Linking actual self-congruence, ideal self-congruence, and functional congruence to donor loyalty: the moderating role of issue involvement

Nonprofit organizations need loyal donors to accomplish their mission and goals. However, organizations lose up to 60% of their first-time donors. Therefore, a deeper understanding of the determinants of donor loyalty is necessary. This study tests whether actual self-congruence, ideal self-congruence, and functional congruence strengthen the tie between nonprofit organizations and donors. In addition, and as an extension of the self-congruence theory, issue involvement is added as a moderator variable. The study results show that ideal self-congruence and functional congruence have a positive effect on donor loyalty, while actual self-congruence does not affect donor loyalty. Moreover, the study results show that issue involvement strengthens (mitigates) the effect of actual self-congruence (ideal self-congruence) on donor loyalty. Regarding the relationship between functional congruence and donor loyalty, issue involvement has no moderating effect. Based on the empirical findings, this study discusses academic implications and presents managerial implications for fundraisers and nonprofit organizations.


Introduction
There is a broad consensus that nonprofit organizations need loyal donors because without these donations, nonprofit organizations would have much fewer or no resources with which to achieve their mission and goals (e.g., Lee and Chang 2007;Traeger and Alfes 2019;Wymer and Rundle-Thiele 2016). However, maintaining donor loyalty requires organizations' efforts and is a challenge that nonprofit organizations must address because competition in the fundraising market has intensified in the past decades (Topaloglu et al. 2018). Many organizations lose up to 60% of their first-time donors (Sargeant and Woodliffe 2007;Topaloglu et al. 2018). It is therefore more important than ever to establish a relationship with donors (von Schnurbein 2016). Hence, nonprofit organizations must find ways to strengthen the tie to their donors.
Consumer psychology research has concluded that self-congruence and functional congruence positively influence individual behavior (e.g., brand loyalty) (see, e.g., Bauer et al. 2006;Huber et al. 2018;Choi and Reid 2016;Sirgy et al. 1991). Self-congruence, which is symbolically driven, describes the perceived similarity between one's personality and the personality of an organization (see, e.g., Malär et al. 2011;Zhu et al. 2019). Thus, self-congruence is based on the perception of the brand as a global construct rather than on the brand's objects (Wymer and Akbar 2017). For-profit organizations have been using this approach to influence consumer behavior positively for decades (Japutra et al. 2019;Nienstedt et al. 2012;Park et al. 2010;Sirgy et al. 1991). Some organizations advertise their brands by cooperating with celebrities to match consumers' ideal self-concept, while other organizations use "common people" in their ads to match their brand personality with consumers' actual self-concept (Malär et al. 2011).
While self-congruence is symbolically driven, functional congruence is utilitarian driven and describes the extent to which an organization's functional attributes fulfill the consumer's expectations with respect to attribute performance (Batra and Ahtola 1991;Choi and Reid 2016;Sirgy et al. 1991). Hence, functional congruence focuses on utilitarian values, while self-congruence focuses on symbolic values (Kressmann et al. 2006;Lee et al. 2017;Sirgy et al. 1991).
Previous studies have especially focused on the effects of self-congruence and functional congruence in the fields of brands and products (Japutra et al. 2019;Karjaluoto et al. 2019;Tan et al. 2019;Zhu et al. 2019), tourism and hospitality (Hung et al. 2019;Lee et al. 2017;Su and Reynolds 2017), and retail (Das 2014;Hedhli et al. 2017;Kwon et al. 2016) to better understand how to attach consumers to a brand, a destination, or a shopping mall. Until now, third-sector research has focused on the fit of participants' own personality and prototypical volunteers' personality and brand/cause fit (Prajecus and Olsen 2004;Van Vianen et al. 2008). Moreover, it has been demonstrated that volunteers who collaborate with different types of nonprofit organizations have different self-concepts. With respect to this, the authors Linking actual self-congruence, ideal self-congruence,… were able to confirm a positive relationship between volunteers' personalities and organizations' personalities (Beerli et al. 2004).
However, there is a need for research in the third-sector to analyze whether (potential) donors can be attached to organizations via self-congruence and/or functional congruence. Wymer and Akbar (2017) have highlighted the importance of this research question by suggesting that self-congruence is demanded to understand how individuals (e.g., donors) perceive an organization. Moreover, this research question is relevant because the resulting knowledge will help answer the question of whether nonprofit organizations should focus on donors' personalities or donors' expectations regarding utilitarian values in fundraising campaigns. Previous consumer psychology research has emphasized the importance of this research question by demonstrating that the relevance of self-congruence and functional congruence can vary depending on the research context and consumer attributes (e.g., consumer health consciousness) (e.g., Choi and Reid 2016;Malär et al. 2011;Su and Reynolds 2017).
