Allan Lee PhD Research Proposal 2010

54
1 PhD Proposal 2010 – Allan Lee PhD Proposal Allan Lee The Development of the Leader-Member Exchange (LMX) Relationship: A Longitudinal Analysis

Transcript of Allan Lee PhD Research Proposal 2010

Page 1: Allan Lee PhD Research Proposal 2010

1

PhD Proposal

Allan Lee

The Development of the Leader-Member Exchange (LMX) Relationship: A

Longitudinal Analysis

Page 2: Allan Lee PhD Research Proposal 2010

2Table of Contents

Page Number(s)

Abstract 3

Overview of the Proposed Research 4-6

Relevance to ABS Research Priorities 6

Research Objectives and Intended Contributions of the Research 6-8

Conceptual Model 9

Extended Literature Review, Theoretical Framework and Development of Hypotheses 10- 27

Hypothesis 1a &b 16

Hypothesis 2a-d 17

Hypothesis 3a & b 19

Hypothesis 4a & b 20

Hypothesis 5a & b 21-22

Hypothesis 6a & b 24

Hypothesis 7a-d 25

Hypothesis 8a & b 27

Proposed Research Methodology 27-30

Reducing the Threat of Common Method Variance 30

Brief Overview of Subsequent PhD Research Studies 30-31

Timetable for Proposed Research 31

References 32-36

Page 3: Allan Lee PhD Research Proposal 2010

3Abstract

Research has demonstrated the importance of leader–member relationships for a host of important

outcomes, including satisfaction, performance, and citizenship behaviours. Yet, relatively little

research has explored how these relationships develop over time; despite constant calls in the

literature. Using longitudinal design and cross-lagged panel analysis the study proposes to examine

the development of leader–member relationships in established dyads. The mechanisms by which

the relationship develops would be tested using established LMX theories as well as relevant

theories from Social Psychology and Relationship Science. In particular the role reciprocation of

relationship effort and performance would be tested along with attributions, expectations and social

comparisons. The use of a cross-lagged panel design would also allow for the examination of the bi-

directional effects between LMX and theoretically-guided variables. Furthermore both leader and

member perspectives would be measured in order to provide a more complete understanding of

how different variable affect each member’s perceptions of the LMX relationship. Specific

suggestions for how subsequent studies can build upon this proposed research design are also

discussed.

Page 4: Allan Lee PhD Research Proposal 2010

4Overview of the Proposed Research

Leadership solves the problem of how to organise collective effort; consequently, it is one of

the keys to organisational effectiveness (Hogan & Kaiser, 2005). The Leader-Member Exchange

(LMX) theory has evolved into one of the more insightful and useful approaches to examine the link

between leadership processes and outcomes. Indeed, LMX is currently the second most researched

theory of leadership and recent years have seen this trend continue (Yammarino, Dionne, Chun &

Dansereau, 2005). The unique focus of LMX is on the quality of the relationship between a leader

and a subordinate. Prominence is also given to the idea that leaders adopt different styles when

dealing with different members; a practice referred to as LMX differentiation (Liden, Erdogan,

Wayne & Sparrowe, 2006). This relationship heterogeneity distinguishes LMX from more traditional

leadership theories, which tend to assume that leaders develop fairly similar relationships with all of

the members that are working below them; known as an average leadership style approach

(Dansereau, Graen & Haga, 1975). The LMX approach to leadership draws upon social exchange

theory (Blau, 1964) and the norm of reciprocity (Gouldner, 1960), and argues that (mutual) influence

is stronger when it is derived through processes of social exchange as it engenders unspecified

obligations between dyadic partners to repay favourable treatment (Henderson, Liden, Glibkowski &

Chaudhry, 2009). There has been considerable cross-sectional research demonstrating that leader-

subordinate relationships have a significant impact on important employee attitudes and

behaviours, such as job satisfaction, job performance and organisational citizenship behaviour (for a

review see Erdogan & Liden, 2002; Martin, Epitropaki, Thomas & Topaka, 2010). In fact, Gerstner

and Day (1997), on the basis of their meta-analytic review, concluded that the relationship with

one’s supervisor acts as a lens through which the entire work experience is viewed.

Despite the wealth of research on LMX there are still substantial gaps in the literature that

merit further research. Typically an over-reliance on cross-sectional designs, a tendency to examine

the relationship solely from the member’s perspective and a concentration on the early stages of

LMX relationship have been observed in the research (Nahrgang, Morgeson & Ilies, 2009). In

particular, there is relatively little understanding of how the LMX relationship, once established, is

actively maintained and developed. This dearth of empirical research investigating the process of

LMX development has been highlighted on numerous occasions in reviews of the literature (e.g.,

Avolio, Walumbwa & Weber, 2009; Erdogan & Liden, 2002; Gerstner & Day, 1997; Graen & Uhl-Bien,

1995; Liden, Wayne & Stilwell, 1993; Martin et al., 2010; Yukl, 2009), and thus represents a very

important area to be addressed in my proposed PhD research.

Page 5: Allan Lee PhD Research Proposal 2010

5The proposed research would draw upon established LMX theories, such as the Role-Making

Model (Graen & Scandura, 1987) and the Leadership Making Model (Graen & Uhl-Bien 1991) as a

framework to investigate the process of LMX development. Although a range of variables have been

posited to influence LMX development, based on these theories there are two antecedents are likely

to be particularly germane: namely, performance and relationship effort. Initial levels of in-role

performance (or competence) has been viewed as the primary determinant (and outcome) of LMX

quality (e.g., Graen & Scandura, 1987; Graen & Uhl-Bien, 1995), whilst perceived effort has been

advanced as underlying all aspects of the social exchange process (Maslyn & Uhl-Bien, 2001). In

addition, the extent to which the leader’s (or member’s) effort directed towards building and

maintaining the LMX relationship has been reciprocated that is likely to play a crucial role in LMX

development. Surprisingly, there has been little longitudinal research that has rigorously tested the

extent to which these antecedents are prospectively associated with changes in the quality of LMX

over time (see Martin et al., 2010). A notable exception, however, is a longitudinal study by Nahrang

et al. (2009) on newly established dyads which found that both performance and certain aspects of

personality influenced the development of LMX quality, but had differential effects over time.

Interestingly, personality was more influential during the first few interactions, whereas

performance had more impact over the first few weeks of the LMX relationship. This result highlights

the different role played by antecedents of LMX quality across time. This study, however, did not

measure relationship effort and only tracked the development of LMX quality over the very early

stage of the relationship. There are good theoretical and empirical reasons (outlined below) to

believe that the predictors of LMX development may vary across the life cycle of the LMX

relationship. Hence, the proposed research intends to build upon the Nahrang et al. (2009) study by

examining the prospective effects of relationship effort and performance on LMX development over

longer periods of time, in the context of more established dyads, and from the perspective of both

leaders and subordinates.

The current proposal also intends to supplement established LMX theories with relevant

theories from Social Psychology and Relationship Science, such as Attribution theory (see Fincham,

1985; Fincham, Beach & Nelson, 1987) and Social Comparison theory (Festinger, 1954), in order to

more closely examine the mechanisms underlying the social exchanges in leader-member

relationships. For example, the attribution-relationship quality hypothesis posits that positive or

benign causal attributions that accentuate the impact of positive relationship behaviour and

minimise the impact of negative relationship behaviour will enhance future relationship quality. In

addition, applying the logic of Social Comparison theory and LMX differentiation (Henderson et al.,

2008), LMX quality is not only likely to be influenced by the amount of relationship effort expended

Page 6: Allan Lee PhD Research Proposal 2010

6within a particular leader-member dyad, but also by the amount of relative or comparative effort

that the leader expends with other LMX relationships within the workgroup (i.e. between-dyad

effort differentiation). That is, if a subordinate perceives that a leader puts comparatively more

effort into relationships with other subordinates than the leader does with his or her own

relationship then the resulting perceptions of unfairness and inequity are likely to undermine the

quality of the LMX relationship. Because membership of work groups and the treatment of work

group members is likely to vary over time, even established LMX dyads should be vulnerable to such

social comparisons, and hence this process of effort differentiation is likely to influence the

development of LMX relationships. Moreover, it is plausible that attributions may also moderate the

relationship between performance and relationship effort and changes in LMX quality over time

(please see below for a more detailed discussion of these moderating effects). To date, these direct

and moderating effects of LMX development in more established dyads has not yet been empirically

tested, and thus this represents an important task that the proposed research will address.

Relevance of Proposed Research to ABS Research Priorities

It is important to note that the research domain of leadership, and in particular LMX, is an

area of research strength in the Work and Organisational Psychology Group. For example, my

supervisor Dr Geoff Thomas, Professor Robin Martin, Dr Olga Epitropaki and Dr Yves Guillaume have

an ongoing research interest in LMX. For example, they have previously published a variety of

research articles, a recent review chapter of LMX research in the International Handbook of

Industrial and Organizational Psychology, and are currently conducting future studies as well as a

meta-analytic review of the outcomes of LMX (of which I am currently involved as a research

assistant). Therefore, my proposed research would fit well with the research priorities of both my

supervisor and the Work and Organizational Psychology Group.

Research Objectives and Intended Contributions to the LMX Literature

To summarise, the research objectives and intended theoretical and empirical contributions

of the proposed PhD research are as follows:

1. Guided by the Role Making and Leadership Making theories of LMX, the proposed research

aims to provide a more rigorous examination than prior research of the effect of

performance and relationship effort on the development of LMX quality over time. This aim

addresses the frequent call in the literature to address the process of LMX development – a

topic that has been largely neglected by previous researchers (see Martin et al., 2010).

