Journal of Organizational Behavior

J. Organiz. Behav. 23, 425–453 (2002) Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/job.148

Boundaryless and traditional contingent employees: worlds apart

JANET H. MARLER1*, MELISSAWOODARD BARRINGER2

AND GEORGE T. MILKOVICH3

1School of Business, University at Albany, S.U.N.Y., U.S.A. 2Isenberg School of Management, University of Massachusetts, U.S.A. 3School of Industrial and Labor Relations, Center for Advanced Human Resource Studies, Cornell University, U.S.A.

Summary We find support for distinguishing between two types of contingent workers: boundaryless and traditional. The existence of these two types helps explain the conflicting views about contin- gent work in the literature. Using hierarchical cluster and multivariate analysis on a national survey of 614, and a regional sample of 276 temporary employees, we also find that these two types exhibit different work attitudes and behaviors. What distinguishes boundaryless tempor- aries from others is their preference for temporary work and level of skill and experience. Results also show that the performance of traditional temporaries is more sensitive to attitudes than boundaryless temporaries. After controlling for level of work satisfaction, traditional temporaries reported higher task and contextual performance. We discuss the implications of these findings for theory development, organization practice and public policy. Copyright # 2002 John Wiley & Sons, Ltd.

Introduction

One of the most notable human resource trends in the past decade has been the extraordinary growth in

contingent employment arrangements. With growth rates significantly higher than conventional

employment, contingent employment relationships have attracted both popular and academic atten-

tion, much of it contradictory. For example, several researchers indicate the trend is creating a worker

underclass (Kalleberg, Reskin, & Hudson, 2000), with low wages, no benefits, negligible job security,

little training and no possibility of advancement (Segal & Sullivan, 1997b; McAllister, 1997;

Kalleberg et al., 1997; Barker & Christiensen, 1998; Parker, 1994; Rogers, 1995, 2000). Symptomatic

of a new ‘political economy of insecurity’ in which ‘labour market flexibility has become a political

* Correspondence to: Janet H. Marler, School of Business, University at Albany, S.U.N.Y. Albany, New York 12222, U.S.A. E-mail: marler@albany.edu

Contract/grant sopnsor: Alfred P. Sloan Foundation. Contract/grant number: #96-6-9. Contract/grant number: #99-6-23.

mantra’ and risk is redistributed ‘away from the state and the economy and towards the individual’

contingent work is viewed as a threat to social welfare (Beck, 2000, p. 3). A competing view suggests

the trend is consistent with the emergence of boundaryless career strategies (Sullivan, 1999; Arthur

and Rousseau, 1995) stemming from employers’ emphasis on employability, flexibility (Capelli,

1999), and network technologies (Miles & Snow, 1996), and from an increasingly diverse work force

that includes growing demand for career flexibility among dual career families, single parents, and

older workers transitioning to retirement (Albert & Bradley, 1997; Bailyn, 1993; Capelli, 1999; Kohli,

1994; Moen and Yu, 1999).

Perhaps both views have merit. As Rogers (2000, p. 2) and others note, ‘temporary employment has

many faces’ and can not be summed up as either all ‘good’ or all ‘bad.’ Rogers (2000, p. 10) cautions

against assuming ‘that the ‘temporary’ component of contingent employment is the primary identify-

ing factor for these workers,’ although it is clearly an important and shared influence. She suggests that

the meaning of these work arrangements depends also on ascriptive attributes of the workers (e.g.,

class, gender, race) and conditions in the specific labor markets of their occupations.

Recent attempts to extend theory beyond the confines of assumed conventional employment rela-

tionships also yield conflicting views. For example, Tsui, Pearce, Porter and Hite (1995) suggest that

the primarily transactional nature of contingent arrangements may be detrimental to organizational

performance. In contrast, Matusik and Hill (1998) and Lepak and Snell (1999) theorize that, contrary

to popular belief, contingent employees may in fact create valuable new knowledge and more effective

organizations. There is agreement, however, that heterogeneity within these alternative employment

arrangements needs to be better understood in order to inform organizational and public policy makers

(Von Hippel, Mangum, Greenberger, Heneman, & Skoglind, 1997; Kunda, Barley, & Evans, 2002;

McLean Parks, Kidder, & Gallagher 1998; Gallagher & McLean Parks, 2000; Kalleberg, Reskin, &

Hudson, 2000).

Theoretical discourse about contingent work is part of a larger discussion about the changing work-

place. For instance, theorists adopting a postmodern perspective examine the implications of a ‘pro-

found transformation in the logic of modern, industrial society’ (Gephart, 1996, p. 34). Although the

meaning of this transformation is context-specific, many of the ‘stories’ told by postmodern theorists

reflect a general trend toward uncertainty and risk in the workplace. Gephart (1996, p. 38), for exam-

ple, posits that, in the postmodern organization, ‘commitment and motivation will likely be trans-

formed into quasi-contractual commitments to particular projects undertaken by transient work

teams composing temporary organizational fiefdoms.’ Smith (2001, p. 3) draws on the traditions of

labor process theory, flexible specialization, and dual labor markets to ‘untangle the contradictory

strands present in the contemporary transformation of jobs and work organizations.’ She challenges

the simplistic division of today’s work into ‘bad’ contingent jobs and ‘good’ permanent standard jobs,

and argues ‘that uncertainty and unpredictability, and to varying degrees personal risk, have diffused

into a broad range of postindustrial workplaces’ (Smith, 2001, p. 7).