Therefore, besides analyzing the effects of self-congruence and functional congruence on donor behavior, we also refine the congruence model by exploring whether issue involvement moderates the effect of congruence. According to consumer psychology research (e.g., Kaufmann et al. 2016;Kressmann et al. 2006;Malär et al. 2011), we expect that issue involvement will have an accelerating effect on the influence of functional congruence and that the moderating effect of issue involvement on the influence of self-congruence on donor behavior will be twofold given that self-congruence can be divided into actual and ideal self-congruence.
To answer our research questions, we begin by first providing the theoretical underpinnings of the formation and effects of self-congruence and functional congruence. In doing so, we theoretically argue why actual and ideal self-congruence as well as functional congruence are important drivers of donor behavior. Moreover, we demonstrate the accelerating and dampening moderating effects of issue involvement, and subsequently, we present an empirical analysis and discuss the results of that analysis.

Self-congruence and functional congruence
According to Rosenberg (1979), a person's self-concept is the sum of thoughts and emotions regarding the person himself or herself. That is, the self-concept is the image a person has about himself or herself (Hong and Zinkhan 1995). On this basis, self-congruence describes the extent to which one`s own self-concept matches the self-concept of another individual or entity (e.g., brands, tourism destinations, products, celebrities) (see, e.g., Huber et al. 2012;Liberatore and Tscheulin 2011;Sirgy et al. 1991;Zhu et al. 2019).
Self-congruence can be divided into actual and ideal dimensions (Malär et al. 2011;Zhu et al. 2019). Hence, if someone compares another's self-concept with his/her actual self-concept (i.e., one's own personality in the here and now), the result of this comparison process is called actual self-congruence. Conversely, if someone compares another's self-concept with his/her ideal self-concept (i.e., the person one would like to be), the result relates to the so-called ideal self-congruence (Huber et al. 2018;Malär et al. 2018).
Previous research has revealed positive effects of actual and ideal self-congruence on consumer behavior (e.g., Dolich 1969;Hu et al. 2017). Actual and ideal self-congruence positively influence consumer behavior because the higher perceived similarity (i.e., self-congruence) compels consumers to feel more consistency and harmony between themselves and the brand (Malär et al. 2011). Consequently, actual and ideal self-congruence strengthen the relationship between consumers and a brand (Huber et al. 2018;Japutra et al. 2019), which results in an increase in brand loyalty (Hung et al. 2019;Kang et al. 2015).
The impact of self-congruence on individuals' behavior can be explained by two central behavioral motives (Claiborne and Sirgy 1990;Sirgy 1982). First, the need for self-consistency motivates individuals to consume brands whose selfconcept matches their actual self-concept (Sirgy 1982). This behavioral driver is based on the need for self-verification to confirm personal values and avoid the risk associated with changes (Liberman et al. 1999;Swann 1983). Second, the need for self-enhancement motivates individuals to consume brands that match their ideal self-concept (Claiborne and Sirgy 1990). Based on this motive, subjects desire to enhance their self-concept and thus increase their self-esteem by closing the gap to their ideal self by means of consumption (Fiske 2001;Malär et al. 2011).
Since previous consumer research has indicated that the effects of actual and ideal self-congruence depend on the considered context (see, e.g., Ahn et al. 2013;Ekinci et al. 2008;Huber et al. 2018;Zhu et al. 2019), the question arises whether these constructs have a determining role in the nonprofit organization sector context (i.e., donor behavior). Previous research regarding nonprofit organizations has concluded that donor behavior is driven by personal values (i.e., verifying personal norms) as well as by the need for self-enhancement (e.g., Konrath and Handy 2018;Ottoni-Wilhelm 2017). This is also supported by Beerli et al. (2004), who concluded that volunteers of different nonprofit organizations have different personalities and that there is a fit between volunteers' personalities and organizations' personalities. Consequently, it is determined that both self-verification (i.e., basis for actual self-congruence) and self-enhancement (i.e., basis for ideal self-congruence) play an important role in predicting donor loyalty. Moreover, recapitulating that actual and ideal self-congruence influence individuals' behaviors in several contexts (e.g., Choi and Reid 2016;Kaufmann et al. 2016;Zhu et al. 2019) leads to the conclusion that self-congruence has a positive effect on donor loyalty. Therefore, hypotheses H1 and H2 are postulated as follows: H1 Actual self-congruence has a positive effect on donor loyalty.
