xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001914857
007 cr mnu|||uuuuu
008 071029s2007 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001987
Klein, Rebecca H.
The "dark side" of OCB :
b examining the relationship between citizenship behavior and work-to-family conflict
h [electronic resource] /
by Rebecca H. Klein.
[Tampa, Fla.] :
University of South Florida,
ABSTRACT: Research on organizational citizenship behaviors (OCB) has focused on the positive aspects of the construct, neglecting the possibility that individuals who engage in OCB may suffer negative consequences. Thus, the present study expands the literature by examining the possibility that OCB is also related to negative individual-level factors, such as work-to-family conflict (WIF). In a replication and extension of Bolino and Turnley's (2005) research, the present study offers an in-depth analysis of the relationship between OCB and WIF, examining two potential mediators (work time and role overload) and two potential moderators (gender and perceptions of OCB as discretionary). Two hundred and ninety-six participants, recruited from the alumni database of a large southeastern university, Craig's List, and a snowball approach, completed surveys. Additionally, supervisor-ratings of OCB were obtained for a sub-sample of 35 participants. Study hypotheses were tested using zero-order correlations and multiple regression analyses. No support was found for a relationship between OCB and WIF, nor was there support for the moderating role of gender and perceptions of OCB as discretionary. Theoretical and practical implications, as well as future directions, are discussed.
Thesis (M.A.)--University of South Florida, 2007.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
Title from PDF of title page.
Document formatted into pages; contains 80 pages.
Advisor: Tammy D. Allen, Ph.D.
Organizational citizenship behavior.
t USF Electronic Theses and Dissertations.
The Â“Dark SideÂ” of OCB: Examin ing the Relationship between Citizenship Behavior and Work-to-Family Conflict by Rebecca H. Klein A thesis submitted in partial fulfillment of the requirements for the degree of Master of Arts Department of Psychology College of Arts and Sciences University of South Florida Major Professor: Tammy D. Allen, Ph.D. Walter C. Borman, Ph.D. Vicky Phares, Ph.D. Date of Approval: March 27, 2007 Keywords: organizational citizenship behavior work-family conflict, role overload, work time, gender, discretionary Copyright 2007, Rebecca H. Klein
To all my friends and loved ones, without whom I could never have accomplished this feat. In particular, I dedi cate this thesis to my parent s, who taught me that I could accomplish anything I set my mind on, and w hose never-ending love and support has made me the person I am today. And also to my fianc, for all the words of encouragement, patience, and technical s upport that he has provided along the way.
Acknowledgements I would like to acknowledge the members of my thesis committee, for their time, assistance, and guidance. My advisor, Dr. Tammy Allen, deserves a special acknowledgement, for graciously accepting me as her student, for providing continual encouragement and support, a nd for being a true mentor.
i Table of Contents List of Tables iii List of Figures v Abstract vi Chapter One Â– Introduction 1 Organizational Citizenship Behavior (OCB): Defining the Construct 3 Organizational Citizenship Behavior (OCB): Positive Biases 6 Work-Family Conflict (WFC): An Overview 8 Work Time as a Mediator to the OCBÂ—WIF Relationship 10 Role Overload as a Mediator to the OCBÂ—WIF Relationship 11 OCBO, OCBI, and WIF 13 The Moderating Effects of Gender 14 The Moderating Effects of Perceptio ns of OCB as Discretionary 16 Hypotheses 18 Chapter Two Â– Method 19 Participants 19 Measures 21 Procedure 24 Chapter Three Â– Results 26 Preliminary Analyses 26 Hypothesis Testing 27 Exploratory Analyses 45 Chapter Four Â– Discussion 56 Theoretical Implications 57 Practical Implications 59 Limitations 60 Future Directions 62 Conclusion 64 References 65 Appendices 74 Appendix A: Hypothesized Relationships 75
ii Appendix B: Organizational Citizen ship Behaviors Scale Items 76 Appendix C: Work-to-Family Conflict Scale Items 78 Appendix D: Role Overload Scale Items 79 Appendix E: Individual Initiative Scale Items 80
iii List of Tables Table 1 Descriptive Statistics 28 Table 2 Intercorrelations among Study Variables 29 Table 3 Regression of OCB (Self-Reported) on WIF 32 Table 4 Regression of OCBI (Self-Reported) on WIF 32 Table 5 Regression of OCBO (Self-Reported) on WIF 33 Table 6 Regression of OCB (S upervisor-Reported) on WIF 33 Table 7 Regression of OCBI (S upervisor-Reported) on WIF 34 Table 8 Regression of OCBO (S upervisor-Reported) on WIF 34 Table 9 Regression of OCB (Sel f-Reported) on Work Time 36 Table 10 Regression of OCBI (S elf-Reported) on Work Time 36 Table 11 Regression of OCBO (S elf-Reported) on Work Time 37 Table 12 Regression of OCB (Super visor-Reported) on Work Time 37 Table 13 Regression of OCBI (Superv isor-Reported) on Work Time 38 Table 14 Regression of OCBO (Superv isor-Reported) on Work Time 38 Table 15 Regression of (Self-Repor ted) OCB on Role Overload 40 Table 16 Regression of (Self-Repor ted) OCBI on Role Overload 40 Table 17 Regression of OCBO (Sel f-Reported) on Role Overload 41 Table 18 Regression of OCB (Supervi sor-Reported) on Role Overload 41 Table 19 Regression of OCBI (Supervis or-Reported) on Role Overload 42
iv Table 20 Regression of OCBO (Supervis or-Reported) on Role Overload 42 Table 21 Moderated Regression of Gender on Work Time and WIF 43 Table 22 Moderated Regression of Gender on OCB (Self-Reported) and Role Overload 46 Table 23 Moderated Regression of Ge nder on OCBI (Self-Reported) and Role Overload 46 Table 24 Moderated Regression of Ge nder on OCBO (Self-Reported) and Role Overload 47 Table 25 Moderated Regression of Gender on OCB (SupervisorReported) and Role Overload 47 Table 26 Moderated Regression of Gender on OCBI (SupervisorReported) and Role Overload 48 Table 27 Moderated Regression of Gender on OCBO (SupervisorReported) and Role Overload 48 Table 28 Moderated Regression of Per ceptions of OCB as Discretionary on OCB (Self-Reported) and Role Overload 49 Table 29 Moderated Regression of Per ceptions of OCB as Discretionary on OCBI (Self-Reported) and Role Overload 50 Table 30 Moderated Regression of Per ceptions of OCB as Discretionary on OCBO (Self-Reported) and Role Overload 51 Table 31 Moderated Regression of Per ceptions of OCB as Discretionary on OCB (Supervisor-Reported) and Role Overload 52 Table 32 Moderated Regression of Per ceptions of OCB as Discretionary on OCBI (Supervisor-Reported) and Role Overload 53 Table 33 Moderated Regression of Per ceptions of OCB as Discretionary on OCBO (Supervisor-Reported) and Role Overload 54 Table 34 Regression of Individual In itiative (Self-Reported) on WIF 54 Table 35 Regression of Individual In itiative (Supervisor-Reported) on WIF 55
v List of Figures Figure 1. Hypothesized relationships between OCB and WIF, including mediating and moderating variables. 76
vi The Â“Dark SideÂ” of OCB: Examin ing the Relationship between Citizenship Behavior and Work-to-Family Conflict Rebecca H. Klein ABSTRACT Research on organizational citizensh ip behaviors (OCB) has focused on the positive aspects of the construct, neglecting the possibility that individuals who engage in OCB may suffer negative consequences. Thus, the present study expands the literature by examining the possibility that OCB is also re lated to negative indi vidual-level factors, such as work-to-family conflict (WIF). In a replication and ex tension of Bolino and TurnleyÂ’s (2005) research, th e present study offers an in -depth analysis of the relationship between OCB and WIF, examin ing two potential mediators (work time and role overload) and two potential moderato rs (gender and perc eptions of OCB as discretionary). Two hundred and ninety-six participants, recruited from the alumni database of a large southeastern univers ity, CraigÂ’s List, and a snowball approach, completed surveys. Additionally, supervisor-ratings of OCB were obtained for a subsample of 35 participants. Study hypotheses we re tested using zero-order correlations and multiple regression analyses. No support wa s found for a relationship between OCB and WIF, nor was there support for the moderating role of gender and pe rceptions of OCB as discretionary. Theoretical and practical implic ations, as well as future directions, are discussed.
1 Chapter One Introduction Organizational citizenship behavior (OCB) and related constructs have received an increasing amount of resear ch attention over the past de cade (Podsakoff, MacKenzie, Paine, & Bachrach, 2000). As a result, mu ch progress has been made in addressing conceptual ambiguities with the construct, as well as uncovering a variety of antecedents and consequences to citizenship behavi or (see Organ & Ryan, 1995; Podsakoff & MacKenzie, 1997; and Podsakoff et al., 2000 fo r reviews). Despite the fieldÂ’s growing knowledge base, the vast majority of research on citizenship behavior still focuses on the positive aspects of such action, clinging to the prevailing assumptions that OCB stems from non-self-serving motives and results in beneficial organizatio naland individuallevel outcomes (Bolino, Turnley, & Niehoff, 2004). However, several researchers have recently suggested that there may also be a Â“dark sideÂ” to ci tizenship behavior, advocating the inclusion of a broader range of criterion variables, such as overload, role stress, and work-family conflict (Bolino et al., 2004; Organ & Ryan, 1995; Podsakoff, MacKenzie, & Hui, 1993). Although little empirical research has tested these propositions, Bolino and Turnley (2005) recently found evidence that OCB is, in fact, associated with negative consequences for employees, including ro le overload, job stress, and work-family conflict (WFC). Moreover, gender was found to moderate the relationship between
2 citizenship behavior and WFC, with female s exhibiting a stronger relationship between the two variables. These findings represent an important step toward understanding how contextual performance relate s to individual-level outcomes; however, more research is clearly needed. Accordingly, the present study aimed to replicate and extend Bolino and TurnleyÂ’s (2005) research, focusing specifica lly on the relationship between OCB and work-interfering-with-family conflict (WIF). Bolino and Turnley (2005) focused on one type of OCB, individual initia tive, and, as they point out in their discussion section, their measure of individual initiative (developed for that particular study) may be more related to work time and WIF than previously developed OCB scales. Thus, the present study utilized a more inclusive, psychometrically es tablished measure of this construct, testing the veracity of the OCB-WIF link. Additiona lly, as is common in the OCB literature, supervisor ratings of OCB were collected, while Bolino and Turnle y (2005) utilized spousal ratings. Because spousal ratings are rare ly used in OCB research, it is unclear the degree to which spouse and supervisor rati ngs converge. However, given that other research has shown that results do vary by source of OCB (Allen, Barnard, Rush, & Russell, 2000), it is important to assess wh ether the relationship between OCB and WIF holds with supervisor-repor ted citizenship behavior. Th e current study also extended upon Bolino and TurnleyÂ’s (2005) research by examining two potential mediating variables Â– work time and role overload Â– to the relationship between OCB and WIF. Finally, two moderators were investigated, including gender, in continuation with Bolino and TurnleyÂ’s (2005) findings, and individua l differences in the degree that OCB is perceived as non-discretionary. By using a mo re fine-grained measure of organizational
3 citizenship behavior, as well as examini ng both mediating and moderating variables affecting its relationship with WIF, the pres ent study intends to shed light on a largely neglected area within the OCB field. Organizational Citizenship Behavior (OCB): Defining the Construct Organizational citizenship behavior, originally define d as Â“individual behavior that is discretionary, not dire ctly or explicitly recognized by the formal reward system, and that in the aggregate promotes the eff ective functioning of the organizationÂ” (Organ, 1988, p. 4), was developed as an extension of Katz and KahnÂ’s (1966, 1978) work. Using an open systems model, Katz and Kahn ( 1966, 1978) proposed that organizations must elicit three patterns of employee behavior in order to mainta in effectiveness: attracting and retaining membership, ensuring an acceptabl e level of role performance consistency, and evoking Â“innovative and spontaneous behavior: performance beyond role requirements for accomplishments of orga nizational functionsÂ” (Katz & Kahn, 1966, p. 337). Thus, Katz and Kahn (1966, 1978) di stinguish between dependable role performance and extra-role behavior, with the latter including such acts as cooperating with co-workers and promoting a favorable work climate. Drawing upon these ideas, Organ and colleagues coined the term Â“organiz ational citizenship be havior,Â” originally introduced as a means of broadening the perf ormance domain, thereby helping to explain the surprisingly low observed correlations between productivity and job satisfaction (Bateman & Organ, 1983; Organ, 1988). Although OCB is considered the most well-known and resear ched among related concepts (Van Dyne, Cummings, & Parks, 1995) several alternativ e constructs have been described as well, including extra-role be havior (Van Dyne et al., 1995), contextual
4 performance (Borman & Motowidlo, 1993), pros ocial organizational behavior (Brief & Motowidlo, 1986), and organizatio nal spontaneity (George & Brief, 1992). While there is much overlap among these terms, subtle differe nces exist. For example, unlike extra-role behavior and the original de finition of OCB, contextual performance, defined as behaviors that Â“do not support the technical core itself so much as they support the broader organizational, social and psychological environment in which the technical core must functionÂ” (Borman & Motowidlo, 1993, p. 73), does not exclude behaviors that are in the job description or that are formally rewarded. Moreover, organizational spontaneity only includes extra-role beha viors, though such behaviors may be rewarded (George & Brief, 1992). Finally, prosocial organizational behavior is described in terms of the employees intentions to benefit the organization, so th e actual behavior can be either functional or dysfunctional (Brief & Motowidl o, 1986). Despite these subtle conceptual distinctions, the items used to measure the various constructs are often very similar, tapping such behaviors as help ing others, staying late or working weekends, performing at levels that exceed minimum requirement s, and tolerating inconveniences on the job (Bolino et al., 2004). Variations among these definitions highlig ht two points of contention that have arisen in conceptualizing these constructs: whether such behaviors are truly discretionary and whether they are recognized by the formal reward system. In the OCB literature, Â“discretionaryÂ” refers to behavi ors that are not in formal job descriptions and that are not an enforceable part of the role. However, se veral researchers have pointed out that there is ambiguity in the boundary between ta sk performance and OCB (Graham, 1991; Morrison, 1994; Van Dyne et al., 1995) and that the distinction between in-role and