Our study examines the changing landscape of the postindustrial workplace by focusing on the fast-

est growing segment of the workforce and contingent employment arrangements: temporary agency

employees (Kalleberg, Reskin, & Hudson, 2000; Segal, 1996). To explore the differences that exist

among workers in this sector of the labor market, we used a national probability sample and a regional

sample of temporary agency employees. Adopting a rich multi-disciplinary framework drawing from

economics, industrial organizational psychology, organizational theory, and sociology, and an explora-

tory empirical methodology cluster analysis, we propose and empirically test the existence of two pri-

mary types of temporary agency employees—traditional and boundaryless. Traditional temporaries

represent the typical notion of temporary employee, someone with limited skills, hoping to find stan-

dard work within a bureaucratic organization (Rogers, 2000). Boundaryless connotes being outside or

opposite to being within one organization (Arthur & Rousseau, 1995) and boundaryless temporaries

426 J. H. MARLER ET AL.

refer to individuals whose current career consists of ‘quasi-contractual’ commitments/temporary pro-

jects voluntarily arranged through temporary employment agencies.

We believe these two types exemplify the different perspectives in the literature on contingent

employment. We show that these two types differ significantly along several theoretically derived

dimensions related to skill and preference, and also to work attitudes and behaviors. These findings

are important because they provide generalizable empirical evidence of significant heterogeneity in

contingent employment, which only recently is acknowledged as a developing trend, and underscores

the need to take diversity into account when developing organizational strategies and when making

pay and public policy decisions.

The paper is divided into three sections. The first section develops a theoretical perspective that sup-

ports the expectation of heterogeneity within temporary labor markets. Section two describes the

methods and results of the empirical analyses. We conclude with a discussion of our findings and

the implications for practice, research and policy.

Definitions and Theoretical Background

The temporary workforce

We use the BLS definitions of contingent work and temporary employment arrangements in this paper.

They capture the most important aspect of this work: the externalization and weak connection between

workers and firms (Pfeffer & Barron, 1988). Contingent work is defined as ‘any job in which an indi-

vidual does not have an explicit or implicit contract for long-term employment’ (Polivka, 1996a, p. 4).

Temporary refers to jobs that are arranged by an employment intermediary or agency. Under this defi-

nition, about two-thirds of temporary employees are also contingent (Polivka, 1996a) and they com-

prise somewhere between 1.5 and 5 per cent of all workers (Cohany, 1998; Segal, 1996).

A demand and supply framework: heterogeneity in skills and preference

Heterogeneity within the temporary work force is influenced by both demand and supply factors.

Demand factors are associated with varying skill requirements, increasing demand for flexibility, cost

reduction, and the growth of intermediaries such as temporary help agencies to satisfy these organiza-

tional demands. Demand factors have implications for supply factors, which we develop in terms of

differences in individual preferences for contingent work.

Demand factors: variable skill requirements, cost reduction and temporary help agencies We propose a two-part explanation for increasing demand for temporary employees. The first relates to

changes in organizations’ competitive strategies that have created demand for both high and low skill

workers. The second links the growth of intermediary institutions to the increased attractiveness of

temporary work arrangements to both organizations and workers.

One view of the decision to use temporary employees suggests that it is part of an organization’s

strategic process aimed at increasing profitability and shareholder return. Lepak and Snell (1999) sug-

gest that organizations adopt staffing strategies that are consistent with their particular human resource

needs. Drawing on resource-based theory (Barney, 1991; Wright, McMahan, & McWilliams, 1994),

they argue that organizations enhance their value by determining which human resource capabilities

should remain within the firm (specific skills) and which should be acquired in the market (general

BOUNDARYLESS AND TRADITIONAL CONTINGENT EMPLOYEES 427

skills). Knowledge, skills, and abilities that contribute to a firm’s unique capabilities should be devel-

oped within the firm and those that don’t should be acquired on demand in the external labor market.

Capabilities categorized as low value do not justify a high level of commitment and are best acquired

in the external market.

This logic is also consistent with the economic transaction cost model (Williamson, 1990). Transac-

tions that are neither unique, nor asset specific, can be easily imitated and supplied in the market. With

many market alternatives, the more costly bureaucratic internal labor market associated with conven-

tional full-time employment arrangements is not justified or competitive. Capabilities that are too spe-

cialized are needed too infrequently to justify incurring full-time employment costs, such as

the overhead associated with recruiting, compensation and benefit administration, training, perfor-

mance evaluation, and so on. When economies of scale are not achieved, these skills are better

acquired in the external labor market on a temporary basis.

Under these perspectives, organizations ‘buy’ more skills in the external labor market by externaliz-

ing certain employment relationships (Pfeffer & Barron, 1988; Abraham & Taylor, 1996). As noted

above, the different circumstances under which these organizations decide to adopt such a strategy will

be associated with different decisions about the skills that are purchased. Lepak and Snell (1999)

identify two types of external employment arrangements: out-sourcing and partnering. Out-sourcing,

which involves acquiring low value, low unique skills, is akin to hiring temporary office staff. Partner-

ing entails engaging someone with unique skills for a short period of time (e.g., a temporary software

engineer). Lautsch (1999) made similar distinctions in her qualitative study of the organizational use

of temporary workers across occupations and skills. She identified two main strategies: cost-reduction

and flexibility. A flexible strategy, like partnering, generally involves higher skilled temporaries.

These workers are integrated with regular employees, performing the same tasks and often earning

greater wages, but have no implicit contract for long-term employment. The cost-reduction strategy

corresponds roughly to outsourcing in that it is achieved by employing temporaries with fewer

skills.

These emerging staffing strategies have resulted in a dramatic increase in the demand for both high

and low skill temporary workers, and, in turn, favorable environments for temporary help agencies.

