Since the 1980’s, human resources (HR) management has been discussed, as it then became part of the work relations’ lexis (Boxall & Purcell, Reference Boxall and Purcell2011). Armstrong (Reference Armstrong2014) defines HR as a planned method for managing an organization’s most valuable resources: the people who contribute to the company’s goals and success. Through HR management, a company can provide development opportunities and implement processes that can increase worker’s retention, commitment, and motivation (Armstrong, Reference Armstrong2014; Nishii & Wright, Reference Nishii, Wright and Smith2008). According to social exchange theory (Blau, Reference Blau1964), HR management is often the method for employer and employee communication (Bowen & Ostroff, Reference Bowen and Ostroff2004). By sending certain messages, which may be from an HR Practice (HRP) system, companies can connect with workers, emphasize their importance and persuade them to respond in the same direction and present the organizations´ values and expectations (e.g., Den Hartog, Boon, Verburg, & Croon, Reference Den Hartog, Boon, Verburg and Croon2013; Nishii & Wright, Reference Nishii, Wright and Smith2008).
Earlier research demonstrated that HRP are positively related to affective commitment (AC) and workers’ and firm’s performance (e.g., Takeuchi, Lepak, Wang, & Takeuchi, Reference Takeuchi, Lepak, Wang and Takeuchi2007; Zacharatos, Barling, & Iverson, Reference Zacharatos, Barling and Iverson2005). However, these reactions are not solely due to the presence of an HRP. Nishii and Wright (Reference Nishii, Wright and Smith2008) showed that it is possible to outline an HRP’s intended (i.e., the practice’s design), actual (i.e., the practice’s delivery and implementation), and perceived (i.e., the employee’s interpretation of the practices) practices. Thus, HRP’s intended and actual effects primarily result in employee’s perceptions (Nishii & Wright, Reference Nishii, Wright and Smith2008). A good fit between practices and individual characteristics (i.e., values, goals, and expectations) can increase the effectiveness of a specific HRP system (e.g., Den Hartog et al., Reference Den Hartog, Boon, Verburg and Croon2013).
Additionally, HR management strategies are not immune to the social context (Boxall & Purcell, Reference Boxall and Purcell2011; Guest, Reference Guest2017). Currently, HR departments are expected to operate as a “business within a business” and in structural alignment with the organization’s model (Ulrich, Younger, & Brockbank, Reference Ulrich, Younger and Brockbank2008), contributing to the company’s financial performance and establishing a stable and sustainable production, while increasing competition, versatility, and work flexibility (de Jong, Reference de Jong2014).
Numerical, functional and mental flexibility are key concepts in work relations. The worker’s perspectives combined with competitive pressure forced companies to assume more flexible work policies (de Jong, Reference de Jong2014). However, in contrast to the 80’s and 90’s, currently using contingent work may not be connected to specific increases or demands (Zimmerman, Gavrilova-Aguilar, & Cullum, Reference Zimmerman, Gavrilova-Aguilar and Cullum2013). Indeed, contingent workers perform their tasks alongside direct employees in a wide range of sectors and activities inside the market (Buch, Kuvaas, & Dysvik, Reference Buch, Kuvaas and Dysvik2010; De Cuyper, De Witte, & van Emmerik, Reference De Cuyper, De Witte and van Emmerik2011). Furthermore, statistics confirm that the numbers of contingent workers have grown since the end of the last century (Eurostat, 2019; Organization for Economic Co-operation and Development, 2018).
Labor flexibility’s growth and development also impacted the configuration of contingent work. From the existing contingent work formats, temporary staffing agencies have the most complex work arrangement (Sobral, Chambel, & Castanheira, Reference Sobral, Chambel and Castanheira2016). Temporary agency workers (TAW) are in a three-way relationship and are capable of creating ‘dual allegiances’ (Bonet, Cappelli, & Hamori, Reference Bonet, Cappelli and Hamori2013). They establish work relations with the temporary agency, with which they have a formal contract, and the client company, with whom they actually work (Sobral et al., Reference Sobral, Chambel and Castanheira2016; Sobral, Lopes, Chambel, & Castanheira, Reference Sobral, Lopes, Chambel, Castanheira and Chambel2015). Given this reality, several studies (e.g., Zimmerman et al., Reference Zimmerman, Gavrilova-Aguilar and Cullum2013) have indicated a need to adjust and adapt academic representations of labor relations to the present context.
