Unit III Scholarly

Article

Highlighting Effort Versus Talent in Service Employee Performance: Customer Attributions and Responses

Fine F. Leung , Sara Kim , and Caleb H. Tse

Abstract Firms often attribute their service employees’ competent performance to either dedicated effort or natural talent. However, it is unclear how such practices affect customer evaluations of service employees and customer outcomes. Moreover, prior work has primarily examined attributions of one’s own performance, providing little insight on the impact of attributions of others’ per- formance. Drawing on research regarding the warmth–competence framework and performance attributions, the current research proposes and finds that consumers expect a more communal-oriented and less exchange-oriented relationship when a service employee’s competent performance is attributed to dedicated effort rather than natural talent, as effort (vs. talent) attribution leads consumers to perceive the employee as warmer. The authors further propose customer helping behaviors as downstream consequences of relationship expectations, finding that effort (vs. talent) attribution is more likely to induce cus- tomers’ word-of-mouth and idea provision behaviors. The findings enrich existing literature by identifying performance attri- butions as a managerially meaningful antecedent of relationship expectations and offer practical guidance on how marketers can influence consumers’ relationship expectations and helping behaviors.

Keywords attribution theory, communal and exchange relationships, competence and warmth, customer helping behaviors, service employee performance

Online supplement: https://doi.org/10.1177/0022242920902722

When firms communicate information about their service

employees’ competent performance, they often attribute it to

either dedicated effort or natural talent. For example, on their

websites, financial services firms such as Citigroup state that

“Citi works tirelessly . . . . We strive to create the best out- comes,” and Partners Group Holding asserts that “we work

hard and deliver outstanding results.” In contrast, Manulife

Financial highlights that the “talent of our employees is what

makes Manulife Financial a successful organization,” and

BlackRock states, “Our best solutions come from the contribu-

tions of a group of talented and smart people” (for more exam-

ples, see Web Appendix W1). We systematically examined the

company websites of the top service firms on the 2018 Forbes

Global 2000 list and found that many top financial and health

care services firms mention these two types of performance

attribution on their websites (see Figure 1). Despite the

real-world prevalence of references to these two types of per-

formance attribution, it is unclear how firms’ promotions of

performance attributions affect customer evaluations of service

employees and customer outcomes.

Academic research suggests two types of performance attri-

bution: one to dedicated effort and the other to natural talent

(Brown et al. 2018; Dweck 2000; Weiner 1972). Psychology

and marketing literature has mainly examined a fixed or malle-

able view of people’s own performance and its impact on how

they judge themselves (e.g., judgments of one’s own intelli-

gence or personality; Dweck 2000; Dweck and Leggett 1988),

their own performance (e.g., academic performance; Hong

et al. 1999; Weiner 1972), or brands/products (Mathur, Block,

and Yucel-Aybat 2014; Murphy and Dweck 2016), providing

little insight on relationship judgments and behaviors toward

others. However, in service relationships, beliefs about others’

Fine F. Leung is Assistant Professor of Marketing, Faculty of Business, Hong

Kong Polytechnic University, Hong Kong (email: [email protected]).

Sara Kim (corresponding author) is Associate Professor of Marketing,

Faculty of Business and Economics, University of Hong Kong, Hong Kong

(email: [email protected]). Caleb H. Tse is Assistant Professor of Marketing,

Nanyang Business School, Nanyang Technological University, Singapore (email:

[email protected]).

Journal of Marketing 2020, Vol. 84(3) 106-121

ª American Marketing Association 2020 Article reuse guidelines:

sagepub.com/journals-permissions DOI: 10.1177/0022242920902722

journals.sagepub.com/home/jmx

performance can influence relationships with those others

(e.g., how a customer views a service employee’s performance

can influence the customer’s relationship with the employee;

Bitner 1995). We recognize this characteristic in service rela-

tionships, as well as a lack of research on the attributions of

others’ performance and their impact on relational aspects;

thus, in this article we examine how attributions of service

employees’ performance influence consumers’ relationship

expectations with and behaviors toward service employees.

We propose that attributions of service employees’ compe-

tent performance can change the extent to which customers

expect a more communal-oriented (or less exchange-oriented)

relationship. Prior work has conceptualized consumers’ rela-

tionship expectations with service employees along the com-

munal–exchange continuum (Aggarwal 2004; Clark and Mills

1993). In a communal relationship, consumers expect a service

employee to take genuine care of them and understand their

needs as a friend or family member would. In an exchange

relationship, consumers consider a service employee strictly

as a business partner and expect the employee to provide ser-

vices that will be worth their money. We propose that

consumers will expect a more communal and less exchange-

oriented relationship when a service employee’s competent

performance is attributed to dedicated effort rather than natural

talent, because effort (vs. talent) attribution leads consumers to

perceive the employee as warmer. We further examine cus-

tomer helping behaviors toward firms (i.e., voluntary and dis-

cretionary behaviors that aid firms beyond those required in the

purchase of products and services; Bettencourt 1997; Bove

et al. 2009) as downstream behavioral consequences of rela-

tionship expectations. In particular, we propose that highlight-

ing service employees’ effort (vs. talent) can increase customer

helping behaviors such as word of mouth (WOM) and idea

provision.

Our investigation of service employees’ performance attri-

butions makes several theoretical and managerial contribu-

tions. First, the current research broadens our understanding

of social judgments in commercial relationships. Although a

considerable body of research has investigated the relation

between judgments of competence and warmth (Judd et al.

2005; Yzerbyt, Kervyn, and Judd 2008), this prior work has

mainly examined how a certain level of competence is related

to warmth perception. The current research examines the attri-

butions of competence as a new dimension that influences

warmth perception, holding the objective level of competence

constant. Second, the marketing literature has focused on the

downstream consequences of a communal versus exchange

relationship with consumers (Aggarwal 2004; Aggarwal and

Zhang 2006; Wan, Hui, and Wyer 2011), but very few studies

have proposed firm tactics that could induce a certain type of

relationship expectation (communal or exchange). For instance,

McGraw, Schwartz, and Tetlock (2012) showed that a com-

pany’s communal obligations (e.g., providing medical care on

the basis of need instead of ability to pay) can influence con-

sumers’ relationships with the company. The current research

enriches the existing literature by examining performance attri-

butions as an antecedent of relationship expectations. Third, as

we have mentioned, whereas most prior work in the marketing

literature has focused on attributions of one’s own performance

and their effect on product evaluation and choice (Mathur,

Block, and Yucel-Aybat 2014; Murphy and Dweck 2016), our

work examines the attributions of others’ performance.

Our findings also provide important marketing insights.

Figure 1 indicates that firms often attribute their employees’

performance to effort or talent. Our research proposes that

firms can strategically implement such performance attribu-

tions to evoke a type of relationship expectation that they want

to promote (e.g., highlighting employees’ effort when a firm

wants to promote a communal-oriented relationship with cus-

tomers). Thus, performance attribution is a managerially mean-

ingful antecedent of relationship expectations, because it can

be embedded in communication messages without requiring

customers to have direct interactions with service employees.

Our research also provides implications on customer attention

to communication messages. We suggest that customers’ rela-

tionship expectations can be manifested in their attention to

service employee information. If firms want to attract

Health Care Services (19 Companies)

Effort Attribution

31.6%

Talent Attribution

15.8%

Neither

31.6%

Both

21.1%

Investment Services (74 Companies)

Talent Attribution

27.0%

Effort Attribution

18.9% Neither

44.6% Both

9.5%

Life and Health Insurance (33 Companies)

Effort Attribution

21.2%

Talent Attribution

24.2%

Neither

36.4% Both

18.2%

Major Banks (64 Companies)

Talent Attribution

23.4%

Effort Attribution

18.8%Neither

40.6% Both

17.2%

Figure 1. Prevalence of service employee performance attributions among top service firms. Notes: The figures represent the percentages of top service firms on the 2018 Forbes Global 2000 list that explicitly communicate either dedicated effort or natural talent (or both) or that do not provide performance attribution information on the company websites. Two independent coders were instructed to code performance attributions on the web pages in which firms deliver communication messages toward their customers. Agreement between the coders was high (83%), and disagreements were resolved by discussion.

Leung et al. 107

customers’ attention to person-related information (e.g., per-

sonal background information about service employees), they

can highlight the effort of their employees, whereas if they

want customers to focus on job-related information (e.g., what

service employees do), they can highlight the talent of their

employees. Finally, our research suggests that promoting dif-

ferent types of performance attribution can shape customer

behavior. Specifically, by highlighting employees’ effort (vs.

talent), firms can increase customer helping behaviors such as

sharing the firms’ information on social networks or providing

new product ideas.

Social Judgments and Performance Attributions

Research in social psychology as well as marketing has sup-

ported the notion that when people form impressions about

others, they tend to make judgments along two fundamental

dimensions: competence (e.g., capability, skillfulness, effi-

cacy) and warmth (e.g., friendliness, helpfulness, trustworthi-

ness) (Abele and Wojciszke 2014; Fiske et al. 2002). For

example, judgments of competence and warmth shape consu-

mers’ relationships with commercial partners, such as non-

profit and for-profit firms (Aaker, Vohs, and Mogilner 2010),

salespeople (Scott, Mende, and Bolton 2013), and brands

(Kervyn, Fiske, and Malone 2012). This line of research has

investigated the relations between judgments of competence

and warmth, mainly by examining how a certain level of

competence is related to warmth perception. Some studies

have reported that a higher level of competence results in

greater warmth perception (Rosenberg, Nelson, and Viveka-

nanthan 1968; Suitner and Maass 2008), whereas others have

shown that a lower level of competence leads to greater

warmth perception (Judd et al. 2005; Yzerbyt, Provost, and

Corneille 2005).

