y Do Donors Donate? Examining the Effects of Organizational Identification and Identity Salience on the Relationships among Satisfaction, Loyalty, and Donation Behavior

y Do Donors Donate? Examining the Effects of Organizational Identification and Identity Salience on the Relationships among Satisfaction, Loyalty, and Donation Behavior

Article

Why Do Donors Donate? Examining the Effects of Organizational Identification and Identity Salience on the Relationships among Satisfaction, Loyalty, and Donation Behavior

Silke Boenigk1 and Bernd Helmig2

Abstract With an empirical study in two nonprofit industries (a money-collecting and blood-collecting organization), the authors investigate how organizational identification and identity salience together function in relation to satisfaction, loyalty, and behavior. They develop and test a model that best represents relationships featuring donor-nonprofit identification and donor identity salience in existing satisfaction-loyalty studies. Overall, the study empirically confirms that donor-nonprofit identification and donor iden- tity salience are distinct constructs and that both have direct positive effects on loyalty, but not that much on donations. Within the money donation context, both identification constructs have stronger total effects on donor loyalty than donor satisfaction, whereas in the blood donation context, donor satisfaction has a stronger effect on loyalty. In testing the causal direction between donor-nonprofit identification and donor satisfaction, the authors also find that the path should be conceptualized from satisfac- tion to identification. The study contributes to the theory of organizational identification and identity salience by highlighting the advantages of taking a combined theoretical approach. Finally, the study suggests several means to implement donor identification management, including group activities, development of online communities, donor events, and more long-term-oriented tactics, all of which treat the donor as a cocreator of value.

Keywords organizational identification, identity salience, customer-company identification, donor-nonprofit identification, donor identity salience, nonprofit services, donor relationship management

In the past decade, satisfaction and loyalty studies have moved

beyond a first-generation perspective, which analyzes direct

links of satisfaction, loyalty, and profit, to a more complex

approach by including psychological constructs (Sen and

Bhattacharya 2001). For example, Homburg, Wieseke, and

Hoyer (2009) assess customer-company identification and cus-

tomer satisfaction simultaneously and find that identification

drives organizational outcomes and financial performance.

As such, they call for more investigations of ‘‘the incremental

explanatory power of customer–company identification

beyond the influence of customer satisfaction’’ (p. 48). In

response, we seek a deeper understanding of identification-

based constructs and their influence on the relationships among

satisfaction, loyalty, and behavior.

From a comprehensive review of service management

research, we find that knowledge about identification constructs

mainly encompasses two research streams. The first stream

focuses on organizational identification, defined as people’s

perceptions of belonging to a group (for details, see Ashforth,

Harrison, and Corley 2008), in selected service settings. For

example, Bhattacharya, Rao, and Glynn (1995) analyze mem-

bers’ identification with an art museum, and Gwinner and

Swanson (2003) examine the impact of high identification

levels among sports fans for sponsorship outcomes. Other stud-

ies have analyzed the identification of customers and/or

employees with a service provider and its impact on loyalty and

outcomes (e.g., Homburg, Wieseke, and Hoyer 2009; Lichten-

stein, Drumwright, and Braig 2004; Netemeyer, Heilman, and

Maxham 2012).

The second stream focuses on identity salience (Chattara-

man, Lennon, and Rudd 2010), defined as ‘‘the relative impor-

tance or centrality of given identity (and thus role) for defining

1 University of Hamburg, Hamburg, Germany 2 University of Mannheim, Mannheim, Germany

Corresponding Author:

Silke Boenigk, University of Hamburg, Von-Melle-Park 5, 20146 Hamburg,

Germany.

Email: silke.boenigk@wiso.uni-hamburg.de

Journal of Service Research 16(4) 533-548 ª The Author(s) 2013 Reprints and permission: sagepub.com/journalsPermissions.nav DOI: 10.1177/1094670513486169 jsr.sagepub.com

oneself’’ (Hoelter 1983, p. 141). For example, Laverie and

Arnett (2000) confirm the importance of identity salience for

sports fans’ behavior, and Arnett, German, and Hunt (2003)

analyze the impact of university identity salience on alumni

donation behavior. Thus, service management literature predo-

minantly features two separate evaluations of identity-based

concepts, and only two articles explicitly combine both

constructs (i.e., Bhattacharya and Sen 2003; Marin, Ruiz, and

Rubio 2008).

Therefore, a first research gap in the limited conceptual and

empirical knowledge is how organizational identification and

identity salience together function in relation to satisfaction,

loyalty, and behavior. Second, we find incomplete information

on the existence, causal direction, and impact of organizational

identification and identity salience on satisfaction, loyalty, and

behavior. In particular, although previous research has

confirmed a link between customer orientation and customer-

company identification (Homburg, Wieseke, and Hoyer 2009),

no studies have addressed the path from customer orientation

to identity salience. The findings on the link between satisfac-

tion and organizational identification also remain contradic-

tory. Mael and Ashforth (1992) report that satisfaction with

the school among alumni has a positive impact on identifica-

tion, but Camarero and Garrido (2011) posit a reverse causal

link, from organizational identification to satisfaction, and both

studies offer empirical evidence.

A third research gap entails the limited and contradictory

knowledge about the role of identity salience. Arnett, German,

and Hunt (2003) offer empirical support for a direct link

between university identity salience and donations by alumni

to their former university. In contrast, Bhattacharya and Sen

(2003) argue conceptually that identity salience moderates the

relationship between identity attractiveness and customer-

company identification, with positive effects on customer

loyalty, company promotion, and customer recruitment. Marin,

Ruiz, and Rubio (2008) analyze both constructs empirically

and find that identity salience moderates the relationship

between consumer-company identification and bank customer

loyalty. These conflicting outcomes leave unanswered the

question whether identity salience should be conceptualized

in service management studies as a moderator.

To the best of our knowledge, no previous study has

combined the two constructs and tested the relevant effects in

a nonprofit relationship context. Therefore, we pursue our

study in two nonprofit sector industries—namely, a nonprofit

organization working to collect money for a mission to prevent

cancer and a blood donation service provider. The nonprofit

sector is an important research setting for several reasons. First,

as Arnett, German, and Hunt (2003) argue, in relationships that

are social in nature, psychological constructs other than satis-

faction are likely relevant, so identification demands particular

consideration in this context. Second, Ahearne, Bhattacharya,

and Gruen (2005) note that people are highly likely to identify

with nonprofit organizations, so focusing on donor-nonprofit

relationships can provide clear empirical answers to

identification-based research questions. Third, donor-nonprofit

relationships are complex, in that in many nonprofit organiza-

tions, people play dual, simultaneous roles as members and as

private donors (Fombelle et al. 2012; Hogg, Terry, and White

1995). Thus, a nonprofit setting features both membership and

individual relationship elements. Finally, deeper knowledge

about the relevance of identification-based constructs in differ-

ent nonprofit relationships could help nonprofit managers

implement successful donor acquisition, retention, and recov-

ery strategies (Helmig and Thaler 2010). With these arguments,

we aim to answer three main research questions:

Research Question 1: How do donor-nonprofit identification

and donor identity salience together function in

satisfaction-loyalty studies?

Research Question 2: Which conceptual model best repre-

sents both new and existing relationships featuring the

two identification constructs in relation to donor orienta-

tion, satisfaction, loyalty, and donation behavior?

Research Question 3: What explanatory power does donor-

nonprofit identification and donor identity salience have,

beyond the influence of donor satisfaction, for donor

loyalty and donor behavior?

In the next section, we propose a conceptual framework to rep-

resent the unexplored and underexplored links we introduce.

We test this framework in two nonprofit industries because

we expect differences in the degree of donor identification with

a money-collecting or a blood-collecting service organization.

After presenting the empirical results from the proposed model,

we compare our findings with three rival models. We conclude

with a discussion of our findings in the two service contexts and

derive several implications for theory, measurement, and non-

profit practice.

Conceptual Framework and Hypotheses

We developed a conceptual framework that represents unex-

plored and underexplored relationships pertaining to identifica-

tion in prior satisfaction-loyalty studies. It features three

simplifying elements. First, we use orientation, satisfaction,

and loyalty and thereby include the most important relationship

drivers of purchase behavior identified by previous service

management studies (Mittal and Kamakura 2001). We adapt

these constructs to the donor-nonprofit context, with money

donations on the one hand and blood donations on the other

hand as the final outcome variables. The selection of the two

contexts is not incidental, but rooted in the different character-

istics of these donation settings. Monetary giving is an organi-

zational process that tends to be relatively transactional and

bureaucratic; the donor must provide bank account informa-

tion, an address, and other such data. In contrast, giving blood

is connected with a very individual, personal intervention and

is an existential issue, because saving lives is the primary goal

of this social exchange, but it also entails bodily contact, fear of

needles, and indisposition (Lee, Piliavin, and Call 1999). Sec-

ond, we model donor-nonprofit identification as an additional

534 Journal of Service Research 16(4)

driver of donor loyalty and donations. Third, we expand the

framework by including donor identity salience, which we

expect has direct effects on donor loyalty and donations. In

contrast to previous research, we do not conceptualize a

moderating effect of identity salience but rather a bidirectional

relationship between donor-nonprofit identification and donor

identity salience. However, we also test for a potential moder-

ating effect with a rival model. Finally, and as mentioned pre-

viously, we conceptualize and test a bidirectional relationship

between donor-nonprofit identification and donor satisfaction.

Because the links between satisfaction and loyalty have

been well explored in various service contexts, we do not

derive explicit hypotheses for all paths of the framework.

Instead, in line with our focus on donor-nonprofit identification

and donor identity salience, we develop five hypotheses

pertaining to the new and underexplored relationships in the

framework.

Relationship Between Donor-Nonprofit Identification and Donor Identity Salience. Existing knowledge about the correlation and causal links between donor-nonprofit identification and donor

identity salience is somewhat vague. The two studies that com-

bine the two constructs in one framework conceptualize and

measure identity salience as a moderator. In their conceptual

study, Bhattacharya and Sen (2003, p. 82) conceptualize

identity salience as a moderator of the link between identity

attractiveness and customer-company identification. Marin,

Ruiz, and Rubio (2008) test three different moderating effects

of identity salience and find support for this conceptualization.

We argue that both constructs may be directly linked and

therefore that a bidirectional rather than a moderating effect

exists. Our assumption is in line with the two underlying

theories relevant in this study—namely, identity theory, which

explains individual behaviors in relation to the self and society

(Stryker and Burke 2000), and social identity theory, which

helps explain group processes and intergroup relationships

(e.g., Tajfel 1974). Both theories assume that identification

positively affects individual behavior (e.g., positive word of

mouth, buying, donating), though identity theory argues that

it results from role salience, whereas social identity theory indi-

cates that it derives from the alignment with group norms

(Hogg, Terry, and White 1995; Stets and Burke 2000). Conse-

quently, both identification constructs should be linked directly

to donor loyalty, and thus identity salience should not be inter-

preted as a moderator. The idea of bidirectionality is based on

organizational research, which suggests that identification

should be regarded as ‘‘a cycle that iterates between organiza-

tional sensebreaking and sensegiving and individual identity

enactment, sensemaking, and identity narrative construction’’

(Ashforth, Harrison, and Corley 2008, p. 359). This cycle argu-

ment indicates that group and individual processes of identifi-

cation are dynamic and interconnected, so bidirectionality is

plausible.

