keen to help managers’ implicit

8/10/2019 Keen to Help Managers Implicit

1/32

PERSONNEL PSYCHOLOGY

2006, 59, 871902

KEEN TO HELP? MANAGERS IMPLICIT

PERSON THEORIES AND THEIR SUBSEQUENT

EMPLOYEE COACHING

PETER A. HESLINManagement and Organizations Department

Southern Methodist University

DON VANDEWALLEManagement and Organizations Department

Southern Methodist University

GARY P. LATHAMRotman School of Management

University of Toronto

Although coaching can facilitate employee development and perfor-mance, the stark reality is that managers often differ substantially intheir inclination to coach their subordinates. To address this issue, wedraw from and build upon a body of social psychology research thatfinds thatimplicit person theories(IPTs) about the malleability of per-sonal attributes (e.g., personality and ability) affect ones willingness

to help others. Specifically, individuals holding an entity theory thathuman attributes are innate and unalterable are disinclined to investin helping others to develop and improve, relative to individuals whohold the incremental theory that personal attributes can be developed.Three studies examined how managers IPTs influence the extent oftheir employee coaching. First, a longitudinal field study found thatmanagers IPTs predicted employee evaluations of their subsequent em-ployee coaching. This finding was replicated in a second field study.Third, an experimental study found that using self-persuasion principlesto induce incremental IPTs increased entity theorist managers willing-ness to coach a poor performing employee, as well as the quantity andquality of their performance improvement suggestions.

The first study in this manuscript was presented at the annual meeting of the Academy ofManagement, Honolulu, HI, August 2005. The third study in this manuscript was presentedat the annual conference of the Society for Industrial-Organizational Psychology, Chicago,IL, April 2004. It was in part based on the doctoral dissertation of the first author. We thankUte-Christine Klehe, Daniel Tzabbar, and Bob Wood for valuable assistance in developingand conducting the third study, as well as Joshua Aronson for the videos used in this study.

We also thank Bob Wood, John Slocum, Ute-Christine Klehe, Glen Whyte, and StephaneCote, for helpful comments on an earlier version of this manuscript.

Correspondence and requests for reprints should be addressed to Peter A. Heslin, CoxSchool of Business, Southern Methodist University, Box 750333, Dallas, Texas 75275-0333; [email protected].

COPYRIGHT C 2006 BLACKWELL PUBLISHING, INC.

871


2/32

872 PERSONNEL PSYCHOLOGY

Employee coaching entails managers providing one-on-one feedback

andinsights aimed at guiding andinspiringimprovements in an employees

work performance (London, 2003; Yukl, 2002). Inherent in this definition

is that coaching, compared to generic training programs, typically fo-cuses more on employees specific workplace challenges (Hall, Otazo, &Hollenbeck, 1999). Large organizations are increasingly expecting man-

agers to coach their employees (Latham, Almost, Mann, & Moore, 2005;

London, 2003). For example, employee coaching is a key managerial task

atMcKinsey & Company(Bartlett & Ghoshal, 2002),YUM! Brands, Inc.(Mike & Slocum, 2003),KPMG(Heslin & Latham, 2004), andMotorola(Latham et al., 2005).

Coaching can facilitate learning to master altered job roles, such as

following a promotion (Goldsmith, 2000) or in response to organizationalchange (Hawkins & Pettey, 2000). As coaching is provided on the job and

is tailored to the employee being coached, it is less prone to the transfer-of-

training issues that typically undermine the utility of most off-site devel-opmental initiatives (cf., Baldwin & Ford, 1988; Tracey, Tannenbaum,

& Kavanaugh, 1995). Consistent with this notion, Olivero, Bane, and

Kopelman (1997) found that employee coaching increased productivity

over and above the effects of a managerial training program. More re-

cently, Smither, London, Flautt, Vargas, and Kucine (2003) reported that

after receiving multi source feedback, managers who worked with a coachwere more likely to set specific (rather than vague) goals, solicit ideas for

improvement from their supervisors, and subsequently to receive improved

performance ratings from both their direct reports and their supervisors.