Extending the existing literature, the current research exam-

ines the attributions of competence as a new dimension of

competence influencing warmth perception, holding the objec-

tive level of competence constant. Research in social psychol-

ogy has corroborated dedicated effort and natural talent as two

internal sources of people’s performance (Brown et al. 2018;

Tsay and Banaji 2011). In the case of dedicated effort, compe-

tent performance is believed to be the result of commitment,

perseverance, and hard work. In the case of natural talent,

competent performance is believed to be the result of innate

aptitude. This typology is also in line with implicit theories

suggesting that people’s performance can be attributed to mal-

leable traits such as effort or to fixed traits such as natural talent

(Dweck 2000; Murphy and Dweck 2016). In addition, prior

work on attribution theory has made it clear that competence

can be attained through either dedicated effort or natural talent.

Weiner (2005) suggests that “when associated with aptitude

[natural talent], the concept of competence is conceived as

mainly uncontrollable, whereas when associated with effort

expended, the attainment of competence is conceived as con-

trollable” (p. 79). Thus, conceptually, effort and talent are two

different attributions of competence. Bridging these two

streams of research on the warmth–competence framework and

performance attributions, we examine how information on dif-

ferent attributions of competent performance changes warmth

judgments and, in turn, relationship expectations.

Employee Performance Attributions and Relationship Expectations

We posit that attributing a service employee’s competent per-

formance primarily to dedicated effort (vs. natural talent)

makes consumers perceive the employee as warmer. Extant

research indirectly supports this proposition. Prior work has

found that when a person’s performance is attributed to effort,

that person is more likely to be seen as “one of us” (Hong and

Lin-Siegler 2012), because most people generally believe that

they also need to exert high effort to succeed (Klein and

O’Brien 2017). Indeed, when students learn about successful

scientists’ hard work in their scientific discoveries, they are

more likely to see the scientists as ordinary people (Lin-

Siegler et al. 2016). Prior work also has indicated that those

who are socially close are perceived to be warmer than those

who are socially distant (Leyens et al. 2000). For instance,

compared with out-group members, in-group members are

rated as having a greater capacity to experience emotions and

being higher in warmth (Harris and Fiske 2006). Therefore, we

propose that compared with talent attribution, effort attribution

will lead the customer to perceive the service employee as

warmer.

In contrast, because most people tend to believe that only a

few individuals possess natural talent (Emerson and Murphy

2015), talent attribution can increase perceived social distance.

Geniuses and exceptionally talented individuals are typically

perceived to “have” something that most people do not have

and, thus, are seen as different (Fuchs 2001). In addition, Lin-

Siegler et al. (2016) note that viewing scientists as individuals

with a special aptitude for science discourages students from

feeling connected with the scientists. When one feels discon-

nected from another individual, one is less likely to attribute the

ability to feel to that person (Leyens et al. 2000). For instance,

compared with in-group members, out-group members are

rated as lacking emotional capacity and as being more self-

centered (Harris and Fiske 2006). Furthermore, people tend

to see naturally talented others as disconnected from human

experiences and emotionally inert (Klein and O’Brien 2017),

and gifted intellectuals are considered to be more antisocial

than others (Persson 2007). Teachers often view gifted and

talented students as emotionless, antisocial, and insensitive to

the feelings of others (Baudson and Preckel 2013; Geake and

Gross 2008). Therefore, compared with effort attribution, talent

attribution that can increase perceived social distance between

a customer and an employee will make customers perceive a

service employee as less warm.

Our research further posits that the perceived warmth of

service employees is the basis for consumers’ relationship

expectations with those employees. Consumers in a communal

108 Journal of Marketing 84(3)

relationship expect a service employee to take care of them and

consider their needs (Aggarwal 2004; Clark and Mills 1993). In

contrast, in an exchange relationship, parties understand that

the benefits received should correspond to the benefits given,

focusing on self-interest (Clark and Mills 1993; Kwak, Puza-

kova, and Rocereto 2015). Although commercial relationships

always involve elements of exchange relationships, such as

monetary exchange, consumers’ relationship expectations can

vary on the communal–exchange continuum, because consu-

mers can expect different degrees of communality in commer-

cial relationships depending on the situation (Aggarwal and

Law 2005; Aggarwal and Zhang 2006).

When consumers perceive a service employee to be

warm, they will likely expect that the employee will be

cooperative, have other-profitable (rather than self-profit-

able) intentions, and show genuine concern for consumers’

needs (Aggarwal and Zhang 2006; Kervyn, Fiske, and Mal-

one 2012). Such expectations are consistent with the norms

of communal relationships. In contrast, people tend to

expect a cold person to show less empathy for others and

care more about him- or herself than about others (Kervyn,

Fiske, and Malone 2012). In addition, when people see oth-

ers as being low in warmth and lacking emotional respon-

siveness, they can more readily perceive those others as

instruments for their own goals (Haslam 2006). For exam-

ple, viewing others as emotionless helps managers make

decisions in difficult situations (e.g., layoff decisions) by

seeing those individuals as objects or instruments to achieve

their goals (Haslam and Loughnan 2014). Perceiving others’

self-centered intentions and focusing on the instrumentality

of others are behaviors in line with the characteristics of

exchange relationships. Thus, when customers perceive an

employee as warmer (less warm), they will expect a more

(less) communal-oriented relationship with him or her along

the communal–exchange continuum. Thus,

H1: Consumers expect a more communal-oriented (i.e.,

less exchange-oriented) relationship when a service

employee’s competent performance is attributed to dedi-

cated effort rather than to natural talent.

H2: The effect of service employees’ performance attri-

butions on consumers’ relationship expectations is

mediated by warmth judgments regarding the service

employees.

Figure 2 depicts our conceptual framework and the flow of the

studies. We first present five studies providing empirical evi-

dence for the link between performance attributions and rela-

tionship expectations. Study 1a shows that when a service

employee’s competent performance is attributed to dedicated

effort rather than to natural talent, consumers expect a more

communal- and less exchange-oriented relationship with the

employee. In Study 1b, we examine simultaneous attribution

to both effort and talent. We then test whether perceived

warmth underlies the effect of performance attributions on

relationship expectations by directly measuring the variable

(Study 2a) and by manipulating the perceived warmth of the

service employee (Study 2b). Study 3 uses eye-tracking tech-

nology to show that effort attribution leads consumers to pay

more attention to person- than to job-related information about

the service employee. We then develop our hypothesis for

customer helping behaviors as downstream consequences of

relationship expectations and present two studies, one using a

real firm context (Study 4) and the other in a field experiment

(Study 5), to support the hypothesis.

Study 1a: Service Employee Performance Attribution

In Study 1a, we attribute a service employee’s competent per-

formance either to dedicated effort or to natural talent and test

whether effort attribution leads participants to expect a more

communal-oriented relationship with the employee. We also

examine a control condition in which no information about

performance attribution is provided.

Method

Two hundred seventy participants (106 women; mean age ¼ 37.59 years) were recruited online from Amazon Mechanical

Turk in exchange for monetary compensation.

Performance Attributions (Natural Talent vs. Dedicated Effort)

Perceived Warmth

Relationship Expectations (Exchange vs. Communal)

Word-of-Mouth Behavior (Study 4)

(Studies 1–3)

Idea Provision Behavior (Study 5)

Customer Helping Behaviors

Figure 2. A conceptual framework of the current research.

Leung et al. 109

Performance attribution manipulation. Participants were told that the medical society in a U.S. city periodically featured the

city’s top physicians and were asked to provide feedback on

an article. All participants read identical information on per-

formance, such that a physician had received a peer review

rating in the top 10% of general physicians in the city. Then, in the effort attribution condition, participants read statements

attributing the physician’s performance to effort (e.g., “[He/

she] puts a lot of effort into the work”), whereas in the talent

attribution condition, they read statements attributing the phy-

sician’s performance to talent (e.g., “[He/she] is naturally skill-

ful at the work”; see Web Appendix W2). The control

condition article only stated the physician’s performance with-

out any information on performance attribution. As a manip-

ulation check, participants indicated the extent to which they

thought the physician had achieved his or her level of perfor-

mance because of effort or talent with three items (e.g., “Put a

lot of effort into his or her work/Was naturally talented at his or

her work”; a ¼ .96).

Dependent variable. Next, participants rated the degree to which they would expect their relationship with the physician to be

communal- or exchange-oriented using eight items adapted

from Aggarwal (2004). Five items tapped into communal rela-

tionship expectation (e.g., “a person with whom I would want

to interact outside of business”) and three tapped into exchange

relationship expectation (e.g., “a person with whom I would

interact only for business purposes”; 1 ¼ “not at all,” and 7 ¼ “very much”). In all studies, we followed prior work (Aggarwal

2004; Scott, Mende, and Bolton 2013) and combined the

reverse-coded items on exchange relationship expectation with

the items on communal relationship expectation (a ¼ .89). Web Appendix W3 lists measurement items for all studies.

Control variables. Participants responded to questions related to the design (“I like the design of the article”), credibility (“I

think the content is credible”), and understandability (“I think

the content is easy to understand”) of the article (1 ¼ “strongly disagree,” and 7 ¼ “strongly agree”), as well as their knowl- edge of health care services (“How much do you know about

health care services in general?”; 1 ¼ “not at all,” and 7 ¼ “very much”), attention to the study (1 ¼ “paid little attention,” and 7 ¼ “paid a lot of attention”), and mood (1 ¼ “feel bad,” and 7 ¼ “feel good”) as control variables. The control variables did not differ across the conditions (ps > .10).

Results

Manipulation check. A one-way analysis of variance (ANOVA) revealed a significant effect of performance attributions among

the conditions (F(2, 267) ¼ 42.85, p < .001, Z 2p ¼ .24). Parti- cipants in the effort attribution condition (M ¼ 2.35, SD ¼ 1.71) were more likely to attribute the physician’s performance

to dedicated effort than those in the talent attribution condition

(M ¼ 4.83, SD ¼ 2.05; t(267) ¼�8.70, p < .001, d ¼�1.31). Moreover, performance attribution in the control condition (M

¼ 2.94, SD ¼ 1.89) scored in the middle and was significantly different from that in the effort attribution condition (M ¼ 2.35, SD ¼ 1.71; t(267) ¼ 2.03, p < .05, d ¼ .33) and that in the talent attribution condition (M ¼ 4.83, SD ¼ 2.05; t(267) ¼ �6.85, p < .001, d ¼�.96). Therefore, neither effort nor talent attribution seems to be a default attribution in the absence of

attribution information.

