1 Introduction

Charitable giving is defined as the donation of money to an organization and is part of a broader family of helping behaviors (Bekkers & Wiepking, 2011). The total value of charitable contributions in the U.S.A. in 2020 was estimated at $474.44 billion (Giving, 2021). This estimate represents a 5.1% increase over 2019, setting a new all-time record in donations and continuing a trend of year-to-year gains. Given the value of charitable giving to society, it is essential that charitable organizations understand factors that may increase or decrease donor intentions and behaviors (da Silva et al., 2020; James, 2018). Recent research supports the importance of examining donor-related factors in explaining donation intentions (Mainardes et al., 2016; Maleki & Hosseini, 2020). In the current research, we focus on donors’ Implicit Theories (hereafter, “mindsets”) as a potential donor-related factor that has largely been unexplored in the area of charitable giving.

Mindsets broadly refer to beliefs people hold about the constancy of different human characteristics, such as intelligence, personality, leadership, selling ability, and health (Dweck, 1999). Whereas some individuals tend to believe such characteristics are relatively fixed and trait-like (“fixed mindset”) and therefore cannot easily change, others believe that those characteristics are state-like and malleable (“growth mindset”) and can substantially change. Often people’s mindsets are measured, which reflect their default, naturally-occurring (“chronic”) beliefs (Hsieh & Yucel-Ayba, 2018), though it is also possible to temporarily induce mindsets via experimental manipulation (Dweck, 1999). However, some researchers note potential complications of inducing mindsets that may be incongruent with their chronic mindset (Mathur et al., 2016), so measuring chronic mindsets may provide clearer and more realistic examination of their impact.

In addition, research shows that a person’s mindset about the changeability of characteristics is domain-specific: holding a fixed mindset (whether chronic or induced) about one characteristic (e.g., intelligence) does not necessarily mean that the person holds a fixed mindset about other characteristics (e.g., personality or selling ability) (Dweck, 1999; Novell et al., 2016). Consequently, domain-specific measures of mindsets have been developed, and it is critical to use the proper mindset measure to the corresponding domain of interest (Burnette et al., 2017).

Researchers have found that the mindsets people hold have many consequences and can prompt a range of motivational, affective, and behavioral outcomes (Dweck & Leggett, 1988; Plaks et al., 2005), and generally find that a growth mindset is associated with adaptive outcomes whereas a fixed mindset is associated with maladaptive outcomes. Although only a few studies have specifically investigated the impact of mindsets on charitable giving (Hsieh & Yucel-Aybat, 2018; Khalil et al., 2020; Septianto, 2020), we believe mindsets may be of high relevance to this area because the act of donating represents an attempt to make change for people. Specifically, we suspect that donors’ mindsets may influence their perceptions about whether others’ statuses/outcomes can change and may affect their willingness to donate. The current research examines the impact of potential donors’ chronic mindsets about personality on their willingness to donate to a homelessness charity.

As we discuss below, we build on the findings of recent studies by investigating the impact of mindsets on charitable giving to a social cause that has ambiguous controllability and examining a combination of cognitive and dual-affective mediators. After presenting relevant literature on mindsets and charitable giving as well as the proposed mediating variables, we review the methodology and results for two studies. Finally, we discuss our contributions, practical implications, and limitations and future research.

2 Mindsets and Charitable Giving

Charitable giving is a potentially important societal application of mindset research. Although little research has been conducted in this area, recent studies by Hsieh & Yucel-Aybat (2018), Khalil et al., (2020), and Septianto (2020) suggest a growing interest in the relationship between mindsets and charitable giving. For example, Hsieh & Yucel-Aybat (2018) investigated the effect of mindsets on charitable giving to different health causes by experimentally manipulating both health mindsets (fixed vs. growth) and health cause controllability (low – cancer vs. high – obesity), and found that mindsets moderated the impact of controllability on charitable giving. Khalil et al., (2020) found that among consumers who felt powerless, those with a growth (vs. fixed) mindset showed an increased willingness to donate, and that negative affect was an underlying mechanism in this relationship. Septianto (2020) found that whereas consumers with a growth mindset valued charities with a malleable (bad-to-good) reputation that showed effort, those with a fixed mindset valued charities that had an unchanging (always-good) reputation.

