This is an open access article distributed under the terms of the Creative Commons
Attribution License (“Work is our sanity, our self-respect, our salvation. So far from being a curse, work is the greatest blessing.” Henry Ford
Working conditions are changing. Fewer people are employed in a standard work arrangement and work part-time and/or have fixed-term or temporary contracts. Thus, there has been a marked increase in non-standard and contingent work (George & Chattopadhyay, 2015; Kalleberg, 2009). For example, data from the Organization for Economic Cooperation and Development (OECD, 2018) demonstrates that, on average, 11.2% of employees have temporary work contracts in OECD countries. However, in Chile this number is as high as 27%, whereas it is about 5.7% in the UK. These figures show that workplaces refrain from offering permanent contracts for their workers (De Cuyper et al., 2008), but that this trend varies across countries. A question that remains open is how these work arrangements impact the relationship between employees and their workplace. To what extent do people on fixed-term or casual contracts identify with an organisation compared to those on permanent contracts and, more importantly, how might this influence well-being and other work-related outcomes? More generally, do we have to rethink the relationship employees have with their organisation based on their unstable work-status?
Identity-based approaches to understanding organisational issues have long recognised that psychological aspects of work have real consequences, including work being understood as a central aspect of people’s identity (Haslam, 2004; Kalleberg, 2009; Thatcher & Zhu, 2006). Additionally, an individual’s social identity can be derived from the organisation in which she works (Ashforth & Mael, 1989). This particular type of social identity, known as organisational identification, has been found to have a positive impact on several individual and organisational outcomes, including job satisfaction, well-being, and organisational citizenship behaviour (Lee, Koo, & Park, 2015; Mael & Ashforth, 1992).
Processes of organisational identification could arguably be challenged and affected by non-permanent – and sometimes precarious – job arrangements (i.e., employees’ work-status). In addition, drivers and consequences of identification could be affected by the condition of being a temporary worker compared to those with a permanent work-status. Although previous studies have linked work-status and organisational identification by assessing their effect on organisational outcomes (Buonocore, 2010; Feather & Rauter, 2004; Veenstra, Haslam, & Reynolds, 2004), the present study expands these efforts in four important ways. Firstly, it incorporates two forms of identification in the workplace: with the organisation as a whole and with the team of which employees are part. This combination provides not only understanding as to whether work-status influences employees’ attachment to organisations as a whole but their identification with their immediate teams. Secondly, we focus on both an important antecedent (communication climate) and outcomes (well-being, OCB, and intention to recommend) of organisational and team identification. Subsequently, we test a model that incorporates both levels of identification and possible moderating effects of work-status. Thirdly, the present research examines two countries with different distributions of non-standard work arrangements and with different cultural contexts, namely Chile and the UK. Thus, we broaden research to non-‘WEIRD’ (Western, Educated, Industrialized, Rich, Democratic) countries (Gelfand, Leslie, & Fehr, 2008; Henrich, Heine, & Norenzayan, 2010) and provide a valuable contribution to Latin American-based psychological and organisational behaviour research (Nicholls-Nixon, Davila Castilla, Sanchez Garcia, & Rivera Pesquera, 2011). Latin America and Chile are underrepresented areas in psychological science; for example, in 2014, only 3 (0.6%) studies and in 2017 not a single study published in Psychological Science (a top journal in the field) came from Latin America; (Rad, Martingano, & Ginges, 2018). This is also true for most research on nonstandard work arrangements that has been conducted in North America, Europe and Australia (George & Chattopadhyay, 2015) but not in Latin America. However, Chile provides us with an interesting context because, overall, 27.7% of dependant employees work in non-standard work arrangements (ILO, 2016; OECD, 2018). In addition, we aim to replicate our findings in a ‘WEIRD’ country, in order to spot similarities and differences between them. In the UK, only about 5.7% of the workforce has a temporary and about 2.5% has so-called ‘zero-hour’ contracts (i.e., casual work arrangements; ILO, 2016). Lastly, we focus our analyses on younger workers that are classified as ‘millennials’ (those born between 1980 and 1995/6). Firstly, millennials are now the largest age group in the workforce (Brownstone, 2014). Secondly, millennials are often depicted as being less loyal to organisations and lacking work ethics (Myers & Sadaghiani, 2010). Thirdly, Ng and McGinnis Johnson (2015) and others (Lyons & Kuron, 2014) report that generational differences in terms of work attitudes exist; these suggest that younger workers value flexible work conditions, open and clear communication but that organisational commitment might be lower (see also Myers & Sadaghiani, 2010). In addition, millennials also experience more job and organisational mobility and a higher proportion of lateral or downward career moves (Lyons, Ng, & Schweitzer, 2014). However, these attitudes could be driven by the fact that younger workers experience greater job instability and are more often affected by non-standard forms of employment that older workers (ILO, 2016; OECD, 2018). All of these facts indicate that it is worthwhile to study the influence of work-status on job-related variables in the context of younger workers (e.g., millennials) in Chile and the UK.
Literature Review
Main Features and Effects of Work-Status
Nonstandard work arrangements have gained interest from researchers as the ‘standard employment relationship’ has been challenged by the emergence of novel forms of work arrangements, such as part-time, temporary, and ‘zero-hour’ jobs (Benach et al., 2014; Spreitzer, Cameron, & Garrett, 2017). Temporary employment has grown since the 1980s, considerably changing traditional working life (De Cuyper et al., 2008). McLean Parks and colleagues defined a contingent workforce as a group of employees that do not have “either an implicit or explicit understanding that employment will be continuous or ongoing, assuming satisfactory performance by both the individual worker and the organisation” (McLean Parks, Kidder, & Gallagher, 1998, p. 701). This contingent workforce often includes a wide variety of workers (e.g., temporary agency workers, consultants, contractors, zero-hour contracts, etc.). The present study considers three broad categories of work-status, based on the explicit existence and duration of the contract: permanent, fixed-term, and casual statuses. More specifically, permanent workers have a contract with no specified ending, while temporary workers do not. The latter group can be divided into two work-statuses: fixed-term and casual workers. Fixed-term workers have a contract that lasts for a particular time frame (e.g., 12 months). This group of employees has been largely studied as “temporary” or “contingent” workers (McLean Parks et al., 1998; Wilkin, 2013). Casual workers are a specific type of temporary workers, which has gained less attention in the literature (Allen, 2011; Veenstra et al., 2004). According to Allen (2011), casual work involves no expectations of continuous work, as well as no paid holiday, notice of dismissal, or redundancy payments, among others.
Importantly, these features apply to a considerable group of workers in Chile who do not have an employment contract with their employers, but an agreement based on the “provision of services”1(see Neira & Rojas, 2015). This means they do not have a legally recognised employer-employee relationship, which is regulated under the labour law. They are not entitled to receive paid holidays, or social security (including healthcare, parental or sick leave, and pension). Although many of these employees may renew their agreements regularly, they experience high levels of uncertainty, as no employment contract protects them against anticipated or unexpected dismissal2. In many other industrialised countries, the diversification of part-time work into “very short hours” or “on-call” work, including “zero-hour” contracts (with no guaranteed minimum hours) parallels with casual work in more developed countries. For example, at the end of 2015, some 2.5 % of employees were on zero-hour contracts in the UK (ILO, 2016) and in 2018 the number increased to roughly 2.7% or 844,000 workers; importantly, people on these contracts tend to be younger, female, and students (Office for National Statistics, 2019).