Relationship expectations. A one-way ANOVA revealed that per- formance attributions had a significant effect on relationship

expectations (F(2, 267) ¼ 8.50, p < .001, Z 2p ¼ .06). Planned contrasts revealed that participants expected their relationship

with the physician to be more communal when the physician’s

performance was attributed to effort (M ¼ 3.62, SD ¼ 1.15) than when it was attributed to talent (M ¼ 2.88, SD ¼ 1.18; t(267) ¼ 4.12, p < .001, d ¼ .64), in support of H1. In addition, participants’ relationship expectations in the control condition

(M ¼ 3.24, SD ¼ 1.25) were significantly lower than those in the effort attribution condition (M ¼ 3.62, SD ¼ 1.15; t(267) ¼ �2.08, p < .05, d ¼ �.32) and higher than those in the talent attribution condition (M ¼ 2.88, SD ¼ 1.18; t(267) ¼ 2.07, p < .05, d ¼ .30).

Discussion

Study 1a offers preliminary evidence for our primary proposi-

tion that individuals expect a more communal-oriented rela-

tionship with a service employee whose performance is

attributed to effort rather than to talent. The findings also show

that either effort or talent attribution changes relationship

expectations, compared with when there is no attribution,

which indicates that neither of the performance attributions

may be the default attribution in consumers’ minds. Rather,

firms can strategically create communication messages to high-

light effort or talent, which can move customers’ relationship

expectations with their service employees along the commu-

nal–exchange continuum. Some might argue that there may be

other more direct ways to develop communal relationships,

such as by treating customers well and satisfying them. How-

ever, these tactics require actual interactions with customers.

The current research suggests that communication messages

that do not involve interactions with customers can still create

a certain type of relationship expectation. In the next study, we

additionally examine a situation in which the performance is

simultaneously attributed to both effort and talent.

Study 1b: Performance Attribution to Both Dedicated Effort and Natural Talent

Although researchers have agreed that effort and talent attribu-

tions are on opposite ends of a continuum (Hong et al. 1999),

and our research focuses on the relative emphasis on effort or

talent, firms might communicate both effort and talent, as Fig-

ure 1 illustrates. Thus, in Study 1b, we examine simultaneous

attribution to both effort and talent. Prior work on attribution

theory has shown that people tend to perceive that naturally

110 Journal of Marketing 84(3)

talented people’s achievements come without effort (Tsay

2016; Tsay and Banaji 2011). Therefore, providing information

about a service employee’s natural talent without any informa-

tion about his or her effort can increase social distance (Lin-

Siegler et al. 2016) and lower warmth perception. However,

prior work has also shown that learning that even talented

people (e.g., great scientists like Einstein) had to exert high

effort to succeed can increase people’s sense of relatedness

with those talented people (Hong and Lin-Siegler 2012; Lin-

Siegler et al. 2016). Thus, we argue that, compared with talent

attribution only, simultaneous attribution to dedicated effort

and natural talent can help consumers understand that even a

talented employee is someone like them—that is, someone who

needs to put in a lot of effort to achieve good performance—

which will enhance warmth judgments of and a communal

relationship expectation toward the employee.

Method

One hundred twenty-five undergraduate students (81 women;

mean age ¼ 20.38 years) from a large university in Hong Kong participated in this laboratory experiment in exchange for mon-

etary compensation. Effort and talent attribution was similar to

that in Study 1a. Participants read an article about an accoun-

tant whose competent performance (e.g., “has ranked Jesse in

the top 15% among CPAs in Hong Kong”) was attributed to either effort (e.g., “Jesse puts a lot of effort into the work”) or

talent (e.g., “Jesse is naturally skillful at the work”). In the

effort-and-talent attribution condition, participants read state-

ments attributing the accountant’s performance to both effort

and talent (e.g., “Jesse puts a lot of effort and is naturally

skillful at the work”; see Web Appendix W4). After reading

the article, participants indicated their relationship expectations

with the accountant as in Study 1a.

Results

A one-way ANOVA revealed that performance attributions had

a significant effect on relationship expectations (F(2, 122) ¼ 3.08, p < .05, Z 2p ¼ .05). In a replication of the previous findings, effort attribution (M ¼ 3.47, SD ¼ 1.00) induced a more communal relationship expectation than talent attribution

(M ¼ 2.99, SD ¼ .96; t(122) ¼ 2.22, p < .05, d ¼ .49), further supporting H1. In addition, the effort-and-talent attribution (M

¼ 3.43, SD ¼ .92) induced a more communal relationship expectation than the talent attribution (M ¼ 2.99, SD ¼ .96; t(122) ¼ 2.11, p < .05, d ¼ .47), but it was not different from the effort attribution (M ¼ 3.47, SD ¼ 1.00; t(122) ¼�.18, p ¼ .86, d ¼�.04).

Discussion

Study 1b reveals that attributing a service employee’s perfor-

mance to both dedicated effort and natural talent yields an

effect similar to that of effort attribution only. As long as effort

is made salient, consumers perceive a more communal (or less

exchange-oriented) relationship with the service employee

compared with a situation in which effort information is not

salient. Note, however, that our findings do not imply that

highlighting both effort and talent is always preferable to high-

lighting only one or the other. For example, compared with

talent attribution only, attribution to both effort and talent can

create expectations of a more communal relationship, and such

expectations may not align with the service propositions of a

firm that tends to engage in exchange-oriented relationships. In

the next study, we test the mechanism for the effect of perfor-

mance attributions on relationship expectations by directly

measuring the perceived warmth of a service employee.

Study 2a: Mediating Role of Perceived Warmth

In Study 2a, we investigate the mechanism underlying the

effect of performance attributions on consumers’ relationship

expectations. We predict that attributing a service employee’s

performance to effort (vs. talent) leads participants to perceive

the employee as warmer and therefore to expect a more

communal-oriented relationship with that employee.

Method

Two hundred thirty-five undergraduate students (150 women;

mean age ¼ 19.98 years) from a large university in Hong Kong participated in this laboratory experiment in exchange for mon-

etary compensation. Participants were told that a bank on cam-

pus was promoting an investment program for university

students. They then read an advertisement featuring an invest-

ment manager whose competent performance (e.g., “winner of

best employee of the year award and ranked in the top 1% in performance”) was attributed to either effort (e.g., “I work very

hard to pick my investments”) or talent (e.g., “I am talented at

picking my investments”; see Web Appendix W5). In this

study, we used “top 1%” to reduce the range of the performance level in participants’ mind to control competence perceptions.

Manipulation check. Participants indicated the extent to which they thought the investment manager had achieved his or her

level of performance because of effort or talent, using a seman-

tic differential scale with three items (e.g., “Put a lot of effort

into his or her work/Was naturally talented at his or her work”;

a ¼ .97).

Measures. Participants then indicated their relationship expec- tations with the investment manager as in Studies 1a and 1b (a ¼ .89). We also measured the extent to which participants perceived the investment manager to be warm with six items

(e.g., “friendly,” “warm”; 1 ¼ “not at all,” and 7 ¼ “very much”; Gershon and Cryder 2018; a ¼ .89).

Control variables. To ensure that the performance attribution manipulation did not induce different competence perceptions,

we measured perceived competence of the investment manager

with six items (e.g., “competent,” “capable”; 1 ¼ “not at all,”

Leung et al. 111

and 7 ¼ “very much”; Gershon and Cryder 2018; a ¼ .89). We also measured perceived attractiveness of the investment man-

ager to check whether the performance attribution manipula-

tion affects attractiveness perceptions. Neither of the variables

differed across conditions (all ps > .20).

Results

Manipulation check. Participants in the effort attribution condi- tion (M ¼ 2.10, SD ¼ 1.22) were more likely to attribute the investment manager’s performance to effort than those in the

talent attribution condition (M ¼ 5.53, SD ¼ 1.53; t(233) ¼ �18.99, p < .001, d ¼�2.48).

Relationship expectations. Again, in support of H1, participants expected a more communal relationship when the investment

manager’s performance was attributed to effort (M ¼ 3.21, SD ¼ 1.21) rather than to talent (M ¼ 2.75, SD ¼ 1.14; t(233) ¼ 2.98, p < .01, d ¼ .39).

Perceived warmth. Participants perceived the investment man- ager as warmer when his performance was attributed to effort

(M ¼ 4.47, SD ¼ 1.05) rather than to talent (M ¼ 3.87, SD ¼ 1.17; t(233) ¼ 4.15, p < .001, d ¼ .54), in support of H2. To establish discriminant validity between perceived warmth and

relationship expectations, we performed a confirmatory factor

analysis. For each construct, the average variance extracted

exceeded .50 (perceived warmth ¼ .51, relationship expectations¼ .57). Fornell and Larcker’s (1981) test also revealed that both average variances extracted were higher than

the shared variance of .15, confirming that they represent dis-

tinct constructs.

Mediation analyses. We tested perceived warmth as a possible mediator with a bootstrapping analysis using PROCESS Model

4 (Preacher and Hayes 2008; see Figure 3). Results revealed

that the indirect effect of performance attributions on relation-

ship expectations through perceived warmth was significant

(indirect effect ¼ .20, SE ¼ .07, 95% confidence interval ¼ [.09, .36]).