Increased outsourcing and lay-offs create economies of scale for the temporary help industry, lowering

their costs and increasing their profits (Segal & Sullivan, 1997a). With increasing profitability, the tem-

porary help services industry has grown dramatically, providing a greater supply and variety of tem-

porary services. Greater variety, availability, and lower costs make temporary employment more

attractive to organizations, and encourage further substitution of temporary workers for permanent

workers at both low and high skill levels. Indeed, the number of temporary help service (‘THS’) com-

panies in the United States grew from 3133 to 10 611 between 1975 and 1987 (Carnoy, Castells, &

Benner, 1997). Further, 20 per cent of new jobs created between 1991 and 1993 in the United States

were temporary jobs (Von Hippel et al., 1997).

In sum, theory and some evidence suggest that the demand for temporary labor is not homogeneous.

Temporary workers with both low and high skills are becoming increasingly attractive to organizations

seeking to minimize labor costs and maximize flexibility. Thus we would expect to see the existence of

not just lower skilled temporaries such as clerks and receptionists but also temps with specialized skills

such as accountants, lawyers, engineers, and medical professionals. We next consider the attributes of

workers who may be willing to accept such employment.

Supply factors: human capital theory, boundaryless careers, and preferences for temporary work Human capital theory suggests that individuals do not invest in skill development unless the present

value of the returns on these investments is greater than the present value of the costs (Becker, 1992).

428 J. H. MARLER ET AL.

Individuals who invest in their own skill development by, for example, attending college or participat-

ing in a lengthy apprenticeship, expect to earn higher salaries to justify the out-of pocket costs of edu-

cation, the foregone earnings, and the effort required for their investment. The investment is made in

expectation of higher earnings over a substantial period of time. Workers who have made such invest-

ments would prefer longer-term employment. According to this logic, the only workers who would be

willing to supply labor to temporary help agencies would be those with few human capital investments.

Generally, the empirical evidence supports this prediction. Cohany (1996) shows that the majority of

temporary workers is younger and less educated than traditional workers, and are therefore more likely

to have less experience and formal training.

Human capital theory also postulates the conditions under which organizations choose to invest in

developing employee skills (Becker, 1992). Few employers will invest in general skills training (those

skills that are standardized and can be easily duplicated in the market, e.g., computer programming)

and if they do, then the employee bears the costs through lower wages (see Bishop, 1994 for a review).

Organization specific training, in contrast, involves developing skills that are valuable primarily to the

organization. Unlike general training, the employee shares both the benefits (e.g. higher wages after

training) and costs (e.g. lower wages during training) of training. In the case of temporary workers,

however, neither the organization nor the worker could expect to reap the benefits and recoup their

costs. Hence, temporaries have fewer opportunities and less incentive to invest in firm-specific skills,

but without such skills they are less productive and earn lower wages. Because of such limitations on

skill and earnings growth, employees may prefer the longer-term employment arrangements of inter-

nal labor markets, where their skills earn higher wages and are the basis for upward mobility

(Osterman, 1984).

In sum, human capital theory implies that few workers should be interested in working in temporary

positions; most will seek positions where specific skills should result in higher wages for a longer

duration. Those who do accept temporary jobs are likely to have few skills and a strong interest

in gaining access to a traditional job, where they can develop specific skills, earn more and gain job

security.

These conclusions rest on two assumptions. The first assumption is that workers’ decisions about

employment are primarily influenced by financial outcomes. A substantial literature, however, finds

that the value, or importance, of work outcomes varies across individuals (Bartol & Locke, 2000).

For example, some workers may place a high value on variety, or autonomy. Bradach’s (1997—unpub-

lished manuscript) interviews of temporary executives and computer programmers provide evidence of

such values. He found that many preferred to maintain their independence from any one organization.

Similarly, Kunda, Barley, and Evans (2002), in their sample of technology contract workers, and

Rogers (2000) in her sample of temporary lawyers found that many liked working outside organiza-

tional boundaries because of the flexibility and the ability to separate themselves from organizational

politics, incompetence and inequities. Finally, a study by the National Association of Temporary Staff-

ing Specialists (NATSS) identified a group that consisted primarily of older females who desired flex-

ibility but were not the family breadwinners (Steinberg, 1994).

The second assumption is that temporary employment prevents high-skill workers from recouping

their investments in human capital. However, as noted above, the outlook for workers in the temporary

help industry is changing. As organizations continue to increase their use of temporary workers, tem-

porary help agencies increase their volume assignments. With greater volume, these agencies are pro-

viding the work continuity and variety necessary to induce more workers to consider temporary

assignments. Hence, the increasing demand for temporary workers has diminished the risks associated

with temporary job insecurity. Individuals with specialized skills can be more assured of a string of

opportunities, which mitigates a major disadvantage that may have served to inhibit them from con-

sidering temporary work in the past.

BOUNDARYLESS AND TRADITIONAL CONTINGENT EMPLOYEES 429

Furthermore, temporary workers may actually increase their value in the marketplace if, by holding

jobs in multiple firms, they are able to accumulate general skills valued by organizations. General skill

accumulation is defined as the level of investment in human capital, general skills, experience, and

transferable competencies (Becker, 1992; Sullivan, 1999). Baker and Aldrich (1996) argue that knowl-

edge accumulation is an important attribute of a career. It represents skill development (general and

firm specific) that evolves from work experience and training, both of which influence opportunities

for further development and wage growth. According to this logic, experience gained from working for

multiple firms facilitates the accumulation of transferable specific skills and increases individuals’

marketability and earnings potential. The accumulation of skills may be particularly favorable for indi-

viduals trained in certain recognized and codified occupations or professions, such as accountants,

lawyers, engineers, where general skills can be augmented with multiple firm-specific experiences

(Tolbert, 1996) or for those in craft-like occupations (Osterman, 1984). Thus, it isn’t the case that valu-

able skill accumulation can only occur within an organizational boundary or that only firm-specific

skills carry wage premiums as human capital theory suggests. Matusik and Hill (1998) point out that

those outside organizational boundaries can provide new knowledge that would not have been intern-

ally developed and this new knowledge carries great value to the organization, particularly in dynamic

and competitive industries. Lautsch (1999—unpublished manuscript) also shows that firms are willing

to pay wage premiums for numerical and skill flexibility.

