Perceived dissimilarity and knowledge sharing with younger workers: the role of generativity striving and knowledge receiving from younger workers

Abstract

Drawing on self-categorization theory, this study examined the impacts of perceived age and deep-level dissimilarities with younger workers on older workers’ tacit and explicit knowledge sharing (KS) with younger workers via generativity striving (GS), and extended the theory by proposing the moderating role of knowledge receiving (KR) from younger workers. This study used a three-wave online survey of 570 older workers in a large Chinese aircraft maintenance company. The results showed that GS mediated both the positive relationships between perceived age dissimilarity with younger workers and older workers’ tacit and explicit KS with younger workers, as well as the negative relationships between perceived deep-level dissimilarity (PDD) with younger workers and older workers’ tacit and explicit KS with younger workers. Moreover, the positive direct impact of GS and the negative indirect impact of PDD with younger workers on older workers’ explicit KS with younger workers were found to be relatively weaker when older workers’ KR from younger workers was high. The findings suggest that perceived age and deep-level dissimilarities with younger workers present both opportunities and challenges for older workers to share knowledge with younger workers.

Organizations are facing significant development issues arising from the aging workforce. One major concern is the impending retirement of a substantial number of older workers, which will inevitably result in the loss of valuable tacit and explicit knowledge (Harvey, 2012). Organizations must ensure that crucial knowledge, especially tacit knowledge, is effectively transferred from older to younger workers (Fasbender & Gerpott, 2021; Harvey, 2012). Explicit knowledge is transmitted in the form of codes or formal language (e.g., rules and procedures), which is often referred to as “know-what” (Smith, 2001). Tacit knowledge refers to the practical experience gained by individuals and stored in the memory of actors, which is rarely expressed publicly and is often referred to as “know-how” (Smith, 2001). The protection of tacit knowledge in the context of changing demographics is a significant challenge for organizations (Schmidt & Muehlfeld, 2017). Older workers are considered a valuable source of tacit knowledge within organizations, which is at a greater risk of being lost as a large number of older workers are set to retire in the future (Joe et al., 2013; Schmidt & Muehlfeld, 2017). To address the risk of organizational memory loss, scholars have been particularly interested in knowledge sharing (KS) between older and younger workers (e.g., Burmeister & Deller, 2016; Fasbender & Gerpott, 2021). In this study, we define KS as the process of transferring knowledge from one person to another (Lee, 2001). Demographic dissimilarity interferes with the flow of knowledge and influences individuals’ decisions on whom they choose to share knowledge with (Ku, 2019). Understanding the demographic dissimilarity between older and younger workers, which can range from surface-level dissimilarity to deep-level dissimilarity, is a pressing management issue (North & Fiske, 2015). Surface-level dissimilarity refers to differences in overt demographic characteristics (e.g., age, gender, and race), while deep-level dissimilarity refers to differences in psychological characteristics (e.g., personality, values, and attitudes; Harrison et al., 2002). In the context of an aging workforce, we focus on perceived age dissimilarity (PAD) and perceived deep-level dissimilarity (PDD). Perceived dissimilarity is a suitable operational approach, aligning with the tenets of the theoretical foundations of existing research on age dissimilarity, such as the self-categorization theory (SCT; e.g., Avery et al., 2007; Williams et al., 2007). Both PAD and PDD with younger workers influence older workers’ self-categorization and social identity processes (Riordan, 2000). Social identity is defined as “the nature or content of an individual’s perceived membership in a social group” (Zacher et al., 2019). SCT emphasizes that individuals define themselves through social identity in the process of social comparison (Spears, 2011). The process of social comparison should be viewed as an individual’s psychological process, rather than as necessarily an objective assessment of reality. Considering the role of individuals in self-categorization, we need to acknowledge that social identity is personal and subjective (Montepare & Lachman, 1989). Individuals align with group norms when they subjectively identify themselves as older workers (Harwood et al., 1995).

Previous research on the impact of age dissimilarity on KS has shown mixed results. Using the D-Score approach as the measuring method, Zenger and Lawrence (1989) suggested that workers tended to communicate more frequently about technical issues with those who were similar in age to themselves than with those who were dissimilar. Calculating the absolute difference in age, Ku (2019) found that age dissimilarity negatively affected KS. However, these objective approaches to measuring age dissimilarity obscure the nuances of direction in age dissimilarity and assume that differences of the same magnitude are equivalent. Taking age as an example, the D-Score approach assumes that a 10-year age dissimilarity has the same impact on an individual, regardless of whether the dissimilarity is younger by 10 years or older by 10 years (Riordan, 2000). If the operationalization of age dissimilarity does not clearly reflect the direction of age dissimilarity, it will overlook the potentially meaningful true impact. For example, Burmeister et al (2018) found that the age of younger workers had a negative indirect impact on older workers’ KS with younger workers. This ingenious method indirectly captures the direction of age dissimilarity, where the younger the workers, the more knowledge older workers share with them. Meanwhile, a qualitative study revealed that older workers felt a responsibility to selflessly share their knowledge with younger workers (De Blois & Lagacé, 2017). We support the view that age dissimilarity does not inherently impede KS, but rather the direction of age dissimilarity is important. Grounded in SCT, perceived dissimilarity in relation to social identity is crucial for understanding KS. In order to better operationalize the direction of age dissimilarity, based on SCT, we operationalize age dissimilarity as older workers’ PAD with younger workers and operationalize deep-level dissimilarity as older workers’ PDD with younger workers. Through this approach, we can better reflect how the two types of social identity processes, based on PAD and PDD with younger workers, differently impact older workers’ KS with younger workers. To provide a clearer understanding of the theoretical framework underpinning our study, Figure 1 illustrates the relationships between the core variables discussed in this study.

Significantly, selecting appropriate mediating variables is an effective way to clarify the relationships between demographic dissimilarity and outcomes (Lawrence, 1997; Riordan, 2000). SCT describes how group membership is self-defined and how this self-definition generates motivation that is oriented toward the group (Haslam et al., 2000; Van Knippenberg & Hogg, 2003). Both PAD and PDD with younger workers activate the social identity of older workers. Generativity striving (GS) refers to an individual’s interest in and enjoyment of organizational practices such as mentoring, training, and sharing skills with young people (Kooij & Van De Voorde, 2011). GS is an age-related motivation that is important for older workers (Kooij et al., 2011). Society expects the elderly to be accountable to the young (McAdams et al., 1993). In this study, we believe that SCT and GS (a mediator) are the key to unlocking the black box of the relationships between PAD/PDD with younger workers and older workers’ KS with younger workers. Moreover, Fasbender et al. (2021) provided empirical support for the relationship between GS and KS. Fasbender et al. (2021) did not classify the type of knowledge. Hau et al. (2013) found that enjoyment (an intrinsic motivation) had a greater impact on tacit KS than on explicit KS. GS is an intrinsic motivation (Fasbender et al., 2021). It is unclear whether the impacts of GS on the two types of KS are consistent. Hence, we conduct empirical tests to examine the difference in the impacts of GS on the two types of KS. SCT suggests that the translation of group-oriented motivation into effective action is influenced by various factors. These factors depend on the level of normative support for this action (Van Knippenberg & Hogg, 2003). It is unclear which factors interfere with the relationship between GS and older workers’ KS with younger workers. Knowledge receiving (KR), also known as knowledge seeking, refers to the act of individuals seeking knowledge from others (Dietz et al., 2022). There are implicit normative expectations, such as the belief that older workers are the best candidates for KS (Burmeister et al., 2018). Older workers’ KR from younger workers goes against normative expectations (Chaudhuri & Ghosh, 2012). We propose that KR from younger workers may moderate the relationship between GS and older workers’ KS with younger workers.