5 extra-role behaviors often varies across persons and tim es (Van Dyne et al., 1995). Additionally, although OrganÂ’s (1988) original definition of OCB stipulated that such behaviors are not directly rewarded, Podsa koff et al. (1993) provide numerous reasons why managers may take them into account, in cluding norms of reciprocity and fairness, the accessibility of OCB in memory, imp licit assumptions about what makes a good employee, and attributional processes. More over, empirical resear ch has supported that managerial performance ratings are, in fact influenced by citizenship behaviors (Allen & Rush, 1998; Conway, 1999; MacKenzie, P odsakoff, & Fetter, 1993; Van Scotter, Motowidlo, & Cross, 2000; Werner, 1994). In fact, after conducting a meta-analysis of relevant lab and field studies, Podsakoff et al. (2000) conclu ded that OCB is at least as influential as is in-role performance in pred icting managerial ratings /personnel decisions, even after controlling for common method vari ance. In response to these findings and criticisms, Organ (1997) recently modified the definition of OCB, removing the stipulations that such behavi or is discretionary and not re warded. Thus, for the purposes of this paper, OCB is defined accordingly. Another aspect of citizenship behavior th at has suffered from a lack of consensus is the dimensionality of the construct. Am ong the various factor structures used to describe OCB, the most common are a two-f actor, a three-factor, and a five factor solution. The two-factor solution differentiate s behaviors aimed at helping individuals (OCBI) from those directed at the orga nization (OCBO; Williams & Anderson, 1991). More recently, Coleman and Borman (2000) f ound support for three fact ors; the first two, interpersonal citizenship performance and organizational citizenship performance, significantly overlap with O CBI and OCBO, respectively. However, they also found
6 support for a third factor, labeled job/task conscientiousness, whic h includes behaviors that benefit the job/task. Although both of th ese solutions define citizenship behavior according to the target of the act, other researchers have used the type of behavior to differentiate among dimensions. For example, the five-factor structure is composed of conscientiousness, courtesy, civic virtue, altruism, and sportsmanship (Organ, 1988). Conscientiousness (termed individual initiativ e by Podsakoff et al., 2000) is defined as carrying out role behaviors at a level well beyond the minimum requirements, while courtesy involves behaviors aimed at prev enting work-related probl ems. Individuals who responsibly participate in the organizationÂ’s political life and c ontribute to corporate governance are displaying civic virtue, and al truism involves helping a specific person with an organizationally releva nt task. Finally, sportsmanshi p is described as tolerating less than ideal situations without complaining. Although researchers have found support fo r each of these factor structures (Coleman & Borman, 2000; Podsakoff, MacK enzie, Moorman, & Fetter, 1990; Williams & Anderson, 1991), the two-factor solution is the most parsimonious. Additionally, the relationship between OCB and other variab les has been found to differ, depending on whether the behavior target s the individual or the orga nization (Illies, Nahrgang, & Morgeson, 2007; McNeely & Meglino, 1994; Somech & Drach-Zahavy, 2004). Thus, William and AndersonÂ’s (1991) two-factor soluti on (OCBI and OCBO) is utilized in this study. Organizational Citizenship Behav ior (OCB): Positive Biases Evidently, a great deal of the OCB literat ure has been devoted toward defining the construct and addressing conceptual ambi guities with its meaning. Additionally, much
7 research has focused on examining the ante cedents and consequences of citizenship behavior. However, the vast majority of this research describes citi zenship behavior as a positive phenomenon, largely ignoring the possib ility that it could result from selfserving motives or lead to negative cons equences. For example, although researchers have examined a wide variety of ante cedents, including job satisfaction, positive affectivity, organizational support, and transfor mational leadership (see Podsakoff et al., 2000 for a review), their focus has generally be en limited to variables that carry a positive connotation, stemming from the assumption that OCB is guided by a desire to help others (Bolino et al., 2004). Alternatively, research ers have suggested that employees may engage in citizenship behaviors in order to manage impressions, gain rewards, or make others look bad (Bolino et al., 2004; Eastman, 1994; Rioux & Penner, 2001). Similarly, the consequences associated w ith OCB are also generally assumed to be positive. Consistent with Katz and Ka hnÂ’s (1966) notion that such behaviors are crucial for effective organizational functioni ng, researchers assume that OCB leads to positive organizational outcomes. Although empiri cal research is generally supportive of this notion, there are also exceptions to the pattern, with various dimensions of OCB being unrelated, or even ne gatively related, to performance (Podsakoff, Ahearne, & MacKenzie, 1994; Podsakoff & MacKenzie, 1994). In terms of individual-level outcomes, OCB has been associated with career advancement and rewards over time (Van Scotter et al., 2000), and it is generall y assumed that employees are attracted to organizations that encourage citizenship be havior (Bolino et al., 2004). However, as described by Bolino et al. (2004), there are reasons to predict ne gative individual-level consequences as well. Specifically, employees may experience diffi culty distinguishing
8 between in-role and extra-role performance (leading to role ambiguity) and feel pressured to display escalating citizenship (enhanci ng overload and work-fam ily conflict). These propositions are consistent with the finding th at individual initiativ e is positively related to role overload, job stress, and work -family conflict (Bo lino & Turnley, 2005). Bolino and TurnleyÂ’s (2005) findings are pa rticularly troubleso me, given that the OCB literature frequently focuses on ways that supervisors can encourage citizenship behavior at their organiza tions (e.g., MacKenzie et al., 1993; Organ, 1988; Podsakoff & MacKenzie, 1997). If such behavior results in negative consequences for employees, supervisors need to recognize the dangers of blindly enc ouraging OCB. Although an indepth examination of the relationship between OCB and a multitude of individual-level consequences is clearly necessary, this pape r focuses on one outcome in particular: workto-family conflict. Hypothesized relationships are described in the subsequent sections and presented in Figur e 1 (see Appendix A). Work-Family Conflict (WFC): An Overview Drawing from Kahn, Wolfe, Quinn, Snoe k, and RosenthalÂ’s (1964) work on role conflict and role ambiguity, Greenhaus and Be utell (1985) define wo rk-family conflict as Â“a form of interrole conflict in which the role pressures from the work and family domains are mutually incompatible in some respect. That is, participation in the work (family) role is made more difficult by virtue of participation in th e family (work) roleÂ” (p. 77). WFC is considered multi-faceted, consisting of time-based (the time requirements of one role impair performance on the second role), strain-based (pressures associated with one role adversely aff ect performance in the other role), and behavior-based conflicts (the behavioral requirements n ecessary for each domain are different or
9 incompatible). Additionally, researchers ha ve acknowledged the bi-d irectionality of the construct; WFC can arise in the work dom ain (work-to-family conflict or work interfering with family, WIF) or in the fam ily domain (family-to-work conflict or family interfering with work, FIW; Frone, Russe ll, & Cooper, 1992). The domain specificity hypothesis posits that situational variables associated with a gi ven domain relate to conflict originating from that domain (F rone, 2003; Frone, Russell, & Cooper, 1992); thus, because OCB is a work-related variable, the current study focused on its effects on work interfering with family. Work-family conflict, and more specifical ly WIF, has been associated with a variety of negative outcomes, ranging from attitudinal, behavioral, and health-related variables (see Frone, 2003 a nd Allen, Herst, Bruck, & Sutton, 2000 for reviews). In a meta-analysis of 67 studies, Allen et al. ( 2000) identified three br oad domains of WIF consequences: nonwork-related (e.g., life sati sfaction, family perfor mance), work-related (e.g., intentions to quit, job satisfaction, or ganizational commitment), and stress-related (e.g., general psychological st rain, somatic/physical sympto ms, depression). Given these detrimental effects, it is esse ntial to pinpoint the antecedents leading to work-to-family conflict. To that end, a multitude of antecedents have been identified and studied (see Bruck & Allen, 2005; Byron, 2005; and Frone, 2003 for review s). Among these variables are demographic factors (e.g., gender, numb er of kids), dispositional factors (e.g., negative affectivity, agreeableness), situationa l characteristics (e.g., social support, job involvement), and crossover effects (e.g., pa rtnerÂ’s level of WFC). While numerous antecedents have been found to predict WI F, Frone, Yardley, and Markel (1997)
10 advocate distinguishing proximal predictors of WFC from thos e that are more distal. In their model, work-related antecedents such as supervisor support are posited to affect WIF through three proximal predictors: work ti me, role overload, and work distress. In support of work time as a proximal antecedent, Major, Klein, and Ehrhart (2002) found that several commonly studied WIF predictors including nonjob respons ibilities, career identity salience, and organi zational rewards, exerted thei r effects through work time. Although their study provides convincing evid ence for work time as an important mediator, Cardenas, Major, and Bernas (2004) caution that simply looking at how much time you spend in a given role may Â“fail to cap ture the subjective experiences that often lead individuals to feel ove rworked and to experience conf lict between work and family domainsÂ” (p. 347). Thus, the current study focused on two of Frone et al.Â’s (1997) proximal predictors, work time and role overloa d, to explain why OCB is likely related to WIF. Work Time as a Mediator to the OCBÂ—WIF Relationship A cursory examination of the OCB construct highlights why engaging in citizenship behaviors often i nvolves devoting more time to work, either directly or indirectly by taking on extra demands. Fo r example, OCBO includes attending extra organizational functions and keeping up w ith developments in the organization. Similarly, behaviors included in the OCBI dimension are adju sting oneÂ’s work schedule to accommodate other employees, helping those th at are absent, and assisting others with their workload. Perlow and W eeks (2002) discovered that em ployees often view helping behavior as an unwanted interruption from th eir Â“real work,Â” with one employee stating: Â“The biggest frustration of my job is always having to help others and not getting my
11 own work doneÂ” (p. 353). Thus, individuals who engage in citizenship behaviors may view such acts as unwanted demands, feeling pressured to work longe r hours in order to fulfill their other work requirements. As suggested by the rational model and th e resource drain model, individuals who work longer hours are likely to face higher leve ls of WIF. The rational model of workfamily conflict predicts a linear relationshi p between the amount of time spent in the work and family domains and the degree of WFC experienced (Greenhaus, Bedeian, & Mossholder, 1987). Overall, research has b een supportive of this notion, with the relationship between work time and WI F being approximately .26 (Byron, 2005). Another reason that OCB may impact WIF is elucidated by the resource drain model, which postulates that individuals have a finite amount of resources (Rothbard, 2001; Staines, 1980). Given that time, one type of resource, is clearly in limited supply, employees are faced with the challenge of balancing their time between the work and family domains. If situational variables, such as engaging in citizenship behaviors, require that more time be spent in one dom ain, fewer hours are available for the other domain, and work-family conflict is likely to ensue. Thus, it is hypothesized that individuals who engage in ci tizenship behaviors tend to spend more time engaged in work-related activities, thereby increasing their levels of WIF. Role Overload as a Mediator to the OCBÂ—WIF Relationship Another mechanism by which citizenship be havior is likely to affect work-tofamily conflict is through role overload. Kahn et al. (1964) de scribe role overload as a type of stressor that occurs when an i ndividual experiences an overwhelming level of demands, which are perceived as exceeding hi s/her capabilities given the time available
12 and/or his/her abilities (Ka hn et al., 1964). Role overload is a function of the number of roles enacted, as well as the demands require d of each role. As previously mentioned, citizenship behaviors may be viewed by em ployees as extra demands placed upon them. Moreover, while all employees inherently en act the job-holder role researchers have proposed that individuals who engage in OCB essentially take on an additional work role: that of organizational-member (Welbour ne, Johnson, & Erez, 1998). Because each of these roles requires unique demands and expectations, employees may experience pressure and strain. Additionally, employees who have partners, ch ildren, and/or other familial obligations are faced with the addi tional challenge of simultaneously handling family role requirements. Not surprisingly, strong empirical support ex ists for a relationship between role overload at work and WIF, w ith one meta-analysis estimati ng the correlation to be .65, correcting for unreliability (Byron, 2005). Th e spillover model sheds light on why role overload is likely to influence work-family conflict. Work-family spillover involves the transfer of thoughts, affect, and behavior be tween the work and family domains; thus, experiences in one domain impact experiences in the other domain (Roehling, Moen, & Batt, 2003). In support of the spillover model, Williams and Alliger (1994) found that moods, stress, and thoughts generated in one domain influenced behavior and cognitions in the other domain; moreover, task demands were associated with distress, which enhanced spillover. Because overload is related to intense affective responses, such as anger, frustration, emotional exhaustion, a nd stress (Chen & Spect or, 1991; Thompson, Kirk, & Brown, 2005), these emotions are likely to carry over into the alternate domain. Additionally, Frone et al. (1997) describe how individuals experiencing role overload are
13 likely to be psychologically preoccupied with one role, while attempting to fulfill the requirements of the other. Because such ruminating thoughts and overpowering affective responses are distracting a nd debilitating, not to menti on resource-draining, interrole conflict is likely to result. Therefore, it is pr edicted that, by taking on an additional role as well as more demands, individuals who engage in citizenship behavior s are more likely to experience role overload, which then leads to WIF. OCBO, OCBI, and WIF It is also important to consider whethe r OCBO and OCBI differentially relate to work-interfering-with-family conflict. Although both types of behaviors are expected to require increased time and demands, thereby leading to WIF, it is possible that the strength of the relationship varies dependi ng on the target of the OCB. Specifically, individuals engaging in OCBI may do so with the expectation that other employees will reciprocate such behaviors (Perlow & Weeks, 2002). To the extent th at these assumptions are correct, the relationship between OCBI a nd WIF may be weakened (compared to the relationship between OCBO and WIF), given th at these individuals may have co-workers willing to help in the event of a family-relate d absence. Conversely, the majority of OCBI items require active behavioral involvement, while the OCBO construct also includes emotion-based behaviors, such as showing pr ide for the organizati on in public. Although these behaviors may not involve an extra tim e commitment, they may require the use of emotion regulation strategies which, as descri bed in the emotional la bor literatu re, can be relatively resource-intensive (Diefendorff & Go sserand, 2003). Thus, to the extent that such behaviors are perceived by employees as stressors requiring resources, they may be particularly likely to relate to work-interfering-with-family conflict.