Page 7: Allan Lee PhD Research Proposal 2010

72. The cross-fertilisation of Social Psychological and Relationship Science theories (e.g.,

Attribution theory; Social Comparison theory) with LMX research will also be used to

advance the understanding of the basic processes of LMX development. Relying on such

theoretical approaches provides a novel set of hypotheses, including the testing of plausible

direct (e.g., attributions; expectations; effort differentiation) and moderating effects of LMX

development.

3. The research intends to use a more appropriate design and analytic strategy to measure

relationship development than has been used in the vast majority of previous LMX studies.

Specifically, by moving beyond a cross-sectional design to a (longitudinal) cross-lagged panel

design allows for the examination of the bi-directional effects between LMX and

theoretically-guided variables (see Fincham, Beach, Harold & Osborne, 1997; Fincham,

Harold & Gano-Phillips., 2000). One of the critical advantages of employing a cross-lagged

panel designs is that it permits a better test of the causal reciprocal effects between the key

variables (both antecedents and consequences) and LMX quality over time (see Figure 1).

4. The research intends to test the process of LMX development within established LMX dyads.

For the most part, research attention has focused on the developmental processes that

occur during the genesis of the LMX relationship because it is assumed that this constitutes a

critical period in the development of the relationship. However, there are good reasons why

LMX relationships continue to develop over more prolonged periods of time. Moreover, it

plausible that the predictors of LMX development in the early stages of the relationship are

different to those in the more mature stages of the relationship. Existing empirical research

is largely silent about the nature of such development; making research with established

dyads a critical area of research.

5. The assessment of both leader and subordinate reports of LMX allows for the investigation

of LMX development from both perspectives. This is important because a recent met-

analytic review concluded only moderate levels of agreement exists between leader and

member judgements of the quality of their relationship (Sin, Nahrang & Morgeson, 2009)

To my knowledge only one study (Nahrang et al., 2009) has examined LMX development

from the perspective of both members of the dyad, and this study only measured two

antecedents (personality and performance). Therefore, including both leader and member

Page 8: Allan Lee PhD Research Proposal 2010

8perspectives should provide a more nuanced view of LMX development as well as helping to

minimize any problems associated with common method variance.

6. An ancillary aim of the research is also to examine the consequences of LMX over time. Prior

LMX research has typically assessed the “consequences” of LMX using cross-sectional deigns,

and therefore the temporal status of such variables (e.g., performance) as consequences is

implied rather than directly assessed. Moreover, those that have measured the effect of

LMX on performance at a later date have often neglected to control for initial levels of

performance. The proposed research will examine the effect of LMX on changes in

important outcomes over time, including certain outcome variables (e.g., relationship effort;

attributions) that, to date, have not been assessed by LMX researchers.

Page 9: Allan Lee PhD Research Proposal 2010

9Conceptual Model

Time 1 Time 2

A

C D

B

Figure 1: A theoretical framework illustrating the focus of the current study: The cross-lagged correlational stability design.

The path labels (A, B, C, and D) are used to describe the key features of the proposed conceptual model. The path from Wave 1 LMX to Wave 2 LMX (Path A) represents the stability or reliability of the measure of LMX across time; this reliability is anticipated to be moderate to moderately high, thus leaving a significant portion of unexplained variance in LMX for the antecedents to explain. The path from the antecedents at Wave 1 to the consequences at Wave 2 (Path B) represents the stability or reliability of the target measure across time. The paths of greater theoretical interest represent the effect of the antecedents (e.g., performance) at Wave 1 on LMX 6 months later (Path C) and the effect of LMX at Wave 1 on the consequences (e.g., performance) 6 months later (Path D). These crossed paths represent the ability of the proposed antecedents to predict changes in LMX over the 6 months and the ability of LMX to predict changes in the proposed consequences over the 6 months.

LMX

Antecedents:

Relationship effort

Effort Differentiation

Performance

Attributions

Expectations

LMX

Antecedents:

Relationship effort

Effort Differentiation

Performance

Attributions

Expectations

Page 10: Allan Lee PhD Research Proposal 2010

10Extended Literature Review and Development of Research Hypotheses

The Leader-Member Exchange Relationship: A Brief Overview

Leader-member Exchange (LMX) theory has become a very important and popular theory

within the leadership literature. In order to accurately conceptualise the developmental trends

within the LMX domain, Graen and Uhl-Bien (1995) employed a categorisation system which

outlined four stages of LMX theory development detailing the way in which research has progressed

since its inception. These stages are the discovery of differential dyads, the focus on the LMX

relationship and its outcomes, the description of dyadic partnership building and the examining of

LMX at the group and networks level. At the core of LMX theory is the recognition of LMX

differentiation, relating to stage one of the theory’s development. LMX theory posits that the quality

of the members exchange relationship with the leader, which is based upon the degree of support

provided by the leader and the exchange of valued resources, is pivotal in determining the member’s

fate within the organisation. LMX relationships can range from high-quality dyadic exchanges which

are based on mutual trust, respect and obligation to low-quality LMX relationships which are simply

based on the terms of the formal employment contract (Dansereau et al. 1975). Extensive research

has supported the prevalence of such differentiation across a variety of team and organisational

settings. For example, Liden and Graen (1980) report that approximately 85-90% of leaders

developed relationships of differing quality with their subordinates.

In terms of the second stage of LMX theory development, a wealth of research has identified

numerous antecedents which have been shown to be significantly associated with the quality of LMX

relationships. Examples of such antecedents include, the member characteristics of locus of control

(e.g. Kinicki & Vecchio, 1994; Martin, Thomas, Charles, Epitropaki & Mcnamara 2005), extraversion

(Phillips & Bedeian, 1994), personality traits (Bernerth, Armenakis, Feild, Giles & Walker, 2007) and

implicit leadership theories (Engle & Lord, 1997; Epitropaki & Martin, 2005) as well as a large range

of leader characteristics, interactional variables and contextual variables (Erdogan & Liden, 2002).

Furthermore a number of key organisational outcomes such as; employee satisfaction, job

performance and intention to remain within an organisation have been linked to LMX relationship

quality (Gerstner & Day, 1997). In addition, high quality LMX relationships have been found to have

significant direct effects upon a number of employee attitudes such as organisational citizenship

behaviour (Liden, Sparrowe & Wayne, 1997) as well as several moderators of performance such as

high conscientiousness (Kamdar & Van Dyne, 2007) and high goal commitment (Klein & Kim, 1998).

It should be noted, however, that the vast majority of such research has been cross-sectional and

Page 11: Allan Lee PhD Research Proposal 2010

11therefore the temporal categorisation of such variables as either antecedents or outcomes is merely

inferred rather than directly assessed – an issue that the proposed research intends to address.

The third stage of theory development is of particular interest to the present research

proposal as it concerns profiling the development of LMX relationships and how leaders work with

each member to develop effective relationships. The key difference of this stage to earlier stages is

that research has moved away from examining leaders having different quality relationship with

different subordinates to how leaders develop these relationships. Rousseau (1998) concluded that

the process by which LMX relationships are developed has not been clearly articulated and has been

treated as a ‘black box’ of LMX research. It is particularly important to understand how LMX

relationships develop as such knowledge would enable both leaders and subordinates to have

greater influence over the way in which their LMX relationship develops. The benefit of being able to

manage the way in which the relationship unfolds would not only improve the workplace experience

for both leaders and members, but would also benefit the organisation as a whole.

Finally, the traditional focus of LMX research has been to investigate the linkages between

leadership processes and organisational outcomes through its unique adoption of the dyadic

relationship between a leader and a member of the team as the level of analysis (Graen & Uhl-Bien,

1995). However, stage four in the theories development recognises that LMX relationships do not

occur in isolation but within a network of other relationships and the wider organisation. Therefore,

recent advancements in the theory have been related to the expansion of research to the group and

network level of analysis, this view will be incorporated in the current research through the

consideration of how social comparison information of the wider work group influences subsequent

LMX development.

Graen and Uhl-Bien (1995) observe that some areas of research such as the identification of

LMX antecedents and outcomes are more advanced whilst the latter stages are still in the initial

stages of development. A recent review of the LMX literature by Martin et al (2010) anticipates that

there will be an increased shift in research emphasis toward the later stages of LMX theory

development, therefore making the current research highly relevant to what is most needed within

the LMX domain. The fairly narrow focus of LMX research to date is not the only criticism that has

been levelled at this area. A further limitation is that the majority of studies investigating the

construct have relied on cross-sectional designs (Gerstner & Day, 1997); as a result the causal nature

of variables associated with LMX quality has been largely inferred. Furthermore, despite the

theoretical emphasis on the reciprocal influence between leader and member and the importance of

both leader’s and member’s perspective of the relationship, this has been rarely examined. Common

Page 12: Allan Lee PhD Research Proposal 2010

12practice in existing research is the examination of the relationship solely from the member’s

perspective (e.g. Scandura & Schriesheim, 1994). Studies that have examined leader perspectives

have often found a lack of convergence between leaders’ and members’ views of the relationship

(for meta-analytic reviews, see Gerstner & Day, 1997; Sin et al., 2009). The lack of understanding of

why leaders and members may view the relationship differently is a major limitation of LMX

considering its emphasis upon LMX relationship being an objective mutually perceived construct

(Brower, Schoorman & Hwee Hoon, 2000). The current study aims to examine the development of

the LMX relationship longitudinally, from the perspective of both members of the dyad. This will

enable a better understanding of how both individuals influence the relationship and the extent to

which perceptions of LMX quality (by both) are related to individual and organisational level

outcomes.