Integrating contingent workers presents a challenge for HR management. It is important to understand how companies can effectively integrate these workers into their workforce (Jiang & Wang, Reference Jiang and Wang2018; Zimmerman et al., Reference Zimmerman, Gavrilova-Aguilar and Cullum2013) and the HRPs that can be applied to match their needs. Contingent work is often associated with uncertainty and insecurity, which, in turn, is related with negative outcomes (Höge, Sora, Weber, Peiró, & Caballer, Reference Höge, Sora, Weber, Peiró and Caballer2015). Moreover, most contingent workers accepted their positions with a goal of obtaining a permanent position inside the client company (e.g., Sobral et al., Reference Sobral, Chambel and Castanheira2016; Sobral et al., Reference Sobral, Lopes, Chambel, Castanheira and Chambel2015). Therefore, these workers do not choose to be contingent. However, previous research has shown that TAW reciprocate and value the investments that are made by their organizations, which may lead them to respond in accord (e.g., Buch et al., Reference Buch, Kuvaas and Dysvik2010; Chambel, Castanheira, & Sobral, Reference Chambel, Castanheira and Sobral2016; Chambel & Sobral, Reference Chambel and Sobral2011; Jiang & Wang, Reference Jiang and Wang2018). Voluntary or involuntary contingent work situations have an influence on the worker’s internal employment prospects (within the current employee) and the external labor market (finding a job elsewhere; De Cuyper et al., Reference De Cuyper, De Witte and van Emmerik2011). Hence, the HRP that companies implement may play a crucial role in responding to TAW’s specific needs. Like recently underlined by Guest (Reference Guest2017), the context variation can make certain HRP more relevant than others. Meaning that the national context, the organizational specificities and arrangements (i.e., sector, size or external competitive environment), and the individual characteristics have impact in the way practices are delivered by the company and received by the workers. However, according with the research team knowledge, there are no HRP measures that have been adapted or psychometrically assessed for contingent workers, and specifically TAW.
Several empirically validated HRP measures have been developed in the context of standard employment. Zacharatos and colleagues (2005) developed a 68-item scale that evaluated high-performance work practices and measured employment security, selective hiring, training, decentralized decision making, status distinctions, information sharing, contingent compensation, leadership, high-quality work, and management practices. Then, Takeuchi and colleagues (2007) developed a 21-item scale that was also for high-performance work systems and measured workers’ autonomy, the selection process, training and development opportunities, performance appraisal, and the compensation system. The HRP was described as a system in both studies.
Although both scales have been validated and recognized by the academic community, none have been tested with contingent workers.
The current study builds on Zacharatos et al. (Reference Zacharatos, Barling and Iverson2005) and Takeuchis et al.’s research (2007) to develop and psychometrically validate an HRP scale that was adjusted for the TAW population. These authors greatly inspired the scale’s item construction and measurement of the HRP system. Furthermore, because each HRP is not independent (Chambel et al., Reference Chambel, Castanheira and Sobral2016), it is examined how TAW’s perceive HRP regardless of the company providing the practices. It is believed that the agency and the client are jointly responsible for the HRP system; thus, TAW’s overall perceptions were measured. Buch et al. (Reference Buch, Kuvaas and Dysvik2010) highlighted that behaviors toward one company may depend on the benefits that are exchanged with the other and vice versa.