H2 Ideal self-congruence has a positive effect on donor loyalty.
Linking actual self-congruence, ideal self-congruence,… In contrast to self-congruence, functional congruence is based on (e.g., brands', products') functional characteristics and reflects the matching of expected and perceived functional characteristics (Choi and Reid 2016;Johar and Sirgy 1991;Sirgy et al. 1991). Functional characteristics can vary in different contexts. For example, in the tourism sector, performance attributes, such as a hotel's interior design and food quality, may contribute to the fulfillment of tourist needs (Ahn et al. 2013), while in the food context, product attributes, such as fiber and vitamin content, have a utilitarian function (Choi and Reid 2016). Consequently, functional congruence is based on individuals' utilitarian needs, which can be satisfied by the performance of products or services (Batra and Ahtola 1991).
Previous research has revealed positive effects of functional congruence on consumer behavior. Lee et al. (2017) and Su and Reynolds (2018) demonstrated that functional congruence plays an important role in the hospitality context by confirming the positive effect of functional congruence on consumer satisfaction and brand attitude. Choi and Reid (2016) confirmed the positive effect of functional congruence on consumer behavior in the food context, while Hung and Petrick (2012) confirmed the positive effect in the tourism context. With respect to nonprofit organizations, Willems et al. (2016) state that the importance of organizations' performance has been increasing for decades. The relevance of organizations' performance is supported by Sargeant et al. (2006), who showed that an organization's performance has a positive effect on donor behavior (i.e., trust and commitment).
Based on the above chain of argumentation regarding the effect of functional congruence on consumer behavior in several behavioral domains and on the importance of organization performance in predicting donor behavior, hypothesis H3 is postulated as follows: H3 Functional congruence has a positive effect on donor loyalty.
Studies that explore congruence effects should consider that actual and ideal self-congruence are generated more subconsciously and that self-congruence evolves before people evaluate the functional characteristics of which they are more aware. This results in an evaluation bias because subjects holding higher self-congruence perceptions (i.e., more positive emotions) evaluate functional characteristics more positively than subjects with lower self-congruence perceptions do (Choi and Reid 2016;Kressmann et al. 2006;Sirgy et al. 1991). Numerous studies have confirmed the positive effect of actual and ideal self-congruence on functional congruence in several contexts. Lee et al. (2017) determined that self-congruence based on the perception of air lounges positively influences functional congruence, while Kang et al. (2015) investigated the effect of self-congruence based on coffee shop images and confirmed the positive effect. The findings of Ryu and Lee (2013) also confirmed that self-congruence based on attendees' perception of a convention positively influences functional congruence. Consequently, it can be stated that actual and ideal self-congruence positively influence functional congruence (Huber et al. 2010 Sirgy et al. 1991). Thus we expect actual and ideal self-congruence, which are measured based on the nonprofit organization's personality, to positively influence the perception of the organization's functional attributes. Hence, hypotheses H4 and H5 are as follows: H4 Actual self-congruence has a positive effect on functional congruence.
H5 Ideal self-congruence has a positive effect on functional congruence.

Issue involvement
The moderator proposed in our study, involvement, represents the perceived relevance of an object, situation, or action for a person (Celsi and Olson 1988). The personal relevance of the same object, situation, or action can differ among individuals and is strongly related to subjects' enthusiasm and personal values (Kaufmann et al. 2016;Malär et al. 2011;Sirgy et al. 2008). While product involvement determines the relevance of a product to an individual (Kaufmann et al. 2016;Malär et al. 2011), issue involvement determines subjects' involvement with, for instance, nonprofit organizations' activities and mission.
As discussed above, confirming one's own personal values and verifying the individual self-concept (i.e., actual self-congruence) as well as enhancing one's own self-concept (i.e., ideal self-congruence) explain the effect of self-congruence on individual behavior (Claiborne and Sirgy 1990;Sirgy 1982). As research shows, the relevance of self-verification (i.e., actual self-congruence) increases when people deal more closely with an issue and are therefore more involved (see, e.g., Kressmann et al. 2006;Malär et al. 2011;Sirgy et al. 2008;Swann et al. 1990). Recapitulating that higher issue involvement means that people deal more intensively with nonprofit organizations' activities and missions and that the relevance of selfverification increases in such situations, it can be assumed that issue involvement strengthens the effect of actual self-congruence on donor loyalty. Hence, hypothesis H6 is as follows: H6 Issue involvement strengthens the relationship between actual self-congruence and donor loyalty.