14 Despite these propositions, ther e is a dearth of literature in this area. Thus, these statements are exploratory, and no explic it hypotheses are included regarding whether OCBO and OCBI differentially relate to WIF. The Moderating Effects of Gender The current study also inve stigated the moderating role of gender. As suggested and tested by Bolino and Turnley (2005), gender is likely to play a role in the relationship between OCB and WFC. Using role theory, they argue that interrole conflict is particularly likely to occur when other dema nds prevent an individual from carrying out the requirements of their most valued role. Although gender ro les are changing, the perception still exists that men associate more strongly with the work domain, while women place the highest value on the family (Gutek, Searle, & Klepa, 1991; McElwain, Korabik, & Rosin, 2005). Therefore, women are likely to be particularly sensitive to work demands that interfere with family ti me. Consistent with their hypothesis, Bolino and Turnley (2005) found that gender modera ted the relationship between individual initiative and work-family conflict, such that the relationship was stronger for women than for men. In order to more fully understand how gender affects the citizenship behaviorÂ– WIF relationship, it is important to examin e more precisely where gender exerts its effects, given the two proposed mechanisms shown in Figure 1. The current study proposes that gender influences both pa thways between OCB and WIF, though in different ways. Specifically, gender is expect ed to moderate the relationship between work time and WIF, as well as the relations hip between citizenship behavior and role overload.
15 The gender role perspective states that Â“people interpret their time expenditures and thus their perceptions of conflict in accordance with sex role expectationsÂ” (Gutek et al., 1991, p. 567). As a result, gender role expectations may distort the rational perspective because time in oneÂ’s Â“ownÂ” do main is viewed as less of an imposition compared to time in the Â“alternateÂ” domain (Gutek et al., 1991). Thus, womenÂ’s WFC may be particularly affected by work time, while family time is more predictive of menÂ’s WFC. In support of the gender role perspec tive, Gutek et al. (1991) found that women reported more WIF, even though the number of hours they worked was comparable to that of men. Additionally, work time interacted with gender such that working long hours was associated with WIF for women only (G utek et al., 1991). Th e nature of this interaction is similar to the one uncovered by Bolino and Turnley (2005), which indicated that the relationship between individual initiative and work-family conflict was stronger for women than for men. If individual initia tive, along with other types of OCB, is positively related to work time, as proposed in this study, then their findings may be explained by the gender role perspective. Sp ecifically, engaging in OCB is expected to increase oneÂ’s work time, thereby leading to more WIF; moreover, the relationship between work time and WIF is likely to be stronger for females than for males, given that women are more likely to perceive work time as interfering with their family role (Gutek et al., 1991). The current study tested th is proposition, hypothesi zing that gender and work time interact to predict WIF. Gender may also influence perceptions of role overload. Just as time is often viewed as more imposing when it arises from the less-valued role, demands in general may be perceived differently depending on with which role they are associated. Thus, the
16 extra demands that arise from engaging in OCBs may be considered manageable when the work role is highly valued but overwh elming when alternate roles are valued more. Consistent with this notion, role theory predic ts that role pressure s are intensified when they interfere with the role most central to oneÂ’s self-concept (Greenhaus & Beutell, 1985; Kahn et al., 1964). Because role overload is a function of oneÂ’s perceptions of the demands placed upon him/her, these heightened ro le pressures may result in higher levels of role overload. Thus, it is expected that women are more likely than men to view citizenship behaviors as exce ssive demands, with role overl oad as a probable outcome. The Moderating Effects of Percep tions of OCB as Discretionary Another potential moderator involves indi vidual differences in how employees view citizenship behaviors. As formerly disc ussed, one of the biggest points of contention in defining the OCB construct has been whet her such behaviors are truly extra-role. Organ (1997) responded to debates and critic isms by removing the word Â“discretionaryÂ” from his original definition of OCB, thereby allowing for the possibility that OCB can be either in-role or extra-role. However, employees are still likely to make this distinction, differentiating between behaviors viewed as Â“above and beyondÂ” the job requirements and those that are an expected part of the pos ition. A closer examination of role theory elucidates why these perceptions are lik ely to be affected by both situational characteristics (e.g., organizational norms for staying late) and indi vidual differences (i.e., different employees perceive the requirements of the job differently). According to Rizzo, House, and Lirtzm an (1970), roles involve a position in a social structure along with a set of expecta tions about behavior. Role expectations are communicated via role-senders, which are th en interpreted by th e person occupying the
17 role (Katz & Kahn, 1966, 1978). This interpretation process is affected by an individualÂ’s perceptions and cognitions of the message, as well as his/her own inherent biases about what behaviors are expected of a given role (Katz & Kahn, 1978); moreover, different role senders may have different role exp ectations, or vary their messages over time, complicating the process further (Kahn et al., 1964). Thus, role expectations are likely to differ across individuals, even for the same position in a single organization. Consequently, numerous researchers have proposed that the same behavior may be considered either extra-role or in-role, depending on the personÂ’s role expectations (Graham, 1991; Morrison, 1994; Van Dyne et al., 1995). In fact, Morrison (1994) found vast differences in how employees defined their jobs, and employees and supervisors differed in their classifications of which behaviors were considered discretionary. Whether an individual perc eives a given behavior as discretionary may have implications for employee outcomes, such as role overload. Morris on (1994) argues that behaviors that are considered in-role will be conceptualized differently than those classified as extra-role. Because they are c onsidered discretionary, extra-role behaviors may be viewed as optional, rather than as demands required of the position. Additionally, role pressures are intensified when noncomp liance with demands is perceived as having negative consequences (Greenhaus & Beutel l, 1985). It seems probable that negative consequences are perceived as a more lik ely outcome of noncompliance with in-role behaviors, compared to behavi ors deemed extra-role. Since ro le overload is a function of Â“legitimate role requirementsÂ” (Herman & Gyllstrom, 1977, p. 320), occurring when demands and expectations exceed a given thre shold (Rizzo et al., 1970), role overload may be less likely to occur when behaviors ar e perceived as discretionary. Accordingly, it
18 is proposed the relationship between OCB and role overload is mode rated by perceptions of OCB, such that the relationship is weaker when citizenship beha viors are perceived as extra-role. Hypotheses 1. Organizational citizensh ip behavior is positively asso ciated with work-interferingwith-family conflict (WIF). 2. Organizational citizenship be havior is positively associated with work time. 3. Organizational citizenship be havior is positively associ ated with role overload. 4. The relationship between organizational ci tizenship behavior and work-interferingwith-family conflict (WIF) is mediated by work time. 5. The relationship between organizational ci tizenship behavior and work-interferingwith-family conflict (WIF) is mediated by role overload. 6. The relationship between work time and wo rk-interfering-with-family conflict (WIF) is moderated by gender, such that the rela tionship is stronger for females than for males. 7. The relationship between organizational citiz enship behavior and role overload is moderated by gender, such that the relationshi p is stronger for females than for males. 8. The relationship between organizational citiz enship behavior and role overload is moderated by perceptions of OCB as extra-ro le, such that the relationship is stronger when OCB is perceived as non-discretiona ry than when OCB is perceived as discretionary.
19 Chapter Two Method Participants Consistent with other work-family conflict research, participants were only included if they were working at least 20 hour s a week and were either married, living with a partner, or a parent with a child livi ng at home. The initial recruitment strategy for the study was to rely on the alumni database of a large southeastern university, which consisted of 12,976 names and email addresses. However, this technique resulted in a dismal response rate: only 295 alumni filled out the survey, with supervisor data being available for a mere 21 respondents. Although a full explanation for the low response rate is indeterminable, one contributing factor wa s the number of invalid email addresses, as 3,340 emails were returned as undeliverable. Additionally, 141 emails were sent directly to SPAM, as indicated by an automatic email that was sent in res ponse to the solicitation email; and 51 participants responded that they did not meet the studyÂ’ s inclusion criteria. Of the 12,976 alumni, only 70 individuals formally declined to participate, by clicking Â“No, I will not participateÂ” on th e solicitation website, or by responding to the solicitation email. Because the alumni database was not as successful as originally hoped, other recruitment methods were employed. Inform ation about the study was posted on CraigÂ’s List, an extensive classifieds website, and a snowball approach was utilized, in which
20 friends, family members, and acquaintances of the principal investigator were asked to forward the study information to interested individuals. Fourteen people responded to the ad on CraigÂ’s List, and the snowball approach generated 37 interested individuals. Of these potential participants, 28 went on to fill out the survey (5 recruited from CraigÂ’s List, and 23 from the snowball approach). In total, 397 individuals responded that they were willing to participate in the study, and 323 went on to fill out the survey. However, 25 of these participants had a significant amount of missing da ta, and two participants did not meet the inclusion criteria; thus, these 27 indivi duals were dropped from the an alysis. This resulted in a sample of 296 participants. A determination of the Â“trueÂ” response rate is difficult, given the uncertainty in the number of emails that were never receiv ed and the number of participants who were ineligible to participate. It can be confir med that at least 467 individuals received the initial solicitation email, denoted by the fact that they followed the link to indicate whether they were willing to participate in the study. Thus, an upper-bound estimate of the response rate is 63.38 pe rcent. Conversely, one might consider that 13,027 emails were originally sent out. Subt racting out the known values of undeliverable emails, letters sent to SPAM, and ineligible participants, the number is reduced to 9,495. Using this number as a total, a lowe r-bound estimate of the respons e rate is 3.12 percent. In actuality, the Â“trueÂ” response ra te probably lies somewhere between these two numbers. The participant sample was 43.9 percent male, composed of 92.9 percent White; 4.1 percent Hispanic or Mexican American; 1 percent Black or Afri can American; and .7 percent Asian or Pacific Isla nder. One percent of the samp le reported Â“otherÂ” as their
21 race. In terms of education, .7 percent of the sample had only a high school diploma; 1.4 percent attended some colle ge; 44.9 percent had a colleg e degree; 38.5 percent had a masterÂ’s degree; and 14.5 percent had a doc toral degree. Moreover, the majority of participants were married (80.1 percent), though 7.8 percent were si ngle, living with a partner, and 12.2 per cent were single. Of the 296 participants, matched supervis or data was provide d for 38 individuals (12.84 percent of the primary participant sample; 21 from the alumni database, 4 recruited from CraigÂ’s List, and 13 from the snowball approach). However, because of missing data, three of the supervisors had to be dropped from analysis, resulting in 35 matched subordinate-supervisor pairs. The supervisor sample was 51.4 percent male, with the majority reporting being White (94.3 percent). The remaining 5.7 percent reported Black or African American as th eir race. The educational background of the supervisor sample was as follows: 2.9 per cent had attended some high school; 5.7 percent had attended some college; 25.7 percent had a college degree; 45.7 percent had a masterÂ’s degree; and 20 percent had a doctoral degree. Additional demographic information for both the primary participants and their supervisors is presented in Table 1. Measures Unless otherwise noted, responses to all scales were on a 7-point Likert-type scale, that ranged from 1 (strong ly disagree) to 7 (strongly ag ree). Scores were calculated by averaging item responses. Scales are prov ided in Appendices B, C, D, and E. Organizational citiz enship behavior OCB ratings were attained from both supervisorand self-r eported data. To measure OCBI a nd OCBO, Lee and AllenÂ’s (2002) measure was utilized. Although the Williams and Anderson (1991) scale is more