Discussion

Study 2a shows that consumers expect a more communal rela-

tionship with a service employee when the employee’s perfor-

mance is attributed to effort rather than to talent, because they

perceive such an employee to be warmer. This study also estab-

lished discriminant validity between perceived warmth and

relationship expectations. Warmth judgment and communal

relationship expectation, though correlated, are conceptually

distinct constructs. Gershon and Cryder (2018) conceptually

and empirically separated warmth perceptions (perceptions of

a trait) and inferred communal intent (perceptions of a motive

behind a trait or action). Perceived warmth of a service

employee is a perceived trait of that employee that is not spe-

cific to a given service context, whereas a communal relation-

ship expectation involves the predicted norms in the

relationship with a service employee in a specific service con-

text. This study also shows that performance attributions do not

necessarily change perceived competence of the service

employee, which is in line with prior work suggesting that

effort and talent are two different types of attribution of com-

petence (Weiner 1972, 2005). In the next study, we test our

proposed mechanism by directly manipulating the warmth of

the employee.

Study 2b: Manipulating Service Employee Warmth

Study 2b uses a moderation-of-process strategy (Spencer,

Zanna, and Fong 2005) to manipulate the warmth of a service

employee to provide further evidence for warmth as a mediator

for the effect of performance attributions on relationship expec-

tations. If effort (vs. talent) attribution leads consumers to

expect a more communal (or less exchange-oriented) relation-

ship because the employee is perceived as warmer, information

signaling that the employee is warm should attenuate the pro-

posed effect. We employ a 2 (performance attribution: effort

vs. talent) � 2 (warmth: yes vs. no) between-subjects design.

Method

Three hundred seventy-one undergraduate students (233

women; mean age ¼ 20.30 years) from a large university in Hong Kong participated in this laboratory experiment. Partici-

pants read website information about a physician whose com-

petent performance (e.g., “Dr. Lee received a peer review

rating in the top 5% among general practitioners in Hong Kong”) was attributed to either his effort (e.g., “Dr. Lee spends

a lot of time [and] works really hard to develop personalized

health improvement programs”) or talent (e.g., “Dr. Lee has a

sharp instinct [and is] naturally skillful at developing persona-

lized health improvement programs”; see Web Appendix W6).

Warmth manipulation. To manipulate the warmth of the physi- cian, we provided additional information that can increase

warmth perceptions but is not directly related to the employee’s

behavior toward his or her customers. Warmth is particularly

relevant to the prosocial domain, because people rely on

warmth judgments to predict whether a person is well-

intentioned toward other people (Fiske et al. 2002). Thus, we

manipulated the warmth of a service employee by informing

Performance Attributions (Natural Talent vs. Dedicated Effort) .19** (.11n.s.)

.26*** .34***

Perceived Warmth

Relationship Expectations (Exchange vs. Communal)

Figure 3. Mediation analysis (Study 2a).

112 Journal of Marketing 84(3)

participants that the employee donates a part of his earnings to

various charity organizations. No such information was men-

tioned in the control condition.

To test the effectiveness of the warmth manipulation, we

conducted an independent pretest (n ¼ 170; 106 women; mean age ¼ 20.88 years). After reading the website information (excluding information on performance attributions), partici-

pants indicated the extent to which they perceived the physi-

cian to be warm and competent, as in Study 2a. A t-test

revealed that participants perceived the physician as warmer

in the warmth condition (M ¼ 5.11, SD ¼ .77) than in the no- warmth condition (M ¼ 4.44, SD ¼ 1.13; t(168) ¼ 4.45, p < .001, d ¼ .69). However, perceived competence did not differ across the two conditions (p ¼ .22).

Measures. We measured relationship expectations with the phy- sician as in the previous studies. We also measured partici-

pants’ expectations about the employee’s service process

quality with four items (e.g., “unfavorable/favorable,” “bad/

good”; a ¼ .93) and service outcome quality with four items (e.g., “unfavorable/favorable,” “bad/good”; a ¼ .92). Perfor- mance attribution manipulation did not change these expecta-

tions (all ps > .30).

Results

We ran a 2 (performance attributions: effort vs. talent) � 2 (warmth: yes vs. no) ANOVA on relationship expectations.

The results revealed a significant main effect of performance

attributions (F(1, 367) ¼ 8.07, p < .01, Z 2p ¼ .02), no signif- icant effect of warmth (F(1, 367) ¼ .49, p ¼ .48, Z 2p ¼ .001), and a significant interaction (F(1, 367) ¼ 4.07, p < .05, Z 2p ¼ .01). Planned contrasts revealed that our previous findings were

replicated in the no-warmth condition; specifically, effort attri-

bution (M ¼ 4.03, SD ¼ 1.15) induced more of a communal relationship expectation than did talent attribution (M ¼ 3.47, SD ¼ 1.05; t(367) ¼ 3.41, p ¼ .001, d ¼ .51), in support of H1. In contrast, this effect was attenuated in the warmth condition

(Meffort ¼ 3.88, SD ¼ 1.09 vs. Mtalent ¼ 3.79, SD ¼ 1.14; t(367) ¼ .59, p ¼ .56, d ¼ .08), in support of H2.

Discussion

In this study, we directly manipulated the mediating variable

(i.e., warmth of a service employee). The results support our

mechanism that effort (vs. talent) attribution leads consumers

to perceive a service employee to be warmer by showing that

information signaling that the employee is warm attenuates the

effect of performance attributions on relationship expectations.

This study also shows that performance attributions do not

change participants’ expectations about the employee’s service

process quality and service outcome quality, thus ruling these

out as possible alternative explanations for our proposed

effects. In the next study, we examine the effect of performance

attributions on customer attention.

Study 3: Customer Attention to Service Employee Information

To enhance the validity of our findings, Study 3 provides fur-

ther evidence for the effect of performance attributions by

using an alternative, more objective measure of relationship

expectations: consumers’ attention to service employee infor-

mation. We argue that consumers’ relationship expectations

can be manifested in their attention while reading advertise-

ments. Prior work has shown that under an exchange relation-

ship, individuals focus on their counterparts’ instrumental

function to ensure that the benefits they are to receive fulfill

their own goals (Abele and Wojciszke 2014; Aggarwal 2004).

Furthermore, Schroeder and Fishbach (2015) argue that when

individuals focus on others’ instrumental function, they tend to

overlook the facts relating to the personal lives and experiences

of those others. Therefore, we predict that if talent attribution

leads to greater expectation of an exchange relationship, con-

sumers will pay more attention to information pertaining to the

service employee’s instrumental function (e.g., what the ser-

vice employee can do for them) than to personal information

about the employee (e.g., personal background information).

We use an eye-tracking technique to capture participants’

attention toward service employee information, which allows

us to measure a subconscious or preconscious reflection of

relationship expectations (Plassmann and Mormann 2017).

Method

One hundred forty-seven undergraduate students (110 women;

mean age ¼ 20.82 years) from a large university in Hong Kong participated in this laboratory experiment. We used an eye-

tracking device, The Eye Tribe, powered by the software Gaze-

Lab (30 Hz), which collects raw eye movement data points

every 33.3 milliseconds. This eye tracker was integrated into

a 15.4-inch monitor at a resolution of 1,680 � 1,050 pixels. As participants viewed the stimuli shown on the screen, a discreet

infrared camera located below the screen unobtrusively

recorded participants’ attention.

Performance attribution manipulation. Participants were told that their university’s medical society was editing a newsletter, and

they were asked to read an article featuring an interview with a

physician from the university’s health clinic. On the first page

of the article, we manipulated performance attributions as in

Study 2b. When we defined the performance attribution infor-

mation as an area of interest (i.e., a selected region of the

stimulus of which eye-movement metrics are extracted), parti-

cipants in the two conditions did not differ in terms of the

attention they paid to the manipulation stimuli (Meffort ¼ 4.58 seconds, SD ¼ 3.70; Mtalent ¼ 4.23 seconds, SD ¼ 3.02; t(145) ¼ .61, p ¼ .54, d ¼ .10). We excluded any participants who did not fix their attention on the performance attribution informa-

tion because they were neither exposed to the effort nor the

talent attribution manipulation.

Leung et al. 113

To test the effectiveness of our manipulation, we conducted

an independent pretest (n ¼ 92; 67 women; mean age ¼ 20.60 years). After reading an article about the physician, participants

indicated the extent to which they thought the physician had

achieved his level of performance because of effort or talent,

using a semantic differential scale as in previous studies. Par-

ticipants in the effort attribution condition (M ¼ 2.41, SD ¼ 1.37) were more likely to attribute the physician’s perfor-

mance to effort than those in the talent attribution condition

(M ¼ 4.60, SD ¼ 1.53; t(90) ¼�7.26, p < .001, d ¼�1.51). Participants also indicated the extent to which they perceived

the physician to be warm and competent, as in Study 2a. We

also measured participants’ expectations about the overall

quality of the physician (1 ¼ “very bad,” and 7 ¼ “very good”). A t-test revealed that participants perceived the phy-

sician as warmer when his performance was attributed to

effort (M ¼ 5.42, SD ¼ .82) rather than to talent (M ¼ 5.07, SD ¼ .83; t(90) ¼ 2.04, p < .05, d ¼ .42). However, perceived competence and expected overall quality did not

differ across the two conditions (ps > .30).

Dependent variable and control variables. Participants were then presented with two columns of additional information about the

physician (see Web Appendix W7). One column presented

person-related information about the physician, such as the

physician’s background (e.g., “Dr. Lam is 32 years old and was

born and raised in Hong Kong”). The other presented job-

related information, such as information about what the physi-

cian could do for the participants (e.g., “Dr. Lam investigates

[students’] current health states and conducts physical exami-

nations to establish risk factor levels”). We counterbalanced

the presentation of each column. Each of the two columns of

service employee information was defined as a separate area of

interest. For each participant, we calculated the ratio of time

spent fixating on person-related information to the time spent

fixating on job-related information. Because this ratio was

positively skewed (skewness ¼ 8.55, SE ¼ .20; Shapiro– Wilk’s W ¼ .32, p < .001), we used the log-transformed ratio as the dependent measure. Moreover, we measured partici-

pants’ knowledge of health care services, mood, and arousal

as control variables and found that these variables did not differ

across conditions (all ps > .40).