To obtain the final 20-item scale that measured TAW’s perceptions of the HRP system, were employed several steps. First, the research team interviewed agencies and client companies HR managers to discuss the practices that were included in Zacharatos et al. (Reference Zacharatos, Barling and Iverson2005) and Takeuchi et al.’s (Reference Takeuchi, Lepak, Wang and Takeuchi2007) scales. Second, the interviews were analyzed to understand which practices are used to manage these workers. Most HR managers agreed that there were four central practices: The recruitment process (i.e., the fit between employees and their work environment; Items 1 through 4, see Table 2), socialization (i.e., identifying with the organization’s context; Items 5 through 8, see Table 2), training and development (employability inside and outside the current company) opportunities (i.e., skill empowerment; Items 9 through 16, in which Items 9, 11, 13 and 14 refer to training that can develop external employability; and Items 10, 12, 15 and 16 refer to training that can develop external employability, see Table 2), and performance appraisal (i.e., a clear and fair evaluation method; Items 17 through 20, see Table 2). Third, items were developed to measure each practice. Items that referred to the recruitment process and performance appraisal were obtained from Zacharatos et al. (Reference Zacharatos, Barling and Iverson2005) and Takeuchi et al.’s scales (2007). However, there were several practices that the HR managers identified that were not included in those scales, specifically socialization and training/employability development. The items that measured training to promote TAW’s employability (e.g., acquiring transferable skills that can be used inside or outside the current organization) were influenced by these authors as well as Chambel and Sobral’s items (2011). The socialization items were based on Ashforth and Sak’s research (Reference Ashforth and Saks1996). These authors suggested that socialization allows newcomers to identify with the organization’s values and norms and learn the organizational culture.
Moreover, certain practices, including employment security and decentralized decision-making, were not useful in this work context. In addition, the compensation system was not transversal to all sectors and functions that were included in the final pool.
Finally, the Messick’s validation framework (1995) was used to assess the psychometrics of the scale. Specifically, this study followed steps for analyzing the response process, internal structure, and relationships to other variables. To examine the scale’s ‘relationship with other variables,’ it was tested the relationship between TAW’s perceptions of HRP and their affective commitment (AC) toward the agency and client companies. Because AC is based on an emotional liaison and identification with organizational values and goals (Chambel et al., Reference Chambel, Castanheira and Sobral2016), when companies want to achieve competitive performance, they must invest in a critical outcome like worker’s AC. Consequently, it was important to determine whether perceptions of the HRP system perception could predict TAW’s AC. In addition, because the literature already demonstrated that TAW’s create two different ACs, one with the agency and another with the client company (Buch et al., Reference Buch, Kuvaas and Dysvik2010; Sobral et al., Reference Sobral, Chambel and Castanheira2016; Sobral et al., Reference Sobral, Lopes, Chambel, Castanheira and Chambel2015), both ACs were considered in the current study.
Method
Sample
The sample consisted of TAW (N = 4,551) from several temporary employment agencies that were operating in Portugal. The sample included workers from services (i.e., commerce, hospitality, catering, and transports; 68.8%) and industry (i.e., manufacturing, heavy machinery handling, storage, and construction; 30.8%), who had different functions. These participants all worked for clients who employed TAW to adapt to current market needs. This decision enabled the organizations to adjust to fluctuations in client requests or services. The respondents completed an online questionnaire. Data were collected using SurveyMonkey, which is a commercial survey service. The agencies emailed a link to the survey to their workers. This email included a message from the research team that explained the purpose of the project, the voluntary nature of participation, and the procedure. Because participation was voluntary and anonymous, participants did not sign an informed consent form. Participant’s response rates ranged from 42% to 58%.
For demographic characteristics, women constituted 53.5% of the sample and the average age was approximately 31 years old. The distribution by age group was: 50.9% were less than 29 years old; 33.9% were between 30 and 39 years old; 11.9% were between 40 and 49 years old; and 3.3% were more than 50 years old. Participant’s academic qualifications included: 8.5% had up to 9th grade; 20.0% completed secondary school; 30.9% attended the university; 17.7% graduated; and 23% completed postgraduate studies. Almost one-quarter of the participants were students (24%). Most of the sample reported that this was their only job (91.1%) and that their previously job situation was unemployment (51.6%). Participant’s agency tenure distribution included: 22.1% had less than 3 months; 20% had between 3 and 6 months; 13% had 6 to 9 months; 11% had 9 months to a year; 30.3% had 1 to 5 years; 2.6% had 5 to 10 years; and .8% had more than 10 years. For tenure distribution in the client company, 21.4% had less than 3 months; 19% had 3 to 6 months; 13% had 6 to 9 months; 11% had 9 to 13 months; 8.6% had 13 to 18 months; and 26.9% had more than 18 months.