Our study has several theoretical contributions. First, this study will expand the antecedents of GS by examining the double-edged impacts of PAD and PDD with younger workers on older workers’ GS. Considering only the in-group/out-group phenomena is insufficient for exploring the consequences of intergroup dissimilarities because out-group discrimination is not the sole means to differentiate (Spears, 2011). To provide a more comprehensive explanation of the social identity phenomenon, it is important to consider the perspectives on group prototypes in SCT. This will allow for a more precise understanding of the cognitive foundations of the social identity phenomenon (Hogg & Terry, 2000). Based on SCT, this study will explain how older workers categorize themselves and younger workers into different social groups through PAD and PDD with younger workers, and how this categorization influences their group-oriented motivation (GS) based on appropriate group prototypical characteristics. Second, this study will provide further empirical support for the mediating role of GS. Prior to this study, there was no empirical research demonstrating whether the relationships between PAD/PDD and KS are mediated by GS. GS has already been theorized in the knowledge literature as one of the main mechanisms linking age dissimilarity and knowledge exchange outcomes (e.g., Fasbender et al., 2021; Gerpott et al., 2020). This study will deepen our understanding of this emerging theory by demonstrating that GS mediates the relationships between PAD/PDD with younger workers and older workers’ KS with younger workers. Third, this study will expand our understanding of a specific contextual condition that may moderate the impact of GS on older workers’ KS with younger workers by investigating KR from younger workers as a moderator. Although previous research has found that GS predicts older workers’ KS with younger workers (Fasbender et al., 2021), our study will further demonstrate that the positive impact of GS on older workers’ explicit KS with younger workers is relatively weaker when older workers actively seek more knowledge from younger workers.

Theoretical background and research hypotheses

Dissimilarity and SCT

Scholars have focused on exploring the consequences of demographic dissimilarity within an immediate work group or a larger work unit (Reinwald & Kunze, 2020). Previous empirical research found that age similarity influenced technical communication (Zenger & Lawrence, 1989), and age dissimilarity affected KS (Ku, 2019). Additionally, the literature on mentoring claimed that deep-level dissimilarity between mentors and protégés can trigger negative mentoring relationships (e.g., Hu et al., 2014; Lankau et al., 2005). These studies used the relational demography approach to investigate the consequences of age dissimilarity and deep-level dissimilarity. The relational demography approach posits that demographic dissimilarity between individuals within a social unit affects individual-level outcomes (Guillaume et al., 2012; Riordan, 2000; Roberson, 2019). The relational demography approach serves as a valuable addition to conventional diversity studies (Reinwald & Kunze, 2020). Conventional diversity studies view diversity as a structural characteristic of units and aim to evaluate how variations in composition between units contribute to differences in outcomes at the unit level (Guillaume et al., 2012). Our study aims to employ the relational demography approach to obtain a comprehensive understanding of how PAD and PDD with younger workers influence older workers’ KS with younger workers. There are two distinct approaches for measuring demographic dissimilarity in relational demography studies: The objective approach and the perceptual approach. The objective approach measures the actual dissimilarities among individual demographic characteristics (Riordan, 2000). The perceptual approach emphasizes that individuals subjectively compare and assess their objective demographic dissimilarities with others and assign unique psychological significance to these dissimilarities (Harrison et al., 2002; Riordan, 2000). In this study, we consider the perceptual approach to be a more suitable approach for measuring age dissimilarity and deep-level dissimilarity. Individual reactions arise from sensations and interpretations of reality rather than reality itself, and each individual may interpret objective dissimilarities differently (Shemla et al., 2016). Especially in terms of age, considering oneself as older or younger is inherently subjective (Montepare & Lachman, 1989). This view aligns with SCT, which suggests that older workers only begin to acknowledge the social identity of an “older worker” when they subjectively perceive significant age and deep-level dissimilarities with younger workers.

The minimal group paradigm is a crucial research methodology in social identity research that is used to investigate the minimal circumstances needed for people to differentiate between groups (Hogg, 2016). Scholars have found through extensive minimal group experiments that simply highlighting the distinction between “us” and “them” leads to intergroup discrimination and in-group favoritism (Hogg, 2016). Social identity theory (SIT) is developed based on the minimal group paradigm (Hogg, 2016; Hornsey, 2008). Thus, SIT is often used to explain out-group discrimination in minimal or quasi-minimal group settings (Brown, 2000; Spears, 2011). However, individuals who are categorized into certain groups are likely to show preferences for their own group, but they may not necessarily discriminate against members of out-groups unless they perceive threats from those groups (Mummendey & Otten, 1998). There may be more benign ways of differentiating between groups in social life (Spears, 2011). To provide a more comprehensive and effective explanation of the cognitive foundations of social identity phenomena, SCT is an appropriate theory (Turner & Reynolds, 2012). SCT posits that when individuals’ social identity becomes salient, their thoughts, feelings, and behaviors align with the prototypes that are relevant to the context (Hogg et al., 1995). The concept of social identity salience is significantly influenced by the notion of “fit” (Haslam et al., 2000). Social identity salience captures how individuals establish and amplify their social identity within a social context. Fit is a critical measure that evaluates the extent to which this social identity corresponds with the social reality of a given situation. According to Hornsey (2008), fit gauges the degree to which dissimilarities in the real world are perceived and assessed. Fit includes two dimensions: comparative fit and normative fit. Comparative fit refers to the degree to which social categorization results from the clustering of individuals who share significant similarities within a group, while also exhibiting notable dissimilarities between groups (Van Knippenberg & Mell, 2020). Comparative fit describes the group dissimilarity per se (Spears, 2011). For example, demographic characteristics such as age, race, and gender are relevant to the principle of comparative fit (Van Knippenberg & Mell, 2020; Van Knippenberg et al., 2004). When older workers perceive significant age dissimilarity with younger workers (high comparative fit), their social identity is activated (social identity salience; Van Knippenberg et al., 2004). Normative fit refers to the meaning or content associated with social groups and determines whether these dissimilarities are consistent with social groups’ stereotypes (Spears, 2011; Van Knippenberg & Mell, 2020). Age-based social categorization commonly carries related stereotypes that imbue them with meaning (Van Knippenberg & Mell, 2020). People generally have common expectations about the typical characteristics of older and younger workers. When the perceived deep-level characteristics of younger workers align with the expectations of older workers regarding how typical younger workers think or behave, there is normative fit. Deep-level dissimilarity captures the extent of normative fit (Adamovic, 2022; Valenzuela et al., 2020). When older workers perceive significant deep-level dissimilarity with younger workers (high normative fit), their social identity is activated. In general, once social identity is activated based on optimal fit, older workers are likely to attribute themselves and younger workers to the related group prototypes (Hogg & Terry, 2000). The group prototypes serve as the cognitive basis for older workers to perceive and emphasize intergroup dissimilarities. Group prototypes are similar, fuzzy collections of attributes (e.g., cognition, emotion, and behavior) that are shared by the in-group and out-group in a certain environment (Hogg, 2016). Group prototypes highlight the greatest dissimilarities between the in-group and out-group (Hogg, 2016; Hogg & Terry, 2000; Van Knippenberg & Hogg, 2003). One way to conceptualize a group prototype is by considering what immediately comes to one’s mind when the phrase “older workers” is mentioned (Hogg, 2016). If one believes that older workers have acquired much experience and knowledge (Finkelstein et al., 2013; Weidner et al., 2024) but may be out of touch and set in their ways (Finkelstein et al., 2013; Posthuma & Campion, 2009), these descriptions can be considered the prototype of older workers. The prototype can also become a stereotype when it is widely accepted by the majority (Hogg, 2016). A group stereotype refers to a set of characteristics that individuals collectively view as typical of a particular social group (Judd & Park, 1993).