If a bidirectional relationship exists, one causal direction

likely dominates in most cases. For this study, we predict that

the direct link from donor identity salience to donor-nonprofit

identification is dominant for the following reasons: First, for

most people, individual identification processes and benefits

are more important than group processes (Frumkin and

Andre-Clark 2000). Second, according to identity theory and

social identity theory, the two constructs differ in their levels

of stability. In identity theory, donor identity salience is a

relatively stable construct; changes to the identity hierarchy

represent responses to a change in the person’s role. For exam-

ple, if a person’s role changes from being a money donor to

becoming a volunteer or employee of the nonprofit organiza-

tion, the salience of being a donor also changes. In contrast,

social identity theory views donor-nonprofit identification as

a dynamic response to the organizational context (Hogg, Terry,

and White 1995). Thus, a donor’s identification with a nonpro-

fit organization might shift quickly, such as when the organiza-

tion suffers a scandal. The stability argument from identity

theory indicates that changes in donor identity salience are

unlikely and occur only when the person’s individual role

changes. Thus:

Hypothesis 1: The donor identity salience ! donor-nonprofit identification causal direction dominates the donor-nonprofit

identification! donor identity salience causal direction in non- profit relationships.

Effects of Donor Orientation on Donor-Nonprofit Identification and Donor Identity Salience. Donor orientation refers to the stable, durable tenor of a nonprofit organization’s general norms and

behavior toward its donor base. A nonprofit organization with

a high donor orientation takes care of donors’ needs, tries to

build trusting relationships, and keeps donors’ best interests

in mind. Service research indicates that a customer-oriented

organizational culture drives several positive outcomes, includ-

ing customer satisfaction. We do not offer hypotheses about

this well-accepted link. However, Homburg, Wieseke, and

Hoyer (2009) indicate that customer orientation has a strong

effect on customer-company identification. We thus predict

that donor orientation is positively related to donor-nonprofit

identification and further anticipate a possible unexplored link

from donor orientation to donor identity salience. According to

psychology studies, people are more likely to prioritize objects

they perceive as positive rather than negative (Fredrickson and

Branigan 2005). Therefore, the individual salience of being a

donor should be greater if the nonprofit organization exhibits

a strong caring strategy toward the donor. Thus:

Hypothesis 2: Higher donor orientation leads to greater

(a) donor-nonprofit identification and (b) donor identity salience.

Effects of Satisfaction on Donor-Nonprofit Identification and Donor Identity Salience. Customer satisfaction is a critical factor for retaining loyal customers, but the relationship between satis-

faction and identification is less certain, especially because pre-

vious findings on the link between organizational identification

Boenigk and Helmig 535

and satisfaction are contradictory. For example, Homburg,

Wieseke, and Hoyer (2009) conceptualize no relationship

between the two constructs. Other models feature a path from

satisfaction to organizational identification (Mael and Ashforth

1992), and still others confirm an opposite path from organiza-

tional identification to satisfaction (e.g., Camarero and Garrido

2011). Thus, the causal direction of the relationship between

donor satisfaction and donor-nonprofit identification is not yet

tested. Bodet and Bernache-Assollant (2011) test five alterna-

tive models to conceptualize the relationship among customer

satisfaction, sports team identification, and customer loyalty

and find that team identification is a mediating construct

between customer satisfaction and customer loyalty. Following

that study, we predict that the link from donor satisfaction to

donor-nonprofit identification dominates the oppositional path.

Thus:

Hypothesis 3a: The donor satisfaction ! donor-nonprofit identification causal direction dominates the donor-nonprofit

identification! donor satisfaction causal direction in nonprofit relationships, and higher donor satisfaction leads to greater

donor-nonprofit identification.

Concerning the relationship between satisfaction and iden-

tity salience, we are aware of only one empirical marketing

study that explicitly analyzed this link. Within the context of

universities and their alumni, the identity salience model of

relationship marketing success showed empirical evidence that

‘‘satisfaction is related significantly to identity salience’’

(Arnett, German, and Hunt 2003, p. 98). The authors assumed

that ‘‘alumni who are satisfied with their university experience

are more likely to place a university identity higher in their

hierarchy of identities’’ (p. 94). At first sight, this result would

support the opinion that the relationship between donor satis-

faction and donor identity salience should in this study also

be hypothesized as ‘‘higher donor satisfaction leads to greater

donor identity salience.’’ However, we want to express our

skepticism on that aspect, because of the following arguments.

On the one hand, we believe that the perceived satisfaction of

an individual could change over time, because positive or neg-

ative service experiences, for example, during blood-collecting

events or fund-raising events, can occur at any time and could

impact the satisfaction of donors. On the other hand, the char-

acter of donor identity salience is completely different to

donor-nonprofit identification, because the relative importance

to be a donor in the evoked set of a different other identities is

not changed easily and is surely not based on one or a few neg-

ative interactions. Thus, it is more plausible that donor satisfac-

tion leads to donor-nonprofit identification, whereas identity

salience changes only if the donor’s role is affected. By com-

bining these arguments, we generally assume that it is possible

that donors who are satisfied with the work and activities of the

nonprofit organization are to some extent likely to place the

donor identity higher in their own role identity hierarchy. How-

ever, we assume the impact of donor satisfaction on donor iden-

tity salience to be a lot lower compared to the path between

satisfaction and donor-nonprofit. Nevertheless, we hypothesize

the following:

Hypothesis 3b: Donor satisfaction is related positively to donor

identity salience.

Effects of Donor-Nonprofit Identification and Donor Identity Salience on Loyalty. The relationship between customer satisfaction and customer loyalty is evident in many service industries and

should transfer to nonprofit relationships (Helmig and Thaler

2010). For example, Wisner et al. (2005) find that satisfied

volunteers are more likely to remain with the organization

longer, donate more, and recommend the volunteer experience

to others. Boenigk, Leipnitz, and Scherhag (2011) confirm that

this mechanism holds among blood donors; the relationship

between blood donors’ satisfaction and their loyalty is even

stronger than the path from altruistic values to donor loyalty.

According to social identity theory and identity theory,

donor-nonprofit identification and donor identity salience also

have additional effects on donor loyalty. Ahearne, Bhattarch-

arya, and Gruen (2005, p. 577) test the link from customer-

company identification to customer extra-role behaviors (e.g.,

positive word of mouth, product improvement suggestions,

recruiting other customers, anticipating problems) and find

empirical support, concluding that ‘‘performing such extra-

role behaviors is a way to express one’s identification.’’ In a

travel agency context, Homburg, Wieseke, and Hoyer (2009)

offer empirical support for the link between customer-

company identification and customer loyalty. Accordingly,

we hypothesize that donor-nonprofit identification has a posi-

tive effect on donor loyalty. Regarding identity salience,

Arnett, German, and Hunt (2003) confirm effects on word of

mouth, which is one dimension of loyalty. Thus:

Hypothesis 4: (a) Higher donor-nonprofit identification

levels and (b) a more salient donor identity increase donor

loyalty.

Effects of Donor-Nonprofit Identification and Donor Identity Salience on Donation Behavior. Finally, service management studies indicate that loyalty leads to purchase behavior (Mittal and

Kamakura 2001; Olsen 2002) and better financial performance

(Rust and Zahorik 1993). In a donor context, loyalty means that

a donor is willing to give again, donate more, or recommend

the nonprofit organization to family and friends (Sargeant and

Woodliffe 2007). According to Sargeant and Jay (2010), even

small improvements in donor loyalty can have profound

impacts on the ‘‘profitability’’ of fund-raising. In line with

identity theory, Arnett, German, and Hunt (2003) also test the

link from university identity salience to donations to a univer-

sity and find a positive connection. We thus predict that both

identification constructs have positive effects on donation

behavior.

536 Journal of Service Research 16(4)

Hypothesis 5: (a) Higher donor-nonprofit identification levels

and (b) a more salient donor identity increase donations.

In the next section, we test our proposed conceptual frame-

work empirically in two nonprofit industries and present the

empirical results. Then, we test three rival models: (1) a baseline

model without identification constructs, (2) a model extended

with donor-nonprofit identification, and (3) a model with donor

identity salience as a moderator.

Research Method

Data Collection

To test the proposed model, we conducted a large-scale, quan-

titative study, in cooperation with two nonprofit organizations.

Sample 1 features data from private donors to a nonprofit orga-

nization working in the field of cancer prevention and therapy.

This organization consists of a national umbrella association

with 20 regional headquarters, and 77% of its annual financial budget comes from private donations, which supports 50–60

cancer projects each year. In addition, it operates an online

shop, cooperates with and invites sponsorships from for-

profit companies, and receives legacies (i.e., money and in-

kind donations from wills). In November 2009, we e-mailed

3,751 questionnaires to the donor database of this nonprofit

organization; 314 donors completed the survey, for a response

rate of 8.4%. Sample 2 data came from blood donors at a nonprofit blood donation service. In March 2011, the same

questionnaire was delivered to 1,000 members of a blood dona-

tion online community; 298 people participated, for a response

rate of 29.8%. We assessed nonresponse bias for each sample by comparing early and late respondents, but it was not a sig-

nificant problem for this study. Table 1 presents an overview

of the sample characteristics.

Measurement

We used existing scales whenever possible and adapted them to

a nonprofit context (see Table 2). The multidimensional con-

structs used 7-point Likert-type scales, ranging from 1 (totally

disagree) to 7 (totally agree).

Donor Orientation. Nonprofit management research provides no accepted scale to measure donor orientation. Instead, previous

studies have approached it as one dimension of the market

orientation of nonprofit organizations (Duque-Zuluaga and

Schneider 2008). Other studies have used the term but actually

measure service quality aspects, such as communication qual-

ity or the payment method used for fund-raising (Sargeant and

Woodliffe 2007; Shapiro 2010). Therefore, we adopted the

items from Homburg, Wieseke, and Hoyer (2009) to measure

customer orientation and transferred them to a nonprofit set-

ting. We reflectively specify and measure donor orientation

using 4 items: (1) ‘‘The nonprofit organization is taking care

of donors’ needs,’’ (2) ‘‘The behavior of the nonprofit organi-

zation toward donors is very relational,’’ (3) ‘‘The nonprofit

organization tries to establish a long-term relationship,’’ and

(4) ‘‘The nonprofit organization has the donor’s best interest

in mind.’’

