In many countries, the COVID-19 pandemic catalyzed nurses’ migration into temporary employment arrangements (EA). The supply of temporary nurses first mirrored demands due to COVID-19 incidence rates and then plateaued at a level more than twice as high as before the pandemic (e.g., Hansen & Tuttas, 2022; swissstaffing, 2023). These nurses responded to market signals by traveling to regions with high demand and high wages (Gottlieb & Zenilman, 2022). Confronted with their temporary counterparts, local core staff often felt the “deal” they received to be much less favorable (e.g., Gan, 2022).

Labor legislation prescribes equal treatment of all employees (e.g., non-discrimination clauses, the principle of “same work, same pay”). Such principles notwithstanding, contractual terms and conditions of temporary employees hired via staffing agencies can differ from those of core staff assuming the same tasks. Often, these differences disadvantage temporary employees, but when labor markets are tight, as is the case in healthcare, it is typically the other way around (Houseman et al., 2003). Regardless of who is better off, introducing alternative EA with distinct rewards reinforces social comparisons for all employees (Chattopahyay & George, 2001), which may entail systematic differences in perceptions of pay unfairness and the underlying evaluation processes.

These processes are complex because pay is not simply a means to acquire goods and services but also carries symbolic value, which is of specific importance, in the context of this study. Pay signals “achievement, recognition, and potential enhancement of self-esteem” (Goodman, 1974, p. 171) and an employee can only gauge the value of these symbolic factors by undergoing a complex evaluation process. To gain more nuanced understanding of the perception of and reactions to pay unfairness, it is therefore important to know which information employees rely on, in this process. The distributive-justice literature mainly focused on the exchange relationship between employee and employer and the related standard that employees can use to determine how fairly they are treated, namely, the psychological contract (Cropanzano & Prehar, 2001). It is unlikely, however, that employees exclusively rely on one source of information (e.g., van den Bos & Lind, 2002). Hence, scholars with a broader institutional perspective (e.g., Blau, 1994; Goodman, 1974) introduced more comprehensive taxonomies distinguishing entities used as bases for comparing, i.e., reference standards or, in short, referents.

In particular, Goodman (1974) differentiated three classes of referents, i.e., system referents, other-referents, and self-referents with specific examples, each. Informed by Goodman’s (1974) referent classes, the present study will examine how three self-referents - linked to recognition and self-esteem - influence the perception of and reactions to pay unfairness in employees with distinct employment arrangements. We thereby extend the literature on distributive justice in several ways. First, we will provide new insights into how employees determine the worth they have for their employer and the fairness of their pay by introducing three internal self-referents that are highly relevant in health care. The first referent is perceived responsibility, i.e., the degree to which an individual is held accountable. The other referents are two types of social appreciation (from colleagues and patients). Second, in addition to the formation of unfairness perceptions per se on which previous studies centered (e.g., Eisnecker & Adriaans, 2023), we will also examine referents’ role regarding reactions to perceived unfairness. For that purpose, we will integrate the Stress-as-Offence-to-Self theory (Semmer et al., 2019) into the perspective of distributive justice. Focusing on implicit messages of recognition or respect referents may send, we will introduce them as boundary conditions of reactions to perceived pay unfairness since negative information about an employee’s worth is a threat to their self-esteem. To complement bottom-line-oriented outcomes (e.g., withdrawal behaviors) that are dominant in the management literature, we use more proximal and employee-oriented outcomes (exhaustion and job satisfaction) as key indicators of employee well-being (e.g., de Jonge & Schaufeli, 1998). Finally, we will distinguish permanent and temporary EA as contexts that differ with respect to the salience of different types of self-referents, and we will explore for whom the use of which referent most strongly affects perceived pay unfairness.

We chose registered nurses as an exemplary profession for this comparison. They represent employees with skills that are in high demand, and they typically choose temporary employment voluntarily because of their high levels of employability (see research context, below). Sharing similar levels of education as well as occupation-specific labor-market and task characteristics, this homogenous group is ideal for our comparisons, and we will investigate perceptions of and strain reactions to pay unfairness within it. Specifically, we will examine to which degree perceived responsibility as well as appreciation from colleagues and patients affect unfairness perceptions and modulate reactions to them within and across the two EAs within which self-referents may be used differentially. In this fashion, we will assemble a more comprehensive and nuanced picture of the processes underlying perceptions of pay unfairness and reactions to them.

Theoretical Background

Reactions to Pay Unfairness

Equity theory (Adams, 1965; Homans, 1961) is a social-exchange theory widely used by organizational-justice scholars to explain differences in consequences of pay unfairness for employees, including sleep quality and exhaustion (e.g., Greenberg, 2006; Van Dierendonck et al., 2001), job satisfaction (Lam et al., 2002), extra-role (Colquitt et al., 2001) as well as several withdrawal behaviors (e.g., Rubenstein et al., 2018). The relative equality rule posits that an exchange is perceived as fair if the relative distributions of output (rewards such as pay, benefits, job security) and input (contributions such as time, effort, skills) between a target individual and a referent (in an equity-theoretical perspective primarily “generalized others”) are equal. If the rule is violated through unequal distributions, psychological and physical strain should arise, and attitudes towards work should become more negative for all parties involved in the exchange. Since the number of hours worked is a central element in the mental processes underlying equity perceptions and may also influence strain (e.g., (O’Driscoll & Roche, 2016), weekly work hours should be held constant when analyzing the effects of pay unfairness.