Next, we will begin a detailed discussion of the hypotheses related to the theoretical framework of this study based on SCT. For a better understanding of our reasoning in developing hypotheses, please refer to Figure 1.

PAD with younger workers and GS

Social categorization provides both in-group and out-group prototypes, which describe the appropriate psychological and behavioral norms of the groups (Hogg, 2016). Comparisons with different referent groups have varying direct impacts on individuals’ self-categorization (Van Knippenberg et al., 2004). Age is one of the readily observable physical characteristics that can be used to classify people socially (Montepare & Zebrowitz, 1998). When older workers perceive significant age dissimilarity with younger workers (high comparative fit), their social identity becomes salient. People cognitively represent social groups with prototypes. As suggested by Harrison et al. (1998), when individuals categorize themselves and others based on surface-level characteristics, their perceptions of the in-group and out-group should be based on group-related public stereotypes (prototypes). A key feature of prototypes is that they maximize in-group similarity and intergroup dissimilarity to the greatest extent (Hogg, 2016). The more salient the social identity of older workers, the more likely they are to describe themselves and younger workers using age-related prototypes (Montepare & Zebrowitz, 1998). Different age categories have both positive and negative prototypical characteristics (Harwood et al., 1995). Older workers are likely to take pride in their experience, or they may feel shame for their decline (Harwood et al., 1995). Based on SCT, people seek positive social identity (Spears, 2011). In the case of social categorization, older workers are more likely to identify with positive prototypical characteristics of the in-group. Furthermore, group members actively promote the belief that “we” are better than “them,” attaching the status, prestige, and social value of the in-group to themselves (Hogg, 2016). Older workers are more likely to adopt prototypical characteristics that favor the in-group over the out-group (younger workers) to maximize intergroup dissimilarity. The prototype of older workers not only includes being knowledgeable, experienced, mature, mentors, and role models, but also being perceived as out of touch, narrow-minded, stubborn, and conservative (Finkelstein et al., 2013). The prototype of younger workers not only includes energy, enthusiasm, tech-savviness, and ambition but also encompasses inexperience, narcissism, arrogance, self-absorption, entitlement, and thoughtlessness (Finkelstein et al., 2013; Rauvola et al., 2019).

Based on existing research, we believe that when social categorization is based on PAD with younger workers, older workers rely on prototypical characteristics that are more realistically associated with age to view themselves and younger workers (De Blois & Lagacé, 2017; Finkelstein et al., 2013). Age attributes can be used to infer differences in experience, skills, and knowledge. There is a practical correlation between age and knowledge, as knowledge increases with age (Buengeler et al., 2016). Knowledge can maximize the age-based dissimilarity between older and younger workers. After socially categorizing themselves and younger workers based on PAD with younger workers, older workers are likely to rely on their experience and knowledge to define themselves, while they tend to view younger workers as inexperienced. Older workers can better highlight the status and value of their group and achieve self-improvement by defining themselves and younger workers in this way. De Blois and Lagacé (2017) and Finkelstein et al. (2013) asked older workers to describe the stereotypes that younger workers held about them based solely on chronological age. De Blois and Lagacé (2017) found that older workers had a positive perception of younger workers’ evaluations of them and relied on the positive stereotype of being experienced to assess themselves. Meanwhile, older workers considered younger workers to be inexperienced, while viewing their own experience as beneficial to young people. Through the use of quantitative methods, Finkelstein et al. (2013) found that older workers believed that younger workers held more positive stereotypes about them than negative stereotypes. Older workers believed that the most significant positive stereotype of their group by younger workers was being experienced. Meanwhile, they perceived the most significant stereotype of younger workers to be a lack of experience. According to SCT, group members strive to achieve a positive social identity to fulfill their motivation for self-improvement (Hogg, 2016). Older workers can leverage positive age stereotypes to promote a favorable group image (Hummert et al., 2004). SCT describes how individuals’ self-definitions, based on their group memberships, generate motivations that are oriented toward the in-group (Van Knippenberg & Hogg, 2003). GS (a motivation) reflects the preference and intrinsic desire of adults to mentor and promote their offspring (Kooij & Van De Voorde, 2011). When older workers believe that their knowledge and experience are valuable and needed by younger workers, they expect to take on responsibilities for less experienced younger workers (McAdams et al., 1993). In summary, when PAD with younger workers activates the social identity of older workers, they are more likely to exhibit higher GS.

 

PDD with younger workers and GS

Social groups can be distinguished not only based on age but also based on deep-level characteristics (Tajfel & Turner, 1986). Studies found that workers of different ages exhibited differences in personality (Diehl et al., 2020), values, and job attitudes (Rhodes, 1983). Additionally, Lazarus (1996) found that younger individuals tended to employ more active and interpersonal problem-centered coping strategies, whereas older individuals tended to rely on passive and intrapersonal emotion-centered coping strategies. Unlike surface-level characteristics (e.g., age), deep-level characteristics are acquired through time and interaction (Harrison et al., 1998). Indeed, significant dissimilarities in deep-level characteristics such as values, interests, personality, outlook on organizational issues, and problem-solving approaches are inherently present between older and younger workers. Older workers should anticipate significant deep-level dissimilarity with younger workers, even in the absence of direct interaction. However, this is merely a typical expectation or stereotype about social groups. As stated by Harrison et al. (1998) and Guillaume et al. (2012), accurate information regarding deep-level characteristics is conveyed through verbal and nonverbal behavior patterns. This information can only be perceived and comprehended through prolonged, personalized interaction and data collection. After a period of interaction, older workers acquire accurate information about the deep-level characteristics of younger workers. When the expectations regarding the content or meaning of social groups are confirmed during daily interactions in the workplace, the social identity of older workers is more likely to be activated. Conversely, if these typical expectations related to social groups are not confirmed after a period of interaction (e.g., if older workers discover that their values do not significantly differ from those of younger workers but rather have common ground), then it is less likely that the social identity’s meaning (truly identifying with one’s social identity psychologically) will be invoked to define themselves. When older workers perceive significant deep-level dissimilarity with younger workers that aligns with typical expectations (high normative fit), their social identity becomes salient.

Social categorization does not necessarily trigger intergroup discrimination without out-group threats (Mummendey & Otten, 1998; Spears, 2011). When older workers socially categorize based on PAD with younger workers, they are less likely to discriminate against younger workers. Instead, older workers’ absolute advantage in experience and knowledge may inspire more GS. SCT posits that intergroup discrimination is typically triggered by the perception of threats from the related out-group (Van Knippenberg et al., 2004). Deep-level characteristics (e.g., value) touch on the core of the self, referring to what one considers fundamentally right or wrong (Urick et al., 2017). The out-group with different values and attitudes can be seen as threats to the core identity of the in-group. When older workers genuinely perceive that younger workers hold a set of work values and attitudes starkly different from their own, they may worry that these emerging values will gradually supplant the ones they have long cherished. For older workers, this may be seen as a signal of threat, indicating that they may feel challenged by younger workers. When older workers engage in social categorization based on PDD with younger workers, they are likely to discriminate against the younger workers. For example, Urick et al. (2017) found that younger workers tended to embrace innovation in terms of values, while older workers leaned toward maintaining the status quo. Older workers observed a mismatch in values with their younger colleagues, noting the younger group’s tendency to eschew established protocols and the perceived hesitance to absorb knowledge from others (Urick et al., 2017). Older workers pointed out that their younger colleagues often labeled them as “old” and overlooked the wealth of experience and knowledge they actually held (Urick et al., 2017). Based on the observations of PDD with younger workers, older workers might perceive younger workers as self-absorbed, arrogant, narcissistic, thoughtless, and entitled. These negative perceptions might stem from younger workers’ disregard for traditional practices and their self-driven work styles. PDD with younger workers may promote older workers to perceive themselves as having a wealth of experience, but also considering that their knowledge may not be appreciated by younger workers (Cheng et al., 2008; Urick et al., 2017). Meanwhile, Urick et al. (2017) found that older workers perceiving insurmountable value dissimilarity with their younger colleagues resorted to a strategy of “self-removal,” which entailed distancing themselves from interactions with their younger colleagues. Similarly, Cheng et al. (2008) found that older people in Hong Kong perceived dissimilarities in values and lifestyles with the youth, who were unwilling to listen to their advice, thereby diminishing the elders’ generativity. In summary, when PDD with younger workers activates the social identity of older workers, they are more likely to exhibit lower GS.

