Conflict, justice, and inequality: Why perceptions of leader–member exchange differentiation hurt performance in teams

• Author: Maria L. Kraimer,Daejeong Choi,Scott E. Seibert
• Published date: 01 July 2020
• DOI: http://doi.org/10.1002/job.2451
• Date: 01 July 2020

RESEARCH ARTICLE

Conflict, justice, and inequality: Why perceptions of leader–

member exchange differentiation hurt performance in teams

Daejeong Choi

1

| Maria L. Kraimer

2

| Scott E. Seibert

2

1

Department of Management and Marketing,

Faculty of Business and Economics, University

of Melbourne, Carlton, Victoria, Australia

2

School of Management and Labor Relations,

Rutgers University, Piscataway, New Jersey, U.

S.A.

Correspondence

Daejeong Choi, PhD, Department of

Management and Marketing, Faculty of

Business and Economics, University of

Melbourne, Level 10, 198 Berkeley Street,

Carlton, Victoria 3010, Australia.

Email: daejeong.choi@unimelb.edu.au

Summary

To better understand why leader–member exchange (LMX) differentiation in teams

may be detrimental to individual and team performance, we propose that team mem-

bers' perception of LMX differentiation (PLMXD) is more important than statistical

measures of LMX differentiation. Specifically, we hypothesize a multilevel model in

which relationship conflict and procedural justice (climate) mediate the relationships

of individual and collective PLMXDs with individual and team performance, respec-

tively. Using a sample of 235 individuals in 53 teams, we found that individual

PLMXD was negatively related to individual performance through relationship con-

flict perceptions, controlling for LMX. At the team level, collective PLMXD was nega-

tively related to team performance through procedural justice climate and

relationship conflict, controlling for a statistical measure of LMXD. Theoretical impli-

cations and directions for future research are explored.

KEYWORDS

conflict, differentiation, justice, leader–member exchange, multilevel model

1|INTRODUCTION

A high-quality leader–member exchange (LMX) relationship with one's

leader—indicated by high levels of mutual trust, liking, respect, and

loyalty (Liden & Maslyn, 1998)—is positively associated with individual

performance (Dulebohn, Bommer, Liden, Brouer, & Ferris, 2012;

Gerstner & Day, 1997; Martin, Guillaume, Thomas, Lee, &

Epitropaki, 2016). However, it is not clear that a team leader can or

should develop a high-quality relationship with every team member,

because not every team member will be equally capable and depend-

able (Bauer & Green, 1996). Instead, team leaders may choose to form

relationships of differing quality with each of the team members, a

process known as LMX differentiation (Dansereau, Graen, &

Haga, 1975; Liden, Erdogan, Wayne, & Sparrowe, 2006).

The effects of LMX differentiation in teams are, however, unclear

due to an inherent tension. Developing higher quality relationships

with the most capable team members is consistent with the equity

principle of reward fairness and is likely to promote individual perfor-

mance, but inconsistent with the equality principle of reward fairness

and may lead to a loss of harmony in the team (Leventhal, 1976). To

better understand this important theoretical and practical issue, much

of the research in this area has shifted from studying dyadic LMX to

the team level and the consequences of LMX differentiation (Anand,

Vidyarthi, & Park, 2015). However, to date, research evidence has

been inconclusive and contradictory (Anand et al., 2015; Martin,

Thomas, Legood, & Dello Russo, 2018; Yu, Matta, & Cornfield, 2018).

For example, the relationship between LMX differentiation and team

performance has been found to be negative (Li & Liao, 2014), positive

(Naidoo, Scherbaum, Goldstein, & Graen, 2011), or curvilinear (Sui,

Wang, Kirkman, & Li, 2016). Research on the impact of LMX differen-

tiation on individual performance has been more limited but has

shown positive effects only under certain conditions, such as low indi-

vidual LMX quality (Liden et al., 2006) and high levels of distributive

justice climate (Haynie, Cullen, Lester, Winter, & Svyantek, 2014).

We argue that these inconsistent findings are due to three prob-

lems in the LMX differentiation literature that our work seeks to rem-

edy. First, LMX differentiation has been operationalized in many

different ways, including objective indirect statistical measures

(i.e., standard deviation or variance of LMX scores within a team), sub-

jective indirect measures of the pattern of the relationships in the

Received: 27 March 2019 Revised: 29 April 2020 Accepted: 30 April 2020

DOI: 10.1002/job.2451

J Organ Behav. 2020;41:567–586. wileyonlinelibrary.com/journal/job © 2020 John Wiley & Sons, Ltd. 567

team, and (less often) subjective direct measures of perceptions of

LMX differentiation (Martin et al., 2018). We discuss this issue in

more detail in the next section. Second, as Martin et al. (2018) noted,

the majority of researchers use indirect statistical measures of LMX

differentiation, which is not well suited to capture the underlying

explanatory concepts, such as justice (e.g., Chen, He, & Weng, 2018;

Cobb & Lau, 2015; Erdogan & Bauer, 2010), social comparisons

(e.g., Henderson, Wayne, Shore, Bommer, & Tetrick, 2008), and social

identity perspectives (e.g., Sui et al., 2016). Such statistical measures

reflect aggregate group properties that are not necessarily known to

any individual team member, making it hard to understand how or

why differentiation will affect team member attitudes and behavior.