LMX Development: Theoretical Perspectives

LMX theory draws upon social exchange theory (Blau, 1964) and the norm of reciprocity

(Gouldner, 1960) to explain how the LMX relationship is likely to develop. Specifically, LMX theory

argues that relationships develop via a process of exchanging a variety of tangible and intangible

commodities within a leader-member dyad (Schriesheim, Castro & Cogliser, 1999). A leader, for

example, may exchange resources such as, information, influence, desired tasks and support in

return for the services of the employee, such as task performance, commitment, loyalty and

organisational citizenship behaviours (Martin et al. 2010).

One early perspective on how LMX develops has been proposed by Liden and colleagues

(Dienesch & Liden, 1986; Liden et al. 1993) who view such exchanges as a series of steps that

commence during initial interactions. This is then followed by a sequence of exchanges in which

both members of the dyad attempt to build the components of trust, respect and obligation that are

required for the existence of high quality relationships. The authors also posit that if the reception of

exchange behaviour is positive and the individual who initiated the exchange is satisfied with the

response then the exchange relationship will be likely to continue. However if the response to an

exchange is negative (i.e. not reciprocated) or if this exchange relationship is not initiated in the first

place then there will be very limited opportunities for a high quality exchange relationship to occur

or be maintained.

Dyadic social exchanges also form the basis of the two major theoretical models that

attempt to explain LMX development; the ‘Leadership Making Model’ (Graen & Uhl-Bien, 1991;

Graen & Uhl-Bien, 1995) and the Role Making Model of Leadership (Graen, 1976). The two models

Page 13: Allan Lee PhD Research Proposal 2010

13have many similarities, mainly because the former developed from the Role-Making Model and has

become the most popular theoretical model to explain LMX development (Martin et al. 2010). The

Leadership-Making Model (also known as the Partnership-Building Model) is comprised of three

stages: the ‘stranger stage’, the ‘acquaintance stage’ and the ‘maturity stage’. The model proposes

that that the nature of the exchanges that take place between the leader and subordinate change

over time as the pair progress through the various stages. In the stranger stage the exchanges are

based upon the job description. From this point in time the leader or the subordinate may make an

offer for an improved working relationship through certain career-orientated social exchange (Graen

& Uhl-Bien, 1995). This initiation of social exchanges may cause the relationship to progress to the

acquaintance stage, where the nature of the leader-member relationship becomes defined through

mutual and reciprocal exchanges of resources. Interactions in the acquaintance stage may become

more regular and extend beyond those stipulated in the formal employment contract. The final

stage, the maturity stage, involves exchanges between the leader and subordinate that are

extremely well established and the relationship is characterised by mutual respect, trust and

obligation.

The Leadership Making model provides a broad explanation of how the LMX relationship

develops and emphasises that not all leader-member dyads progress through the process in a linear,

orderly fashion or in accordance with a uniform timetable. There is therefore a theoretical argument

to suggest there is variation in the way that LMX relationships develop. For example, the model

suggests that some relationships never progress beyond the stranger stage of development. For

those dyadic relationships that do progress, the stranger stage is thought to span the first few

months of the leader-member relationship, whilst the acquaintance stage may last as much as a year

or longer (Boyd & Taylor 1998). According to Graen and Uhl-Bien (1995) it is the acquaintance stage

of LMX development that is a critical stage in the development process since dyads that do not

progress beyond this stage will eventually return to the lower stage of LMX development.

Given the importance of the acquaintance stage to LMX development, it is vital that

research focuses on this stage and investigates the processes which will ultimately determines how

the relationship evolves. As mentioned earlier social exchanges are central to the model’s

explanation of how dyads progress through this stage. This phase of relationship building is

characterised as one of role-making and is described in terms of a sequence of exchanges between

leader and follower that are linked to task behaviour (Graen & Uhl-Bien, 1995). The proposed

research aims to examine established dyads at two time points over a 6 month period and thus

should enable a closer examination of LMX relationships during the acquaintance stage of

Page 14: Allan Lee PhD Research Proposal 2010

14development. Graen and Uhl-Bien (1995) suggest there is a limit to the time at which dyads may stay

within this developmental stage, therefore suggesting that even in established dyads there is

potential for change as a function of social exchanges. Such changeability of relationship quality is

proposed in the current research to be due to both the leaders and members perceptions of a

number of key variables, all of which are governed through social exchanges, which are essentially

the mechanism for change.

LMX Development: The Empirical Evidence

A plethora of studies have indentified variables that are concurrently associated with LMX

quality, however, the majority of this research is cross-sectional in nature, therefore no matter how

robust the correlations between such variables it does not permit causal inferences to be made.

Therefore, more prospective studies are needed to obtain a more comprehensive understanding of

which variables account for the development of the LMX relationship and how and under what

conditions this process occurs. Many authors have alluded to the value of exploring LMX

development further although only a few studies have attempted to answer this question by

adopting longitudinal designs, despite frequent calls within the literature. The prospective studies

conducted to date have identified a number of antecedents found to predict early LMX quality;

including leader and member expectations, similarity and liking (Liden et al. 1993); positive

affectivity, performance and delegation (Bauer & Green, 1996); role expectations (Major, Kozlowski,

Chao & Gardner,1995); member extroversion and leader agreeableness (Nahrgang et al. 2009).

These studies found that LMX relationships are established reasonably quickly, and can begin as

early as the recruitment and selection process (Parsons, Liden & Bauer 2001).

The aforementioned research suggests that early LMX quality is influenced by a number of

variables including salient characteristics such as personality similarity and liking. It has been

suggested, however, that these variables are significant in the initial leader-member interactions

whereas other variables play a larger role in the subsequent development of LMX (Nahrgang et al.

2009). For example, Bauer and Green (1996) noted that performance and delegation displaced

personality similarity after the initial inclusion phase as a predictor of LMX. Thus similarity may

predispose a leader to see a member in a positive light, giving the developmental process a nudge in

the right direction; however variables such as performance and delegation may be pivotal in the

subsequent development of the LMX relationship. This view is supported by a recent longitudinal

study conducted by Nahrgang et al. (2009). The study used growth curve modelling to examine the

trajectory of the development of the leader-member relationship over the first 8 weeks. It is one of

few studies that have tested the proposition that different variables are important for relationship

Page 15: Allan Lee PhD Research Proposal 2010

15quality at different stages of relationship development. The findings showed that member

extraversion and leader agreeableness influenced the ratings of LMX quality at the start of the

relationship, however after this early interaction, it was both leader and member performance

which influenced subsequent development.

It is important to note that one of the prerequisites for LMX development to occur is the

existence of a significant degree of instability in perceptions of LMX quality across time. In this

regard, the few longitudinal studies that have measured LMX (in)stability have revealed a reasonably

consistent pattern of results. For example, Bauer and Green (1996) report a correlation of .54

between the LMX relationship at the start of the study and 9 months later, whilst a correlation of .45

was found over a shorter period of 6 months (Liden et al. 1993). Nahrgang et al. (2009) examined the

first 8 weeks of the relationship and found LMX quality increased considerably over the first few

weeks and then became moderately stable. One notable feature of the literature is that such results

have frequently been interpreted as evidence for LMX development being confined to the very early

stages of the relationship (see Martin et al., 2010). In other words, LMX quality has been interpreted

as a highly stable construct that plateaus after the first few months of the relationship. This

conclusion is surprising considering when correlations of a similar magnitude are found in

prospective studies conducted within the area of Developmental Psychology and Relationship

Science they are interpreted as evidence for a significant degree of instability, and therefore they are

typically followed by empirical investigation of how other variables may account for the remaining

unexplained variance. For example, two papers in the domain of romantic relationship development

report similar correlations (between .40 and .60) between relationship quality at two points in time

(Fincham et al. 2000; Fincham et al. 1997). Both studies successfully explained a significant portion

of the unexplained variance that exists between relationship quality over time by investigating the

influence of variables such as depression and relationship attributions.

Applying this logic to the domain of LMX development it can argued that the existence of

moderate stability coefficients provides evidence for the fact that there is a significant amount of

relationship development that occurs beyond the early stages of the relationship life cycle. Indeed,

as previously discussed this alternative view is consistent with theories of LMX development (e.g.,

Leadership Making Model). Hence, in order to fill an important niche in the LMX literature, the

primary goal of my proposed PhD research is to explain the antecedents of LMX development,

particularly in the context of more established dyads. This aim will be achieved through the use of a

cross-lagged panel design and structural equation modelling (SEM). This design will allow for the

examination of a number of theoretically-guided variables over time and for longitudinal relations

Page 16: Allan Lee PhD Research Proposal 2010

16between these constructs to be explored whilst controlling for the stability of LMX. Such a

methodology will reveal whether these variables are correlated with changes in LMX quality over

time. This type of statistical technique is regularly applied within the developmental psychology

domain in order to tease out the unique influence and bidirectional interplay of key variables. In this

way, the current study will aim to explain significant amounts of unexplained variance that exists

between measurements of LMX over time. As the study will investigate influences on the changes of

both leader and member perceptions of LMX it is necessary to assess the variance in both these

measures longitudinally. Measurements will be taken from established dyads (with a dyadic tenure

of at least 6 months) at two time points with a six month lag in-between measurements. Established

dyads will be the focus of the current research for two reasons. First, prior research has largely

overlooked relationship development beyond the early stages of the relationship. Second, based on

the Leadership Making Model of LMX development, it is expected that considerable variation and

instability in LMX quality will be found in established dyads because they there is likely to be more

progressive and regressive transitions between the various stages of relationship development.