Measures
Perceptions of HRP. Perceptions of the HRP system was measured using a 20 item scale that measured TAW’s socialization, recruitment, training (i.e., promotion of internal and external employability), and performance appraisal (see Table 2). The items were answered on 7-point Likert-type rating scale that ranged from I strongly disagree (1) to I strongly agree (7).
AC. TAW’s AC toward the agency and the client company was measured using a Portuguese version of Meyer, Allen, & Smith’s scale (1993). Example items included, “I would be happy if I developed the rest of my career in this company,” and “I do not feel emotionally attached to this company” (reverse scored). The items were answered on 7-point Likert-type rating scale that ranged from I strongly disagree (1) to I strongly agree (7). This scale had already been translated, adapted and tested with Portuguese TAW (Chambel et al., Reference Chambel, Castanheira and Sobral2016; Chambel & Sobral, Reference Chambel and Sobral2011).
Statistical procedures
This study followed Messick’s data analysis steps (1995) for assessing the validity of a construct. First, in line with the author’s conceptualization, a correct response process was used to confirm the data integrity. Thus, the response process was designed to eliminate or at least maximally control for all potential sources of error that were connected to administering the instrument (Downing, Reference Downing2003). The data collection and analysis process used clear instructions and disclosed the research aims, team, and technical procedures’ to increase quality assurance.
Second, it was necessary to ensure the instrument’s psychometric characteristics (i.e., the internal structure). This step examined the descriptive statistics, item characteristics (difficulty and discrimination), and scale reliabilities. Additionally, exploratory and confirmatory factor analyses were performed. As stated by Jöreskog (Reference Jöreskog, Cydeck and MacCallum2007), factor analysis does not have to be exclusively exploratory or confirmatory, especially in studies where there are variables of known and unknown composition. Therefore, by cross-validating the results from the exploratory factor analysis, through the conduction of a confirmatory factor analysis on a different data set (e.g. using random split samples), it was possible to obtain more reliable results. Following Gerbing and Hamilton’s (Reference Gerbing and Hamilton1996), and Jöreskog (Reference Jöreskog, Cydeck and MacCallum2007) recommendation, SPSS was used to randomly split the sample and run an exploratory factor analysis on one sample (50% of the original sample) as a precursor to the confirmatory factor analysis on the other sample (the remaining 50% of the original sample).
The exploratory factor analysis was performed with SPSS version 21.0 (SPSS Inc., Chicago, IL, USA). This study followed Takeuchi and colleagues’ (2007; see also Zacharatos et al., Reference Zacharatos, Barling and Iverson2005) procedures to evaluate a uni-dimensional conceptualization of the HRP. A principal axis factor analysis with a single-factor solution was performed. The confirmatory factor analysis was conducted with AMOS version 21.0 (Amos, Chicago, IL, USA). The overall goodness of the fit was based on combinations of several fit indexes. The model had adequate fit to the data when there was a significant chi-square, .90 or higher for the Tucker Lewis (TLI) and fit indices (CFI), .06 or less for the root mean square error of approximation (RMSEA), and .08 or less for the standardized root mean square (SRMR) (Arbuckle, Reference Arbuckle2003).
After the confirmatory factor analysis, it was also tested the model´s fit across different sectors (services and industry). First, the baseline model’s fit was determined (Byrne, Reference Byrne2010). Then, the baseline model was compared to (a) a fully constrained model, in which all coefficient paths were constrained to be identical in both sectors; and (b) an invariant measurement model, in which only the factor loadings for the observed variables were constrained to be identical in both sectors. These procedures were conducting using multiple group analysis, which included both samples (N = 4,551).
Finally, to test the relationship between the HRP measure and other variables, it was necessary to perform a correlation analysis between the new and a well-known measure. Thus, the correlation coefficients between the HRP system and TAW’s AC (Meyer et al., Reference Meyer, Allen and Smith1993) toward the agency and client company were analyzed.