Being a temporary worker implies certain characteristics that distinguishes these workers from permanent employees. Temporary work might be less desirable and might provide less satisfaction than permanent work; it is often associated with lower job security, reduced access to benefits, training, lower wages, and fewer opportunities to participate in decision making (Booth, Francesconi, & Frank, 2002; Foote, 2004; Kauhanen & Nätti, 2015; Veenstra et al., 2004). Furthermore, temporary employment has been recognised as precarious, because it is characterised by uncertainty, unpredictability and risk (Kalleberg, 2009). It has also been suggested that continuous precarious employment could elicit higher levels of stress or employment strain when uncertainty about the future is constant. (Lewchuk, de Wolff, King, & Polanyi, 2003; Pirani & Salvini, 2015). Ek, Sirviö, Koiranen, and Taanila (2014) found that the impact of precarious work was associated with symptoms of depression, which was higher in the case of young, highly educated men. In addition, a recent study by Van Aerden, Gadeyne, and Vanroelen (2017) found in a Belgian sample that precarious work (that was defined by low income, flexible work hours and involuntary part-time work) was associated with a decrease in general and mental health.
Importantly, these temporary conditions could not only impact people’s general and mental health but also organisation-related outcomes. De Cuyper et al. (2008) conducted a systematic literature review focused on the impact of temporary work on well-being, job satisfaction, organisational commitment, and productive behaviours. These authors emphasised that classic theories – such as social comparison and social exchange – would support a negative relationship between work-status and job-related outcomes. However, the reviewed results were not consistent and they found both negative and positive relationships between work-status and, for example, organisational commitment. Further, when De Cuyper and De Witte (2010) explored the reasons behind these puzzling results, they suggested, for example, that workforce composition (in terms of the number of temporary employees) or the type of non-standard work could play a role in understanding how work-status impacts job-related outcomes (see also Wilkin, 2013 for a meta-analysis about contingent work and job satisfaction). In the light of these inconclusive results we aim to explore how work-status influences antecedents and consequences of identification processes at work.
Organisational Identification: Antecedents and Consequences
Mael and Ashforth (1992, p. 104) defined organisational identification as “the perception of oneness with or belongingness to an organisation, where the individual defines him or herself in terms of the organisation(s) in which he or she is a member”. From a Social Identity perspective, individuals build part of their identities based on their membership of different social groups or categories (Tajfel & Turner, 1979). They navigate this process by internalising and enacting norms and behaviours from in-group members (Hogg, 2006). Thus, the organisation for which people work would provide a relevant source for the process of self-categorisation and subsequent self-definition based on that social category. Organisational identification raises individuals’ self-related motives (e.g., self-distinctiveness, self-continuity), fulfilment of basic human needs (e.g., affiliation, safety) and self-enhancement (Ashforth, Harrison, & Corley, 2008). Additionally, it benefits individuals in terms of their health and well-being (Steffens et al., 2017). Organisational identification was found to be a key contributor to people’s well-being as both organisational and/or team identification are positively associated with well-being. In the present work we focus on job affective well-being to assess the affective states derived directly from the work experience (Van Katwyk, Fox, Spector, & Kelloway, 2000).
The relevance of organisational identification not only comes from its impact at the individual level, but also from its impact on organisation-related outcomes, such as in-role and extra-role performance or organisational citizenship behaviour (Kreiner & Ashforth, 2004; Mael & Ashforth, 1992; Riketta, 2005). Additionally, cooperation, effort, participation, motivation, and coordinated action are among the most referenced outcomes (Ashforth et al., 2008). A recent meta-analysis confirmed organisational identification’s impact on attitudes and behaviours. Interestingly, it also showed that the effect of organisational identification on behaviours would exist above and beyond the effect of attitudes on behaviours (Lee et al., 2015). Thus, an important consequence of organisational identification on which we focus is organisational citizenship behaviour, expecting that both are positively correlated.
In terms of the drivers of organisational identification, previous work has found that communication climate is positively associated with organisational identification (Ashforth et al., 2008; Edwards & Peccei, 2010; Fuller, Hester, Barnett, & Frey, 2006; Ngo, Loi, Foley, Zheng, & Zhang, 2013; Smidts, Pruyn, & Van Riel, 2001). Communication Climate (CC) is defined as a dimension of the psychological climate that involves perceptions about the work environment, especially in relation to trust and openness in communications, participation in decision making, and supportiveness (Smidts et al., 2001), covering relevant aspects of other predictors. The connection between CC and the other predictors can be sustained by Fuller and colleagues’ (2006) work, showing that perceived internal respect was given by the actions of management that communicate the inclusion or belongingness of workers, such as the degree to which they can participate in decision making. Thus, in the present work we focus on CC as an important antecedent for organisational and team identification in the way that CC should be positively associated with both foci of identification. However, we are also interested in whether and how fixed-term and casual workers’ identification with the organisation might be different from those in permanent jobs; thus, we are interested in the effects of work-status on identification and its antecedents and consequences.
Organisational Identification and Its Connection to Work-Status
Based on the previously exposed characterisation of temporary employment, the question of whether fixed-term and casual workers identify with their organisation can be raised. For example, from a psychological contract perspective (Rousseau, 1995), temporary workers have been found to have a more transactional (and less relational) orientation towards work (Millward & Hopkins, 1998), which could also impact the degree to which they identify with the organisation (Koene & Van Riemsdijk, 2005). In support of these ideas, Buonocore (2010) found that work-status was negatively associated with organisational identification.
On the other hand, it could be argued that temporary workers can still identify with their organisations, but maybe in a different way. According to Rousseau (1998), organisational identification can occur in two forms: deep-structure and situated. While deep-structure identification would be associated with a more relational form of psychological contract, situated identification would be more transactional. As such, the latter would depend on contingent cues, which create a sense of “we” that lasts while the cues are present. Therefore, Rousseau argues that this form of identification could also be found in short-term work settings. The latter found support in Feather and Rauter’s (2004) study, which showed that the levels of identification did not differ between permanent and temporary teachers. They further argued that temporary workers had a situated identification that might not differ from the deep-structure in terms of the measure used. In spite of the similar levels of organisational identification, they did find an impact of identification on organisational citizenship behaviour (OCB) in permanent workers, and not in temporary workers, suggesting that work-status had a moderating effect on identification and its relationship with OCB, an important observation that we will address in the following.
However, Veenstra et al. (2004) found differences in the level of organisational identification between permanent, fixed-term and casual roles that was not significant after including the variable job security. The authors suggested the relevant role of security in determining the impact of casualisation was explained by the fact that they were themselves determinants of social identification (see also De Cuyper et al., 2008).
In sum, previous work on the impact of work-status on attitudinal and behavioural outcomes is not conclusive (De Cuyper & De Witte, 2007). Consequently, our paper aims to further explore the role of work-status on organisational identification both in terms of its antecedents and consequences.