Hypothesis 1a: The quality of the Leader-Member relationship as measured from the member’s

perspective will be moderately stable over a 6 month period in established LMX dyads.

Hypothesis 1b: The quality of the Leader-Member relationship as measured from the leader’s

perspective will be moderately stable over a 6 month period in established dyads.

The Effect of Relationship Effort on LMX Development

The corollary of the above predictions is that there will be a substantial amount of variance

in LMX relationship quality and implies that this variance can be explained by variables above and

beyond prior levels of LMX quality. LMX theory emphasises social exchanges as key to the

development of the relationship. Underlying this perspective is the assumption that these exchanges

are based on effort exerted by both members of the dyad to the relationship (Maslyn & Uhl-Bien,

2001). The LMX theory would therefore suggest that the reciprocation of effort will influence the

development of leader-member relationships. However empirical investigation into the relative

effort that individuals put into relationships and the effect this has on their development is largely

absent from the literature (Maslyn & Uhl-Bien 2001). It is not known whether contributions to the

relationship need to be seen as equal or whether one party contributes more. It is also not clear

whether the balance of effort is perceived in the same way by both leaders and members.

Page 17: Allan Lee PhD Research Proposal 2010

17To date, the most informative study on the role of effort on LMX quality was conducted by

Maslyn & Uhl-Bien (2001). The study used questionnaires designed to directly measure retrospective

judgements of both self-effort and the effort that the other member of the dyad retrospectively put

into the relationship. Effort made by both dyad members was found to be related to current

measures of LMX quality, but that the perception of effort made by the dyad partner appeared to be

the critical factor associated with high-quality LMX relationships. The reported levels of self-effort

played a smaller role in light of the perceptions of the other member’s effort. Moreover effort made

by the dyad partner was positively related to reports of one’s own effort. The findings are consistent

with LMX theory in terms of reciprocity and the social exchange perspective in that the perception of

others effort was found to be highly significant in shaping LMX quality. The findings were consistent

across leader and member suggesting both parties are responsible for relationship development

through the operation of the social exchange process of perceived effort.

A major limitation of the above study was that it was cross sectional and therefore did not

allow for testing of the causal nature of effort over time. Perceptions of effort and LMX quality

appear to be strongly linked but the direction of effect cannot be determined (i.e. whether effort

influences the development of LMX quality or whether LMX quality influences the effort put in to the

relationship) As depicted in Figure 1, the use of cross lagged design in the current study will enable

the examination of these potential bidirectional effects that exists between self and other’s effort

and LMX quality in order to build upon the findings of Masyln & Uhl-Bien (2001).

Hypothesis 2a: Leader perceptions of high effort made by the member towards the dyadic

relationship will positively predict changes in leader ratings of LMX quality over time.

Hypothesis 2b: Member perceptions of high effort made by the leader towards the dyadic

relationship will positively predict changes in leader ratings of LMX quality over time.

Hypothesis 2c: Member ratings of high self effort will positively predict changes in member rating

of LMX quality over time.

Hypothesis 2d: Leader rating of high self effort will positively predict changes in leader rating of

LMX quality over time.

Page 18: Allan Lee PhD Research Proposal 2010

18Effort Differentiation and Social Comparison Theory

A concern regarding LMX theory is that the majority of research remains located at the

dyadic level. A consequence of such an approach is the assumption that individuals evaluate their

own LMX relationship in an absolute sense (Avolio, Walumbwa & Weber, 2009). Increasingly LMX

researchers are considering the potentially important role that other individuals such as co-workers

(Sherony & Green, 2006), who are in close proximity to the LMX relationship, may play in shaping

LMX development. Martin et al. (2010) highlight in their review the need for research to consider

dyadic and group level perspectives as complementary approaches to examine LMX processes. Social

Comparison theory (Festinger, 1954) suggests that individuals are highly sensitive to social

comparison information and will actively compare their exchanges with those of other team

members. A few studies have applied social comparison theory to LMX relationships. Hooper and

Martin (2006) found evidence that such differentiation of relationship quality awarded to different

members is observable to those outside of the focal dyad and this had implications for their own

LMX quality relationship. Furthermore, Sias and Jablin (1995) provided evidence that the favourable

treatment of others was often perceived as unfair by those with low LMX relationships, promoting

feelings of distrust and disliking. Typically research examining the role of social comparison

information has looked at the focal member’s perception of the different overall quality

relationships afforded to other members within the work group relative to their own, and has not

focussed specifically on the differentiation of relationship effort. In addition, research has focused on

member’s social comparisons of the leader’s effort and largely ignored the role of leader’s social

comparisons of member’s effort.

The proposed study aims to investigate the role that social comparisons made by both

leaders and members play in the development of LMX quality. As such, this research acknowledges

that LMX relationships do not develop in isolation and rather within the context of the wider group

in which the relationship is embedded and that both members and leaders may be influenced by

social comparative perceptions. Specifically this research will examine whether members

perceptions and therefore social comparisons of the amount of effort the leader puts into other LMX

relationships within the work group may have implications for their own LMX quality relationship.

Similarly it may be the case that leader’s perceptions of the effort of other members with whom

they have LMX relationships outside of the focal dyad may have implications for the quality of

relationship they hold with that particular member. It is expected that favourable comparisons

relating to perceptions of effort will play a role in either facilitating the development of the LMX

relationship. Unlike other LMX research which examines overall LMX quality differentiation the

Page 19: Allan Lee PhD Research Proposal 2010

19current research looks at an aspect of that differentiation that being the different levels of effort the

leader and other members put into their relationships. The use of cross lagged design will allow for

examination of how much of the variance in LMX quality over time can be accounted for by such

social comparisons.

Hypothesis 3a: Member’s favourable perception of leader effort exerted by leaders into their own

relationship, compared to leader’s effort in other LMX relationships, will be associated with an

increase in member’s LMX quality over time.

Hypothesis 3b: Leaders’ favourable perception of effort exerted by members within the focal

dyad, compared to effort exerted by member outside the dyad, will be associated with an increase

in leader’s LMX quality over time.

The Effect of Performance on LMX Development

Unlike relationship effort, the effect of performance on LMX development has been

empirically tested with prospective studies (e.g. Nahrgang et al, 2009; Bauer & Green, 1996)

demonstrating that it is an important variable in determining the nature of the leader-member

relationship and therefore should not be overlooked in the current research. However as with the

majority of LMX research these studies have mainly focused on the members perceptions of LMX

quality. The current study will aim to address the balance and investigate the reciprocal nature of

members and leaders performance and the influence this has on LMX development. This study is

again testing a key component of LMX theory, that being social exchange processes, with

performance constituting the currency of the exchange in this example.

Performance can be seen as an exchange behaviour that must be reciprocated in order for

the LMX relationship to develop. For example member’s performance may be reciprocated by the

delegation of tasks by the leader or vice-versa. Member performance may also be reciprocated by

leader performance (Bauer & Green, 1996). One prospective study which investigated the influence

of performance strongly suggests that it is an integral part of LMX development (Bauer & Green

1996). Furthermore in support of the LMX theories mentioned early it seems that performance

interacted with delegation as both where strongly correlated and predicted later LMX quality. The

findings showed that leader’s judgment about a member’s performance was a predictor of LMX

quality. In addition the increased responsibility and latitude the leader grants to the member

through delegation was strongly associated with better quality LMX relationships. These findings

point to the importance of reciprocity between leaders and members behaviours. However, the

study did not measure either leader performance or leader perceptions of LMX and we therefore do

Page 20: Allan Lee PhD Research Proposal 2010

20not have an understanding of whether there is reciprocity between leader and member

performance and whether this influences the leader’s perceptions of LMX quality, therefore the

current research would extend the findings of this influential piece of research and address aspects

of LMX theory not yet empirically explored.

There is empirical evidence that suggests leader performance can have an influence on LMX

relationship quality. Concurrent studies have shown that leader performance has a positive

association with both member and leader perceptions of LMX relationship quality (e.g Deluga, 1998;

Varma, Srinvas & Stroh, 2005). Nahrgang, Morgeson and Ilies (2009) conducted a prospective study

that examined the effect of leader and member performance. They found that for both leaders and

members, the performance of their dyadic partner was a key predictor of relationship quality. Whilst

this is an important finding, the duration of the study was relatively short (8 weeks) and it is

therefore not possible to draw any conclusions as to the effect that performance has over a longer

period of time. The current study will aim to build upon this research in order to investigate the role

that leader and member performance plays within established dyads where the most variability in

the stage of the LMX relationship is likely to be captured. The above study clearly demonstrates that

early performance influences early LMX development, but what is also of interest is whether later

changes in the level of performance will go on to influence later LMX development and account for

some of the variation in LMX over time. The use of cross-lagged designs and SEM will allow for the

unique effect of performance on the development of LMX to be determined as illustrated in the

conceptual model.

Hypothesis 4a: Leader performance will be positively related to changes in LMX relationship

quality from the member’s perspective.

