Introduction
Trust is considered a key construct within the organizational context (Fulmer & Gelfand, Reference Fulmer and Gelfand2012) and plays an important role in research by different disciplines, such as Psychology, Ethics, Sociology and Economics (Colquitt et al., Reference Colquitt, Scott and LePine2007). Tan and Lim (Reference Tan and Lim2009) define organizational trust as “the willingness of employees to be vulnerable to the actions of their organization, without employees having control over these actions and behaviours” (p. 46). Key ideas from this definition include the vulnerability of workers to the actions of the organization and the willingness or voluntariness to trust.
Organizations are considered multilevel systems, so organizational trust must be analyzed from different levels: (a) at the individual level, where most research has focused; (b) at the group or team level, considering the shared perceptions resulting from the interaction between teammates; and (c) at the organizational level itself (Fulmer & Gelfand, Reference Fulmer and Gelfand2012; Salanova et al., Reference Salanova, Acosta-Antognoni, Llorens and Le Blanc2021). Therefore, trust is crucial as it facilitates the formation of shared perceptions, expectations, and behaviors among workers, contributing to the creation of a positive organizational environment (Costa & Anderson, Reference Costa and Anderson2011; Dirks & Ferrin, Reference Dirks and Ferrin2002).
Fulmer and Gelfand (Reference Fulmer and Gelfand2012) further consider trust by taking into account the key actors in organizations, i.e., trust in the leader or immediate supervisor (i.e., vertical trust) and trust among coworkers (i.e., horizontal trust). According to Omar (Reference Omar, Littlewoord and Vega2015), trust is important in organizations because it is considered an essential variable for organizations to achieve their goals. Previous studies highlighted the key role of organizational trust and its positive impact on organizational outcomes (De Jong et al., Reference De Jong, Dirks and Gillespie2016; Morrissette & Kisamore, Reference Morrissette and Kisamore2020).
Addressing the role of trust at the team level has become increasingly important due to the evolving characteristics of organizations, which are moving toward structures where teams play a pivotal role (Costa & Anderson, Reference Costa and Anderson2011). An example of this is the healthcare organization, characterized by its complexity and multilevel structure, where work teams occupy an important place within the organizational framework.
Although research on horizontal trust at the work team level in the healthcare context is limited, some studies provide evidence of its important mediating role between social resources and organizational outcomes in the healthcare sector, positioning it as a strategic variable within healthcare organizations (Olvera et al., Reference Olvera, Llorens, Acosta and Salanova2017), as suggested by the HEalthy & Resilient Organizations (HERO) Model (Salanova et al., Reference Salanova, Llorens, Cifre and Martínez2012).
In this context, the complexity of healthcare organizations and their organizational demands requires that these demands do not negatively affect the well-being of workers, healthcare teams, and organizational outcomes. Additionally, healthcare organizations must manage multiple processes, such as role allocation, procedural design, decision-making, resource allocation, and reward systems, among others. To ensure effective healthcare management, it is essential that health professionals perceive these processes as fair. Therefore, fostering organizational justice is a fundamental goal.
On the other hand, healthcare organizations face critical environments and periods of instability and adversity. In these challenging contexts, the importance of resilience emerges as the capacity of workers, teams, and organizations to adapt, recover, and maintain performance in the face of difficulties (Salanova et al., Reference Salanova, Llorens, Cifre and Martínez2012). This research aims to delve into the antecedents and consequences of horizontal trust at the team level in the healthcare context based on the HERO model (Salanova et al., Reference Salanova, Llorens, Cifre and Martínez2012).
The objective of this study is to assess at the team level, the relationship between perceived organizational justice and perceived team performance, considering the mediating role of horizontal trust and collective resilience perceived by workers in the context of the healthcare organization. In this research, we introduce significant innovations, including (a) examining the interactions between team-level variables, considering the collective perceptions of team members in the healthcare context; (b) exploring the impact of horizontal trust in fostering well-being, specifically collective resilience, and its influence on organizational outcomes such as job performance; and (c) assessing the mediating role of horizontal trust and collective resilience in the relationship between perceived organizational justice and job performance.
Theoretical background and research hypothesis
Decades of research have highlighted the importance of trust in the organizational context (Costa, Reference Costa2017). Among its antecedents are studies showing how human resource management practices (i.e., organizational support) promote trust (Tremblay et al., Reference Tremblay, Cloutier, Simard, Chênevert and Vandenberghe2010). Similarly, Acosta et al. (Reference Acosta, Salanova and Llorens2011), in a study with aggregated data at the work team level, provide evidence of how organizational practices (i.e., work-family reconciliation, prevention of workplace bullying, psychosocial health, organizational communication) promote organizational trust (i.e., vertical trust). Furthermore, organizations use their organizational resources (i.e., team resources, transformational leadership, organizational justice) to foster organizational trust (i.e., vertical trust, horizontal trust) among workers or work teams, and there is evidence to support this (Acosta et al., Reference Acosta, Llorens, Escaff, Díaz-Muñoz, Troncoso, Salanova and Sanhueza2019; Chen et al., Reference Chen, Wu, Chang, Lin, Kung, Weng, Lin and Lee2015; Olvera et al., Reference Olvera, Llorens, Acosta and Salanova2017; Omar, Reference Omar2011; Yue et al., Reference Yue, Men and Ferguson2019).
As consequences of organizational trust, research reveals its impact on organizational outcomes such as job performance, commitment, continuous improvement, and openness to perceived change processes of both workers and work teams (Khattak et al., Reference Khattak, Zolin and Muhammad2020; Olvera et al., Reference Olvera, Llorens, Acosta and Salanova2017; Salanova et al., Reference Salanova, Acosta-Antognoni, Llorens and Le Blanc2021; Tremblay et al., Reference Tremblay, Cloutier, Simard, Chênevert and Vandenberghe2010; Vanhala et al., Reference Vanhala, Heilmann and Salminen2016; Vásquez-Pailaqueo et al., Reference Vásquez-Pailaqueo, Inostroza-Naranjo and Acosta-Antognoni2021; Yue et al., Reference Yue, Men and Ferguson2019). Moreover, research suggests that organizational trust has a positive relationship with elements that contribute to well-being at work, such as work engagement, indicating that the greater the trust of workers or work teams, the higher their levels of work engagement (i.e., vigor, dedication, and absorption) (Acosta et al., Reference Acosta, Torrente, Llorens and Salanova2015; Lin, Reference Lin2010; Ugwu et al., Reference Ugwu, Onyishi and Rodríguez-Sánchez2014).
