Age-inclusive HR practices and older workers' voice behavior: the role of job crafting toward strengths and negative age-based metastereotypes Free

Due to continuously falling fertility rates, and the improvement of medical technology, as well as increasing life expectancy and retirement ages, the proportion of older workers in the labor force continues to increase (Kooij et al., 2020). Such as, relevant data show that by 2030, nearly half of China's working-age population will be those over 45 years old (Peng et al., 2022). Consequently, this demographic change has sparked increased academic research on how to motivate aging workers and use their value (Kooij et al., 2020; Peng et al., 2022). Especially, given that older workers tend to have a large reserve of knowledge and experience (Kanfer and Ackerman, 2004), as well as more familiar with the organization (Zhao et al., 2023), it is easier for them to find out the problems or deficiencies of the organization, thus the opinions they put forward tend to be more scientific and practical. From this, it can be seen that promoting the voice behavior of older employees is an important way to effectively utilize their value and promote the success of the enterprise (Morrison, 2011). However, due to the potential risks of voice behavior and coupled with the fact that older employees face more complex external and internal situations compared with young employees (Zacher, 2015), the willingness of older employees to participate in voice behavior may be undermined. More specifically, in the external situations, the age-related discrimination, such as “35 years old crisis” and “40 and 50 years old to be unemployed”, occur frequently in the workplace, which seriously inhibit older workers’ emotional commitment to the organization and subsequently reduce their willingness to contribute to the organization (Ye et al., 2023). What is more, as a result of age discrimination, older workers are more sensitive to risk and seek more job stability (Ye et al., 2023) for fear of that if their suggestions touched on the interests of the leaders or co-workers, they would probably meet with resistance from other people or be ostracized, or even face the problem of being fired or replaced (Turner et al., 2020). With these in mind, older workers may assess voice as high-risk behavior and thus be reluctant to participate. Additionally, in terms of internal cognition, compared to young workers, older people perceive time as limited-holding a “time until death” perspective-and will, therefore, be especially motivated by achieving short-term goals, such as maintaining one's existing status rather than taking risk behavior (e.g. voice behavior) to achieve long-term goals (Wang et al., 2022). From the above description, we can see that strengthening the specialized research on the voice of older employees is the key to improving management effectiveness (Zacher, 2015). But unfortunately, this important question still has not yet been solved, that is, organizations how to effectively motivate older employees to voice?

Signaling theory suggests that individuals' voice behavior will be influenced by the signals from the organization's human resource management (Nyfoudi et al., 2024). Similarly, the workplace aging literature also suggests that older employees are more likely to engage in proactive behavior and reach optimal levels of performance at work when they receive supportive signals from the organization's HR practices (Kooij et al., 2020; Fasbender and Gerpott, 2022). Thus, combining literature of both fields, we speculate HR practices may be an answer to the above question. In this respect, Age-Inclusive Human Resource Practices (AIHRP), which focuses on age equality, has been proved to be an effective way to manage and motivate older employees in the face of two major problems of how to manage organizational age diversity and age discrimination that accompanies aging, such as increasing older workers' work engagement and occupational future time perspective (Fan et al., 2023; Boehm et al., 2014; Fasbender and Gerpott, 2022). In particular, evidence shown that AIHRP has a positive impact on the value recognition and identity security of older employees (Oliveira, 2021a; Oliveira and Cabral-Cardoso, 2018b), which may satisfy the two basic considerations of employees to engage in voice behavior: effective and safety (Morrison, 2011). For example, AIHRP's activities, such as encouraging employees to speak up and valuing the contribution of all age groups, can send important signals to older workers that voice behavior in the organization is safe and effective, which implies that AIHRP may have an effect on voice, but there is still a gap in empirical research on the relationship between them. Thus, overall, the present study intends to answer the questions that why, how and when does AIHRP affect older workers' voice behavior.

Further, how does AIHRP affect older workers' voice behavior? In this regard, signaling theory proposes that individuals will exert their subjective initiative to interpret signal after receiving it (Connelly et al., 2011), which means workers, as proactive players rather than as passive recipients of HR practices, can translate perceived HR practices into behavioral/attitudinal outcomes through individual initiative, such as job crafting (Guan and Frenkel, 2018; Meijerink et al., 2020; Song et al., 2024). Especially, based on the literature in personality development over the lifespan, researchers found that among the ways of job crafting, job crafting toward strengths (JCS) has significant applicability for older workers (Kooij et al., 2017), because over the lifespan, individuals will learn more about their own strengths and weaknesses (Bosma and Kunnen, 2001) and become more able and motivated to play to these with aging (Helson et al., 1995). Similarly, the empirical research also demonstrated above views by showing that compared to younger employees, older workers are more able and motivated to craft their job in line with their strengths and JCS will fit the goals of older workers in particular and will therefore be beneficial for them (Kooij et al., 2017; Zhao et al., 2023). Thus, combined with the HR practices literature (Meijerink et al., 2020) and empirical evidence on job crafting playing a role in facilitating workers' voice behavior (Rofcanin et al., 2019), we believe that when older workers receive positive signals from the AIHRP that they are valued and their future is still full of possibilities (Xu and Wang, 2023), they are more likely to perform JCS to express themselves better, which in turn, inspires more confidence and sense of control to engage in voice behavior. Accordingly, we intend to test the mediating role of JCS to address the question how AIHRP influence older workers' voice behavior.

