Article body

The industrial relations system first developed by Dunlop (1958) is considered to be a grounding theory of mainstream industrial relations teaching and research in Canada (Meltz and Adams, 1993). In Canada, Craig’s model (1967), which is grounded in that of Dunlop, is often used. The essence of Craig’s system is that the actors (e.g., labour, management and government) of the system, through a series of conversion mechanisms (e.g., bargaining, grievances, etc.), take inputs from both the external environment (e.g., economic factors, legislation, etc.) and the actors themselves (e.g., goals, values, etc.), convert these inputs into organizational (e.g., union/management rights) and worker-oriented (e.g., wages, affect, etc.) outputs. These outputs, through a feedback loop, can then flow into the external environment, impacting the actors of the system.

In this system model, downsizing represents a conversion mechanism in the system as it leads to changes in the terms and conditions of employment. While current demographic issues and skills shortages mean that many organizations are struggling with labour shortages, this is not the case in all sectors of the economy. For example, the highly unionized areas of manufacturing/processing, natural resources, and the broader public sector have all recently faced significant challenges related to employment losses, downsizing and restructurings (see CBC, 2006; CLC, 2006; Fenton, Ip and Wright, 2001).

There are several reasons why human resource practices, including downsizing, would occur differently in unionized firms (see Freeman and Medoff, 1984; Ng and Maki, 1994; Slichter, Healy and Livernash, 1960; Verma, 2005; Wagar, 1997). First, unionized employees often prefer seniority-based practices as they are seen as less subjective and less influenced by managerial favouritism. Second, unionized workers often remain with a firm for a longer period of time given the advantages associated with seniority, thus reinforcing the desire for seniority-based practices/benefits. Third, building on exit-voice theory (Hirschman, 1970), unionized workers can voice their discontent through their union versus exiting the workplace. Through collective bargaining, workers often have increased power to convince employers to adopt practices that reflect worker preferences, compared to non-union employees in individual employment contracts. Overall, this results in human resources practices where senior workers are given preference over junior workers in pay, promotion, layoff, and other employment practices (Gersuny, 1982). These differences have led to a special kind of downsizing process known as bumping. Bumping is a process whereby union members with greater seniority, who are about to be laid off, can use their seniority rights to remove (or bump) more junior union members who otherwise would have been unaffected by the layoff, from their jobs (Brown and Beatty, 2005). In essence, the more senior employee takes the job of the more junior employee. Note that this often results in a cascading process whereby employees who are bumped from their positions can, in turn, use their seniority rights to bump more junior employees. In many cases, this bumping process will continue until there are no, qualified, junior employees left to bump. This means that despite the numerous layoff notices that may exchange hands during a bumping process, it is usually the most junior employees, who have no one to bump, who are truly laid-off and lose their employment with the firm.

While a complicated process, bumping is a way to clarify procedures concerning layoffs. People often join unions as a way to have influence on workplace decisions, to make workplace procedures more transparent, as well as to minimize managerial favouritism in terms of inconsistent application of rules, practices and standards (Hebdon and Brown, 2008; Freeman and Medoff, 1984; Slichter, Healy and Livernash, 1960). Thus, the bumping process adheres to long-standing union philosophies concerning the need to minimize favouritism, ensure clear workplace rules/practices, as well as the desire to provide increased benefits/protection for more senior employees.

While all bumping clauses discuss the ability of senior workers to bump more junior workers, there are several different types of bumping clauses. Our analyses of both the arbitration literature and a national database of collective agreements (Brown and Beatty, 2005; Government of Canada, 2006) revealed that one of the key differences in bumping clauses is the relative importance of seniority versus skill (e.g., the ability to perform the job). In our review, three types of clauses were found. In the first type, which we called ‘seniority bumping,’ seniority is the only factor examined. As long as an employee is more senior, he or she can bump the more junior employee. In the second type of bumping, which we called ‘minimum qualifications bumping,’ more senior employees can bump more junior employees as long they meet the minimum job requirements for the position in question. In the third type of bumping, ‘equal qualifications bumping,’ the issue of seniority only comes into play when the more senior employee and the more junior employee have equal job-related skills. In this case, employees can only use their seniority rights if they have equal (or better) job related skills than more junior employees.