While we expect self-verification (i.e., actual self-congruence) to become more important when issue involvement increases, we assume that self-enhancement (i.e., ideal self-congruence) becomes less relevant when issue involvement increases. Our reasoning is based on the finding that individuals tend to avoid self-enhancing interactions in high-involvement situations (Malär et al. 2011;Swann et al. 1990). It has been shown that people do not want to interact with ideals (e.g., ideal brands; ideal celebrities) when they are highly involved in a situation (Malär et al. 2011).
This behavioral tendency can be explained by considering the self-evaluationmaintenance model (Malär et al. 2011;Tesser 1988). This model confirms that ideals cause negative emotions and threaten self-evaluation when people deal with something (more) intensively. Due to the anticipation of these negative emotions, in such situations (i.e., high-involvement situations), an individual puts distance between herself/himself and an ideal person or an ideal organization (Malär et al. 2011). Transferring this line of reasoning to donor behavior means that an "ideal organization" will threaten donors' self-evaluation by showing that the donor has not achieved her/his ideals in donating. However, if donors' emotions are not aroused (i.e., low-involvement), the donors will not consider in-depth whether they have acted as they (ideally) should. Accordingly, we expect issue involvement to mitigate the effect of ideal self-congruence on donor loyalty. Hypothesis H7 therefore reads as follows: H7 Issue involvement mitigates the relationship between ideal self-congruence and donor loyalty.
As discussed above, functional congruence is based on utilitarian values (e.g., Choi and Reid 2016). Research has shown that utilitarian values become more important when individuals perceive a situation or action to be of high importance or relevance (e.g., Celsi and Olson 1988;Kressmann et al. 2006;Sirgy et al. 2008). In such high-involvement situations, individuals invest more cognitive efforts in thinking of an action (e.g., purchasing an expensive product) than they do in low-involvement situations (e.g., purchasing an affordable product) (e.g., Fritz et al. 2017;Malär et al. 2011). Consequently, individuals are more aware of utilitarian values such as attribute performance in high-involvement situations (e.g., Kressmann et al. 2006). Therefore, it can be expected that subjects who deal more intensively with issues of nonprofit organization (i.e., highly involved subjects) attach more importance and relevance to utilitarian attributes. Thus, we postulate hypothesis H8 as follows: H8 Issue involvement strengthens the relationship between functional congruence and donor loyalty.

Data collection and sample
To test our hypotheses, we conducted a survey study with 257 German donors. Data were collected via the online panel provider Qualtrics (http://www.qualt rics.com). In 2018, 53% of German donors were female, and 65% of German donors were older than 50 years (Deutscher Spendenrat e.V. 2019). In our sample, 51% of donors were female; however, only 51% of donors were older than 50 years. In 2018, the average donation amount per donor was approximately 260€ (Deutscher Spendenrat e.V. 2019). In our sample, the revealed average donation amount per person and per year was 283€. Table 1 presents the sample characteristics.
At the beginning of the online survey, a definition of nonprofit organizations was given to the interviewees. We informed the respondents that nonprofit organizations have a charitable purpose. Then, we asked the subjects whether they had donated to a nonprofit organization in the past 3 years. The subjects could select between different types of organizations (e.g., charitable organizations, disaster organizations, development aid associations, political organizations). We were interested in charitable organizations as well as disaster relief and development aid organizations, as the involved issues relate to the same context (the overall goal of those organizations was helping beneficiaries: e.g., disaster victims, care-dependent people). If the subjects answered that they had donated to such organizations, they were asked to give the name of the organization. If the subjects answered that they had not donated to such organizations in the past 3 years, they were excluded from the survey.

Research design and measures
The measurement scales used were previously empirically validated by other studies and are depicted in Table 2. All considered measurement instruments used sevenpoint Likert scales (1 = strongly disagree, 7 = strongly agree). First, the predictor variables self-congruence and functional congruence were measured. We measured self-congruence according to Malär et al. (2011) and Sirgy et al. (1997). Sirgy et al. (1997) showed that measuring self-congruence in a direct, global way without using predetermined attributes is more accurate than measuring it by discrepancy indices and using specific attributes. However, subjects should first take time to think about the object's personality; therefore, following the introduction but before asking about self-congruence, we instructed the subjects (Malär et al. 2011;Sirgy et al. 1997) to "Take a moment to think about organization 'XY'. Think about this organization's personality. Imagine this organization and describe this organization in your mind's eye using one or more personality characteristics, such as responsible, active, aggressive, simple, emotional, or other adjectives you would use to describe this organization". Next, we measured the actual and ideal self-congruence with three items each.