Results
Response process
Respondents were assured that their answers were confidential and anonymous. Participants were informed that they would have the opportunity to receive the overall results. These instructions were written on the questionnaire’s cover letter. The instructions explained that the questions were directly related to several parts of their daily work, specifically their perceptions of employment relationships. Participants were informed that the questionnaire was not a test and that there were no right or wrong answers. Workers were also assured that the companies would only have access to a final report and not to the raw data, as the data were used exclusively for academic research. The lead researcher’s email address was included in the cover letter in addition to a website address where respondents could find more information about the research project, including the involved academic organizations, its goals, outcomes, partners and other researchers included in the process. There was no incentive (cash or otherwise) for participating in this project.
Internal process
Descriptive statistics (mean, standard deviation, median, and variance) for the HRP scale and both types of AC were calculated and are presented in Table 1. The HRP scale items characteristics, including difficulty and discriminatory power, are presented at Table 2. The item difficulty showed that TAWs responded using all values on the Likert scale and that 6 and 7 were the most commonly chosen answers (Table 2). For item discrimination, the results indicate that all 20 items were good discriminants, as all values were greater than .40 (Table 2).
Table 1. Descriptive Statistics and Correlations
Note: ACA = AC toward the agency; ACC = AC toward the client company. n = 4,551.
* p < .01.
Table 2. HRP Scale Items
Note: Item disc. = Item Discrimination; n = 4551. (a) Exploratory factor analysis with a random sample of approximately 50% of all cases (n = 2,273).
An exploratory factor analysis with principal axis factoring that imposed a single-factor solution analyzed TAW’s responses on the HRP scale (Table 2). All 20 items had factor loadings of .45 or above, and the eigenvalue was 9.70 and explained 48.48% of the variance. These results are in accord with (a) Kaiser (1960; in Field, Reference Field2013), who establishes that the eigenvalues should be greater than 1; (b) Stevens (2002; in: Field, Reference Field2013, p. 645), who recommends, “interpreting only factor loadings with an absolute value greater than .4 (that explain approximately 16% of the variance in the variable)”; and (c) Tabachnick and Fidell (Reference Tabachnick and Fidell2001), who suggested that the minimum acceptable loading is .32 (approximately 10% of the variance in the variable). In addition, the percent of variance explained by the one-factor solution was very close to 50%. Although there are no strict guidelines for this parameter, it is an acceptable target (Table 2). The HRP scale also had strong internal-consistency reliability with a Cronbach’s alpha value of .94.
These findings are acceptable and comparable to Takeuchi and colleagues’ results (2007). The items from their HRP system scale had factor loadings of .37 or above that loaded on a single factor, which explained 35.82% of the variance and had an eigenvalue of 7.88. Moreover, they had a reliability value of .90, which was similar to Zacharatos and colleagues’ study (α = .89; 2005).
For the confirmatory factor analysis, the model with one latent factor (i.e., HRP system) had good fit, χ2(149) = 2429.40, p < .001; SRMR = .06; CFI = .93; TLI = .91; RMSEA = .08 (see Figure 1).
Figure 1. Confirmatory Factor Analysis Model, with One Latent Factor
HRP = Human Resources Practices.
The internal consistency for AC toward the agency and client company were also adequate at .89 and .90 (Table 1).
For the multiple group analysis between the two sectors (industry and services), a separate baseline model was examined for each sub-samples. As shown in Table 3, these two models had acceptable fit to the data. Subsequently, the two sectors were constrained to compare the baseline and fully constrained models. There was a significant difference between the models chi-square values, which indicated a significant decrease in model fit for the fully constrained model (Cheung & Rensvold, Reference Cheung and Rensvold2002). The baseline multiple group model had acceptable fit, χ2(298) = 5,230.98, p. 001, CFI = .92; RMSEA = .06 (Table 3).
Table 3. Fit statistic for the Measurement of HRP System of TAW
Note: Df = degree of freedom; SRMR = Standardized root mean square; CFI = Bentler comparative fit index; TLI = Tuckler Lewis index; RMSEA = root mean square error of approximation; ns = non-significant.
** p < .001.
Finally, the baseline model was compared to an invariant measurement model (factor loadings) and there were no differences between the two sectors on the HRP system. The chi-square and CFI differences were not significant, which supports invariance across the two samples (Table 3).