Donor-Nonprofit Identification. Three dominant approaches have served to measure organizational identification in prior liter-

ature. Mael and Ashforth’s (1992) scale contains 6 items and

has been widely adopted (e.g., Bhattacharya, Rao, and Glynn

1995; Gwinner and Swanson 2003; Homburg, Wieseke, and

Hoyer 2009). However, some critics note that all 6 items are

cognitive in nature (Van Dick et al. 2006). Therefore, a mul-

tidimensional approach instead differentiates cognitive, affec-

tive, evaluative, and behavioral items to measure

organizational identification (Van Dick et al. 2006; Van Dick

and Wagner 2002). Van Dick et al. (2006) indicate strong

empirical support for this approach, though they exclude the

evaluative item from their employee identification model. To

determine the differences in measurement, we first applied

Mael and Ashforth’s scale and then contrasted it with the

multidimensional scale. We adopted the measurement

approach that provides better measurement quality for

donor-nonprofit identification—the one-dimensional cogni-

tive approach.

Donor Identity Salience. Callero (1985) developed an early scale to measure blood donor identity salience, and many service

management studies have adapted this scale to their purposes

and contexts. For example, to measure university identity

salience, Arnett, German, and Hunt (2003) use 4 items from

Callero’s scale (two positive and two negative). We adopted

this 4-item approach to measure donor identity salience on a

7-point scale: (1) ‘‘Giving is an important part of who I am,’’

(2) ‘‘Giving is something about which I have no clear feeling,’’

(3) ‘‘[Giving] means more to me than just donating money/

blood,’’ and (4) ‘‘[Giving] is something I rarely think about.’’

Table 1. Sample Characteristics.

Overall Characteristics Sample 1 Sample 2

Donation Money (cancer nonprofit) 314 (51%) — Blood (blood donation service) — 298 (49%)

Gender Female 169 109 Male 133 188 Missing responses 12 1

Age Young/middle-age donors (18–54 years) 148 226 Older donors (55 years and older) 154 70 Missing responses 12 1

Income Lower income (less than US$60,000) 163 168 Higher income (more than US$60,000) 139 101 Missing responses 12 29

Nature of the relationship Sporadic (first-time donor or a few times) 164 71 Continuous (regular donor) 150 223 Missing responses 0 4

Boenigk and Helmig 537

Table 2. Measurement Quality Report, Discriminant Validity, and Multigroup Comparison.

Measurement Quality of All Constructs (7-point Likert-Type Scale, 1 ¼ Totally Disagree, 7 ¼ Totally Agree)

Sample 1 (Money Donations) N ¼ 314 Sample 2 (Blood Donations) N ¼ 298

M (SD)

Factor Loading AVE

Composite Reliability

M (SD)

Factor Loading AVE

Composite Reliability

Donor orientation (Homburg, Wieseke, and Hoyer 2009) .65 .87 .75 .92 The NPO is taking care of donors’ needs 4.2 (1.7) .818*** 5.6 (1.4) .811*** The NPO’s behavior to donors is very relational 5.7 (1.3) .815*** 5.8 (1.3) .917*** The NPO tries to establish a long-term donor

relationship 6.0 (1.2) .711*** 5.9 (1.3) .830***

The NPO has the donor’s best interest in mind 4.8 (1.5) .865*** 5.4 (1.4) .907*** Donor-nonprofit identification (Mael and Ashforth 1992) .60 .90 .72 .94

When someone criticizes NPO, it feels like a personal insult

2.8 (1.9) .762*** 3.0 (1.9) .881***

I am very interested what other think about NPO 3.5 (1.9) .754*** 3.0 (1.8) .864*** When I talk about NPO, I usually say ‘‘we’’ rather than

‘‘they’’ 2.2 (1.7) .834*** 2.4 (1.8) .881***

This NPO’s success is my success 3.0 (1.8) .781*** 2.9 (2.0) .878*** When someone praises this NPO, it feels like a personal

compliment 2.5 (1.8) .854*** 2.7 (1.9) .912***

If a story in the media criticized the NPO, I would feel embarrassed

2.4 (1.8) .626*** 2.6 (1.8) .631***

Donor identity salience (Arnett, German, and Hunt 2003; Callero 1985; 2 items excluded)

.82 .90 .84 .91

Giving is an important part of who I am 3.4 (1.9) .903*** 4.7 (2.0) .906*** Giving means more to me than just donating money/

blood 4.1 (1.9) .912*** 4.5 (2.0) .926***

Donor satisfaction (Anderson and Fornell 1999) .96 .98 .91 .95 Overall, I am very satisfied with this nonprofit

organization 5.6 (1.2) .982*** 5.8 (1.1) .954***

When I reflect on my expectation before I started a relationship and donated, this NPO fulfills my entire expectations

5.5 (1.3) .980*** 5.5 (1.3) .956***

Donor loyalty (Sargeant and Woodliffe 2007) .52 .76 .46 .71 Willingness to donate again to this NPO 5.3 (1.7) .694*** 6.6 (0.9) .586*** Willingness to donate more to NPO 2.4 (1.5) .653*** 3.9 (2.3) .545*** Recommendation to family and friends 4.2 (2.1) .814*** 5.0 (1.8) .858***

Donations I donated money to this NPO. The total amount of giving

in the last 2 years was (6-point scale from 1 ¼ little money/50 SFr to 6 ¼ a lot of money/more than 1000 SFr)

2.3 (1.0) 1.00 1.00 1.00 5.1 (1.4) 1.00 1.00 1.00

I donated blood to this NPO. The total numbers of events in the last 2 years was (6-point scale from 1¼ 1 time to 6 ¼ more than 5 times)

Discriminant validity Sample 1 (money donations) 1 2 3 4 5 6

Donations 1.0 Donor identity salience .00 .82 Donor loyalty .04 .30 .52 Donor-nonprofit identification .00 .36 .27 .60 Donor orientation .00 .14 .14 .12 .65 Donor satisfaction .00 .09 .12 .06 .15 .96

Sample 2 (blood donations) 1 2 3 4 5 6 Donations 1.0 Donor identity salience .03 .84 Donor loyalty .00 .18 .46 Donor-nonprofit identification .01 .25 .19 .72 Donor orientation .00 .07 .12 .07 .75 Donor satisfaction .00 .08 .22 .09 .43 .91

Note. n.s. ¼ not significant; NPO ¼ nonprofit organization; AVE ¼ average variance extracted. SmartPLS bootstrapping with 600 iterations. Boldface values on the diagonal are AVE; values below the diagonal represent squared correlation values. *p < .05. **p < .01. ***p < .001.

538 Journal of Service Research 16(4)

The measurement quality results led us to exclude 2 items;

thus, we measured donor identity salience only on the first and

third items.

Donor Satisfaction. A widely accepted measure of customer satisfaction uses 2 items, as in national customer satisfaction

surveys such as the American Customer Satisfaction Index

(Anderson and Fornell 1999) and the European Performance

Satisfaction Index (Eklöf and Selivanova 2008). Accord-

ingly, we asked donors to evaluate the following items:

‘‘Overall, I am very satisfied with this nonprofit organiza-

tion’’ and ‘‘When I reflect on my expectation before I

started a relationship and donated, this nonprofit organiza-

tion fulfilled my entire expectations’’ (1 ¼ totally disagree, 7 ¼ totally agree).

Donor Loyalty. A 3-item measure of customer loyalty offers high reliability, even in nonprofit marketing contexts (Sargeant and

Woodliffe 2007). To ensure good measurement quality, we

used all 3 items in this study. Donors indicated their willingness

to donate again to the nonprofit organization, donate more

often, and engage in positive word-of-mouth behavior about

the organization.

Donations. We measured donating to the nonprofit organization using a single item, which asked the donors about the amount

of money they gave to the organization in the previous 2 years.

For the blood donation sample, we asked how many times the

respondents had donated blood to the blood service provider in

the previous 2 years.

Control Variables. To clarify the conceptual framework, we controlled not only for the two nonprofit industry differences

(money vs. blood) but also for context dynamics (recently

occurred scandals), donor characteristics (gender, age,

income), and relationship characteristics (sporadic vs. continu-

ous relationship; see Table 3).

Analytical Approach. We used partial least squares (PLS) path modeling to test our predictive model (Lohmöller 1989;

Wold 1982). Structural equation modeling would require

large sample sizes and relatively few indicators or constructs,

whereas PLS can apply to relatively small sample sizes and

complex models (Hair, Ringle, and Sarstedt 2012; Hair

et al. 2012). We used SmartPLS 2.0 software to test all mod-

els (Ringle, Wende, and Will 2005) and employed nonpara-

metric bootstrapping through SmartPLS to obtain the t-

values to test for significance. Missing values were not an

issue; when they rarely emerged, we used a mean replace-

ment procedure.

Results

In general, PLS path modeling typically entails a two-step

procedure that evaluates the measurement model and then the

structural model (Henseler, Ringle, and Sinkovics 2009).

However, this study required three steps because we had to

clarify the most appropriate causal direction of the bidirec-

tional paths first.

Step 1: Cohen Path Analysis

The path between donor-nonprofit identification and donor

identity salience has gone unexplored; Cohen’s path analysis

can support empirical tests of which of the two causal relation-

ships is dominant (Cohen et al. 1993). Wilson et al. (2007)

were, to the best of our knowledge, the first marketing

researchers who used Cohen’s path method within PLS model-

ing (see additionally, Sattler et al. 2010). For this study, Alter-

native 1 refers to the path from donor identity salience to

donor-nonprofit identification (Hypothesis 1) and Alternative

2 moves from donor-nonprofit identification to donor identity

salience. The underlying idea of Cohen’s path analysis is that

the estimated correlations (all direct and indirect effects¼ total effects) should be as close as possible to the actual correlation.

The results of Cohen’s path analysis for Sample 1 (money

donations) revealed that the total squared error (TSE1) for

Alternative 1 was .113 and that for Alternative 2 (TSE2) was

.105. The error change from Alternative 1 to Alternative 2 was

�7.1% (.105 to .113/.113 ¼ �.071). In Cohen’s terminology, this negative sign of the error change means that the TSE

decreases in the shift from Alternative 1 to Alternative 2. We

also calculated the d-value (Cohen 1988) by applying ([TSE2� TSE1]/s), where s is the pooled standard deviation of TSE val-

ues, which equaled .053 (Sample 1) and .044 (Sample 2) in this

study. A d-value of .20 would indicate a small effect, .50 a

medium effect, and greater than .80 a large effect (Cohen

1988). In Sample 1, we calculated a small Cohen d-value of

�.152 and, in Sample 2, no effect with a d-value of .006. Thus, the data do not support Hypothesis 1. In this study, the donor-

nonprofit identification! donor identity salience causal direc- tion dominates the opposite direction in nonprofit relationships.

The results appear in Figure 1.

We next tested the causal direction between donor satisfac-

tion and donor-nonprofit identification. Here, Alternative 1

refers to the path from donor satisfaction to donor-nonprofit

identification (Hypothesis 3a), and Alternative 2 moves from

donor-nonprofit identification to donor satisfaction. For

Sample 1, we found a TSE for Alternative 1 (TSE1) of .105 and

that for Alternative 2 (TSE2) of .135 (TSE change ¼ .29). For Sample 2, TSE1 is .104 and TSE2 is.159 (TSE change ¼ .52). From these results, the d-value within the money donation con-

text is medium, with a value of .59, and the d-value within the

blood donation context is large at 1.03. Thus, the data support

Hypothesis 3a. In satisfaction-loyalty studies, the link should

be conceptualized and measured from satisfaction to

identification.