Salanova et al. (Reference Salanova, Llorens, Cifre and Martínez2012) introduce the HERO concept and propose a heuristic model backed by empirical evidence that supports its theoretical validity, showing among its strengths the ability to respond to the need for data analysis at different levels (i.e., work teams). These authors consider a healthy organization to be “one that makes systematic, planned and proactive efforts to improve employee and organizational processes and outcomes” (p. 788). Furthermore, they consider it to be resilient because it knows how to emerge stronger from unfavorable situations and critical circumstances, maintaining its functioning and its results in terms of financial performance (Salanova et al., Reference Salanova, Llorens, Cifre and Martínez2012).
The HERO model identifies three key interacting components: (a) resources (i.e., social and task resources) and healthy organizational practices as strategies for work organization; (b) healthy employees related to psychosocial well-being (i.e., organizational trust, collective resilience); and (c) healthy organizational outcomes related, among others, to high job performance (Salanova, Reference Salanova2008; Salanova et al., Reference Salanova, Llorens, Cifre and Martínez2012, Reference Salanova, Martínez and Llorens2014).
According to the HERO model (Salanova et al., Reference Salanova, Llorens, Cifre and Martínez2012), organizational trust is included in the component “healthy employees or teams” and distinguishes two dimensions of organizational trust: vertical trust and horizontal trust. Vertical trust refers to the degree to which employees trust the actions of their superiors or the organization in which they work, whereas horizontal trust refers to the degree to which employees trust the people they work with and enjoy their company.
From the above, it can be inferred that organizational practices and resources have a clear influence on the trust and well-being of employees and work teams, which in turn impacts organizational outcomes. Consequently, organizational trust is regarded as a mediating variable (Olvera et al., Reference Olvera, Llorens, Acosta and Salanova2017; Tan & Lim, Reference Tan and Lim2009; Vásquez-Pailaqueo et al., Reference Vásquez-Pailaqueo, Inostroza-Naranjo and Acosta-Antognoni2021; Zhu et al., Reference Zhu, Newman, Miao and Hooke2013). From this perspective, taking into account social exchange theory (Blau, Reference Blau1964), organizational practices and resources can generate social exchange relationships between workers and work teams, which act on trust levels and, by reciprocity, also on performance (Salanova et al., Reference Salanova, Acosta-Antognoni, Llorens and Le Blanc2021).
Organizational justice, horizontal trust, and job performance
The concept of organizational justice has its origin in the approaches of philosophers such as Plato and Aristotle, who focused on how to find justice in the different facets of life (Ryan, Reference Ryan1993, cited by Rodriguez et al., Reference Rodríguez, Martínez and Salanova2014). Nevertheless, independent of this philosophical perspective, it is in the studies of Adams (Reference Adams1966) and his Theory of Equity, that the impact of perceptions of equity in the labor context is addressed, understanding justice as a social construction.
The term organizational justice is attributed to Greenberg (Reference Greenberg1987), who defines it as the perceptions workers have about what is fair within the organization they work for, examining the understanding of those perceptions and how it impacts organizational outcomes (Cropanzano et al., Reference Cropanzano, Bowen and Gilliland2007; Moorman, Reference Moorman1991). Research on organizational justice has historically evolved into a multidimensional construct. It mainly contemplates three dimensions: (a) distributive justice, which refers to how fair workers perceive the distribution of rewards to be according to their contribution to the organization (Adams, Reference Adams1966); (b) procedural justice, which determines how fair the decision-making procedures are perceived to be (Thibaut & Walker, Reference Thibaut and Walker1975); and (c) interactional justice, which focuses on the worker’s perception of the quality of treatment received during the process of implementing procedures within the organization (Bies & Moag, Reference Bies, Moag, Lewicki, Sheppard and Bazerman1986). Greenberg (Reference Greenberg1993) distinguishes two different dimensions of interactional justice, which he identifies as interpersonal justice and informational justice. Interpersonal justice refers to the degree to which the organization’s workers perceive that their superiors treat them with respect and dignity in relation to procedures and decisions. On the other hand, informational justice refers to the degree to which superiors provide workers with information on the procedures and decisions adopted.
The measurement of organizational justice has been approached differently by researchers. Among the instruments developed are the contributions of Moorman (Reference Moorman1991), Niehoff and Moorman (Reference Niehoff and Moorman1993) that measure perceptions of justice in three dimensions (i.e., distributive justice, procedural justice, and interactional justice). On the other hand, Colquitt (Reference Colquitt2001) operationalizes it into four dimensions by differentiating two additional dimensions within interactional justice: interpersonal and informational justice. Furthermore, Dai and Xie (Reference Dai and Xie2016), in an article on the review and perspective of interactional justice, state that Colquitt’s scale is mainly used in research on interactional justice and show that researchers consider it in one or two dimensions, depending on the research.
In relation to the HERO model (Salanova et al., Reference Salanova, Llorens, Cifre and Martínez2012), organizational justice is included in the first component, corresponding to healthy organizational practices and resources. Specifically, this block identifies actions that promote the organization’s task resources (i.e., distributive justice and procedural justice) and social or interpersonal resources (i.e., interpersonal justice and informational justice) (Rodríguez et al., Reference Rodríguez, Martínez and Salanova2014).