Moreover, modern, flexible working relationships could promote the adoption of different types of sub-identities (Ashforth & Johnson, 2001; Ashforth et al., 2008). Accordingly, social identification does not only occur with the organisation as a whole, but also with a sub-group of it, such as a department, team, union, or age cohort (Ashforth & Mael, 1989). Riketta and Van Dick (2005) concluded that identification with a team was higher than identification with the organisation as a whole. They explained that the work group would be a more salient social group that would provide more instances of socialisation and control, and would consequently be suited to satisfy people’s needs for distinctiveness. This could be especially pertinent for employees on non-contingent contracts.
Regarding the outcomes of the different foci of identification, previous work highlighted that team identification had a higher correlation with job satisfaction, job involvement, motivation, and turnover, compared to identification with the organisation as a whole (Knippenberg & Van Schie, 2000). However, a recent meta-analysis by Steffens, Haslam, Schuh, Jetten, and van Dick (2017) indicated that the positive correlation between organisational and work-place (e.g., team) identification and well-being did not differ in strength. However, what remains unclear is whether these relationships are influenced by non-standard work arrangements compared to those with permanent contracts. One could assume that non-standard work environments disrupt social psychological processes that underlie identification; thus, when the continuous employment in one organisation turns out to be less important, more ‘fluid’ notions of work identity - such as team identification - become more relevant (Thatcher & Zhu, 2006) and therefore influence the relationship between identification and work-related outcomes.
The Present Study
This study formulates and analyses a conceptual model that explores antecedents and consequences of organisational and team identification in permanent, fixed-term, and casual workers in Chile and the UK (see Figure 1). In our conceptual model, we expect communication climate to predict organisational and team identification to capture the impact of context on both foci of identification (Bartels, Pruyn, De Jong, & Joustra, 2007). As such, we predict a positive association between communication climate and identification. Although previous literature on the relationship between work-status, communication climate and identification is scarce, we suggest that CC will be an important driver for all workers despite their different work status; thus, we do not expect a moderating role for work-status on the relationship between CC and identification. Further, corresponding to Riketta and Van Dick’s (2005) meta-analysis, we expect a medium-sized correlation between both foci of identification. In addition, in line with previous work, we predict that team-identification will be higher than organisational identification. Three outcomes of identification are presented and positive associations with both forms of identification are expected. Firstly, we chose job-affective well-being as a way to assess the affective states derived directly from the work experience. Secondly, OCB is placed in the model in order to assess the impact of identification on a relevant behavioural outcome. Finally, the intention to recommend the organisation (that is, whether employees recommend the organisation they work for to others) is an important indicator for employee engagement and similar to the employee net promoter score (eNPS), which is often used by practitioners (Almquist, Leiman, Rigby, & Roth, 2013). As suggested by Mael and Ashforth (1992), organisational identification would increase the willingness to advise others to join the organisation.
In line with the work by Feather and Rauter (2004), who found that the impact of identification on organisational citizenship behaviour (OCB) was positive for permanent workers but not for temporary workers, we suggest that work-status has a moderating effect on identification and its relationship with OCB. However, due to the inconsistent nature of previous findings on the relationship between work-status, identification, well-being and intentions to recommend, we do not posit more specific hypotheses regarding the influence of work-status on foci of identification and their consequences; we explore whether work-status moderates these associations by analysing how the overall model differs between permanent, fixed-term, and casual workers.
In a second step, we pre-register Study 2 to replicate our empirical model in a different context (the UK). The full conceptual model is summarised in Figure 1.
Figure 1
Conceptual model for the relationship between variables. JAWS = job affective well-being, OCB = Organisational Citizenship Behaviour.
Study 1
Method
Participants and Design
We collected data via an online questionnaire that was distributed via online social networks and mailing lists. No monetary incentive was offered for participation. 631 participants completed the survey (375 women, 236 men, 20 not given) who were professionals aged between 23 and 36 (Mage = 29.46, SD = 3.15), working in Chile (426 private sector, 161 public sector, and 43 other sectors). 371 (58%) were employed on a permanent contract (Mtenure = 30.55 months, SD = 28.85); 129 (20.4%) on a fixed-term contract, (Mtenure = 23.99 months, SD = 25.90) and 131 (20.6%) on casual arrangements; (Mtenure = 22.7 months, SD = 27.89).
Procedure and Measures
The study was approved by the ethics committee at the researchers’ institution. We obtained informed, electronic consent from all participants. The first page of the questionnaire specified the conditions of participation and provided a general description of the study. By clicking on a box, participants confirmed that they: 1) had been born in 1980 or after, 2) were working for an organisation in Chile, 3) had a bachelor’s degree or professional title, 4) were part of a work-team, and 5) were not part of upper executive management. All measures were translated (and back translated) from English to Spanish3.
Outcome variables
A one-item measure (“Would you recommend your friends to work for this organisation?”) was used as an indicator of the general evaluation of the organisation made by the respondent, thus measured Intention to recommend their organisation.
Job-affective well-being
A 12-item version (Schaufeli & Van Rhenen, 2006) of the Job-related Affective Well-Being Scale (JAWS) (Van Katwyk et al., 2000) was used to assess the frequency of certain emotions experienced at work in the last 30 days (5-point Likert scale, from “Never” to “Always”). The reliability of JAWS in this study was Cronbach’s α = .88.
Organisational citizenship behaviour
was measured by a 5-item scale adopted by van Dick, van Knippenberg, Kerschreiter, Hertel, and Wieseke (2008). Participants were asked to state their level of agreement (5-point Likert scale) for the statements presented (e.g., “I follow rules very thoroughly”). Reliability for the scale was low, Cronbach’s α = .61 and results for this variable have to be interpreted with caution.
Organisational and team identification
The key construct of this study was measured with a 5-item scale composed by items selected from three previous studies (Knippenberg, Knippenberg, Monden, & De Lima, 2002; Smidts et al., 2001; van Dick et al., 2008). Replicating van Dick and colleagues’ (2008) methodology, this scale was presented twice in the questionnaire, using the word “organisation” in one and the word “team” in the other (e.g. “When I talk about my organisation [team], I usually say ‘we’ rather than ‘they’”). This variation measured identification both at the team and the organisation-as-a-whole level. Both scales showed high reliability; Cronbach’s αOID=.88, Cronbach’s αTID=.86. The order of the scales was randomised, so half of the participants answered the items related to Organisational Identification (OID) first and the other half answered the items related to Team Identification (TID) first.
Predictor Variables
Communication climate
was measured with Smidts and colleagues’ (2001) 15-item scale that alludes to perceptions and interpretations of the work environment related to its communicative elements. The original measure was composed by three dimensions: trust and openness in communications (e.g., “When my direct boss tells me something, I trust him/her to be candid and honest”), participation in decision making (e.g., “Our general management is open to suggestions we put to them”), and supportiveness (e.g., “The information we receive here is often about trivial matters” [reversed]). For the purposes of this research, we compiled the 15 items into one scale, Cronbach’s α =.88, that was used as an observed variable in the model.
Work-status
One question was used to identify the respondents’ work-status (WS). This question was placed in-between other demographic items.
Demographic and control variables
Several measures were included in the questionnaire to allow the option to control for relevant aspects. Perceived Job Security, measured with one question, recorded the respondents’ perceived certainty about the possibility of continuing working for that organisation. Respondents could choose a value in a scale of 0% to 100%.