Hence, both theory and recent evidence suggest that workers in the temporary help industry are

likely to vary in their preference for this type of employment and in their opportunities for skill

accumulation and wage growth. Consequently, while government data indicate that over 60 per cent

of temporary workers wanted a permanent work arrangement, there is also a sizeable minority who

actually prefer their temporary status (Cohany, 1996). Evidence also suggests that although blue

and pink-collar jobs comprise the majority of temporary work, there are also a growing number of

professional, technical and managerial positions (Cohany, 1996; Segal & Sullivan, 1997a). We believe

that the latter group is part of a growing group of workers who have adopted an alternative, ‘boundary-

less’ career path, one that is opposite to bureaucratic career, one of independence from, rather than

dependence on, traditional organizational career arrangements (Arthur & Rousseau, 1996). In contrast

to careers that involve a sequence of jobs within one internal labor market, boundaryless careers

reflect career paths that go beyond the boundaries of single employment settings (DeFillippi & Arthur,

1996).

As job security and promotional opportunities within larger organizations decline, individuals may

view multiple employer experiences in a positive light because it supports skill development, increases

marketability, shifts career control to the employee, and perhaps results in better matching career and

family life-cycle demands. As such, boundarylessness represents a different conception of job security.

The key benefit of working within an organization is the sense of social and economic security it

affords (Kunda, Barley, & Evans, 2002). When individuals become boundaryless their security is

rooted in their own skills and ability to sell those skills in the external labor market. Their worth is

validated outside the organization. For example, in their study of technology contractors, Kunda,

Barley, and Evans (2002) noted that security came with experience in selling skills and that security

was not about keeping a job but about how easily one could find another job. Skill development is

clearly an important component of the boundaryless career.

Boundaryless careers take many forms, and may involve the ‘transitions across organizational

boundaries’ that are characteristic of temporary employment (Sullivan, 1999, p. 464). Hence, many

workers who may prefer a boundaryless career may be drawn to the temporary help industry. Such

workers value general skill accumulation and have invested more than other temporaries in their

human capital. They prefer temporary work arrangements with multiple firms because of the oppor-

tunities for skill development.

430 J. H. MARLER ET AL.

We believe that those who prefer boundaryless careers are an emerging group different from our tra-

ditional notion of temporary workers because they perceive their careers differently. Aldrich & Baker

(1996) suggest that boundaryless careerists are comfortable with multiple employer assignments

because they provide opportunities to increase skills and knowledge. They are less likely to look for

a permanent job because they increase their value and marketability by accumulating knowledge across

employers. They also identify with their occupation or skill set and not with the organization that they

work for. We expect, therefore, that boundaryless temporaries will perceive their work differently, as a

way to improve their marketability and reduce their risk of dependence on any single organization.

A bifurcated temporary labor market

The above discussion of demand and supply factors influencing the labor market suggests that the tem-

porary workforce will vary along two key dimensions: skill accumulation and employment prefer-

ences. Further, both theory and evidence suggests that these dimensions are not orthogonal, they

are positively related. Higher skill will coincide with preference for temporary work similar to a boun-

daryless career. Low skilled individuals are more likely to seek a standard job with a single employer.

Thus, there should be, and we propose, two primary types of temporary worker. We expect one type to

be the ‘traditional’ temporary worker: low skill and a preference for traditional, long-term employment

that is consistent with the predictions based on human capital theory. The other type we refer to as a

‘boundaryless’ temporary worker: high skill and a preference for non-traditional employment arrange-

ments, that is consistent with the notion of boundaryless careers and changing organizational strategies

based on economic transaction theory. Therefore, we propose:

Hypothesis 1a: Temporary employee skill level and preference for temporary employment are not

independent.

Hypothesis 1b: Temporary employees will form two distinct groups that will differ significantly in

skill and preference for temporary employment arrangements. One group will comprise individuals

with low skill and low preference for temporary employment and the other group will consist pri-

marily of higher skilled individuals with a high preference for temporary employment arrangements.

In addition to this combination of skill and preference, we expect that these two types will have

several other related characteristics derivative of the same demand and supply factors.

As discussed above, the growth of the temporary staffing industry has resulted in more favorable

conditions for skilled workers. Temporary help or staffing agencies lower the risks and costs of job

search and placement, increasing the opportunities for multiple contracts. Higher wages may attract

higher skilled individuals, but to induce workers to switch to contract work, they must perceive oppor-

tunities for multiple assignments. Thus we further expect that:

Hypothesis 2: There will be a significant difference between how the two types perceive alternative

job opportunities and this difference will be such that boundaryless types perceive that they have

more job alternatives than traditional types.

Human capital theory predicts that in general individuals will prefer a traditional job to a temporary

assignment because the wages are likely to be higher for specific, organizationally valued skills.

Historically, workers in temporary jobs have earned lower wages than workers with equivalent skills

in longer-term jobs (Segal, 1996). The opportunities and wage rates of higher skilled temporary work-

ers, however, should increase as more organizations choose to externalize highly, skilled work. Wages

for particular, unique general skills are likely to be higher to attract higher skilled workers to accept

BOUNDARYLESS AND TRADITIONAL CONTINGENT EMPLOYEES 431

shorter, riskier employment contracts. Thus, it is likely that boundaryless types earn higher relative

wages than the traditional group. Further, as job security within single firms appears less assured, indi-

viduals may become more focused on the transactional aspects of their relationship (Robinson Kraatz

& Rousseau, 1994; Rousseau, 1995), and therefore expect to earn higher wages. We hypothesize,

therefore, that:

Hypothesis 3: There will be a significant difference between the two types with respect to relative

wages, whereby boundaryless types earn higher relative wages for their temporary work (compared

to standard work) than do traditional types.