 

GS and KS with younger workers

Older workers with higher GS have a stronger desire to leave a positive legacy for younger workers and aim to mentor younger workers in their own image (Kooij & Van De Voorde, 2011; McAdams, 2013). GS drives older workers to share their knowledge with younger workers (Fasbender et al., 2021). Knowledge includes both tacit knowledge and explicit knowledge. Tacit KS is a type of KS that goes beyond job requirements and emphasizes the achievement of exceptional performance (Zhao et al., 2023). Explicit KS is a type of KS that occurs within the scope of an individual’s job responsibilities and is influenced by job characteristics (Zhao et al., 2023). Significantly, the impacts of various antecedents on tacit KS and explicit KS differ. For example, Wang et al. (2022) found that virtual rewards positively influenced explicit KS but exhibited an inverse U-shaped relationship with tacit KS. Santos et al. (2023) found that trust and cooperation norms directly affected tacit KS but only indirectly affected explicit KS. Zhao et al. (2023) found no impact of extrinsic rewards on explicit KS but observed a negative impact on tacit KS. Fasbender et al. (2021) reported that GS was a key motivation for knowledge senders. Fasbender et al. (2021) did not distinguish between two types of knowledge, and it is unclear to us whether these two types of KS processes are consistent.

GS serves as a significant intrinsic motivation for older workers to share their knowledge with younger workers (Fasbender et al., 2021). Hau et al. (2013) found that the positive impact of enjoyment (an intrinsic motivation) on tacit KS was greater than that on explicit KS. Intrinsic motivation refers to the motivation to engage in an activity because one enjoys the task itself (Ryan & Deci, 2000). Tacit KS is a complex process that necessitates extensive face-to-face communication, rendering it more time-consuming and demanding compared to explicit KS (Hau et al., 2013). The inherent desire to mentor younger workers is even more crucial for older workers to impart their tacit knowledge to their younger colleagues. We hypothesize that GS has a more positive impact on tacit KS than on explicit KS.

   

The mediating role of GS

According to Lawrence (1997), it is challenging to draw consistent conclusions regarding the relationships between demographic dissimilarity and outcomes. Hence, it is very important to choose the appropriate theory and mediator. Based on SCT, when older workers perceive significant age dissimilarity with younger workers, it promotes meaningful social categorization. Meaningful social categorization requires considering the direction of age dissimilarity, as the impacts of age dissimilarity on outcomes may differ without this consideration. Some empirical research has found that age dissimilarity hinders KS without specifying the age ranges of the individuals involved in sharing and receiving knowledge (i.e., without clearly identifying the direction of age dissimilarity; e.g., Ku, 2019; Zenger & Lawrence, 1989). However, other empirical research has clarified this issue (e.g., Burmeister et al., 2018), demonstrating that age dissimilarity between older and younger workers promotes KS from older to younger workers. Burmeister et al. (2018) specified that the individuals sharing knowledge were older workers, while the recipients were younger workers. Based on the organizational theory of age effects, Burmeister et al. (2018) demonstrated that perceived ability to receive mediated the negative relationship between the age of younger workers and older workers’ KS with younger workers. It is necessary to gain insights into the intervening processes (mediators). Drawing on SCT, we suggest that PAD with younger workers facilitates older workers’ KS with younger workers via GS. The social identity of older workers becomes salient when they perceive significant age dissimilarity with younger workers. When older workers use age as a basis for social categorization, they may emphasize their prototypical characteristics—experience and knowledge—while highlighting the prototypical characteristics of younger workers, such as inexperience (De Blois & Lagacé, 2017; Finkelstein et al., 2013). The favorable comparison, which effectively enhances their collective reputation and social status, encourages older workers to develop a concern for younger workers and to foster their growth (Dunham & Burt, 2011; Finkelstein et al., 2013). Generativity is a concept that is graded by age and is considered normative (McAdams, 2013). Society expects more experienced older workers to be responsible for mentoring younger, less experienced workers. Thus, we expect that GS is more likely to arise when workers identify themselves as older individuals based on PAD with younger workers. In addition, motivation is the primary factor influencing KS among workers of different ages (Burmeister et al., 2018). Fasbender et al. (2021) found that GS was an important intrinsic motivation for older workers’ KS with younger workers. Hence, we suggest that higher PAD with younger workers is related to higher GS, thus encouraging older workers to engage in more KS with younger workers.

  

Given that mentoring involves the transfer of knowledge from more experienced workers to less experienced workers (Burmeister et al., 2020), we may be able to identify some empirical evidence drawn from the mentoring literature that supports the negative relationship between PDD with younger workers and older workers’ KS with younger workers. Eby et al. (2000) found that protégés were more likely to report that their mentors had different attitudes, values, and beliefs from them when describing their most negative mentoring relationships. The social identity of older workers becomes salient when they perceive that they exhibit significant dissimilarities with younger workers in terms of personality, values, and attitudes. Higher PDD with younger workers is more likely to make older workers feel threatened, and therefore, they are more likely to discriminate against younger workers. Hence, older workers may rely on younger workers’ negative prototypical characteristics—such as being self-absorbed, arrogant, narcissistic, thoughtless, and entitled—which can weaken their GS. If older workers’ motivation to leave a positive legacy for younger workers is undermined, it also appears to hamper their KS with younger workers. We propose that older workers who perceive higher deep-level dissimilarity with younger workers demonstrate lower GS and, as a result, are hesitant to share their knowledge with younger workers.

  

The moderating role of KR from younger workers

SCT has proposed that the degree to which group-oriented motivation translates into effective behavior is influenced by a number of factors. These factors depend on the degree of normative support for this behavior (Van Knippenberg & Hogg, 2003). According to the traditional view, expertise, and wisdom are associated with older workers (Gerpott et al., 2017). In the process of knowledge exchange between older and younger workers, age generates normative expectations regarding the roles that individuals play (e.g., Burmeister et al., 2018; Finkelstein et al., 2003). Burmeister et al. (2018) found that older workers had more experience and occupied the role of knowledge sender, while younger workers had less experience and occupied the role of knowledge receiver. Due to the expectation that age and experience are positively correlated, older individuals are perceived to be tasked with acting as mentors (Dunham & Burt, 2011). Based on these studies, we know that people typically have normative expectations regarding the KS role of older workers. This role is not only integral to internal role allocation and expectations but also essential for their sense of self-worth. This positioning often sees older workers as repositories of experience and knowledge. However, as dynamics change, especially in industries with rapid technological advancement and quick turnover of knowledge, older workers may find themselves seeking help for knowledge and skills from younger workers (Gerpott et al., 2017). In other words, older workers seeking knowledge from younger workers breaks the norm in traditional organizational culture where knowledge flows from older to younger workers. The reversal of roles might prompt them to reassess the relevance and value of their knowledge. As older workers increasingly turn to younger ones for assistance, they might start doubting the current value of their knowledge, especially in rapidly evolving fields. Even if they are motivated to share knowledge, this reassessment of self-worth and uncertainty about the value of their knowledge might diminish their enthusiasm for sharing. Moreover, frequently seeking knowledge from younger workers might make older workers feel in a state of debt, believing they owe their younger colleagues some form of reciprocation. This feeling could make them more cautious and selective in sharing knowledge, hoping to repay this debt by offering more valuable knowledge. Even though older workers are motivated to share knowledge, they are likely to tend to be reserved in doing so because they carefully consider the value of the knowledge they are sharing. Therefore, while older workers may be motivated to share knowledge, the increase in seeking knowledge from younger workers might psychologically reduce their efficiency in turning this motivation into action. Thus, we suggest that the positive impact of GS on older workers’ KS with younger workers is relatively weaker when older workers’ KR from younger workers is high (vs. low).