Previous research reveals a significant and positive relationship between organizational justice and organizational trust, that is, how perceptions of justice promote workers’ trust in their organization. For example, Brockner and Siegel (Reference Brockner and Siegel1996) demonstrated how trust is influenced by workers’ perception of procedural justice. In this sense, when a worker perceives a decision to be fair, they assume that future decisions will also be fair. Similarly, Chen et al. (Reference Chen, Wu, Chang, Lin, Kung, Weng, Lin and Lee2015) conducted a study in the healthcare context, specifically on nurses at a university hospital in Taiwan, concluding that professionals’ perceptions of organizational justice in its three dimensions (i.e., distributive justice, procedural justice, and interactional justice) positively influence organizational trust, trust in management, and trust among coworkers, increasing their commitment to the organization. Furthermore, several meta-analyses on the role of justice in organizations reveal positive correlations between justice and organizational trust; thus, the more organizational justice is perceived, the more trust develops (Akar, Reference Akar2018; Cohen-Charash & Spector, Reference Cohen-Charash and Spector2002; Colquitt et al., Reference Colquitt, Conlon, Wesson, Porter and Ng2001; Yorulmaz et al., Reference Yorulmaz, Püsküllüoğlu, Colak and Altınkurt2021).
The literature also reveals the impact of perceived fairness and organizational trust on job performance at both the individual and team levels, and how trust is positioned as a mediating variable between fairness and job performance (Colquitt et al., Reference Colquitt, LePine, Piccolo, Zapata and Rich2012; Ha & Lee, Reference Ha and Lee2022; Hanif & Endang, Reference Hanif and Endang2018). Regarding job performance, Rotundo and Sackett (Reference Rotundo and Sackett2002, p. 66) define it as “the actions and behaviors that are under the individual’s control and that contribute to the organization’s goals.” This definition, extended to the level of work teams, is understood as those team behaviors that contribute toward the organization’s goals (Borman & Motowidlo, Reference Borman and Motowidlo1997). Other authors define team performance more generally as the degree to which a work team achieves its mission or goal (Devine & Philips, Reference Devine and Philips2001).
In the HERO model, the team’s job performance is covered by the component “healthy organisational outcomes” and has two dimensions (Goodman & Svyantek, Reference Goodman and Svyantek1999): (a) Intra-role performance, which refers to activities related to formal work and task description that contribute to the technical basis of the organization and (b) Extra-role performance, referring to those activities that exceed the given tasks that workers must perform, which are not formally part of the job and that workers voluntarily carry out. These two dimensions of job performance are considered complementary and, together, provide a comprehensive view of team job performance (Devine & Philips, Reference Devine and Philips2001; Olvera et al.,Reference Olvera, Llorens, Acosta and Salanova2017; Salanova et al., Reference Salanova, Acosta-Antognoni, Llorens and Le Blanc2021).
Regarding horizontal trust at the team level, researchers consider it as a key construct to explain how organizational resources (i.e., organizational justice) promote high team performance. According to Costa (Reference Costa2003), addressing horizontal trust at the work team level, understood as the degree to which team members trust each other, implies considering work teams as organizational groups with a common goal in which all team members are stakeholders. In these teams, there is an interdependence of tasks and, therefore, the development of perceptions, shared understandings, and expected patterns of behaviors. Currently, research reveals how horizontal trust at the work team level is related to team performance and how it mediates between organizational resources and job performance of work teams (De Jong et al., Reference De Jong, Dirks and Gillespie2016; Olvera et al., Reference Olvera, Llorens, Acosta and Salanova2017; Salanova et al., Reference Salanova, Acosta-Antognoni, Llorens and Le Blanc2021).
Taking the HERO model as a reference and based on the above, the following research hypothesis is proposed:
Hypothesis 1: At the team level, horizontal trust fully mediates the relationship between organizational justice (i.e., distributive justice, procedural justice, informational justice, and interpersonal justice) and job performance.
Organizational justice, horizontal trust, collective resilience, and job performance
Organizations, as well as teams and professionals, are exposed to critical environments, periods of instability, and adversity. It is in this context of crisis that the concept of resilience emerges. According to Horne and Orr (Reference Horne and Orr1998), resilience is essential for organizations to respond productively to significant changes that disrupt the expected pattern of events without engaging in a prolonged period of regressive behavior. Concretely, resilience has been defined at the team level as “a team’s belief that it can absorb and cope with strain, as well as a team’s capacity to cope, recover and adjust positively to difficulties” (Carmeli et al., Reference Carmeli, Friedman and Tishler2013, p.149).
In the HERO model, resilience is part of the “healthy employees” component. It refers to the situation in which workers have positive psychological resources with which to feel good at work which is related to well-being at work and, in turn, to desirable outcomes such as performance (Salanova et al., Reference Salanova, Llorens, Cifre and Martínez2012, Reference Salanova, Llorens and Martínez2019). Although there is limited empirical evidence on the antecedents and consequences of resilience, existing studies indicate that job resources positively influence its development (Meneghel et al., Reference Meneghel, Salanova and Martínez2013). For instance, Vera et al. (Reference Vera, Rodríguez-Sánchez and Salanova2017) conducted a multilevel study involving 194 teams from 38 organizations. Their findings demonstrated that resources at both the team and organizational levels are significantly associated with greater resilience within teams.
Other research has shown a significant positive relationship between resilience and job performance at the both individual and team levels (Beuren et al., Reference Beuren, dos Santos and Theiss2021; Kim, Reference Kim2020; Luthans et al., Reference Luthans, Avolio, Avey and Norman2007). These studies recognize that resilience is a psychological mechanism that explains how teams manage collective resources and job demands for better performance (Meneghel et al., Reference Meneghel, Martínez and Salanova2016, Reference Meneghel, Salanova and Martínez2016). On the other hand, studies show the need to deepen the analysis of perceptions of organizational justice and its relationship with resilience (Oliveira & Ferreira, Reference Oliveira and Ferreira2016; Raffety, Reference Raffety2021).
Finally, previous research has highlighted the relevance of trust (i.e., horizontal trust) as a predictor of team resilience in adverse situations. In this sense, organizational trust enhances interpersonal relationships among team members by improving coordination and cooperation. This results in a collaborative work environment, where resources to promote resilience are acquired (Pavez et al., Reference Pavez, Gómez, Laulié and González2021). Based on the above, the proposed hypotheses are as follows (see Figure 1):
Hypothesis 2: At the team level, horizontal trust is positively related to collective resilience.
Hypothesis 3: At the team level, collective resilience fully mediates the relationship between organizational justice and job performance.