Hypothesis 4b: Member performance will be positively related to changes in LMX relationship

quality from the leader’s perspective.

Prior Expectations as a Moderator of the Performance-LMX Development Link

Although prior longitudinal research suggests that performance has a low to moderate effect

on LMX quality; less is known about factors which might moderate this link. Given that cross-

sectional research has found good evidence for moderators (e.g., empowerment; perceived

organisational support) of the link between performance and LMX quality (see Martin et al., 2010), it

seems plausible that there will also be moderators of the longitudinal association between

performance and the development of LMX quality. One such variable that might act as such a

Page 21: Allan Lee PhD Research Proposal 2010

21moderator is prior expectations. For example, LMX quality may suffer if a member’s performance

does not live up to the leader’s expectations. The emphasis is that both leader and member

performance is relative to what is expected by the other member of the dyad. In this way

performance is not absolute but rather evaluated in the light of previous expectations. Individuals

often form expectations of targets based on available information (Liden et al. 1993). Information

available to leaders about members may come from secondary sources such as test scores,

recommendations or interviews (e.g. Phillips & Dipboye, 1989) or directly from the member during

the first few days on the job (Hollander & Offerman, 1990). Similarly member’s information relevant

to the leader is likely to come from co-workers, interviews or directly during the early days of

working in the organisation (Fisher, 1986; Jablin, 1987). Although expectations have not been

investigated as a moderator between performance and LMX, research has shown that they are

important in the development of the leader-member relationship. Positive leader expectations of a

member appear to be translated into such leader behaviours as delegating challenges task

assignments (Graen, Orris & Johnson, 1973; Leana, 1986) and providing constructive feedback,

desirable awards and training (Feldman, 1986). Major, Kozlowski, Chao and Gardner (1995)

investigated the expectations that newly hired members brought into the organisation when they

began work. The authors measured the extent to which these various expectations were met after

the newcomers had begun their jobs and found that the degree to which expectations, related to

role clarity and personal acceptance, were met was associated with current LMX quality.

Furthermore, Liden et al. (1993) found that leader and member expectations for each other

measured in the first 5 days in the life of the dyad predicted LMX at both 2 weeks and at 6 weeks.

The previously mentioned research demonstrates that expectations made by both the

leader and member are related to LMX quality. The limited prospective work has largely focused

upon previously held role expectations and their influence on early socialisation in initial stages of

LMX. This research is the first of its kind to explore expectations in relation to a particular variable,

namely performance, with the intention of detailing the potentially important moderating role that

expectations held throughout the development of the relationship play in shaping LMX quality. The

current study aims to investigate whether the specific expectations of both leaders and members

relevant to the performance of the other member of the dyad will moderate the performance-LMX

link.

Hypothesis 5a: Leader’s prior expectations will moderate the effect that performance has on

leader LMX quality such that performance and changes in LMX quality will be positively related

when leader expectations are exceeded as opposed to when they are unmet.

Page 22: Allan Lee PhD Research Proposal 2010

22Hypothesis 5b: Member’s prior expectations will moderate the effect that leader performance has

on member LMX quality such that performance and changes in LMX quality will be positively

related when member expectations are exceeded as opposed to when they are unmet.

Attribution Theory and LMX Development : Direct and Moderating Effects

Relationship-specific Attributions: Direct effect on LMX Development. The theory of Attributions is

a stalwart within Social Psychology and has increasingly been used to explain the development and

maintenance of romantic relationships. Attributions involve observing the behaviour of others and

seeing stable characteristics in the individuals we observe. In this way dispositions or traits are

perceived as being responsible for people’s behaviour. Kelley (1967) suggests that by observing an

individual’s behaviour in a variety of situations, we infer that the behaviour has an internal or an

external cause (something within or outside of the control of the individual). Whilst attributions

focus on causal explanations for single events (i.e. a task that a member has performed well on),

relationship-specific attributions refer to individuals tendencies to make the same attributions in

response to incidents/behaviours in a particular relationship. Numerous studies have documented

the robust associations between causal and responsibility attributions and relationship satisfaction

(reviewed by Banbury and Fincham, 1990). Prospective studies within other relevant domains have

directly examined the impact of attributions on relationship satisfaction. One such study found that

wives’ initial causal and responsibility attributions predicted later satisfaction (for a review see

Fincham & Bradbury, 1987). Far less research has examined attributions made in the context of the

leader-member relationship and the effect they have on LMX development. As there is usually a

large amount of interaction between a leader and a member in the work setting there is ample

opportunity for each to make attributions about each other. The present research would aim to

investigate whether relationship-specific attributions will significantly predict LMX quality.

A number of authors have recognised that attributions may play a key role in the leader-

member relationship, with Green and Mitchell (1979, cited in Dasborough & Ashanasy, 2002) being

the first to suggest the link. According to the leadership making model of LMX the influence of

attributions is expected to be most influential during the acquaintance stage (Boyd & Taylor 1998).

In the earlier stages of relationship development, attributions may be influenced by upward

impression management on the part of the member. However if the relationship progresses then

the increased communication and reduced barriers to communication that are theorised to occur

between the leader and member may reduce or alter the attributional bias as both dyad members

gain a better understanding of each other’s point of view (Martinko & Gardner, 1987). Conversely if

Page 23: Allan Lee PhD Research Proposal 2010

23the LMX relationship fails to progress beyond the first stage, the attributional bias may be

reinforced. Dienesch and Liden (1986) stressed that attributions and categorisations serve as critical

inputs to the development of the relationship between leaders and those who follow them. They

proposed a model where attributions influence LMX development via one of two pathways. The first

involves salient behaviours that occur during the initial interaction between a leader and a member.

These behaviours may lead to early categorisation which will have an immediate influence on the

nature of the leader-member exchanges. The second and more likely pathway does not involve

immediate categorisation; instead attributions are made after the initial delegation of

responsibilities made by the leader. Members make attributions about the leaders attitudes and

behaviour based upon the nature of these initial assignments. According to the model the attributed

characteristics go on to influence how members perform for the leader and determine whether they

will respond with loyalty and high performance. Conversely the leader will make attributions related

to member’s performance on these initial assignments and will then influence assignments of

further responsibilities to the member. Whilst the aforementioned model is intuitive the role that

attributions play in the development of the leader-member relationship does not have empirical

support due to the lack of prospective studies testing the influence they have on LMX development.

There is some empirical research to support the theory that attributions are related to the

LMX relationship; for example one cross-sectional study found that leaders tend to make more

favourable attributions to their in-group members (Heneman, Greenberger & Anonyuo, 1989).

Kinicki & Vecchio (1994) reported that members who displayed a tendency to make internal

attributions for their own behaviour were more likely to have a high LMX relationship with their

leader. A particular relevant study, conducted by Martinko, Moss, Douglas and Borkowski (2007)

investigated attribution styles and the effect that these have on LMX quality. Attribution Styles are

similar to relationship-specific attributions as they suggest that individuals will explain different

events in the same way; the difference being that this tendency extends beyond a specific

relationship to all contexts (e.g. Abramson, Seligman & Teasdale, 1978). Martinko et al. (2007) found

that that leaders’ and members attribution styles had interactive effects on members’ perceptions of

LMX quality. The results of the study showed that members’ perceptions of poor LMX relationships

were accentuated when they had an optimistic (positive) attribution style, whilst their leader had a

pessimistic (negative) style. This study is particularly interesting and is the first to examine

attribution styles in the domain of LMX. The current study will aim to build upon prior theory and

research by determining whether the positivity of relationship-specific attributions influences the

development of the LMX relationship.

Page 24: Allan Lee PhD Research Proposal 2010

24An important concern with the suggested link between attributions and LMX quality is the

fact that the link may be bi-directional. Thus rather than attributions having a causal effect upon the

subsequent development of LMX, the quality of the LMX relationship may in fact influence the

attributions made by one member of the dyad about the other. The few studies that have

investigated the link between LMX and attributions have not made use of appropriate longitudinal

methodologies in order to determine the direction of effects between the two variables. The use of

cross-lagged design will also allow for examination of the direction of the effects and whether

attributions influence LMX or whether LMX influences attributions.

Hypothesis 6a: Positive Relationship-specific attributions made by leaders will positively influence

changes in Leader LMX quality.

Hypothesis 6b: Positive Relationship-specific attributions made by members will positively

influence changes in Member LMX quality

Incident-specific Attributions: A Moderating Effect. The preceding hypotheses predicted that the

positivity of attributions made by each member of the dyad will positively influence the

development of the leader-member relationship. However the effect of attributions can also be

investigated at a lower level, in terms of the incident specific attribution made by individuals. For

example if leaders make internal attributions about member’s good performance and external

attributions about poor performance, the development of high quality leader-member relationship is

likely (Mitchell, Green & Wood., 1981; Steiner & Dobbins, 1989). In this way the causal attributions

made relative to the performance on a particular task will moderate the influence that performance

has on LMX quality.

A few studies have investigated attributions made in response to specific incidents. A

particularly interesting study had nursing supervisors read incidents of subordinate poor

performance. The incident reports were designed to produce internal or external attributions. The

attributions made on the basis of these reports were associated with the action that the supervisors

took in dealing with the performance issues. Internal attributions were associated with supervisor’s

indicating that action should be taken in response. These actions ranged from counselling to

termination of contracts (Mitchell & Wood 1980). Other studies report a similar pattern, whereby

internal attributions are associated with supervisors taking stricter action towards member’s whose

performance was poor (Trahan & Steiner, 1994; Wood & Mitchell, 1981).