According to several research articles (e.g., Choi and Reid 2016;Kressmann et al. 2006;Sirgy et al. 1991), functional congruence is a multi-attribute construct, and these attributes vary between different objects of evaluation (Choi and Reid 2016;Hung and Petrick 2012). Because we surveyed functional congruence related to nonprofit organizations' performance, we used seven functional characteristics according to an approach developed by Willems et al. (2016). These characteristics can be used to assess nonprofit organizations' performance, and the construct is specified as a reflective construct (see, e.g., Kressmann et al. 2006;Willems et al. 2016). In a next step, the subjects evaluated the functional characteristics by answering how closely the respective nonprofit organization had met these functional attributes. Moreover, we used four items based on (Malär et al. 2011) to measure the moderator issue involvement. We then measured donor loyalty by using two items based on measurement approaches of Kuppelwieser and Sarstedt (2014), Suh and Youjae (2006) and Zogaj et al. (2019) who used word-of-mouth and future behavior items (i.e., considering the company as first choice in future) to measure loyalty.
Additionally, we considered the subjects' age, gender, familiarity with the organization, and two social desirability bias factors as control variables. As Lee and Chang (2007) show, age and gender influence donation behavior. Social desirability bias is also an important factor when querying subjects regarding their giving and helping behavior. This is because giving and helping are classified as prosocial behaviors, and the motivation to give socially desirable answers may influence subjects' answers in survey research (Lee and Sargeant 2011). We used a two-dimensional approach to measure the social desirability bias (i.e., self-deceptive enhancement and impression management) developed by Winkler et al. (2006).
Each of the dimensions was measured by one item (i.e., "My first impression of people usually turns out to be correct"; "I have received too much change before and have said nothing" (reverse coded)). Nonprofit brand familiarity was also considered as a covariate because it has been shown that familiarity might have an influence on the perception of nonprofit organizations (Farwell et al. 2019) and on the effects of Actual self-congruence (based on Malär et al. 2011;Sirgy et al. 1997) 1. The personality of "nonprofit organization x" is consistent with how I see myself (my actual self) 2. "Nonprofit organization x" reflects who I am 3. The personality of "nonprofit organization x" is a mirror image of me (my actual self) Ideal self-congruence (based on Malär et al. 2011;Sirgy et al. 1997) 1. The personality of "nonprofit organization x" is consistent with how I would like to be (my ideal self) 2. "Nonprofit organization x" reflects who I would like to be 3. The personality of "nonprofit organization x" is a mirror image of the person I would like to be (my ideal self) Functional congruence (based on Willems et al. 2016) 1. "Nonprofit organization x" uses available resources optimally 2. "Nonprofit organization x" deploys the predefined strategy 3. "Nonprofit organization x" brings the planned strategic actions to a good end 4. "Nonprofit organization x" keeps the vision and mission up to date 5. "Nonprofit organization x" creates efficient output from the limited means available 6. "Nonprofit organization x" brings positive changes to our society 7. "Nonprofit organization x" realizes services to beneficiaries 1. In the future, I will be loyal to "nonprofit organization x" 2. I will tell my friends and acquaintances only good things about "nonprofit organization x" Issue involvement (based on Malär et al. 2011) 1. Because of my personal values, I feel that this is an issue that ought to be important to me 2. This issue is very important to me personally 3. Compared with other issues, this issue is important to me 4. I'm interested in this issue self-congruence on individuals' behavior (Malär et al. 2011). We measured familiarity by using one item (i.e., "I know the organization XY and its activities") based on Malär et al. (2012).

Reliability, validity, and model fit
We ran a reliability analysis in SPSS 25 as well as a confirmatory factor analysis in AMOS 25 to check the constructs' validity and the goodness-of-fit of the hypothesized five-factor models. Tables 3 and 4 show the results of the reliability and validity analyses.