Hypothesis 4: There will be a significant difference between the two types in their expectations for

higher wages such that boundaryless types expect to earn higher wages; whereas, traditional types

do not.

We have argued that the growth of staffing agencies reduces the risk and costs for both temporary

workers and their employers, hence increasing the supply of high skilled temporary workers who actu-

ally prefer temporary work. It is also possible that changes in the labor force have contributed to chan-

ging career choices over the life course (Baker & Aldrich, 1996; Elder & Pavalko, 1993).

Demographic shifts in the labor supply, most notably greater participation of married women in the

labor force and an associated increase in dual income families, may lead to increased opportunities to

experiment with alternative work arrangements. Because of their higher skills, boundaryless workers

are likely to have employment opportunities in both traditional and temporary work arrangements.

That they choose temporary work may, in some cases, reflect a lower need for steady income and

employer-provided benefits. We therefore expect that boundaryless types are more likely to have work-

ing spouses. Hence,

Hypothesis 5: There will be a significant difference between the two types in the working status of their

spouse such that boundaryless types are more likely to have working spouses than traditional types.

In summary, we have developed a boundaryless profile and distinguished this particular profile

against the more ‘traditional’ category of temporary worker. We use this set of hypotheses to empiri-

cally test whether theory and evidence coincide. We suggest that high skilled temporary workers who

prefer temporary work represent an example of the growing acceptance of boundaryless careers, which

coincides not only with changing organizations’ boundaries and employment security policies but also

with changing demographic characteristics of workers. And from a demand perspective, an increased

demand for high skilled temporary work, coupled with the growth of staffing intermediaries, has low-

ered the cost for both organizations and higher skilled individuals who might otherwise spurn contract

work. Organizations with strategies that emphasize employability rather than commitment and longer-

term job security also encourage a form of boundaryless career.

Differences within temporary employment arrangements: does it matter?

Several studies have found differences in satisfaction with temporary work with respect to variation in

voluntariness or preference. Ellingson, Gruys, and Sackett (1998) found a positive correlation between

voluntariness and overall satisfaction with working as a temporary employee. Feldman, Doerpinghaus,

and Turnley (1995) also found significant differences between self-reported voluntary and non-

voluntary temporaries in satisfaction with temporary life. These results are to be expected because

one would expect voluntariness and satisfaction with temporary work to be highly correlated. Both

studies, however, report conflicting results when measuring facets of job satisfaction. Ellingson and

432 J. H. MARLER ET AL.

colleagues found no significant relationship between voluntariness and specific facets of job satisfac-

tion. In contrast, Feldman et al.’s study showed voluntary temporaries reported higher job facet satis-

faction, in particular, general satisfaction and pay satisfaction.

It is important to note, however, that in both these studies, samples on which these results rested

controlled for occupational and assignment variation by using homogenous samples. Ellingson and

colleagues used a sample of clerical temporaries located in one city. Similarly, Feldman and collea-

gues’ sample included clerical and blue-collar temporaries.

We explore whether these same results will occur in a broader, more varied sample of temporary

workers. We expect that boundaryless temporaries are likely to experience greater job satisfaction

because, in addition to the volition, the nature of the work should be more challenging thus enhancing

perceived intrinsic value of the work.

We therefore propose:

Hypothesis 6: Boundaryless temporaries will be more satisfied with their jobs than traditional

temporaries.

Consistent with our arguments concerning the differences across the two temporary types regarding

differences in wages, we expect boundaryless temporaries to report higher levels of pay satisfaction

than traditional temporaries.

Hypothesis 7: Boundaryless temporaries will be more satisfied with their pay than traditional

temporaries.

Gallagher and McLean Parks (2000) argue that the notion of organizational commitment is rendered

more complex when considered in the context of alternative employment relationships. They theorize

that not only are there different commitment foci such as organization, job, employment, and occupa-

tion but, in the context of temporary agency employment arrangements, one must also consider

commitment to the employing temporary help agency as well as to the client organization. Further-

more, commitment to the client is likely to represent a cumulative experience of assignments with

multiple clients. Given this perspective and given the antecedents and correlates of organizational

commitment are primarily under the control of the client firm rather than the employer temporary help

agency (Gallagher & McLean Parks, 2000), we would expect that boundaryless temporaries would

exhibit greater client commitment than traditional temporaries. Boundaryless temporaries are more

likely to encounter client assignments that have environmental antecedents such as enriched jobs that

are correlated with greater organizational commitment (Mathieu & Zajac, 1990). Further we have

already hypothesized that boundaryless temporaries will have higher job satisfaction, and job satisfac-

tion is positively correlated with organizational commitment (Mathieu & Zajac, 1990). Finally,

Gallagher and McLean Parks (2000) suggest that there may be an overlap of commitment foci across

organization, job and occupation, which may be particularly true for ‘professionalized’ occupations

such as those likely to represented by boundaryless temporaries. Thus we would hypothesize:

Hypotheses 8: Boundaryless temporaries will have higher client commitment than traditional

temporaries.

There is less theory or consistent evidence regarding performance differences within temporary

employment. Pearce (1993) found higher extra-role performance reported by independent contractor

engineers compared to traditional core employees, whereas Van Dyne and Ang (1998) found loan

officers and nurses had lower citizenship behavior than their core employee counterparts. There is very

little evidence, however, regarding performance differences within different temporary types.