Results

A 2 (performance attributions: effort vs. talent) � 2 (presenta- tion order: person-related information on the left vs. right)

ANOVA revealed that the log-transformed ratio of fixation

time was higher when the physician’s performance was attrib-

uted to effort (M ¼ .26, SD ¼ 1.12) rather than to talent (M ¼ �.05, SD ¼ 1.24; F(1, 143) ¼ 4.00, p < .05, Z 2p ¼ .03), in support of H1. Thus, when the performance was attributed to

effort (vs. talent), participants spent a relatively greater propor-

tion of time attending to the physician’s person-related infor-

mation than to the physician’s job-related information. The

main effect of the presentation order was significant; the log-

transformed ratio of fixation time was higher when person-

related information was presented on the left (M ¼ .63, SD ¼ 1.23) than on the right (M ¼ �.44, SD ¼ .85; F(1, 143) ¼ 38.53, p < .001, Z 2p ¼ .21), consistent with the tendency to read English text from left to right (Spalek and Hammad 2005).

However, the interaction between performance attributions and

presentation order was not significant (F(1, 143) ¼ .02, p ¼ .90, Z 2p < .001).

Discussion

This study validates the theoretical and managerial importance

of relationship expectations by showing that it can be reflected

in consumers’ attention to advertisements, not just in self-

reported relationship expectation measures. Specifically, effort

attribution leads consumers to spend a greater proportion of

time attending to person-related information compared with

job-related information about the service employee, consistent

with the norms of communal relationships. This study provides

practical insights on how to utilize performance attributions in

communication messages. For instance, firms often communi-

cate their service employees’ personal background information

to enhance consumers’ connection with the employees (Wang,

Hsu, and Fang 2009). Our findings suggest that in such a sit-

uation, firms can attribute their employees’ performance to

effort rather than to talent. We also showed that the observed

effects cannot be attributed to changes in competence or quality

perceptions. In the next section, we develop a hypothesis

regarding downstream consequences of relationship expecta-

tions and present two studies to provide empirical evidence

supporting the hypothesis.

Service Employee Performance Attributions and Customer Helping Behaviors

To demonstrate the managerial and practical importance of

service employee performance attributions, we examine down-

stream consumer behaviors resulting from relationship expec-

tations. Specifically, we examine customer helping behaviors

for firms as a result of relationship expectations. Drawing on

prior work, we define customer helping behaviors as voluntary

and discretionary behaviors toward firms that aid the firms

beyond those required in the purchase of products and services

(Bettencourt 1997; Bove et al. 2009), which can include

spreading WOM (e.g., sharing product/service information on

one’s social networks), providing suggestions for product and

service improvements, participating in firm activities, and

helping other customers (Bettencourt 1997; Garma and Bove

2011; Groth 2005). Although the link between relationship

expectations and customer helping behaviors has not been

directly tested, prior research has suggested that customers are

more likely to engage in helping behaviors when they believe a

service employee places the welfare of the customers above the

employee’s own immediate self-interest (Bove et al. 2009),

which is consistent with characteristics in communal relation-

ships (Aggarwal 2004). Therefore, we predict that when an

114 Journal of Marketing 84(3)

employee’s performance is attributed to effort, thus inducing

more of a communal relationship expectation, consumers will

have a higher likelihood of engaging in helpful behaviors.

Formally,

H3: Consumers are more likely to engage in customer

helping behaviors toward a firm when its service employ-

ees’ competent performance is attributed to dedicated

effort rather than to natural talent.

Prior research has identified both WOM and idea provision as

important customer helping behaviors that can promote firm

interests. Scholars have found that WOM can influence the way

consumers make purchase decisions and, thus, affect sales

(Babić Rosario et al. 2016), and that customers’ participation

in idea provision can enhance new product financial perfor-

mance (Chang and Taylor 2016). In the next two studies, we

test the effect of performance attributions on these two cus-

tomer helping behaviors. In Study 4, we used a real firm con-

text and measured individuals’ WOM behaviors. We show that

customers are more likely to help a firm share information on

social networks when the employees’ performance is attributed

to effort than to talent. In Study 5, we conducted a field experi-

ment to examine customers’ provision of new product ideas.

The findings indicate that effort attribution makes customers

more likely to provide new product ideas.

Study 4: Performance Attributions and WOM Behaviors

In Study 4, we explore WOM behaviors as a downstream con-

sequence of relationship expectations. We predict that when a

firm highlights its service employees’ dedicated effort (vs. nat-

ural talent), thus inducing a more communal-oriented relation-

ship expectation, customers will be more likely to share the

firm’s information on social networks.

Method

One hundred fifty-five undergraduate students (98 women;

mean age ¼ 20.21 years) from a large university in Hong Kong participated in this laboratory study for monetary compensa-

tion. To increase realism of the experimental context, we used a

real fitness center in Hong Kong, which operates in multiple

locations and offers two types of classes with trainers: one

combining yoga and fitness training, and the other combining

Thai boxing and fitness training.

Performance attribution manipulation. Participants were given website information about this fitness center and its trainers.

They were told that the fitness classes were instructed by a

team of highly qualified fitness trainers who have won awards

and championships in Hong Kong and overseas. We attributed

these performances to either effort (e.g., “A group of hardwork-

ing trainers . . . will dedicate their efforts”) or talent (e.g., “A group of talented trainers . . . have good natural skills”; see Web Appendix W8).

We also conducted an independent pretest (n ¼ 80; 55 women; mean age ¼ 20.69 years). Participants in the effort attribution condition (M ¼ 2.75, SD ¼ 1.37) were more likely to attribute the fitness trainers’ performance to effort than were

those in the talent attribution condition (M ¼ 4.89, SD ¼ 1.36; t(78) ¼ �6.97, p < .001, d ¼ �1.57), indicating that our manipulation was successful. Participants also indicated the

extent to which they perceived the fitness trainers to be warm

and competent, as in Studies 2a and 3, and how experienced the

trainers seemed to be (1 ¼ “not at all,” and 7 ¼ “very much”). A t-test revealed that participants perceived the fitness trainers

as warmer when their performance was attributed to effort (M

¼ 4.85, SD ¼ .92) than to talent (M ¼ 4.26, SD ¼ 1.25; t(78) ¼ 2.44, p < .05, d ¼ .54). However, perceived competence and experience did not differ across the two conditions (ps > .10).

Sharing of website on social networks (WOM behaviors). Partici- pants then read a message from the fitness trainers asking par-

ticipants for their help to share the fitness center’s website on

social networks. Following Consiglio, De Angelis, and Costa-

bile’s (2018) measure of WOM behaviors, participants were

led to believe that by clicking a share button, they would share

the website on a social network of their choice. After choosing

their favored social network(s), participants were informed that

they would not actually share the website. As an incentive,

customers who chose to share the website could enter a lucky

draw for a chance to win a free trial class at the fitness center

(worth HK$200 or US$25).

Measures. We measured relationship expectations with the fit- ness trainers as in the previous studies. We also measured

participants’ general tendency to share information on social

media (1 ¼ “never,” and 7 ¼ “very frequently”), which did not differ across conditions (p > .50).

Results

Sharing of website on social networks (WOM behaviors). A cross- tabulation analysis revealed that participants in the effort attri-

bution condition (58.97%) were more likely to share the fitness center’s website on social networks than those in the talent

attribution condition (42.86%; w2(1) ¼ 4.03, p < .05), in sup- port of H3.

Relationship expectations. A t-test analysis revealed that partici- pants expected a more communal relationship when the fitness

trainers’ performance was attributed to effort (M ¼ 3.93, SD ¼ 1.30) than to talent (M ¼ 3.31, SD ¼ 1.30; t(153) ¼ 2.95, p < .01, d ¼ .48), in support of H1.

Mediation analyses. We tested relationship expectations as a mediator for the effect of performance attributions on sharing

behavior with a bootstrapping analysis using PROCESS Model

4 (Preacher and Hayes 2008). Results revealed that the indirect

effect of performance attributions on sharing behavior through

relationship expectations was significant (indirect effect ¼ �.41, SE ¼ .19, 95% confidence interval ¼ [�.85, �.13]).

Leung et al. 115

Discussion

Using a real firm context, Study 4 offers important marketing

implications by examining WOM behaviors as a customer out-

come of relationship expectations. The findings support our

prediction that when a firm highlights its service employees’

dedicated effort as opposed to their natural talent, thus inducing

a more communal relationship expectation, customers are more

likely to engage in helpful behaviors by sharing the firm’s

information on social networks. Instead of providing an exact

performance level as in previous studies, we offered a descrip-

tion of the fitness trainers’ achievements (i.e., a team of highly

qualified fitness trainers who have won awards and champion-

ships) to generalize our findings. In this study, we also showed

that the trainers described as hardworking were perceived to be

warmer, but not more competent or more experienced, than

those described as talented. In the next study, we examine the

effect of performance attributions on another type of customer

helping behaviors.

Study 5: A Field Experiment on Customer New Product Idea Provision

In Study 5, we conducted a field experiment at the coffee shops

of an international coffee chain to test the effect of performance

attributions on customers’ provision of new product ideas. This

coffee chain employs user-design philosophies to generate new

product ideas through its website and has implemented many

crowdsourced ideas. We predict that customers will be more

likely to provide new product ideas when firms highlight their

service employees’ dedicated effort (vs. natural talent), thus

inducing more of a communal relationship expectation.

Method

Procedure. Over a two-week period, we launched a “Share Your Ideas” campaign (hereinafter, “campaign”) at two locations of

the coffee chain. In the shops, we prominently displayed mar-

keting materials (e.g., posters on walls, poster stands, table

stickers) highlighting the baristas’ dedicated effort for one

week, and those highlighting their natural talent for another

week. To control any confounding effects associated with par-

ticular dates, we simultaneously ran the campaign at two coffee

shops, each located in a different large university in Hong

Kong. We counterbalanced the performance attribution condi-

tions between the two shops (i.e., talent attribution condition in

Shop A and effort attribution condition in Shop B in the first

week, and vice versa in the second week).