All Cronbach's alpha values and composite reliability values are above 0.80, indicating appropriate internal consistency (Nunnally 1978). All standardized regression weights are significant and above the commonly used thresholds: Actual self-congruence ≥ 0.853, ideal self-congruence ≥ 0.882, functional congruence ≥ 0.679; donor loyalty ≥ 0.773, and issue involvement ≥ 0.796. Next, we verified discriminant validity based on three approaches. First, we performed χ 2 difference tests (i.e. phi difference test). To do this, we constrained the covariance between two constructs to the value of 1 and sequentially assessed in pairs whether the χ 2 -value of the unconstrained model differs from that of the constrained models (see, e.g., Anderson and Gerbing 1988;Franke and Sarstedt 2019). The results indicate that the χ 2 -values of the restricted models are higher than the χ 2 -value of the freely estimated model. However, as not all of the χ 2 -values related to the constrained models are significantly worse, we conducted an HTMT (heterotrait-monotrait) analysis to assess construct discriminant validity (Franke and Sarstedt 2019). According to the HTMT analysis, all constructs are distinguishable as all values are below the threshold of 0.850 (Henseler et al. 2015). Third, we used an alternative heuristic approach to confirm discriminant validity. To do so, we estimated AVE values and evaluated the Fornell-Larcker criteria. The AVE for each construct is higher than the threshold of 0.50. Furthermore, the square root of the AVEs is higher than the inter-construct correlations (Fornell-Larcker criterion), indicating discriminant validity. Consequently, according to these three approaches (i.e., phi test, HTMT analysis, and Fornell-Larcker criteria), it was determined that the constructs exhibit good discriminant validity. Further, we conducted a test on multi-collinearity in SPSS (version 25) to check for independence between the analyzed constructs. As proposed in the theoretical model, we selected donor loyalty as the dependent variable, while actual self-congruence, ideal self-congruence, functional congruence, and issue involvement were selected as the independent variables. First, we conducted the Durbin-Watson test to assess the independence of the observations. The Durbin-Watson value was 1.846 (i.e. greater than 1.5 and less than 2.5), thus confirming that the observations were independent. Moreover, we examined the VIF (variance inflation factors) values and found that the VIF values were less than 3, indicating that low multi-collinearity does not negatively impact the results (see, e.g., Hair et al. 2010): actual self-congruence = 2.791; ideal self-congruence = 2.646; functional congruence = 1.624; and issue involvement = 2.013.
Next, we checked the measurement model's overall goodness-of-fit using the thresholds recommended by Byrne (2016) and Schreiber et al. (2006). The model yielded a χ 2 value of 323.954, with 142 degrees of freedom. The χ 2 /df ratio value is 2.281 and therefore is lower than the threshold of 3.0. The standardized RMR is 0.035 (< 0.05) and indicates a well-fitting model. The CFI value was 0.957 (> 0.95), which also indicated that the model fits the data well. Consistently, the IFI value of 0.957 and the TLI value of 0.948 indicate a good model fit. Finally, RMSEA = 0.071 is smaller than the threshold of 0.08, also indicating appropriate model fit. Because of the good overall model fit, we did not conduct any post hoc modifications of measurement scales used.

Hypotheses testing
A basic model for testing direct effects and the effects of actual and ideal self-congruence on functional congruence (i.e., hypotheses H1-H5) as well as a moderated model for checking moderating effects (i.e., hypotheses H6-H8) were estimated. We used AMOS 25 to estimate these models. We estimated the basic model in a first Linking actual self-congruence, ideal self-congruence,… step and analyzed the unconditional effects of actual self-congruence, ideal selfcongruence and functional congruence on donor loyalty. Moreover, we considered actual and ideal self-congruence to have a direct effect on functional congruence (as shown in Fig. 1). Furthermore, we added the predicted moderating variable as a covariate to control for its influence on subjects' behavior when estimating the effects of self-congruence and functional congruence on donor loyalty, as recommended by the literature (Hayes 2018). In the following table (Table 5), the basic model's results are reported.
The results reveal significant positive effects of ideal self-congruence (b = 0.148, p = 0.004) and functional congruence (b = 0.237, p = 0.000) on donor loyalty. However, actual self-congruence (b = 0.067, p = 0.237) has no significant influence on donor loyalty. Consequently, H2 and H3 can be confirmed. In contrast, H1 must be rejected. Furthermore, supporting other studies' results and in support of the external validity of our analysis, actual (b = 0.213, p = 002) and ideal (b = 0.190, p = 0.003) self-congruence have a positive effect on functional congruence. Consequently, hypotheses H4 and H5 can be confirmed.
To validate the hypothesized moderating effects, we eliminated the constraint of unconditionally according to Hayes (2018) by expanding the model with issue involvement as a moderating variable. The expanded model is identical to the basic  Table 6.