Page 25: Allan Lee PhD Research Proposal 2010

25The current study would aim to test a number of interaction hypotheses, whereby the causal

attributions made by both leaders and members related to the performance and relationship effort

exerted by the other member of the dyad will moderate the influence these variable have on later

LMX quality. For example if member’s high relationship effort is attributed to an internal cause by

the leader it will have a positive effect on the leader’s perception of the LMX relationship. However if

internal attributions are made relating to low relationship effort then this will have a negative effect

on LMX perceptions.

Hypothesis 7a: Leader attributions concerning member’s performance will moderate the influence

of this variable on Leader LMX quality. Examples of this interaction between performance and

LMX quality:

High Performance x Internal Attributions = Positive effect on changes in LMX Quality

Low Performance x Internal Attributions = Negative effect on changes in LMX Quality

Hypothesis 7b: Leader attributions concerning member’s effort will moderate the influence of this

variable on Leader LMX quality.

High Relationship Effort x Internal Attributions = Positive effect on changes in LMX

Quality

Low Relationship Effort x Internal Attributions = Negative effect on changes in LMX

Quality

Hypothesis 7c: Member attributions concerning leader’s performance will moderate the influence

of this variable on Member LMX quality.

High Performance x Internal Attributions = Positive effect on changes in LMX Quality

Low Performance x Internal Attributions = Negative effect on changes in LMX Quality

Hypothesis 7d: Member attributions made by the member concerning leader’s effort will

moderate the influence of this variable on Member LMX quality.

High Relationship Effort x Internal Attributions = Positive effect on changes in LMX

Quality

Low Relationship Effort x Internal Attributions = Negative effect on changes in LMX

Quality

Page 26: Allan Lee PhD Research Proposal 2010

26The Consequences of LMX Quality

One of the main reasons why LMX has attracted so much attention and research over the

years is that high quality relationships have frequently been associated with a number of positive

outcomes. A meta-analysis conducted by Gerstner & Day (1997) showed that LMX quality (perceived

by the member) was related to supervisor rating of performance, objective performance,

satisfaction with the leader, overall satisfaction, commitment, role conflict, role clarity and turnover

intentions. Variables like these are clearly highly beneficial to both the individuals involved and the

organisation which they work for. The fact that LMX is associated with numerous important

outcomes clearly justifies the need to further understand how the relationship evolves and the

mechanisms through which high-quality relationship are achieved. In doing so leaders and members

can be equipped with the power to have greater influence over their relationship and shape a more

positive working situation for themselves and the organisation.

The proposed study will be able to gain an understanding of the unique influence that LMX

quality has on performance, effort, attributions and expectations. This type of methodological design

suggested in the present research proposal is all but absent from the LMX literature, however two

studies that have been discussed previously have examined changes in both LMX and performance

over time (Bauer & Green, 1996; Nahrgang et al. 2009). Bauer and Green (1996) found that change

in LMX quality over a six month period significantly predicted performance ratings.

The suggestion from LMX theorists is that LMX quality should have a positive effect on

numerous outcomes. High LMX quality is associated with increased communication and with

exchanges that go beyond the formal job contract where the aim is to develop the individual’s ability

to perform on the job. Furthermore Graen and Uhl-Bien (1995) proposed that high quality LMX

relationships consist of a number of underlying dimensions; namely mutual respect, trust and

obligation. Similarly Liden and Maslyn (1998) posit that such relationships are characterised by affect

loyalty, contribution and professional respect. These type of attitudinal and behaviour factors are all

likely to have a positive impact on work-related outcomes. For example increased communication is

likely to reduce task ambiguity and increase clarity in terms of expectations, which in turn would

have a positive effect on task performance. It is also likely that a number of attributional biases may

be evident as a result of having a high-quality LMX relationship. Supervisors, for example, may be

more likely to attribute the poor performance of a member to an external cause when their

relationship is characterised by high affect (high-LMX) in order to maintain a positive view of the

person (see Maio & Thomas, 2007). Based on previous research and the theoretical suggestions it

Page 27: Allan Lee PhD Research Proposal 2010

27can be predicted that LMX quality will have a positive impact on the variable measured in the

current study.

Hypothesis 8a: LMX quality measured from the members perspective will positively predict

changes in performance, effort exerted, attributions and expectations of members over time.

Hypothesis 8b: LMX quality measured from the leaders perspective will positively predict changes

in performance, effort exerted, attributions and expectations of leaders over time.

Proposed Methodology

The proposed research is looking to empirically investigate the development of the leader-

member exchange relationship. A longitudinal approach will be adopted with measurements taken

at two time points over a period of 6 months. A cross-lagged panel design will be used to examine

the causal nature of key variables on LMX quality and account for the unexplained variance in the

stability of LMX quality that has been found in previous research. Furthermore, when considering

LMX quality, both leader and member perspectives on the relationship will be measured.

Sample

An organisation sample will be required with approximately 250 leader-member dyads. This

relatively large sample is needed in order to increase the predictive power of the analysis. In order to

capture most variance in the LMX relationships that exist within the sample, established dyads (at

least 6 months dyadic tenure) will be used.

Measures

Leader-member exchange (LMX)

The quality of the leader–member relationship will be assessed using the LMX-7 scale (Graen & Uhl-

Bien, 1995). Reviews of the literature suggest that this measure offers the most appropriate measure

of dyadic relationship quality (Gerstner & Day, 1997; Graen & Uhl-Bien, 1995). This measure consists

of seven items with single responses recorded on a 7-point likert scale, ranging from 1 (strongly

disagree) to 7 (strongly agree). The items will be adapted for use with both leaders and members

and each will rate the quality of the relationship at both time points. Example items for leaders will

include ‘‘I recognize this team member’s potential” and ‘‘He/she would characterize our working

Page 28: Allan Lee PhD Research Proposal 2010

28relationship as effective.” Example items for members will include ‘‘This team leader recognizes my

potential” and ‘‘This team leader and I have an extremely effective working relationship.”

Leader and member performance

Member performance will be assessed by the leader using four items from Liden et al.’s (1993)

performance measure. Example items include ‘‘The overall level of performance that I have

observed for this team member is outstanding” and ‘‘My personal view of this team member is that

he or she is very effective.” Leader performance will be rated by the member using an adapted

version of the same 4 items; an example item being: ‘‘This team leader’s performance is very high”.

Relationship Effort

Perceptions of self and others effort will be measured using a scale developed by Maslyn and Uhl-

Bien (2001). This will require a response to two questions; "How much effort have you put into

developing a good relationship with your manager [subordinate]?" and "How much effort has your

manager [subordinate] put into developing a good relationship with you?" Both items used a 5-point

response format ranging from 1 (none at all) to 5 (a great deal).

Effort Differentiation (Relative Effort)

I will adapt the Hooper & Martin (2006) measure of LMX differentiation to measure relative

relationship effort. Relative effort exerted by the member into the LMX relationship will be assessed

from the leader’s perspective. The above measure will be adapted to measure leaders perceptions

of the effort put into the relationship by the member relative to other members under the leaders

span of supervision. Leaders will be asked to rate the amount of effort each of their team members

have put into developing a good relationship. Specifically, they will be required to indicate the

number of people in their team whose effort level can be described as either: “very poor” (1), “poor”

(2), “satisfactory” (3), “good” (4) or “very good” (5). The relative effort exerted by the focal member

will be measured by subtracting member’s individual-level effort score within a group from each

group member’s composite effort score. Using the same scale, the relative effort that leaders put

into relationship with other members will be measured by subtracting member’s perceptions of

leader’s effort score from effort scores with other members in the work group.

Relationship-Specific Attributions

Page 29: Allan Lee PhD Research Proposal 2010

29Relationship-specific attributions will be measured using the Relationship Attribution Measure (RAM;

Fincham & Banbury, 1992); adapted for use in the context of the leader-member relationship. Both

leaders and members will rate how much they agree with four statements. An example item is: ‘Your

boss does not appreciate the work that you do’, respondents will rate their agreement with the

statement on a 6-point likert scale ranging from 1 (disagree strongly) to 6 (agree strongly).

Incident-Specific Attributions

Leaders and members will be asked to think of occasions where the other has shown particularly

high/low performance and incidents where the other has exerted little/high relationship effort.

Individuals will then be asked to rate the cause of these events using a scale adapted from the

Causal Attributtion Index (RAM-C, Fincham & Bradbury, 1992).

Performance Expectations

A measure of performance expectations will be made based upon items adapted from Liden et al.

(1993) in which six items will be used to assess the subordinates expectations of the new supervisor

and visa versa. An example item is: ‘I think that my supervisor is a high performer’. Responses to

items will be captured on a 7-point likert scale ranging from 1 = strongly disagree to 7= strongly

agree.

Analysis

The proposed research will take a quantitative approach in the statistical analysis. Cross-

lagged stability models allow examination of longitudinal relations between constructs whilst

controlling for their stability. Significant cross-lagged effects reflect the presence of a relationship

beyond that which can be accounted for by the stability of the construct and the magnitude of their

association at time 1. A common limitation of longitudinal research is that they typically use

correlations as estimates for examining relationships but the predictor variable often correlates

cross-sectionally with the predicted variable and failure to control for this association precludes the

study of change. Therefore the current research will adopt the statistical technique of Structural

Equation Modelling (SEM) as it allows for the simultaneous estimation of all parameters in the causal

system, something which other methods such as correlation or multiple regression cannot achieve

(Fincham et al, 1997).