Despite the widespread endorsements of employee coaching as an

important managerial activity (e.g., Bartlett & Ghoshal, 2002), managersvary in their willingness to coach their employees (Campbell, Dunnette,

Lawler, & Weick, 1970; Heslin & Latham, 2004; London, 2003; Smither,

London, Vasilopoulos, Reilly, Millsap, & Salvemini, 1995). Indeed, as

early as 1964, it was observed that:

Although good coaching is basic to managerial productivity, most organiza-tions have difficulty getting their managers to be effective coaches (Mahler,1964, p. 28).

Yet, 40 years later, several scholars (e.g., Latham & Latham, 2000;

Orenstein, 2002; Smither & Reilly, 2001; Sue-Chan & Latham, 2004)

lamented that there still remains a paucity of research and empiricallyvalidated theory about the antecedents of employee coaching.

Overview

Following Smither and Reillys (2001) proposal that social psychol-

ogy may provide a fruitful source of insights for research on employee


3/32

PETER A. HESLIN ET AL. 873

coaching, these studies explored the potential usefulness to the employee

coaching literature of Dwecks social psychological theory and research

onimplicit person theory(IPT). Lab research with children and students

has established that compared to people holding the incrementalIPT thatpersonal attributes are malleable and hence can be developed, individualswho hold the entity IPT that personal attributes do not change are lessinclined to help others improve their performance.

Our first study examined whether managers IPT about the stability of

personal attributes is linked to the extent to which they provide develop-

mental coaching to employees. A second field study replicated our first

study. Both of these studies examined employee ratings of the frequencywith which managers coach them. Our third study investigated whether the

incremental training intervention reported by Heslin, Latham, and Vande-Walle (2005, Study 4) increased managers inclination to coach, as well

as the quality of the suggestions they provided to a hypothetical employee

whose video-recorded performance they observed.1 The results of thesestudies have practical implications for organizational initiatives aimed at

encouraging managers to coach their employees, as well as suggesting sev-

eral potentially fruitful avenues for future research. We begin by briefly

examining the nature of employee coaching.

Employee Coaching

Coaching has been defined by scholars who have focused on executivecoaching by professional coaches, as well as those concerned with em-

ployee coaching by managers. Hall et al. (1999) defined executive coach-

ing as a practical, goal-focused form of personal, one-on-one learning

that may be used to improve performance, enhance a career or prevent

derailment, and work through organizational issues or change initiatives.

In contrast, based on their annotated review of the literature, Douglas andMorley (2001) defined executive coaching somewhat more narrowly as a

short-term developmental relationship created to achieve specific, mutu-

ally agreed-upon performance goals.2

Within the domain of employee coaching conducted by managers, Cor-

coran, Petersen, Baitch, and Barrett (1995) concluded thatsales coaching

1Although the participants and incremental intervention are common between studies,Heslinet al.(2005, Study4) reportedon how theintervention affected change in participantsinitial performance appraisals. That study did not present the coaching-related data reportedin the present study.

2In contrast to professional, external coaches, who develop an identity and incomebased on their provision of coaching, coaching is just one of a wide range of managerialfunctions (Yukl, 2002). This study thus investigated a potential individual difference sourceof variability in managers propensity to coach their employees.


4/32


entails a sequence of conversations and activities that provide ongoing

feedback and encouragement to improve a salespersons performance.

Brocato (2003, p. 18) argued that the coach is the vehicle that facilitates

performance change.Despite variation in the scope of these definitions, their common theme

is thatas mentioned in the opening sentencecoaching entails a man-

ager providing one-on-one feedback and insights aimed at guiding and

inspiring improvements in an employees work performance (London,

2003; Yukl, 2002). Conceptually, the performance improvement focus of

coaching suggests it may be an initiating structure leadership behavior,

whereby managers define and organize followers roles, facilitate goalattainment, and establish well-defined patterns and channels of communi-

cation (Fleishman, 1957). However, coaching is likely to be distinct frominitiating structure insofar as helping individuals to improve is also an ex-

pression of consideration for employees, that is, showing them concern,

respect, appreciation, and support (Fleishman, 1957). An initial empiricaltest of our conceptualization of coaching as positively related to, though

distinct from, the initiating structure and consideration leadership behav-

iors is reported in the pilot study conducted in Study 1. We now address

the nature of IPT and its potential role in managers willingness to coach

their employees.