  

As we have proposed, the theoretical hypotheses suggest that GS mediates the relationships between PAD/PDD with younger workers and older workers’ tacit/explicit KS with younger workers. Additionally, KR from younger workers moderates the relationships between GS and older workers’ tacit/explicit KS with younger workers. The theoretical rationales behind these hypotheses also suggest that KR from younger workers should influence the strength of the indirect relationships, indicating a pattern of moderated mediation.

    

The role of control variables

As noted by previous studies, the informal and intimate relationship between older and younger workers may be one of the key factors affecting KS (Dietz & Fasbender, 2022; Zhang et al., 2022). According to Sluss and Ashforth (2008), there are two types of identity: GS, which is associated with an age-related motivation among older workers (social identity), and workplace friendship, which refers to a voluntary relationship between older and younger workers who are mutually attracted to and choose each other (relational identity; Nielsen et al., 2000). Workplace friendship with younger workers may reinforce older workers’ generative behaviors (Dietz & Fasbender, 2022). Therefore, we consider that workplace friendship with younger workers should be controlled in our model. Moreover, time pressure refers to either subjectively perceived time pressure or the imposition of a deadline (Škerlavaj et al., 2018). We believe that time pressure should be controlled because it can result in older workers being too occupied with their job duties to assist their younger colleagues (Fasbender et al., 2021; Gerpott et al., 2020).

In addition to workplace friendship with younger workers and time pressure, we considered several demographic variables that should be controlled, including age, organizational tenure, gender, and education. As adults grow older, they tend to perceive time in the future as limited. Consequently, generativity (motivation and behavior) increases accordingly (Gerpott et al., 2020). Similarly, older workers with longer organizational tenure are closer to retirement and become more focused on meaningful goals (Rudolph et al., 2018), such as generativity. Gender has been further considered as a control variable. Women are expected to take on more responsibility for raising offspring. Generativity theory suggests that women may exhibit more generativity (McAdams & de St. Aubin, 1992). Finally, we believe that education should be controlled because a meta-analysis conducted by Doerwald et al. (2021) found a significant and positive correlation between education and generative motivation. Individuals with a bachelor’s degree are more likely to share knowledge, as they possess a greater amount of knowledge resources (Fasbender & Gerpott, 2021; Gerpott et al., 2020). Hence, we suggest that age, organizational tenure, gender, and education should be considered alongside the impacts of PAD and PDD with younger workers on older workers’ GS. Meanwhile, age, organizational tenure, gender, education, workplace friendship with younger workers, and time pressure should be considered in conjunction with the impact of GS on older workers’ KS with younger workers.

Methodology

Samples and procedures

During the period between April 2022 and May 2022, we collected data from a large aircraft maintenance company in southwestern China. The company had multiple departments, each of which was further subdivided into numerous subdepartments. Regular collaboration and communication took place among the different subdepartments. Different subdepartments within the same department worked in the same building. Different departments were geographically separated. Our research focused on individuals who were over 45 years old (including 45 years old) and possessed high levels of knowledge and expertise. Based on previous studies on knowledge transfer between older and younger workers (e.g., Burmeister et al., 2018, 2020), we set 45 years as the appropriate age cutoff. According to the employee list, there were 717 workers who fell within the age range of older workers. We adopted an online questionnaire and released the survey link to the participants through email. To minimize the potential for common method biases, a temporal separation method was adopted to collect data at three different time points (Podsakoff et al., 2012). The time lag was two weeks, and the employee ID, kept confidential from researchers, was used as the labeling code.

In the first stage (Time 1), older workers reported their age, organizational tenure, gender, education, PAD with younger workers, PDD with younger workers, and KR from younger workers. Out of 717 participants, 657 returned the questionnaire, resulting in a response rate of 91.63%. Two weeks later (Time 2), older workers were invited to complete a questionnaire on GS, workplace friendship with younger workers, and time pressure. A total of 663 participants completed the second-stage survey, resulting in a response rate of 92.47%. Two weeks after the second stage (Time 3), older workers were asked to complete a questionnaire on tacit KS with younger workers and explicit KS with younger workers. A total of 627 participants finished the T3 survey, resulting in a response rate of 87.45%. Finally, after removing the unmatched samples, the total number of valid samples was 570, resulting in a response rate of 79.50%. The age of the respondents ranged from 45 to 60 years (M = 51.910, SD = 3.895), and their organizational tenure ranged from 12 to 44 years (M = 32.260, SD = 4.689). Of the participants, 458 (80.4%) were male, and 231 (40.5%) held a university degree.

Measures

All the measures used for this study were translated into Chinese following the procedure of translation and back-translation (Brislin, 1980). Except for the demographic variables and PAD with younger workers, participants completed the measures using a 5-point Likert scale (1 = strongly disagree, 5 = strongly agree).

PAD with younger workers

PAD with younger workers was assessed using the item “How dissimilar were you to younger colleagues of your department in terms of age?” The response scale ranged from 1 (very similar) to 5 (very dissimilar). This single-item measure of PAD with younger workers was adapted from Williams et al.’s (2007) scale for measuring PAD with teammates. Williams et al. (2007) asked participants about their perceived age similarity with teammates. Meanwhile, they assigned a low score to indicate similarity and a high score to indicate dissimilarity. First, to ensure consistency in the question and response format and to facilitate participant responses, we changed the type of judgment from similarity to dissimilarity. Second, we adapted the referent group. The referent group in this study differs from that of Williams et al. (2007). The referent group in Williams et al. (2007) is the teammates within the team, whereas our referent group is the younger workers within the department (a subgroup). Following the suggestions of Riordan and Wayne (2008), in measuring perceived dissimilarity entries, we guided participants (older workers) to (a) consider the department as the referent range, and (b) compare themselves with the younger workers within the department. On the one hand, this approach more clearly reflects how knowledge sharers perceive age dissimilarity with the beneficiaries, better illustrating the activation process of social identity. On the other hand, it allows for better comparison with the results of previous studies (Ku, 2019; Zenger & Lawrence, 1989). As noted by Riordan (2000), one primary reason for the contradictory results observed in many previous studies on demographic dissimilarity may be that some research focused on the impact of demographic dissimilarity within a workgroup, while other studies examined its impact at the departmental or entire organizational level. Comparing these studies may be inappropriate because workgroups, departments, as well as entire organizations, differ in terms of size, the level of interaction among group members, and group lifespan. To better compare with the results of previous research, it is more appropriate to adopt a similar organizational level when investigating the impacts of demographic dissimilarities. To better compare our findings with those of Ku (2019) and Zenger and Lawrence (1989), we opted to assess PAD with younger workers within a department, rather than within a work team.