Ellingson et al. (1998) found that overall assignment satisfaction and satisfaction with temporary work

BOUNDARYLESS AND TRADITIONAL CONTINGENT EMPLOYEES 433

were positively and significantly related to performance but found no significant difference in perfor-

mance between those who preferred and did not prefer temporary work. Lautsch (1999—unpublished

manuscript) on the other hand, found evidence of better performance among traditional temporaries

than their full-time counterparts, which she attributed to differences in how organizations use tempor-

aries. Some organizations use temporary work as a method of screening job applicants (Lautsch,

1999—unpublished manuscript). In this case, traditional temporaries have an incentive to work hard.

In both cases, the sample of temporaries performed clerical or low skilled assignments. In contrast,

boundaryless temporaries are more likely to be in enriched jobs, which are theoretically more satisfy-

ing and motivating (Hackman & Oldham, 1976). Further, Kunda et al. (2002) found that high techno-

logy temporaries were aware that better performance is likely to lead to referrals and repeat business.

Since theory and evidence seem to offer competing predictions about the relative performance of the

two types of temporary worker, we propose no hypotheses a priori with respect to performance and

conduct an exploratory study to determine empirically whether differences exist across the two diverse

type of temporary worker.

Organizational Context

The country In 1995, the Bureau of Labor Statistics began counting the number of contingent work and alter-

native employment arrangements in the U.S. workforce. According to their estimates, approxi-

mately 5 per cent of the U.S. workforce were in contingent work, defined as any job in which an

individual does not have an implicit or explicit contract for long-term employment. In 1997, about

12.6 million, or 1 in 10 workers were estimated to be in alternative employment arrangements.

These arrangements included those classified as independent contractors, on-call workers, tempor-

ary agency help workers and workers provided by contract firms. About two-thirds of alternative

employment arrangements were independent contractors, and ten per cent were temporary help

agency workers. Temporary help agency workers were also more likely than any other category

to be also classified as contingent and continue to be the fastest growing segment of the workforce.

They comprise about 1.5 to 5 per cent of the total workforce.

The participants in the national sample The Bureau of Labor Statistics scientifically selected participants in the national sample from

households in the nation as a whole. Participants represented anyone 16 years or older who had

a job or was looking for a job and were available for work during the month of February 1995

and 1997 and who further identified themselves as being paid by a temporary help agency. Fifty-

seven per cent were women and 38 per cent were married. The average temporary agency worker

was 36 years old and earned on average $371 per week.

The participants in the regional sample Participants in the regional sample consisted of individuals working on assignments in 1995 and

sampled from four temporary help agencies and one employment broker located in the Northeast

United States. Participants were working in assignments in 95 client organizations representing

local, regional and international companies. Fifty-four per cent were women and 54 per cent were

married. The average age was 38 years old and average hourly earnings were $17 per hour.

434 J. H. MARLER ET AL.

Methods

To test our propositions, we use two samples: a national probability sample that has external validity

and a regional sample that contains extensive measures of temporary worker attitudes and behaviors.

We apply a hierarchical clustering technique to the national sample and replicate these results on the

regional sample. We then use the regional sample’s more extensive measures of attitudes and perfor-

mance to conduct further analyses.

National sample

The national sample is derived from the Current Population Survey’s (CPS) Contingent Work

Supplement, collected in February 1995 and 1997. The CPS is a monthly survey series collected by

the Bureau of the Census for the Bureau of Labor Statistics (BLS) and is the source for official govern-

ment statistics on employment. Households are scientifically selected to represent the nation as a

whole and household participants (U. S. Dept. of Commerce, 1995). The 1995 and 1997 Contingent

Work Supplement consists of a set of additional questions about contingent work and alternative work

arrangements that were asked of household participants who had a job or were looking for a job and

were available for work during the month of February. The sample used in this study consisted of 614

temporary workers with no missing data1 who were employed at the time of the survey and included in

the supplemental questions.

National sample cluster variables

Ten variables were included in the cluster analysis. These variables are described below and are

included to test whether they cluster in the theoretically derived pattern across the two proposed types.

Skill accumulation is measured in three ways: by level of formal education, age and occupation. Level

of education is a measure of cognitive ability and general skill accumulation. Age captures the degree

of accumulated work experience, both specific and general, and occupation captures the specific skill

accumulation.

Education is a categorical scale that indicates 16 levels of educational attainment. It is ordinal with the

lowest attainment level being less than first grade and the highest, a doctorate degree. The actual mea-

sure converts this scale into number of years of education. Less than 9th grade educational attainment

is treated as an average of eight years of education. A high school diploma represents 12 years; a col-

lege degree represents 16 years and a doctorate 21 years.

Age is measured as number of years from birth and is an accepted measure of experience in the study of

human capital.

Occupation was measured by 13 indicator variables. Individual occupation is coded according to the

single digit Census Bureau major occupation coding. For example occupation code 1 equals executive

and managerial occupations, occupation code 2 equals professional occupations. Codes increase with

1An analysis of those cases with missing data show that there were no significant differences in preference for contingent work, marital status, race, location or reasons for contingent work. Temporary employees with missing data were more likely to be in occupations with a higher rate of unemployment.

BOUNDARYLESS AND TRADITIONAL CONTINGENT EMPLOYEES 435

decreasing level of skill required for the occupation. For example occupation 12 are helpers and

laborers.

Preference for temporary employment arrangement is measured by two variables.

Preference for temporary work. Respondents make judgments about their work arrangement: do they

prefer or not. Those who worked for temporary help agencies were asked: ‘Earlier you said you were

paid by a temporary help agency. Would you prefer a job with a different type of employer?’ Response

choices were bi-polar qualitative response categories: yes, depends, or no.