In addition, note that we cannot test for bidirectionality of

the relationship between donor satisfaction and donor identity

salience. Changing the causal direction from donor identity sal-

ience to donor satisfaction would lead to a closed-loop system,

Boenigk and Helmig 539

T a b

le 3 .

R es

u lt s

o f th

e P ar

ti al

Le as

t Sq

u ar

es (P

LS )

E st

im at

io n

fo r

R iv

al M

o d el

s an

d M

u lt ig

ro u p

C o m

p ar

is o n .

P LS

E st

im at

io n

fo r

R iv

al M

o d el

s

Sa m

p le

1 (M

o n ey

D o n at

io n s,

N ¼

3 1 4 )

Sa m

p le

2 (B

lo o d

D o n at

io n s,

N ¼

2 9 8 )

B as

e lin

e M

o d el

O I

M o d el

IS M

o d .

M o d el

P ro

p o se

d M

o d el

B as

el in

e M

o d el

O I

M o d el

IS M

o d .

M o d el

P ro

p o se

d M

o d el

P at

h s

fr o m

d o n o r

o ri

en ta

ti o n

D o n o r

o ri

en ta

ti o n !

d o n o r

sa ti sf

ac ti o n

.4 0 **

* .3

9 **

* .3

9 **

* .3

9 **

* .6

6 **

* .6

6 **

* .6

6 **

* .6

6 **

* D

o n o r

o ri

en ta

ti o n !

d o n o r-

n o n p ro

fit id

en ti fic

at io

n .2

9 **

* .2

9 **

* .2

9 **

* .1

4 *

.1 4 *

.1 4 *

D o n o r

o ri

en ta

ti o n !

d o n o r

id en

ti ty

sa lie

n ce

.1 5 **

.0 9

n .s

. P at

h fr

o m

d o n o r

sa ti sf

ac ti o n

D o n o r

sa ti sf

ac ti o n !

d o n o r-

n o n p ro

fit id

en ti fic

at io

n .1

4 **

.1 4 **

.1 4 *

.2 2 **

* .2

2 **

* .2

2 **

D o n o r

sa ti sf

ac ti o n !

d o n o r

id en

ti ty

sa lie

n ce

.1 1 *

.1 0

n .s

E ff ec

ts o n

d o n o r

lo ya

lt y

D o n o r

sa ti sf

ac ti o n !

d o n o r

lo ya

lt y

.3 6 **

* .2

3 **

* .1

8 **

* .1

8 **

* .4

7 **

* .3

7 **

* .3

4 **

* .3

4 **

* D

o n o r-

n o n p ro

fit id

en ti fic

at io

n !

d o n o r

lo ya

lt y

.4 9 **

* .4

1 **

* .2

7 **

* .3

3 **

* .2

3 **

* .2

3 **

* D

o n o r

id en

ti ty

sa lie

n ce !

d o n o r

lo ya

lt y

.3 3 **

.2 0 **

* D

o n o r

id en

ti ty

sa lie

n ce

as a

m o d er

at o r

.1 8

n .s

. .0

2 n .s

. E ff ec

ts o n

d o n at

io n s

D o n o r

lo ya

lt y !

d o n at

io n s

.2 0 **

* .1

9 **

* .1

9 **

* .2

6 **

* .0

4 n .s

. .0

6 n .s

. .0

3 n .s

. .1

5 n

.s .

D o n o r-

n o n p ro

fit id

en ti fic

at io

n !

d o n at

io n s

� .0

4 n

.s .

.0 9

n .s

. D

o n o r

id en

ti ty

sa lie

n ce !

d o n at

io n s

� 0 8

n .s

.1 9 *

R 2

d o n o r-

n o n p ro

fit id

en ti fic

at io

n —

1 3 %

1 3 %

1 3 %

— 1 1 %

1 1 %

1 1 %

R 2

d o n o r

id en

ti ty

sa lie

n ce

— —

— 4 1 %

— —

— 2 8 %

R 2

d o n o r

sa ti sf

ac ti o n

(% )

1 6

1 6

1 6

1 6

4 4

4 4

4 4

4 4

R 2

d o n o r

lo ya

lt y

(% )

1 3

3 3

3 9

4 0

2 2

3 2

3 5

3 5

R 2

d o n at

io n s

(% )

4 4

4 4

0 0

0 5

G o o d n es

s o f fit

(G o F)

.2 8

.3 4

.3 6

.4 0

.3 9

.3 9

.4 1

.4 3

f2 (f

o r

R 2

d o n o r

lo ya

lt y)

— .3

0 .4

3 a /

.1 0

b .4

5 a /. 1 2

b —

.1 5

.2 0

a / .0

5 b

.2 0

a /. 0 5

b

Sa m

p le

1 Sa

m p le

2

M u lt ig

ro u p

co m

p ar

is o n

(H en

se le

r te

st )

b 1 vs

. b 2

p H e n se

le r

b 1 vs

. b 2

p H

e n se

le r

D o n o r

sa ti sf

ac ti o n

o n

d o n o r

lo ya

lt y

G en

d er

: fe

m al

e vs

. m

al e

.2 3

vs . .1

1 .1

1 n .s

. .2

9 vs

. .3

5 .6

6 n .s

. A

ge : m

id d le

ag e

vs . o ld

er d o n o rs

.2 6

vs . .1

2 .0

8 n .s

. .3

0 vs

. .6

2 .9

6 *

In co

m e:

Lo w

er vs

. h ig

h er

in co

m e

.2 0

vs . .1

8 .4

3 n .s

. .3

2 vs

. .3

2 .5

2 n .s

. R

el at

io n sh

ip b re

at h : sp

o ra

d ic

vs . co

n ti n u o u s

.2 2

vs . .1

5 .2

5 n .s

. .3

4 vs

. .3

9 .6

4 n .s

. D

o n o r-

n o n p ro

fit id

en ti fic

at io

n o n

d o n o r

lo ya

lt y

G en

d er

: fe

m al

e vs

. m

al e

.2 6

vs . .3

4 .7

3 n .s

. .2

7 vs

. .1

8 .2

5 n .s

. A

ge : m

id d le

ag e

vs . o ld

er .3

4 vs

. .2

3 .1

9 n .s

. .2

6 vs

. �

.1 1

.0 5 *

In co

m e:

lo w

er vs

. h ig

h er

in co

m e

.2 1

vs . .3

2 .8

2 n .s

. .3

2 vs

. .1

0 .0

4 *

R el

at io

n sh

ip b re

at h : sp

o ra

d ic

vs . co

n ti n u o u s

.2 7

vs . .3

1 .6

5 n .s

. .2

2 vs

. .2

1 .4

8 n .s

. D

o n o r

id en

ti ty

sa lie

n ce !

d o n o r

lo ya

lt y

G en

d er

: fe

m al

e vs

. m

al e

.3 2

vs . .3

4 .5

6 n .s

. .1

5 vs

. .2

7 .7

8 n .s

. A

ge : m

id d le

ag e

vs . o ld

er d o n o rs

.2 1

vs . .4

3 .9

6 *

.2 1

vs . .2

5 .6

5 n .s

. In

co m

e: lo

w er

vs . h ig

h er

in co

m e

.3 6

vs . .3

0 .3

2 n .s

. .1

0 vs

. .3

8 .9

7 *

R el

at io

n sh

ip b re

at h : sp

o ra

d ic

vs . co

n ti n u o u s

.2 8

vs . .3

5 .7

1 n .s

. .2

2 vs

. .1

9 .3

8 n .s

.

N ot

e. B as

el in

e m

o d el ¼

m o d el

w it h o u t

an y

id en

ti fic

at io

n co

n st

ru ct

; O

I m

o d el ¼

o rg

an iz

at io

n al

id en

ti fic

at io

n m

o d el

; IS

m o d . m

o d el ¼

id en

ti ty

sa lie

n ce

m o d er

at o r

m o d el

. n .s

. ¼

n o t

si gn

ifi ca

n t.

Sm ar

tP LS

b o o ts

tr ap

p in

g w

it h

6 0 0

it er

at io

n s.

A ll

st at

is ti ca

l si

gn ifi

ca n ce

s o f b o ld

fa ce

va lu

es ar

e gi

ve n , w

h er

ev er

ap p ro

p ri

at e.

E ff ec

t si

ze s

fo r

f2 :

a C al

cu la

ti o n

o f f2

fo r

th e

m en

ti o n ed

m o d el

in co

m p ar

is o n

to th

e b as

el in

e m

o d el

. b C

al cu

la ti o n

o f f2

fo r

th e

m en

ti o n ed

m o d el

in co

m p ar

is o n

w it h

th e

o rg

an iz

at io

n al

id en

ti fic

at io

n m

o d el

. p h

e n se

le r in

d ic

at es

th e

p ro

b ab

ili ty

th at

th e

se co

n d

gr o u p ’s

p o p u la

ti o n

p ar

am et

er is

gr ea

te r

th an

th at

o f th

e fir

st gr

o u p

(p >

.0 5 ¼

n o t

si gn

ifi ca

n t

(n .s

.) .I

n th

e ca

se th

at th

e se

co n d

p o p u la

ti o n

p ar

am et

er is

gr ea

te r

th an

th e

fir st

, p

< .9

5 , w

h ic

h is

n o t

si gn

ifi ca

n t.

T h e

re su

lt s

fr o m

H en

se le

r (2

0 0 7 )

w er

e el

ig ib

le fo

r a

o n e-

si d ed

te st

. *p

< .0

5 . **

p <

.0 1 . **

*p <

.0 0 1 .

540

and consequently we would not be able to estimate the model

with PLS.

Step 2: PLS: Evaluation of Measurement Quality

In Step 2, we evaluated the measurement quality of the reflec-

tive measurement models by determining their item reliability,

construct reliability, and discriminant validity. For the item

reliability assessment, the factor loadings of the reflective

constructs should be greater than .7, which would indicate that

more than half the variance in the observed variable was due to

the construct. The average variance extracted (AVE) should be

above the critical value of .5 (Fornell and Larcker 1981), and

composite reliability should be equal to or greater than .7

(Nunnally and Bernstein 1994). Table 2 presents an overview

of measurement quality.

The relationships between the reflective items and the con-

structs were all significant; most items indicated factor load-

ings greater than .8. The results in Table 2 further show that

except for donor loyalty, all factor loadings achieved reliability

and validity. For loyalty, two factor loadings were slightly

below the threshold: .694 for ‘‘willingness to donate again’’

and .653 for ‘‘willingness to donate more.’’ In Sample 1, the

AVE for donor loyalty was slightly above the critical value

(.52), but in Sample 2, it was only .46. We interpret this weak

AVE result as a first hint of the need to measure blood donor

loyalty further. However, the composite reliability was satis-

factory in both samples.