Third, perhaps related to the conceptualization of differentiation as a

statistical group property, there has been little theoretical work

explaining why and how LMX differentiation is related to both team

and individual performance. Instead of focusing on explanatory mech-

anisms (i.e., mediators), much of the research has examined boundary

conditions (i.e., moderators). As Yu et al. (2018) noted, although the

moderator approach is informative, it usually relies on implicit

assumptions about the logic relating differentiation to performance

but does not directly test such assumptions.

To address these limitations, we develop a novel approach to

conceptualizing and operationalizing LMX differentiation at both the

individual and team levels. We then test a multilevel mediated model

explaining the LMX differentiation–performance relationship. Building

upon Martin et al.'s (2018) argument, we introduce perception of

LMX differentiation (PLMXD) as a construct that can be defined and

measured at the individual and team levels. We define individual per-

ception of LMX differentiation (individual PLMXD) as an individual's

perception of the extent to which the leader treats some team mem-

bers better than others. At the team level, we define collective per-

ception of LMX differentiation (collective PLMXD) as members'

shared perceptions of the extent to which the leader treats some

team members better than others. We argue that individual and col-

lective PLMXDs negatively relate to individual and team performance,

respectively, mediated by justice (climate) and relationship conflict.

Our study contributes to the LMX differentiation literature in

three ways. First, we develop and validate the PLMXD construct.

PLMXD is a construct that directly captures team members' percep-

tions that the leader treats some members better than others. From a

theoretical perspective, it is better suited, than statistical measures of

LMX differentiation, to test the effects of LMX differentiation on

social construction processes, such as justice and conflict perceptions

(Martin et al., 2018). Given that most LMX differentiation research

relied on social construction processes to explain the LMX

differentiation–performance relationship, statistical measurements of

LMXD are misaligned with the theory. PLMXD resolves this mis-

alignment as it is a direct measure of LMX differentiation perceptions.

Due to the ubiquitous use of indirect statistical measures, it has been

overlooked how important individual perceptions of LMX differentia-

tion are to explaining performance and to what extent indirect statisti-

cal measures overlap with actual perceptions of LMX differentiation.

The potential gap between direct perceptual and indirect statistical

measures is not uncommon in other literatures such as person–

environment fit (see Kristof-Brown & Guay, 2011, for a review). Our

development and empirical validation of the new PLMXD construct

should thus facilitate future empirical testing of other theoretical

mechanisms relating the consequences of LMX differentiation at the

individual and team levels.

Second, our work extends the LMX differentiation literature by

examining LMX differentiation in a way consistent with the subjective

and emergent processes associated with differentiation at the individ-

ual and team levels. Our theoretical model is built upon an equality

perspective of allocation preference theory (Leventhal, 1976), which

argues that individual and collective PLMXDs are negatively related to

individual and team performance, respectively, because team mem-

bers tend to prefer having resources equally, rather than equitably,

allocated among team members. We provide a conceptual account

that explains why PLMXD captures the equality perspective better as

well as clear and consistent empirical evidence with regard to the neg-

ative relationship between LMX differentiation and performance. We

recognize that Yu et al. (2018) found a positive direct relationship

between LMX differentiation and team performance (in addition to a

negative relationship); however, none of their mediators could explain

this positive relationship. Further, their meta-analysis clearly showed

that LMX differentiation is negatively related to team performance

through several social construction processes such as justice and con-

flict. Consistent with this, we propose and test a negative relationship

between PLMXD and performance through justice and conflict

because PLMXD is likely to capture such social construction pro-

cesses (Martin et al., 2018).

Third, we test the relationships of PLMXD not only with team

performance but also with individual performance, which has been

examined in only a handful of studies (e.g., Gooty &

Yammarino, 2016; Haynie et al., 2014; Henderson et al., 2008; Liden

et al., 2006). We note that Yu et al.'s (2018) meta-analysis is also lim-

ited to the team level. Because PLMXD is a perceptual measure, it can

be examined at the individual or team level, with the latter as an

emergent team perception. This helps validate our new PLMXD con-

struct and produces greater clarity in the literature regarding the

empirical effects of LMX differentiation in teams. Together, these

contributions will help us understand the consequences of LMX for

individuals and teams.

2|PERCEPTION OF LMX

DIFFERENTIATION

Theoretically, PLMXD captures individual team members' perceptions

regarding the degree to which the leader treats team members differ-

ently through their behaviors and actions. Consistent with LMX the-

ory (Liden & Maslyn, 1998), the leader's differential treatment may be

based on liking, loyalty, support, and/or professional respect. In com-

parison, LMX differentiation has been defined as the “process by

which a leader, through engaging in differing types of exchange pat-

terns with subordinates, forms different quality exchange

568 CHOI ET AL.