Implicit Person Theory

An IPT is ones implicit belief about the malleability of the personal

characteristics (e.g., ability and personality) that affect human behavior

(Dweck, 1999). Individuals who hold the IPT that personal attributes are

essentially a fixed entity are classified by Dweck asentity theorists; thosewho implicitly believe that personal attributes can change and be devel-

oped are labeled as incremental theorists. Implicit theories can be domain-

specific, such that people sometimes hold different implicit theories aboutthe malleability of ability, personality, and morality (Dweck, 1999). How-

ever, an individuals IPT reflects their implicit assumptions regarding the

stability of the collection of personal attributes that determine the overallkind of person that someone is and how they tend to behave (Chiu, Dweck,Tong, & Fu, 1997).

Dweck and Leggett (1988) theorized that implicit theories create an

analytic framework for interpreting and responding to the events that an

individual experiences. Consequently, they proposed that implicit theo-

ries have important motivational implications. Considerable subsequentresearch has found support for many facets of Dweck and Leggetts ar-

gument, most commonly studies conducted with grade school and un-

dergraduate students that were focused on the realm of self-regulation.


5/32


Within an organizational context, several studies have found that com-

pared to entity theorists, incremental theorists are more likely to personally

adopt learning goals (e.g., VandeWalle, 1997), develop high self-efficacy

(e.g., Martocchio, 1994), maintain their self-efficacy following setbacks(e.g., Wood & Bandura, 1989), and exhibit high performance on complexdecision-making tasks (e.g., Tabernero & Wood, 1999).

These studies have documented implicit theory effects on an individ-

uals self-regulation (i.e., achievement striving and resulting performance).To the authors knowledge, no published studies have examined the effect

of managers IPTs on their willingness to help other people to improve theirperformance. Nonetheless, an extension of Dweck and Leggetts (1988)theory suggests that managers IPTs could affect the extent to which they

coach their employees.Specifically, Dweck, Chiu, and Hong (1995a) proposed that implicit

theories influence how one perceives and relates to others. Viewing per-sonal attributes as largely fixed in nature sets up an emphasis on stabletraits for understanding and predicting others behavior (Erdley & Dweck,

1993). Thus, Dweck et al. argued that entity theorists view the behavior

of others as reflective of their enduring, static personal qualities. In con-

trast, incremental theorists tend to view others behavior as likely to be

the product of more malleable and dynamic factors, such as effort and

strategies. Thus, Dweck et al. theorized that entity theorists belief thathuman attributes are innate and unalterable makes them disinclined to in-

vest in helping others to develop and improve. Four studies have supported

Dweck et al.s theory in this regard.

First, Chiu, Dweck, et al. (1997) reported that in response to a profes-

sor who made a seemingly unfair last-minute change in grading policy,undergraduate entity theorists primarily wanted to punish the professor. In-

cremental theorists, on the other hand, were more likely to want to educate

the wrongdoer. Second, Heyman and Dweck (1998) found that holding

an incremental IPT was associated with the extent to which students gavea hypothetical struggling fellow student extensive and helpful suggestionsabout precisely what that person should do to improve. Third, Gervey,

Chiu, Hong, and Dweck (1999) observed that entity theorists indicated

that the prime purpose of imprisonment was either punishment or retri-

bution. Incremental theorists were more likely to state that imprisonment

is also to rehabilitate wrongdoers for leading a moral and productive life,reflecting a presumption that they have potential to grow and develop.

Finally, Karafantis and Levy (2004) reported that compared to 9-

to 12-year-old entity theorists, holding an incremental implicit theorywas positively associated with active participation in collecting money

for a UNICEF event, enjoying doing so, and intending to volunteer

again.