While it seems clear from these studies that mindsets have some impact on charitable giving, the extant research is scant and highly contextualized, which leaves many questions about the foundation for this relationship and underlying processes. The current research seeks to expand on previous research regarding mindsets and charitable giving in several ways. First, by examining our hypotheses in a homelessness context, our research builds on past work by investigating the effect of mindsets on charitable giving to a human condition that is social in nature. In prior research, the cause behind the charity has focused on either a medical (Hsieh & Yucel-Aybat, 2018; Septianto, 2020) or animal/pet (non-human) cause (Khalil et al., 2020).

Second, to our knowledge, this is the first study to examine a foundational relationship between chronic mindsets and charitable giving. Previous studies differed in the method used to examine mindsets, with Hsieh & Yucel-Aybat (2018) and Septianto (2020) manipulating mindsets and only Khalil et al., (2020) measuring donors’ chronic mindsets. Further, the mindset domain examined in each study varied, ranging from health mindset (Hsieh & Yucel-Aybat, 2018) to personality mindset (Septianto, 2020) to emotion mindset (Khalil et al., 2020). We believe chronic mindsets will be particularly appropriate for investigating the relationship between mindsets and charitable giving to human social conditions.

Third, previous research has varied in defining what cause controllability entails, often conflating blame and attribution measures (Brickman et al., 1982; Burnette et al., 2017; Hsieh & Yucel-Aybat, 2018). Here, we define perceived controllability as the ability to control/change homelessness once it occurs, rather than whether or not homelessness can be prevented, or describing the source of cause. In addition, previous research that examined perceived cause controllability experimentally manipulated this construct (Hsieh & Yucel-Aybat, 2018). By examining a cause whose perceived controllability is left ambiguous to donors and is measured instead of manipulated, we are able to examine the possibility that chronic mindsets naturally influence the perceived controllability of homelessness.

Fourth, by testing positive and negative affective mediators, our research expands on previous work that investigated the role of affect on charitable giving (Hsieh & Yucel-Aybat, 2018; Khalil et al., 2020). Previous studies have primarily focused on negative feelings toward those in need, such as blame (Hsieh & Yucel-Aybat, 2018). Another study used a single-item measure, the scale of which ranged from positive to negative emotion (Khalil et al., 2020). Notably, other research finds that positive and negative affect may have differential influence on behavior and decision-making (Herr et al., 2012; Kuhbandner et al., 2010). Further, because it is possible to experience ambivalence – both positive and negative emotions simultaneously (Giner-Sorolla, 2001; Mikulincer et al., 1998) – it is important to measure both valences separately to assess their role in charitable giving. Thus, we include measures for positive affect (sympathy) and negative affect (blame) to expand the literature in this area. To our knowledge, this is the first study on mindsets and charitable giving that separately examines dual-valence mediators.

In Study 1, we seek to establish a foundational model for the relationship between mindsets and charitable giving to a homelessness charity and identify a primary cognitive process (i.e., perceived donation efficacy). Study 2 seeks to replicate and build upon this foundational model by proposing a process model that includes cognitive and affective mediators.

3 Hypothesis Generation

Given the previous research on mindsets and charitable giving, it is evident that mindsets are influential. For example, Hsieh & Yucel-Aybat (2018) found that mindsets moderated the impact of controllability on charitable giving. Specifically, those induced to have a fixed (growth) mindset were (in)sensitive to whether a charity was for a low vs. high controllable medical condition, donating more (equally) to medical conditions for which the victim could not be blamed. Similarly, Khalil et al., (2020) examined the moderating effect of consumers’ emotion mindset on pet donation. Specifically, when consumers were placed in a low power situation, having a growth emotion mindset was associated with greater intention to donate to a pet charity.