Hypothesis 4: At the team level, horizontal trust and collective resilience fully mediate the relationship between organizational justice and job performance.
Research model and study hypothesis.
Note: “+” positive and significant relationship; “ns” nonsignificant relationship.
Method
Participants and procedure
The sample consisted of 927 workers distributed into 100 work teams from seven healthcare centers in Spain. The total population of professionals in these centers was 6304, resulting in a participation rate of 15%. The sample is representative with a margin of error of .03 for a confidence level of 95%. In terms of employment, 63% of the participants had a permanent employment contract and an average length of service in the company of between five and ten years (SD = 1.51). Additionally, the average age of the workers who participated in the study was 44 years (SD = 10.3), and 68.2% were women. Finally, the average team size was nine members (SD = 6.8).
The research focuses on work teams within the healthcare context made up of groups of workers who share responsibility for performance results and have the leadership of a referent or supervisor (George, Reference George1990). The inclusion criterion for the sample was that they had been with the company for at least six months, which, according to McCarthy (Reference McCarthy1992), would allow them to reach the first stages of the labor socialization process. They could, therefore, answer the questionnaire with more realistic information. It is important to note that the sample explicitly excludes supervisors and employees with managerial status to focus exclusively on workers’ perceptions.
In terms of the percentage of representation by professional category of the participants in the research study, the most represented category is nurses (34%), followed by doctors (22%), auxiliary nursing care technicians (16.2%), administrative personnel (15.7%), orderlies (7%), and social worker (5%).
In relation to the procedure and after the acceptance of study participation by the management of the healthcare centers during initial contact, informative meetings were held with key agents of the organization (i.e., managers and employees) in which information was provided regarding the procedure and the ethical and legal aspects of the research project. Those workers who agreed to participate voluntarily in the research completed an online questionnaire. It should be noted that confidentiality was guaranteed in the treatment of the data. Compliance with current regulations related to the Protection of Personal Data was ensured and the study was approved by the Ethics Committee.
Measures
All items were evaluated on a 7-point Likert scale ranging from 0 (Strongly Disagree) to 6 (Strongly Agree).
Organizational Justice. It was assessed using the Spanish adaptation of Colquitt’s (Reference Colquitt2001) organizational justice scale (Díaz-Gracia et al., Reference Díaz-Gracia, Barbaranelli and Moreno-Jiménez2014), which was included in the HERO questionnaire (Salanova et al., Reference Salanova, Llorens, Cifre and Martínez2012). The scale comprises 20 items distributed across four dimensions: (a) informational justice (five items; e.g., “Our supervisor is honest in communicating with us”; ω = .96); (b) interpersonal justice (four items; e.g., “Our supervisor treats us appropriately”; ω = .95); (c) distributive justice (four items; e.g., “Rewards are fair given our performance”; ω = .98); and (d) procedural justice (seven items; e.g., “We have been able to express our views and feelings about the procedures used to reward us”; ω = .97).
Horizontal Trust. It was measured using McAllister’s (Reference McAllister1995) questionnaire, adapted and integrated into the HERO questionnaire (Salanova et al., Reference Salanova, Llorens, Cifre and Martínez2012). The scale comprises four items (ω = .92), distributed into two observable indicators (with two items each). An example item is “The workers in my work team can talk freely with colleagues about possible difficulties experienced at work.”
Resilience. It was assessed using a nine-item scale based on Mallak’s (Reference Mallak1998) principles and was included in the HERO questionnaire (Salanova et al., Reference Salanova, Llorens, Cifre and Martínez2012). These items were organized into three observable indicators (with three items each; ω = .92). An example item is “We try to look for the positive side of difficult situations.”
Team Job Performance. It was measured using an adaptation of Goodman and Svyantek’s (Reference Goodman and Svyantek1999) performance scale, integrated into the HERO questionnaire (Salanova et al., Reference Salanova, Llorens, Cifre and Martínez2012). The scale consists of six items grouped into two dimensions: (a) intra-role performance (three items; e.g., “We, the workers in my work team, achieve the established performance criteria”; ω = .95); and (b) extra-role performance (three items; e.g., “The workers in my team perform functions that are not required but improve the image of the organisation”; ω = .88).
Data Analysis
Preliminary Analyses
Using IBM SPSS Statistics version 27.0 and the individual database, descriptive analyses (means, standard deviations, and correlations) were first performed. These analyses provided a first approximation of the behavior of the variables and served as a basis for further analyses. Next, the reliability of the scales was assessed using McDonald’s Omega (ω). This coefficient provides a robust estimate of internal consistency in multifactor measurement models. Values above .70 are considered acceptable, while values above .80 indicate good reliability (McDonald, Reference McDonald1999).
The variance inflation factor (VIF) and the Condition Index were calculated to identify possible multicollinearity problems. Values above 10 suggest significant multicollinearity, values between 5 and 10 indicate moderate problems (Kline, Reference Kline2015), and Condition Index values above 30 indicate significant collinearity (Belsley et al., Reference Belsley, Kuh and Welsch1980). As a corrective measure, the variables were centered with respect to their mean. Additionally, principal component analysis (PCA) was performed to reduce the redundancy between correlated dimensions.
Subsequently, analysis of variance (ANOVA) was performed to evaluate the significant differences in the scales evaluated between work teams (Fisher, Reference Fisher1925). This analysis allowed us to compare the means of the different teams and to assess whether the observed differences were statistically significant.
Finally, to justify the aggregation of individual data to the team level, intraclass correlation coefficients (ICC1 and ICC2) and the intragroup agreement index (Rwg) were calculated. ICC1 values above .12, and ICC2 values above .60 support the validity of the aggregation (Bliese, Reference Bliese, Klein and Kozlowski2000; Glick, Reference Glick1985), and Rwg values above .70 indicate a high level of intragroup consensus (James et al., Reference James, Demaree and Wolf1984).