Expectations regarding compensation can differ among individuals performing similar activities within the same work unit (e.g., Dabos & Rousseau, 2004), and we argue that EA provide a lens to explain these differences (e.g., De Cuyper et al., 2008). Temporary nurses’ limited future time perspective within a unit (Polivka, 1996) may result in their placing a larger discount on returns receivable in the future than staff nurses with an open-ended time perspective. Since relational outcomes typically take more time to develop, psychological contracts of employees with different EA differ in breadth with those of temporary employees more likely to be solely transactional in nature (i.e., focusing on simple, measurable economic outcomes, e.g., Connelly & Gallagher, 2004; Peel & Boxall, 2005). Consequently, a discrepancy between what is received and what was promised is more salient for them, at any point in time. There is, indeed, evidence showing that temporary employees are more responsive to employers’ contributions than permanent employees (e.g., Coyle-Shapiro & Kessler, 2002) whose psychological contracts have longer time perspectives and typically are both transactional and relational in nature.

Summing up, we expect perceived pay unfairness to be positively related to affective strain, as indicated by exhaustion, and negatively to work-related attitudes, as indicated by job satisfaction. Moreover, the effects of pay unfairness should be stronger in temporary as compared to staff nurses. Since the hours temporary and staff nurses put in may differ and work hours affect strain, the overall activity level should be controlled.

Hypothesis 1

Pay unfairness is positively associated with exhaustion (H1A) and negatively associated with job satisfaction (H1B). The effects of pay unfairness should be stronger in temporary as opposed to staff nurses (exhaustion: H1C, job satisfaction: H1D).

The picture may be more complex, however, than the bulk of evidence from studies guided by equity theory suggests. To refine the approach, we will look more closely at additional assumptions that have received less attention in the organizational-justice literature. We will examine three referents as both predictors of pay unfairness and boundary conditions of its effects. Referents have an important role since, from an equity theory perspective, an imbalance between contributions and rewards per se is not deemed to be stressful. Instead, an imbalance should lead to strain only if the target individual detects a difference relative to a referent via comparison processes (Adams, 1965; see also Blau, 1964; Crosby, 1976; for similar arguments). Therefore, our aim is to determine for which group of nurses the use of which referent most strongly affects the perception of pay unfairness. Considering the differential selection of referents can yield more nuanced insights regarding the fairness process. It helps us to understand who feels treated unfairly, based on which information, and this knowledge can be used to deduce interventions regarding the well-being and retention of nurses and other professions with highly demanded skills.

Extending the distributive-justice perspective, we argue that nurses can use referents interchangeably or in conjunction as a frame of reference in dynamically evaluating their pay. Goodman (1974) proposed three major classes of referents. In addition to other-referents (referring to the input-output ratio of other individuals within or outside the focal person’s organization) and system referents such as the well-investigated psychological contract, he introduced the class of self-referents including pay history or the capacity to provide for one’s family. Internal self-referents are a third type of self-referent representing an employee’s individual conception of their worth (to their employer) that is closely linked to self-esteem (Goodman, 1974).

These internal standards can be shaped by external factors, such as education and occupation (Goodman, 1974). Since registered nurses are a homogenous group regarding occupation and largely education, we argue that other factors are more important in shaping their view of their worth. The first factor we propose is perceived responsibility. The higher the responsibility the higher an employee’s worth to their employer and this is why responsibility is aligned with salary, in many organizations (Manove, 1997). The two other factors we introduce are colleagues’ and patients’ appreciation. Both play an important role in the worth an employee perceives to possess for their employer (e.g., Semmer et al., 2016).

Which combination of referents a nurse selects is a function of the specific context in which they act (e.g., Goodman, 1977). As a context, we believe employment arrangements to be a seminal choice. While individuals largely perform the same tasks across EAs, they receive different levels of rewards. Moreover, with their differing psychological contracts (Coyle-Shapiro et al., 2019), the fairness process may not be identical for temporary and staff nurses. We argue that this is a function of their differential use of internal self-referents, which may not possess the same salience for both groups. In the following, we will delineate the effects of the referents introduced above, perceived responsibility, colleagues’ and patients’ appreciation. For each of them, we will explain why we hypothesize differential effects for temporary vs. staff nurses.

The Role of Perceived Responsibility for Pay Unfairness

Nurses’ actions significantly impact hospital performance. For instance, they contribute to healthcare quality (Oldland et al., 2020) by preventing adverse events (Montalvo, 2007). More generally, the extent to which an individual’s effort (or inattentiveness, errors) can prevent (lead to) serious damage or the degree to which an individual is held accountable for the consequences of their actions or decisions is called responsibility (Jackson et al., 1993). In most organizations, responsibility is aligned with wages (Manove, 1997; for a more general consideration of principal-agent relationships, see Holmstrom, 1979). Therefore, we argue that nurses can use the extent of the responsibility they perceive to have to determine their worth to the employer. The higher a nurse perceives their level of responsibility, the more exigent they should be when judging pay fairness. Since there is no disrespect (Semmer et al., 2016) implied in perceived responsibility, we do not expect it to accentuate effects of pay unfairness.