Despite the number of unionized workplaces that exist, we could find no study in the broader literature that investigated union member reaction to bumping practices. This was surprising to us. Turning to the industrial relations system that grounds much of IR teaching and research, we hypothesized that workers’ previous experience with bumping, as a conversion mechanism, have attitudinal consequences on workers, given that an output of the system is worker reactions. We also hypothesized that, given the feedback loop, both direct experience in bumping as well as indirect experience with bumping (e.g., working in a department where bumping occurred) would also have an impact on worker attitudes. Thus, the objective of this study was to examine the impact that exposure to bumping (as a conversion mechanism) would have on the worker reaction outputs of the IR system. In particular, we sought to examine the attitudinal consequences that exposure to bumping had on perceived job security, job satisfaction, organizational commitment, union commitment and organizational justice. Given the lack of representation of bumping in the IR literature, we turned to the downsizing literature, more prevalent in psychology-based journals, to ground our study and our choice of variables. We did so because bumping is in essence a form of downsizing. Thus, it could be argued that union members experiencing bumping may experience similar affective reactions as non-unionized workers facing job loss through downsizing.

Effects of Downsizing

Employees who remain with the firm after downsizing are referred to as survivors and the negative effects they experience are often referred to as survivor syndrome (Baruch and Hind, 1999). Some commonly known adverse effects of survivor syndrome include reductions in perceived job security, job satisfaction, organizational commitment and perceptions of organizational justice (or fairness). In terms of the IR System, these adverse effects, or workers’ reactions, are considered to be outputs (Hebdon and Brown, 2008). In this section of the paper, we provide a brief overview of some of the research concerning these outputs, most of which has been conducted in non-unionized environments. In addition to these preceding measures, and given the unionized context of the present study, we also briefly overview the issue of union commitment.

Job Security

Arguably, one of the biggest impacts of downsizing on surviving employees is reduced job security. Despite its importance to the downsizing literature, Hellgren, Sverke and Isaksson (1999) argued that job insecurity has often been measured in an ad hoc manner. More specifically, they argued that a person’s perceived level of job insecurity could be conceptualized in two ways: quantitative or qualitative job insecurity. They stated that quantitative job insecurity included perceived threats to the future existence of one’s present job; qualitative insecurity referred to the perceived threats of impaired quality in one’s employment relationship (e.g., decline in working conditions, lack of career opportunities, decrease in salary, etc.). In a longitudinal study of a Swedish organization, these researchers found support for their two categories. They further found that the most important dimension, in terms of subsequent well-being and health, was quantitative insecurity. For these reasons, we used a quantitative measure of perceived job insecurity.

In a second longitudinal study, Moore, Grunberg and Greenberg (2004) compared over 1,200 white and blue-collar workers who reported 0, 1 or 2 contacts with layoffs throughout their careers. They found that workers with a greater exposure to direct (e.g., targeted for layoffs) and indirect (e.g., witnessed layoff of co-workers) downsizing reported lower levels of job security than workers with no experience with layoffs. This concept of direct and indirect experience to downsizing is germane to our study as workers can have direct (e.g., they were bumped) or indirect (e.g., bumping took place in their workplace) experience with bumping. An unanswered question is whether similar exposure to a bumping process, which by design provides increased job security for senior workers, impacts job security.

The concept of job security, rather than employment security, is crucial in this study. Bumping procedures are different from layoff procedures in that people who are bumped do not necessarily lose their employment, though they may be bumped out of their current job into a lesser one within the same organization. Due to the cascading effect of senior union members bumping junior union members, more people will experience job loss than employment loss. Moreover in this study, none of the participants: (a) experienced employment loss—all were still employed, (b) had ever been laid-off from the organization, and (c) had current layoff notices.

Job Satisfaction

Job satisfaction has traditionally been one of the most studied variables in the organizational psychology/behaviour literature (Brief and Weiss, 2002) and was included in the present study. Research has found that following organizational restructuring or downsizing, job satisfaction will decrease (Armstrong-Stassen and Cameron, 2003; Lee and Teo, 2005).

In a longitudinal study, Armstrong-Stassen and Cameron (2003) examined job satisfaction among healthcare workers in Canadian hospitals. They found that following a second wave of downsizing, nurses reported a significant decrease in job satisfaction concerning their work, their career future, and their hospital. This suggests that repeated exposure to downsizing has a more profound effect on employees than does a single exposure.