As hypothesized, the empirical results show that issue involvement has different moderating effects on the influence of actual and ideal self-congruence on donor loyalty. While issue involvement has a positive moderating effect on the relation between actual self-congruence and donor loyalty (b = 0.096, p = 0.049), it has a negative moderating effect on the relation between ideal self-congruence and donor loyalty (b = − 0.117, p = 0.012). Therefore, hypotheses H6 and H7 can be supported. However, the results do not confirm our hypothesis that issue involvement strengthens the impact of functional congruence on donor loyalty (b = 0.014, p = 0.178). Accordingly, there is no support for hypothesis H8.
In a next step we performed a median split along the values of issue involvement to define two sub samples (low and high involved donors) according to Kaufmann et al. (2016) and Malär et al. (2011). We then analyzed the basic model in both subsamples. The estimates support the results of the above conducted moderating analyses, where issue involvement was considered as continuous variable. Hence, it can be stated that the influence of actual (ideal) self-congruence is higher (lower) for donors with high (low) issue involvement than for donors with low (high) issue involvement, supporting hypotheses H6 and H7 (Table 7).

Discussion
Previous consumer research has shown that the influence of actual and ideal selfcongruence depends on the considered behavioral domain (see, e.g., Ahn et al. 2013;Ekinci et al. 2008;Huber et al. 2018;Zhu et al. 2019). Malär et al. (2011) showed in their research article that there is no unconditional effect of ideal self-congruence on individual behavior while there is a strong effect of actual self-congruence. Recently, Zhu et al. (2019) found that the effects of actual and ideal self-congruence depend on the type of brand.
More precisely they found that actual self-congruence (ideal self-congruence) has a stronger effect on attitudes towards functional brands (symbolic brands). Our results confirm the importance of the considered context regarding the self-congruence effects by showing that ideal self-congruence predicts donor loyalty, while there is no (unconditional) effect of actual self-congruence on donor loyalty. This means that, in the case of nonprofit organizations, ideal self-congruence is more important than actual self-congruence in predicting individual behavior. Therefore, matching donors' ideal personality has a stronger effect on their loyalty than matching donors' actual personality. Considering that ideal self-congruence, which is based on self-enhancement, positively influences donor behavior, and that self-enhancement serves to close the gap between the actual self in the here and now and the ideal self (see, e.g., Claiborne and Sirgy 1990;Malär et al. 2018), it can be postulated that donors aim to enhance their selves by supporting nonprofit organizations through donations and word-ofmouth (i.e., donor loyalty). This supports the notion that egoistic motives, such as "donating makes me feel needed" (Konrath and Handy 2018) and "enhancing the one's own self-esteem" (Ottoni-Wilhelm 2017) have a significant role in predicting donor behavior.
Functional congruence represents donors' expectations and satisfaction regarding an organization's performance (see, e.g., Choi and Reid 2016;Kressmann et al. 2006;Willems et al. 2016). As our results reveal, functional congruence predicts donor loyalty. Connecting this result to the finding that ideal self-congruence drives donor loyalty shows that donors want to support organizations that are ideal but that also have high organizational performance and use donations efficiently to accomplish the stated ideals and goals. In addition, the direct positive effect of functional congruence on donor loyalty confirms and extends the findings of Sargeant et al. (2006), who revealed that an organization's performance influences giving behavior indirectly via trust and commitment.
Furthermore, our research aimed to analyze whether issue involvement influences the effects of actual self-congruence, ideal self-congruence, and functional congruence on donor loyalty. Our results indicate that issue involvement strengthens (mitigates) the influence of actual (ideal) self-congruence on donor loyalty. Previous research has led to divergent effects: While Malär et al. (2011) were able to confirm that involvement strengthens (weakens) the effect of actual (ideal) self-congruence on individual behavior, Kressmann et al. (2006) were not able to support the moderating role of involvement regarding the influence of actual and ideal self-congruence on relationship quality.
A recent study by Kaufmann et al. (2016) also tested the moderating role of involvement. As the authors conclude in their article, the sign of the moderator product involvement was in line with the findings of Malär et al. (2011). Our results give new insights for the donating context by showing that in the case of nonprofit organizations issue involvement strengthens (mitigates) the need to verify the actual (ideal) self-concept. That is, the more intensively donors deal with the organization's issue, the more important self-verification becomes (e.g., verifying personal norms). Furthermore, the more intensively donors deal with the organization's issue, the less important self-enhancement becomes.