There should be, however, differential effects for temporary and staff nurses. Temporary nurses have been described as performing a “relatively narrow set of defined duties” (Lien, 2023) that may be monitored more easily and as having a “limited necessity to take on full responsibility” (Bajorek & Guest, 2019). More importantly, due to the short-term nature of their assignments, they may not have to face the full consequences of mistakes they may make – ranging from a bad reputation with another department to patient harm and resulting regulator visits. At the same time, the level of responsibility of permanent staff was shown to increase when they work alongside temporary employees, and the former may even be held responsible for errors committed by the latter (e.g., Pearce, 1993). Hence, perceived responsibility should be more salient and, therefore, weigh more strongly in fairness perceptions for staff nurses as compared to temporary nurses. Again, the number of hours worked is the central element in the mental processes underlying fairness perceptions and may also influence exhaustion. Hence, the weekly work hours should be controlled.

Hypothesis 2

Perceived responsibility is positively related to pay unfairness (H2a). The effects of responsibility should be stronger in staff than in temporary nurses (H2b).

The Role of Colleagues’ Appreciation for Perceiving and Reacting to Pay Unfairness

To determine whether they are paid fairly, nurses can also rely on social appreciation as an internal self-referent. Social appreciation reflects nurses’ worth to their employer and can be expressed by different stakeholders; one example is the public that applauded nurses for their efforts during the pandemic. Appreciation expressed by colleagues is a particularly important type of appreciation since it signals peer acceptance, involves recognition of a nurse’s qualities and achievements (Semmer et al., 2019). It is directly related to self-esteem, well-being, and indicators of physical health (e.g., Auer et al., 2024; Stocker et al., 2010) and has additionally been conceptualized as protective resource (e.g., Stocker et al., 2019). In contrast to this notion, we argue that when the symbolic currency is recognition or worth – both closely linked to the self – appreciation may accentuate the stress of feeling underpaid. Nurses should factor the positive social message sent via colleagues’ recognition of their qualities into the formation of fairness perceptions, on the input side and this increases the weight of their inputs and should, therefore, augment the level of output considered fair.

Low levels of pay implicitly signal that an employer deems an employee to be of low value. Negative information about an employee’s worth is a threat to their self-esteem, and this can be a stressful experience for them (e.g., Semmer et al., 2019). Therefore, we suggest that – beyond its main effect on fairness perceptions – coworker appreciation leads to stronger reactions to pay unfairness because it emphasizes the discrepancy and the offense of not being rewarded adequately (see Semmer et al., 2016, for a similar argument). Hence, nurses with higher levels of appreciation should react more strongly to perceived pay unfairness than nurses with lower levels.

Based on uncertainty-management theory (van den Bos & Lind, 2002), we additionally propose differential effects of appreciation in temporary and staff nurses. Evidence reported by Goodman (1974) did not support the assumption that employees occupying an inside role in the organization rely more on “other-inside” referents than employees with a boundary role. We argue, however, that temporary nurses who can gauge their worth in terms of their being in great demand on the labor market and, at the same time, have more insights into contributions and rewards of other nurses in a range of hospitals (“outside other-referents” according to Goodman, 1974) should use appreciation by colleagues less to determine their worth. Staff nurses, in contrast, should rely more on this heuristic substitute because of the higher level of uncertainty they may experience regarding their worth for the hospital (van den Bos et al., 1997). Again, the number of hours worked is the central quantitative element in the mental processes underlying equity perceptions and may also influence exhaustion. Hence, the weekly work hours should be controlled.

Hypothesis 3

Perceived appreciation from colleagues is positively related to pay unfairness (H3a). The effect of colleagues’ appreciation on pay unfairness should be stronger in staff than in temporary nurses (H3b). For nurses who perceive more (less) appreciation from their colleagues, the effects of pay unfairness on exhaustion (H3c) and job satisfaction (H3d) are stronger (weaker).

The Role of Patients’ Appreciation for Perceiving and Reacting to Pay Unfairness

Next to colleagues, patients are a further important source of appreciation. We argue that the basic process of factoring the positive social message sent by patients who appreciate nurses’ qualities into the formation of pay-unfairness perceptions resembles that described above for colleagues’ appreciation. Hence, this referent should make a nurse more exigent in what they consider as fair compensation. Beyond its main effect on fairness perceptions, patient appreciation may also lead to stronger reactions to pay unfairness because it emphasizes the offense of being insufficiently valued which threatens self-esteem (Semmer et al., 2016). Hence, nurses with higher levels of patient appreciation should react more strongly to pay unfairness.

As for colleagues’ appreciation, we propose differential effects for temporary and staff nurses, but they should deviate from those described for colleagues’ appreciation. Since temporary nurses are often less respected, appreciated, and integrated in their units than staff nurses (Viitala & Kantola, 2016), they should rely more on patient appreciation than staff nurses to assess their worth. Again, the number of hours worked is the central quantitative element in the mental processes underlying fairness perceptions and may also influence exhaustion. Hence, the weekly work hours should be controlled.

Hypothesis 4

Perceived appreciation from patients is positively related to pay unfairness (H4a). The effect of patients’ appreciation on pay unfairness should be stronger in temporary as opposed to staff nurses (H4b). For nurses who perceive more (less) appreciation, the effects of pay unfairness on exhaustion (H4c) and job satisfaction (H4d) are stronger (weaker).