22 typically used to assess these factors, Lee and Allen (2002) criticize their measure, pointing out that it includes items that refl ect workplace deviance behavior rather than OCB. Thus, Lee and Allen (2002) modified Williams and AndersonÂ’s (1991) measure, removing such items as Â“Takes undeserved wo rk breaksÂ” and Â“Great deal of time spent with personal phone conversations.Â” The result ant scale is composed of 16 items, eight per dimension. Additionally, the OCBI scal e was supplemented with two items from another OCB scale (Schneider, Goff, Ande rson, & Borman, 2003). A sample OCBI item is Â“Willingly give your time to help others who have work-related problems,Â” while Â“Attend functions that are not re quired but that help the organizational imageÂ” is an item designed to measure OCBO. Res ponses were reported on a 7point scale, that ranged from 1 (never) to 7 (always). Lee and Allen (2002) reported s ub-scale reliabilities to be .83 (OCBI) and .88 (OCBO). The alpha coeffi cients in the present study were .86 (OCBI), .91 (OCBO), and .90 (combined) for th e self-reported data ; and .95 (OCBI), .91 (OCBO), and .96 (combined) for the supervisor-reported data. Work-to-family conflict Netemeyer, Boles, and McMurrianÂ’s (1996) WIF scale was used, which is composed of five items. A sample item includes Â“The amount of time my job takes up makes it difficult to fulfill family responsibilities.Â” Netemeyer et al. (1996) reported high alpha co efficients for the scale ( =.88-.89 for WIF and =.83-.89 for FIW across three diverse samples), and slightly higher alpha coefficients were observed in the present study, at .93. Work time One item was used to assess work time: Â“How many hours do you work in an average week? Include time spent doing job-related work at home.Â”
23 Role overload Employees provided self-reports of their role overload at work, using the role overload scale from th e Michigan Organizational Assessment Questionnaire (MOAQ; Cammann, Fichman, Jenki ns, & Klesh, 1983). The scale contains three items, including Â“I have too much wo rk to do to do everything well,Â” and the reported reliability is .65 (Ca mmann et al., 1983). Items were s lightly modified to specify the work domain; for example, Â“The amount of work I am asked to do is fairÂ” was altered to Â“The amount of work I am asked to do at my job is fair.Â” The alpha coefficient for the present study was .80. Perception that OCB is extra-role Consistent with MorrisonÂ’s (1994) method, for each OCB item, employees classified the behavior into one of two categories: Â“You see this as an expected part of your jobÂ” or Â“You see this as above and beyond what is expected for your job.Â” Each behavior clas sified under the first category was coded as one, while behaviors in the second category we re coded as two. Then, as done in Allen and Rush (1998), an overall score was co mputed by summing res ponses and dividing by the total number of behaviors. This met hod yielded a reliability of .86, as reported by Allen and Rush (1998), and an alpha co efficient of .83 in the present study. Demographics/control variables Gender was coded as follows: 1 (male) and 2 (female). Race was also collecte d, asking participants to identify themselves as American Indian or Alaskan Native; Asian or Paci fic Islander; Black or African American; Hispanic or Mexican American; White (nonHispanic); or other. Additionally, age, marital status (married; single, living with pa rtner; or single), or ganizational tenure, and salary were included as potential control va riables. Responsibility for Dependents (RFD), an index that statistically combines the num ber and age of dependents according to level
24 of responsibility (Rothausen, 1999), was in cluded as well. Although the majority of work-family research focuses solely on the number of children living at home and/or the age of the youngest child liv ing at home, the Responsib ility for Dependents index provides a more thorough assessment by considering the age of each child. Individual initiative The fifteen individual initia tive items developed by Bolino and Turnley (2005) were include d for exploratory purposes, an d both self and supervisor ratings were gathered. A sample item is Â“Checks back with the office even when he/she is on vacation.Â” Responses were reported on a 7-poin t scale, that ranged from 1 (never) to 7 (always). Bolino and Turnley (2005) reported a scale reliability of .91, and the alpha coefficient in the present study was .91 (s elf-report) and .92 (supervisor-report). Procedure Participants were contacte d via email or CraigÂ’s List asking them to voluntarily participate in the study. The solicitation email included a link to a website, whereby participants indicated whether they were will ing to participate. Specifically, they were provided with two options: Â“Yes, I am willing to participateÂ” and Â“No, I am not willing to participate.Â” Those who responded Â“noÂ” were thanked for their time and no longer contacted; those who responded Â“yesÂ” were se nt an additional email, which included a link to the survey. Survey instructions and informed consent were provided on the first page of the website, assuring participants that their responses were confidential and anonymous. Those who chose to continue were di rected to a webpage with survey items, and upon completion of the survey, participan ts were asked to submit their responses electronically. Once the survey was submitted, a debriefing page was displayed, asking the participant to give a similar survey to hi s/her supervisor (for another source of OCB
25 ratings). A separate link was provided, and the electronic link was designed such that the supervisorÂ’s responses were automati cally linked to that participant. Participants were also given the option of receiving survey pa ckets in the mail. Six participants elected this option, following a similar procedure as the online respondents. However, they filled out hard copies of study materials, were asked to manually give their supervisor the survey, a nd were provided with an envelope/prepaid postage for mailing study material s back. In order to link the tw o sources of data (self and supervisor), hard copies of surveys were ma rked with a code number. The identification numbers was unrelated to part icipant data but rather gene rated randomly to protect the anonymity of the participants. Those recruited from CraigÂ’s List were asked to contact the lead investigator, and the remaining proce dure matched that of th e other participants. Participants who indicated that they were w illing to participate in the study were sent reminder emails approximately two weeks after the initial email if they had not yet completed the survey.
26 Chapter Three Results Preliminary Analyses Descriptive data for all study variables, including means, st andard deviations, observed minimums/maximums, and alpha coefficients, are provided in Table 1. Intercorrelations among study variab les are provided in Table 2. Before conducting the primary analyses, the data was inspected to determine whether any assumptions had been violat ed. The first assumption of regression, independence, is a methodological questi on. The study design provides no reason to suppose that the participantsÂ’ responses de pended upon each other; thus, this assumption was assumed to be met. Scatterplots of variable pairs were inspected to test the assumptions of linearity and homoscedasticit y. The graphs did not indicate that nonlinear relationships were present. Additionally, the variance of the independent variables appeared relatively constant across all levels of the depe ndent variable, supporting the assumption of homoscedasticity. The data was also assessed for normali ty and outliers. Specifically, skewness and kurtosis values were computed, and histogr ams and box plots were graphed. Several of the variables were skewed or kurtotic. For ex ample, supervisor reports of OCBI, OCBO, and overall OCB were negativel y skewed. Given the positive nature of OCB, this skewness was not unexpected and is typical in the literature. Thus, it was deemed
27 inappropriate to transform the data. The dist ributions of work hours and number of kids exhibited some kurtosis, and income and or ganizational tenure were positively skewed. Again, these results are not uncommon in the lit erature, so the data was not normalized. Outliers were defined as data points falling more than three standard deviations away from the mean. Accordingly, two indivi duals had extreme scor es on self-reported overall OCB, with scores of 2.72 and 2.83. This was driven by a very low OCBO score for the first individual (1.75) and a very low OCBI score fo r the second individual (2.70). Additionally, another particip ant had an outlying score of 2.60 on self-reported OCBI, and two individuals had a low score of 2.25 on self-reported OC BO. Two individuals reported working a very high number of hour s a week, at 85 and 90, and there were multiple outliers regarding salary, including f our individuals who reported an income of over $350,000. Finally, there was one outlier for organizational tenure at 48 years, and two individuals with five kids living at home. Despite the si gnificant number of outliers, all of these values are plausible and were thus kept in the data set. Hypothesis Testing Hypotheses 1, 2, and 3 predicted that O CB would be positively associated with WIF, work time, and role overload, resp ectively. These hypotheses were tested by examining zero-order correlations between outcome variables and each dimension of OCB. Additionally, hierarchical regression was used to test whether these relationships remained significant after controlling for ge nder, age, salary, te nure, and number of children living at home. Alt hough initially the Responsibilit y for Dependents index was intended to be used rather th an number of children living at home as a control variable, the two were highly correlated ( r = .92, p < .01). Moreover, an initial assessment of the
28 Table 1. Descriptive Statistics Variable N # of Items M SD Obs. Min. Obs. Max. Study Variables OCB (self) 296 18 .90 5.31 .79 2.72 7.00 OCBI 296 10 .86 5.16 .84 2.60 7.00 OCBO 296 8 .91 5.49 1.03 1.75 7.00 OCB (supervisor) 35 18 .96 5.67 .89 3.00 7.00 OCBI 35 10 .95 5.65 .89 3.00 7.00 OCBO 35 8 .91 6.07 .76 4.13 7.00 WIF 296 5 .93 4.03 1.54 1.00 7.00 OCB Discretionary* 293 18 .83 1.38 .24 1.00 2.00 Role Overload 296 3 .80 4.09 1.49 1.00 7.00 Work Time (hrs/wk) 293 1 -46.15 11.32 20 90 II** (self) 295 15 .91 4.17 1.20 1.00 7.00 II** (supervisor) 35 15 .92 4.44 1.14 2.60 6.64 Sample Demographics Age 290 1 -46.58 9.33 22 67 Salary 277 1 -90,810 83,403 2,400 800,000 Tenure (years) 280 1 -9.72 8.74 .08 48 # Kids at Home 178 1 -1.82 .92 0 5 RFD 170 --8.57 4.57 0 21 Supervisor Demographics Age 33 1 -48.33 7.65 34 65 Tenure (years)*** 34 1 -4.07 4.32 .25 16.83 All non-demographic variables are measured on a 7-point scale unl ess otherwise noted *Measured on a 2-point response scale **Individual Initiative ***Length of time as supervisor of study participant
29 Table 2. Intercorrelati ons among Study Variables Variable 1 2 3 4 5 6 7 8 1 OCB (self) -2 OCBI (self) .86Â† -3 OCBO (self) .85Â† .47Â† -4 OCB (sup.) .04 -.00 .07 -5 OCBI (sup.) .04 .06 -.00 .97Â† -6 OCBO (sup.) .04 -.09 .16 .94Â† .84Â† -7 WIF -.05 .02 -.10 -.09 -.09 -.08 -8 OCB Discr. -.07 .03 -.15Â†.15 .15 .13 -.13* -9 Role Ovld. -.12*-.07 -.14*-.19 -.22 -.13 .65Â† -.09 10 Work Time .30Â† .26Â† .25Â† -.27 -.30 -.21 .40Â† -.04 11 Gender1 .04 .06 .01 -.22 -.29 -.11 .12* .08 12 Age .04 -.02 .09 .11 .11 .11 -.12* -.13* 13 Salary .05 -.02 .12 .03 .04 .01 .13* -.07 14 Tenure .02 -.04 .08 -.28 -.28 -.24 -.08 -.09 15 # Kids .10 .06 .11 .12 .20 -.03 .04 -.05 16 RFD .03 .00 -.05 .17 .22 .07 .10 -.13 17 II (self) .39Â† .25Â† .43Â† .06 .04 .09 .32Â† -.09 18 II (sup.) -.14 -.39*.17 .42* .37* .45Â† .05 .00 p <.05; Â† p <.01; NÂ’s ranged from 35 to 296 1Male=1; Female=2
30 Variable 9 10 11 12 13 14 15 16 17 1 OCB (self) 2 OCBI (self) 3 OCBO (self) 4 OCB (sup.) 5 OCBI (sup.) 6 OCBO (sup.) 7 WIF 8 OCB Discr. 9 Role Ovld. -10 Work Time .36Â† -11 Gender1 .21Â† -.07 -12 Age -.10 .02 -.20Â†-13 Salary -.03 .34Â† -.31Â†.08 -14 Tenure .07 .06 -.11 .37Â† .17Â† -15 # Kids .10 .12 -.12 -.09 .11 -.05 -16 RFD -.05 .11 -.08 -.31Â†.10 -.10 .92Â† -17 II (self) .44Â† .44Â† .00 .11 .16Â† .07 -.05 -.07 -18 II (sup.) .06 .12 -.03 .03 -.26 -.06 .03 -.04 .57Â† p <.05; Â† p <.01; NÂ’s ranged from 35 to 296 1Male=0; Female=1
31 regression analyses revealed that results we re similar, regardless of which index was used. Because of missing data there were more responses for number of children living at home, compared to Responsibility for Depende nts. Thus, to maximize power, number of children living at home was used as a control variable in the present study. The remaining control variables were chosen based on th eir inclusion in Bolino and TurnleyÂ’s (2005) study. The present study aimed to maximize the similarity between the two studies, in order to replicate their research. However, gi ven that many of the control variables were unrelated to the studied depende nt variable, analyses were also performed with no control variables entered into the regression equations. The two methods yielded convergent findings; thus, only the results with the contro l variables entered in step one are reported here. Relationships between OCB and the depende nt variables were assessed using both the selfand supervisor-reports of OCB. It is important to note that the correlations between self and supervisor ratings of OC B, OCBI, and OCBO were not significant ( r = .04, r = .06, and r = .16, respectively). This finding, while noteworthy, is consistent with prior research, which reported a non-signi ficant correlation of .11 between selfand supervisor-reported OCB (Allen et al., 2000). Regression results for Hypothesis 1 are pr esented in Tables 3 through 8. Tables 3, 4, and 5 report results for self-reported ove rall OCB, OCBI, and O CBO, respectively; and Tables 6, 7, and 8 report results for superv isor-reported overall O CB, OCBI, and OCBO, respectively. As presented in the tables, control variables we re entered in step one, and OCB was entered in step two. The numbers presented in the table are standardized beta weight coefficients, unl ess otherwise specified.