Reasons for temporary work represented 16 nominal categorical variables indicating which of the 16

choices reflected the main reason for choosing temporary work. These 16 reasons grouped into three

categories: job, economic, and personal factors. Reasons that were job related include: to be own boss,

money is better, training, and flexible schedule.

Reasons representing economic factors were: this was the only type of work I could find, hope leads

to a permanent job, laid-off and rehired as contingent worker, and other economic. Personal factors

included child-care, other family/personal obligations, in school, health limitations, and social security

limitations.

Other characteristics consider job alternatives and marital status.

Number of job alternatives was measured by the level of occupational unemployment, which repre-

sents a ratio of the number of unemployed and looking for employment divided by the number of

employed, categorized by a two-digit occupation code. The data represent current unemployment as

of the time of the survey, February 1995 or 1997.

Wage rates are reported weekly wage.

Marital status is an indicator variable recorded as one if the individual was married with a spouse pre-

sent, otherwise zero.

Married woman is measured as an interaction of two indicator variables, female and marital status.

Regional sample

The Regional Sample consists of survey data collected from individuals working on assignments in the

Northeast United States. Four temporary help agencies and one employment broker2 agreed to distri-

bute questionnaires (by mail or with paychecks) to all individuals currently on their active list. Com-

pleted surveys were returned by mail to the researchers. A total of 960 surveys were successfully

delivered and 276 were returned for a response rate of 29 per cent.3

Respondents represent a broad cross-section of temporary workers and work assignments. They

were working in assignments at 95 client organizations representing local, regional, national and inter-

national companies.

2An employment broker is an agency that makes arrangements with temporary help agencies to provide contract workers to client organizations. 3This response rate is about average for sample surveys of this population. Miller (1995) sampled one large temporary agency located in the West and had a 17 per cent response rate. Von Hippel et al. (1997) had a 54 per cent response rate with a survey fielded by a large Midwestern agency.

436 J. H. MARLER ET AL.

Regional sample cluster variables

We match the measures used in the national sample to those collected in the regional sample.

Education is measured as a 6-item ordered categorical variable. The lowest level of education was high

school or equivalency diploma and the highest level was Ph.D. The mean level of education was an

associate’s degree.

Age is measured as number of years from birth.

Occupation was measured by seven indicator variables. Individuals reported their occupation from a

choice of clerical, technician, engineering, accounting, software/systems, industrial, and managerial/

executives. These correspond to the occupational designations used by the temporary help industry and

when aggregated to Census Bureau coding that categorizes occupations as managerial, professional,

clerical and industrial.

Preference for temporary work. A two-item measure was created to estimate each individual’s prefer-

ence for temporary work. Response choices were ordered categorically: strongly disagree, disagree,

neutral, agree, and strongly agree. The two items were: ‘I have a choice and I prefer temporary/con-

tract work’ and ‘I have little choice; I would prefer a permanent, regular job’ (reverse coded). The

alpha coefficient for this scale is 0.84.

Reasons for temporary work. Paralleling the measures in the national sample, we identified job, eco-

nomic and personal reasons. For job factors, individuals were asked to rate the importance of a parti-

cular factor to making temporary work attractive. The response choices ranged from ‘not important at

all’ (1) to ‘extremely important’ (5) and the factors evaluated included variety, challenge, flexibility

and independence/autonomy. The alpha coefficient for this scale is 0.70. For economic reasons we

used a one-item scale that asked about the likelihood of taking a permanent job if one were available

at another company over the next six months. Response choice was a 7-item Likert-type scale where

1¼ very unlikely and 7¼ very likely. Respondents also indicated whether higher pay was a reason for accepting temporary employment. Personal reasons included a one-item measure in which respon-

dents indicated how important balancing work and family is to making temporary work attractive.

Other variables Number of job alternatives. Respondents were asked to estimate their present alternative job opportu-

nities on a 5-point scale ranging from no opportunities (1) to many opportunities (5).

Wage rates represented respondents’ reported hourly wage rate.

Marital status is an indicator variable recorded as one if the individual was married with a spouse pre-

sent, otherwise zero.

Married woman. This is measured as an interaction of two indicator variables: gender (female) and

marital status.

Other measures Relative wages is measured as the difference between what the temporary worker is currently earning

and the wage of an equivalent conventional work arrangement. The following equation depicts how

this measure was derived based on wage equations used by Krueger & Summers (1988)

Wpi ¼ Xi� þ Yi� þ Zi� þ “i ð1Þ

BOUNDARYLESS AND TRADITIONAL CONTINGENT EMPLOYEES 437

where Wpi is the predicted log hourly wage for employed individual i; Xi a vector of human capital

variables for individual i that include education, age, and its square; Yi is a vector of geographic vari-

ables including region of the country and dummies indicating central city, mixed or rural location; Zi is

a vector of individual characteristics such as gender, race, marital status, and number of children; “i is an error term, and � are parameters to be estimated based on a national sample population of employed men and women,

Ri ¼ Ai � Wpi ð2Þ

where Ri is the difference between the predicted outcome, Wpi and the individual’s actual log hourly

wage, Ai. The difference, Ri, represents a measure of relative outcome.

Work satisfaction is a 9-item scale. Individuals indicate their level of satisfaction, (1¼ very dissatis- fied, 5¼ very satisfied) with attributes of their work assignment. These attributes include, for example, ‘amount of freedom on the job’, ‘opportunity to accomplish something worthwhile’, and ‘friendliness

of the people you work with.’ The mean of the summation has an alpha of 0.90.