Data and Methods

Research Context

In our study, we distinguish staff nurses with standard permanent contracts from nurses with different kinds of temporary contracts (see Gan, 2020; Mattmann et al., 2017; Schmelzer et al., 2021). Staff nurses have open-ended employment contracts at the same hospital and typically stay in the same unit. Due to their higher tenure and identification with the organization, this group is characterized by high perceived responsibility (Parker et al., 2002). Temporary nurses have classic fixed-term contracts with staffing agencies by which they are placed in hospitals for longer-term assignments (typically one to three months or more).

Contributing a small percentage of the hours nurses work in Swiss hospitals (Lampart & Bühler, 2019), there is also a small but fast-growing group of nurses with on-demand work arrangements which were included in the group of temporary nurses. They have contracts with staffing agencies or platforms through which they are placed in different hospitals, typically for short-term assignments (“gigs”). From the perspective of the in-group, on-demand nurses are often stereotyped as lacking competence (Fiske et al., 2002). Educated equally well, they may be less respected, appreciated, and integrated in their units than staff nurses (Viitala & Kantola, 2016). Moreover, being paid better than their permanent counterparts, they tend to be resented for their “privileged” conditions (e.g., Manias et al., 2003).

Because nursing skills are highly demanded, temporary nurses who prefer a permanent position are likely to find one. In additional analyses, not a single temporary nurse reported an inability to find a permanent position as a reason for their temporary EA and perceived employability was high in both groups (M = 4.5, SD = 0.8 for permanent, M = 4.4, SD = 1.1 for temporary nurses on a scale ranging from 1 to 5). Therefore, we conclude that nurses with a temporary EA in our sample chose it voluntarily (Krausz et al., 1995; see also Hom & Griffeth, 1991).

Participants and Procedure

In the German-speaking part of Switzerland, we recruited registered nurses through intermediaries (hospitals, staffing agencies, professional associations) and directly (in-person distribution of flyers in front of hospitals, social media), between spring 2021 and fall 2022, i.e., in the middle of the pandemic. For the recruitment through intermediaries, we prepared various informational materials to fit different communication channels, including blog posts, newsletters, and social media posts. In the case of temporary nurses, we were able to send a recruitment message to all the nurses placed by the largest staffing agency in the healthcare sector in the German-speaking part of Switzerland to invite them to participate. To incentivize nurses to participate, we used two strategies: First, feedback on study results, and second, a chance to win one of ten vouchers (each worth 200 Swiss francs). Upon following the link to access the online survey, nurses were presented with detailed information about the study and the conditions of participation. If they declared their informed consent, nurses proceeded to participate in the initial assessment. Here, we enquired about socio-demographic and work-related information and presented different questionnaires regarding work- and employment-related demands and resources. At the end of the first online survey, we provided a link to another survey platform to collect nurses’ contact information. Using separate survey platforms guaranteed that participants’ responses could not be linked to sensitive personal information. After two weeks, we mailed participants a link to access the second part of the assessment regarding work-related outcomes. We sent an additional reminder seven days after the first email. The local IRB has approved the project (protocol number 2021 − 706).

A total of 262 staff nurses and 120 temporary nurses (55 travel, and 59 per-diem nurses working on demand, six did not provide this information) completed the initial as well as the follow-up assessment (60% of initial participants). The average age of the participants was 41.9 years (SD = 11.4, range = 20–65), and 85.1% were female. Regarding educational levels, the largest group were registered nurses (vocational training, 46.5%), and another large group (30.3%) were registered nurses with postgraduate specialization, e.g., intensive care. The mean work experience was 19.5 years (SD = 11.0, range = 1–45). Nearly two thirds of the participants (63.6%) were married or cohabitating. The average number of children was 0.9 (SD = 1.2, range = 0–4). Assuming strong measurement invariance, we tested via multigroup models if participants who only completed t1 differed from those who also finished t2 regarding the latent constructs or manifest variables reported in this paper. This was neither the case for the means, Δχ²(6) = 11.1, p = .08, nor for the variances, Δχ²(6) = 3.7, p = .72.

Measures in Initial Assessment

Type of work Arrangement

Depending on the type of EA nurses held (employed permanently vs. temporarily), they chose different links leading to the respective set of items. In addition, each group provided specific information on the scheduling flexibility of their EA to distinguish between classic temporary nurses and nurses who work on-demand.

Perceived Pay Unfairness

We used the four injustice items of the full range scale of distributive justice (Colquitt et al., 2015) to assess nurses’ evaluation of pay unfairness. The item stem referred to the nursing director for staff nurses, and to the agency for temporary nurses. A sample item is “Is your compensation inconsistent with the effort you have put into your work?”. The Likert scale ranged from 1 ‘not at all’ to 5 ‘very much’ (Cronbach’s Alpha .90).

Colleagues’ Appreciation

Perceived appreciation was measured with the 5-item subscale of the Bern Appreciation at Work scale that refers to colleagues’ appreciation (Stocker et al., 2014). Responses indicated the extent to which each item applied to the participants’ work situation: “My colleagues ask me for advice and that shows me that they value my opinion.” Deviating from the original 7-point scale, the Likert scale only ranged from 1 ‘not at all’ to 5 ‘very much’ (Cronbach’s Alpha 0.82).