32 Table 3. Regression of OCB (Self-Reported) on WIF Dependent Variable: WIF Variable Step 1 Step 2 Control Variables Gender .15 .15 Age -.02 -.02 Salary .31Â† .31Â† Tenure -.16 -.16 # Kids -.05 -.04 Independent Variable OCB (self) -.04 F1 3.43Â† 2.89* Overall R2 .10 .10 Adjusted R2 .07 .07 in Adjusted R2 .00 1Step 1 df = 5, 151; Step 2 df = 6, 150 p <.05; Â† p <.01 Table 4. Regression of OCBI (Self-Reported) on WIF Dependent Variable: WIF Variable Step 1 Step 2 Control Variables Gender .15 .14 Age -.02 -.02 Salary .31Â† .31Â† Tenure -.16 -.15 # Kids -.05 -.06 Independent Variable OCBI (self) .05 F1 3.43Â† 2.93* Overall R2 .10 .10 Adjusted R2 .07 .07 in Adjusted R2 .00 1Step 1 df = 5, 151; Step 2 df = 6, 150 p <.05; Â† p <.01
33 Table 5. Regression of OCBO (Self-Reported) on WIF Dependent Variable: WIF Variable Step 1 Step 2 Control Variables Gender .15 .15 Age -.02 -.03 Salary .31Â† .32Â† Tenure -.16 -.15 # Kids -.05 .03 Independent Variable OCBO (self) -.13 F1 3.43Â† 3.36Â† Overall R2 .10 .12 Adjusted R2 .07 .08 in Adjusted R2 .01 1Step 1 df = 5, 151; Step 2 df = 6, 150 p <.05; Â† p <.01 Table 6. Regression of OCB (S upervisor-Reported) on WIF Dependent Variable: WIF Variable Step 1 Step 2 Control Variables Gender .12 -.20 Age .30 .64 Salary .35 .12 Tenure -.24 -.97 # Kids -.05 .52 Independent Variable OCB (supervisor) -.61 F1 .92 1.03 Overall R2 .31 .41 Adjusted R2 -.03 .01 in Adjusted R2 .04 1Step 1 df = 5, 10; Step 2 df = 6, 9 p <.05; Â† p <.01
34 Table 7. Regression of OCBI (Supervisor-Reported) on WIF Dependent Variable: WIF Variable Step 1 Step 2 Control Variables Gender .12 -.07 Age .30 .49 Salary .35 .23 Tenure -.24 -.69 # Kids -.05 .33 Independent Variable OCBI (supervisor) -.38 F1 .92 .80 Overall R2 .31 .35 Adjusted R2 -.03 -.09 in Adjusted R2 -.06 1Step 1 df = 5, 10; Step 2 df = 6, 9 p <.05; Â† p <.01 Table 8. Regression of OCBO (Supervisor-Reported) on WIF Dependent Variable: WIF Variable Step 1 Step 2 Control Variables Gender .12 -.10 Age .30 .55 Salary .35 .16 Tenure -.24 -.69 # Kids -.05 .25 Independent Variable OCBO (supervisor) -.46 F1 .92 1.06 Overall R2 .31 .42 Adjusted R2 -.03 .03 in Adjusted R2 .06 1Step 1 df = 5, 10; Step 2 df = 6, 9 p <.05; Â† p <.01
35 As indicated by both the zero-order correlations and the regression results, Hypothesis 1 was not supported. Neither self -reported nor superv isor-reported OCB were related to WIF ( r = -.05, r = .02, r = -.10 for self ratings of OCB, OCBI, and OCBO, respectively; r = -.09, r = -.09, r = -.08 for supervisor ratings of OCB, OCBI, and OCBO, respectively). Moreover, none of the forms of OCB (across both selfand supervisorreport) accounted for significant varian ce in WIF after accounting for the control variables ( = -.04, = .05, and = -.13 for self-reported tota l OCB, OCBI, and OCBO, respectively; and = -.61, = -.38, and = -.46 for supervisor-reported total OCB, OCBI, and OCBO, respectively). Tables 9 through 14 present the regr ession results for Hypothesis 2, which predicted a positive association between OCB and work time. Results for self-reported OCB, OCBI, and OCBO are reported in Tabl es 9, 10, and 11, while Tables 12, 13, and 14 presents results for supervisor-reported OCB, OCBI, and OCBO, respectively. Selfreported OCB significantly related to work hours ( r = .30, p < .01 for total OCB; r = .26, p < .01 for OCBI; and r = .25, p < .01 for OCBO), while su pervisor-reported OCB did not ( r = -.27 for total OCB; r = -.30 for OCBI; and r = -.21 for OCBO). While nonsignificant, it is interesting to note that the correlations between supervisor-rated OCB and work time were in the negative directio n; this is counter to both the hypothesized relationship and the results for self-repor ted OCB. The regression analyses produced similar results. Self-reported OCB and OCBI had significant positive standardized beta weights ( = .21, p < .01 and = .24, p < .01, respectively), while self-reported OCBO and supervisor-reported OCB, OCBI and OCBO were not significant ( = .11, = -.12, = -.30, and = .09, respectively). Thus, Hypothe sis 2 received partial support.
36 Table 9. Regression of OCB (S elf-Reported) on Work Time Dependent Variable: Work Time Variable Step 1 Step 2 Control Variables Gender -.02 -.03 Age -.00 .00 Salary .42Â† .41Â† Tenure -.02 -.02 # Kids .10 .07 Independent Variable OCB (self) .21Â† F1 7.62Â† 8.01Â† Overall R2 .20 .24 Adjusted R2 .18 .21 in Adjusted R2 .03 1Step 1 df = 5, 151; Step 2 df = 6, 150 p <.05; Â† p <.01 Table 10. Regression of OCBI (S elf-Reported) on Work Time Dependent Variable: Work Time Variable Step 1 Step 2 Control Variables Gender -.02 -.04 Age -.00 -.00 Salary .42Â† .41Â† Tenure -.02 -.00 # Kids .10 .07 Independent Variable OCBI (self) .24Â† F1 7.62Â† 8.56Â† Overall R2 .20 .26 Adjusted R2 .18 .23 in Adjusted R2 .05 1Step 1 df = 5, 151; Step 2 df = 6, 150 p <.05; Â† p <.01
37 Table 11. Regression of OCBO (S elf-Reported) on Work Time Dependent Variable: Work Time Variable Step 1 Step 2 Control Variables Gender -.02 -.02 Age -.00 .00 Salary .42Â† .41 Tenure -.02 -.02 # Kids .10 .08 Independent Variable OCBO (self) .11 F1 7.62Â† 6.79Â† Overall R2 .20 .21 Adjusted R2 .18 .18 in Adjusted R2 .00 1Step 1 df = 5, 151; Step 2 df = 6, 150 p <.05; Â† p <.01 Table 12. Regression of OCB (Supe rvisor-Reported) on Work Time Dependent Variable: Work Time Variable Step 1 Step 2 Control Variables Gender .51 .44 Age .34 .41 Salary .45 .41 Tenure .16 .01 # Kids .29 .40 Independent Variable OCB (supervisor) -.12 F1 1.01 .77 Overall R2 .34 .34 Adjusted R2 .00 -.10 in Adjusted R2 -.10 1Step 1 df = 5, 10; Step 2 df = 6, 9 p <.05; Â† p <.01
38 Table 13. Regression of OCBI (Supervisor-Reported) on Work Time Dependent Variable: Work Time Variable Step 1 Step 2 Control Variables Gender .51 .36 Age .34 .49 Salary .45 .36 Tenure .16 -.21 # Kids .29 .59 Independent Variable OCBI (supervisor) -.30 F1 1.01 .83 Overall R2 .34 .36 Adjusted R2 .00 -.07 in Adjusted R2 -.07 1Step 1 df = 5, 10; Step 2 df = 6, 9 p <.05; Â† p <.01 Table 14. Regression of OCBO (Supe rvisor-Reported) on Work Time Dependent Variable: Work Time Variable Step 1 Step 2 Control Variables Gender .51 .55 Age .34 .29 Salary .45 .49 Tenure .16 .24 # Kids .29 .23 Independent Variable OCBO (supervisor) .09 F1 1.01 .77 Overall R2 .34 .34 Adjusted R2 .00 -.10 in Adjusted R2 -.10 1Step 1 df = 5, 10; Step 2 df = 6, 9 p <.05; Â† p <.01
39 Hypothesis 3 proposed that OCB would be positively re lated to role overload. This hypothesis was not supported. Self-repo rted OCB and OCBO were significantly related to role overload but in th e opposite direction as hypothesized ( r = -.12, p < .05 and r = -.14, p < .05, respectively). Self-reported O CBI and supervisor-reported OCB, OCBI, and OCBO were not significan tly correlated with role overload ( r = -.07, r = -.19, r = -.22, and r = -.13, respectively). Tables 15, 16, and 17 present the regression results for self-reported OCB, OCBI, and OCBO; a nd Tables 18, 19, and 20 present regression results for supervisor-reported OCB, OCBI and OCBO, respectively. As shown in the tables, only OCBO (self-reported) had a sign ificant standardized beta weight in the regression analyses ( = -.19); self-reported OCB ( = -.11), self-reported OCBI ( = .00), supervisor-reported OCB ( = -.25), supervisor-reported OCBI ( = -.19), supervisor-reported OCBO ( = -.15) did not explain si gnificant variance in role overload, according to the regr ession analyses. Thus, the resu lts did not support a positive association between OCB and role overload; conversely, the relationships were either non-significant or negative. Hypotheses 4 and 5 proposed that the relationship between OCB and WIF would be mediated by work time and role over load, respectively. Following the strategy outlined in Baron and Kenny (1986), three st eps are required in an assessment of mediation: (1) the dependent variable (DV) is regressed onto the IV ; (2) the mediator is regressed onto the IV ; and (3) the DV is regressed ont o both the IV and the mediator. Support for mediation requires significant relation ships in the first two steps, as well as a nonsignificant relationship betw een the IV and DV when the mediator is present (step three). As previously mentioned, the re lationship between O CB and WIF was not
40 Table 15. Regression of OCB (Sel f-Reported) on Role Overload Dependent Variable: Role Overload Variable Step 1 Step 2 Control Variables Gender .18 .19* Age -.06 -.06 Salary .10 .11 Tenure .06 .06 # Kids -.13 -.12 Independent Variable OCB (self) -.11 F1 1.63 1.69 Overall R2 .05 .06 Adjusted R2 .02 .03 in Adjusted R2 .01 1Step 1 df = 5, 151; Step 2 df = 6, 150 p <.05; Â† p <.01 Table 16. Regression of OCBI (Sel f-Reported) on Role Overload Dependent Variable: Role Overload Variable Step 1 Step 2 Control Variables Gender .18 .18 Age -.06 -.06 Salary .10 .10 Tenure .06 .06 # Kids -.13 -.13 Independent Variable OCBI (self) -.00 F1 1.63 1.35 Overall R2 .05 .05 Adjusted R2 .02 .01 in Adjusted R2 -.01 1Step 1 df = 5, 151; Step 2 df = 6, 150 p <.05; Â† p <.01
41 Table 17. Regression of OCBO (Sel f-Reported) on Role Overload Dependent Variable: Role Overload Variable Step 1 Step 2 Control Variables Gender .18 .19* Age -.06 -.07 Salary .10 .12 Tenure .06 .08 # Kids -.13 -.11 Independent Variable OCBO (self) -.19* F1 1.63 2.34* Overall R2 .05 .09 Adjusted R2 .02 .05 in Adjusted R2 .03 1Step 1 df = 5, 151; Step 2 df = 6, 150 p <.05; Â† p <.01 Table 18. Regression of OCB (Superv isor-Reported) on Role Overload Dependent Variable: Role Overload Variable Step 1 Step 2 Control Variables Gender .22 .09 Age .48 .62 Salary .02 -.08 Tenure -.27 -.57 # Kids .41 .64 Independent Variable OCB (supervisor) -.25 F1 .38 .32 Overall R2 .16 .17 Adjusted R2 -.26 -.38 in Adjusted R2 -.12 1Step 1 df = 5, 10; Step 2 df = 6, 9 p <.05; Â† p <.01
42 Table 19. Regression of OCBI (Supervisor-Reported) on Role Overload Dependent Variable: Role Overload Variable Step 1 Step 2 Control Variables Gender .22 .12 Age .48 .57 Salary .02 -.04 Tenure -.27 -.50 # Kids .41 .61 Independent Variable OCBI (supervisor) -.19 F1 .38 .30 Overall R2 .16 .17 Adjusted R2 -.26 -.39 in Adjusted R2 -.13 1Step 1 df = 5, 10; Step 2 df = 6, 9 p <.05; Â† p <.01 Table 20. Regression of OCBO (Superv isor-Reported) on Role Overload Dependent Variable: Role Overload Variable Step 1 Step 2 Control Variables Gender .22 .15 Age .48 .56 Salary .02 -.05 Tenure -.27 -.42 # Kids .41 .51 Independent Variable OCBO (supervisor) -.15 F1 .38 .31 Overall R2 .16 .17 Adjusted R2 -.26 -.38 in Adjusted R2 -.12 1Step 1 df = 5, 10; Step 2 df = 6, 9 p <.05; Â† p <.01
43 significant across all forms of OCB and both sour ces of data. Thus, the first step was not satisfied for either of the two hypotheses, a nd no subsequent analyses were performed. Finally, the interactions proposed in H ypotheses 6, 7, and 8 were tested with moderated hierarchical regres sion (James & Brett, 1984). For these regression equations, control variables were added in first, followed by the independent variable. The interaction term was entered in the third step. Results were inspected to determine whether the interaction term added incremen tal variance over the i ndependent variable, thereby indicating support for moderation. Table 21. Moderated Regression of Gender on Work Time and WIF Dependent Variable: WIF Variable Step 1 Step 2 Step 3 Control Variables Gender .15 .16 -.01 Age -.02 -.02 -.03 Salary .31Â† .12 .13 Tenure -.16 -.15* -.15* # Kids -.05 -.09 -.09 Independent Variable Work Time .46Â† .32 Interaction Gender X Work Time .20 F1 3.43Â† 9.16Â† 7.86Â† Overall R2 .10 .27 .27 Adjusted R2 .07 .24 .23 in Adjusted R2 .17 -.01 1Step 1 df = 5, 151 Step 2 df = 6, 150; Step 3 df = 7, 149; p <.05; Â† p <.01
44 Hypothesis 6 proposed that the relationshi p between work time and WIF would be moderated by gender, such that the relations hip would be stronger for females than for males. The moderated regression results ar e presented in Table 21. Although work time was significantly associated with WIF ( = .46, p < .01), the interaction term did not increase the variance explained in WIF. Thus, Hypothesis 6 was not supported. Hypothesis 7 predicted that the relati onship between OCB and role overload would be moderated by gender, such that the relationship would be stronger for females than for males. Tables 22, 23, and 24 presen t the regression results using self-reported OCB, OCBI, and OCBO; and Tables 25, 26, and 27 present the regression results using supervisor-reported OCB, OCBI, and OCBO, respectively. Across both sources of OCB ratings and all three types of OCB, the interaction term was not significant, with no incremental variance in role overload being explained ( = -.40 for self ratings of OCB; = -.06 for self ratings of OCBI; = .45 for self ratings of OCBO; = -4.57 for supervisor ratings of OCB; = .40 for supervisor ratings of OCBI; and = 10.10 for supervisor ratings of OCBO). No support was found for Hypothesis 7. Hypothesis 8 proposed that the relati onship between OCB and role overload would be moderated by percepti ons of OCB as extra-role, such that the relationship would be stronger when OCB is perceived as non-discretionary than when OCB is perceived as discretionary. Results for self-r eported OCB (Table 28), OCBI (Table 29), and OCBO (Table 30), as well as superviso r-reported OCB (Table 31), OCBI (Table 32), and OCBO (Table 33), are presented. As i ndicated by the tables, the interaction terms were not significant for self-rated OCB ( = -.90), OCBI ( = .33), or OCBO ( = -.02),
45 or supervisor-rated OCB ( = 3.65), OCBI ( = 1.02) or OCBO ( = .72). So Hypothesis 8 was unsupported. Exploratory Analyses Although the present study aimed to replicate Bolino and TurnleyÂ’s (2005) findings that OCB and WIF were positively associated, no such relationship was found. Thus, it is important to consid er potential reasons for the di vergent findings. One possible explanation is that the relationship between OCB and WIF only holds for certain types of OCB, mainly individual initiative. Alternativ ely, sample differences may be the cause of the inconsistency. To explore the first possibi lity, it is necessary to assess whether the relationship between individual initiative and WIF holds in the present sample. As shown in Table 2, WIF was, in fact, significantly correlated with self-reported individual initiative ( r = .32, p < .01). Conversely, WIF and supervisor-reported individual initiative were not significantly related ( r = .05). Table 34 and Table 35 present results for the regression of WI F on self-reported and supervisor-reported individual initiative, respectively. Unlike the more general measure of OCB, self-reported individual initiative does explain significan t variance in WIF ( = .36, p < .01). This is very similar in magnitude to the standardiz ed beta weight of .37 observed by Bolino and Turnley (2005). The beta weight for supe rvisor-reported individual initiative, though similar in magnitude, was not significant ( = .32).