Analyses with Aggregated Data
First, the means, standard deviations, and correlations of the study variables were calculated using the database aggregated at the team level (N = 100 teams). Subsequently, to ensure the stability and accuracy of the structural model estimates, a comprehensive multicollinearity analysis was performed using the VIF and Condition Index. To mitigate the effects of multicollinearity, Ridge Regression (Hoerl & Kennard, Reference Hoerl and Kennard1970) was implemented using R software (version 4.4.1) and the glmnet package. This regularization technique penalizes the magnitude of the estimated coefficients, thus improving the stability of the model. Ridge fitting includes the selection of the regularization parameter (λ) by cross-validation
First, the means, standard deviations, and correlations of the study variables were calculated using the database aggregated at the team level (N = 100 teams). Subsequently, to ensure the stability and accuracy of the structural model estimates, a comprehensive multicollinearity analysis was performed using the VIF and Condition Index. To mitigate the effects of multicollinearity, Ridge Regression (Hoerl & Kennard, Reference Hoerl and Kennard1970) was implemented using R software (version 4.4.1) and the glmnet package. This regularization technique penalizes the magnitude of the estimated coefficients, thus improving the stability of the model. Ridge fitting includes the selection of the regularization parameter (λ) by cross-validation
Subsequently, confirmatory factor analysis (CFA) was conducted to assess the factor structure of the constructs. The calculation of AVE, CR, and discriminant validity of the dimensions included in the proposed model was also included. CR values above .70 indicate adequate internal consistency (Bagozzi & Yi, Reference Bagozzi and Yi2012), and AVE values above .50 ensure satisfactory convergent validity (Hair et al., Reference Hair, Anderson, Tatham and Black1995).
Since the data were collected through a single questionnaire, the possible influence of common method variance (CMV) was addressed following the recommendations of Podsakoff et al. (Reference Podsakoff, MacKenzie, Lee and Podsakoff2003). Harman’s test was used to check whether a single factor explained more than 50% of the total variance. Subsequently, the fit indices of the base model and the model with a common method factor (CMF) were compared. This comparison allowed us to determine whether CMF significantly affected the relationships between latent variables.
Structural equation modeling (SEM) was performed using AMOS (version 26.0) to analyze the hypothesized relationships. This analysis tested the study hypothesis by comparing the two mediation models. Model 1 (M1) proposed full mediation of horizontal trust and collective resilience in the relationship between organizational justice and team-level work performance, including a direct relationship between horizontal trust and collective resilience. Model 2 (M2) proposed a partial mediation that included an additional direct relationship between organizational justice and work performance (intra- and extra-role).
Additionally, following Kline’s (Reference Kline2015) recommendations, two alternative models were tested: Alternative Model 1 (MA1), in which organizational justice and collective resilience mediate the relationship between horizontal trust and work performance, and Alternative Model 2 (MA2), in which resilience and work performance mediate the relationship between organizational justice and horizontal trust.
To assess the goodness of fit of the models, absolute and relative measures of fit were considered: the chi-square index (χ2), chi-square ratio over the degrees of freedom (χ2/df), root mean squared error approximation (RMSEA), and root mean squared standard residual (SRMR). A p-value associated with the chi-square greater than .05 indicates a good fit; similarly, the χ2/df indicates a good model fit for values less than 2. RMSEA values less than .05 are indicative of a very good fit; values between .05 and .08 are considered an acceptable fit; values between .08 and .10 indicate a moderate fit and values greater than .10 are indicative of a poor model fit (Browne & Cudeck, Reference Browne, Cudeck, Bollen and Long1993). For the SRMR, values below .08 are indicative a good fit (Hu & Bentler, Reference Hu and Bentler1999). Relative measures of fit include the normalized fit index (NFI), non-NFI (also called the Tucker-Lewis index, TLI, or NNFI), comparative fit index (CFI), and incremental fit index (IFI). For the TLI, IFI, CFI, and NFI, values above .90 are considered a good fit (Hoyle, Reference Hoyle and Hoyle1995). The Akaike Information Criterion (AIC; Marsh et al., Reference Marsh, Balla, Hau, Marcoulides and Schumacker1996) was calculated to compare non-nested competitive models; the lower the AIC values, the better the fit (Akaike, Reference Akaike1987). Finally, the indirect effects of team-perceived organizational justice on team job performance through horizontal trust and collective resilience were estimated using the bootstrap method with 5,000 samples, adhering to the recommendations of Preacher & Hayes (Reference Preacher and Hayes2008). Bias-corrected confidence intervals (CIs) and percentiles were analyzed for statistical significance. An indirect effect was considered significant if the CI did not encompass the value zero. This methodological approach facilitated the assessment of both direct and mediated effects, thereby ensuring the statistical validity of the estimates.
Results
Preliminary Analyses
Table 1 summarizes the descriptive statistics (means and standard deviations), correlations, intraclass correlation coefficients (ICC1 and ICC2), and intragroup agreement index (Rwg) for the study variables. The results showed that the ICC1 values ranged from .57 to .94, and the ICC2 values ranged from .87 to .98, justifying the aggregation of the variables to the team level (Chen et al., Reference Chen, Mathieu, Bliese, Yammarino and Dansereau2004). Furthermore, the intragroup agreement index (Rwg) exceeded the threshold of .70 (James et al.,Reference James, Demaree and Wolf1984) for all variables, supporting the aggregation of individual responses at the team level. Internal consistency of the scales was assessed using McDonald’s Omega (ω), with coefficients ranging from .88 to .98, indicating high reliability of the measures (McDonald, Reference McDonald1999).
Mean (M), standard deviation (SD), correlations, internal consistency, intraclass correlation coefficient (ICC1 and ICC2), and within-group interrater reliability (Rwg) for variables under study
Note. ω = McDonald Omega Coefficient; correlations are presented at the individual level (below the diagonal) and at the team level (above the diagonal).
** p < .001
The ANOVA revealed statistically significant differences between teams in all dimensions. The F-statistics, degrees of freedom, and significance levels for each variable were as follows: informational justice, F(99, 827) = 5.33, p < .001; interpersonal justice, F(99, 827) = 3.81, p < .001; distributive justice, F(99, 827) = 6.71, p < .001; and procedural justice, F(99, 827) = 7.30, p < .001; collective resilience, F(99, 827) = 6.23, p < .001; horizontal trust, F(99, 827) = 4.26, p < .001; intra-role performance, F(99, 827) = 4.06, p < .001; and extra-role performance, F(99, 827) = 3.52, p < .001.