In addition to studying why and how AIHRP motivate older employees to voice, it is also necessary to study which conditions affect AIHRP's effectiveness, because research has shown that due to individuals apply different schemas in perceiving and interpreting HR-related information, there is always the possibility that employees will not react to HR practices similarly (Bos-Nehle and Veenendaal, 2019). Within the signaling theory, the characteristics of the receiver have also been conceptualized as important factors that affect the signaling effectiveness (Connelly et al., 2011). Among older workers' characteristics, individuals' metastereotypes have been found to affect their mindset and influence the way they interpret the workplace conditions (Shiu et al., 2015; Oliveira and Cabral-Cardoso, 2018a). As a type of metastereotypes, negative age-based metastereotypes (NABM) which refer to individual negative beliefs concerning stereotypes other age groups hold about one's ingroup, provide a potential explanatory mechanism for behavior (Finkelstein et al., 2013; Oliveira, 2023). Indeed, recent research has found that NABM would interact with HR practices in predicting employee outcomes (Oliveira and Cabral-Cardoso, 2018b). Thus, from the signaling theory and previous studies about NABM, we suggest that for the signals sent by AIHRP, employees with different levels of NABM will also react and interpret the same AIHRP differently. Such as, older workers with high NABM, are prone to self-doubt and lack of confidence in their own abilities (Weiss and Perry, 2020; Owuamalam and Zagefka, 2011), and thus, they tend to interpret AIHRP as useless practices and ignore the signals it sends, ultimately lowering the evaluation of AIHRP. Therefore, our research furtherly aims to answer the question of when AIHRP effect voice and JCS by unraveling the moderating role of NABM.

Taken together, our study aims to make several contributions to the literature. First, by taking the lead in linking AIHRP and older employees' voice behavior, on the one hand, we broaden the outcome research of AIHRP and prove that AIHRP is an effective way to use older workers' value. On the other hand, it also effectively answers the question of organizations how to motivate older employees to voice and provide more complete explanations about the phenomenon of older workers' voice behavior. Second, in line with the view of employees as proactive players of HR practices (Guan and Frenkel, 2018), we examined JCS as a key mediator, which not only contribute to signaling theory by answering the call for more research on the role of works' proactivity in translating effectiveness of HR practices (Meijerink et al., 2020), but also address the question of how AIHRP influence older workers' voice behavior. Besides, despite HR practices increasing mention in the job crafting field (Guan and Frenkel, 2018), JCS's role has been examined rarely as a mediator between HR practices and work-related outcomes. In this respect, by shedding further light on the important antecedents and consequences of JCS in the workplace, this study also bridges the connection gap between HR practices and the JCS literature and furtherly constructing a more comprehensive view of the nomological network of JCS. Third, this study verified the moderating effect of NABM, which offers insights into the AIHRP and NABM literature that NABM is important to understand the boundary conditions when AIHRP facilitates older workers' voice behavior, thus responding to researchers' call for identifying individual differences' effects on the AIHRP's effectiveness (Xu and Wang, 2023; Fasbender and Gerpott, 2022). We present our hypothesized research model in Figure 1.

Voice behavior, which is defined as proactively challenging the status quo and making constructive suggestions (Van Dyne and LePine, 1998). Due to the possibility of challenging an authority's status, engaging in voice behavior may cause potential risky (Weiss and Zacher, 2022). Therefore, it is understandable that in general, employees will judge whether voice is likely to be effective and safe based on the signals sent by their organization (Morrison, 2011). According to this, research investigating the role of HR practices in voicing, as a crucial part of organization's signaling, has gradually gained traction (Nyfoudi et al., 2024). Along this line and based on the importance of age-related practices for older employees (Zacher, 2015), we focus on AIHRP in this study, which refer to the practices aiming at provide equal opportunities for employees of all age groups, including three domains: age-neutral recruiting activities as well as equal access to training and further education for all age groups; equal opportunities to be promoted, transferred and to make further career steps irrespective of one's age; the promotion of an age-friendly organizational culture that promotes and values the contribution of all age groups (Boehm et al., 2014). Such inclusive practices have been proved to be effective in stimulating initiative and positive states in older employees, particularly has a positive impact on the value recognition and identity security of older employees, which may satisfy the above two basic considerations of employees to engage in voice behavior: effective and safety (Morrison, 2011). Accordingly, based on this logic and signaling theory, it is possible that AIHRP can impact older workers' voice behavior.