6/32


Implications for Employee Coaching

Although quite suggestive, three characteristics of previous IPT re-

search limit the basis it provides for concluding that managers holding anincremental IPT are more inclined to coach employees than their entitytheorist colleagues. First, with the exception of the self-reported volun-

teering assessed by Karafantis and Levy (2004), previous research on the

role of implicit theories in helping behavior has been limited to responses

to hypotheticalpaper people (i.e., vignettes). It cannot be assumed that re-actions to paper people will generalize to reactions toactualpeople whoare seen or heard (Cook & Campbell, 1979, p. 66). Moreover, meta-analytic research by Murphy, Herr, Lockhart, and Maguire (1986) found

that vignettes are low salience stimuli that consistently yield higher effectsizes than studies based on observation of a persons behavior. Whether

the findings from Chiu, Dweck, et al. (1997), Heyman and Dweck (1998),

and Gervey et al. (1999) can be replicated using a higher salience stimulushas yet to be examined. Either smaller or nonsignificant effects from such

research would suggest that low stimulus salience may account for the

implicit theory effects reported in these studies.

Second, the extant evidence linking incrementalism and helping is

based on self-report data. Suchperceptperceptresearch designs typically

contain a common-method bias that tends to inflate the observed relation-ship between predictor and criterion variables (Podsakoff, MacKenzie,

Lee, & Podsakoff, 2003). Research with independent ratings of partici-

pants actual behaviors, as a function of their IPT, is needed to address

this threat to the putative causal effect implied within the literature on how

IPT is related to subsequent helping behavior.Third, with reference to the results of child- and adolescent-based im-

plicit theory research, Tabernero and Wood (1999, p. 124) concluded that

studies are needed to establish if the same is true for adults. Bernardin,

Buckley, Tyler, and Wiese (2000), as well as Arvey and Murphy (1998),emphasized the importance of empirically assessing whether findings fromstudies conducted with student samples generalize to the behavior of man-

agers performing tasks that are part of their organizational role.

Study 1

Holding the entity perspective that people generally do not change

has been associated with children and students being disinclined to in-vest in other peoples performance improvement. In contrast, holding an

incremental IPT has been found to positively predict helping behavior.

Although the generalizability of this research to the domain of employee


7/32


coaching cannot be assumed and has yet to be tested, it provides the basis

for examining:

Hypothesis 1: Managers incrementalism is positively related to the extentto which they coach their employees.

To assess coaching, we were guided by Atkins and Woods (2002) findingthat compared to ratings provided by self, supervisors, and peers, sub-

ordinates were the best predictors of assessment center competency at

coaching people (p. 895). Thus, we used 170 subordinate observations

as the criterion of managers coaching behavior in this longitudinal field

study. Specifically, 6 weeks after we measured managers IPTs, we col-lected subordinate evaluations of their employee coaching in order to test

Hypothesis 1.

Method

Participants. Based on the mediumlarge effect size of IPT on chil-drens advice-giving reported by Heyman and Dweck (1998, i.e., d= .67),3

using the power guidelines of Cohen (1992), we estimated that 42 partic-

ipants were required to achieve an acceptable level of power (.80) in this

study. The participants were 45 managers who were engaged in a 6-week

professional MBA elective module on leading organizational change ata private, southwestern U.S. university. Participants mean age was 31.1

years (SD= 4.1); 21 were women and 24 were men. On average, they had3.0 years managerial experience (SD = 2.8) and their employee coachingwas assessed by 5.0 subordinates (SD = 1.8).

Procedure. Four weeks before beginning a 6-week professional MBAelective module, all the managers completed a range of Web-based surveys

for the purpose of later class discussion. Given the equivalence of Web-based and paper-and-pencil ratings (Smither, Walker, & Yap, 2004), as

well as to the convenience of the Internet, we assessed participants IPTvia the Web in order to optimize the rate of IPT survey completion.

Six weeks later, participants requested 3 to 10 of their employees to

provide them with anonymous feedback about their effectiveness at lead-ing organizational change, via an anonymous Web-based survey. As all

the participants worked in different organizations, this approach to rater se-

lection, which is consistent with previous upward feedback research (e.g.,

Brett & Atwater, 2001; Heslin & Latham, 2004), was necessary to identify

suitable respondents. In order to minimize the possibility that participants

would select raters who were most likely to provide them with favorable

3Which is typical of the effect sizes reported in the extant IPT literature.