Pay satisfaction was measured using the pay level satisfaction dimension of the Heneman and Schwab

(1985) Pay Satisfaction Questionnaire (PSQ). The 8-item pay level satisfaction measure was highly

reliable, with an alpha level of 0.92.

Client commitment measurement was adapted from the Mowday, Porter and Steers (1982) 9-item

measure of organizational commitment. The measure was revised slightly to make explicit that the

questions referred to the client organization rather than the hiring agency. The alpha level for the mea-

sure of organizational commitment to the client was 0.92.

Task and contextual performance measures come from a 12-item self-report scale that factored into two

dimensions shown on Table 2. The first factor relates to performance efficiency and consists of four

items about quantity and quality of work output, accuracy of work, and customer service provided.

The second factor describes reported work group cooperation and company level citizenship and con-

sists of the remaining eight items. Items include questions about working as part of a team or group,

seeking information from others in the group, responding to the needs of others in the group, and doing

things that help others when it’s not part of the job. The alpha on each is 0.65 and 0.88 respectively.

Analyses

Cluster analysis is a descriptive analysis that involves measuring the distance between clusters, in this

case profiles of temporary employees, and comparing within cluster and across cluster distance. Using

this methodology, we group each individual into clusters based on an individual profile formed by the

ten cluster variables described above. We chose these measures specifically to test hypotheses 1–5 in

which we derive from theory a specified pattern of characteristics. The cluster methodology derives

groupings based on a set of variables; however, each variable individually does not necessarily differ

significantly across clusters. Once clusters are formed, we used Tukey’s studentized range t-tests and

chi-square statistics to investigate individual variable differences across clusters.

An advantage of using a clustering methodology is that it uses an objective data driven process to

determine similarities between temporary profiles through an agglomerative process that builds up

clusters by putting similar case profiles, in this case individual temporary employees, in the same clus-

ter. Clusters emerge and distinguish themselves based on comparisons of similarities of within-cluster

multivariate profile differences to across-cluster multivariate profile differences (Griffeth, Hom,

438 J. H. MARLER ET AL.

DeNisi, & Kirchner, 1985). In our hierarchical clustering method we used Ward’s minimum variance

method, which joins clusters using the criteria of minimum within cluster variance. We used Gower’s

resemblance coefficient, which is a weighted combination of the Jaccard coefficient for nominal vari-

ables and the Euclid distance measure. The number of clusters was chosen on the basis of parsimony

and maximum distance between levels of the solution.

We used the same clustering process for both the national probability sample and the regional sample.

We then performed multivariate regression analyses and MANOVA on the regional sample to esti-

mate the effect of temporary worker type on measures of work attitudes and behaviors proposed in

hypotheses 6–8 and to explore differences in task performance.

Results

A correlation analysis of the national sample variables appears in Table 1 and, for the regional sample,

in Table 2. These results provide mixed but generally supportive evidence in favor of hypothesis 1a,

which proposed that skill accumulation and preference for temporary employment are not indepen-

dent. Preference is significantly correlated with age, sales, other service occupations, expectation of

higher wages and flexibility. Expectation of higher wages is significantly correlated with education and

professional occupations. And flexibility is significantly correlated with age, clerical occupations, and

preference. In general, it appears that the various measures of skill (i.e., education, age and occupation)

are not independent of measures of preference (i.e. preference, expectation of higher wages, flexibility

and desire of finding a standard job). Similar patterns of relationships emerge in Table 2, which shows

a correlation of these measures using the regional sample.

The results of the cluster analyses, ANOVA and chi-square tests appear in Tables 3 and 4. Table 3

contains the results using the national sample and Table 4 the results for the regional sample. Each

table presents the cluster profile, depicting the cluster variables, their means, and standard deviation.

Also reported are the results of the ANOVA using Tukey’s studentized range test and chi-square in the

case of categorical variables.

In both samples, a two-cluster solution emerged as the most stable pattern, initially supporting

hypothesis 1b although a more detailed analysis of the combination of variables in each cluster is needed

to eliminate the rejection of the hypothesized profile of both types. In both samples, type 1 consisted of

significantly higher proportions of blue-collar and pink-collar occupations and type 2, a significantly

higher (all) proportion of white-collar occupations. Further in both samples, type 1 had significantly

lower levels of preference for temporary work than type 2 and provides further support for hypothesis

1b which predicted both preference and skill accumulation were not orthogonal but linked.

As shown in Table 3, the two clusters depict two distinct temporary types. Type 1 consists of tem-

poraries who generally do not prefer temporary work. Only 22 per cent preferred their work arrange-

ment. Type 1 also had a significantly lower proportion of individuals who valued temporary work for

its flexibility compared to type 2 (10 versus 20 per cent). This group on average had a high school

education and was in exclusively pink-collar and blue-collar occupations. In contrast, type 2 consisted

of a greater proportion of individuals who preferred temporary work and its work-related attributes.

Over 35 per cent preferred their work arrangement compared to 22 per cent in type 1. Type 2 also had a

greater level of education and were primarily in managerial, professional and technical occupations.

Type 2 corresponds to a boundaryless temporary and type 1 represents the traditional temporary.

Replicating the procedure with the regional sample, we again found a stable two-cluster solution.

The two clusters found in this sample parallel the results obtained with the national sample. Type 1

BOUNDARYLESS AND TRADITIONAL CONTINGENT EMPLOYEES 439

T ab

le 1

. N

at io

n al

sa m

p le

co rr

el at

io n

1 2

3 4

5 6

7 8

9 1

0 1

1 1

2 1

3 1

4 1

5 1

6 1

7 1

8 1

9 2 0

2 1

2 2

2 3

Order from us and get better grades. We are the service you have been looking for.

Order essay

Type of paper needed:

Academic level:

Deadline:

Pages:

1650 words

Total price: $0.00