46 Table 22. Moderated Regression of Gender on OCB (Self-Reported) and Role Overload Dependent Variable: Role Overload Variable Step 1 Step 2 Step 3 Control Variables Gender .18 .19* .53 Age -.06 -.06 -.06 Salary .10 .11 .11 Tenure .06 .06 .06 # Kids -.13 -.12 -.11 Independent Variable OCB (self) -.11 .05 Interaction Term Gender X OCB (self) -.40 F1 1.63 1.69 1.50 Overall R2 .05 .06 .07 Adjusted R2 .02 .03 .02 in Adjusted R2 .01 .00 1Step 1 df = 5, 151; Step 2 df = 6, 150; Step 3 df = 7, 149 p <.05; Â† p <.01 Table 23. Moderated Regression of Gender on O CBI (Self-Reported) and Role Overload Dependent Variable: Role Overload Variable Step 1 Step 2 Step 3 Control Variables Gender .18 .18 .23 Age -.06 -.06 -.06 Salary .10 .10 .10 Tenure .06 .06 .06 # Kids -.13 -.13 -.13 Independent Variable OCBI (self) -.01 .02 Interaction Term Gender X OCBI (self) -.06 F1 1.63 1.35 1.15 Overall R2 .05 .05 .05 Adjusted R2 .02 .01 .01 in Adjusted R2 -.01 .00 1Step 1 df = 5, 151; Step 2 df = 6, 150; Step 3 df = 7, 149 p <.05; Â† p <.01
47 Table 24. Moderated Regression of Gender on OCBO (Self-Repor ted) and Role Overload Dependent Variable: Role Overload Variable Step 1 Step 2 Step 3 Control Variables Gender .18 .19* .57 Age -.06 -.07 -.07 Salary .10 .12 .11 Tenure .06 .08 .06 # Kids -.13 -.11 -.10 Independent Variable OCBO (self) -.19* .04 Interaction Term Gender X OCBO (self) -.45 F1 1.63 2.34* 2.10* Overall R2 .05 .09 .09 Adjusted R2 .02 .05 .05 in Adjusted R2 .03 .00 1Step 1 df = 5, 151; Step 2 df = 6, 150; Step 3 df = 7, 149; p <.05; Â† p <.01 Table 25. Moderated Regression of Gender on OCB (Supervisor-Reported) and Role Overload Dependent Variable: Role Overload Variable Step 1 Step 2 Step 3 Control Variables Gender .22 .09 5.03 Age .48 .62 .511 Salary .02 -.08 .28 Tenure -.27 -.57 -.13 # Kids .41 .64 .39 Independent Variable OCB (supervisor) -.25 2.12 Interaction Term Gender X OCB (supervisor) -4.57 F1 .38 .32 .30 Overall R2 .16 .17 .21 Adjusted R2 -.26 -.38 -.49 in Adjusted R2 -.12 -.23 1Step 1 df = 5, 10; Step 2 df = 6, 9; Step 3 df = 7, 8; p <.05; Â† p <.01
48 Table 26. Moderated Regression of Gender on OCBI (Supervisor-Reported) and Role Overload Dependent Variable: Role Overload Variable Step 1 Step 2 Step 3 Control Variables Gender .22 .12 3.77 Age .48 .57 .47 Salary .02 -.04 .20 Tenure -.27 -.50 -.06 # Kids .41 .61 .34 Independent Variable OCBI (supervisor) -.19 1.87 Interaction Term Gender X OCBI (supervisor) .40 F1 .38 .30 .36 Overall R2 .16 .17 .24 Adjusted R2 -.26 -.39 -.42 in Adjusted R2 -.13 -.03 1Step 1 df = 5, 10; Step 2 df = 6, 9; Step 3 df = 7, 8; p <.05; Â† p <.01 Table 27. Moderated Regression of Gender on OCBO (Supervisor-Rep orted) and Role Overload Dependent Variable: Role Overload Variable Step 1 Step 2 Step 3 Control Variables Gender .22 .15 -10.31 Age .48 .56 .33 Salary .02 -.05 -.65 Tenure -.27 -.42 -.27 # Kids .41 .51 .32 Independent Variable OCBO (supervisor) -.15 -4.46 Interaction Term Gender X OCBO (supervisor) 10.10 F1 .38 .31 .33 Overall R2 .16 .17 .22 Adjusted R2 -.26 -.38 -.46 in Adjusted R2 -.12 -.08 1Step 1 df = 5, 10; Step 2 df = 6, 9; Step 3 df = 7, 8; p <.05; Â† p <.01
49 Table 28: Moderated Regressi on of Perceptions of OCB as Discretionary on OCB (SelfReported) and Role Overload Dependent Variable: Role Overload Variable Step 1 Step 2 Step 3 Control Variables Gender .19 .22 .24 Age -.07 -.08 -.08 Salary .04 .04 .04 Tenure .08 .08 .09 # Kids -.06 -.05 -.04 Independent Variables OCB (self) -.14 -.71 OCB Discretionary -.16 -.85 Interaction Term OCB Discretionary X OCB -.90 F1 2.18 2.92Â† 2.82Â† Overall R2 .05 .10 .11 Adjusted R2 .03 .06 .06 in Adjusted R2 .03 .00 1Step 1 df = 5, 192; Step 2 df = 7, 190; Step 3 df = 8, 189 p <.05; Â† p <.01
50 Table 29: Moderated Regression of Perceptions of OCB as Discretionary on OCBI (SelfReported) and Role Overload Dependent Variable: Role Overload Variable Step 1 Step 2 Step 3 Control Variables Gender .19* .21* .21* Age -.07 -.08 -.08 Salary .04 .04 .04 Tenure .08 .08 .08 # Kids -.06 -.06 -.05 Independent Variables OCBI (self) -.05 -.25 OCB Discretionary -.07 -.31 Interaction Term OCB Discretionary X OCBI .33 F1 2.18 1.78 1.59 Overall R2 .05 .06 .06 Adjusted R2 .03 .03 .02 in Adjusted R2 .00 -.01 1Step 1 df = 5, 192; Step 2 df = 7, 190; Step 3 df = 8, 189 p <.05; Â† p <.01
51 Table 30: Moderated Regressi on of Perceptions of OCB as Discretionary on OCBO (Self-Reported) and Role Overload Dependent Variable: Role Overload Variable Step 1 Step 2 Step 3 Control Variables Gender .19* .20Â† .21Â† Age -.07 -.07 -.07 Salary .04 .04 .03 Tenure .08 .09 .09 # Kids -.06 -.06 -.06 Independent Variables OCBO (self) -.19Â† -.18 OCB Discretionary -.19Â† -.19* Interaction Term OCB Discretionary X OCBO -.02 F1 2.18 3.83Â† 3.34Â† Overall R2 .05 .12 .12 Adjusted R2 .03 .09 .09 in Adjusted R2 .06 .00 1Step 1 df = 5, 192; Step 2 df = 7, 190; Step 3 df = 8, 189 p <.05; Â† p <.01
52 Table 31: Moderated Regre ssion of Perceptions of OC B as Discretionary on OCB (Supervisor-Reported) and Role Overload Dependent Variable: Role Overload Variable Step 1 Step 2 Step 3 Control Variables Gender .49 .48 .48 Age .37 .41 .38 Salary .13 .09 .07 Tenure -.35 -.46 -.57 # Kids .43 .49 .62 Independent Variables OCB (supervisor) -.06 -1.72 OCB Discretionary -.10 -2.78 Interaction Term OCB Discretionary X OCB 3.65 F1 .70 .46 .43 Overall R2 .21 .23 .25 Adjusted R2 -.09 -.27 -.34 in Adjusted R2 -.18 -.07 1Step 1 df = 5, 13; Step 2 df = 7, 11; Step 3 df = 8, 10 p <.05; Â† p <.01
53 Table 32: Moderated Regressi on of Perceptions of OCB as Discretionary on OCBI (Supervisor-Reported) and Role Overload Dependent Variable: Role Overload Variable Step 1 Step 2 Step 3 Control Variables Gender .49 .44 .46 Age .37 .37 .35 Salary .13 .10 .09 Tenure -.35 -.39 -.38 # Kids .43 .46 .48 Independent Variables OCBI (supervisor) -.12 -.57 OCB Discretionary .12 -.66 Interaction Term OCB Discretionary X OCBI 1.02 F1 .70 .46 .37 Overall R2 .21 .23 .23 Adjusted R2 -.09 -.27 -.39 in Adjusted R2 -.18 -.12 1Step 1 df = 5, 13; Step 2 df = 7, 11; Step 3 df = 8, 10 p <.05; Â† p <.01
54 Table 33: Moderated Regressi on of Perceptions of OCB as Discretionary on OCBO (Supervisor-Reported) and Role Overload Dependent Variable: Role Overload Variable Step 1 Step 2 Step 3 Control Variables Gender .49 .43 .19 Age .37 .47 .45 Salary .13 -.05 -.29 Tenure -.35 -.50 -.35 # Kids .43 .41 .12 Independent Variables OCBO (supervisor) .03 -.24 OCB Discretionary -.36 -.80 Interaction Term OCB Discretionary X OCBO .72 F1 .70 .67 1.04 Overall R2 .21 .30 .45 Adjusted R2 -.09 -.15 .02 in Adjusted R2 -.06 .17 1Step 1 df = 5, 13; Step 2 df = 7, 11; Step 3 df = 8, 10 p <.05; Â† p <.01 Table 34. Regression of Individual Initiative (Self-Reported) on WIF Dependent Variable: WIF Variable Step 1 Step 2 Control Variables Gender .15 .15 Age -.01 -.04 Salary .31Â† .24Â† Tenure -.16 -.14 # Kids -.06 -.04 Independent Variable Individual Initiative (self) .36Â† F1 3.43Â† 7.20Â† Overall R2 .10 .23 Adjusted R2 .07 .19 in Adjusted R2 .12 1Step 1 df = 5, 150; Step 2 df = 6, 149 p <.05; Â† p <.01
55 Table 35. Regression of Individual In itiative (Superviso r-Reported) on WIF Dependent Variable: WIF Variable Step 1 Step 2 Control Variables Gender .12 .30 Age .30 .37 Salary .35 .53 Tenure -.24 -.16 # Kids -.05 -.04 Independent Variable Individual Initiative (supervisor) .32 F1 .92 .93 Overall R2 .31 .38 Adjusted R2 -.03 -.03 in Adjusted R2 .00 1Step 1 df = 5, 10; Step 2 df = 6, 9 p <.05; Â† p <.01
56 Chapter Four Discussion The current study was designed to addre ss a limitation in the OCB literature: the consistent trend of treating the construc t as a wholly positive phenomenon. Although citizenship behaviors have been related to positive outcomes for individuals and for organizations (see Podsakoff et al., 2000 for a review), few studies have examined whether negative consequences are also asso ciated with engaging in OCB. Bolino and TurnleyÂ’s (2005) research represents an important exception, and the present study was designed to replicate and ex tend their findings. Additiona lly, this study intended to address some limitations of their study, by utilizing a more general OCB measure, collecting supervisor ratings of citizenship performance, and examining potential mediators and moderators to the OCBÂ—WIF relationship. Despite these intentions, the study Â’s hypotheses were largely unsupported. Neither supervisornor self-reported O CB were significantly related to WIF. Additionally, self-re ported, but not supervis or-reported, OCB was significantly related to work hours. Contrary to expect ations, self-reported OCBO was negatively related to role overload, and the relationship between all ot her types of OCB and role overload was nonsignificant. Although the study hypothesized that work tim e and role overload would mediate the relationship between OCB and WI F, the conditions of mediation were not met, as the direct relationship between OCB and WIF was not si gnificant. Finally, none
57 of the proposed interactions were significant, as gender did not moderate the relationship between OCB and role overload; perceptions of OCB as discretionary did not moderate the relationship between OCB and role ove rload; and gender did not moderate the relationship between work time and WIF. Because these findings diverged from that of Bolino and Turnley (2005), it is fruitful to consider poten tial reasons for the discrepancy. As Bolino and Turnley (2005) mention in their article, their measure of i ndividual initiative, de signed specifically for their study, may be more strong ly related to WIF than othe r OCB measures. Thus, it is possible that the relationsh ip between OCB and WIF is limited to the specific OCB dimension of individual initiative. Consis tent with this hypothesis, self-reported individual initiative and WIF were positiv ely related in the present sample. Theoretical Implications The present study highlight s several important theo retical implications. Specifically, OCB should not be treated as a wholly positive phenomenon. Although WIF was not related to OCBI or OCBO in the present study, the relationship between WIF and the specific dimension of i ndividual initiative was signifi cant. The divergent findings may reflect the nature of the individual initi ative factor, which incl udes behaviors that are particularly likely to interf ere with oneÂ’s work-life balance (e.g., bringing work home; working during vacations). While negative i ndividual-level consequences may only be associated with certain types of citizenship behavior, it is important for researchers to broaden the scope of OCB research, consider ing potential positive and negative aspects of the construct.