Organizational Commitment

Organizational commitment has also been extensively examined in the organizational psychology/behaviour literature. A recent paper has referenced almost 1000 studies on the topic (Cooper-Hakim and Viswesvaran, 2005). Organizational commitment refers to the connection that an individual has to an organization, and essentially involves three elements: (1) identification with the values and goals of the organization; (2) desire to remain with the organization; and (3) willingness to exert effort on behalf of the organization (Snape, Redman and Chan, 2000). These components make up what has been termed ‘affective’ commitment, or how closely a person relates to and is interested in being a part of, his or her organization (Meyer and Allen, 1984). A second form of organizational commitment, ‘continuance’ commitment, refers to how easy or difficult it is for an individual to leave his or her organization for another, based on previous personal investments into the organization (Meyer and Allen, 1984). As affective commitment has been more extensively studied in the organizational behaviour literature, and has dominated the union commitment literature (Snape, Redman and Chan, 2000), we have focused entirely on affective commitment.

Organizational commitment is important to employers given the well-established relationship between this measure and job satisfaction (Cooper-Hakim and Viswesvaran, 2005). Given the previously mentioned negative impact of downsizing on satisfaction, it should not be surprising that downsizing has been shown to have a negative impact on the organizational commitment levels of surviving employees (Kets de Vries and Balazs, 1997; Knudsen et al., 2003). A potential reason for the negative relationship between these measures was provided by Knudsen et al. (2003), who argued that affect may drop following downsizing due to increased employee workload. Thus, we included a measure of organizational commitment in this study to see if bumping experience would have any relationship with organizational commitment.

Union Commitment

Researchers have discussed the similarity between union and organizational commitment (Kuruvilla, Gallagher and Wetzel, 1993; Snape, Redman and Chan, 2000). Researchers have often presented a ‘parallels model’, with Snape et al.’s (2000) review of the literature concluding that organizational commitment was positively associated with union commitment. This model suggests that research concerning union commitment can be grounded in research from the organizational commitment literature, such as Mowday, Steers and Porter’s (1979) organizational commitment questionnaire.

In terms of union commitment, affective commitment is based on a perception of shared values, identity and pride in the union; whereas, instrumental commitment is based on the perceived benefits that come from the union or being a part of the union. Instrumental commitment is likely to be shorter term relative to affective commitment, which is seen as more long lasting (Snape et al., 2000). As we did with organizational commitment, we focus on affective commitment, measuring loyalty to the union, willingness to work for the union, and responsibility to the union.

Unlike the previously discussed measures, limited research has examined the relationship between union commitment and downsizing. One study by Turnley et al. (2004) investigated the relationship between breach of psychological contract, but not specifically downsizing, and union commitment among union employees. In essence, a psychological contract can be defined as employees’ beliefs concerning the mutual obligations between them and their organization. As such, it examines what employees believe their organization has promised them (e.g., competitive pay and benefits) in exchange for their contribution to the organization (e.g., a fair day’s work). A breach of such a contract occurs when an employee perceives that the organization has failed to deliver its promises (e.g., when it cuts jobs during downsizing). These researchers found that the breach of this contract, between the employee and the employer, can lead to an increase in union commitment. Limitations of that study include that it did not specifically examine union commitment in a downsizing situation.

We would assert that downsizing could be seen as a breach of psychological contract. Given that unionization represents a mechanism to restore equity in the employment relationship, one could argue that such a breach of psychological contract, between employee and employer (based on the findings of Turnley et al., 2004), could increase union commitment post-downsize, as unionized employees will usually turn to their union to seek recourse. In contrast, we could also argue the reverse. The fact that the bumping clause was negotiated in part by the union may result in decreases in union commitment. For example, workers bumped out of their jobs may blame their union for their being bumped, when the target was originally another employee. This could consequently reduce their commitment to the union. To date, the downsizing literature provides little guidance as to whether bumping would have a positive or negative impact on union commitment.

Organizational Justice

Organizational justice theory has become increasingly important in the workplace literature. In essence, it refers to employees’ perceptions of the fairness of procedures and outcomes related to workplace decisions (Greenberg, 1990). Justice is also important to union members as their union represents a voice mechanism through which they can bring about fairness in their workplaces (Freeman and Medoff, 1984).