Page 30: Allan Lee PhD Research Proposal 2010

30Reducing the Threat of Common Method Variance

A correlation between two variables can be inflated if both are obtained from the same

person at the same time using the same data collection technique. Such common method variance is

cited as frequent limitation of LMX research. In the present proposal a number of features within the

research design have been included in order to reduce the potential threat to the validity of the

results as a function of common method variance. Firstly, through using a longitudinal design the

data collection of measures is separated across time points. Secondly the data will be collected from

both the leader and member of the dyad therefore reducing the concern associated with common

method bias.

Brief Overview of Subsequent PhD Research Studies

The current proposal represents a shift in research focus to how the leader-member

relationship develops; a direction highlighted as necessary by Graen and Uhl-Bien (1995).

Subsequent PhD studies would be needed to build on this platform and further investigate LMX

relationship development. Although the proposed two wave longitudinal study is appropriate for

measuring changes in LMX quality, and represents a valuable starting point for examining LMX

development, , a subsequent study could employ an even stronger design in order to replicate and

extend this research by using a 3 wave longitudinal design. The additional benefits of measuring LMX

development over 3 waves are threefold. First, it represents a stronger design for measuring

changes in LMX quality over more measurement points. Second, it allows for a stronger test of the

reciprocity of social exchanges over time and their effect on subsequent LMX development. Finally, it

allows for the addressing more interesting theoretical questions such as the testing of time-ordered

meditational mechanisms that help explain the process of how antecedents lead to changes in LMX

quality over time. For example, Time 1 antecedents may predict time 2 mediators which in turn

predict time 3 LMX quality, controlling for initial levels of LMX quality.

In addition, potential further research directions could include research which goes beyond

survey methods and adopts well-established social psychological methods such as diary research

(e.g.Bolger, Davis & Rafaeli, 2003) in order to capture subordinates perceptions of their own and

their leader’s cognition, emotion and behaviour as it unfolds during the development of LMX.

Fairhurst (1993) calls for more longitudinal research with methods which record the pattern of

interaction over time in detail and probe into each party’s changing perceptions to resolve

inconsistencies. The introduction of diary methodology to be used over a period of time may offer

some fine-grained analyses of the way in which this dynamic interactive relationship develops (e.g.

Page 31: Allan Lee PhD Research Proposal 2010

31Bolger et al, 2003). Such methodology provides the unique advantage of measuring over time

(typically daily) to test what types of perceptions, cognitions, emotions and behaviours predict LMX

development at the formative stage of the leader-member relationship. Diary methodology also

allows the collection of both qualitative and quantitative data, because open ended descriptions can

be qualitatively coded and then combined to form quantitative measures of member’s social

cognition and emotion. Such measures can then be used for inferential statistical analysis; therefore

rich detail following a critical incident may be obtained but can then be statistically analysed. Given

that LMX is a relational construct a more diverse approach to methodology may be appropriate in

order to potentially capture more complex aspects of the concept (Martin et al, 2010). Therefore,

such a mixed methods approach would be well suited and could provide a rich data source to

complement the questionnaire based empirical approach currently prevailing in this domain.

Timeline for the Completion of the Proposed Research

Months 1-12: Literature review and further development of research idea

Identify and contact potential organisations for data collection

Complete quarterly reports

Months 13-24: Data Collection and Analyses

Prepare measures and materials for study

Collect data from 1st study – 2 waves over 6 months

Analyse results of 1st study

Collect data from 2nd study – diary study over 8 weeks

Analyse results of 2nd study

Months 25-36: Data Collection, Analysis and Write-Up

Collect data from 3rd study – 3 waves over 6 months

Analyse results of 3rd study

Write up findings of study 1 -3 and complete first draft of thesis

Make changes and complete final write-up of thesis.

Page 32: Allan Lee PhD Research Proposal 2010

32References

Abramson, L. Y., Seligman, H. E. P., & Teasdale, J. D. (1978). Learned helplessness in humans: Critique and reformulation. Journal of Abnormal Psychology,

Avolio, B. J., Walumbwa, F. O., & Weber, T. J. (2009). Leadership: Current Theories, Research, and Future Directions. Annual Review of Psychology, 60(1), 421-449.

Bauer, T. N., & Green, S. G. (1996). Development of leader-member exchange: A longitudinal test. Academy of Management Journal, 39(6), 1538.

Bernerth, J. B., Armenakis, A. A., Feild, H. S., Giles, W. F., & Walker, H. J. (2007). Is personality associated with perceptions of LMX? An empirical study. Leadership & Organization Development Journal, 28(7), 613-631.

Blau, P. M. (1964). Exchange and power in social life. New York: Wiley

Bolger, N., Davis, A., & Rafaeli, E. (2003) Diary methods: Capturing life as it is lived. Annual Review of Psychology, 54, 579-616

Boyd, N. G., & Taylor, R. R. (1998). A developmental approach to the examination of friendship in leader-follower relationships. The Leadership Quarterly, 9(1), 1-25.

Bradbury, T.N. & Fincham, F. D. (1990). Attributions in marriage: Review and critique. Psychological Bulletin, 107(1), 3-33.

Brower, H. H., Schoorman, F. D., & Hwee Hoon, T. (2000). A model of relational leadership: the integration of trust and leader-member exchange. Leadership Quarterly, 11(2), 227-250.

Dansereau, F., Graen, G.B., & Haga, W. (1975). A vertical dyad linkage approach to leadership in formal organizations. Organizational Behavior and Human Performance, 13, 46-78.

Dasborough, M. T., & Ashkanasy, N. M. (2002). Emotion and attribution of intentionality in leader-member relationships. The Leadership Quarterly, 13(5), 615-634.

Dienesch, R.M. and Liden, R.C. (1986). Leader-member exchange model of leadership: A critique and further development. Academy of Management Review, 11, 618-634.

Deluga, R. J. (1998). Leader-Member Exchange Quality and Effectiveness Ratings. Group & Organization Management, 23(2), 189-216.

Engle, E. M., & Lord, R. G. (1997). Implicit Theories, Self-Schemas, And Leader-Member Exchange. Academy of Management Journal, 40(4), 988-1010.

Epitropaki, O., & Martin, R. (2005). From ideal to real: A longitudinal study of the role of implicit leadership theories on leader-member exchanges and employee outcomes. Journal of Applied Psychology, 90, 659−676.

Erdogan, B., & Liden, R.C. (2002). Social exchanges in the workplace: A review of recent developments and future research directions in Leader-Member Exchange theory. In L.L. Neider & C.A. Schriescheim (eds.), Leadership (p. 65-114). Greenwich, CT: Information Age Publishing.

Page 33: Allan Lee PhD Research Proposal 2010

33Fairhurst, G. T. (1993). The leader-member exchange patterns of women leaders in industry: A discourse analysis. Communication Monograph, 60, 321-351.

Feldman, J. M. (1986). A note on the statistical correction of halo error. Journal of Applied Psychology, 71, 173–176.

Festinger, L. (1954). A theory of social comparison processes. Human Relations, 7, 117−140.

Fincham, F. D., Beach, S. R. H., Harold, G. T., & Osborne, L. N. (1997). Marital satisfaction and depression: Different causal relationships for men and women? Psychological Science, 8, 351-357.

Fincham, F. D., Harold, G. T., & Gano-Phillips, S. (2000). The longitudinal association between attributions and marital satisfaction: Direction of effects and role of efficacy expectations. Journal of Family Psychology, 14(2), 267-285.

Fincham, F.D., & Bradbury, T.N. (1987). The impact of attributions in marriage: A longitudinalanalysis. Journal of Personality and Social Psychology, 53, 510-517.

Fincham, F.D., & Bradbury, T.N. (1992). Assessing attributions in marriage: The Relationship Attribution Measure. Journal of Personality and Social Psychology, 62, 457-468

Fisher, C. D. (1986). Organizational socialization: An integrative review. Research in Personnel and Human Resources Management, 4, 101–145.

Gerstner, C. R., & Day, D. V. (1997). Meta-Analytic Review of Leader-Member Exchange Theory: Correlates and Construct Issues. Journal of Applied Psychology, 82(6), 827-844.

Graen, G. B. (1976). Role-making processes within complex organizations. In M. D. Dunnette (ed.), Handbook of Industrial and Organizational Psychology. Chicago: Rand McNally, 1201-1245.

Graen, G. B., & Cashman, J.F. (1975). A role-making model of leadership in formal organisations: a developmental approach. In J.G. Hunt & L.L Larson (eds.), Leadership frontiers (pp. 143-165). Kent, Ohio: Kent State Univ. Press.

Graen, G. B., & Scandura, T. (1987). Toward a psychology of dyadic organizing. In B. Staw & L.L Cummings (eds.), Research in organizational behavior (pp. 175-208). Greenwich, CT: JAI Press.

Graen. G.B., & Uhl-Bien, M. (1991). Partnership-making applies equally well to team mate-sponsor team mate-competence network, and team mate-team mate relationships. Journal of Management Systems, 3, 33-48.

Graen. G.B., & Uhl-Bien, M. (1995). Relationship-based approach to leadership: Development of leader-member exchange (LMX) theory of leadership over 25 years: Applying a multi-level multi-domain perspective. Leadership Quarterly, 6, 219-247.