Relationship to other variables
The results also indicate that there may be different relationships between perceptions of HRP systems and AC for the agency and client company. As expected, the HRP system was positive and significantly related to both types of AC. The correlations ranged from .45 to .54 (p < .001; Table 1).
Discussion
Increases in contingent work and changes in complexity and configuration reflects a vast shift in the employment landscape. Prior research has indicated that it is important to examine existing ideas, theories, and instruments before approaching new work arrangements (Zimmerman et al., Reference Zimmerman, Gavrilova-Aguilar and Cullum2013;) that has new rules, expectations, goals, and, most of all, new roles for the involved actors.
The present study responded to this challenge by developing and assessing the psychometric proprieties of a new HRP measure that was specifically designed for TAWs. The measure was developed in line with Messick’s (Reference Messick1995) framework for construct validation. For the response process, the data were comprehensively processed. The internal structure of the measure was also acceptable. First, the descriptive analyses and item characteristics confirmed the importance of each item to the measure. Second, consistent with earlier results (Takeuchi et al., Reference Takeuchi, Lepak, Wang and Takeuchi2007; Zacharatos et al., Reference Zacharatos, Barling and Iverson2005), the exploratory and confirmatory factor analyses confirmed a one-factor structure for the overall sample, as well as both sector sub-samples (services and industry). The diversity in job sectors was not reflected in TAW’s perception of HRP system. TAW in services and industry sectors had homogeneous interpretations of the four-general HRPs. These findings are important because services and industry are the most representative sectors in hiring TAWs (World Employment Confederation, WEC, 2018). Future studies should replicate this invariance analysis with different sectors to replicate this pattern.
For the last step in Messick’s (Reference Messick1995) framework, determining if there is a relation with other variables, TAW’s perceptions of the HRP system were positively related to both types of AC. This confirms that investing in TAWs can lead to high AC (Chambel et al., Reference Chambel, Castanheira and Sobral2016; Chambel & Sobral, Reference Chambel and Sobral2011). Further, the shared variance between the HRP system and both ACs was stronger for the client company (.52), than the agency (.45). Consistent with previous outcomes, this shows that (a) TAWs can create two relationship exchanges and affectively connect with both companies (Buch et al. Reference Buch, Kuvaas and Dysvik2010; Sobral et al., Reference Sobral, Chambel and Castanheira2016); and (b) TAWs tend to create closer relations with the client company because they hope to obtain a permanent job there (George, Levenson, Finegold, & Chattopadhyay, Reference George, Levenson, Finegold and Chattopadhyay2010).
Overall, the results demonstrate that the practices that are used to create this HRP system are consistent and coherent with the literature and the context. As Cagliano, Caniato, Longoni, and Spina (Reference Cagliano, Caniato, Longoni and Spina2014) have suggested, using temporary work may be associated with more than just flexibility, as companies often use this process to select permanent workers. Therefore, the recruitment process is a fundamental step in hiring these workers, which can be maximized through an effective socialization. Socialization is a key practice inside the HRP system, on one hand it allows the return made in the recruitment and training processes, leveraging the potential competitive advantage of human capital; and on the other hand, effectively socialized workers may be more skilled and loyal (Batistič, Reference Batistič2018). A good example of this practice occurs when workers are well integrated and feel that they are fairly treated. When there is distance between contingent workers and others in the workplace, feelings of alienation may occur (Halbesleben & Clark, Reference Halbesleben and Clark2010). In contrast, fair treatment and support reduces counterproductive behaviors (Bonet et al., Reference Bonet, Cappelli and Hamori2013). Another form of TAW alienation is related to the lack of training opportunities (Ulrich et al., Reference Ulrich, Younger and Brockbank2008). Training and employability development opportunities, as shown by Chambel and Sobral (Reference Chambel and Sobral2011), contribute to TAW’s feelings of value and support from the organization. Finally, the performance appraisal practice directs workers to the companies’ goals by shaping their outcomes in accord with defined targets (Takeuchi et al., Reference Takeuchi, Lepak, Wang and Takeuchi2007). Additionally, if correct feedback is provided, it can help workers to better perform and adjust themselves to the organization.