Signaling theory proposes that employees usually regard the signals sent by HR practices as organization's attitude and then guide their behaviors according to these signals (Song et al., 2024). From the views of this, we predict that the signals sent by AIHRP contained the information about safety and effectiveness of voice behavior, which in turn encourage older workers to be more willing to perform it. Regarding effectiveness, through valuing the contribution and input from all employees regardless of age, AIHRP signal to older workers that organizations recognize their wisdom and consider them to be valuable (Li et al., 2021; Boehm et al., 2014), which is beneficial for boosting their confidence in providing quality and valid advice, and increasing the belief that organization will be receptive to their ideas (Soll and Larrick, 2009). Besides, the function of AIHRP equaling access to training and promotion for all age groups signal to older workers that they are wanted in the workplace and organizations still have a lot of expectation in their work ability (Oliveira, 2021a; Li et al., 2021). In this context, older workers tend to believe their opinions or views of effectively solving organizational problems will be considered as well as adopted (Soll and Larrick, 2009). What's more, equal training and transferring also give older workers opportunity to learn new knowledge, enhance work ability and broaden horizons (Boehm et al., 2014; Rudolph and Zacher, 2021), which also furtherly increase beliefs in their ability to offer insightful opinions (Liu et al., 2017). Regarding safety, AIHRP's promotion of an age-friendly organizational culture signal to the old workers that they can speak up and bring in their own ideas to the organization freely (Boehm et al., 2014). From this, older workers believe that even if they raise a different opinion, they will not be attacked and retaliated against by their leaders, which reduce their risk expectations, thereby tending to interpret voicing as safety. Moreover, because AIHRP are founded on recognition and respect for all age groups workers, these practices provide identity safe for older workers and make them feel the organization is trustworthy (Oliveira and Cabral-Cardoso, 2018b; Oliveira, 2021a; Li et al., 2021), which, in turn, may allay the worry caused by voicing. Following the aforementioned rationale, a hypothesis is formulated:

Signaling theory proposes that individuals will exert their subjective initiative to interpret signal after receiving it (Connelly et al., 2011), which suggests employees will respond to signals from HR practices with proactive behaviors such as job crafting (Meijerink et al., 2020; Song et al., 2024). As a typical job crafting behavior, JCS embodies the self-initiated efforts that individuals make in their jobs to make better use of their strengths, and usually influenced by organization's activities (Kooij et al., 2017). Following this logic and previous study on HR practices as an important factor influencing job crafting (Guan and Frenkel, 2018), we contend that older workers will perform JCS according to the signals sent by AIHRP. Firstly, AIHRP send a positive signal that all age groups are valuable and useful (Xu and Wang, 2023), which help encourage older workers to proactivity explore their strengths and express full potential in the workplace (Oliveira, 2021a). Secondly, AIHRP include equal opportunities to be promoted, trained and to make further career steps irrespective of one's age. Such practices send out signals to older workers that the organization is concerned with their successful aging (Oliveira, 2021a), and that their future is still full of endless possibilities and opportunities (Xu and Wang, 2023; Fasbender and Gerpott, 2022). As such, older workers tend to work harder (Xu and Wang, 2023) and be more proactive in finding opportunities to use their strengths at work to better express themselves in the organization when they receive these inspiring signals. Thirdly, AIHRP help to foster mutual support among employees and promote the formation of good colleague relationships (Boehm et al., 2014). Under such workplace, older workers are more willing to try to craft the job to make it better fit their strengths as they perceive that they can receive help and cooperation from colleagues (Fasbender and Gerpott, 2022). Finally, bottom-up job redesign may be accompanied by errors and risks (Kooij et al., 2017). Whereas, when older workers perceived the signals from the AIHRP that organizations take care of older workers and offer them secure long-term career development opportunities (Xu and Wang, 2023; Boehm et al., 2014; Rudolph and Zacher, 2021), they tend to be more proactive in taking full advantage of organization's emotional investment and support (Xu and Wang, 2023; Fan et al., 2023) to engage in JCS because they predict that they will not suffer organizational ostracism even if the JCS fails (Tian and Liu, 2017). Therefore, we propose that:

As theorized above, AIHRP facilitate older employees to engage in JCS. We further explain how the JCS is related to voice behavior. First, scholars clearly note that JCS contains pro-social and pro-organizational components, and can encourage employees to go the extra mile for their organization (Tian and Liu, 2017). That is, when older employees craft their work according to their own advantages, they tend to expect to bring benefits to the organization and show more extra-role behaviors such as voicing (Morrison, 2011). Indeed, evidence from empirical research has also shown that JCS help older employees make better use of their abilities (Kooij et al., 2017), which are known to be correlated with employees increased extra-role performance (Hai and Park, 2024). Second, voice behavior usually requires employees to bear risk and spend resources for it (Morrison, 2011). In this regard, JCS could create abundant resources that stimulate employees to voluntarily do more than that is required (Zhang et al., 2021; Tian and Liu, 2017). To illustrate, employees may make full use of their strengths and perform at their personal best by engaging in JCS (Zhang et al., 2021). In this process, employees will not only experience positive work significance, but also generate positive psychological experience such as a sense of self-efficacy and competence (Kuijpers et al., 2020; Zhang et al., 2021), ultimately giving them enough motivation and confidence to voice for the organization's development (Dedahanov et al., 2019). Third, through JCS, older employees are likely to have a broader conception of their role and strengths, facilitating their fit to the job (Kooij et al., 2017) and subsequently participating in more voice behavior (Kim et al., 2020). Additionally, although the effect of JCS on voice behavior has not been directly examined, the relationship has been indirectly supported. For example, Rofcanin et al. (2019) found that relational crafting has an effect on voice behavior. Thus, we propose that:

In summary, according to signaling theory, AIHRP signal to older workers that they are valuable and their future is full of endless possibilities and opportunities by implementing a serious of age equality measures. Such inspiring signals effectively motivate older workers to proactively seek opportunities to use their strengths to better express themselves at work. Furtherly, under the perceived support from AIHRP, older workers are likely to engage in more extra-role behavior after performing JCS (Tian and Liu, 2017). And the process of JCS also enhance individual perception of personal control over their work (Zhang et al., 2021), which will make older workers dare to express innermost thoughts. Therefore, based on Hypothesis 2 and 3, we expect that:

Noticeably, despite the proposed routes and the corresponding theoretical justification, the effects may not be supposed to be absolute or universal. Signaling theory notes that signaling effectiveness is determined in part by the characteristics of the receiver (Connelly et al., 2011). In other words, employees with different characteristics will interpret the HR practices' signals differently, and thus the function of practices will be influenced (Bos-Nehle and Veenendaal, 2019). As a typical cognitive characteristic of individuals, NABM has been conceptualized as an important factor influencing older workers' thoughts and feelings to organizational environment (Shiu et al., 2015; Oliveira and Cabral-Cardoso, 2018b; Von Hippel et al., 2019). Thus, we can see, despite AIHRP can send positive signals to older workers, NABM may affect the interpretation of these signals in the process, thereby moderating the relationship between AIHRP and voice behavior.

Firstly, a large number of literature have found that NABM will trigger older workers' concerns that target both age group reputation and older workers' self-image (Oliveira and Cabral-Cardoso, 2018a). Such negative belief may weaken the signals of organization recognition and age inclusiveness that employees perceive from AIHRP, and then prevent them from taking the initiative to take risks at work. Consequently, older workers with high NABM generally tend to keep silent and refrain from suggesting ideas to the organization to avoid possible bad results for the challenges and risks of voice behavior (Li and Sun, 2015). Secondly, previous studies have demonstrated that employees' NABM is usually couple with self-doubt and lacking of confidence in their abilities (Weiss and Perry, 2020; Owuamalam and Zagefka, 2011). This implies that older workers with high NABM are less likely to believe that they can propose insightful recommendations by taking advantage of the horizons-expanding and skill-enhancing opportunities provided by AIHRP, thus undermining the effectiveness of AIHRP on voice. Thirdly, since older workers with higher NABM are likely to psychologically distance from their organization and always skeptical of the organization's values or ideologies (Oliveira and Cabral-Cardoso, 2018a). As such, they may doubt the authenticity of signals sent by AIHRP and argue that the AIHRP does not sincerely care about older workers in nature, which undermine the appraisal about voice behavior's safety and effectiveness, thus weakening the possibility to perform it. In contrast, older workers with lower NABM may perceive that other age groups regard them as experienced and knowledgeable (Finkelstein et al., 2020). Therefore, the external signals sent by AIHRP that organizations value and recognize older workers' value may be consistent with their internal cognitive evaluation, thus they will feel the organization's practices are trustworthy (Li et al., 2021). As such, they are more likely to interpret AIHRP as a chance of showing ability and wisdom, and then engage in more voice behavior. Thus, we propose that:

According to previous studies, the influence of HR practices on job crafting largely depends on individual factors (Song et al., 2024). From the view of this, we predict that the relationship between AIHRP and JCS will also be moderated by older workers' NABM. Specifically, older workers with higher NABM tend to believe that younger employees regard them as negative group and hold a negative age identity for themselves (Oliveira, 2023), which lead to these older workers generally feeling that they are unvalued or incompetent organizational members (Oliveira and Cabral-Cardoso, 2018a). Therefore, they are more likely to interpret the AIHRP's signal of age equality and value recognition as attaching importance to other age groups, ultimately ignoring the significance of AIHRP to their own work, reducing the possibility for them to find and use their strengths in the workplace. Further, previous research has shown that older workers with higher NABM see their future opportunities and time at work as very limited (Bal et al., 2015). This may not only restrict older workers desire for self-enhancement, make them uninterested in AIHRP's development activities (Oliveira and Cabral-Cardoso, 2018a), such as promotion, but it may also make them not pay attention to the signals sent by the AIHRP that their future is still full of endless possibilities, which in turn reduce their motivation to engage in JCS to express themselves better. Additionally, older workers with higher NABM perceived low support from younger groups (Oliveira, 2021b), which results in they diminishing their interaction and cooperation with other age groups colleagues, but this weakened the channels and quality of information acquisition (Oliveira and Cabral-Cardoso, 2018a), which may ultimately lead to none of them being aware of AIHRP's support measures, such as job transferred and training opportunities, consequently losing the chance to identify their strengths. Likewise, even when they perceive support from AIHRP, they may still fail to use these opportunities to successfully craft the job in line with their strengths because of lacking essential social resources to change the work boundary (Oliveira and Cabral-Cardoso, 2018a), such as colleagues' support and cooperation. In contrast, older workers with lower NABM are more likely to experience positive age identity (Oliveira, 2021b). Thus, when they perceive the signal from AIHRP that organization believes in their future, they will evaluate AIHRP positively and trust its message of caring for aging employees to achieve successful aging at work, thus boosting the effectiveness of the AIHRP, and be more willing to perform JCS. Moreover, older workers with lower NABM prefer to interact with other people (Oliveira, 2021b), which not only helps them to obtain organizational information and understand the signals sent by AIHRP, but also provides external support for the successful combination of strengths and work. Combined with the above derivation, although the positive signal sent by AIHRP is conducive to promoting the JCS of older workers, higher NABM of individuals may tend to interpret AIHRP as useless practices and ignore the signals sent by the AIHRP. This hinders employees' job crafting and thus weakens the positive influence of AIHRP on voice behavior transmitted through JCS. Considering this, we propose that:

Data were collected from different organization in China. We obtained agreement and support from the management team of their company via the MBA/EMBA/EDP students in the author team's university. After we were granted permission, the HR department of each organization assisted with the data collection. One staff member in the HR department helped us sort out a list providing the phone number contacts of the company who were age 45–65. We invited all eligible employees to take part in the research project, and a total of 380 older workers agreed to participate. With the assistance of the human resources department, participants were gathered in conference rooms to complete the questionnaire. When collecting the data, we first introduced the study procedures and assured the participants of data confidentiality. Besides, we assured respondents that there were no right or wrong answers to survey items to decrease evaluation apprehension. Then, we distributed hard-copy surveys, which had been prepared in advance by the researchers. And, we told the participants that in order to match the three-wave date they need to write down the last four digits of their phone number in the last questionnaire. All questionnaires were completed and collected on site. All participation was voluntary, and participants were allowed to quit the research at any time.

To reduce common method variance (CMV) (Podsakoff et al., 2012), we measured the independent variable, the mediators, the moderator and the outcome variable at three time points with a two-week interval. Following best practice, we included three attention checks in each wave of the survey (worded as, for example, “this is an attention checker, please select ‘totally agree’”). During Time 1 (T1), participants reported AIHRP (independent variable) and NABM (moderator) and provided basic demographic (age, gender, education level and tenure). Among the 380 participants who completed the survey, 25 were removed because they failed to pass at least one attention checker. The remaining 355 were invited to participate in the Time 2 (T2) survey, in which they were asked to report JCS (mediators). Among the 355 responses received (T2), we removed 12 participants who failed attention checks, three participants who did not write down the last four digits of their phone number. The remaining 340 participants were invited to participate in Time 3 (T3), in which they completed measures of voice behavior (outcome variable). Among 340 responses, consistent with previous procedures, 321 valid responses were finally obtained after removing 19 participants failed the attention check items. The final matched sample consisted of 321 older workers. About 40.19% of employees were men, 9.03% held Middle school or below, 14.33% held High school, 27.11% held College; 49.53% held Bachelor or above. The average age was 50.03, average organizational tenure was 17.21 years.