There are three commonly identified types of organizational justice, all of which are associated with downsizing and survivor syndrome: distributive justice (e.g., fairness of the outcomes of decision-making), procedural justice (e.g., fairness of the procedures used to make decisions), and interpersonal justice (e.g., fairness of how employees were treated during downsizing; Kernan and Hanges, 2002; Mishra and Spreitzer, 1998; Thornhill and Saunders, 1998).

Two outcomes commonly associated with distributive justice during downsizing are: (1) the outcomes of the victims, since survivors often identify with the victims as either friends or colleagues, and (2) the distribution of resources among the remaining survivors (Mishra and Spreitzer, 1998).

Procedural justice is important as the procedures used to arrive at layoff decisions should be seen as fair by employees (Thornhill and Saunders, 1998). For example, when the layoff decision is based on individual success or ability, a survivor is more likely to evaluate downsizing as predictable and less threatening. If the decision is seen as being politically driven or random, it is less likely to be perceived as fair and survivors are likely to react negatively (Mishra and Spreitzer, 1998).

Finally, employee perceptions about the fairness of the way that they, and victims of the downsizing, were treated during downsizing are important for organizational justice (Thornhill and Saunders, 1998). People value their relationships with groups and organizations and expect to be treated fairly in such relationships (Kernan and Hanges, 2002). Thus, survivors react more positively to the downsizing process if they are treated with dignity by the organization (Mishra and Spreitzer, 1998).

As such, the current study has used a measure of organizational justice that includes all three elements: distributive, procedural and interpersonal. More specifically, we examined issues related to the outcomes, procedures and interpersonal treatment of downsizing decisions related to bumping.

Purpose of the Present Study

We could not find a study that examined union member reaction to bumping-based downsizing. Our purpose was to investigate the reactions of union members to such procedures. Specifically, we set out to investigate employee reactions in terms of job security, organizational justice, job satisfaction, organizational commitment and union commitment in a sample of union members who were employed in a workplace that used a bumping process during a recent reorganization. Based on a review of the downsizing literature, we examined two research questions:

  • First, is there a relationship between an employee’s total experience (or exposure) with bumping and each of the five employee affects under investigation?

  • Second, is there a significant difference between workers who have direct bumping experience and those with indirect bumping experience in terms of the five affective measures?

Methodology

Sample and Procedure

The current study was conducted in summer 2006 in the province of Newfoundland and Labrador, the easternmost province of Canada. All participants were members of a provincial public sector union who were employed by a province wide distribution and retail operation. All were currently employed with the organization where the study took place. The organization consisted of over twenty retail stores, one central warehouse and several distribution centres located throughout the province. All participants worked either in a retail or warehouse capacity.

This particular workplace was chosen as the organization was undergoing reorganization and a bumping policy had been utilized for the first time. Thus, bumping was a salient issue for all members of the organization (i.e. union leaders, employees and managers). Note that all participants were covered by the same collective agreement and the same bumping policy. The policy being utilized was essentially a ‘minimum qualifications’ bumping policy, protecting senior union members from layoffs by allowing them to bump employees of lesser seniority out of their job, provided that the senior member’s qualifications met the minimum required standards for the job.

The typical respondent was, on average, 40 years old, with a median tenure of four years; 72% of respondents were female. Since an anonymous survey was used, we could not compare respondents to non-respondents in order to assess non-response bias. However, the human resources manager of the organization in question informed us that the demographics of our sample were representative of the unionized workforce.

Prior to sending out a survey, members of the research team met with representatives of union and management to discuss the project. Both were very supportive and endorsed the project, as they too hoped to understand how workers react to bumping. Subsequently, personalized cover letters and surveys were distributed to all 411 union members of the organization. Fourteen uncompleted surveys were returned, as the addressees were no longer employed with the organization, leaving 397 valid surveys having been sent out. Completed surveys were received from 100 union members (N = 100), for a response rate of 25.2%. This response rate is consistent with other studies involving Canadian union members, where response rates approached 30% (Barling et al., 1992; Chaulk and Brown, 2008).

Completed surveys were returned directly to the research team (at their university address) using a prepaid postage envelope attached to each survey. Seven days after the surveys had been distributed to union members, we sent an e-mail message to all employees (via the organization’s intranet) encouraging them to complete surveys if they had not already done so.