To identify the highest measurement quality for donor-

nonprofit identification, we first estimated the model using six

cognitive items from Mael and Ashforth’s (1992) scale and

then contrasted the outcome with the results from the previ-

ously mentioned multidimensional measurement approach

(Van Dick et al. 2006). We found that the multidimensional

approach did not improve measurement quality for donor-

nonprofit identification. For example, in Sample 2, the weak

AVE reached only .37, in contrast to the .72 value obtained

using Mael and Ashforth’s scale. Composite reliability was

also much weaker (.69 vs. .94) when we used the multidimen-

sional measure. In addition, Cronbach’s a value was only .52 for the multidimensional measurement approach, compared

with .89 for the Mael/Ashforth scale. Therefore, we finally

applied the six traditional items of Mael and Ashforth’s scale

documented in Table 2 to evaluate donor-nonprofit

identification.

The evaluation of measurement quality for donor identity

salience led to some concerns with both AVE and composite

reliability. First, we tested the 4-item measurement approach

that Arnett, German, and Hunt (2003) suggest. The AVE values

for both samples were .53, marginally above the critical value,

but the composite reliability attained only unacceptable values

of .16 (Sample 1) and .13 (Sample 2). With a reverse coding of

the two negative items (see Arnett, German, and Hunt 2003),

the AVE decreased further to .4. Therefore, we deleted the 2

items with low factor loadings (Items 2 and 4), which produced

satisfactory measurement quality for donor identity salience.

With the 2-item approach, the AVEs were .82 and .84, and the

composite reliability scores were .90 and .91 for Samples 1 and

2, respectively (for the same problems on measurement quality,

see Lee, Piliavin, and Call 1999).

Donor-NPO Identification

R2 .13 .11

H1 Donor Donor L lt

.29*** .14*

.39*** .66*** Donor

H3a.14* .22***

18*** 34*** Donations.26*** .15 n.s. .53*** 44***

H2a H4a .27*** .23***

H5a -.04 n.s. .09 n.s.

H1 Orientation Loyalty

.15** .09 n.s.

Satisfaction

H3b .11* .10 n.s.

. ..44

R2 .16 .44 R2 .40 .35 R 2 .04 .05

H2b H4b .33*** .20***

H5b -.08 n.s. .19**

Donor Identity Salience

R2 .41 .28

H

Figure 1. Results of the proposed model: Samples 1 and 2. Note. Control variables: (1) Context dynamics: no indication of scandals during the study period. (2) Donor and relationship characteristics have been tested applying multigroup comparison analysis. Results are shown in Table 3. (3) Nonprofit industry differences: heterogeneity across industries exist, see results above. Assessment of two PLS Models: Sample 1: money donations, N ¼ 314 (Sample 2: blood donations, N ¼ 298); *p < .05; **p < .01; ***p < .001; n.s ¼ not significant. SmartPLS Algorithm Settings: Path Weighting Scheme (additional test with factor weighting scheme show no influence on the results); Data metric mean ¼ 0, variance ¼ 1; maximum iterations ¼ 300; abort criterion ¼ 1.0E-5. SmartPLS bootstrapping settings: cases ¼ 600 iterations; sample ¼ 314 (298). Dotted paths: Bidirectional flow has been tested applying Cohen’s d path analysis. The causal direction indicated by a dotted path is not supported by the data.

Boenigk and Helmig 541

Second, we tested the discriminant validity of the model

using Fornell and Larcker’s (1981) criterion, which requires

that each construct shares more variance with its own indica-

tors than with any other constructs’ indicators in the model

(AVE > squared correlation). The results in Table 2 show that

this criterion was fulfilled, in support of discriminant validity.

Step 3: PLS: Evaluation of the Structural Model

The central criteria for evaluating the structure of the PLS

model are the path coefficient estimation for each link and the

rate of reliability (R2) for the endogenous variables. In nonpro-

fit marketing studies, no accepted opinion on threshold values

for a weak, moderate, and substantial R2 exist; but the closer to

1, the better. Therefore, we follow the recommendations of

Chin (1998) and argue that R2 values less than .19 indicate

weak statistical power, those up to .33, moderate, and those

of .67 or greater, substantial. The results derived from the

proposed model (all path coefficients, R2 values, p values with

*p < .05; **p < .01; ***p < .001) appear in Figure 1. Because

the empirical results vary with the service industry being

studied (Rigdon et al. 2011), we report the results of the two

samples separately in Figure 1 (with the blood donation

Sample 2 in brackets).

Although we reject Hypothesis 1, which postulates a causal

direction from donor identity salience to donor-nonprofit

identification, we found based on the SmartPLS bootstrapping

procedure with 600 iterations significant path coefficients of

.53 (money donations) and .44 (blood donations) for the reverse

direct relationship, from donor-nonprofit identification to

donor identity salience. In both samples, the R2 values for

donor identity salience were moderate, at .41 and .28,

respectively.

In support of Hypothesis 2a, the data show significant path

coefficients for the link from donor orientation to donor-

nonprofit identification in both samples. However, in Sample

1, the effect was stronger, with a value of .29 compared with

.14 for Sample 2. Moreover, the links from donor orientation

to donor identity salience were not homogeneous: There was

a significant path coefficient of .15 for the money donation

context, but a smaller, nonsignificant path coefficient of .09 for

the blood donation context. Therefore, Hypothesis 2b receives

only partial support from the data. The strongest effects

emerged in the path from donor orientation to donor satisfac-

tion, with coefficients of .39 (Sample 1) and .66 (Sample 2).

We also found strong differences in the R2 values for donor

satisfaction in the two industries. In Sample 1, the weak R2 only

reached .16 and in Sample 2, it was moderate at .44.

In support of Hypothesis 3a (higher donor satisfaction leads

to greater donor-nonprofit identification), we found support in

both samples and significant path coefficients (Sample 1: .14;

Sample 2: .22) from donor satisfaction to donor-nonprofit iden-

tification. For Hypothesis 3b, which posits that donor satisfac-

tion has a positive effect on donor identity salience, we found a

significant path coefficient of .11 in Sample 1, but could not

confirm this relationship with Sample 2. In Sample 2, the path

is not significant with a path coefficient of .10. Therefore,

Hypothesis 3b is partially supported by the data.

Hypotheses 4a and 4b receive support from the data; both

identification constructs exerted positive impacts on donor loy-

alty. Unexpectedly, and in contrast to the studies that indicate

that satisfaction is the most influential driver of loyalty, we

found in the money donation context that the highest path

coefficient for donor loyalty was the one from donor identity

salience to donor loyalty (path ¼ .33). Although the two other drivers were significant, their lower values indicated weaker

effects. A contrary outcome arose in the blood donation

context: Donor satisfaction had the greatest impact on donor

loyalty, with a path coefficient of .34, followed by donor-

nonprofit identification (.23) and donor identity salience (.20).

We also calculated the total effects of satisfaction, identifi-

cation, and identity salience on donor loyalty. For the calcula-

tion of the total effects for a relationship, we used the sum of

direct and indirect path relationships, as implemented in

SmartPLS software as a standard routine. The indirect effect

represented the multiplication of all possible path coefficients

pertaining to this relationship. However, the total effects for

donor loyalty confirmed that in Sample 1, donor-nonprofit

identification exerted the greatest total effect on loyalty, with

a value of .45, followed by donor identity salience (.33) and

donor satisfaction, with its total effect of .28. In Sample 2, the

importance of the three constructs differed. Donor satisfaction

had the strongest total effect on loyalty (.43), followed by

donor orientation (.35), donor-nonprofit identification (.32),

and donor identity salience (.20).

Finally, studying the effects on donation behavior, we found

no support for Hypothesis 5a. The path coefficients for the link

from donor-nonprofit identification to donations were low and

nonsignificant (Sample 1: �.04; Sample 2: .09). Hypothesis 5b receives only partial support; we found a path coefficient of .19

in the blood donation context but a nonsignificant effect (�.08) in the money donation context. In regard to R2, we found

critical values of .04 and .05 for donations. Consequently and

besides the identification constructs that are here in the focus

of the analysis, other factors such as income should be taken

into account to explain real donation behavior of individuals.

Control Variables

First, and in line with social identity theory, we found that iden-

tification with an organization is not always stable; therefore,

we controlled for context dynamics. In interviews with the

managers in the two nonprofit organizations, we asked whether

any scandals or image problems had occurred in the past 12

months that might have influenced donors’ identification; both

respondents reported that no such event occurred. Furthermore,

searches of popular media for negative articles about the two

organizations produced no results. Thus, because there was

no indication of scandals during the study period, related con-

textual factors that might affect the level of identification can

be neglected.

542 Journal of Service Research 16(4)

Second, previous research confirms that customer character-

istics (Mittal and Kamakura 2001) and relationship characteris-

tics (Bolton, Lemon, and Verhoef 2004) are important drivers,

and thus we controlled for three paths from donor characteris-

tics (age, gender, and income) to donor loyalty. In the PLS path

modeling analysis, we cannot include the control variables

directly in the proposed model; therefore, we tested them by

applying a multigroup comparison analysis (Sarstedt, Henseler,

and Ringle 2011). We followed Henseler’s (2007) recommen-

dation and applied a bootstrapping approach for the multigroup

comparison.

This approach compares specific paths in two subsamples

by applying separate bootstrap analyses; the bootstrap outcome

then serves as a basis for testing the probability that group

differences exist. The pHenseler value expresses ‘‘the probability

that the second group’s population parameter will be greater

than that of the first’’ (Sarstedt, Henseler, and Ringle 2011,

p. 202). We compared several donor characteristics: female

versus male, middle age (18–54 years) versus older people

(55þ years), and lower income (less than US$60,000) versus higher income (more than US$60,000). We did not include

younger people (under 18 years) in this study because blood

donation is permissible in Europe only for those aged 18–68.

Furthermore, we controlled for the character of the relationship

by differentiating sporadic (first-time donation or few dona-

tions) versus continuous (regular donations) donor relationships.

As Table 3 shows, not many significant group differences

emerged, and most of the pHenseler values were not significant.

For Sample 1, the multigroup analysis indicated that donor

identity salience exerted a greater impact on donor loyalty for

older donors than for middle-age donors. For Sample 2, the

same age effect emerged for the links from donor satisfaction

to donor loyalty and from donor-nonprofit identification to

donor loyalty. Moreover, donor loyalty was stronger for the

group of blood donors with higher incomes in both identifica-

tion paths.

Estimation and Evaluation of Three Rival Models

As mentioned previously, the model documented in Figure 1 is

the first to combine both identification constructs to indicate

the direct effects on donor loyalty and donations. To compare

the results of the proposed model with previous research

models and findings, we estimated three rival models. Table 3

provides an overview of the results of the PLS estimation.