58 By testing for the moderation of gender in the relationship between OCB and role overload, and between work time and WIF, the present study provided a test of the gender role perspective. This theory proposes that peopleÂ’s perceptions of time and role conflict depend upon sex role expectations (G utek et al., 1991); in light of this perspective, the null findings in the present study are surprisin g. It is likely that societal views about sex role expectations have changed considerably in the past sixteen years, as an increasing number of women have en tered the workforce. Women may have internalized the work role more than they did previously, resulting in diminished gender differences in how work-related demands are perceived. Thus, as compared to the gender role perspective, the rational model of WFC may receive increasing support in the future, as the notion that work is a manÂ’s domain is diminished. The present study supported this notion, finding that work time was positively rela ted to WIF but that the relationship was consistent for men and women. Another theoretical implication elucidat ed by the present study concerns the dimensionality of the OCB construct. As previously mentioned, OCB has suffered from conceptual ambiguity, and there is a lack of consensus rega rding its factor structure. Although the present study ca nnot speak to the number of factors underlying the construct, the results do support the multidim ensionality of OCB, as the relationship between OCB and outcome variables varied acr oss different dimensions of citizenship behavior. For example, OCBO but not OCBI was negatively associated with role overload. Additionally, individual initiative, bu t OCBI and OCBO, related to WIF. Thus, the present study provides some support for di fferentiating citizenship behavior according to target as well as type of behavior.
59 Researchers have also disagreed on whethe r citizenship behavior can be described as discretionary. Contrary to OrganÂ’s (1988) original definition of the construct, the present study supports the notion that citizenship behaviors ar e frequently viewed as an expected part of oneÂ’s job. Moreover, th ere was considerable variance in these perceptions. Such findings are consistent wi th past research (Morrison, 1994). Thus, researchers are encouraged to refrain from defining OCB as discretionary behavior. From a methodological standpoi nt, the present study high lights the importance of collecting OCB ratings from multiple sources. The correlation between self and supervisor ratings of OCB ranged from .04 to .16 in the present study, confirming that the two sources are offering distin ct information. Interestingly, the relationship between selfand supervisor-ratings of individual initia tive was much higher, at .57. This may stem from the fact that certain types of OCB are more observable and sali ent to supervisors. A sample OCBO is keeping up with developmen ts in the organization, and a sample OCBI is assisting others with their duties. Supe rvisors may be less aware of these types of behaviors as compared to those encompasse d by the individual initiative factor, which includes working during vacations and coming in to work early. Thus, while collecting OCB ratings from multiple sources is always useful, the amount of unique information provided by each source may depend upon the spec ific type of citizenship behavior. Practical Implications The present study has practical implications as well. Organizations may implicitly or explicitly encourage employees to engage in citizenship behaviors, through the culture, norms, policies, and/or managerial attit udes and behaviors. If employees experience negative outcomes as a result of engaging in OCB, then organizations that encourage
60 such behaviors may face adverse consequences in the long run. The results of this study suggest that only certain types of OCB, mainly individual initi ative, are associated with WIF. Conversely, OCBO was negatively related to role overload. The reasons for this finding are unclear. Perhaps individuals who engage in OCBO are more committed to their organization and less likely to see th e extra behaviors as a demanding burden. In accordance with this supposition, support for a positive relationship between citizenship behaviors and organizational commitment has been found in many studies (see Podsakoff, MacKenzie, & Bommer, 1996). A lternatively, perhaps the direction of causality is reversed, such that indivi duals who experience role overload respond by engaging in fewer citizenship behaviors, and those with low role overload perform more OCBs. While the present study cannot speak to the viability of these potential explanations, it does suggest that orga nizations should pay attention to the type of OCB that is encouraged; some behaviors may have positive implications, while others may be associated with negative outcomes. Limitations It is important to highlight the limitations of the present study. Perhaps the biggest problem involved sampling issues. The study su ffered from a poor response rate, which creates concerns regarding the representa tiveness of the sample. One of the major constructs of interest in the present study was OCB. Filling out the survey, as well as asking oneÂ’s supervisor to do the same, may itself be considered an example of a prosocial or citizenship behavior. Thus, the sample as a whole, and the subset with supervisor data in particular, may differ from the population in ways that are central to the tenets of the study. Moreover, the samp le was highly educat ed, with 53 percent
61 having a graduate degree, and of high SES, with a mean salary of over $90,000. Additionally, 94.3 percent of the sample repor ted being of white race. These sample characteristics are likely a function of the recruitment strategies, which relied on an alumni database, CraigÂ’s List and a snowball approach. Gi ven the non-representativeness of the sample, it is unclear the extent th at these findings generalize to the broader population. Another problem with the sample was the small number of supervisor data that was available. With only 35 matched partic ipant-supervisor pair s, the current study suffered from a lack of power to detect rela tionships, particularly with respect to the interactions. Although study hypotheses were also tested using self ratings of OCB, this approach raises concerns of common method variance, whic h can serve to inflate the relationship between variables. Thus, a larger sample of ma tched data would have been optimal. Finally, the cross-sectiona l nature of the study precl udes any statements of causality. In fact, this may provide another pote ntial explanation for the lack of relationship between OCB and WIF. While the present study hypothesized that engaging in OCB leads to enhanced WIF, other resear chers have proposed that the direction of causality is reversed. Specifically, it has been suggested that in dividuals experiencing work-family conflict will be less likely to engage in citizenship behaviors (Bragger, Rodriguez-Srednicki, Kutcher, Indovino, & Ro sner, 2005). In support of this proposition, two studies uncovered a negativ e relationship between OCB and WFC (Bragger et al., 2005; Netemeyer, Maxham, & Pullig, 2005). Thus it is possible that the relationship between OCB and WIF is cyclical, where by engaging in OCB increases WIF, which
62 causes the employee to respond by performing fewer citizenship behaviors. A crosssectional approach cannot speak to this possibility. Future Directions Because this area of research has been largely neglected, there are many avenues available for future research. First, futu re research should address some of the methodological limitations of the present st udy, by replicating this study with a more representative sample of partic ipants and a larger sample of supervisors. The field would also benefit from an exploration of thes e relationships using non-cross-sectional methodologies, such as longitudinal or experience sampling approaches. Conceptually, a more in-depth investigat ion of the relationship between OCB and WIF is needed. The present study highlighted that the relationship may be limited to certain types of citizenship behaviors, such as individual initiative. Future research should explore more specifically which behavi ors are detrimental to balancing work and family life. In addition to a more fine-grained analysis of individual initiative, it may be fruitful to explore other dimensions of OCB, using an OCB scale that conceptualizes the construct according to type of behavior, rather th an target. Similarly, more information may be gained by using a WIF scale that differentiates between time-, strain-, and behavior-based conflict. For example, sports manship, defined as tole rating less than ideal situations without complaining (Organ, 1988) may be emotionally draining, leading to stain-based WIF, while individual initiative is more likely related to time-based WIF. It is also necessary to explore condi tions that may mitigate or exacerbate the relationship between OCB and WI F. Several researchers have suggested that there are a variety of motives for engaging in OCB, some of which may be self-serving (Bolino et
63 al., 2004; Eastman, 1994; Rioux & Penner, 2001). In fact, engaging in certain types of OCB may actually be an effective strategy fo r balancing work and family roles. For example, an individual who engages in OCB ma y do so with the hope that the behavior is later reciprocated, to build up a supportive network of co -workers, or to win favor with his/her supervisor. Each of these outcomes coul d lead to future benefits for the employee should a conflict arise between his/her wo rk and family roles. Moreover, the organizational culture may come into play, as organizations vary in the extent to which cooperative behavior is encouraged. In a coope rative climate, citizenship behavior may be expected, but it is also likely to be r eciprocated, thereby miti gating the relationship between OCB and WIF. Thus, to fully understa nd the nature of th e relationship between WIF and OCB, a systems approach may be necessary, which considers the dynamic relationship between multiple systems Â– including coworkers, supervisors, and the organizational culture Â– rather than the indi vidual in isolation. Qu alitative research may enhance our understanding in this regard. Other factors, such as individual differe nces, may also impact the relationship between OCB and WIF. Researchers have suggested that the relationship between stressors and strain may be moderated by time management behaviors (Jex & Elacqua, 1999). Although Jex and Elacqua (1999) did not find support for this claim in their study, time management skills may be particularly useful for employees struggling to balance multiple roles, including that of organizati onal-member, job-holder, and family-member. Thus, the relationship between OCB and WI F may be moderated by time management skills, such that the relationship is weakened for individuals that ar e able to effectively manage their time. Another potential moderator is need for affiliation. Individuals with a
64 high need for affiliation require harmonious re lationships with other people (McClelland, 1987), so they may be particularly likely to en gage in altruistic behaviors. However, these individuals are less likely to view OCB as a burden; on th e contrary, by satisfying their needs, citizenship behaviors may reduce their strain and enhance their ability to juggle multiple roles. Future research should e xplore this possibility and consider other personality factors that may also impact the relationship between OCB and WIF. Research in this area should not be lim ited to WIF, as OCB may lead to other negative individual-level consequences as we ll, such as job stress (Bolino & Turnley, 2005) and role ambiguity (Bolino et al., 2004) Moreover, researchers should focus on steps that organizations can take to prevent the potential negative consequences of engaging in OCB, such as training employees on how to balance multiple demands and training supervisors to be aware of the type s of behaviors they are encouraging their subordinates to conduct. Conclusion Although the hypotheses were largely unsupported, the present study addressed an important research question that has receiv ed limited attention in our field. Moreover, the present study has signifi cant theoretical and practical implications. Because few studies have been done on this topic, additional research is needed. By fully exploring the nature of the OCB construct, including both positive and negative aspects, we can shed light on a neglected topic.
65 References Aguinis, H. (2002). Estimation of intera ction effects in organization studies. Organizational Research Methods, 5 (3), 207-211. Allen, T. D., Barnard, S., Rush, M. C., & Russell, J. E. A. (2000). Ratings of organizational citizenship behavior: Does the source make a difference? Human Resource Management Review, 10 (1), 97-114. Allen, T. D., Herst, D. E. L., Bruck, C. S., & Sutton, M. (2000). Consequences associated with work-to-family conflict: A review and agenda for future research. Journal of Occupational Health Psychology, 5 (2), 278-308. Allen, T. D., & Rush, M. C. (1998). The effect s of organizational citizenship behavior on performance judgments: A field stud y and a laboratory experiment. Journal of Applied Psychology, 83 (2), 247-260. Baron, R. M., & Kenny, D. A. (1986). The m oderator-mediator variab le distinction in social psychological research: Conceptual, strategic and statistic al considerations. Journal of Personality and Social Psychology, 51 1173-1182. Bateman, T. S., & Organ, D. W. (1983). Job satisfaction and the good soldier: The relationship between affect and employee Â“citizenship.Â” The Academy of Management Journal, 26 (4), 587-595.
66 Bolino, M. C., & Turnley, W. H. (2005). The personal costs of citizenship behavior: The relationship between individual initiative and role overload, job stress, and workfamily conflict. Journal of Applied Psychology, 90 (4), 740-748. Bolino, M. C., Turnley, W. H., & Niehoff, B. P. (2004). The other side of the story: Reexamining prevailing assumptions about organizational citizenship behavior. Human Resource Management Review, 14 229-246. Borman, W. C., & Motowidlo, S. J. (1993). Expanding the criterion domain to include elements of contextual performance. In N. Schmitt & W. C. Borman (Eds.), Personnel selection in organizations (pp. 71-98). San Francisco, CA: JosseyBass. Bragger, J. D., Rodriguez-Srednicki, O., Kutcher, E. J., Indovino, L., & Rosner, E. (2005). Work-family conflict, work-family culture, and organiza tional citizenship behavior among teachers. Journal of Business and Psychology, 20 303-324. Brief, A. P., & Motowidlo, S. J. (1986) Prosocial organizational behaviors. The Academy of Management Review, 11 (4), 710-725. Bruck, C. S, & Allen, T. D. (2005). Theoretical antecedents to work-family conflict: A comprehensive investigation. Unpublished manuscript. Byron, K. (2005). A meta-analytic review of work-family conflict and its antecedents. Journal of Vocational Behavior, 67, 169-198. Cammann, C., Fichman, M., Jenkins, G. D., & Klesh, J. (1983). Mich igan organizational assessment questionnaire. In S. E. Seashor e, E. E. Lawler, P. H. Mirvis, & C. Camman (Eds.), Assessing organizational change: A guide to methods, measures, and practices (pp. 71Â–138). New York: Wiley-Interscience.
