullinger and Treisch, 2015;Ronda et al., 2020) using somewhat non-standard techniques and a constructivist methodology to which this study adds. By treating early applicant attraction as a social phenomenon, the study redirects attention to the local conditions of recruitment activities where applicant expectations are influenced by institutional logics and socialization in various social groups. As we identify a dominant job-oriented preference type and a minor career-oriented preference type, we find applicant expectations to be heterogenous and complex. Anticipatory socialization, especially personal prior work experience and the occupation of friends, decreases adherence to the logic of profession and increases the relevance of the family logic. In this regard, our results stress how important the (re-)interpretations of HR practices based on practical experiences are for soon-to-be graduates to form job pursuit intentions (Harris and Pattie, 2020). Future research may thus look at the micro-factors explaining adherence to and defection from logics in early careers. Exploring the role of organizational and peer reference groups seems to be a particular promising venue. Finally, by illustrating the institutional complexity of applicant attraction, an institutional logics perspective suggests recruitment to be an issue of governancemanaging the multiple identities of the organizational self towards external stakeholders in a coherent way (Pache and Santos, 2013).