As we were using a cross-sectional survey, we were aware that common-method variance could be an issue. The survey method relies on participants to provide as reliable responses as possible to a number of questions pertaining to a number of different issues (Kline, Sulsky and Rever-Moriyama, 2000). There is often concern that relationships, which are found through conducting correlations, may be partly due to a “spurious” relationship. For example, if a participant reports a high amount of experience with bumping, they may report a low level of organizational commitment due to some common third, unmeasured, variable. Following the guidance of Kline et al.’s (2000), we used a number of reverse coded items in our key measures in an effort to minimize this issue.

Key Measures

The key variables in the current study included bumping experience, job security, job satisfaction, organizational commitment, union commitment and organizational justice. Each measure is presented below. Note that the employer and union, who fully supported the project, were very concerned that a long survey would reduce response rates. This was particularly important as many of the employees being surveyed worked in retail positions with limited opportunities to complete the survey. Following the advice of the union and the employer, we removed some of the items from the original scales used to create our key measures. This decision was consistent with other studies conducted in unionized workplaces where scales have been modified to reduce survey length (Bemmels, Reshef and Stratton-Devine, 1991), with some researchers using only single measures of employee affect in unionized settings (e.g., Fryxell and Gordon, 1989).

Bumping Experience. Total experience with bumping acted as an independent variable. Given that we could find no studies that assessed bumping experience, we adapted our measure from the layoff contact measure of Moore et al.’s (2004). However, we adapted the questions to better reflect the bumping processes of the organization being studied. Specifically, a total bumping experience score was created by summing each employee’s experience across four survey items: the indirect experience of bumping occurring at an organizational level; the indirect experience of bumping at the work place level; the direct experience of being bumped out of a position; and the direct experience of bumping another union member out of their job. We adopted the yes-or-no response format used in Moore et al.’s (2004) study, and subsequently created three groups from those responses: (1) no bumping experience, (2) indirect bumping experience only, and (3) direct bumping group.

Job Security. Perceived job security was measured using three items adapted from Oldman et al. (1986). These questions assessed quantitative job (in)security, or perceived threats to the future existence of one’s present job, using a 5–point Likert type scale.

Job Satisfaction. Job satisfaction was measured by adapting 10 items (two selected from each facet) from the Job Descriptive Index (JDI) developed by Smith (1974). Each question was assessed using a 7–point Likert type scale.

Organizational Commitment. Organizational commitment was assessed using a 7–point Likert type scale. In total, 12 items from Mowday, Porter and Steers (1982) measure were used.

Union Commitment. Union commitment was measured using 10 items adapted from Bayazit, Hammer and Wazeter’s (2004) and a 5–point Likert type scale. The current study included selected items from each facet of union commitment identified by Bayazit et al.’s (2004), namely, feelings of loyalty to the union, willingness to work for the union, and responsibility to the union.

Organizational Justice. Organizational justice was assessed using a 5–point Likert type scale and 11 items adapted from Colquitt (2001). The wording of the items was adjusted to: (a) relate to the bumping process, and (b) represent union members’ feelings of the bumping process and outcomes in general, rather than only focusing on participants’ personal experiences. This latter decision was made as not all union members were directly impacted by the bumping procedure in our study. For example, the question “Is your outcome justified, given your performance” was changed to “The decision to bump a particular person has usually been justified given his/her performance.”

Results

Preliminary Analysis and Correlations

Scales were created for the five attitudinal measures and reliabilities were assessed using Cronbach’s alphas. Cronbach’s alphas for organizational commitment (α = .83), union commitment (α = .83), job security (α = .74), organizational justice (α = .85) and overall bumping experience (α = .74) measures were all acceptable. The job satisfaction scale obtained a Cronbach’s alpha of .62. Analysis revealed that removing items did not bring Cronbach’s alpha above .70 for the latter measure, thus, no items were removed.

Once scales were created, two-tailed Pearson Correlations were conducted. The means, standard deviations and inter-correlations are reported in Table 1. In response to research question 1, results revealed a negative correlation between amount of experience with bumping and both organizational commitment (r = –.27, p < .01) and job satisfaction (r = –.24, p < .05). Thus, the more experience a union member had with bumping, the lower his/her organizational commitment and job satisfaction. Furthermore, there was a negative relationship (at the .10 level) between amount of bumping experience and both organizational justice (r = –.19, p = .09) and union commitment (r = –.20, p = .06). This suggests that increased experience with bumping is associated with decreases in union commitment and perceptions of fairness (as measured by organizational justice). However, the relationships between bumping experience and job security was not significant (r = –.08, p > .05).