8/32


ratings, the managers and subordinates were informed that the evaluations

would have no impact on course grades and that the feedback would be

anonymous and confidential. The request for feedback was framed as a

valuable learning opportunity, and thus, subordinates should be as accu-rate and honest as possible in their feedback. Feedback providers werealso assured of their anonymity because doing so probably improves their

ability to make accurate ratings (Murphy, Cleveland, & Mohler, 2001).

We also assured subordinates anonymity by informing them that partic-

ipants would only receive a feedback report if at least three subordinates

provided their manager with feedback.

Two weeks after soliciting feedback, all participants received an up-ward feedback report that presented their mean subordinate evaluation for

each coaching survey item. Subordinates comments were also presentedverbatim. None of the managers reported having been aware of the hy-

pothesized link between IPT and coaching, even after it was subsequently

explained for pedagogical purposes.Employee coaching measure. To assess coaching behaviors, we first

conducted an extensive literature review to identify the fundamental el-

ements of coaching behavior (e.g., Ahern, 2003; Ellinger & Bostrom,

1999; Graham, Wedman, & Garvin-Kester, 1994; Heslin & Latham, 2004;

London, 2003; Orth, Wilkinson, & Benfari, 1987; Smither et al., 1995,

2003; Yukl, 2002). Based on our review, we conceptually derived threeintegral components of coaching employees for increased performance:

1. Guidance: The communication of clear performance expectations andconstructive feedback regarding performance outcomes, as well as

how to improve.

2. Facilitation: Helping employees to analyze and explore ways to solve

problems and enhance their performance.

3. Inspiration: Challenging employees to realize and develop their po-

tential.

To operationalize coaching behavior, we next developed a 10-item behav-

ioral observation scale (BOS; Latham & Wexley, 1994) to assess these

dimensions (see Table 1). Consistent with Smither et al. (1995), subordi-nates rated each item on a 5-point Likert-type scale with the anchors 1 =

Not at all to 5 = To a very great extent. We prefaced the items with thefollowing introduction, To what extent does the person to whom you are

providing feedback. Subordinates were also given the option to indicate

that their manager had no opportunity to demonstrate each behavior.To initially assess the coaching behavior scale items, a Web-based pilot

study asked 160 MBA students about the coaching, leadership, and sup-

portive behavior of their current or most recent manager. The participants


9/32


TABLE 1

Completely Standardized Solution for Confirmatory Factor Model

of Coaching Behavior

Coaching behavior items Guidance Facilitation Inspiration

1. Provide guidance regarding performance

expectations?

.89

2. Help you to analyze your performance? .90

3. Provide constructive feedback regarding

areas for improvement?

.92

4. Offer useful suggestions regarding how

you can improve your performance?

.87

5. Act as a sounding board for you to

develop your ideas?

.85

6. Facilitate creative thinking to help solve

problems?

.90

7. Encourage you to explore and try out new

alternatives?

.83

8. Express confidence that you can develop

and improve?

.93

9. Encourage you to continuously develop

and improve?

.93

10. Support you in taking on new challenges? .93

Higher-order factor loadings .83 .93 .92

Note.All parameter estimates are significant atp < .01.

average age was 29.95 years (SD = 4.67), they had an average of 7.36years of work experience (SD = 4.6), and 35% were women.

To analyze the coaching data, we used LISREL 8.54 for confirmatory

factor analysis (CFA). Based on our three-dimensional conceptualization

of coaching behavior, we modeled coaching behavior as a higher-order

factor with three first-order factors. As shown in Table 1, the CFA of theitem responses indicated that the theoretical model provided a good fit to

the data, 2(32, N= 160) = 134.52, p < .01, standardized root-mean-square residual (SRMR) = .04; comparative fit index (CFI) = .96. All

of the coaching behavior item loadings were significant at the p < .01level. Moreover, the three coaching behavior dimensions had significant

loadings (p < .01) with the second-order factors of .83 (guidance), .93 (fa-cilitation), and .92 (inspiration). We next tested a three-factor model with

no higher-order factor. This second model produced fit statistics that were

identical to the higher-order model. Rindskopf and Rose (1988) noted thatwhen a higher-order model has only three first-order factors, it is equiva-

lent to the three-factor first-order model. They explain that such models are

not discriminable and will have the same fit statistics. Given this model


10/32


equivalency, traditional comparisons such as a chi-square difference test

do not apply. Rather, Rindskopf and Rose indicate that the choice of model

form is to be based on substantive grounds. If there are substantive grounds

for a higher-order model, then it is preferred based on parsimony. Giventhat we did nota prioriexpect that the three coaching dimensions wouldhave differential relationships with our criterion variable and that we did

not empirically find differential relationships, we selected the more parsi-

monious higher-order model for our hypothesis testing.