Patients’ Appreciation

Perceived appreciation by patients was measured with the adapted 5-item subscale of the Bern Appreciation at Work scale on client appreciation (Stocker et al., 2014). Responses indicated the extent to which each item applied to the participants’ work situation: “I get compliments from my patients about my work.” Deviating from the original 7-point scale, the Likert scale only ranged from 1 ‘not at all’ to 5 ‘very much’ (Cronbach’s Alpha 0.81).

Perceived Responsibility

We used a three-item scale to assess perceived responsibility (Jackson et al., 1993). Participants indicated to which degree a statement applied to them. A sample item is “A mistake on my part can cause great harm.” The five-point scale ranged from “not at all” to “very much” (Cronbach’s Alpha 0.79).

Control Variable: Weekly work Hours

Employment below full-time can be seen as another alternative EA (e.g., Spreitzer et al., 2017), but at the same time, different contractual arrangements can be freely combined with varying numbers of work hours. Hence, in each EA, nurses can work full-time or different levels of part-time, resulting in nearly all possible combinations. Working more hours leaves less time for recovery (e.g., Sonnentag et al., 2010) and results in more exposure to work-related demands (e.g., Sparks et al., 1997), as well, so they should accentuate the effects of working conditions in a nurse’s employment exchange. For staff nurses, a single item assessed contractual working hours in 5 ranges (1-19%, 20-39%, 40-59%, 60-79%, 80-100%). Hence, each unit translates to 20% of a full-time equivalent. Temporary nurses reported their average weekly working hours.

Outcomes Measured Two Weeks After the Initial Assessment

Exhaustion

In the second assessment, we used items 1, 5, and 16, representing the general-fatigue scale of the Multidimensional Fatigue Inventory (Smets et al., 1995) to assess the level of exhaustion. A sample item is “I feel tired.”. The Likert scale ranged from 1 ‘not at all’ to 5 ‘very much’ (Cronbach’s Alpha 0.80).

Affective Job Satisfaction

In the second assessment, we used the four items of Thompson and Phua’s (2012) brief index to assess affective job satisfaction. A sample item is “I find real enjoyment in my job.” The Likert scale ranged from 1 ‘not at all’ to 5 ‘very much’ (Cronbach’s Alpha 0.78).

Analytic Strategy

All analyses were conducted with Mplus 8.9 (L. K. Muthén & Muthén, 2017).

Measurement Invariance

The analysis started by testing all latent constructs for measurement invariance to ensure comparability across staff and temporary nurses. The corresponding analyses were conducted within a multigroup confirmatory factor analysis (CFA) framework. The CFA treated all item responses as ordered categories and used the weighted least squares algorithm WLSMV (Muthén et al., 1997) for model estimation. The resulting model parametrization is equivalent to the item response theory’s graded-response model (Samejima, 1969). To identify the models, the factors in the group of the staff nurses were fixed to mean m = 0 and variance s² = 1. The assessment started with a configural invariance model, progressively constrained to test scalar (equal factor loadings) and metric (equal item thresholds) measurement invariance. Given adequate fit (CFI ≥ 0.90 and RMSEA ≤ 0.08; see McDonald & Ho, 2002) of the configural invariance model, tests for higher levels of invariance consisted of stepwise fixation of parameter classes across groups (Millsap & Yun-Tein, 2004). The change in model fit was evaluated with χ²difference tests using the MPlus function DIFFTEST and with the change in CFI (Chen, 2007). Changes in the invariance level with less than or equal to ΔCFI = − 0.01 and p < .01 were considered acceptable. Confronted with significant differences on any level, we would have aimed to achieve partial invariance by removing parameter constraints on the smallest set of items required for the fit difference to be within acceptable limits. To facilitate model stability and reduce the number of free parameters, adjacent rare response categories (< 10 in any group) were combined into a single category.

Main Analytic Strategy

For testing H1 and H[2–4][a-b], the multigroup CFA from the previous step was limited to the constructs in the respective hypothesis and extended by the relevant paths. Specifically, by two regressions per model, one being the hypothesized effect and the other the control of the endogenous constructs by weekly hours.

For all other hypotheses (H[2–4][c-d]), we used single level random slope models (depicted in Fig. 1; Hildreth & Houck, 1968; Johnston, 1984). The models shared a common structure, defining the effect of pay unfairness on either exhaustion or job satisfaction as a random slope, which is regressed on by a third, thereby moderating, construct. Again, all endogenous variables (exhaustion, job satisfaction, and random slope) were controlled for weekly work hours.

Fig. 1
figure 1

Schematic illustration of the moderating effects of referents as boundary conditions: Random slope models for testing the hypotheses H[2–4][c-d]

Full size image

Results

Measurement Invariance

The configural invariance model yielded a good fit (CFI = 0.972 and RMSEA = 0.040). Restricting it to metric invariance did not significantly worsen the fit, ΔCFI = + 0.007 and Δχ²(18) = 25.8, p = .10. The next restriction step to scalar invariance, again, indicated no significant differences between groups, ΔCFI = -0.002 and Δχ²(54) = 83.4, p = .014. This final measurement model fitted the data well (CFI = 0.977 and RMSEA = 0.039). This scalar measurement model was the base for the subsequent analyses. In those, due to the identification approach (s² = 1) and measurement invariance, all paths (b) are standardized to the variances in the group of staff nurses.