The presence of multicollinearity among the predictors was assessed using the VIF and Condition Index. Initial diagnostics, with the individual database, revealed moderate multicollinearity, with VIF values ranging from 2.89 to 7.55 and a maximum Condition Index of 29.04, approaching the established thresholds (Kline, Reference Kline2015; Belsley et al., Reference Belsley, Kuh and Welsch1980). To address this, variables were mean-focused, which reduced VIF values to a range of 2.01 to 7.50 and the Condition Index to 7.52. Although this adjustment improved the results, the moderate multicollinearity persisted. Subsequently, PCA was applied to consolidate the correlated dimensions, resulting in two composite indicators: interpersonal-informational justice and distributive-procedural justice. This adjustment reduced redundancy between predictors, yielding final VIF values between 2.16 and 4.07 and a Condition Index of 5.19, both within acceptable thresholds.
Analyses with Aggregated Data
The multicollinearity diagnostic with the aggregated base revealed VIF values higher than five for two variables, i.e. horizontal trust (VIF = 7.40) and resilience (VIF = 6.07) and a condition index of 8.22. Ridge regression was applied by selecting a λ = .16 by cross-validation. The results showed a mean squared error (MSE) of .06, an RMSEA of .25 and an R2 of .75, indicating that the fitted model explains 75% of the variance and presents an acceptable accuracy in predicting the dependent variable job performance.
Confirmatory factor analysis demonstrated that the measurement model fit the data adequately (χ2(21) = 43.08, p = .003, CFI = .98, TLI = .97, RMSEA = .10, Standardized RMR = .027). Reliability and convergent validity were confirmed, with CR values ranging from .92 to .97 and AVE values from .70 to .95. Discriminant validity, assessed using the Fornell–Larcker criterion, indicated that the square root of the AVE for each construct exceeded the inter construct correlations. Standardized factor loadings ranged from .81 to .97, supporting the representativeness of the indicators. These results validate the robustness of the measurement model for subsequent structural analyses.
Structural equation modelling analyses were performed using an aggregated team-level database (N = 100). Specifically, four latent variables were used as follows: (a) organizational justice, consisting of two indicators: distributive-procedural justice and interpersonal-informational justice; (b) horizontal trust, consisting of two indicators with two items each; (c) collective resilience, consisting of three indicators with three items each; and (d) job performance, consisting of two indicators: intra- and extra-role performance.
To assess the potential common method bias, Harman’s single-factor test and the CMF approach were applied. The single-factor test identified five factors with eigenvalues above 1.0, collectively explaining 78.59% of the total variance, with the first factor accounting for 56.39%, thus exceeding the commonly accepted threshold of 50%. The results of the CMF analysis confirmed that the adjusted model without CMF exhibited a significantly better fit to the data. The adjusted model demonstrated good fit indices (χ2(22) = 42.43, χ2/df = 2.02, RMSEA = .10, CFI = .98, TLI = .97, AIC = 90.43) compared to the FMC model, which showed poor fit (χ2(44) = 405.52, χ2/df = 9.21, RMSEA = .30, CFI = .79, TLI = .74, AIC = 449.52). These findings indicate that the impact of CMV is not substantial and that the proposed model provides a valid and parsimonious representation of the data.
Table 2 presents the results of the structural equation models. The results obtained for model M1, which proposes a full mediation of horizontal trust and collective resilience in the relationship between organizational justice and team-level job performance, are as follows: χ2(22) = 42.43, p = .002, χ2/df = 1.93, TLI = .97, IFI = .98, CFI = .98, NFI = .97, RMSEA = .10, SRMR = .02, and AIC = 90.43. Although no statistically significant differences were found with the partial mediation model M2 (Δχ2(1) = 1.68, n.s.), in which a direct relationship between organizational justice and job performance was included, the results of this direct relationship were not significant (β = - .15, p = .25). The results support the proposed model, M1, because of the following findings: a) organizational justice is positively and significantly related to horizontal trust (β = .84, p < .001) and collective resilience (β = .65, p < .001); b) horizontal trust shows a positive and significant relationship with collective resilience (β = .28, p < .001); and c) both horizontal trust (β = .43, p < .001) and collective resilience (β = .57, p < .001) are positively related to job performance.
Fit indices for Structural Equation Models (N = 100 work teams)
Note. χ2 = Chi-square; df = degrees of freedom; χ2/df = Absolute Goodness of Fit Index; TLI = Tucker-Lewis Index; IFI = Incremental Fit; CFI = Comparative Fit Index; NFI = Normed Fit Index; RMSEA = Root Mean Square Error of Approximation; SRMR = Standardized Root Mean Square Residual; AIC = Akaike Information Criterion.
a Combined all variables into one factor.
b Full mediation of horizontal trust and collective resilience in the relationship between organizational justice and job performance, including a direct relationship between horizontal trust and resilience
c Partial mediation, incorporating a direct relationship between organizational justice and job performance.
d Mediation of organizational justice and resilience in the relationship between horizontal trust and job performance
e Mediation of resilience and job performance in the relationship between organizational justice and horizontal trust.
To demonstrate that the proposed model was not chosen arbitrarily, two alternative models (M A1 and M A2) were analyzed alongside the proposed model (M 1). The alternative model M A1 introduced a full mediation of the relationship between organizational justice and job performance through resilience (χ2(22) = 65.37, p < .001, CFI = .96, TLI = .94, RMSEA = .14, SRMR = .04), whereas M A2 proposed full mediation of the relationship between trust and job performance through collective resilience (χ2(22) = 46.73, p = .002, CFI = .98, TLI = .96, RMSEA = .11, SRMR = .03).
Chi-squared difference tests confirmed significant differences between M 1 and M A1 (Δχ2(0) = 22.94, p < .001), as well as between M 1 and M A2 (Δχ2(0) = 4.30, p = .038), reinforcing the superior fit of M 1. Additionally, the proposed model, M 1, demonstrated higher parsimony, as evidenced by its lower AIC value. Standardized path coefficients in M A1 and M A2 revealed nonsignificant direct effects of justice on job performance (β = -.12, p = .145) in M A1 and from trust to job performance (β = .00, p = .991) in M A2, further supporting the robustness of the proposed model M 1.