Using other data collected from our Web-based pilot study, we also

assessed the convergent and divergent validity of the coaching scale (see

Table 2). As expected, the composite coaching scale (as well as its three di-mensions) had positive relationships with the initiating structure and con-

sideration leadership styles (assessed using the LBDQ-Form XII, Stogdill,1963), as well as with social support from ones supervisor4 (Caplan, Cobb,

French, Van Harrison, & Pinneau, 1980). Evidence for the divergent va-

lidity of our coaching scale was that responses to it were unrelated tosocial support received from either (a) others at work, or (b) ones spouse,

friends, and relatives (Caplan et al., 1980). Finally, employee coaching

by managers was also unrelated to subordinate gender, age, or work ex-

perience. In this study, the composite coaching scale data had acceptable

reliability ( = .89).

Social support measure. The related Social Support from SupervisorIndex, the Social Support from Others at Work Index, and the Social Sup-

port from Wife, Friends, and Relatives Index (Caplan et al., 1980) each

contain four related questions such as How much does each of these peo-

ple go out of their way to do things to make your work life easier for you?(a) your immediate supervisor, (b) other people at work, and (c) your wife

[husband], friends and relatives. A second sample item from each scale is

How much can each of these people be relied on when things get tough at

work? (a) your immediate supervisor (boss), (b) other people at work, and

(c) your wife [husband], friends and relatives. Responses are obtained ona 5-point Likert-type scale ranging from 4 = very much through to 0 =dont have any such person. Evidence for the divergent validity of thesescales, which are among the most established scales used to measure social

support in a job (Lim, 1996), has been provided by Caplan et al. (1980).

Regarding construct validity, as expected, supervisor and coworker sup-

port has been shown to correlate positively with overall job satisfaction andwork group cohesiveness (Repetti & Cosmas, 1991), as well as negatively

with job insecurity, job dissatisfaction, and noncompliant work behaviors

(Lim, 1996).

4Using the social support measure described below.


11/32


TABL

E2

DescriptiveStatisticsandCorrelationsforC

oachingBehaviorNomologi

calNetwork

Variable

M

SD

1

2

3

4

5

6

7

8

9

10

11

12

1.

Coachingguidance

2.9

4

.96

(.92)

2.

Coachingfacilitation

3.1

8

1.0

6

.68

(.87)

3.

Coachinginspiration

3.5

8

1.1

1

.68

.76

(.93)

4.

Coaching(higherorder)

3.2

0

.93

.89

.90

.90

(.94)

5.

Initiating

structure

3.4

3

.76

.55

.49

.50

.58

(.90)

6.

Consideration

3.2

8

.45

.41

.47

.53

.52

.49

(.71)

7.

Socialsu

pportfromsupervisor

3.8

3

.77

.46

.52

.54

.56

.59

.59

(.79)

8.

Socialsu

pportfromothers

3.9

1

.62

.0

1

.0

1

.04

.0

2

.09

.0

7

.25

(.68)

9.

Socialsu

pportfromwife,

4.5

9

.50

.1

0

.0

7

.02

.0

7

.06

.0

4

.11

.32

(.77)

friends,andrelatives

10.

Gender

1.6

5

.48

.03

.0

6

.01

.0

1

.15

.13

.16

.00

.09

()

11.

Age

29.9

5

4.6

7

.12

.10

.03

.10

.00

.04

.01

.0

2

.02

.18

()

12.

Workexperience

7.3

6

4.6

0

.16

.10

.06

.12

.0

1

.00

.00

.00

.00

.09

.91

()

Note.N=

160.

p

keen to help managers’ implicit

Documents