Baseline Model. In both samples, we tested a baseline model that contained no identification constructs but rather links from

donor orientation to donor satisfaction, from donor satisfaction

to donor loyalty, and from donor loyalty to donations. Similar

to previous satisfaction and loyalty studies, we found signifi-

cant path coefficients from donor orientation to donor satisfac-

tion (Paths1 ¼ .40; Paths2 ¼ .66) and from donor satisfaction to donor loyalty (Paths1 ¼ .36; Paths2 ¼ .47). Regarding the effects from donor loyalty to donations, we found a significant

path coefficient in the money donation context (Paths1 ¼ .20)

but a low, nonsignificant one in the blood donation context

(Paths2 ¼ .04). The explanatory power of the baseline model for donor satisfaction was weak in Sample 1 (R2 ¼ 16%) and moderate in Sample 2 (R2 ¼ 44%). We also found weak R2 values for donor loyalty, at 13% and 22%, respectively. These R2 values for donations in Sample 1 were equivalent in the pro-

posed and baseline model (4%) but below the limit of detection for Sample 2 (R2 ¼ .00).

Organizational Identification Model (OI Model). We also tested a second rival model, similar to that which Homburg, Wieseke,

and Hoyer (2009) present. This model integrated donor-

nonprofit identification (but excluded donor identity salience),

and we assessed the effects on donor satisfaction, donor loy-

alty, and donations. In both samples, donor orientation exerted

an effect not only on satisfaction but also on donor-nonprofit

identification (Paths1 ¼ .29; Paths2 ¼ .14). Moreover, the links from donor satisfaction to donor-nonprofit identification were

significant (Paths1 ¼ .14; Paths2 ¼ .22). Regarding the effects on donor loyalty, in Sample 1, the path coefficient from

donor-nonprofit identification to donor loyalty was higher

(.49) than the one from satisfaction to loyalty (.23). In Sample

2, the effect of donor-nonprofit identification on loyalty was

not strong, though the path coefficient was significant (.33

vs. .37). Overall, the empirical results for this second rival

model were generally similar to those of the proposed model,

but the R2 values for donor loyalty were lower, at 33% and 32%, than in our proposed model (40% and 35%).

Identity Salience Moderator Model (IS Mod. Model). A rival model with donor identity salience as a moderator of the relationship

between donor-nonprofit identification and donor loyalty pro-

vided another comparison for analysis. We tested for a poten-

tial moderating effect by applying the product indicator

approach that Henseler and Chin (2010) recommend. The

results in Table 3 reveal the low, nonsignificant path coeffi-

cients of the interaction variable (Moderating effects1 ¼ .18, n.s.; Moderating effects2 ¼ .02, n.s.). In contrast to previous research, we did not find any moderating effect of donor

identity salience on the relationship between donor-nonprofit

identification and donor loyalty.

Model Fit in PLS. Beyond these results, we require more justifi- cation to confirm that the proposed model is preferable to the

three noted rival models. Unlike structural equation modeling

with LISREL or AMOS, PLS path modeling does not offer a

global goodness-of-fit (GoF) criterion. Tenenhaus et al.

(2005) propose a GoF index that uses the geometric mean of

the average communalities and the average R2 of endogenous

latent variables. For our proposed model, the GoF values were

.40 (Sample 1) and .43 (Sample 2); that is, the quality of

the explanation was nearly the same in both nonprofit service

industries. Furthermore, the GoFs for the proposed model were

slightly higher than those derived from the other rival models.

Considering the limitations of the GoF index (Hair et al.

2012; Henseler and Sarstedt 2012), we cross-validated the

Boenigk and Helmig 543

proposed model by calculating the effect size (f 2¼ R2included� R2excluded)/(1 � R2included; Cohen 1988). For both samples, we evaluated whether the integration of the two identification con-

structs into the baseline model had a weak (f 2� .02), moderate (f 2 � .15), or substantial (f 2 � .35) effect on donor loyalty. In the money donation context, we found a moderate f2 value of

.30 when we included donor-nonprofit identification (Table

3, Sample 1: OI model). The effect size (.43) increased when

we integrated donor identity salience as a moderator of the

relationship between donor-nonprofit identification and donor

loyalty, but again the moderating effect was not significant in

the third rival model. Finally, we found a substantial effect size

(.45) when we included both donor-nonprofit identification and

donor identity salience in the baseline model. For the identity

salience moderator model and the proposed model, we also

compared the effect size on loyalty with that in the organiza-

tional identification model; there was a weak effect size of

.10 for the former model and a slightly stronger one (.12) for

the proposed model.

In the blood donation context, the moderate f 2 value of .15

showed that the organizational identification model outper-

formed the baseline model (Table 3, Sample 2: OI model). For

the identity salience moderator model and the proposed model,

we found moderate effect sizes in comparison with the baseline

model, at .20 in both cases. In comparison with the organiza-

tional identification model, the effect of integrating donor iden-

tity salience as a moderator or direct path was weak. Overall,

the performance of the proposed model in the money donation

context was slightly clearer than that in the blood donation

context. For completeness, we also calculated the effect sizes

for donations but found no effects.

Discussion

Only partial and incomplete knowledge exists about the role of

organizational identification and identity salience on the rela-

tionship among satisfaction, loyalty, and donations. To answer

our first research question, we developed a theoretical frame-

work that combines both constructs in relation to donor loyalty

and donations and evaluated a possible bidirectional relation-

ship. We did this within two nonprofit service contexts: money

donations versus blood donations. Overall, the key findings

reveal that donor-nonprofit identification and donor identity

are distinct constructs and that both explain donor loyalty, but

not donations. For the first time, we test the causal direction

between donor-nonprofit identification and donor identity sal-

ience and confirm a causal link from identification to identity

salience. Finally, and in contrast to previous studies, we find

that both identification constructs should be conceptualized

according to their direct links to loyalty.

The second research question prompted us to propose the

conceptual framework in Figure 1, which combines donor-

nonprofit identification and donor identity salience. The most

striking difference from previous studies is that we find empiri-

cal support for direct links from donor-nonprofit identification

and donor identity to donor loyalty. Our data do not support

donor identity salience as a moderator. Moreover, we test the

causal direction between donor-nonprofit identification and

donor satisfaction and find that the path from satisfaction to

identification is dominant and should be conceptualized as such

in future service management studies. To provide a better test

of the proposed model, we compared it with three rival models:

a baseline model without identification constructs, an organiza-

tional identification model with one identification construct

(donor-nonprofit identification), and an identity salience

moderator model in which we conceptualized donor identity

salience as a moderator of the link between identification and

loyalty. Although all the models featuring identification

explained a significant amount of the variance in donor loyalty,

the effect sizes improved in the proposed model compared with

the organizational identification and baseline models.

In addition, we conducted this research to assess our third

research question. Although in general the proposed model fits

with money and blood giving, we found differences in the

relative size of the path coefficients in the model. Therefore,

heterogeneity across industries must be taken into account. In

the blood donation context, the path coefficients of donor

orientation on satisfaction are much higher (.66 vs. .39) than

in the money-giving context. This result is not surprising,

because service quality aspects (e.g., physician competence)

more strongly affect blood donation. As a consequence of the

confirmation-disconfirmation of donors’ expectations, the total

effect of donor satisfaction on donor loyalty is the strongest of

the three tested paths. In contrast, in the money donation

setting, it is challenging for donors to be satisfied because they

cannot easily evaluate the extent to which their donated money

is used to support the social mission of the nonprofit organiza-

tion, rather than cover administrative/marketing costs.

The larger satisfaction effect in the blood donation context

may also be influenced by the positive word of mouth of family

members. Thus, to some extent, the decision to donate blood is

more a group behavior, whereas money giving is more an indi-

vidual decision. This may explain why the path coefficient

from donor identity salience to donor loyalty in money giving

is larger than in blood giving. Furthermore, the path from donor

identity salience to donations is nonsignificant in the money

donation context but, similar to Arnett, German, and Hunt’s

(2003) findings, significant in the blood donation context. This

difference likely stems from the differing salience of donors in

each context, which, according to sociological studies, is much

higher for blood than monetary donations (Lee, Piliavin, and

Call 1999).

Furthermore, we find a remarkable difference in the path

from donor loyalty to donations, which is nonsignificant in

the blood donation and significant in the money-giving con-

text. We posit that the main reason for this empirical result

is the limited amount of blood donations allowed per person

in a year combined with the already high loyalty level in the

blood donor base. For example, women between the ages of

18 and 68 may donate blood up to 4 times a year, and men

of the same age may do so 6 times per year. For our blood

donor sample, the mean donation level for the willingness

544 Journal of Service Research 16(4)

to donate blood again is 6.6 (on a 7-point scale). Thus,

increasing donor loyalty among this sample would not have

a powerful or significant effect, because donors would not

have the capacity to donate more.

Limitations

As is true for any study based on survey data, this research has

some limitations. First, this investigation did not include

employee identification (Maxham, Netemeyer, and Lichten-

stein 2010), so the results are limited to external identification

effects. Further research could replicate and extend the model

by integrating internal factors. Second, the research included

only two nonprofit industries. Additional research in other ser-

vice sectors could provide a deeper understanding of organiza-

tional identification and identity salience, particularly in regard

to time giving and volunteering. The role of donors is also not

equivalent to that of customers; donors seem to have a more

complex identity than customers. For example, one person

might be a donor, customer, and volunteer simultaneously

(Helmig et al. 2004). Third, all the measures were self-

reported by donors, which creates a potential for mono-

method bias. Further research could collect objective data

about fund-raising performance to obtain a fuller picture of the

outcomes. Fourth, other influential factors could be tested, such

as competitive fund-raising intensity (e.g., campaigns by other

organizations). Fifth, with the closed-loop effects in the PLS

analysis, we were not able to test the potential bidirectionality

between donor satisfaction and donor identity salience, which

thus provides a potential objective in future studies. Finally,

we did not examine the negative consequences of identification

processes in organizations or identity synergy (Fombelle et al.

2012), which suggests potential avenues for future research

studies.

Implications for Theory and Measurement

This research contributes to the theory on organizational iden-

tification and identity salience. Indeed, our study suggests that

service researchers should take a combined theoretical perspec-

tive and use identity theory and social identity theory to explain

the relationship behavior in satisfaction-loyalty studies. In

doing so, our findings direct attention to both the differences

in the theoretical approaches and the potential learning for

service researchers in combining them (Brickson 2012; Hogg,

Terry, and White 1995; Stets and Burke 2000).

First, the results of this study indicate that identity theory

focuses on individual roles and the individual’s relationship with

an object of identification (e.g., organization, brand, project, per-

son); social identity theory is about group characteristics and

group behavior. In most identity theory research, individual,

relationship-inducing factors, such as the degree of personal

satisfaction or engagement in the organization, represent the cen-

tral focus of the analysis (Arnett, German, and Hunt 2003). In

contrast, member characteristics (e.g., gender, age, nationality)

or organizational characteristics (e.g., size, country, and culture)

take greater precedence in social identity theory studies. From an

integrative theoretical perspective, service marketing researchers

should consider both individual relationship factors and group/

membership factors to address identification issues.