PDD with younger workers

PDD with younger workers was measured by adapting the six-item version of the Deep-level Similarity Scale (Lankau et al., 2005; e.g., “Younger colleagues of my department and I were similar in personal values”). We reversed the encoding of ratings on the items so that higher ratings corresponded to higher PDD with younger workers.

KR from younger workers

The four-item KR Scale was adapted to assess older workers’ KR from younger workers (Wilkesmann et al., 2009; e.g., “I learned a lot by observing younger colleagues of my department doing their job”).

GS

Three items from Kooij and Van De Voorde (2011) were adapted to measure older workers’ GS, such as “It was important for me to have the chance to teach and train younger colleagues of my department.”

KS with younger workers

A seven-item measure of KS with younger workers was adapted from Lee’s (2001) measure of KS, which consists of tacit KS with younger workers (e.g., “I shared know-how from work experience with younger colleagues of my department”) and explicit KS with younger workers (e.g., “I shared business proposals and reports with younger colleagues of my department”).

Control variables

We measured workplace friendship with younger workers by adapting Nielsen et al.’s (2000) six-item scale. For example, one item was “I formed strong friendships with younger colleagues of my department at work.” A three-item measure of time pressure was adopted from Fasbender et al.’s (2021) scale (i.e., “My job required me to work fast”; “I did not have enough time to do my job”; “I did not have enough time to finish my work”). Age and organizational tenure were measured in years. Gender was coded as a dummy variable (binary coded with 0 = male and 1 = female). Education was measured using 4 options: 1 = high school or below, 2 = junior college, 3 = bachelor, 4 = master or above.

Results

Preliminary analyses

Confirmatory factor analysis was used in Mplus 8.0 to examine the discriminant validity of the seven core latent variables. The 7-factor structure displayed a good model fit (χ² (356) = 965.067, χ²/df = 2.711, RMSEA = 0.056, CFI = 0.951, TLI = 0.944, SRMR = 0.049), which was superior to alternative models (see Table 1). Hence, the discriminant validity of the present measures was supported. The Cronbach’s alpha, descriptive statistics, and correlation analysis of all variables adopted in our study were performed using SPSS 26.0. The results are shown in Table 2.

Hypothesis testing

Test of mediation effects

Our study employed structural equation modeling (SEM) with control variables in Mplus 8.0 to simultaneously test H1–H5. We used maximum likelihood (ML) estimation with bootstrapping (5,000 draws) to estimate the indirect effects (Preacher & Hayes, 2008). H1 to H3 addressed the direct relationships between PAD, PDD, GS, tacit KS, and explicit KS. The basic model provided a good model fit (χ² (372) = 1093.764, χ²/df = 2.940, RMSEA = 0.058, CFI = 0.938, TLI = 0.929, SRMR = 0.084). First, the impact of PAD on GS was positive and significant (β = 0.068, SE = 0.030, p = .026), thus supporting H1. Second, the impact of PDD on GS was negative and significant (β = –0.207, SE = 0.064, p = .001), thus supporting H2. Third, the impacts of GS on tacit KS (β = 0.301, SE = 0.055, p < .001) and explicit KS (β = 0.346, SE = 0.062, p < .001) were both significantly positive. Thus, H3a and H3b were supported. Moreover, we tested the difference between the impacts of GS on tacit KS and explicit KS. The results showed that the difference in path coefficients was not significant (difference = −0.045, 95% CI = [−0.097, 0.002]). Thus, H3c was not supported.

H4 and H5 addressed the indirect relationships between PAD, PDD, GS, tacit KS, and explicit KS. PAD had positive indirect impacts on tacit KS (indirect effect = 0.020, 95% CI = [0.004, 0.045]) and explicit KS (indirect effect = 0.023, 95% CI = [0.005, 0.050]) via GS. Hence, H4a and H4b were supported. PDD had negative indirect impacts on tacit KS (indirect effect = −0.062, 95% CI = [−0.115, −0.025]) and explicit KS (indirect effect = −0.072, 95% CI = [−0.132, −0.029]) via GS. Hence, H5a and H5b were supported.

Test of moderated mediation effects

We used SEM in Mplus 8.0 to simultaneously test H1, H2, H3, H6, H7, and H8. To test the moderated mediation effects, our study applied ML estimation and INTEGRATION command. We used the XWITH command to calculate the interaction term involving GS and KR. The results of SEM with control variables are shown in Figure 1 (without indirect effects).

H6 addressed the moderating role of KR on the impacts of GS on tacit KS and explicit KS. The results showed that GS × KR had a nonsignificant impact on tacit KS (β = −0.055, SE = 0.038, p = .152). Thus, H6a was not supported. Furthermore, H7a and H8a were not supported. Moreover, our study found that GS × KR had a negative and significant impact on explicit KS (β = −0.093, SE = 0.040, p = .021). Through a slope difference test, we found that the positive impact of GS on explicit KS was stronger for participants with low levels of KR (−1SD) (simple slope = 0.405, SE = 0.068, p < .001) in comparison with participants with high levels of KR (+1SD) (simple slope = 0.233, SE = 0.072, p = .001; slope difference = −0.172, SE = 0.081, p = .033). The two-way interaction is shown in Figure 2. These findings supported H6b.

Finally, H7b and H8b addressed the moderating role of KR on the indirect impacts of PAD and PDD on explicit KS via GS. When KR was lower, the indirect impact of PAD on explicit KS via GS was significant (indirect effect = 0.029, 95% CI = [0.003, 0.056]). In contrast, when KR was higher, the indirect impact was nonsignificant (indirect effect = 0.016, 95% CI = [−0.001, 0.033]; difference = −0.013, 95% CI = [−0.029, 0.003]). Hence, H7b was not supported. Moreover, when KR was lower, the indirect impact of PDD on explicit KS via GS was significant (indirect effect = −0.088, 95% CI = [−0.138, −0.038]). In contrast, when KR was higher, the indirect impact was significant (indirect effect = −0.049, 95% CI = [−0.086, −0.012]; difference = 0.039, 95% CI = [0.001, 0.077]). Thus, H8b was supported.

Supplementary analysis

First, according to Spector and Brannick (2011), the results should be tested both with and without control variables. Hence, we tested the theoretical model without control variables. We used SEM in Mplus 8.0 to test the mediation effects and the moderated mediation effects. The impact of PAD on GS was positive and significant (β = 0.063, SE = 0.031, p = .044), thus supporting H1. The impact of PDD on GS was negative and significant (β = −0.194, SE = 0.053, p < .001), thus supporting H2. The impacts of GS on tacit KS (β = 0.373, SE = 0.046, p < .001) and explicit KS (β = 0.414, SE = 0.048, p < .001) were both significantly positive. Thus, H3a and H3b were supported. PAD had positive indirect impacts on tacit KS (indirect effect = 0.023, 95% CI = [0.001, 0.051]) and explicit KS (indirect effect = 0.026, 95% CI = [0.001, 0.056]) via GS. Hence, H4a and H4b were supported. PDD had negative indirect impacts on tacit KS (indirect effect = −0.071, 95% CI = [−0.125, −0.026]) and explicit KS (indirect effect = −0.080, 95% CI = [−0.142, −0.028]) via GS. Hence, H5a and H5b were supported. The results showed that GS×KR had nonsignificant impacts on tacit KS (β = −0.043, SE = 0.039, p = .266) and explicit KS (β = −0.081, SE = 0.041, p = .050). Thus, H6a and H6b were not supported. Furthermore, H7 and H8 were not supported. Supplemental analysis regarding control variables indicated that the mediation effects remained significant regardless of whether control variables were included. These outcomes attest to the stability of the explanatory path of GS. However, if control variables were not included, the originally significant moderation effect (H6b) and the moderated mediation effect (H8b) became insignificant. These findings underscore the importance of the control variables that were theoretically selected in the theoretical background section. To more precisely interpret the impact of GS on older workers’ KS with younger workers, as previously mentioned, it is meaningful to include the control variables selected for our study to eliminate interference from other factors.