Table 1

Means, Standard Deviations and Correlations

 

Mean (sd)

1

2

3

4

5

6

Bumping experience

1.04

(1.25)

.74

 

 

 

 

 

Organizational commitment

62.88

(12.79)

‑.27**

.83

 

 

 

 

Union commitment

34.55

(8.11)

‑.20+

.28**

.83

 

 

 

Job security

10.03

(2.85)

‑.08

.30**

.10

.74

 

 

Job satisfaction

32.21

(6.10)

‑.24*

.72**

.49**

.37**

.62

 

Organizational justice

31.26

(8.24)

‑.19+

.40**

.27*

.25*

.46**

.85

Note: Cronbach’s alpha coefficients reported on the diagonal. ** Correlation is significant at the 0.01 level (2–tailed); * Correlation is significant at the 0.05 level (2–tailed); + Correlation is significant at the 0.10 level (2–tailed).

-> See the list of tables

Direct vs. Indirect Bumping Experience

Research question two examined whether there was any difference between union members who directly experienced bumping, those who only had indirect experience and those participants with no experience with bumping at all, with regard to all five dependent variables. In order to determine whether or not such differences existed, three separate levels of bumping were created: those who had no experience with bumping (no experience, n = 48), those who experienced bumping only as it occurred around them in the workplace (indirect experience, n = 26) and those who had directly experienced bumping by having been bumped themselves or having bumped someone else (direct experience, n = 22).[1] Note that four respondents did not complete this question. An analysis of variance (ANOVA) was conducted on each of the five dependent variables.

ANOVA revealed a significant difference between groups for both organizational commitment (F = 7.33, p < .01) and job satisfaction (F = 3.29, p < .05). ANOVA also found a main effect (at the .10 level) for union commitment (F = 2.40, p = .10). The means and standard deviations for each of these three variables (by level of bumping experience) are presented in Table 2. Note that bumping experience did not have significant effects on job security (F = .63, p > .05) or organizational justice (F = 1.14, p > .05).

Table 2

Means and Standard Deviations by Bumping Experience

 

Type of Bumping Experience

Mean

Standard Deviation

Organizational commitment

No experience

67.70

10.02

Direct experience

59.10

15.37

Indirect experience

57.44

12.16

Union commitment

No experience

36.27

7.96

Direct experience

33.19

8.28

Indirect experience

32.16

8.09

Job satisfaction

No experience

33.79

5.74

Direct experience

30.41

6.82

Indirect experience

30.85

5.88

-> See the list of tables

After determining that a significant difference existed between bumping experience and organizational commitment (.01 level), job satisfaction (.05 level) and union commitment (.10 level), two-tailed, independent sample t-tests were conducted between the levels of the bumping variable to better understand the main effects revealed by the ANOVAs.

When considering organizational commitment, there was a significant difference found between those participants with no bumping experience and those with both indirect bumping experience (t = 3.82, p < .001) and those with direct bumping experience (t = 2.74, p < .01). No significant difference existed between those who experienced bumping indirectly and those who had direct bumping experience (t = .41, p > .05). This suggests that it is not so much the type of bumping experience a participant had, but rather, whether or not a participant had any experience with bumping that made a difference to feelings of organizational commitment.

When examining job satisfaction, there was a significant difference found again between those who had no experience with bumping and both those who had only indirect experience with bumping (t = 2.09, p < .05) and those who had direct experience with bumping (t = 2.16, p < .05). Again, no significant difference existed between those with indirect and those with direct experience with bumping (t = .24, p > .05). Thus, these results support that what is important to feelings of job satisfaction is whether or not you have had any bumping experience, and not the type of experience you have had.

Independent sample t-tests were also conducted between groups on the union commitment measure, since a difference at the .10 level had been found. Similar to organizational commitment and job satisfaction, there was no significance found between those with indirect versus direct bumping experience (t = .43, p > .05) and a significant difference was found between those participants with no bumping experience and those with only indirect experience (t = 2.06, p < .05). However, no significant difference was found between those with no bumping experience and those with direct bumping experience (t = 1.44, p > .05).