Our research has both theoretical and practical implications. First it highlights the need to design empirical studies focused on the workers’ perceptions. Previous research as Nishii and Wright (Reference Nishii, Wright and Smith2008) suggested that the variability associated with workers specificities may act as a moderator between HRP and labor variables. In the present research we propose to study as an element of variability the workers contractual typology, namely by focusing on a specific sample of TAW. There are several measures available to evaluate workers’ perceptions over the HRP, but none of them in the context of TAW. Therefore, this study is innovative because it developed a measure specifically built for contingent workers in tripartite employment relationships. HRP such as recruitment, socialization, training or performance evaluation gain in this context a new meaning, both by what they incorporate and by the results they intend to bring, once two companies are involved in the development and application of these practices.
The research design is another significant outcome of the current study. In the same vein of previous research (Takeuchi et al., Reference Takeuchi, Lepak, Wang and Takeuchi2007; Zacharatos et al., Reference Zacharatos, Barling and Iverson2005), the HRP were considered as a global system and not as a set of isolated procedures within the organizational dynamics. Although recognizing the existence of different themes and fields of action, the appreciation of the HRP as a system offers a more tangible reading over the perceptions made by individuals. Workers do not feel the practices separately from each other, but as a set of interconnected actions working as a cohesive structure.
Additionally, this study confirms the existence of a positive relationship between the perception of the HRP system and the double affective commitment of TAW. This result supports the existence of an exchange relationship, even when workers have a contingent contract; and emphasizes that TAW have an overall perception of the investments made in the HR, regardless of the company that executes it (i.e., agency or client company). Also, this study shows that the relationship between the HRP system and the affective commitment with the agency and the client company may be different. In fact, we found a stronger relationship between the HRP system and the commitment with the client than the commitment with the agency. This evidence supports the claim that TAW’s create a closer relationship with the client company, where TAW expect to obtain a permanent or regular contract (George et al., Reference George, Levenson, Finegold and Chattopadhyay2010).
Overall, the results show that the HRP influence each other and, in the case of TAW, even if applied by agency and/or the client company, can be associated with the same HRP system (Chambel et al., Reference Chambel, Castanheira and Sobral2016). In addition, the perception that TAW create of each company (i.e., agency and client company), generates a spillover effect due to the relationship between: worker/agency and worker/client company. The perception that each of these company generates in the individual, can spread to the perception that the worker has regarding the other company. By acting together and creating common and complementary strategies for managing TAW, each company can contribute to the success of the other company. In the end, the development of a positive perception over a joined HRP system, will lead to a positive reinforcement of TAW double affective commitment.
Limitations and directions for future research
First, this study focused on Portuguese TAW and is not representative. Cross-cultural interpretations or generalizations beyond Europe should take caution. Despite this limitation, the present study had a large sample and respondents were from several agencies, client companies, and several professions. Additionally, to the research team knowledge, there are no other HRP instruments that were developed and tested with TAWs and validated in different sectors. An important strength of the present study is that it indicates that it is important to examine TAW’s HRP perceptions in an employment relationship. Future studies should use this measure with cross-cultural and representative samples. Moreover, this measure should be tested, and, if necessary, adapted, to other contingent workers.
Second, using cross-sectorial and self-report data raises questions about the common-method bias. To address this limitation, several methodological recommendations were followed (Podsakoff, MacKenzie, Lee, & Podsakoff, Reference Podsakoff, MacKenzie, Lee and Podsakoff2003). Future research should adopt a longitudinal design and explore (a) how TAWs react to this measure over time; and (b) how it relates to other variables beyond AC.
By examining the most important HRPs in a system, it was possible to identify how these distinct but interrelated practices contribute to communicating that TAWs are important to the company, which may lead them to reciprocate in the same way. This is an important contribution to the literature, not just by adjusting existing conceptions to the labor market’s actual configuration; but also because few studies have recommended investments in contingent workers. Although these workers are formally temporary, they become part of the company in which they work, as they work together with other employees and pursue the company’s goals. TAW’s work in client’s facilities, and therefore, represent the agencies. TAW’s behaviors and attitudes are crucial for agencies and client companies strategies. Consequently, companies should treat them fairly and invest resources in them.