Second, we tested the moderating role of KR on the impacts of PAD and PDD on GS. Both PAD and PDD with younger workers highlight the social identity salience of older workers. KR from younger workers may disrupt the activation process of social identity. KR may moderate the impacts of PAD and PDD on GS. The results showed that PAD × KR had a nonsignificant impact on GS (β = 0.047, SE = 0.029, p = .099). Furthermore, the results showed that PDD × KR had a nonsignificant impact on GS (β = −0.038, SE = 0.041, p = .353). Thus, KR did not moderate the impacts of PAD and PDD on GS.

Third, we tested for the interaction between PAD and PDD. Although we did not formulate a hypothesis regarding the possible interaction between PAD and PDD, in light of the existing literature, there may be an interaction effect between PAD and PDD on GS. Based on SIT and SCT, Williams et al. (2007) found that lower perceived work-style dissimilarity with teammates threatened category uniqueness. Hence, lower perceived work-style dissimilarity with teammates reinforced the negative impact of PAD with teammates on perspective taking. According to the findings of Williams et al. (2007), higher PDD with younger workers should strengthen the positive impact of PAD with younger workers on older workers’ GS. Hence, we conducted an analysis of this interaction effect. The results regarding the two-way interaction of PAD and PDD indicated a nonsignificant effect on GS (β = 0.004, SE = 0.030, p = .887).

Discussion

To address the challenges of effectively transferring knowledge from older to younger workers, our study aims to gain a deeper understanding of how PAD and PDD with younger workers influence older workers’ KS with younger workers. Building on SCT, and using a sample of 570 older workers, our study found that higher PAD with younger workers triggered older workers’ higher GS to participate in more KS with younger workers. However, when older workers perceived that they significantly differed from younger workers in terms of deep-level characteristics, it inhibited older workers’ GS and promoted reduced KS with younger workers. Furthermore, when older workers were in a work environment where they frequently sought knowledge from younger workers, the positive impact of GS on older workers’ explicit KS with younger workers was weakened.

Additionally, based on previously reported empirical evidence and theories, we selected several control variables, including age, organizational tenure, gender, education, workplace friendship with younger workers, and time pressure. The results of the SEM with control variables (see Figure 1) revealed some special relationships. First, according to generativity theory, women are more likely than men to exhibit higher generativity (McAdams & de St. Aubin, 1992). However, our findings indicated that men exhibited higher GS than women. One possible explanation is that our study focused on an aircraft maintenance company that emphasized complex job experience and technical skills. Male workers held positions that required more expertise. Therefore, older male workers may consider it necessary and their responsibility to pass on their valuable experience to younger workers. Second, the relationship between education and GS was consistent with our expectations. Older workers with higher education exhibited higher GS. Finally, the relationships between workplace friendship with younger workers and older workers’ tacit/explicit KS with younger workers were in line with theory. The review by Dietz and Fasbender (2022) suggested that fostering high-quality workplace friendships among individuals of different ages can enhance older workers’ engagement in KS with younger workers. Although it is merely a control variable in our study, we have confirmed through empirical research that workplace friendship between older and younger workers appears to promote older workers’ KS with younger workers. Therefore, our study further supports the viewpoint of Dietz and Fasbender (2022).

Theoretical implications

Our study provides several theoretical contributions. First, based on SCT, the dual impacts of PAD and PDD with younger workers on older workers’ GS have been confirmed. GS, as a motivation related to age, is closely associated with older workers (Kooij & Van De Voorde, 2011; Kooij et al., 2011). Unlike previous studies, we do not explain the impacts of PAD and PDD with younger workers on older workers’ GS based on the in-group/out-group phenomenon. Our study highlights the importance of considering the group prototypes when investigating the impacts of PAD and PDD with younger workers on older workers’ GS to better explain the social identity phenomenon (Hogg & Terry, 2000). We found that PAD with younger workers prompted older workers to actively recognize their social identity and the vulnerability of younger workers. However, PDD with younger workers appeared to trigger older workers to treat younger workers negatively, which hindered the older workers’ motivation to be generative. Our study further validates the theoretical perspective of SCT, specifically indicating that social categorization does not necessarily promote out-group discrimination. One aggravating condition for social categorization to promote out-group discrimination is the requirement that the in-group feels threatened by the out-group (Mummendey & Otten, 1998). PAD with younger workers is just a time scale used for social categorization and does not directly threaten older workers. However, PDD with younger workers may pose a threat to older workers. Older workers may feel that what they consider correct is being challenged. In order to maintain a positive self-concept, older workers in these circumstances may discriminate against younger workers. Our findings further support the research of Urick et al. (2017). Urick et al. (2017) discovered through qualitative research that PDD with younger workers may promote older workers to avoid interacting with them. In line with this, our empirical study has demonstrated that PDD with younger workers reduces older workers’ motivation to share knowledge with them. When older workers perceive that younger workers consistently “sing a different tune” from them in terms of values and attitudes, they are likely to rely on negative prototypical characteristics about younger workers in their minds (Urick et al., 2017). They may perceive younger workers as narcissistic and presumptuous. Older workers may perceive themselves as experienced but no longer able to be appreciated by younger workers (Cheng et al., 2008; Urick et al., 2017).

Additionally, according to the results of our study, we have offered additional empirical support for Fasbender and Gerpott’s (2022) theoretical model. The theoretical model proposed by Fasbender and Gerpott (2022) suggested that older workers’ perceived high similarity with younger workers may reverse the negative relationship between unfavorable temporal social comparison to younger workers and GS. Based on our research findings, we believe that when older workers perceive high (vs. low) deep-level dissimilarity with younger workers, the negative relationship between unfavorable temporal social comparison to younger workers and GS should be stronger. Older workers’ PDD with younger workers may trigger negative social comparisons, promoting the older workers to feel that their values and beliefs are no longer acknowledged by the younger workers. When faced with younger workers who do not recognize their values and attitudes, older workers who believe they will be surpassed by the younger workers in the future are even less likely to want to share their knowledge with them.

Second, our study provides empirical support that GS mediates the relationships between PAD/PDD with younger workers and older workers’ KS with younger workers. By simultaneously testing two different yet complementary dissimilarities, our study provides a comprehensive understanding of the underlying mechanisms that influence KS. Although there is a wealth of literature on KS (Nguyen et al., 2019) and similarity is widely recognized as an important factor affecting older workers’ KS (Burmeister & Deller, 2016), relatively fewer studies specifically examined the relationship between dissimilarity and KS in organizations. In particular, previous research ignored exploring how older workers engage in KS with younger workers to cope with PAD and PDD with younger workers.

Previous studies suggested that age similarity positively affected workers’ KS (Lazazzara & Za, 2016; Zenger & Lawrence, 1989). Our research has verified that the relationship between PAD with younger workers and older workers’ KS with younger workers can also be positive by testing the mediating role of GS. Based on the prototypical characteristics of the in-group (e.g., experienced) and the out-group (e.g., inexperienced), PAD with younger workers enhances older workers’ KS with younger workers by strengthening GS. Conversely, PDD with younger workers inhibits older workers’ KS with younger workers by weakening GS. PDD with younger workers appeared to prompt older workers to focus on the negative prototypical characteristics of younger workers, such as arrogance, narcissism, and thoughtlessness. Considering the perspectives on group prototypes in SCT, our findings validate that PAD and PDD with younger workers have double-edged sword impacts on older workers’ KS with younger workers via GS. Our study addresses the call made by Riordan (2000) and Lawrence (1997) to consider intervening processes when examining the relationships between demographic dissimilarity and outcomes.