However, we do not know of any studies that point to a foundational, direct relationship between mindsets and donation intentions. We thus look to other domains of mindset research that may provide insight. In the education domain, research on response to academic failure consistently finds that the two mindsets strongly differ in the perceive utility of effort (Dweck, 1999). Specifically, when in position of deficiency, those with a fixed mindset about intelligence tend to quit and/or withdraw effort because they believe changing this deficit (i.e., their academic performance) is not possible (Hong et al., 1999). In contrast, those with a growth intelligence mindset tend to respond to deficiencies and failure with resilience and increased efforts to improve because they believe change is possible (Murphy & Dweck, 2016). Another study in the sales domain found a main effect of mindset on sales outcomes, where those with a fixed selling mindset were more likely to avoid feedback about their sales performance (Novell et al., 2016). Across research in these other domains, there is a favorable pattern between holding a growth mindset and perseverance or continuation behaviors and an unfavorable pattern of holding a fixed mindset with avoidance or stopping behaviors. We believe that these patterns will extend to the donation domain regarding the investment of effort to change an unfavorable condition like homelessness. In other words, compared with holding a growth mindset, holding a fixed mindset should be associated with a lower intention to donate to the charity. We, therefore, hypothesize that:

H1

A more fixed mindset will be associated with a lower intention to donate to a homelessness charity.

3.1 Perceived Donation Efficacy

To build our foundational model, we wanted to identify a primary cognitive process mechanism. A search of the donation literature suggested that perceived donation efficacy would be a likely potential mediator. Perceived donation efficacy refers to donors’ perceptions that their contributions will make a difference to the cause they are supporting (Bekkers & Wiepking, 2011). A consistent, positive relationship between perceived donation efficacy and charitable giving has been documented in Bekkers and Wiepking’s meta-review (2011), as well as subsequent studies (Cao & Jia, 2017, Carroll & Kacharsky, 2019). In addition, experimental research has found that giving donors information about contribution effectiveness positively affects philanthropy (Parsons, 2007). Given the widespread finding in the literature that perceived donation efficacy is related to charitable giving, we believe this result will be corroborated in our research. Furthermore, since individuals with a fixed mindset generally do not believe change is possible, we propose that they will be less likely than those with a growth mindset to perceive that donations to a homelessness charity will be effective. This was evident in Hsieh and Yucel-Aybat’s study (2018), in which the relationship between mindsets and donation activity was mediated by perceived donation efficacy. Thus, we hypothesize that:

H2a

A more fixed mindset will be associated with lower perceived donation efficacy.

H2b

Perceived donation efficacy will be positively associated with donation intentions.

H2c

Perceived donation efficacy will mediate the relationship between mindsets and intention to donate to a homelessness charity.

4 Pretest

Following the protocol in Hsieh & Yucel-Aybat (2018), a pretest using an undergraduate sample (N = 99) aged 18 or older was conducted to measure perceived controllability of homelessness, cancer, and obesity (the latter two being the ‘less’ and ‘more’ controllable health conditions in Hsieh and Yucel-Aybat [2018], respectively). Participants completed a 3-item adapted assessment for perceived controllability (Hsieh & Yucel-Aybat, 2018) using a 7-point Likert scale (1 = highly disagree to 7 = highly agree). An example item was “I believe that cancer (obesity, homelessness) can be controlled.“ As expected, respondents’ perceptions of the controllability of homelessness (M = 4.63, SD = 1.44) fell in between cancer (M = 4.12, SD = 1.31) and obesity (M = 5.89, SD = .97), significantly varying from each (t[98] = 2.69, p < .01) (t[98] = -9.00, p < .001), respectively. As homelessness’ perceived controllability fell between the ‘high’ and ‘low’ controllability social conditions used by Hsieh & Yucel-Aybat (2018), we considered it to have ambiguous perceived controllability and a suitable social condition to test our hypotheses on the impact of mindsets.