Graen, G. B., & Uhl-Bien, M. (1997). Relationship-based approach to leadership: Development of Leader Member Exchange (LMX) Theory of leadership over 25 years: Applying a multi-level multi-domain perspective. Journal of Applied Psychology, 82, 6: 927-947.

Graen, G. B., Orris, J. B., & Johnson, T. W. (1973). Role assimilation processes in a complex organization. Journal of Vocational Behavior, 3, 395–420.

Heneman, R. L., Greenberger, D. B., & Anonyo, C. 1989. Attributions and exchanges: The effects of interpersonal factors on the diagnosis of employee performance, Academy of Management journal, 32: 466-476.

Page 34: Allan Lee PhD Research Proposal 2010

34Henderson, D., Liden, R., Glibkowski, B., & Chaudhry, A. (2009). LMX differentiation: A multilevel review and examination of its antecedents and outcomes. Leadership Quarterly, 20(4), 517.

Hogan, R Kaiser, R. B. (2005) "What we know about leadership", Review of General Psychology, Vol. 9 pp.169 - 180.

Hogg, M. A., Martin, R., Epitropaki, O., Mankad, A., Svensson, A., & Weeden, K. (2005). Effective Leadership in Salient Groups: Revisiting Leader-Member Exchange Theory From the Perspective of the Social Identity Theory of Leadership. Personality and Social Psychology Bulletin, 31(7), 991-1004.

Hollander, E. P., & Offermann, L. R. (1990). Power and leadership in organizations: Relationships in transition. American Psychologist, 45, 179–189.

Hooper, D., & Martin, R. (2006). Beyond personal LMX quality: The effects of perceived LMX variability on employee reactions. In A. I. Glendon B. Myors & B. M. Thompson (eds.), Advances in organizational psychology: An Asia-Pacific perspective. Brisbane: Australian Academic Press.

Howell, J.M., & Hall-Merenda, K.E. (1999). The ties that bind: The impact of Leader-Member Exchange, transformational and transactional leadership, and distance on predicting follower performance. Journal of Applied Psychology, 84, 680-694.

Jablin, F. M. (1987). Organizational entry, assimilation, and exit. In F. M. Jablin, L. L. Putnam, K. H. Roberts, & L. W. Porter (Eds.), Handbook of organizational communication (pp. 679–740). Newbury Park, CA: Sage.

Kamdar, D., & Van Dyne, L. (2007). The Joint Effects of Personality and Workplace Social Exchange Relationships in Predicting Task Performance and Citizenship Performance. Journal of Applied Psychology, 92(5), 1286-1298.

Kelley, H. H. (1967). Attribution theory in social psychology. In D. Levine (ed.), Nebraska Symposium on Motivation (Volume 15, pp. 192-238). Lincoln: University of Nebraska Press.

Kinicki, A. J., & Vecchio, R. P. (1994). Influences on the quality of supervisor-subordinate relations: The role of time-pressure, organizational commitment, and locus of control. Journal of Organizational Behavior, 15, 75−82.

Klein, H. J., & Kim, J. S. (1998a). A field study of the influence of situational constraints, leader-member exchange, and goal commitment on performance. Academy of Management Journal, 41(1), 88-95.

Leana, C. R. (1986). Predictors and consequences of delegation. Academy of Management Journal, 29, 754–774.

Liden, R.C., Wayne, S.J., & Stilwell, D. (1993). A longitudinal study on the early development of leader-member exchanges. Journal of Applied Psychology, 78, 662-674.

Liden, R.C., & Graen, G. (1980). Generalizability of the Vertical Dyad Linkage model of Leadership. Academy of Management Journal, 23, 451-465.

Page 35: Allan Lee PhD Research Proposal 2010

35Liden R. C., & Maslyn, J. M. (1998). Multidimensionality of Leader-Member Exchange: An empirical assessment through scale development. Journal of Management, 24, 43-72.

Liden, R. C., Sparrowe, R. T., & Wayne, S. J. (1997). Leader-member exchange theory: The past and potential for the future Research in personnel and human resources management, Vol. 15. (pp. 47-119): US: Elsevier Science/JAI Press.

Liden, R. C., Erdogan, B., Wayne, S. J., & Sparrowe, R. T. (2006). Leader-member exchange, differentiation, and task interdependence: implications for individual and group performance. Journal of Organizational Behavior, 27(6), 723-746.

Maio, G.M., & Thomas, G. (2007). The epistemic-teleologic model of deliberate self persuasion. Personality and Social Psychology Review, 11, 1-22.

Martin, R., Epitropaki, O., Thomas, G., & Topaka, A. (2010). A critical review of Leader-Member Relationship (LMX) research: Future prospects and directions.

Martin, R., Thomas, G., Charles, K., Epitropaki, O., & McNamara, R. (2005). The role of leader-member exchanges in mediating the relationship between locus of control and work reactions. Journal of Occupational & Organizational Psychology, 78(1), 141-147.

Martinko, M. J., & Gardner, W. L. (1987). The leader member attribution process. Academy ofManagement Review, 12, 235-249.

Martinko, M. J., Moss, S. E., and Douglas, S. C., & Borkowski, N. (2007). Anticipating the Inevitable: When Leader and Member Attribition Styles Clash. Organizational Behavior and Human Decision Processes: 102: 158-174

Maslyn, J. M., & Uhl-Bien, M. (2001). Leader-Member Exchange and Its Dimension: Effects of Self-Effort and Other's Effort on Relationship Quality. Journal of Applied Psychology, 86(4), 697-708.

Major, D.A., Kozlowski, S.W.J., Chao, G.T., & Gardner, P. D. (1995). A longitudinal investigation of newcomer expectations, early socialization outcomes, and the moderating effects of role development factors. Journal of Applied Psychology, 80, 418-431.

Mitchell, T.R., & Wood, R.E. (1980). Supervisor’s responses to subordinate poor performance: A testof an attributional model. Organizational Behavior and Human Performance, 25, 123–138.

Mitchell, T.R., Green, S.G., Wood, R.E. (1981), "An attributional model of leadership and the poor performing subordinate: development and validation", in Cummings, L.L., Staw, B.M. (Eds),Research in Organizational Behavior, Vol. 3, JAI Press, Greenwich, CT, pp.197-234.

Nahrgang, J., Morgeson, F., & Ilies, R. (2009). The development of leader-member exchanges: Exploring how personality and performance influence leader and member relationships over time. Organizational Behavior and Human Decision Processes, 108(2), 256.

Parsons, C.K., Liden, R.C., & Bauer, T.N. (2001). Employment interviews. In M. London (Ed.), How people evaluate others in organizations (pp. 67-90). Mahwah, NJ: Lawrence Erlbaum.

Phillips, A. P., & Dipboye, R. L. (1989). Correlational tests of predictions from a process model of the interview. Journal of Applied Psychology, 74, 41–52.

Page 36: Allan Lee PhD Research Proposal 2010

36Phillips, A.S., & Bedeian, A.G. (1994). Leader-Follower Exchange Quality: The role of personal and interpersonal attributes. Academy of Management Journal, 37, 990-1001.

Rousseau, D. M. (1998). LMX meets the psychological contract: Looking inside the black box of leader–member exchange. In F. Dansereau & F. J. Yammarino (Eds.), Leadership: The multilevel approaches (Vol. B, pp. 149–154). Greenwich, CT: JAI Press.

Sias, P. M., & Jablin, F. M. (1995). Differential superior-subordinate relations, perceptions of fairness, and coworker communication. Human Communication Research, 22, 5−37.

Scandura, T. A., & Schriesheim, C. A. (1994). Leader-member exchange and supervisor career mentoring as complementary concepts in leadership research. Academy of Management Journal, 37, 1588-1602.

Schriesheim, C. A., Castro, S. L., & Cogliser, C. C. (1999). Leader-member exchange (LMX) research: A comprehensive review of theory, measurement, and data-analytic practices. Leadership Quarterly, 10(1), 63.

Sin, H. P., Nahrgang, J. D. & Morgeson, F. P. (2009) Understanding Why They Don’t See Eye to Eye: An Examination of Leader–Member Exchange (LMX) Agreement. Journal of Applied Psychology, 94: 4, 1048- 1107.

Sherony, K. M., & Green, S. G. (2006). Coworker exchange: Relationships between coworkers, leader-member exchange, and work attitudes. Journal of Applied Psychology, 87, 542−548.

Steiner, D.D., & Dobbins, G. H. (1989). The role of work values in leaders’ attributions and the development of leader–member exchanges. International Journal of Management, 6, 81–90.

Trahan, W.A., & Steiner, D.D. (1994). Factors affecting supervisors’ use of disciplinary actions following poor performance. Journal of Organizational Behavior, 15, 129–139.

Varma, A., Srinivas, E. S., & Stroh, L. K. (2005). A Comparative Study of the Impact of Leader-Member Exchange in US and Indian Samples. Cross Cultural Management, 12(1), 84-95.

Wood, R.E., & Mitchell, T.R. (1981). Manager behavior in social context: The impact of impression management on attributions and disciplinary actions. Organizational Behavior and Human Performance , 28, 356–378.

Yammarino, F.J., Dionne, S.D., Chun, J.U., & Dansereau, F. (2005). Leadership and levels of analysis: A state-of-the-science review. Leadership Quarterly, 16, 879-919.

Yukl, G. (2009). Leadership in organizations (6th ed.). Pearson: New Jersey.