While the preceding analyses suggest that bumping experience impacted organizational commitment, union commitment and satisfaction, we need to take into consideration that all three variables were moderately correlated (all greater than r = .25 as shown in Table 1). Thus, consistent with Moore et al. (2004), we ran a multivariate analysis of co-variance (MANCOVA), controlling for age and tenure by using them as covariates. The means and standard deviations resulting from this MANCOVA are presented in Table 3. The Wilks’ Lambda value (.83) indicated that the multivariate effect of bumping experience was significant (F = 2.57, p < .05). Moreover, a review of the between subjects effects for the model produced results that mirrored those from the previously reported ANOVAs. Specifically, there was a main effect for job satisfaction (F = 5.44, p < .001) and organizational commitment (F = 5.41, p < .001) while the effect for union commitment approached the .10 level (F = 1.86, p < .13).

Thus the remainder of our MANCOVA discussion focuses on the variables of job satisfaction and organizational commitment. The pairwise comparisons produced by the MANCOVA analysis were examined. These comparisons demonstrated that levels of job satisfaction differed for those with no direct experience and both indirect and direct experience at the .05 and .01 levels, respectively. Job satisfaction did not differ between people who had direct versus indirect experience (p > .05). Similarly, the organizational commitment levels of workers with direct and indirect bumping experience was lower (at the .01 level) than that of workers with no bumping experience, however, they did not differ from one another (p > .05).

Table 3

Key Variable Means and Standard Deviations by Bumping Experience (MANCOVA)

 

Type of Bumping Experience

Mean

Standard Deviation

Organizational commitment

No experience

67.66

9.95

Direct experience

58.20

15.20

Indirect experience

57.67

12.36

Union commitment

No experience

36.61

7.70

Direct experience

32.70

8.18

Indirect experience

32.33

8.21

Job satisfaction

No experience

34.48

5.14

Direct experience

29.75

6.61

Indirect experience

30.79

6.13

-> See the list of tables

As these MANCOVA results now confirm, employees’ experience with bumping can impact employee affect. Moreover, these results suggest that the effect of bumping on both job satisfaction and organizational commitment is negative.

Discussion

Implications of the Study

From an IR perspective, the IR system has grounded much of the teaching and research in North America. The current study set out to investigate how the conversion mechanism of a bumping system could impact the outputs of the IR system related to worker reactions. More specifically, our primary goal was to investigate the impact that experience with bumping had on worker affect (e.g., organizational commitment, union commitment, job satisfaction, perceptions of job security and organizational justice). Consistent with the systems view, we did find that the differing exposures to the bumping conversion mechanism resulted in differing outputs (e.g., worker reactions).

Perhaps the most significant implication of our investigation is that we fill current gaps in both the industrial relations and downsizing literatures. The former has not examined the role of employee reactions to downsizing. In contrast, the latter has largely focused on employee reactions to downsizing in non-unionized workplaces. Thus, it has neglected the effects of downsizing through a bumping procedure. This is of key importance as bumping is essentially a special type of downsizing procedure and is often used in unionized workplaces. As such, our findings have a number of implications for researchers, managers and union leaders alike.

First, consistent with previous studies on the effects of downsizing on organizational commitment (Kets de Vries and Balazs, 1997; Knudsen et al.’s, 2003) and job satisfaction (Armstrong-Stassen and Cameron, 2003; Lee and Teo, 2005), the current study revealed that experience with bumping had a negative effect on organizational commitment and job satisfaction. The more experience union members had with bumping, the lower their organizational commitment and job satisfaction. The results further revealed that the negative effects were true for union members with either direct or indirect bumping experience. For management, it is clear that those employees who experience bumping, even if only indirectly through the awareness of bumping occurring in the workplace, will have lower job satisfaction and organizational commitment.

Second, of importance to researchers and union leaders alike, are the results pertaining to union commitment. The relationship between experience with bumping and union commitment appears to be somewhat more complex than that of organizational commitment and job satisfaction. The ANOVA results revealed a significant difference between union members with no experience with bumping and those with indirect experience. No significant differences were found between members with no experience and direct experience, or between workers with direct and indirect experience. However, the effect of union commitment was not significant in the multivariate analysis. Taken together, the best case scenario for unions is represented by the MANCOVA results, suggesting that the overall impact on union commitment is neutral—it does not change based on bumping experience. Worst case scenario, the ANOVA results suggest that it is the group of workers who observe the impacts of bumping on co-workers (i.e., the indirect experience group) who experience a reduction in commitment to their union. Given that similar decreases in organizational commitment have been found among survivors post-layoff, this suggests that union leaders and researchers need to examine the survivor literature concerning ways to potentially address this commitment issue.