Managerial implications
The purpose of our research was to develop a better understanding of how nonprofit organizations can strengthen donor loyalty. This is of vital importance because without loyal donors, nonprofit organizations would have great difficulties in accomplishing their mission (e.g., Lee and Chang 2007; Traeger and Alfes 2019; Wymer and Rundle-Thiele 2016).
The first managerial implication is based on the positive effect of ideal selfcongruence on donor loyalty. This finding indicates that nonprofit organizations should focus on donors' ideal self-concept to strengthen the organization-donor tie. To accomplish this goal, nonprofit organizations should first examine donors' ideal self-concept. Then, they should create marketing campaigns to match donors' ideal self-concept and to show donors that they can enhance their selfconcept by donating. A viable way to achieve this goal is to cooperate and advertise with celebrities. However, first organizations should examine which celebrities are perceived as ideals by prospective donors because many organizations do not know their customers (Bruhn et al. 2012). Subsequently, they could create marketing campaigns with the accordant celebrities to match donors' ideal self-concept.
The second implication is based on the positive effect of functional congruence on donor loyalty. This finding suggests that donors care about the organization's performance when donating money or spreading word-of-mouth (i.e., donor loyalty). Therefore, nonprofit organizations should in particular communicate success in the achievement of their mission. A viable way to do that would be to attract certificates and good ratings to signal to prospective donors that their donations are being effectively used. Another way would be to communicate successful projects. In this way, organizations could at least influence donor behavior via functional congruence. However, it should be noted that nonprofit organizations should enhance transparency and communicate ratings that are known to communicate organizational performance because transparency is associated with higher levels of trust (Farwell et al. 2019). Moreover, it should be noted that ratings that are not commonly recognized do not affect donor behavior (Sloan 2009).
The third implication is based on the positive (negative) moderation effect of issue involvement on the influence of actual (ideal) self-congruence on donor loyalty. Accordingly, highly involved donors (e.g., donors who are engaged in the activities of several organizations) want to interact with ideals less than less involved donors (e.g., donors who donate once a year and have no other connection to the organization or other organizations). Therefore, when issue involvement is low, donors perceive "ideals" to be more positive. However, when issue involvement is high, donors are more aware of and perceive the gap between their (i.e., donors') actual personality in the here-and-now and their ideal self-concept.
Thus, in order not to cause negative emotions and threaten donors' self-evaluation marketing managers need to be careful with aspirational branding strategies when donors are highly involved. Consequently, aspirational branding strategies should rather be implemented when communicating with donors with low involvement. Otherwise, we recommend authentic branding strategies when donors are highly involved because issue involvement strengthens the relevance of actual selfcongruence in predicting donor loyalty.
The fourth implication is based on the finding that issue involvement does not moderate the effect of functional congruence on donor loyalty. Consequently, nonprofit organizations should emphasize functional attributes (see above: successful projects, organizational performance certificates, ratings, etc.) regardless of donor involvement. That is, nonprofit organizations should communicate their functional attributes in mass-marketing approaches as well as in nonprofit organization magazines.

Limitations and avenues for future research
In this study, we focused on charitable organizations and disaster relief and development aid organizations. Other studies could test the self-/functional congruence model based on other types of nonprofit organizations to increase results' generalization. Moreover, we examined donor loyalty as a dependent intention variable. Because there are also other outcome variables (e.g., volunteering, trust, brand preference, donor satisfaction), other studies should examine the effects on different dependent variables and examine the differences that may result. Moreover, although the intention variables reflect real behavior well, there is no perfect correlation between behavioral intentions and actual behaviors. Further, our analysis is based on cross-sectional data and we theoretically argued causal relations between constructs which were based on individuals' perceptions. Thus, to strengthen the internal validity and confirm the causal links, other studies could conduct experiment and examine real-behavior contexts. Further, considering that, volunteering is associated with higher time efforts, volunteering could be related to higher involvement. Therefore, actual self-congruence could have a stronger effect on volunteering than on donor loyalty.
In this study, we used a global construct to measure actual and ideal self-congruence. Other studies could examine whether perceived congruity with different personality characteristics (e.g., responsibility vs. emotionality) leads to different effects since the predictive role of actual and ideal self-congruence changes when brand type (i.e., functional vs. symbolic) changes (Zhu et al. 2019).
This study was conducted in a rather individualistic culture where the construct of "personality" and the related need for uniqueness are stronger than in collectivistic cultures. Therefore, other studies could examine the effects of actual self-congruence, ideal self-congruence and functional congruence on giving behavior in collectivistic cultures.
Linking actual self-congruence, ideal self-congruence,… material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/.