Unless stated otherwise, participants were asked to indicate their level of agreement on a five-point Likert scale ranging from 1 (totally disagree) to 5 (totally agree) with statements regarding:

Age-inclusive HR practices. Following Boehm et al. (2014), five items with a five-point scale (1 = very low intensity; 5 = very high intensity) were used to measure workers' perceived AIHRP. A sample item is “With how much intensity does your organization offer equal access to training and further education for all age groups?” Cronbach's alpha was 0.889.

Voice behavior. We measured older workers voice behavior using Van Dyne and LePine's (1998) six-item voice scale. A sample item is “I communicate my opinions about work issues to others even if my opinion is different and others disagree with me.” Cronbach's alpha was 0.899.

Job crafting toward strengths. We used a four-item scale developed by Kooij et al. (2017) to measure JCS. A sample item is “I organized my work in such a way that it matches my strengths.” Cronbach's alpha was 0.891.

Negative age-based metastereotypes. Four items were adapted from Finkelstein et al. (2020) to measure NABM held by older workers regarding younger workers beliefs about them. A sample item is “I think members of our age group were viewed out of touch by members of young colleagues.” Cronbach's alpha was 0.920.

Previous research has indicated that demographic attributes may affect employees' voice behaviors (Detert and Burris, 2007). Consequently, we controlled for several variables – specifically, gender, age, education and organizational tenure.

All analyses were conducted using AMOS 24.0 and SPSS 26.0 Statistics. Specifically, the hypotheses were tested with the Hayes macro PROCESS v4.1 for SPSS Statistics. We tested the hypotheses by using bootstrapping (5,000 times) and reported the confidence intervals (95%): Lower Limits (LL) and Upper Limits (UL). If the confidence intervals (95%) did not include zero, the hypothesis was supported.

Before testing our hypotheses, we conducted a series of confirmatory factor analyses (CFA) using AMOS to examine the discriminant validity of the study variables.

The results indicated that the hypothesized four-factor model with AIHRP, JCS, NABM and VB fitted the data better (χ² = 321.044, df = 146, CFI = 0.957, TLI = 0.949, RMSEA = 0.061, SRMR = 0.0461) than alternative models (Table 1). These results provided support for our hypothesized measurement model.

Although our data were collected from multiple waves, the variables were collected from the same source. Thus, our data may still be subject to common method biases. To examine whether CMV was likely to have biased our results, we tested the extent of CMV by refitting the model with an additional latent variable (i.e. method factor) representing a single-source bias (Schermuly and Meyer, 2016). Specifically, we first loaded all observed variables in the model onto their respective latent variable and the latent method factors. Then, following Schermuly and Meyer's (2016) and Khan et al.'s (2021) approach, we fixed all unstandardized factor loadings associated with this method factor to 1 and made it uncorrelated with other latent variables. The resulting model with a method factor fitted to the data well (χ² = 318.646, df = 145, CFI = 0.957, TLI = 0.949, RMSEA = 0.061, SRMR = 0.0464). However, it did not fit to the data better than the previous model, Δχ²(1) = 2.398, p = 0.121. Specifically, we followed Castanheira (2016) to calculate the CFI difference between this model and the original four-factor model. The change of CFI between the two models was below the suggested Strunk and Lane's (2017) rule of thumb of 0.02. We also assessed differences in TLI, RMSEA and SRMR. The TLI difference was below the cut-off value of 0.05 recommended by Little (1997). The RMSEA difference was aligned with the recommended value of less than 0.015 (Chen, 2007). The SRMR difference was 0.0003, which was below the cut-off value of 0.03 (Chen, 2007). Given that the differences in these indexes were below recommended cut-off values, on balance, these results suggested that including the method factor in the model did not appear to improve the overall fit of the model. Therefore, we can conclude that CMV is unlikely to be a major concern in this study.

Table 2 presented means, standard deviations and bivariate correlations for all study variables. The results show that, as hypothesized, (a) the AIHRP were positively related to voice behavior (r = 0.273, p < 0.01), (b) the AIHRP were positively related to JCS (r = 0.421, p < 0.01), (c) JCS was positively related to voice behavior (r = 0.410, p < 0.01).

Results showed (Tables 3 and 4) that AIHRP was significantly positively related to voice behavior (β = 0.257, 95% CI = [0.159, 0.354]) and JCS (β = 0.429, 95% CI = [0.326, 0.531]). And JCS also was significantly positively related to voice behavior (β = 0.318, 95% CI = [0.219, 0.418]). These results support Hypothesis 1, 2 and 3. As predicted, results in Table 4 show that AIHRP had an indirect positive effect on voice behavior (β = 0.137, 95% CI = [0.078, 0.209]) through JCS. Thus, hypotheses 4 was supported. Hence, Hypotheses 1, 2, 3 and 4 were supported by our data as these confidence intervals (95%) did not include zero.