Second, we show that the stability of the two identification

construct varies across the two theories. In identity theory, peo-

ple’s identification with an organization is relatively stable, and

any changes are primarily responses to a change in their role (or

identity salience). In contrast, social identity theory views iden-

tification as a dynamic response to the context in which the

organization functions. Thus, a person’s identification with the

organization might change quickly in response to scandals or

other image problems. From our combined theoretical perspec-

tive, we propose that researchers should always recognize the

relative importance of the identification object to a person,

such that they integrate the construct of identity salience into

any identification-based research frameworks. Moreover,

researchers should recognize the potential dynamics of the con-

text and thus control for the potential effects of image problems

that alter the level of identification with an organization.

Third, in taking a combined theoretical perspective, this

study also contributes to the measurement of organizational

identification and identity salience. Identity theory explains

in detail the cognitive process of identifying the self, which

relies mainly on cognitive dimensions. Therefore, our finding

that the best measurements for organizational identification are

those in Mael and Ashforth’s (1992) scale fits with theory.

Fourth, the results indicate that identity salience can also be

measured by applying a 2-item approach without a decrease of

measurement quality—for example, by adopting Callero’s

(1985) scale. Researchers should use ‘‘Giving is an important

part of who I am’’ and ‘‘Giving means more to me than just

donating money/blood.’’

Fifth, the results indicate that researchers should be careful

in adopting the well-accepted items ‘‘willingness to donate

more,’’ ‘‘willingness to donate again,’’ and ‘‘recommendations

to family and friends’’ to measure blood donor loyalty; we

found low-quality measures. Therefore, additional research

on the measurement of blood donor loyalty is necessary.

Implications for Nonprofit Management

Our study also has several implications for managing donors of

nonprofit organizations in general. Our study reveals findings

that help derive managerial recommendations according to the

specific donation context (money or blood). First, for nonprofit

organizations that collect donations, relationships with their

donors are crucial. Our results across both nonprofit contexts

clearly prove that donor orientation, specifically in terms of

enhancing long-term relationships and acting in a relational

way, has a strong impact on donor satisfaction, identification,

and identity salience. Consequently, maintaining long-lasting

relationships with donors is of utmost importance. Therefore,

the instruments applied for relationship marketing must be

carefully adapted to the donation context. Organizations should

implement activities and initiatives that strengthen donors’

Boenigk and Helmig 545

perceptions that it is taking care, has the donors’ best interests in

mind, and works to establish long-term relationships. For exam-

ple, a nonprofit-oriented online community could encourage in-

depth discussions about topics of interest to donors and thereby

foster feelings of belonging to a group with shared values.

Second, our study shows that donor-nonprofit identification

(group level) and donor identity (individual level) are distinct

constructs and that identification influences identity salience;

thus, management should concentrate primarily on influencing

donor-nonprofit identification. We recommend that managers

integrate donor-nonprofit identification and donor identity

salience measures into their regular donor satisfaction surveys.

Annual measurement of identification constructs and an in-

depth investigation of their impact on donor loyalty and

donations would provide a foundation for additional activities.

Managers should also define identification goals for each target

group and develop strategies and actions to stimulate identifi-

cation. Note that across the two nonprofit industries, donor

satisfaction proved an important driver of donor identification.

Consequently, nonprofit managers should sensitize their

employees and volunteers to donor behavior to ensure high lev-

els of donor satisfaction, identification, and loyalty and imple-

ment training programs and incentive systems to achieve this.

Nonprofit organizations should also consider implementing

identification-oriented corporate communication strategies and

media campaigns. For example, to increase identification, the

media campaign by Doctors without Borders uses a virus

metaphor, such that all donors ‘‘infected’’ with the virus can

help spread the mission of the organization worldwide.

Third, regarding the instruments to strengthen donor identity

salience, managers should actively stimulate individual rela-

tionships with donors. For example, by issuing invitations to

events or providing quick and personal responses to questions,

organizations can treat donors as cocreators of value, not as

anonymous providers of resources. Through value cocreation,

nonprofit managers and donors become equally responsible for

decisions about fund-raising projects, campaigns, or new

strategies, which in turn can lead to strong identity salience.

In terms of context-specific recommendations, we stress that

in the blood donation setting, donor satisfaction plays a much

more important role in achieving donor loyalty than in the

money donation context. Thus, active management and excel-

lent service quality during the interaction between the donor

and the organization are crucial. In addition, both structure

quality and process quality management aspects should be

taken into account.

In the money donation context, to achieve donor loyalty,

identity salience is of greater importance than in the blood

donation context. Consequently, communications from non-

profit organizations that collect money should stress a good

‘‘fit’’ between the organization and the donor. Organizations

that are well informed about their donors can meet their needs

and wants more effectively. Here, communication and tone

should focus on targeting the individual donor, not the group.

In turn, achieving a high degree of donor loyalty will translate

into future donations more or less directly.

Finally, the direct influence of nonprofit organization activ-

ities on donor levels remains limited. To increase donations,

nonprofit organizations could advocate charitable giving as a

tool for transforming public well-being. In doing so, they could

emphasize that giving is an important part of people’s lives and

encourage them to donate to their organization.

Acknowledgments

The authors thank the three anonymous reviewers for their helpful

comments on previous versions of the article. Special thanks are due

to Kay Lemon for her balanced advice and recommendations during

this process. The management of this review process was superb. They

also thank Christian Ringle for his support and recommendations

regarding Cohen’s d analysis. The authors gratefully acknowledge the

Swiss National Science Foundation for providing funding to support

the research and the two nonprofit partners for support in data

collection.

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: Swiss

National Science Foundation (Grant No. FN 1606).

References

Ahearne, Michael, C. B. Bhattacharya, and Thomas Gruen (2005),

‘‘Antecedents and Consequences of Customer-Company Identifi-

cation: Expanding the Role of Relationship Marketing,’’ Journal

of Applied Psychology, 90 (3), 574-585.

Anderson, Eugene W. and Claes G. Fornell (1999), ‘‘The Customer

Satisfaction Index as a Leading Indicator,’’ in Handbook of Ser-

vices Marketing and Management, 255-270. Dawn Iacobucci and

Terri Swartz, eds. New York: Sage.

Arnett, Dennis B., Steve D. German, and Shelby D. Hunt (2003), ‘‘The

Identity Salience Model of Relationship Marketing Success. The

Case of Nonprofit Marketing,’’ Journal of Marketing, 67 (2),

89-105.

Ashforth, Blake E., Spencer H. Harrison, and Kevin G. Corley (2008),

‘‘Identification in Organizations: An Examination of Four Funda-

mental Questions,’’ Journal of Management, 34 (3), 325-374.

Bhattacharya, C. B. and Sankar Sen (2003), ‘‘Consumer–Company

Identification: A Framework for Understanding Consumers’ Rela-

tionships with Companies,’’ Journal of Marketing, 67 (2), 76-88.

Bhattacharya, C. B., Hayagreeva Rao, and Mary Ann Glynn (1995),

‘‘Understanding the Bond of Identification: An Investigation of Its

Correlates among Art Museum Members,’’ Journal of Marketing,

59 (4), 46-57.

Bodet, Guillaume and Iouri Bernache-Assollant (2011), ‘‘Consumer

Loyalty in Sport Spectatorship Services: The Relationships with

Consumer Satisfaction and Team Identification,’’ Psychology and

Marketing, 28 (8), 781-802.

Boenigk, Silke, Sigrun Leipnitz, and Christian Scherhag (2011),

‘‘Altruistic Values, Satisfaction and Loyalty among First-time

546 Journal of Service Research 16(4)

Blood Donors,’’ International Journal of Nonprofit and Voluntary

Sector Marketing, 16 (4), 356-370.

Bolton, Ruth N., Katherine N. Lemon, and Peter Verhoef (2004),

‘‘The Theoretical Underpinnings of Customer Asset Management:

A Framework and Propositions for Future Research,’’ Journal of

the Academy of Marketing Science, 32 (3), 271-293.

Brickson, Shelley L. (2012), ‘‘Athletes, Best Friends, and Social Acti-

vists: An Integrative Model Accounting for the Role of Identity in

Organizational Identification,’’ Organization Science, 24 (1), 226-

245.

Callero, Peter L. (1985), ‘‘Role-Identity Salience,’’ Social Psychology

Quarterly, 48 (3), 203-215.

Camarero, Carmen and Ma José Garrido (2011), ‘‘Incentives, Organi-

sational Identification, and Relationship Quality among Members

of Fine Arts Museums,’’ Journal of Service Management, 22 (2),

266-287.

Chattaraman, Veena, Sharron J. Lennon, and Nancy A. Rudd (2010),

‘‘Social Identity Salience: Effects on Identity-Based Brand

Choices of Hispanic Consumers,’’ Psychology and Marketing, 27

(3), 263-284.

Chin, Wynne W. (1998), ‘‘The Partial Least Squares Approach to

Structural Equation Modeling,’’ in Modern Methods for Business

Research, G. A. Marcoulides, ed. Mahwah, NJ: Lawrence

Erlbaum, 295-358.

Cohen, Jacob (1988), Statistical Power Analysis for the Behavioral

Sciences, 2nd ed. Hillsdale, NJ: Lawrence Erlbaum.

Cohen, Paul R., Adam Carlson, Lisa Ballesteros, and Robert Amand

(1993), ‘‘Automating Path Analysis for Building Causal Models

from Data,’’ in Proceedings of the Tenth International Conference

on Machine Learning, Technical Report 93-38. Amherst, MA:

Department of Computer Science, University of Massachusetts,

Amherst, 57-64.

Duque-Zuluaga, Lola C. and Ulrike Schneider (2008). ‘‘Market Orien-

tation and Organizational Performance in the Nonprofit Context:

Exploring Both Concepts and the Relationship Between Them,’’

Journal of Nonprofit & Public Sector Marketing, 19 (2), 25-47.

Eklöf, Jan and Irina Selivanova (2008), ‘‘Human Aspects in Service

Quality: EPSI Benchmark Studies,’’ Total Quality Management

and Business Excellence, 19 (9-10), 827-841.

Fombelle, Paul W., Cheryl Burke Jarvis, James Ward, and Lonnie

Ostrom (2012), ‘‘Leveraging Customers’ Multiple Identities: Iden-

tity Synergy as a Driver of Organizational Identification,’’ Journal

of the Academy of Marketing Science, 40 (4), 587-604.

Fornell, Claes and David F. Larcker (1981), ‘‘Evaluating Structural

Equation Models with Unobservable Variables and Measurement

Error,’’ Journal of Marketing Research, 18 (February), 39-50.

Fredrickson, Barbara L. and Christine Branigan (2005), ‘‘Positive

Emotions Broaden the Scope of Attention and Thought-Action

Repertoires,’’ Cognition and Emotion, 19 (3), 313-332.

Frumkin, Peter and Alice Andre-Clark (2000), ‘‘When Missions,

Markets, and Politics Collide: Values and Strategy in the Nonprofit

Human Services,’’ Nonprofit and Voluntary Sector Quarterly, 29

(1), 141-163.

Gwinner, Kevin and Scott R. Swanson (2003), ‘‘A Model of Fan Iden-

tification: Antecedents and Sponsorship Outcomes,’’ Journal of

Services Marketing, 17 (3), 275-294.