Indirect effects were assessed using the bootstrap method with 5,000 samples, following Preacher and Hayes (Reference Preacher and Hayes2008). The results confirmed that organizational justice significantly influenced team job performance through resilience (β = .56, 95% CI [.39, .70], p < .001) and the combined mediation of horizontal trust and collective resilience (β = .28, 95% CI [.18, .42], p < .001). Additionally, justice significantly affected resilience through horizontal trust (β = .38, 95% CI [.25, .52], p < .001). The direct effect of justice on team performance (β = −.12, p = .145) was not significant, thus supporting the full mediation model. Bias-corrected CIs for all indirect effects did not include zero, ensuring the statistical robustness of the findings.
In summary, while alternative models provided reasonable fit, their theoretical coherence and empirical support were inferior to the proposed model M 1, which emerged as the most parsimonious, theoretically grounded, and empirically supported representation of the relationships under study.
Additionally, organizational justice explains the 66% of the variance in horizontal trust, R2 = .66 and .88 of the variance in collective resilience and, in turn, horizontal trust and collective resilience explain the 93% of the variance in team-level job performance, R2 = .93 (R2 intra-role performance = .91, R2 extra-role performance = .91). The results of this model are graphically presented in Figure 2.
Final model: Structural model of organizational justice, horizontal trust, collective resilience and job performance (N = 100 work teams).
Note: ***p < .001; ns = nonsignificant.
Discussion
This study aimed to analyze the mediating role of horizontal trust and collective resilience between organizational justice in each of its dimensions (i.e. distributive, procedural, informational, and interpersonal justice) and job performance in its two dimensions (i.e. intra-role and extra-role) at the work team level based on the HERO model (Salanova et al., Reference Salanova, Llorens, Cifre and Martínez2012).
The present study provides strong support for the HERO model (Salanova et al., Reference Salanova, Llorens, Cifre and Martínez2012, Reference Salanova, Llorens and Martínez2019) by confirming its three components: (a) healthy organizational practices and resources (i.e. organizational justice), (b) healthy employees (i.e. horizontal trust and collective resilience), and (c) healthy organizational outcomes (i.e. job performance) through aggregated data at the work team level. However, this study goes further by providing a more nuanced understanding of how these variables interact at the team level, highlighting the integral mediation of horizontal trust and collective resilience between organizational justice and job performance, and the importance of these elements in complex organizational contexts such as the healthcare sector.
The results indicated that organizational justice (distributive, procedural, and interactional dimensions) significantly influences job performance in healthcare teams. Combining the four dimensions of organizational justice into two (distributive-procedural justice and informational-interpersonal justice) showed a better model fit, which is consistent with the literature suggesting flexibility in measuring these dimensions (Dai & Xie, Reference Dai and Xie2016). This finding implies that perceptions of work justice enhance teams’ job performance (intra-role and extra-role).
Previous studies have demonstrated the mediating role of organizational trust between organizational practices and job performance (Acosta et al., Reference Acosta, Salanova and Llorens2011; Costa & Anderson, Reference Costa and Anderson2011; Olvera et al., Reference Olvera, Llorens, Acosta and Salanova2017; Palanski et al., Reference Palanski, Kahai and Yammarino2011; Tan & Lim, Reference Tan and Lim2009). The present investigation extends these findings by showing that horizontal trust also mediates the relationship between organizational justice and collective resilience, underscoring the importance of building horizontal trust to strengthen resilience and improve the work performance of healthcare teams. Furthermore, horizontal trust demonstrated a significant mediating effect between organizational justice and collective resilience, underscoring its critical role as a catalyst for healthy dynamics within health teams. Previous literature has already established the relationship between organizational justice and employee well-being (Omar, Reference Omar, Littlewoord and Vega2015), but the present study extends this evidence by showing how horizontal trust enhances resilience, a key resource for coping with the challenges of a healthcare environment characterized by high uncertainty and constant demands (Britt et al., Reference Britt, Shen, Sinclair, Grossman and Klieger2016).
On the other hand, structural analysis confirmed that both horizontal trust and collective resilience fully mediate the relationship between organizational justice and job performance. This highlights how perceptions of organizational justice not only directly influence job performance but also operate through collective empowerment mechanisms.
Compared to the alternative models M A1 and M A2 evaluated, the proposed model not only showed a better fit but also greater parsimony, as evidenced by a lower AIC. This reaffirms that the combination of justice, trust, and resilience is a solid theoretical framework for understanding and improving health team performance (intra-role and extra-role performance), where these elements have an important impact on organizational outcomes such as quality of care (Rego et al., Reference Rego, Lopes and Nascimento2016).
In summary, this study reinforces the importance of organizational justice, horizontal trust, and collective resilience as fundamental pillars of teams’ job performance in the health sector. The results not only strengthen the HERO model but also provide empirical evidence for designing organizational strategies to improve both worker well-being and patient care outcomes.
Theoretical implications
Results of the present study have significant theoretical implications that broaden and deepen the understanding of organizational dynamics and workplace behavior. Social exchange theory (Blau, Reference Blau1964) posits that social interactions are based on a system of reciprocity, where individuals’ actions are influenced by the rewards they expect to receive in return. This study confirms that the perception of organizational justice acts as a valuable resource in social exchange, fostering horizontal trust and collective resilience within work teams. Interactional justice, when perceived as fair, strengthens trust among team members, creating a more collaborative and less conflictual work environment (Cropanzano & Mitchell, Reference Cropanzano and Mitchell2005).
Equity theory (Adams, Reference Adams1966) posits that individuals evaluate the fairness of social exchanges by comparing the rewards they receive with their contributions and those of others. This study extends this theory to the organizational context, demonstrating that distributive and procedural justice are central to the development of horizontal trust and collective resilience. Procedural justice, which relates to the fairness of organizational processes and procedures, and distributive justice, which refers to fairness in the distribution of resources and rewards, are essential for fostering a shared perception of fairness that promotes collective resilience. This collective resilience, in turn, enhances the team’s ability to adapt and maintain high performance in adverse situations (Lengnick-Hall et al., Reference Lengnick-Hall, Beck and Lengnick-Hall2011).