As shown in Table 5, the interaction between AIHRP and NABM was negatively associated with voice behavior (β = −0.121, p < 0.01, 95% CI = [−0.210, −0.031]). The nature of these interaction effects is shown in Figure 3. The simple slope tests further showed that the relationship between AIHRP and voice behavior was stronger at low levels of NABM (simple slope = 0.261, 95% CI = [0.120, 0.401]) than at high levels (simple slope = 0.044, 95% CI = [−0.085, 0.174]). And the difference of effect of AIHRP on voice behavior is significant at different levels of NABM (Δβ = −0.217, p < 0.05, 95% CI = [−0.402, −0.045]). Thus, Hypothesis 5 was supported. Similarly, As shown in Table 5, the interaction between AIHRP and NABM was negatively associated with JCS (β = −0.134, p < 0.01, 95% CI = [−0.230, −0.038]). The nature of these interaction effects is shown in Figure 2. The simple slope tests further showed that the relationship between AIHRP and JCS was stronger at low levels of NABM (simple slope = 0.432, 95% CI = [0.288, 0.576]) than at high levels (simple slope = 0.192, 95% CI = [0.053, 0.330]). Also, the difference of effect of AIHRP on JCS is significant at different levels of NABM (Δβ = −0.240, p < 0.05, 95% CI = [−0.392, −0.082]). Thus, Hypothesis 6 was supported. As shown in Table 6, the index of moderated mediation was significant for the hypothesized indirect relationships between AIHRP and voice behavior via the mediators: JCS (index = −0.041, SE = 0.018, 95% CI = [−0.082, −0.009]). And the conditional indirect effect of AIHRP on voice behavior via JCS was stronger at low levels of NABM (β = 0.133, 95% CI = [0.070, 0.219]) than at high levels (β = 0.059, 95% CI = [0.012, 0.120]). Also, the difference of conditional indirect effect of AIHRP on voice behavior via JCS is significant at different levels of NABM (Δβ = −0.074, 95% CI = [−0.147, −0.016]). Thus, these results supported Hypotheses 7.

First, this study takes the lead in exploring the influence of AIHRP on older workers' voice behavior, which enriches the antecedents of older workers' voice behavior and expands the research perspective of AIHRP. With the increase of older workers in the workplace, researchers have gradually realized the importance of their crystallized intelligence such work experience in giving suggestions to develop enterprises. However, there is little known about how to promote older workers' voice behavior. Thus, by exploring the relationship between AIHRP and older workers' voice behavior, this study not only effectively fills this research gap, identifies a new important factor for motivating older workers' value, but also respond to the call for more research on the role of HR practices in employees' voice (Hu and Jiang, 2018). Additionally, unlike previous studies based on social exchange theory that identify the influence of AIHRP as a process of social exchange (Oliveir, 2021a), we focus on explaining why AIHRP influence voice behavior based on signaling theory, and think AIHRP signal to employee that voice is safe and effective, rather than merely reciprocating organizational support. By doing so, this research reaffirmed the suggestion that researchers can utilize other theoretical perspectives in addition to the predominantly used social exchange perspective when theorizing the relationship between AIHRP and employee behavior (Burmeister et al., 2018), and also broadens AIHRP's outcome research.

Second, through exploring the mediating role of JCS, it not only reveals the mechanism of AIHRP affecting voice behavior, but also contributes to the JCS literature. On the one hand, previous studies mainly took employees as passive receivers to explore the mediating mechanism of AIHRP, while neglected employees as active or proactive players (Fan et al., 2023). Therefore, in line with the view of employees as active or proactive players rather than as passive recipients of HR practices (Meijerink et al., 2020), we explore the mediating role of JCS on AIHRP affecting voice behavior by drawing on signaling theory. Consequently, this finding reveals both the “black box” between AIHRP and voice behavior, and expands the transmission mechanism of AIHRP's impact utility in terms of employee initiative. On the other hand, most previous studies on the influencing factors of JCS focused on personal or job intervention (Zhang et al., 2021; Kooij et al., 2017), but paid little attention to the role of HR practices. In fact, HR practices were proposed as potential sources of older workers' self-regulation behaviors, such as job crafting (Kooij et al., 2020). Besides, despite HR practices increasing mention in the job crafting field (Guan and Frenkel, 2018), JCS's role has been examined rarely as a mediator between HR practices and work-related outcomes. Therefore, discussing the mediating role of JCS not only bridges the gap between HR practices and JCS literature, but also contribute to constructing a more comprehensive view of the nomological network of JCS by shedding further light on the important antecedents and consequences of JCS.