We contracted a professional graphic designer to create the

campaign’s marketing materials (for sample materials, see

Web Appendix W9). The marketing materials in the effort

attribution condition highlighted the baristas’ effort (e.g., “We

are a group of hardworking baristas! Please share your bever-

age ideas with us. We put a lot of effort into creating perfectly

composed drinks”), whereas those in the talent attribution con-

dition highlighted the baristas’ talent (e.g., “We are a group of

talented baristas! Please share your beverage ideas with us. We

are naturally skillful in creating perfectly composed drinks”).

We displayed the marketing materials throughout the shops

(see Web Appendix W10).

Feedback forms. We placed feedback forms throughout the shops that customers could voluntarily pick up, fill out with

their ideas and suggestions, and submit to a collection box. The

feedback forms included a performance attribution manipula-

tion (for samples of the feedback forms, see Web Appendix

W11). We measured participants’ general liking of the coffee

chain (1 ¼ “not at all,” and 7 ¼ “very much”) and frequency of visits (1 ¼ “never,” and 7 ¼ “very frequent”) as control vari- ables. We also measured perception of the baristas’ beverage-

making skill level (1 ¼ “not good at all,” and 7 ¼ “very good”) to ensure that the performance attribution manipulation did not

lead to differences in perceived competence of the baristas. As

an incentive for their participation, customers who submitted a

feedback form could enter a lucky draw for a chance to win a

HK$300 (US$38) coffee chain coupon.

Results

Submission of feedback forms. To test the effect of performance attributions on customers’ likelihood of submitting a feedback

form, we examined the number of submitted feedback forms

as a percentage of the total number of sales transactions. We

obtained the numbers of weekly sales transactions of the two

coffee shops from their managers and found that the number

of total transactions was not significantly different across the

two shops.

We conducted three types of analyses on customers’ like-

lihood of submitting a feedback form. First, a cross-tabulation

analysis indicated that customers were more likely to submit a

feedback form when they were exposed to effort attribution

information than to talent attribution information (5.24% vs. 3.36%; w2(1) ¼ 40.21, p < .001), in support of H3. In addition, two separate analyses showed that the finding was consistent

for both Shop A (4.38% vs. 2.90%; w2(1) ¼ 16.99, p < .001) and Shop B (6.48% vs. 3.99%; w2(1) ¼ 24.49, p < .001).

Second, we ran a binary logistic regression of the submis-

sion of feedback forms (1 ¼ submitted, 0 ¼ not submitted) on performance attributions (dedicated effort vs. natural talent),

shop dummy (Shop A vs. Shop B), and their interaction. There

was a significant main effect of performance attributions (b ¼ �.51, SE ¼ .10, Wald(1) ¼ 24.04, p < .001, Exp(B) ¼ .60). Thus, customers in the effort attribution condition were more

likely to submit a feedback form than were those in the talent

attribution condition, in support of H3. There was a main effect

of the shop dummy (b ¼�.41, SE ¼ .09, Wald(1) ¼ 20.05, p < .001, Exp(B) ¼ .66) and a nonsignificant interaction (b ¼ .08, SE ¼ .15, Wald(1) ¼ .31, p ¼ .58). Two separate logistic regression analyses (one for each shop) indicated that custom-

ers in the effort attribution condition were more likely to submit

a feedback form than were those in the talent attribution con-

dition for both Shop A (b ¼�.43, SE ¼ .11, Wald(1) ¼ 16.76,

116 Journal of Marketing 84(3)

p < .001, Exp(B) ¼ .65) and Shop B (b ¼ �.51, SE ¼ .10, Wald(1) ¼ 24.04, p < .001, Exp(B) ¼ .60).

To further enhance the robustness of our findings, we

adopted the rare events logistic regression method (ReLogit;

King and Zeng 2001). Given that our binary event of interest

(i.e., submission of feedback forms) was relatively rare (4.31% of the sample), ReLogit corrects for rare event biases and stan-

dard error inconsistency, thus providing more accurate esti-

mates than traditional logistic regression models. The

ReLogit results were consistent with those from the logistic

regression models.

Number of suggestions provided. We further tested the effect of performance attributions on the number of suggestions pro-

vided. Two research assistants blind to the research hypotheses

independently counted the number of suggestions provided on

the submitted feedback forms (Cohen’s kappa ¼ .71, p < .001), and disagreements were resolved through discussion. They

were instructed to count only the related suggestions and

exclude suggestions unrelated to the given question on the

coffee chain’s beverage offerings (e.g., “I love you [the name

of the coffee chain]”).

We used a Poisson regression, because the dependent vari-

able was count data (Coxe, West, and Aiken 2009). We

regressed the number of suggestions on performance attribu-

tions, shop dummy, and their interaction. The results revealed

a significant main effect of performance attributions (b ¼ .28, SE ¼ .10, z ¼ 2.91, p < .01), in support of H3. There was also a significant main effect of shop dummy (b ¼�.23, SE ¼ .11, z ¼ �2.07, p < .05) and a significant interaction (b ¼ �.41, SE ¼ .15, z ¼ �2.73, p < .01). Split-group Poisson regres- sions showed that participants at Shop A provided a greater

number of suggestions in the effort attribution condition

(M ¼ 1.62, SD ¼ 1.33) than in the talent attribution condition (M ¼ 1.23, SD ¼ 1.21; b ¼ .28, SE ¼ .10, z ¼ 2.91, p < .01). However, the effect was not significant at Shop B (Meffort ¼ .86, SD ¼ 1.14; Mtalent ¼ .97, SD ¼ .95; b ¼�.13, SE ¼ .11, z ¼�1.13, p ¼ .26).

These effects persisted after we controlled for participants’

liking of the coffee chain and frequency of visits. Therefore, the

effects could not be attributed to individual differences in these

factors. Moreover, performance attributions did not change the

extent to which participants perceived the baristas to be skillful

(p ¼ .44).

Discussion

In a natural field setting, Study 5 shows that when a firm high-

lights its service employees’ dedicated effort (vs. natural

talent), customers (1) are more likely to submit feedback forms

and (2) provide a greater number of suggestions, though the

latter effect was significant at only one shop. Moreover, in this

study we did not provide the exact performance level so as to

generalize our findings, although we believe that customers

consider the coffee chain’s baristas to be a competent group

among coffee shop employees in general (especially among our

participants, who actually visited the coffee chain). In addition,

our performance attribution manipulation did not change per-

ceptions of the baristas’ beverage-making skill level. Thus, our

effect cannot be attributed to participants’ perception that bar-

istas depicted as naturally talented (vs. hardworking) were

more skillful and competent and, thus, were less likely to need

suggestions from customers. We replicated the findings in a

laboratory experiment, in which we also measured relationship

expectations (for details, see Web Appendix W12).

General Discussion

The current research demonstrates that message cues that attri-

bute a service employee’s competent performance to dedicated

effort (vs. natural talent) lead consumers to expect a more

communal and less exchange-oriented relationship due to an

increase in the perceived warmth of the employee. Study 1a

showed that participants expected a more communal relation-

ship with a service employee whose competent performance

was attributed to effort rather than to talent, whereas Study 1b

revealed that simultaneous attribution to both effort and talent

yields an effect similar to that of effort attribution only. In

directly measuring the perceived warmth of an employee,

Study 2a showed that the effect of performance attributions

on relationship expectations is mediated by this construct.

We manipulated the perceived warmth of an employee in Study

2b and showed further support for the mediating role of

warmth. In Study 3, we used eye-tracking technology and

found that effort attribution led participants to pay more atten-

tion to person- than job-related information about the service

employee, reflecting expectation of a more communal-oriented

relationship.

Studies 4 and 5 explored customer helping behaviors as

downstream consumer outcomes of relationship expectations.

In Study 4, we used a real firm context (i.e., fitness center) and

showed that participants were more likely to spread WOM for a

firm when its service employees’ performance was attributed to

effort than to talent. Finally, in Study 5, we conducted a field

experiment and showed that effort attribution, which induced a

more communal relationship expectation, made participants

more likely to provide new product ideas.

Theoretical Contributions

The marketing literature has focused on attributions of one’s

own performance and demonstrated their impact on brand or

product evaluations (Mathur, Block, and Yucel-Aybat 2014;

Murphy and Dweck 2016). The current research highlights the

importance of studying attributions of others’ performance,

because even for the same level of performance, people’s

beliefs about performance attribution can change judgments

of those others (Brown et al. 2018; Tsay 2016). For instance,

people expect hardworking others to perform better on novel

tasks (Brown et al. 2018). We suggest that it is also important to

understand the role of attributions of others’ performance in

consumer outcomes in service relationships, because how a

Leung et al. 117

customer views a service employee’s performance can deter-

mine the customer’s relationship with the employee (Bitner

1995). Therefore, this study fills the gap in prior work by

examining how attributions of service employees’ performance

influence consumers’ relationship expectations with and beha-

viors toward the service employees.

The current research also augments existing knowledge on

the two fundamental dimensions of social judgment—compe-

tence and warmth—by linking the warmth–competence frame-

work (Fiske et al. 2002) with the literature on performance

attribution (Dweck 2000; Weiner 2005). Prior work has inves-

tigated relationships between judgments of competence and

warmth (Judd et al. 2005; Yzerbyt, Kervyn, and Judd 2008),

mainly by examining how a certain level of competence is

related to warmth perception. Extending the existing literature,

the current research examines the attributions of competence as

a new dimension of competence influencing warmth percep-

tion, holding the objective level of competence constant.