Descriptive Information

Table 1 provides information about the means, standard deviations, and intercorrelations of study variables. In temporary nurses, pay unfairness was lower than in staff nurses, but it correlated more strongly with job satisfaction. Moreover, temporary nurses perceived slightly less appreciation by colleagues as well as slightly more appreciation by patients than staff nurses. Still, they reported the more appreciation from colleagues, the more weekly hours they worked. The correlations between both types of appreciation and job satisfaction, as well as exhaustion, were higher in staff nurses as compared to temporary nurses. While perceived responsibility was associated with more exhaustion and less job satisfaction in staff nurses, it was associated with more job satisfaction and with more perceived appreciation by colleagues and patients in temporary nurses. Finally, staff nurses reported to work around 6 hours per week more than temporary nurses.

Table 1 Descriptive statistics and correlations for study variables per employment arrangement
Full size table

Reactions to Pay Unfairness

According to the first hypothesis, pay unfairness should be related to both strain and work-related attitudes, as indicated by exhaustion and job satisfaction. With weekly work hours held constant, we found a positive effect of pay unfairness on exhaustion (b = 0.28, SE = 0.07, p(1-sided) < 0.001) and a negative effect on job satisfaction (b = -0.26, SE = 0.07, p(1-sided) < 0.001). This is in line with H1A and H1B. The effect for exhaustion did not differ between groups (staff: b = 0.25, SE = 0.08, p < .001; temporary: b = 0.35, SE = 0.14, p = .009; Δχ²(1) = 0.46, p(1-sided) = 0.25), which contradicts H1C. But with regard to job satisfaction, it was stronger for temporary than for staff nurses (staff: b = -0.19, SE = 0.07, p = .007; temporary: b = -0.34, SE = 0.09, p < .001; Δχ²(1) = 3.80, p(1-sided) = 0.03). There is, thus, support for H1D.

Perceived Responsibility and Pay Unfairness

In the second hypothesis, we proposed perceived responsibility to be associated with pay unfairness (H2a). The data supported this (b = 0.22, SE = 0.07, p(1-sided) < 0.001). The group comparison indicated a more differentiated mechanism, showing that this positive effect was only present in staff nurses (b = 0.30, SE = 0.08, p < .001) but not in temporary nurses (b = -0.04, SE = 0.10, p = .70). The groups hence differed (Δχ²(1) = 6.30, p(1-sided) = 0.006) in the expected direction, and thus H2b was supported by the data as well.

Colleagues’ Appreciation and Pay Unfairness

In the third hypothesis, we proposed perceived appreciation by colleagues to be positively associated with pay unfairness (H3a). This overall effect was not significant (b = 0.03, SE = 0.06, p(1-sided) = 0.30), and there was no difference (H3b) between staff and temporary nurses (Δχ²(1) = 1.69, p(1-sided) = 0.10). Beyond this effect, perceived appreciation was expected to accentuate the effects of pay unfairness on exhaustion (H3c) and job satisfaction (H3d, see Table 1). Neither the effect of pay unfairness on exhaustion (b = 0.02, SE = 0.06, p(1-sided) = 0.38) nor on job satisfaction (b = 0.02, SE = 0.08, p(1-sided) = 0.37) was moderated by colleagues’ appreciation. Hence, no H3 hypothesis was supported by the data.

In exploratory follow-up analyses (see Table 2 in Supplementary Materials) in which EA were distinguished, we found an interesting pattern of differences between EA, however. For staff nurses, there was a tendency for colleagues’ appreciation to accentuate the effects of pay unfairness on exhaustion and job satisfaction. For temporary nurses, in contrast, there was a tendency for colleagues’ appreciation to be negatively related to the effects of pay unfairness on exhaustion and job satisfaction.

Table 2 Referents as boundary conditions: Directional paths of the random slope models used to evaluate the hypotheses H[3-4][c-d]
Full size table

Patient Appreciation and Pay Unfairness

In the fourth hypothesis, we proposed perceived patient appreciation to be associated with pay unfairness (H4a). There was no overall effect (b = 0.04, SE = 0.06, p(1-sided) = 0.27), but the group comparison (H4b) revealed that there was just no effect in staff nurses (b = 0.06, SE = 0.08, p = 1), whereas, for temporary nurses, there was an unexpected negative effect (b = -0.22, SE = 0.11, p = .02). The effect was stronger for temporary nurses, but in the direction opposite of the hypothesized one (Δχ²(1) = 4.92, p = 1; p-2sided = 0.03). Beyond this finding, perceived appreciation was expected to accentuate the effects of pay unfairness on exhaustion (H4c) and job satisfaction (H4d, see Table 1). Neither the effect of pay unfairness on exhaustion (b = -0.05, SE = 0.06, p(1-sided) = 0.21) nor the one on job satisfaction (b = 0.02, SE = 0.06, p(1-sided) = 0.37) was moderated by patients’ appreciation. Hence, H4 was not supported by the data. In exploratory follow-up analyses (see Table 2 in Supplementary Materials), patient appreciation attenuated the effects of pay unfairness on exhaustion for temporary nurses.