Third, by verifying the moderating effect of NABM in the process of AIHRP influencing voice, our understanding of the mechanism of older workers' voice behavior is deepened. Although previous studies have argued AIHRP can benefit older workers, the boundary conditions of AIHRP have not been paid enough attention (Xu and Wang, 2023). To address this gap, based on the logic of the interaction between employees' individual factors and management practices shaping individual behavior from signaling theory (Connelly et al., 2011), this research examined the moderating effect of NABM. Results show that older workers with high NABM tend to ignore the signals of trust and recognition sent by the AIHRP, then attenuate the effect of AIHRP on voice behavior, and JCS. These findings not only respond to the call for more studies exploring the boundary conditions of AIHRP (Xu and Wang, 2023), but also further illustrated that NABM may interact with AIHRP in predicting employee behaviors, which make a valuable supplement to the research framework of the relationship between AIHRP, JCS and voice, and deepen our understanding of the mechanism of voice behavior.

Three major practical implications can be derived from our findings. First, enterprise leaders should be conscious of the crucial role of AIHRP (Burmeister et al., 2018). Our results suggest that AIHRP positively affect older workers' voice behavior. When managers conduct AIHRP, they should be attentive and respectful to older workers, recognize the efforts they make in the development of the organization, and encourage them to voice. This will make them to feel that voice is an effective and safe behavior. At the same time, as the new generation employees join organizations, the integration of old and new mindsets and behaviors become a key point of organizational management. Thus, organizations should also train managers in how to effectively adapt to the age-diverse organizational environment, foster high levels of a pro-age diverse climate (Boehm et al., 2014), improve the quality of intergenerational contact and provide a supportive environment for older workers.

Second, consistent with previous findings that job crafting positively affects employees' extra behavior beyond their own job duties (Guan and Frenkel, 2018), this study's results supported the positive role of JCS in encouraging older workers to participate in voice behavior. Thus, managers should provide older workers with opportunities to craft their jobs to use their strengths and achieve a better person-job fit, which will result in a series of positive outcomes (Kooij et al., 2017). Additionally, job crafting has a certain risk of failure, thus organizations should also actively foster an organizational culture that tolerates mistakes and correctly views failure, so as to provide psychological resources for older workers to engage in JCS. Furthermore, organizations could conduct job crafting interventions to encourage older workers to engage in crafting (Kooij et al., 2017) and ensure that older workers' strengths are used optimally to achieve positive outcomes (Zhang et al., 2021).

Finally, we suggest that organizational interventions should be taken to reduce NABM effects. For example, organizations should blur intergenerational boundaries within the organization and provide older workers with mentoring opportunities to transfer their knowledge. Opportunities such as these make older workers feel useful and respected, which help them to build a positive social identity (Hennekam and Herrbach, 2015), thereby diminishing NABM. Further, interventions should strive to against age discrimination. By doing so, older workers will feel recognized and accepted by others, thus, NABM may lose its power.

Like all research, there are certain limitations to our study. Firstly, employees have a clearer understanding of their own behaviors and cognition than supervisors, therefore, all data were reported by older workers themselves in this study. Although relevant tests show that the data quality of this study is acceptable, there are inevitable problems of social desirability. Thus, self-report and others' evaluation should be combined, when possible, in future studies. Secondly, as an exploratory study, our research demonstrates the positive relationship between AIHRP and voice. Future research could continually consider other HR practices' role in activating older workers' voice, such as work-family-balanced HR practices. This is because work-family-balanced HR practices can effectively enhance employees' work-family promotion and organizational commitment (Gallie et al., 2001), boost employees' motivation to contribute to the organization. Thirdly, this study confirms that AIHRP affects older workers' voice behavior through the proactive mechanism of JCS, but this does not mean JCS is the only mediating path. Therefore, future research can explore whether there are other mediating paths from the perspectives of emotional and cognitive mechanism, to deepen our understanding of the relationship between AIHRP and older workers' voice. Finally, although we have demonstrated that the internal individual characteristics of NABM moderate the relationship between AIHRP and employee voice, it is likely that other external moderators exist. For example, an authoritarian leader emphasizes absolute authority and control, and punishes subordinates when they do not follow his or her rules (Chan et al., 2013), which transmits signals inconsistent with AIHRP. Without “safe passages” provided by leaders, employees may doubt the effectiveness of AIHRP, and believe that voice behavior is risky and high costs, thereby reducing motivation to perform it. Thus, future research could explore such potential moderator to gain a deeper understanding of the effect of AIHRP on employee voice behavior. Additionally, the control variables of this study are mainly demographic variables. Future research should control for other factors that have been found to positively relate to voice behavior, such as psychological safety (Lee et al., 2023), so as to more comprehensively understand the influence of AIHRP on older workers' voice behavior.