The current research also enriches the existing literature by

identifying performance attributions as an antecedent of rela-

tionship expectations. The marketing literature has focused

mainly on the downstream consequences of a communal versus

exchange relationship with consumers—for example, whether

consumers’ perceptions of a communal versus exchange rela-

tionship influences their evaluation of brands (Aggarwal 2004;

Aggarwal and Law 2005), loss aversion tendency (Aggarwal

and Zhang 2006), and responses to service failures (Wan, Hui,

and Wyer 2011). However, given the lack of research on the

antecedents of relationship expectations, marketers may have

little practical guidance on how they can shape expectations

about a particular type of relationship in the minds of consu-

mers. Addressing this gap, we find that the attributions of ser-

vice employees’ competence can alter consumers’ expectations

about their relationships with the employees along the commu-

nal–exchange continuum. In addition, the current research sug-

gests that relationship expectations can be reflected in

consumers’ attention, not just in self-reported relationship

expectation measures. Our use of eye-tracking technology

allowed us to measure the subconscious or preconscious reflec-

tion of relationship expectations.

In addition, the current research contributes to the literature

on customer helping behaviors by identifying performance

attributions as a new antecedent of such behaviors (Bettencourt

1997; Garma and Bove 2011; Groth 2005). As customer help-

ing behaviors (e.g., spreading WOM, providing new product

ideas) are becoming notable marketing goals for brands and

firms, factors that encourage such behaviors are both theoreti-

cally and managerially important. We have shown that effort

attribution, as opposed to talent attribution, increases the like-

lihood of customer helping behaviors.

Marketing Implications

Our findings offer practical implications, because firms can

highlight either effort or talent as the primary source of service

employees’ competent performance to induce a relationship

expectation that corresponds to their service propositions. For

instance, firms that emphasize communality in their services

(e.g., Disneyland, Starbucks) can attribute their employees’

performance to effort, leading consumers to expect a more

communal relationship with their employees. In contrast, if

these firms attribute employee performance to talent, thus indu-

cing a more exchange relationship expectation, the discrepancy

between consumers’ relationship expectations and their actual

service experience may hurt service satisfaction.

Our findings also demonstrate that, depending on whether a

firm attributes its service employees’ performance to effort or

talent, consumers will pay attention to different types of service

employee information, reflecting their expected relationships

with the employees. This helps guide firms in designing their

marketing materials. For example, when firms want their

consumers to pay attention to a service employee’s personal

(job-related) information, they might want to attribute the

employee’s performance to effort (talent).

Moreover, this research shows that the effect of perfor-

mance attributions on relationship expectations has conse-

quences for customer helping behaviors that offer managerial

insights. Specifically, we gathered empirical evidence suggest-

ing that marketers can implement effort or talent attributions in

their communication messages to influence customers’ actual

WOM and idea provision behaviors. Marketers regard WOM—

electronic WOM in particular—as “one of the most significant

developments in contemporary consumer behavior” due to its

ability to influence the way consumers make purchase deci-

sions and affect sales (Babić Rosario et al. 2016, p. 297). Mar-

keters are also increasingly involving customers in idea

generation for new products, because such a tactic can enhance

new product financial performance (Chang and Taylor 2016).

As firms strive to achieve these marketing goals, our research

findings offer insights into how firms can motivate these cus-

tomer helping behaviors using their communications messages.

According to our findings, firms are advised to attribute their

employees’ performance to effort, rather than talent, when they

want to encourage customers to share firm information on

social networks or to suggest new products or services. We

believe our proposed effect of performance attributions on rela-

tionship expectations can also influence other types of cus-

tomer helping behaviors, such as participating in firm

activities and helping other customers.

What factors shape consumers’ expectations about their

relationship with a service employee is an important practical

question, because it can have significant consequences on con-

sumer outcomes (Aggarwal 2004; Aggarwal and Law 2005;

Wan, Hui, and Wyer 2011). Although it is true that firms can

develop communal relationships through other methods—for

example, by generally treating customers well and satisfying

them—these tactics require actual interactions with customers.

The current research suggests that communication messages

that do not involve interactions with customers also can move

customers’ relationship expectations along the communal–

exchange continuum, in turn influencing consumer behaviors.

Marketing practitioners can utilize this knowledge about

118 Journal of Marketing 84(3)

highlighting effort and/or talent to design their website com-

munications, print advertisements, and social media strategy

going forward, or to reevaluate the effectiveness of their cur-

rent communication strategies.

Future Research

This article offers several fruitful directions for future research.

First, future studies can examine whether the performance attri-

bution effects can be extended to other contexts. For instance,

our proposed effects may not be limited to person perception.

Because people tend to view a relationship with a brand, prod-

uct, or firm similarly to a relationship with a person (Fournier

1998; MacInnis and Folkes 2017), the attributions of brands’ or

firms’ competent performance might influence consumers’

perceived relationships with those brands or firms. Future stud-

ies could also explore service failure contexts. For example,

researchers can investigate whether attributing poor service

performance or negative service outcomes to an employee’s

lack of effort (or natural talent) can lead to differences in a

consumer’s willingness to forgive. In addition, future studies

could explore how consumers might interpret information on

performance attributions of firms whose performance is uncer-

tain (e.g., startups).

Even though our last two studies show that effort (vs. talent)

attribution is more likely to increase customer helping beha-

viors, we do not argue that effort attribution is always more

beneficial to firms than talent attribution. In a supplementary

study (Web Appendix W13), we measured membership sign-

up behavior as a different downstream behavior in the same

fitness training context. The findings show that because cus-

tomers who generally do not want to proactively interact with

service employees during a service process (e.g., by offering

their own opinions about the training program) prefer a more

exchange-oriented (i.e., less communal-oriented) relationship

with a service employee, firms are more likely to acquire them

if the service employee’s performance is attributed to talent

rather than effort. Future studies could explore other conse-

quences of service employees’ performance attributions for

consumer behaviors, such as loyalty to the same service

employee and reactions to service recovery, as well as other

individual and situational factors that influence customers’

relationship preferences.

In addition, future research could explore how relationship

expectations may interact with actual service experience to

affect customer satisfaction. For instance, customers who expe-

rienced an exchange-oriented relationship with a service

employee in digital interactions may be less satisfied with the

same experience when they are exposed to effort attribution

(vs. talent attribution) that induces a more communal relation-

ship expectation. Future research could also explore consumer

heterogeneity in terms of attributions of service employee per-

formance. Because the focal point of this research was to

delineate the effects of service firms’ performance attributions,

we did not directly explore consumers’ heterogeneity in their

attributions, which might depend on the industry or context.

This heterogeneity may be presumed to interact with service

firms’ endogenous attribution decisions.

Associate Editor

Wayne Hoyer

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to

the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for

the research, authorship, and/or publication of this article: This

research is supported by a grant from the Hong Kong SAR Research

Grants Council awarded to the second author (HKU17500715).

ORCID iDs

Fine F. Leung https://orcid.org/0000-0002-7988-2944

Sara Kim https://orcid.org/0000-0003-0105-0184

Caleb H. Tse https://orcid.org/0000-0003-0572-8829

References

Aaker, Jennifer, Kathleen D. Vohs, and Cassie Mogilner (2010),

“Nonprofits Are Seen as Warm and For-Profits as Competent:

Firm Stereotypes Matter,” Journal of Consumer Research, 37

(2), 224–37.

Abele, Andrea E. and Bogdan Wojciszke (2014), “Communal and

Agentic Content in Social Cognition: A Dual Perspective Model,”

Advances in Experimental Social Psychology, 50, 195–255.

Aggarwal, Pankaj (2004), “The Effects of Brand Relationship Norms

on Consumer Attitudes and Behavior,” Journal of Consumer

Research, 31 (1), 87–101.

Aggarwal, Pankaj and Sharmistha Law (2005), “Role of Relationship

Norms in Processing Brand Information,” Journal of Consumer

Research, 32 (3), 453–64.

Aggarwal, Pankaj and Meng Zhang (2006), “The Moderating Effect of

Relationship Norm Salience on Consumers’ Loss Aversion,” Jour-

nal of Consumer Research, 33 (3), 413–19.

BabićRosario, Ana, Francesca Sotgiu, Kristine DeValck, and Tammo

H.A. Bijmolt (2016), “The Effect of Electronic Word of Mouth on

Sales: A Meta-Analytic Review of Platform, Product, and Metric

Factors,” Journal of Marketing Research, 53 (3), 297–318.

Baudson, Tanja Gabriele and Franzis Preckel (2013), “Teachers’

Implicit Personality Theories About the Gifted: An Experimental

Approach,” School Psychology Quarterly, 28 (1), 37–46.

Bettencourt, Lance A. (1997), “Customer Voluntary Performance:

Customers as Partners in Service Delivery,” Journal of Retailing,

73 (3), 383–406.

Bitner, Mary Jo (1995), “Building Service Relationships: It’s All

About Promises,” Journal of the Academy of Marketing Science,

23 (4), 246–51.

Bove, Liliana L., Simon J. Pervan, Sharon E. Beatty, and Edward Shiu

(2009), “Service Worker Role in Encouraging Customer Organi-

zational Citizenship Behaviors,” Journal of Business Research, 62

(7), 698–705.

Leung et al. 119

Brown, Christina M., Nicole S. Troy, Katie R. Jobson, and Jennifer K.

Link (2018), “Contextual and Personal Determinants of Preferring

Success Attributed to Natural Talent or Striving,” Journal of

Experimental Social Psychology, 78, 134–47.

Chang, Woojung and Steven A. Taylor (2016), “The Effectiveness of

Customer Participation in New Product Development: A Meta-

Analysis,” Journal of Marketing, 80 (1), 47–64.

Clark, Margaret S. and Judson Mills (1993), “The Difference Between

Communal and Exchange Relationships: What It Is and Is Not,”

Personality and Social Psychology Bulletin, 19 (6), 684–91.

Consiglio, Irene, Matteo De Angelis, and Michele Costabile (2018),

“The Effect of Social Density on Word of Mouth,” Journal of

Consumer Research, 45 (3), 511–28.

Coxe, Stefany, Stephen G. West, and Leona S. Aiken (2009), “The

Analysis of Count Data: A Gentle Introduction to Poisson Regres-

sion and Its Alternatives,” Journal of Personality Assessment, 91

(2), 121–36.

Dweck, Carol S. (2000), Self-Theories: Their Role in Motivation,

Personality and Development. Philadelphia: Psychology Press.