Relative Importance of Referents

In our final and explorative analyses, we examined the relative importance of referents and the differences between staff and temporary nurses. Figure 2 plots the b-coefficients with 95-percent confidence intervals for both groups, respectively. In staff nurses, all the referents had more positive effects on pay unfairness, and the rank order differed between the two groups of nurses. The strongest predictor of pay unfairness for staff nurses was perceived responsibility, while the absolute effect of patients’ appreciation was largest in temporary nurses.

Fig. 2
figure 2

Relative importance of referents: b-coefficients for perceived pay unfairness with 95-per cent confidence intervals for both employment arrangements (EA). The b-coefficients differ statistically between EA for perceived responsibility (Δχ²(1) = 6.30; p = .01) and patients’ appreciation (Δχ²(1) = 4.92; p = .03), but not for colleagues’ appreciation (Δχ²(1) = 1.69; p = .19)

Full size image

Discussion

Informed by Goodman’s (1974) taxonomy of reference standards and integrating the Stress-as-Offence-to-Self theory (Semmer et al., 2019), our research added nuance to existing research by showing that the fairness process is influenced by context and, hence, more complex than currently acknowledged. We found pay unfairness to be more pronounced in staff nurses as compared to temporary nurses and we replicated the established effect of pay unfairness on exhaustion and job satisfaction. While our data show that temporary nurses’ level of pay unfairness is lower as compared to staff nurses, its effect on job satisfaction - but not on exhaustion - was stronger for them.

Of the proposed referents, we found that perceived responsibility is related to pay unfairness, albeit only in staff nurses. Hence, as expected, this referent proved to be more important for staff nurses than for temporary nurses. In the latter, the effect was negligible. Appreciation by colleagues and patients, however, had no effect on pay unfairness in staff nurses, while in temporary nurses, the direction of the effect of patient appreciation was opposed to the expected ones. Overall, the strength of pay-unfairness effects did not vary with any of the referents. We will discuss the detailed result patterns in the following.

Reactions to Pay Unfairness

Corroborating basic equity theory assumptions (e.g., Adams, 1965), our study found the expected positive relationship of pay unfairness with exhaustion and its negative relationship with job satisfaction. While in both groups of nurses the strain indicator exhaustion was similarly affected, the effect on job satisfaction was stronger in temporary as compared to staff nurses. Hence, job satisfaction is more responsive to pay unfairness in temporary nurses, and this is in line with evidence showing the threshold of reacting to contract breach to be lower for temporary employees since their predominantly transactional contracts are more limited in time and breadth than those of staff nurses (e.g., Schalk et al., 2010). While staff nurses may have reasons to be disappointed not only with regard to the transactional part of their employment exchange (e.g., their extra effort to help temporary colleagues may not be repaid) but also regarding hospitals’ concern for their long-term interests and well-being (e.g., Virtanen et al., 2010), in our study, their job satisfaction was less affected by perceived pay unfairness than that of temporary nurses.

The pattern of results we found differs from some of the evidence on temporary employees, and some of the explanations put forward in the literature may not hold for this highly skilled sample. For example, the argument that temporary employees justify advancing input with the aim of securing long-term employment (e.g., De Cuyper & De Witte, 2010) cannot be put forward here. The majority of temporary nurses do not aim to use temporary employment as stepping stone to permanent employment. Hence, it is plausible that we found temporary nurses to be sensitive to imbalances.

Perceived Responsibility and Pay Unfairness

While there was no overall effect of perceived responsibility on pay unfairness, data suggested that staff nurses used perceived responsibility as a referent in their fairness judgments. Generally, responsibility may be more salient for staff nurses, and this could explain why they rely on it to assess their worth. In addition, while they may not have as many insights into other hospitals as temporary nurses, they may compare themselves to individuals in their private context (e.g., Eisnecker & Adriaans, 2023) working in various professions with comparable levels of responsibility. Provided their wages are aligned with their levels of responsibility (Manove, 1997), these individuals are likely to earn more than staff nurses, leading to unfavorable comparisons.

Colleagues’ Appreciation and Pay Unfairness

As to the effects of perceived appreciation by colleagues, for neither group of nurses, this internal self-referent seemed to be factored in when perceiving equity. Also, regarding the effect of colleagues’ appreciation as a boundary condition, there was no overall effect. The tendency for colleagues’ appreciation to be positively related to the effects of pay unfairness on exhaustion and job satisfaction that emerged for staff nurses in our follow-up analyses could mean that they did perceive the indirect message of recognition signaled by colleagues’ appreciation. This message makes the offense of not being appropriately paid more salient for them, as implied by the Stress-as-Offence-to-Self theory (Semmer et al., 2019). For those temporary nurses who did perceive more appreciation from colleagues, in contrast, we found the effect of pay unfairness to be attenuated. This finding aligns with evidence that perceived appreciation partially offsets the negative effects of inequitable exchanges (e.g., Gordon et al., 2022; Klumb et al., 2006). These opposite tendencies in the two groups show that their use of information provided by referents is not uniform.