Of additional significance to the issue of union commitment is the work of Mellor (1992). That study suggested that the relationship between union commitment and downsizing can be influenced by the perceived responsibility of the union in causing the layoffs. Unfortunately, we did not assess whether workers perceived the layoff to be the responsibility of union or management. This decision was made given the previously discussed request of both management and union regarding survey length. Nevertheless, future research on bumping should now include this variable.

Third, our findings differed from the previously presented downsizing literature on two key variables. Specifically, our analyses failed to detect significant differences in the job security or organizational justice levels of study respondents by bumping experience. This was surprising given that research we cited previously: (a) suggests that reduced job security is one of the most common effects following downsizing (Hellgren et al.’s, 1999); and (b) highlights the importance of organizational justice in layoffs. Perhaps these differences reflect the context of the present study. Our study was conducted in a unionized environment, where there may be inherent differences from the more commonly studied non-unionized environment which grounds most of the downsizing literature. For example, the significance that has been given to seniority in a unionized environment as a way to provide job security, the fact that all downsizing procedures are included in collective agreements (and thus known to all workers), etc., would not hold true in non-union workplaces.

In terms of perceptions of fairness, as assessed by organizational justice, we did find a negative correlation between the amount of bumping experience and perceptions of fairness. Thus, it is possible that the lack of significant ANOVA results is due to the small sample size (N = 100) and, therefore, type II error. Moreover, the questions themselves may have been seen as threatening or sensitive to study participants. Upon examination of the data, we noted that 14 participants did not answer any (or all) of the survey items pertaining to organizational justice. This non-response bias was much higher than all other survey questions and resulted in a sample size of only 86 people for that measure. It is possible that participants found these questions pertaining to the fairness of management particularly threatening due to the fact that management had been investigating union members’ feelings and opinions toward the bumping policy and its implementation, prior to the start of the current study.

Overall, we believe that the inconsistencies between our results and the broader downsizing literature indicate that not all relationships found previously will generalize to a bumping-based downsizing process in a unionized workplace. Clearly, this suggests the need for more research in this area.

Fourth, the current study also has practical benefits for managers and unions. It is important to both as it can help them identify employee experiences with the current bumping policy, which may lead to mutually agreed upon changes to the policy. Of practical importance is that the identification and awareness of any potentially negative effects of a bumping policy can enable the stakeholders to develop interventions to address such effects. For example, by investigating and understanding the potentially negative reactions to bumping, management and union members may openly communicate the rationale for, and specific procedures used in such processes in their discussions with workers and/or creation of contract language.

Limitations and Future Research

The current study has a number of limitations. First, the relatively small sample size (N = 100). This small sample size may potentially explain the null results found in this study. Moreover, relationships found at the .10 level may have been strengthened to the .05 level if a larger sample had been used. This suggests the need for our study to be replicated with a larger sample.

A second limitation is non-response bias. The current response rate, although consistent with previous studies, was approximately 25%. The possibility exists that those who completed the survey differed in some way from those who did not complete it. Furthermore, when considering those participants who did complete the survey, there tended to be a lower response rate on the items pertaining to organizational justice as compared to the other items on the survey. As discussed previously, a higher response rate could have strengthened some of the results that were approaching the .05 level.

A third limitation, common to all cross-sectional survey studies, is common method variance. As we discussed previously, surveys rely on participants to provide as reliable responses as possible to a number of questions pertaining to different issues (Kline et al., 2000). Arguably, the significant correlation between union and organizational commitment could suggest that common method variance is a problem in this study. That being said, correlations between these measures are not uncommon (see Snape et al., 2000) and the correlation matrix revealed a number of insignificant correlations, suggesting that common method variance may not be an issue in this study. Note that we also followed the advice of Kline et al. (2000) to reduce common method variance through the use of reverse coded items. Nevertheless, this limitation cannot be ruled out in this study, suggesting that the time may be ripe for longitudinal surveys in this field of exploration.