67 Cardenas, R. A., Major, D. A., & Bernas, K. H. (2004). Exploring work and family distractions: Antecedents and outcomes. International Journal of Stress Management, 11 (4), 346-365. Chen, P. Y., & Spector, P. E. (1991). Negativ e affectivity as the underlying cause of correlations between stressors and strain. Journal of Applied Psychology, 76 (3), 398-407. Cohen, J. (1988). Statistical power analysis fo r the behavioral sciences ( 2nd ed .). New York: Academic Press. Coleman, V. I., & Borman, W. C. (2000). I nvestigating the underl ying structure of the citizenship performance domain. Human Resource Management Review, 10 (1), 25-44. Conway, J. M. (1999). Distinguishing contextu al performance from task performance for managerial jobs. Journal of Applied Psychology, 84 (1), 3-13. Diefendorff, J. M., & Gosserand, R. H. ( 2003). Understanding th e emotional labor process: A control theory perspective. Journal of Organizational Behavior, 24, 945-959. Eastman, K. K. (1994). In the eyes of the beholder: An attribu tional approach to ingratiation and organizati onal citizenship behavior. Academy of Management Journal, 37 1379-1391. Frone, M. R. (2003). Work-family balance. In J. C. Quick & L. E. Tetrick (Eds.), Handbook of Occupational Health Psychology (pp. 143-162) Washington, DC: American Psychological Association.
68 Frone, M. R., Russell, M., & Cooper, M. L. (1992). Antecedents and outcomes of workfamily conflict: Testing a mode l of the work-family interface. Journal of Applied Psychology, 77 (1), 65-78. Frone, M. R., Yardley, J. K., & Markel, K. S. (1997). Developing and testing an integrative model of th e work-family interface. Journal of Vocational Behavior, 50 145-167. George, J. M., & Brief, A. P. (1992). Feeli ng good-doing good: A conc eptual analysis of the mood at work-organizational spontaneity relationship. Psychological Bulletin, 112 (2), 310-329. Graham, J. W. (1991). An essay on or ganizational citizenship behavior. Employee Rights and Responsibilities, 4 (4), 249-270. Greenhaus, J. H., Bedeian, A. G., & Mossholde r, K. W. (1987). Work experiences, job performance, and feelings of personal and family well-being. Journal of Vocational Behavior, 31 (2), 200-215. Greenhaus, J. H., & Beutell, N. J. (1985). Sources of conflict between work and family roles. Academy of Management Review, 10 76-88. Gutek, B. A., Searle, S., & Klep a, L. (1991). Rational versus gender role explanations for work-family conflict. Journal of Applied Psychology, 76 (4), 560-568. Herman, J. B., & Gyllstrom, K. K. (1977). Working men and women: Interand intrarole conflict. Psychology of Women Quarterly, 1 (4), 319-333. Ilies, R., Nahrgagn, J. D., & Morgeson, F. P. (2007). Leader-member exchange and citizenship behaviors: A meta-analysis. Journal of Applied Psychology, 92 (1), 269-277.
69 James, L. R., & Brett, J. M. (1984). Medi ators, moderators, and tests for mediation. Journal of Applied Psychology, 69 307-321. Jex, S. M., & Elacqua, T. C. (1999). Time management as a moderator of relations between stressors and employee strain. Work & Stress, 13 (2), 182-191. Kahn, R. L., Wolfe, D. M., Quinn, R., S noek, J. D., & Rosenthal, R. A. (1964). Organizational stress. New York: John Wiley & Sons. Katz, D., & Kahn, R. L. (1966). The social psychology or organizations New York: John Wiley & Sons. Katz, D., & Kahn, R. L. (1978). The social psychology of organizations (2nd ed.). New York: John Wiley & Sons. Lee, K., & Allen, N. J. (2002). Organiza tional citizenship behavior and workplace deviance: The role of affect and cognitions. Journal of Applied Psychology, 87 (1), 131-142. MacKenzie, S. B., Podsakoff, P. M., & Fett er, R. (1993). The imp act of organizational citizenship behavior on evaluations of salesperson performance. Journal of Marketing, 57 (1), 70-80. Major, V. S., Klein, K. J., & Ehrhart, M. G. (2002). Work time, work interference with family, and psychological distress. Journal of Applied Psychology, 87 (3), 427436. McClelland, D.C. (1987). Human motivation Cambridge, NY: Cambridge University Press.
70 McElwain, A. K., Korabik, K., & Rosin, H. M. (2005). An examination of gender differences in work-family conflict. Canadian Journal of Behavioural Science, 37 (4), 283-298. McNeely, B. L., & Meglino, B. M. (1994). Th e role of dispositional and situational antecedents in prosocial organizational be havior: An examination of the intended beneficiaries of pr osocial behavior. Journal of Applied Psychology 79 836Â–844. Morrison, E. W. (1994). Role definitions a nd organizational citizen ship behavior: The importance of the employeeÂ’s perspective. The Academy of Management Journal, 37 (6), 1543-1567. Netemeyer, R. G., Boles, J. S., & McMurri an, R. (1996). Development and validation of work-family conflict and family-work conflict scales. Journal of Applied Psychology, 81 (4), 400-410. Netemeyer, R. G., Maxham, J. G. I., & Pull ig, C. (2005). Conflicts in the work-family interface: Links to job stress, custom er service employee performance, and customer purchase intent. Journal of Marketing, 69 130-143. Organ, D. W. (1988). Organizational citizenship behav ior: The good soldier syndrome Lexington, MA: Lexington Books. Organ, D. W. (1997). Organizatio nal citizenship behavior: It Â’s construct clean-up time. Human Performance, 10 (2), 85-97. Organ, D. W., & Ryan, K. (1995). A meta-analy tic review of attitudi nal and dispositional predictors of organizatio nal citizenship behavior. Personnel Psychology, 48 (4), 775-802.
71 Perlow, L., & Weeks, J. (2002). WhoÂ’s help ing whom? Layers of culture and workplace behavior. Journal of Organizational Behavior, 23, 345-361. Podsakoff, P. M., Ahearne, M., & MacKenzie, S. B. (1997). Organizational citizenship behavior and the quantity and qua lity of work group performance. Journal of Applied Psychology, 82, 262-270. Podsakoff, P. M., & MacKenzie, S. B. ( 1994). Organizational citi zenship behaviors and sales unit effectiveness. Journal of Marketing Research, 31 351-363. Podsakoff, P. M., & MacKenzie, S. B. ( 1997). Impact of organi zational citizenship behavior on organizational performance: A review and suggestions for future research. Human Performance, 10 (2), 133-151. Podsakoff, P. M., MacKenzie, S. B., & Bo mmer, W. H. (1996). Meta-analysis of the relationships between Kerr and JermierÂ’s substitutes for leadership and employee job attitudes, role perceptions, and performance. Journal of Applied Psychology, 81 (4), 380-399. Podsakoff, P. M., MacKenzie, S. B., & Hui, C. (1993). Organizational citizenship behaviors and managerial evaluations of employee performance: A review and suggestions for future research. In G. R. Ferris (Ed.), Research in personnel and human resources management (Vol. 11, pp. 1-40). Greenwich, CT: JAI Press Inc. Podsakoff, P. M., MacKenzie, S. B., Moorman, R. H., & Fetter, R. (1990). Transformational leader behaviors and thei r effects on followers' trust in leader, satisfaction, and organizationa l citizenship behaviors. The Leadership Quarterly, 1(2) 107-142.
72 Podsakoff, P. M., MacKenzie, S. B., Paine, J. B., & Bachrach, D. G. (2000). Organizational citizenship be haviors: A critical review of the theoretical and empirical literature and suggestions for future research. Journal of Management, 26 (3), 513-563. Rothausen, T. J. (1999). Â‘FamilyÂ’ in organiza tional research: A revi ew and comparison of definitions and measures. Journal of Organizational Behavior, 20 (6), 817-836. Rioux, S. M., & Penner, L. A. (2001). The cause s of organizational ci tizenship behavior: A motivational analysis. Journal of Applied Psychology, 86 (6), 1306-1314. Rizzo, J. R., House, R. J., & Lirtzman, S. I. (1970). Role conflict and ambiguity in complex organizations. Administrative Science Quarterly, 15 (2), 150-163. Roehling, P., Moen, P., & Batt, R. (2003). Spillover. In P. Moen (Ed.), ItÂ’s about time (pp. 101-121). Ithaca, NY: Corn ell University Press. Rothbard, N. P. (2001). Enriching or deplet ing? The dynamics of engagement in work and family roles. Administrative Science Quarterly, 46 655-684. Schneider, R. J., Goff, M., Anderson, S ., & Borman, W. C. (2003). Computerized adaptive rating scales for measur ing managerial performance. International Journal of Selection and Assessment, 11 (2/3), 237-246. Somech, A., & Drach-Zavavy, A. (2004). Explor ing organizational ci tizenship behaviour from an organizational perspective: The relationship between organizational learning and organizational citizenship behaviour. Journal of Occupational and Organizational Psychology, 77, 281Â–298. Staines, G. L. (1980). Spillover versus compen sation: A review on the literature between work and non-work. Human Relations, 33 111-129.
73 Thompson, B. M., Kirk, A., & Brown, D. F. (2005). Work based support, emotional exhaustion, and spillover of work stress to the family environment: A study of policewomen. Stress and Health, 21, 199-207. Van Dyne, I., Cummings, L. L., & Parks, J. M. (1995). Extra-role beha viors: In pursuit of construct and definitional clarity (a bridge over muddied waters) In L. L. Cummings & B. M. Staw (Eds.), Research in organizational behavior (Vol. 17, pp. 215-285). Greenwich, CT: JAI Press, Inc. Van Scotter, J. R., Motowidlo, S. J., & Cross, T. C. (2000). Effects of task performance and contextual performance on systemic rewards. Journal of Applied Psychology, 85 (4), 526-535. Welbourne, T. M., Johnson, D. E., & Erez, A. (1998). The role-based performance scale: Validity analysis of a theory-based measure. The Academy of Management Journal, 41 (5), 540-555. Werner, J. M. (1994). Dimensions that make a difference: Examining the impact of inrole and extrarole behavior s on supervisory ratings. Journal of Applied Psychology, 79 (1), 98-107. Williams, K. J., & Alliger, G. M. (1994). Role stressors, mood spillover, and perceptions of work-family conflict in employed parents. The Academy of Management Journal, 37 (4), 837-868. Williams, L. J., & Anderson, S. E. (1991) Job satisfaction and organizational commitment as predictors of organizati onal citizenship and in-role performance. Journal of Management, 17 (3), 601-617.
75 Appendix A Hypothesized Relationships Figure 1 Hypothesized relationships between O CB and WIF, including mediating and moderating variables. Organizational Citizenship Behavio r Work Time Work-Interfering-withFamil y Conflict Role Overload Gender Perceptions of OCB as Discretionar y Gender
76 Appendix B Organizational Citizenship Behaviors Scale Items* OCBI Items 1. Helps others who have been absent. 2. Willingly give your time to help others who have work-related problems. 3. Adjust your work schedule to accomm odate other employeesÂ’ requests for time off. 4. Go out of the way to make newer empl oyees feel welcome in the work group. 5. Show genuine concern and courtesy to ward coworkers, even under the most trying business or personal situations. 6. Give up time to help others who have work or nonwork problems. 7. Assist others with their duties. 8. Share personal property with ot hers to help their work. 9. Always shows consideration for others, even when especially busy or stressed. 10. Goes out of his or her way to congrat ulate others for their achievements. OCBO Items 1. Attend functions that are not required but that help the organizational image. 2. Keep up with developments in the organization. 3. Defend the organization when ot her employees criticize it. 4. Show pride when representi ng the organization in public. 5. Offer ideas to improve the f unctioning of the organization. 6. Express loyalty toward the organization.
77 7. Take action to protect the organi zation from potential problems. 8. Demonstrate concern about the image of the organization. *Lee & Allen (2002); Schneider, Goff, Anderson, & Borman (2003)
78 Appendix C Work-to-Family Conflict Scale Items* 1. The demands of my work interfere with my home and family life. 2. The amount of time my job takes up ma kes it difficult to fulfill family responsibilities. 3. Things I want to do at home do not get done because of the demands my job puts on me. 4. My job produces strain that makes it difficult to fulfill family duties. 5. Due to work-related duties, I have to make changes to my plans for family activities. *Netemeyer, Boles, & McMurrian (1996)
79 Appendix D Role Overload Scale Items* 1. I never seem to have enough time at work to get everything done. 2. I have too much work to do to do everything well at my job. 3. The amount of work I am asked to do at my job is fair. (R) *Cammann, Fichman, Jenkins, & Klesh (1983)
80 Appendix E Individual Initiative Scale Items* 1. Checks his/her e-mail or voice mail from home. 2. Works on his/her days off (e.g., weekends). 3. Brings things home to work on. 4. Takes work-related phone calls at home. 5. Carries a cell phone or pager for work so he/she can be reached after normal business hours. 6. Stays at work after normal business hours. 7. Works late into the night at home. 8. Attends work-related functions on his/her personal time. 9. Travels whenever the company asks him/ her to, even though technically he/she doesnÂ’t have to. 10. Works during his/her vacations. 11. Goes into the office before normal business hours. 12. Volunteers for special projects in ad dition to his/her normal job duties. 13. Rearranges or alters his/her pe rsonal plans because of work. 14. Checks back with the office even when he/she is on vacation. 15. Participates in community activities fo r the benefit of his/her company or organization. *Bolino & Turnley (2005)