Dweck, Carol S. and Ellen L. Leggett (1988), “A Social-Cognitive

Approach to Motivation and Personality,” Psychological Review,

95 (2), 256–73.

Emerson, Katherine T.U. and Mary C. Murphy (2015), “A Company I

Can Trust? Organizational Lay Theories Moderate Stereotype

Threat for Women,” Personality and Social Psychology Bulletin,

41 (2), 295–307.

Fiske, Susan T., Amy J.C. Cuddy, Peter Glick, and Jun Xu (2002), “A

Model of (Often Mixed) Stereotype Content: Competence and

Warmth Respectively Follow from Perceived Status and

Competition,” Journal of Personality and Social Psychology, 82

(6), 878–902.

Fornell, Claes and David F. Larcker (1981), “Structural Equation

Models with Unobservable Variables and Measurement Error:

Algebra and Statistics,” Journal of Marketing Research, 18 (3),

382–88.

Fournier, Susan (1998), “Consumers and Their Brands: Developing

Relationship Theory in Consumer Research,” Journal of Consumer

Research, 24 (4), 343–53.

Fuchs, Stephan (2001), “Beyond Agency,” Sociological Theory, 19

(1), 24–40.

Garma, Romana and Liliana L. Bove (2011), “Contributing to Well-

Being: Customer Citizenship Behaviors Directed to Service

Personnel,” Journal of Strategic Marketing, 19 (7), 633–49.

Geake, JohnG. and Miraca U.M. Gross (2008), “Teachers’ Negative

Affect Toward Academically Gifted Students: An Evolutionary

Psychological Study,” Gifted Child Quarterly, 52 (3), 217–31.

Gershon, Rachel and Cynthia Cryder (2018), “Goods Donations

Increase Charitable Credit for Low-Warmth Donors,” Journal of

Consumer Research, 45 (2), 451–69.

Groth, Markus (2005), “Customers as Good Soldiers: Examining Citi-

zenship Behaviors in Internet Service Deliveries,” Journal of Man-

agement, 31 (1), 7–27.

Harris, Lasana T. and Susan T. Fiske (2006), “Dehumanizing the

Lowest of the Low—Neuroimaging Responses to Extreme Out-

Groups,” Psychological Science, 17 (10), 847–53.

Haslam, Nick (2006), “Dehumanization: An Integrative Review,”

Personality and Social Psychology Review, 10 (3), 252–64.

Haslam, Nick and Steve Loughnan (2014), “Dehumanization and

Infrahumanization,” Annual Review of Psychology, 65, 399�423. Hong, Huang-Yao and Xiaodong Lin-Siegler (2012), “How Learning

About Scientists’ Struggles Influences Students’ Interest and

Learning in Physics,” Journal of Educational Psychology, 104

(2), 469–84.

Hong, Ying-Yi, Chi-Yue Chiu, Carol S. Dweck, Derrick M.-S. Lin,

and Wendy Wan (1999), “Implicit Theories, Attributions, and

Coping: A Meaning System Approach,” Journal of Personality

and Social Psychology, 77 (3), 588�99. Judd, Charles M., Laurie James-Hawkins, Vincent Yzerbyt, and

Yoshihisa Kashima (2005), “Fundamental Dimensions of Social

Judgment: Understanding the Relations Between Judgments of

Competence and Warmth,” Journal of Personality and Social Psy-

chology, 89 (6), 899–913.

Kervyn, Nicolas, Susan T. Fiske, and Chris Malone (2012), “Brands as

Intentional Agents Framework: How Perceived Intentions and

Ability Can Map Brand Perception,” Journal of Consumer Psy-

chology, 22 (2), 166–76.

King, Gary and Langche Zeng (2001), “Logistic Regression in Rare

Events Data,” Political Analysis, 9 (2), 137–63.

Klein, Nadav and O’Brien Ed (2017), “The Power and Limits of

Personal Change: When a Bad Past Does (and Does Not) Inspire

in the Present,” Journal of Personality and Social Psychology, 113

(2), 210–29.

Kwak, Hyokjin, Marina Puzakova, and Joseph F. Rocereto (2015),

“Better Not Smile at the Price: The Differential Role of Brand

Anthropomorphization on Perceived Price Fairness,” Journal of

Marketing, 79 (4), 56–76.

Leyens, Jacques-Philippe, Paola M. Paladino, Ramon Rodriguez-

Torres, Jeroen Vaes, Stéphanie Demoulin, and Armando

Rodriguez-Perez, et al. (2000), “The Emotional Side of Prejudice:

The Attribution of Secondary Emotions to Ingroups and Out-

groups,” Personality and Social Psychology Review, 4 (2), 186–97.

Lin-Siegler, Xiaodong, Janet N. Ahn, Jondou Chen, Fu-Fen Anny

Fang, and Myra Luna-Lucero (2016), “Even Einstein Struggled:

Effects of Learning About Great Scientists’ Struggles on High

School Students’ Motivation to Learn Science,” Journal of Edu-

cational Psychology, 108 (3), 314–28.

MacInnis, Deborah J. and Valerie S. Folkes (2017), “Humanizing

Brands: When Brands Seem to Be Like Me, Part of Me, and in a

Relationship with Me,” Journal of Consumer Psychology, 27 (3),

355–74.

Mathur, Pragya, Lauren Block, and Yucel-Aybat Ozge (2014), “The

Effects of Goal Progress Cues: An Implicit Theory Perspective,”

Journal of Consumer Psychology, 24 (4), 484–96.

McGraw, A. Peter, Janet A. Schwartz, and Philip E. Tetlock (2012),

“From the Commercial to the Communal: Reframing Taboo

Trade-Offs in Religious and Pharmaceutical Marketing,” Journal

of Consumer Research, 39 (1), 157–73.

Murphy, Mary C. and Carol S. Dweck (2016), “Mindsets Shape Con-

sumer Behavior,” Journal of Consumer Psychology, 26 (1),

127–36.

120 Journal of Marketing 84(3)

Persson, Roland S. (2007), “The Myth of the Antisocial Genius: A

Survey Study of the Socio-Emotional Aspects of High-IQ

Individuals,” Gifted and Talented International, 22 (2), 19�34. Plassmann, Hilke and Milica Mormann (2017), “An Interdisciplinary

Lens on Consciousness: The Consciousness Continuum and How to

(Not) Study It in the Brain and the Gut, A Commentary on Williams

and Poehlman,” Journal of Consumer Research 44 (2), 258–65.

Preacher, Kristopher J. and Andrew F. Hayes (2008), “Asymptotic and

Resampling Strategies for Assessing and Comparing Indirect

Effects in Multiple Mediator Models,” Behavior Research Meth-

ods, 40 (3), 879–91.

Rosenberg, Seymour, Carnot Nelson, and P.S. Vivekananthan (1968),

“A Multidimensional Approach to the Structure of Personality

Impressions,” Journal of Personality and Social Psychology, 9

(4), 283–94.

Schroeder, Juliana and Ayelet Fishbach (2015), “The ‘Empty Vessel’

Physician: Physicians’ Instrumentality Makes Them Seem Person-

ally Empty,” Social Psychological and Personality Science, 6 (8),

940�49. Scott, Maura L., Martin Mende, and Lisa E. Bolton (2013), “Judging

the Book by Its Cover? How Consumers Decode Conspicuous

Consumption Cues in Buyer–Seller Relationships,” Journal of

Marketing Research, 50 (3), 334–47.

Spalek, Thomas M. and Sherief Hammad (2005), “The Left-to-Right

Bias in Inhibition of Return is Due to the Direction of Reading,”

Psychological Science, 16 (1), 15–18.

Spencer, Steven J., Mark P. Zanna, and Geoffrey T. Fong (2005),

“Establishing a Causal Chain: Why Experiments Are Often More

Effective Than Mediational Analyses in Examining Psychological

Processes,” Journal of Personality and Social Psychology, 89 (6),

845–51.

Suitner, Caterina and Anne Maass (2008), “The Role of Valence in the

Perception of Agency and Communion,” European Journal of

Social Psychology, 38 (7), 1073–82.

Tsay, Chia-Jung (2016), “Privileging Naturals over Strivers: The

Costs of the Naturalness Bias,” Personality and Social Psychology

Bulletin, 42 (1), 40–53.

Tsay, Chia-Jung and Mahzarin R. Banaji (2011), “Naturals and Stri-

vers: Preferences and Beliefs About Sources of Achievement,”

Journal of Experimental Social Psychology, 47 (2), 460–65.

Wan, Lisa C., Michael K. Hui, and Robert S. Wyer Jr. (2011), “The

Role of Relationship Norms in Responses to Service Failures,”

Journal of Consumer Research, 38 (2), 260–77.

Wang, Chao-Hung, Li-Chang Hsu, and Shyh-Rong Fang (2009),

“Constructing a Relationship-Based Brand Equity Model,” Service

Business, 3, 275–92.

Weiner, Bernard (1972), “Attribution Theory, Achievement Motiva-

tion, and the Educational Process,” Review of Educational

Research, 42 (2), 203–15.

Weiner, Bernard (2005), “Motivation from an Attribution Perspective

and the Social Psychology of Perceived Competence,” in Hand-

book of Competence and Motivation, Andrew J. Elliot and Carol S.

Dweck, eds. New York: Guilford Publications.

Yzerbyt, Vincent, Nicolas Kervyn, and Charles M. Judd (2008),

“Compensation Versus Halo: The Unique Relations Between the

Fundamental Dimensions of Social Judgment,” Personality and

Social Psychology Bulletin, 34 (8), 1110–23.

Yzerbyt, Vincent, Valerie Provost, and Olivier Corneille (2005), “Not

Competent but Warm . . . Really? Compensatory Stereotypes in the

French-Speaking World,” Group Processes & Intergroup Rela-

tions, 8 (3), 291–308.

Leung et al. 121

Copyright of Journal of Marketing is the property of American Marketing Association and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.