Further, we included individual (age, gender, tenure, contract length, job type) and organisational factors (public vs. private sector, size of the organisation, organisation type) as demographic variables. Nevertheless, no relevant impact of these variables was found and they were therefore excluded from the main analysis, aiming for simplicity in the model.
Results
Descriptive Statistics and Differences Dependent on Work-Status
First, we aimed to understand whether different kinds of people (in terms of gender, age, tenure, part-time, etc.) had different work-statuses (WS). A higher percentage of women comprised fixed-term (n = 80; 72%) and casual (n = 91; 62%) contracts in comparison to permanent contracts (n = 204; 52%; χ2(611) = 9.86, p = .007). The percentage of part-time work was higher among fixed-term (n = 18; 14%) and casual contracts (n = 53; 40%; χ2(611) = 9.86, p = .007) compared to permanent contracts (n = 20; 5%). Fixed-term (n = 69; 53%) and casual (n = 65; 40%; χ2(611) = 97.12, p < .001) contracts were more prevalent among public sector jobs (compared to private sector) as opposed to permanent jobs (n = 27; 7%). People on casual contracts were slightly younger (M = 28.72; SD = 3.03) than people on permanent contracts, (M = 29.64; SD = 3.31; ∆M = 0.914, p = .006) and on fixed-term contracts, (M = 29.50; SD = 3.31; ∆M = 0.78, p = .05).
People with permanent contracts tend to be male, older, working full-time and in the private sector, whereas people with fixed-term or casual contracts are more likely to be female, younger, working part-time and in the public sector.
Second, we conducted multivariant analysis of variance (MANOVA) using R with WS as the between-factor (permanent, fixed, casual contracts) and all variables of interest as outcome measures. This analysis revealed an overall effect of WS on our variables; F(7, 615) = 10.11, p < .001, = .10). Means and standard deviations were computed for all variables and analyses of variance (ANOVA) were conducted to assess the differences among the groups for each variable (see Table 1). Statistically significant differences were found only in the level of job security, F(1, 622) = 16.34, p < .001, = .050. Inspection of means showed that workers in permanent jobs felt most secure about their jobs (M = 79.86, SD = 24.36). Casual workers felt significantly less secure (M = 64.87, SD = 27.94; ∆M = 14.89; p < .001; 95% CI [9.65, 20.13]) than workers on fixed–term contracts (M = 72.84, SD = 28.72; ∆M = 7.02; p < .001; CI [1.77, 12.27]). Fixed-term workers felt less secure than permanent workers, ∆M = 7.87; p < .001; CI [1.49, 14.26].
Table 1
Means and Standard Deviations in Permanent, Fixed-Term, and Casual Workers: Study 1
| Variable | Permanent (P)
|
Fixed-Term (FT)
|
Casual (C)
|
F | p | η2 | |||
|---|---|---|---|---|---|---|---|---|---|
| M | SD | M | SD | M | SD | ||||
| Communication Climate (CC) | 3.31 | 0.62 | 3.26 | 0.62 | 3.26 | 0.68 | 0.421 | .656 | .001 |
| Job Security (JS) | 79.86 | 24.36 | 72.84 | 28.73 | 64.97 | 27.94 | 16.344 | < .001 | .052 |
| Organisational Identification (OID) | 3.34 | 0.94 | 3.31 | 0.83 | 3.25 | 0.93 | 0.492 | .612 | .001 |
| Team Identification (TID) | 3.84 | 0.81 | 3.69 | 0.78 | 3.73 | 0.85 | 1.918 | .148 | .006 |
| Job AffectiveWell-being (JAW) | 3.27 | 0.59 | 3.36 | 0.61 | 3.40 | 0.67 | 2.548 | .079 | .009 |
| Organisational Citizenship Behaviour (OCB) | 4.02 | 0.53 | 4.13 | 0.49 | 4.00 | 0.56 | 2.554 | .079 | .008 |
| Intention to recommend (IR) | 3.89 | 0.92 | 3.88 | 0.92 | 3.82 | 1.07 | 0.313 | .731 | <.001 |
Note. Post-hoc Tukey test shows statistically significant differences in the level of Job Security between P and FT (p = .024), P and C (p < .001), and FT and C (p = .042).
Additionally, a t-test showed that across the whole sample identification with the team was higher than identification with the organisation, (t(621) = -9.55, p < .001, Mdiff = .47, 95% CI [-.37, -.56]). This result supports the relevance of including both foci of identification in the present study and confirms previous findings about them (e.g., Knippenberg & Van Schie, 2000; Veenstra et al., 2004). In addition, and in line with previous findings, we found that OID and TID were correlated; r(629) = .492, p < .001. For the full correlation matrix, see Table 2.
Table 2
Correlation Matrix: Study 1
| Variable | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
|---|---|---|---|---|---|---|---|
| 1. Communication Climate | – | .33** | .54** | .40** | .65** | .20** | .56** |
| 2. Job Security | – | .32** | .27** | .39** | .08 | .37** | |
| 3. Organisational Identification | – | .49** | .54** | .26** | .56** | ||
| 4. Team Identification | – | .42** | .36** | .36** | |||
| 5. Job Affective Well-Being Scale | – | .22** | .62** | ||||
| 6. Organisational Citizenship Behaviour | – | .19** | |||||
| 7. Intention to Recommend the organisation as a place to work | – |
**p < .01.
Hypothesis Testing: Path Analyses
Although we found surprisingly small mean-level differences between workers with different WS, the main focus of our work was the conceptual model depicted in Figure 1 and thus the question of how antecedents and consequences of OID and TID might differ across participants with different work-status. We conducted path analyses using R and the lavaan package (Rosseel, 2012). We chose this method because it allowed for multiple regression models to be computed and inter-related at the same time. Additionally, it simultaneously and easily incorporated moderating effect of WS, enabling the comparison of regression coefficients between different groups; something that is not easily done in multiple regressions. As there was no clarity regarding where the differences between the three work-statuses would be, we used modification indices to explore whether the relationship between variables changed for different groups depending on their work-status.
The sample was divided into three groups according to the WS: Permanent (n = 369), Fixed-Term (n = 129), and Casual workers (n = 131). The conceptual model was fitted and direct paths between predictors and outcomes of OID and TID were added, as a way to assess the magnitude of the unique effect of these variables on the outcomes. Job security was not included in the model as a previous analysis suggested it to be a proxy for work-status.
Initially we ran the model that was hypothesised in Figure 1 across the full sample (thus, not estimating whether there were differences between WS and assuming that all relationships between variables (hence, regression weights, variances, etc.) are identical across the three groups. However, the overall fit of that model was not satisfactory, χ2(3) = 71.95, p < .001, RMSEA = .191 [.154, .231], CFI = .948, AIC = 7065.23, SRMR = .04; this indicates that the fully-constraint model, which assumes no differences across all three groups, is not a good depiction of the data.
We then fitted a model across all three groups but left all parameters unconstrained. Such a model would indicate that all relationships between variables differ across the three groups. The overall fit of this model was χ2(9) = 78.334, p < .001, RMSEA = .192 [.154, .232], CFI = .949, AIC = 7070.490, SRMR = .035; again not a satisfactory fit. This means that neither a fully constrained model nor a fully unconstrained model adequately fits the data, which suggests that some relationships between variables differ for workers with different work-status and some are the same.