Hair, Joe F., Christian M. Ringle, and Marko Sarstedt (2012), Partial

Least Squares: The Better Approach to Structural Equation Model-

ing?’’ Long Range Planning, 45 (5-6), 312-319.

Hair, Joe F., Marko Sarstedt, Christian M. Ringle, and Jeannette A.

Mena (2012), An Assessment of the Use of Partial Least Squares

Structural Equation Modeling in Marketing Research,’’ Journal

of the Academy of Marketing Science, 40 (3), 414-433.

Helmig, Bernd and Julia Thaler (2010), ‘‘Nonprofit Marketing,’’ in

Third Sector Research, Taylor, Rupert, ed. New York: Springer,

151-170.

Helmig, Bernd, Marc Jegers, Irvine Lapsley, and Rupert Taylor

(2004), ‘‘Challenges in Managing Non-Profit Organisations:

A Research Overview,’’ Voluntas—International Journal of

Voluntary and Nonprofit Organizations, 15 (2), 101-116.

Henseler, Jörg and Marko Sarstedt (2012), ‘‘Goodness-of-Fit Indices

for Partial Least Squares Path Modeling,’’ Computational Statis-

tics, (March 4), 1-16.

Henseler, Jörg and Wynne W. Chin (2010), ‘‘A Comparison of

Approaches for the Analysis of Interaction Effects Between Latent

Variables Using Partial Least Squares Path Modeling,’’ Structural

Equation Modeling, 17 (1), 82-109.

Henseler, Jörg, Christian M. Ringle, and Rudolf R. Sinkovics (2009),

‘‘The Use of Partial Least Squares Path Modeling in International

Marketing,’’ in Advances in International Marketing, New

Challenges to International Marketing, Rudolf R. Sinkovics and

Pervez N. Ghauri (Volume Eds.), Vol. 20, 277-320.

Henseler, Jörg (2007), ‘‘A New Simple Approach to Multi-Group

Analysis in Partial Least Squares Path Modeling,’’ in Proceedings

of the 5th International Symposium on PLS and Related Methods

(PLS’07), Harald Martens, Tormod Naes and Magni Martens, eds.

Norway: Matsforsk, 104-107.

Hoelter, Jon W. (1983), ‘‘The Effects of Role Evaluation and Commit-

ment on Identity Salience,’’ Social Psychology Quarterly, 46 (2),

140-147.

Hogg, Michael A., Deborah J. Terry, and Katherine M. White (1995),

‘‘A Tale of Two Theories: A Critical Comparison of Identity

Theory,’’ Social Psychology Quarterly, 58 (4), 255-269.

Homburg, Christian, Jan Wieseke, and Wayne D. Hoyer (2009),

‘‘Social Identity and the Service-Profit Chain,’’ Journal of Market-

ing, 73 (March), 38-54.

Laverie, Debra and Dennis B. Arnett (2000), ‘‘Factors Affecting Fan

Attendance: The Influence of Satisfaction and Identity Salience,’’

Journal of Leisure Research, 32(2), 225-246.

Lee, Lichang, Jane Allyn Piliavin, and Vaughn R. A. Call (1999),

‘‘Giving Money, Time, and Blood: Similarities and Differences,’’

Social Psychology Quarterly, 62 (3), 276-290.

Lichtenstein, Donald R., Minette E. Drumwright, and Bridgette M

Braig (2004), ‘‘ The Effect of Corporate Social Responsibility on

Customer Donations to Corporate-Supported Nonprofits,’’ Journal

of Marketing, 68 (4), 16-32.

Lohmöller, Jan-Bernd (1989), Latent Variable Path Modeling with

Partial Least Squares. Heidelberg, Germany: Physica.

Mael, Fred and Blake E. Ashforth (1992), ‘‘Alumni and Their Alma

Mater: A Partial Test of the Reformulated Model of Organizational

Identification,’’ Journal of Organizational Behavior, 13 (2),

103-123.

Boenigk and Helmig 547

Marin, Longinos, Salvador Ruiz, and Alicia Rubio (2008), ‘‘The Role of

Identity Salience in the Effects of Corporate Social Responsibility

on Consumer Behavior,’’ Journal of Business Ethics, 84 (1), 65-78.

Maxham, James G., Richard G. Netemeyer, and Donald R. Lichten-

stein (2010), ‘‘The Retail Value Chain: Linking Employee Percep-

tion to Employee Performance, Customer Evaluations, and Store

Performance,’’ Marketing Science, 27 (2), 147-167.

Mittal, Vikas and Wagner A. Kamakura (2001), ‘‘Satisfaction,

Repurchase Intent, and Repurchase Behavior: Investigating the

Moderating Effect of Customer Characteristics,’’ Journal of Mar-

keting Research, 38 (1), 131-142.

Netemeyer, Richard G., Carrie M. Heilman, and James G. Maxham

III (2012), ‘‘Identification with the Retail Organization and

Customer-Perceived Employee Similarity: Effects on Customer

Spending,’’ Journal of Applied Psychology, 97 (4), 1049-1058.

Nunnally, Jum C. and Ira H. Bernstein (1994), Psychometric Theory,

3rd ed. New York: McGraw-Hill.

Olsen, Svein Ottar (2002), ‘‘Comparative Evaluation and the Relation-

ship Between Quality, Satisfaction, and Repurchase Loyalty,’’

Journal of the Academy of Marketing Science, 30 (3), 240-249.

Rigdon, Edward E., Christian M. Ringle, Marko Sarstedt, and Siegfried

P. Gudergan (2011), ‘‘Assessing Heterogeneity in Customer Satisfac-

tion Studies: Across Industry Similarities and within Industry Differ-

ences,’’ Advances in International Marketing, 10 (22), 169-194.

Ringle, Christian M., Sven Wende, and Alexander Will (2005),

SmartPLS 2.0 (beta), available at [www.smartpls.de.] Accessed

date, February 18, 2013

Rust, Roland T. and Anthony J. Zahorik (1993), ‘‘Customer Satisfac-

tion, Customer Retention, and Market Share,’’ Journal of Retail-

ing, 69 (Summer), 193-215.

Sargeant, Adrian and Elaine Jay (2010), Fundraising Management:

Analysis, Planning and Practice, 2nd ed. London, UK: Routledge.

Sargeant, Adrian and Lucy Woodliffe (2007), ‘‘Building Donor Loy-

alty: The Antecedents and Role of Commitment in the Context of

Charity Giving,’’ Journal of Nonprofit & Public Sector Marketing,

18 (2), 47-68.

Sarstedt, Marko, Jörg Henseler, and Christian M. Ringle (2011),

‘‘Multigroup Analysis in Partial Least Squares (PLS) Path Model-

ing: Alternative Methods and Empirical Results,’’ in Measurement

and Research Methods in International Marketing, Advances in

International Marketing, Vol. 22, Marko Sarstedt, Manfred

Schwaiger, and Charles R. Taylor, eds. London, UK: Emerald

Group Publishing, 195-218.

Sattler, Henrik, Franziska Völckner, Claudia Riediger, and Christian

Ringle (2010), ‘‘The Impact of Brand Extension Success Factors

on Brand Extension Price Premium,’’ International Journal of

Research in Marketing, 27 (4), 319-328.

Sen, Sankar and C. B. Bhattacharya (2001), ‘‘Does Doing Good

Always Lead to Doing Better? Consumer Reactions to Corporate

Social Responsibility,’’ Journal of Marketing Research, 38 (2),

225-244.

Shapiro, Stephen L. (2010), ‘‘Does Service Matter? An Examination

of Donor Perceptions of Service Quality in College Athletics,’’

Sport Marketing Quarterly, 19 (3), 154-165.

Stets, Jan E. and Peter Burke (2000), ‘‘Identity Theory and Social

Identity Theory,’’ Social Psychology Quarterly, 63 (3), 224-337.

Stryker, Sheldon and Peter Burke (2000), ‘‘The Present, Past, and

Future of an Identity Theory,’’ Social Psychology Quarterly, 63

(4), 284-297.

Tajfel, Henri (1974), ‘‘Social Identity and Intergroup Behaviour,’’

Social Science Information, 13 (2), 65-93.

Tenenhaus, Michel, Vincenzo Esposito Vinzi, Yves-Marie Chatelin,

and Carlo Lauro (2005), ‘‘PLS Path Modeling,’’ Computational

Statistics & Data Analysis, 48 (1), 159-205.

Van Dick, Rolf, Michael W. Grojean, Oliver Christ, and Jan Wieseke

(2006), ‘‘Identity and the Extra Mile: Relationships between

Organizational Identification and Organizational Citizenship

Behaviour,’’ British Journal of Management, 17 (4), 283-301.

Van, Dick and Ulrich Wagner (2002), ‘‘Social Identification among

School Teachers: Dimensions, Foci, and Correlates,’’ European

Journal of Work and Organizational Psychology, 11 (2), 129-149.

Wilson, Bradley, William Callaghan, Christian M. Ringle, and Jörg

Henseler (2007), ‘‘Exploring Causal Path Directionality for a Mar-

keting Model Using Cohen’s Path Method,’’ in Causalities

Explored by Indirect Observation: Proceedings of the 5th Interna-

tional Symposium on PLS and Related Methods (PLS’07), H. Mar-

tens, T. Naes, M. and Martens eds. Norway: Matforsk, 57-61.

Wisner, Priscilla S., Anne Stringfellow, William E. Youngdahl, and

Lenore Parker (2005), ‘‘The Service Volunteer: Loyalty Chain:

An Exploratory Study of Charitable Not-for-Profit Service Organi-

zations,’’ Journal of Operations Management, 23 (2), 143-161.

Wold, Herman (1982), ‘‘Soft Modeling: The Basic Design and Some

Extensions,’’ in Systems Under Indirect Observations: Part II,

Karl G. Jöreskog and Herman O. A. Wold, eds. Amsterdam,

Netherlands: North-Holland, 1-54.

Author Biographies

Silke Boenigk is a professor of business administration and chair of

management of public, private, and nonprofit organizations at the

University of Hamburg, Germany. Her current research interests are

in the area of nonprofit service management, donor relationship man-

agement, fund-raising, blood donation management, and cause-related

marketing. Her work has been published in journals such as Nonprofit

Voluntary Sector Quarterly, Nonprofit Management & Leadership,

International Journal of Nonprofit and Voluntary Sector Marketing,

Public Administration, Public Administration Review, Journal of

Marketing Management, and Journal of Relationship Marketing.

Bernd Helmig is a professor business administration and chair of

public and nonprofit management at the University of Mannheim,

Germany. His primary areas of research interest are management

of public and nonprofit institutions, service management, and sta-

keholder relationship management. His work has been published in

journals such as Business & Society, European Management Jour-

nal, Journal of Advertising Research, Journal of Marketing Man-

agement, Journal of Public and Nonprofit Sector Marketing,

Journal of Relationship Marketing, Nonprofit and Voluntary Sector

Quarterly, Public Management Review, Schmalenbach Business

Review, and Voluntas: International Journal of Voluntary and

Nonprofit Organizations.

548 Journal of Service Research 16(4)


Comments are closed.