Third, by testing the moderating role of KR from younger workers, our study contributes to the literature on generativity. Little is known about the contextual conditions that may moderate the impacts of generativity (Doerwald et al., 2021). Our research addresses the call made by Doerwald et al. (2021) to provide a more detailed explanation of the boundary conditions of generativity. Specifically, we aim to explore the circumstance in which generativity is more or less beneficial. GS can be conceptualized as resulting from a commitment to age-related norms and goals. Experienced older workers are expected to provide guidance and support to younger workers who have less experience (Haslam et al., 2000; McAdams & de St. Aubin, 1992; McAdams et al., 1993). By identifying KR from younger workers as a moderator, we demonstrate that seeking knowledge from younger workers appears to undermine the older workers’ internalization of and compliance with group norms (Chaudhuri & Ghosh, 2012). This understanding helps us identify the context in which generativity motivates older workers to share less or more knowledge with younger workers. Significantly, we found that the moderating role of KR from younger workers acted only on the relationship between GS and explicit KS with younger workers, but the relationship between GS and tacit KS with younger workers was not moderated. This observation may be because older workers typically seek more explicit knowledge from younger workers, such as new information technologies, new ideas, and methods (Gerpott et al., 2017). When older workers actively seek knowledge from younger workers, this is more likely to challenge the traditional norms of explicit KS. This dynamic shift can be seen as a departure from the established hierarchy of explicit KS. Older workers possess higher levels of crystallized intelligence and are adept at utilizing their accumulated experience to identify issues within their field and anticipate potential problems from a more comprehensive perspective (Li et al., 2021). When searching for knowledge from younger workers, older workers may still have an advantage with their tacit knowledge. Therefore, the norm that older workers are expected to share tacit knowledge with younger workers should not be undermined. Although Fasbender et al. (2021) demonstrated the positive impact of GS on KS, they overlooked the specific type of knowledge being shared. Our study presents new insights into how GS and KR from younger workers work together to influence two types of KS differently. Furthermore, our findings suggest that to minimize the negative indirect impact of PDD with younger workers on older workers’ explicit KS with younger workers, KR from younger workers is essential, as both GS and KR from younger workers interact to influence older workers’ explicit KS with younger workers. In addition, previous research found that there was a difference in the impacts of intrinsic motivation on tacit KS and explicit KS (e.g., Hau et al., 2013). Although GS is an intrinsic motivation for older workers to share knowledge with younger workers, we did not find a significant difference between the two paths through which GS influenced tacit KS and explicit KS. A possible explanation for this is that our study was conducted in an aircraft maintenance company that prioritized normative and process-oriented explicit knowledge.

Practical implications

This study has several important practical implications for organizations and managers. First, it is crucial to recognize and acknowledge the expertise and contributions of older workers. Providing opportunities for older workers to mentor and share their valuable knowledge with younger workers can help maintain their sense of worth and motivation to engage in KS with younger workers. Organizations can help older workers embrace positive age-related prototypical characteristics through HR practices, such as mentoring projects and succession planning (Harvey, 2012). Second, organizations need to focus on the contradictions between younger and older workers in terms of personality, attitudes, work styles, and values. Managers can arrange for older workers to collaborate with younger workers who share similar deep-level characteristics (Roberson, 2019). Organizations can convey the concept of ideological diversification to older workers through formal training and by acknowledging and appreciating their values and attitudes. Finally, our study found that the direct positive impact of GS on older workers’ explicit KS with younger workers was weakened when older workers sought high levels of knowledge from younger workers. By emphasizing the distinct strengths of explicit knowledge in younger and older workers during daily team activities, managers can help older workers perceive that their explicit knowledge complements that of younger workers (Li et al., 2021). This recognition reinforces the value of older workers’ contributions and motivates them to continue sharing their explicit knowledge with younger workers.

Limitations and future research

Our study has some limitations. First, to minimize common method biases, our study collected data at three different time points (Podsakoff et al., 2012). However, it is important to note that the data relied on participants’ online self-reports. To better mitigate common method biases, future studies should consider collecting data from multiple sources. In addition to the relatively subjective way of obtaining data through questionnaires, KS can also be collected through objective data (Fasbender & Gerpott, 2021). Moreover, as mentioned by Shemla et al. (2016), judgments of perceived similarity and perceived dissimilarity are not necessarily complementary. Future research should consider alternative methods for measuring PDD instead of using similarity measures and subsequent reverse coding.

Second, contrary to the findings of previous empirical research (e.g., Ku, 2019; Zenger & Lawrence, 1989), our study found that PAD with younger workers facilitated older workers’ KS with younger workers via GS. This impact may be because the older workers’ referent group was the younger workers in the department. The emphasis on the inheritance of experience and knowledge within the department enhanced the positive group prototypical characteristics of older workers. The results may differ if the target is replaced with a broader range of referent groups, such as younger workers in other departments. In the future, the referent group for older workers can be set against those of younger workers outside the department, and further exploration of the relationship between PAD with younger workers and older workers’ KS with younger workers should be conducted.

Third, Dietz and Fasbender (2022) called for research on the consequences of workplace friendship that spans different age groups. Although we did not investigate workplace friendship with younger workers as a primary variable in our study, we found a significant positive impact of workplace friendship with younger workers on older workers’ KS with younger workers. Older and younger workers can develop informal interpersonal relationships, such as friendships, in the workplace (Dietz & Fasbender, 2022). Future research can explore how workplace friendship between older and younger workers affects knowledge exchange. Unlike relationships that are mandated by formal work roles (e.g., relational coordination), workplace friendship is voluntary and informal (Zhang et al., 2022), making it more susceptible to shocks. Future research may provide additional insights into the potential relationships mentioned above.

Finally, older workers are increasingly playing the role of knowledge receivers. The role expectations associated with age may affect KS and KR among younger and older workers. Older workers are often associated with the role of knowledge sender, while younger workers are often associated with the role of knowledge receiver (Burmeister et al., 2018). However, this one-way perspective on knowledge transfer has been criticized. Fasbender et al. (2021) demonstrated that both older and younger workers could serve as knowledge senders and receivers. Nonetheless, the implicit age norms (e.g., the expectation that younger workers should listen to older workers’ opinions and guidance) observed in organizations are still likely to influence younger workers’ perception regarding whether sharing knowledge with older workers is encouraged (Gerpott & Fasbender, 2020). In the future, it may be possible to explore how PAD with older workers affects younger workers’ KS with older workers.

Conclusion

In the context of an aging workforce, where older workers need to work alongside younger workers, it is important to understand how older workers respond to PAD and PDD with younger workers. In studies on relational demography, it is generally found that age similarity promotes positive outcomes. Our study supports the opinion that age dissimilarity between older and younger workers can also promote positive outcomes. This finding helps dispel the traditional stereotype that age dissimilarity promotes negative outcomes. In addition, this study indicates that differences in personality, attitudes, and values between younger and older workers hinder the motivation and behavior of older workers in providing support and guidance to younger workers. This study represents an effort to bridge two crucial areas of research within the organizational context: Research on relational demography and research on knowledge exchange between different age groups.

Author contributions

Yunyan Lu (Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Writing—original draft, Writing—review & editing) and Hao Zhou (Conceptualization, Funding acquisition, Investigation, Project administration, Resources, Supervision, Writing—review & editing)

Funding

The study received financial support from the National Natural Science Foundation of China (71872119) and the Sichuan Science and Technology Program (2023NSFSC1005).

Conflicts of interest: None declared.

Data availability

The data underlying this article will be shared on reasonable request to the corresponding author.

References