We then ran a model in which we set all regression parameters as equal across the three groups (but allowed further variation). This constrained model fit was χ2(31) = 122.62, p < .001, RMSEA = .191 [.097, .141], CFI = .93, AIC = 7070.785, SRMR = .074.
In order to test whether differences in the parameters are statistically significant, two nested models should be compared in terms of their cross-group invariance (Molina, Alegría, & Mahalingam, 2013). As such, the unconstrained model was compared to the constrained model for the regression weights, where all the regression paths were forced to be equal between the groups. These models differed in terms of their fit, χ2diff (22) = 44.295, p = .003, suggesting a preference for the unconstrained model over the constrained model. From these three analyses we can conclude that there are some group differences in terms of WS in relation to the overall model but that some relationships are similar across the groups.
To explore these possible group differences further and to see which specific regression paths might be moderated by WS, we inspected modification indices for the constrained model (we identified those parameters with a modification index < 5). These firstly suggested the allowance for a correlation between JAWS and intent to recommend. We fitted an overall model where regression paths were set as equal between groups but allowed for a correlation between JAWS and recommend; the fit improved, χ2(28) = 48.90, p < .001, RMSEA = .060 [.03, .08], CFI = .985 AIC = 7003.06, SRMR = .092. However, modification indices also suggested allowing the paths between OID→JAWS, TID→JAWS, and OID→OCB to vary across groups. We then fitted a model that constrained all regression paths to be equal across groups except for the relationship between OID→JAWS, TID→JAWS, and OID→OCB. The model had an excellent fit with χ2(22) = 27.36, p = .198, RMSEA = .034 [.00, .07], CFI = .996, AIC = 6993.51, SRMR = .052. In addition, this model had a better fit in comparison to the fully constrained model, χ2diff (6) = 21.54, p = .001. The full model’s results are summarised in Table 3.
Table 3
Direct Effects (Unstandardised Coefficients) Reported for Each of the Paths in the Model for Study 1
| Relationship between variables | Permanent
|
Fixed-Term
|
Casual
|
|||
|---|---|---|---|---|---|---|
| b (SE) | z | b (SE) | z | b (SE) | z | |
| CC→OID | 0.78 (0.049) | 16.14** | 0.78 (0.049) | 16.14** | 0.78 (0.049) | 16.14** |
| CC→TID | 0.53 (0.047) | 11.26** | 0.53 (0.047) | 11.26** | 0.53 (0.047) | 11.26** |
| CC→JAWS | 0.46 (0.033) | 13.95** | 0.46 (0.033) | 13.95** | 0.46 (0.033) | 13.95** |
| OID→JAWS | 0.17 (0.029) | 5.96** | 0.02 (0.051) | 0.39 | 0.21 (0.044) | 4.69** |
| TID→JAWS | 0.05 (0.030) | 1.64 | 0.26 (0.056) | 4.78** | 0.07 (0.048) | 1.51 |
| OID→OCB | 0.08 (0.032) | 2.62** | -0.08 (0.050) | -1.52 | 0.05 (0.049) | 1.01 |
| TID→OCB | 0.21 (0.028) | 7.42** | 0.21 (0.028) | 7.42** | 0.21 (0.028) | 7.42** |
| CC→OCB | 0.03 (0.037) | 0.69 | 0.03 (0.037) | 0.69 | 0.03 (0.037) | 0.69 |
| CC→IR | 0.54 (0.055) | 9.26** | 0.54 (0.055) | 9.26** | 0.54 (0.055) | 9.26** |
| TID→IR | 0.05 (0.042) | 1.18 | 0.05 (0.042) | 1.18 | 0.05 (0.042) | 1.18 |
| OID→IR | 0.37 (0.040) | 9.32** | 0.37 (0.040) | 9.32** | 0.37 (0.040) | 9.32** |
Note. Paths that varied across groups are in bold. CC = Communication Climate; JS = Job Security; OID = Organisational Identification; TID = Team Identification; JAWS = Job Affective Well-Being Scale; OCB = Organisational Citizenship Behaviour; IR = Intention to Recommend the organisation as a place to work.
*p < .05. **p < .001.
To summarise, communication climate has a strong and positive relationship with organisational and team identification; the more employees experience a positive communication climate, the stronger their identification with the organisation and the team. In addition, communication climate is also directly related with job-affective well-being and intention to recommend, but not with organisational citizenship behaviour. Importantly, these relationships were stable across groups, hence work-status did not influence the importance of communication climate for the organisational outcome variables.
However, the effect of organisational and team identification on job-affective well-being is influenced by work-status. The results indicate that organisational identification is positively associated with job-related well-being for employees with permanent and casual work contracts but not for those with fixed-term contracts, for whom we did not see a meaningful influence of organisational identification on job-related well-being. On the contrary, team identification has a positive relationship with well-being for employees with fixed-term contracts but not those with permanent and casual ones; thus, work-status also moderates the relationship between team identification and well-being. In addition, work-status influenced the relationship between organisational identification and OCB; organisational identification and OCB were positively (but relatively weakly) linked for employees with a permanent contract but not for those with a fixed-term or casual contract. Team identification has an overall positive relationship with OCB. Hence, work-status does influence how organisational and team identification are associated with well-being and OCB but did not influence the relationship between both foci of identification and intentions to recommend, which were positively related across all surveyed employees. However, we have to keep in mind that the OCB measure had a low Cronbach’s alpha and the results should be treated with caution. In addition, job-related well-being and intentions to recommend are positively correlated for people with fixed-term and casual contracts but not those with permanent ones. This means for people on permanent contracts job-related well-being and intentions to recommend seemed to be unrelated.
Study 2
To replicate and compare the previous findings in a different context, we conducted a second study in the UK. In the UK only about 6% of workers have a temporary work-status and 2.9% have ‘zero-hour’ contracts (ILO, 2016; OECD, 2018; Office for National Statistics, 2019). Chile and the UK also differ on cultural dimensions. Whereas Chile is a nation that is far-West but highly collectivistic, receiving a score of 23 on Hofstede’s cultural dimensions (Hofsted, 2019), the UK scores very high on individualism (89). In addition, the UK is also much lower on Power Distance (35 vs. 63), higher in long-term orientation (51 vs. 31) and much lower on uncertainty avoidance (35 vs. 86). Thus, the two countries we compare provide different cultural contexts both in terms of their work environments and their national cultures. However, the question is whether these will influence the psychological processes and, hence, the relationship between variables. We are not aware of any study that explicitly compares identification processes in Chile and the UK. However, there is work that suggests that general principles of organisational behaviour should hold across cultures (Gelfand, Erez, & Aycan, 2007). In addition, the meta-analysis by Steffens and colleagues (2017) showed that the relationship between identification and health/well-being was stable across cultures. Thus, one aim of our work is to examine whether the effects we found in Chile can be replicated in a different cultural context. This aim responds to calls for more replications (Earp & Trafimow, 2015; Rhode, Voyer, & Gleibs, 2016) and more inclusion and diversity (Kitayama, 2017), addressing the question of whether contextual factors (e.g., location, culture) are associated with reproducibility (Van Bavel, Mende-Siedlecki, Brady, & Reinero, 2016).