Patients’ Appreciation and Pay Unfairness

As expected, the EA groups differed regarding the internal self-referent “patient appreciation.” This referent affected perceptions of pay unfairness only for temporary nurses, but rather than stronger, they were weaker with higher levels of appreciation. Temporary nurses may not have used patient appreciation to gauge their worth but rather as an output substitute.

Regarding the effect of patients’ appreciation as a boundary condition, there was no overall effect. In exploratory follow-up analyses, we found a group difference regarding the moderation effects. In temporary nurses, patient appreciation attenuated – rather than accentuated – the effects of pay unfairness on exhaustion. In this group, patient appreciation represented a resource that buffered the effects of pay unfairness on exhaustion. This buffering effect is consistent with but more pronounced than the effect regarding colleagues’ appreciation reported above, and it is also consistent with evidence that perceived appreciation partially offsets the negative effects of inequitable exchanges (e.g., Klumb et al., 2006) as well as with evidence of attenuated effects of work interruptions on well-being (Stocker et al., 2019). Consistently for both sources of appreciation, staff nurses seemed to allocate social appreciation on the input side of their pay evaluation, whereas temporary nurses figured it in on the output side.

Relative Importance of Referents

We interpret the overall pattern of findings as support for our contention that pay-evaluation processes are not necessarily uniform in employees across different contexts, such as distinct employment arrangements. This highlights the importance of “beholder” and context characteristics for fairness evaluations, on which today’s researchers in the field of distributive justice agree (e.g., Jasso et al., 2016). Staff nurses relied on perceived responsibility and colleagues’ appreciation to gauge their worth, which is consistent with uncertainty-management theory (van den Bos & Lind, 2002) that proposes the use of referent information to reflect higher levels of uncertainty. For temporary nurses, who are paid considerably better than staff nurses (e.g., Hilgers, 2022), referent information may be less relevant. Determining their worth may generally not be as important an issue for them or they may see their salary as less of an indication of their own worth. Patients’ appreciation had an effect in this group, but in a direction indicating that it is perceived on the output rather than the input side. This was not the case for staff nurses for whom perceived appreciation tended to accentuate unfairness. We do not suggest a reduction in appreciation as a solution, but it seems evident that hospital management cannot rely on raising appreciation to compensate for pay that staff nurses perceive as unfair.

Strengths and Limitations

The stream of research on the minority of temporary workers who voluntarily choose their alternative EA is still rather small. Recruiting registered nurses allowed us to provide insights into the effects of pay unfairness in a highly skilled and coveted group of employees. The opportunity to invite the participation of all on-demand and temporary nurses placed by one of the largest agencies in the Swiss healthcare sector was deemed a means of achieving sample representativeness.

Generalization to the population is not warranted, however. Selection bias may have influenced results since levels of pay unfairness, strain, and work-related attitudes could have affected the decision to participate. At least regarding gender distribution, a comparison with data from the National Office of Statistics (BfS) showed that with 85,1% of female nurses, our sample resembles the target population of registered nurses (84,3% of whom were female in Switzerland, between 2017 and 2019). Since the BfS only now started to differentiate between temporary and staff nurses, further comparisons were impossible. Moreover, while all the information used in our analyses originated from one source, predictor and outcome assessments were separated by a period of 14 days, as recommended by Podsakoff et al. (2012). Finally, the effects of perceived responsibility would be more convincing if the construct were assessed from a different source, for example, a head nurse.

Avenues for Future Research

To examine how fairness perceptions and their antecedents co-develop over time, future research could adopt a cross-lagged panel design (cf. De Cuyper et al., 2012) across three or more measurement occasions separated by at least several weeks. Such a design would also allow to clarify the role of hourly pay in the unfairness evaluation – information that we did not assess in our study. Differences in hourly pay may be a more proximal explanation for some of our EA-related findings in that temporary nurses start their mental accounting from higher pay levels and may be more exigent regarding pay. Furthermore, since perceptions of pay unfairness depend on comparisons with other employees within the respective organization or work unit, we recommend considering unit composition in terms of permanent and temporary personnel. Cobb et al. (2022) showed unit composition to be a factor that influences fairness perceptions. Finally, since neither staff nor temporary nurses are homogenous groups, future studies should sample sufficiently large subgroups and they could assess unfairness perceptions both with regard to the agency and to the host hospital, for temporary nurses.

Practical Implications and Conclusion

In line with existing evidence (e.g., Robbins et al., 2012), our data suggest that promoting fairness has leverage to reduce strain and improve work-related attitudes in a profession with an increased risk of developing depressive symptoms (e.g., Wei et al., 2022). While improving nurses’ physical and mental health is a worthwhile goal in itself, reducing pay unfairness may also be instrumental for hospitals in achieving their goal of retaining experienced nurses (Griffeth & Gaertner, 2001; Twigg & McCullough, 2014), and this may even lead to financial benefits since the cost of turnover is high (the US recruitment service provider NSI estimated the average cost of turnover for a registered nurse to be $ 46.100,- according to Colosi, 2023).

More generally, our study showed that employees draw on different referents as information sources for their evaluations of pay unfairness rather than on a single comparison target. Our findings are relevant for organizations in which pay unfairness may lead to withdrawal behaviors due to chronic pay-unfairness-related strain or dissatisfaction. Particularly in sectors where it is easy for employees to change jobs because their skills are highly coveted, these effects can have disruptive consequences for organizations.