The HERO model (Salanova et al., Reference Salanova, Llorens, Cifre and Martínez2012) proposes that healthy organizational practices and resources, healthy employees, and healthy organizational outcomes are interrelated. Our study provides empirical evidence supporting this interrelationship, highlighting how organizational justice (a healthy organizational resource) influences horizontal trust and collective resilience (healthy employees), which in turn improves job performance (healthy organizational outcomes). This finding is crucial for understanding how organizational interventions can be designed to improve the well-being and performance of work teams.
Furthermore, the results of this study underline the mediating role of horizontal trust and collective resilience between organizational justice and team job performance. This means that to maximize the positive impact of organizational justice on job performance, organizations must simultaneously foster horizontal trust and collective resilience. This finding refines existing theoretical frameworks by demonstrating that trust and resilience are not only outcomes of fair perception but also act as mechanisms through which organizational justice enhances team performance.
Finally, the present study also contributes to organizational resilience theory by demonstrating that collective resilience can be cultivated through fair and trustworthy workplace practices. Organizational resilience refers to an organization’s ability to anticipate, prepare for, respond to, and adapt to unexpected and disruptive events (Sutcliffe & Vogus, Reference Sutcliffe, Vogus, Cameron, Dutton and Quinn2003). The findings suggest that organizations implementing fair practices and fostering trust can develop more resilient teams capable of maintaining high performance even under adverse conditions.
Practical implications
The present study provides valuable evidence for the management of healthcare organizations, highlighting the importance of implementing strategies that promote the perception of organizational justice among workers and work teams.
Organizations should, therefore, design and implement policies that ensure distributive, procedural, and interactional justice. This may include creating clear and transparent processes for decision-making, equitable distribution of resources and rewards, and promoting open and honest communication between team members and management. Perceptions of organizational justice not only improve job satisfaction but also foster horizontal trust and collective resilience, which in turn improves the job performance of health teams.
Trust between team members (i.e., horizontal trust) is crucial for job performance and collective resilience. Organizations should promote interventions that build and maintain horizontal trust. These efforts can help strengthen bonds between team members, increasing their ability to work together efficiently and cope with challenges.
In addition, collective resilience is essential for teams to adapt and thrive in adverse situations. Organizations should provide resources and support that foster resilience, such as access to wellness programs, opportunities for professional and personal development, and the establishment of an organizational culture that values and rewards adaptability and innovation. Fostering collective resilience not only improves the team’s ability to overcome difficulties but also contributes to a more positive and productive working environment.
Implementing interventions based on the HERO model can be an effective strategy for improving the well-being and performance of work teams. These interventions can include promoting healthy organizational practices, developing programs that foster employee well-being, and creating a work environment that supports both individual and collective development and growth (Salanova et al., Reference Salanova, Llorens, Cifre and Martínez2012). These interventions can help build a strong and resilient organizational culture capable of facing and overcoming the challenges of the healthcare environment.
Given that the healthcare context has specific characteristics, including a high-pressure, dynamic, and often unpredictable environment, it is essential that management strategies and interventions are specifically tailored to this context. Policies need to consider high workload, emotional stress, and the need for clear and effective communication in critical situations. Implementing psychological support programs, fostering a culture of mutual care, and providing resources for stress management can be particularly beneficial.
Limitations and Future Research
This study has several limitations that should be considered when interpreting the results. First, the cross-sectional design of the study precludes establishing causal relationships between the variables analyzed. Future research could use longitudinal designs to explore the evolution of these relationships over time and their causal nature.
Second, although the data were collected using self-report measures, which could introduce response bias, in this study, they were analyzed at the team level, thus increasing the validity of the scores obtained. In addition, Harman’s single-factor test indicated that the common variance bias did not significantly affect the study data. Complementarily, the inclusion of a CMF in the structural equation model confirmed that methodological bias did not significantly influence the relationships between the latent variables, ensuring the validity of the results obtained.
Third, the use of a convenience sample may limit the generalizability of the results. However, the diversity of healthcare organizations included in this study provides a broader and more robust representation of the healthcare context.
Another significant limitation is the presence of moderate multicollinearity among some predictor variables, as evidenced by the high VIF values (Kline, Reference Kline2015). This was addressed through techniques such as the creation of composite indices and the application of Ridge regression.
In terms of future research, this study highlights the importance of adopting multilevel approaches to capture the complex interactions between different organizational levels in healthcare institutions.
For future research, given the characteristics of healthcare organizations, it is advisable to consider studies that incorporate different levels of analysis and adopt a multilevel perspective. In addition, the design of longitudinal studies is essential to expand knowledge about the role of horizontal trust and collective resilience in an organization as complex as healthcare. These approaches will deepen our understanding of the temporal and structural dynamics that influence work performance and team well-being in this context, highlighting the importance of trust in healthcare organizations.
Finally, in a healthcare context characterized by high levels of complexity and pressure, horizontal trust and collective resilience emerge as key elements for improving the quality of organizational outcomes. This study paves the way for further research into the development of interventions aimed at reinforcing these dynamics in healthcare teams, with the aim of improving both the well-being of workers and the quality of patient care.
Data availability statement
Data are available upon request due to privacy and ethical considerations.
Author contribution
JO: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Software, Validation, Writing – original draft, Writing – review & editing. SL: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Funding acquisition, Project administration, Resources, Supervision, Writing – review & editing. HA: Conceptualization, Investigation, Methodology, Supervision, Writing – review & editing. MS: Funding acquisition, Project administration, Resources, Supervision, Writing – review & editing.
Funding statement
This research was funded by the Spanish Ministerio de Ciencia e Innovación (MCIN/ AEI/10.13039/501100011033, project #PID2020-119993RB-I00).
Competing interest
The authors declare no conflict of interest. The funders had no role in the design of the study, the collection, analyses, or interpretation of data, the writing of the manuscript, or the decision to publish the results.
Open access