Method
Participants and Design
Data for this study was collected via an online questionnaire that was completed by 520 people. The vast majority accessed the survey via Prolific Academic and were paid £1.30 (based on £7.80/h) for the completion (n = 501). To match our inclusion criteria for Study 1, we pre-screened participants from Prolific Academic. Thus, participants had to be born in 1980 or after, have at least a BSc degree, were in full-time or part-time employment, worked within a team or with other people, and lived and worked in the UK. These criteria left us with a participant pool of 2790 on Prolific Academic of which 501 answered our survey.
Participants (338 women, 181 men, 1 non-binary) were knowledge workers aged between 18 and 38 (Mage = 30.44, SD = 4.66), working in the UK (338 private sector, 181 public sector, and 43 other sectors). 433 (83.1%) were employed on a permanent contract (Mtenure = 62.98 months, SD = 53.97); 53 (10.2%) on a fixed-term contract (Mtenure = 29.26 months, SD = 32.08) and 35 (6.7%) on casual arrangements (Mtenure = 48.03 months, SD = 39.55). 30% (n = 160) of participants had an income below £20k per year; 47.6% (n = 248) between £20-35k; 13% reported an income of between £35-50k (n = 68), 3.3% (n = 17); 1.8% (n = 9) and the rest preferred not to say or left the field blank.
Procedure and Measures
The study was approved by the researchers’ institution and was preregistered at Aspredicted (see Supplementary Materials). We obtained electronic informed consent from all participants. We used the English version of the same survey that we used in Study 1 (see Supplementary Materials). Cronbach’s Alpha for all scales are summarised in Table 4.
In addition to the previously used measures, we also included a measure of Organisational Citizenship Behaviour (Lee & Allen, 2002) because we found that the OCB measure in Study 1 had low reliability and, thus, we included eight items that addressed behaviours that were beneficial for individuals (e.g., “I willingly give my time to help others who have work-related problems) and eight items that tapped into behaviours that were more beneficial for the organisation as a whole (e.g., “I keep up with developments in the organisation”). We analysed a model using these two OCB measures but, because these measures were not included in Study 1, we will not report the details in the main paper (additional analyses can be found in the Supplementary Materials).
Results
Descriptive Statistics and Differences Dependent on Work-Status
The percentage of women did not differ much between fixed-term (n = 35; 66%) and casual contracts (n = 18; 52%; z-value = 1.3, p = .20) or in comparison to permanent contracts (n = 285; 62%; z = .05, p = .58). The percentage of part-time workers was higher among fixed-term (n = 21; 39%) and casual contracts (n = 18; 52%) compared to permanent contracts (n = 84; 19%).
Fixed-term (n = 29; 55%) contracts were more prevalent among public sector jobs compared to permanent jobs (n = 129; 29%; z = 3.7, p < .001) and casual contracts (n = 10; 29%; z = 3.7, p = .02). In the private sector, 63% of participants (n = 274) had permanent jobs, compared to 16 (30%; z = 4.5, p < .001) with fixed-term contracts. 18 employees with casual contracts (52%) reported working in the private sector.
People on casual contracts were slightly younger (M = 29.50; SD = 5.50) than people on permanent contracts, (M = 30.77; SD = 5.50; ∆M = -1.27, p = .18) and older than people on fixed-term contracts (M = 28.98; SD = 5.01; ∆M = 0.52, p = .62). These differences were small. People on permanent contracts were slightly older than people on fixed-term contracts (∆M = 1.79, p = .10). People on permanent contracts were slightly older, worked full-time and in the private sector, whereas people with fixed-term or casual contracts were more likely to be younger, working part-time and in the public sector. Unlike in the Chilean sample, we did not observe any gender differences here.
We conducted multivariate analysis of variance (MANOVA) using R with WS as the between-factor (permanent, fixed, casual contracts) and all variables of interest as outcome measures. This analysis revealed an overall effect of work-status on our variables, F(7, 507) = 2.70, p = .009, = .029. Then, means and standard deviations were computed for all the variables and an analysis of variance (ANOVA) was conducted to assess the differences between the groups for each of variable (Table 4). Statistically significant differences were found only in the level of job security, F(1, 513) = 11.775, p < .001, = .050. Inspection of means showed that workers in permanent jobs felt most secure about their jobs (M = 85.41, SD = 18.94). Fixed-term workers felt significantly less secure (M = 74.28, SD = 24.00, ∆M = 11.08; p < .001; CI [4.04, 18.11]). The differences between permanent and casual workers (M = 78.29, SD = 26.99; ∆M = 6.44; p = .20, CI [-2.05, 14.94]) were not statistically significant. There were also no significant differences between fixed-term workers and casual workers (∆M = -4.63; p = .87, CI [-15.15, 5.89]).
Table 4
Means and Standard Deviations in Permanent, Fixed-Term, and Casual Workers: Study 2
| Scale (Cronbach's Alpha) | Permanent (P) (n = 428) |
Fixed-Term (FT) (n = 53) |
Casual (C) (n = 34) |
F | p | η2 | |||
|---|---|---|---|---|---|---|---|---|---|
| M | SD | M | SD | M | SD | ||||
| Communication Climate (.85) | 3.29 | 0.61 | 3.34 | 0.52 | 3.31 | 0.55 | 0.16 | .851 | .001 |
| Job Security | 85.41 | 18.94 | 74.28 | 24.00 | 78.29 | 26.99 | 8.58 | < .001 | .032 |
| Organisational Identification (.91) | 3.45 | 1.03 | 3.38 | 1.01 | 3.38 | 0.95 | 0.19 | .820 | .001 |
| Team Identification (.89) | 3.87 | 0.91 | 3.83 | 0.80 | 3.58 | 0.87 | 1.68 | .187 | .007 |
| Job Affective Well-being (.90) | 3.28 | 0.72 | 3.33 | 0.69 | 3.17 | 0.69 | 0.52 | .592 | .002 |
| Organisational Citizenship Behaviour (.64) | 4.12 | 0.56 | 4.21 | 0.45 | 4.11 | 0.60 | 0.69 | .499 | .003 |
| OCB__Individual (.80) | 3.95 | 0.57 | 3.80 | 0.49 | 3.83 | 0.59 | 2.18 | .110 | .008 |
| OCB__Organisation (.88) | 3.63 | 0.75 | 3.42 | 0.78 | 3.50 | 0.78 | 2.01 | .130 | .008 |
| Intention to recommend | 3.54 | 1.16 | 3.59 | 0.89 | 3.56 | 1.13 | 0.49 | .615 | .002 |
Note. Post-hoc Tukey test shows statistically significant differences in the level of Job Security between P and FT (p = .001), but not P and C (p = .87), and FT and C (p = .20).
Additionally, for the whole sample, a t-test showed that identification with the team was higher than identification with the organisation, t(519) = -10.76, p < .001, Mdiff = -.407, 95% CI [-.33, -.48]. This result supports the relevance of including both foci of identification in the present study and confirms previous findings. In addition, we found that OID and TID were correlated, r(519) = 0.604, p < .001. For full correlation matrix, see Table 5.
Table 5
Correlation Matrix, Study 2
| Scale | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
|---|---|---|---|---|---|---|---|
| 1. Communication Climate | – | .21** | .58** | .45** | .65** | .27** | .61** |
| 2. Job Security | – | .25** | .32** | .26** | .11* | .24** | |
| 3. Organisational Identification | – | .60** | .64** | .37** | .66** | ||
| 4. Team Identification | – | .53** | .41** | .32** | |||
| 5. Job Affective Well-Being Scale | – | .33** | .26** | ||||
| 6. Organisational Citizenship Behaviour | – | .31** | |||||
| 7. Intention to Recommend the organisation as a place to work | – |
**p < .01.
Hypothesis Testing: Path Analyses
Similar to the Chilean sample we found small mean-level differences between workers with different WS. For further analyses and as specified in our pre-registration, we conducted path analyses using R and the lavaan package (Rosseel, 2012).
The sample was divided into three groups according to the WS: Permanent (n = 433), Fixed-Term (n = 53), and Casual workers (n = 35).
In light of our preregistration, we aimed to test the same model that we established in Study 1 in the UK sample. We therefore fitted a model that constrained all regression paths to be equal across groups except for the relationship between OID→JAWS, TID→JAWS, and OID→OCB and allowed for a correlation between JAWS and intent to recommend. The model had an excellent fit with χ2(22) = 25.81, p = .260, RMSEA = .032 [.000, .074], CFI = .997, AIC = 5995.41, SRMR = .026. Results are summarised in Table 6. In sum, we could mainly replicate our empirical model in our UK sample that we found in our Chilean sample.
Table 6
Direct Effects (Unstandardised Coefficients) Reported for Each of the Paths in the Model for Study 2
| Relationships between variables | Permanent (P)
|
Fixed-Term (FT)
|
Casual (C)
|
|||
|---|---|---|---|---|---|---|
| b (SE) | z | b (SE) | z | b (SE) | z | |
| CC→OID | 1.001 (0.069) | 16.27** | 1.001 (0.069) | 16.27** | 1.001 (0.069) | 16.27** |
| CC→TID | 0.687 (0.059) | 11.63** | 0.687 (0.059) | 11.63** | 0.687 (0.059) | 11.63** |
| CC→JAWS | 0.488 (0.044) | 10.97** | 0.488 (0.044) | 10.97** | 0.488 (0.044) | 10.97** |
| OID→JAWS | 0.211 (0.030) | 7.06** | 0.305 (0.093) | 3.27** | 0.215 (0.117) | 1.84 |
| TID→JAWS | 0.142 (0.031) | 4.45** | -0.042 (0.109) | -0.04 | 0.19 (0.142) | 1.34 |
| OID→OCB | 0.109 (0.031) | 3.48** | -0.025 (0.075) | 0.33 | 0.004 (0.097) | 0.04 |
| TID→OCB | 0.187 (0.031) | 6.01** | 0.187 (0.031) | 6.01** | 0.187 (0.031) | 6.01** |
| CC→OCB | 0.029 (0.046) | 0.62 | 0.029 (0.046) | 0.62 | 0.029 (0.046) | 0.62 |
| CC→IR | 0.612 (0.070) | 8.87** | 0.612 (0.070) | 8.87** | 0.612 (0.070) | 8.87** |
| TID→IR | 0.155 (0.048) | 3.23** | 0.155 (0.048) | 3.23** | 0.155 (0.048) | 3.23** |
| OID→IR | 0.450 (0.046) | 9.87** | 0.450 (0.046) | 9.87** | 0.450 (0.046) | 9.87** |
Note. Paths that varied across groups are in bold. CC = Communication Climate; JS = Job Security; OID = Organisational Identification; TID = Team Identification; JAWS = Job Affective Well-Being Scale; OCB = Organisational Citizenship Behaviour; IR = Intention to Recommend the organisation as a place to work.
*p < .05. **p < .001.
Discussion
These two studies aimed to explore (Study 1) and replicate (Study 2) whether work-status has a moderating effect on the antecedents and consequences of organisational and team identification. We demonstrated that the interaction between work-status and identification can offer interesting insights into the not-yet-clear impact of temporary work on employees’ relationship with their workplace. In two independent samples we found a moderating effect of work-status on the relationship between both foci of identification and job-affective well-being. Moreover, we saw a moderating effect of work-status on the relationship between organisational identification and OCB. All other relationships were equivalent across groups. With these results we provide the first evidence that work-status influences some consequences of organisational and team identification. We also found that the positive influence of communication climate and intention to recommend seem to be stable across groups and contexts; this means that in Chile and the UK and across different work status, there is a positive relationship between communication climate and intentions to recommend. This insight highlights the complex influence of unstable work-status on psychological variables across different work conditions, requiring further attention.
Contrary to previous work, we did not find mean differences across groups (except in JS) in either of the two studies. This similarity adds to the heterogeneous findings reported in previous literature (De Cuyper & De Witte, 2010) and suggests that non-standard work arrangements do not make individuals, per se, less identified, happy or committed. This finding might suggest that the contract individuals have with their employer might have less of an influence on their work attitudes compared to other contextual factors, such as the immediate work environment or relationships with colleagues. Job security, however, was clearly different across employees with different work-status, confirming previous studies that found lower levels of job security in temporary roles (Feather & Rauter, 2004; Veenstra et al., 2004), which was true for our Chilean and UK samples.
The relationships between variables were mainly reflective of our predictions. We found that communication climate has a positive relationship with foci of identification as well as job-affective well-being and intentions to recommend (but not with OCB). These results correspond to earlier work by Bartels and colleagues (2007) and strengthen the view that communication climate is crucial when creating an effective organisation. The effects of communication climate were not moderated by work-status, an effect found in both samples. This could be due to the fact that when organisations communicate honestly and effectively, those employees with a more unstable work-status know where they stand and can ‘trust’ the organisation. This finding has important implications for practitioners, as employers have control over the communication climate and is a parameter they can directly influence.
In relation to Organisational Citizenship Behaviour, Feather and Rauter’s (2004) findings were replicated as organisational identification had a small but significant influence on the level of OCB. Yet, this effect was found for permanent workers, but not in fixed-term and casual workers. Team identification had a similar impact on the level of OCB across all the groups. For practitioners, this finding implies that organisations should focus their efforts on strengthening team identification rather than organisational identification when attempting to increase OCB.
Veenstra and colleagues’ (2004) findings about the effect of casualisation on extra-role behaviour were challenged, as no differences in the level of OCB were found among the different work-statuses. This finding emphasises the importance of integrating experimental and observational research as well as data from different social contexts as results may vary considerably if people are responding based on their circumstances.
An association between organisational identification and the Intention to Recommend (IR) the organisation as a place to work existed across all work-statuses and its magnitude did not differ among them. In contrast, no association between team identification and IR was found. This result suggests that although team identification can have a useful impact in terms of OCB, it is not necessarily aligned with a positive opinion about the job conditions that the organisation offers. Therefore, promoting identification with the whole organisation by ensuring a positive communication climate could enhance intentions to recommend. On the other hand, promoting team identification positively influences OCB.
We allowed a covariance between job-affective well-being and intentions to recommend, which indicated that job-affective well-being and IR were positive correlated for fixed-term and casual workers (but not for permanent ones). This difference could be because permanent workers might take job-affective well-being for granted and hence make no strong connections between well-being and recommending their workplace. For employees with a less stable work-status, a workplace that fosters their well-being is one that they can recommend; they are less inclined to do so when their workplace compromises their well-being.
Interestingly, the influence of organisational and team identification on well-being differed across work-status groups and across countries. In Chile, organisational identification had a significant effect on well-being only in permanent and casual workers, but not in fixed-term workers; the opposite occurred with team identification, which only had an impact on fixed-term workers’ job well-being. These results differ somewhat from previous meta-analytic findings (Steffens et al., 2017) that did not find a significantly different correlation between organisational and team-identification and well-being and could also not detect an influence of culture.
To explain these results for the Chilean sample, we can return to Rousseau’s (1998) theory of identification (situated vs. deep-structure). We suggest that this result could correspond to different forms of organisational and team identification in casual and permanent workers: casual workers may derive their job-well-being from a situated identification with the organisation, as relying on their identification with the team could be less robust (e.g., they could leave soon and may expect that their team is not going to be there anymore). Next, fixed-term workers may have more options to extend or renew their contracts for a long-enough period to derive their well-being from their teams, but not sufficiently long to derive it from a deep-structure identification with the organisation. Finally, permanent workers may derive their well-being from a deep-structure identification with the organisation, as longer-term relationships that go beyond particular teams could be established.
In the UK sample, organisational identification had a significant relationship with well-being for permanent and temporary workers, but not for casual workers. Team identification was only significantly linked with well-being for permanent workers (but not for temporary and casual workers). That means, for casual workers, we could not see any significant relationship between identification and well-being. This absence could be due to the nature of the contract (e.g., often zero-hours), which might make it more difficult for workers on casual contracts to draw the benefits of a connection with the organisation or the team. Hence, as a zero-hour contract could mean that people have irregular working hours and relationships at work, even a situated identification with the organisation that provides them with a base for job-affective well-being might be difficult to achieve.
Another potential reason for this lack of effect could be the work environment itself. Earlier work by Broschak and Davis-Blake (2006) suggested that workplaces with high proportions of nonstandard work arrangements were associated with more negative attitudes towards the work environment. As such, it might be that people with casual contracts also work in an environment with a high proportion of non-standard contracts, which in itself creates a more challenging context that hinders a positive connection between identification and well-being. However, we only had a relatively small sample of casual workers and we did not have information on the proportion of others in non-standard work arrangements. Therefore, more research is needed to provide a more comprehensive answer.
Interestingly, the impact of OID on job well-being is stable across cultures for permanent employees, whose contracts are also more similar between both nations. In contrast, the nature of work in temporary and casual contracts differs more between countries, as it might be more affected by contextual particularities, like labour laws, culture and organisational practices. For example, many casual workers in Chile are not in zero-hour contracts, but in contracts based on the provision of services (see page 8), which could result in an experience that is very similar to that of a permanent worker, but with higher levels of job insecurity. This suggests that the work-status experienced by workers affects the dynamics of identification.
Taken together, the results highlight that we need better theoretical models to capture the dynamics and complexity inherent to identity processes in different organisational settings. The present study serves as only one stepping-stone toward this goal by highlighting how identification processes can be variable and stable across different groups. However, more work, especially longitudinal work, is needed to draw more attention to these processes.
In addition, our work answers the call for a more global perspective on organisational behaviour (Gelfand et al., 2008) by providing evidence from an underexplored context, namely Latin American (Nicholls-Nixon et al., 2011) and specifically Chile. Despite the slow rise of research from Latin America (see Nicholls-Nixon et al., 2011), there is still a paucity of work from these emerging economies countries such as Chile (Perez Arrau, Eades, & Wilson, 2012; Ronda-Pupo & Díaz-Contreras, 2014). For example, in Tsui, Nifadkar, and Ou’s (2007) review of cross-cultural work in top management journals, they found only one out of 93 papers that included a Chilean sample; in addition, only 13 among 517 authors were from Latin America and none from Chile. Perez Arrau et al. (2012) noted research from Chile is scarce and disarticulated and that Chile as a country has received little attention from researchers abroad. With the present study, we fill some of that gap and provide much needed empirical evidence for the idea that some elements of organisational behaviour are relatively stable across different cultures. Thus, across two different contexts, our work showed that people’s identification with the organisation as a whole, as well as with the team, was positively linked with job-affective well-being and intentions to recommend (and, to a certain extent, also OCB). Thus, despite differences in work-status and societal contexts, people’s experiences with work are grounded in their relationships with the organisation and the team. Identification is strongly connected with a positive communication climate, a relationship that is unaffected by work-status or context. That is, over and above contextual factors, communication climate can create structures that help to develop identities as it influences (and is influenced by) employees’ perception of relationship and communication within the organisation (Bartels, Douwes, De Jong, & Pruyn, 2006).
Limitations and Recommendations for Future Research
Firstly, although we presented a conceptual model showing specific associations between concepts, both studies were cross-sectional and therefore cannot be used to provide any evidence for this causality (Howitt & Cramer, 2014). Valuable insights would be obtained from a longitudinal study examining the directionality of effects (Gleibs, Noack, & Mummendey, 2010) as well as the stability and change of relationships between concepts (Gleibs, Mummendey, & Noack, 2008).
Additionally, all measures were obtained by self-report, which may have affected the results through response biases and/or common method variance (Donaldson & Grant-Vallone, 2002). For this reason, future research should aim to include assessments of supervisors or peers about the respondents’ level of OCB and other performance indices.
In addition to the limitations above, we also acknowledge that our data provided unbalanced groups and small samples sizes for the group of fixed-term and casual workers (especially in the UK). Although this asymmetry was to be expected given the population distributions (see ILO, 2016; OECD, 2018), this could lead to biased estimates (see, for example, Hox & Maas, 2001). Accordingly, our results should be interpreted with caution and should be further replicated in a larger sample.
Finally, the present study failed to provide clear evidence to support the need for regulating (and improving) the conditions of casual work across two countries. The obtained results challenge both common sense and explicit demands made by casual workers, which generally allude to conditions that would negatively impact their well-being (e.g., low pay, rights violations, and job insecurity; Valdebenito, 2016). Actually, Barría (2003) suggested that the levels of burnout in social workers in the Chilean public service were partially due to their WS (fixed-term or casual). In this sense, an explanation for the similar levels of job well-being is needed. However, we did not include or control for other variables such as ‘level of pay’, perception of rights, or other health and well-being indicators. In particular, measuring either well-being in broader terms (e.g., health), or the levels of burnout, could provide interesting data and drive changes in public policies because other researchers (Van Aerden et al., 2017) found that precarious work-status seriously harmed workers’ general health and well-being.
Our work offered mainly theoretical insights into the role of social identity in the context of temporary work. Further research is needed to identify practical courses of action and suggest concrete recommendations in terms of organisational management and public policies.
Conclusion
At the beginning of our paper we asked whether we have to rethink the relationship employees have with their organisation based on their unstable work-status. We can conclude that this is only the case for the relationship between organisational and team identification and job-affective well-being. Consequently, we only need a partial re-think. The relationship between communication climate, foci of identification and our outcome variables remains stable. Our work provided a first step in understanding the antecedents and consequences of social identification in the workplace that depend on work-status and examined this in two culturally different contexts and in a sample that is particularly hit by precarious work arrangements. As such, it offered useful guidelines to continue the development of this research field, particularly in the context of Chile and the UK where prolonged “temporary” roles can be found or are increasing.