5 Study 1

Study 1 tested the hypothesized foundational relationships between mindsets, perceived donation efficacy, and charitable giving outcomes (H1-H2).

5.1 Method

Three hundred and forty people were recruited via Amazon’s Mechanical Turk (MTurk) to participate in the study. All participants had English as their native language. Additional MTurk worker requirements included residing in the U.S.A and being over 24 years old, as participants under this age would be less likely to have steady income to donate. Because of increasing concerns about fraud and/or the quality of data from MTurk workers (Chmielewski & Kucker, 2020), we implemented screening techniques to ensure the quality of the data collected. These included a review for duplicate I.P. addresses (Kennedy et al., 2020) and both fixed alternative and open-ended attention checks to identify “unusual comments” (Chmielewski & Kucker, 2020), bots, or inattentive respondents. Forty-seven of the 340 participants failed one or more of these screeners and were removed from the analysis. Of the 293 participants whose data was retained, 63% were male, and 37% were female. Ages ranged from 24 to 74 years (M = 38.07, SD = 12.18). Participants completed an online survey that contained the measures below. Participants received monetary compensation for their time.

Mindsets were measured using the 8-item ‘Kind of Person Implicit Theories’ scale developed by Dweck (1999). This validated scale was the appropriate mindset measure because we were testing mindset in the domain of personality. An example item includes “The kind of person someone is, is something very basic about them and it can’t be changed very much.“ All mindset items used a 6-point Likert format (1 = highly disagree to 6 = highly agree). Four items were reverse-coded so that higher numbers indicated a more fixed mindset for all items (M = 3.36, SD = 1.61, α = 0.95). Perceived donation efficacy was measured using a 4-item scale adapted from Cao & Jia (2017) with a 7-point Likert format (1 = strongly disagree to 7 = strongly agree) that assessed the ability of donation to make a difference. An example item is, “Making a donation to a homelessness charity is an effective way to help homeless people.“ Higher numbers indicate greater perceived donation efficacy (M = 5.01, SD = 1.52, α = 0.93). Charitable giving was measured with two items adapted from Hsieh & Yucel-Aybat (2018) that assessed the likelihood to donate to a homelessness charity using a 7-point Likert format (1 = extremely unlikely to 7 = extremely likely). An example item is “How likely are you to donate money to a homelessness charity?“ Higher numbers indicate greater likelihood to donate (M = 4.97, SD = 1.72, α = 0.86). Additional demographic items, including gender and ability to donate, were included as control variables. We report the descriptive statistics and the correlation analysis in Table 1.

Table 1 Study 1: Mean, standard deviation, and correlation
Full size table

5.2 Results

Before running our hypothesized model, we ran a simple model looking at the direct effect of mindsets on charitable giving with no mediation pathway. The simple model explained 24% of the variance in charitable giving. We then tested our hypothesized structural model using Amos maximum likelihood method. Indirect effects were also tested using the bootstrap procedure with 2,000 bootstrap samples (Fritz & Mackinnon, 2007). To assess the model fit, we looked at multiple indices. The relations in the path model explained 49% of the variance in charitable giving, indicating a relatively good improvement from the simple model. The model’s CFI was close to 0.9 (CFI = 0.92), indicating a relatively good fit (Bentler, 1990). SRMR was close to the 0.08 threshold (SRMR = 0.09), indicating adequate fit (Hu & Bentler, 1999). The model’s RMSEA was higher than 0.1 (RMSEA = 0.12), which may indicate poor fit. However, Kenny et al., (2015) note that the likelihood of rejecting a correctly specified model based on RMSEA for models with very small degrees of freedom (here, df = 5) and smaller sample sizes (here, N = 293) could be unacceptably high, and the interpretation of the RMSEA value in isolation could be misleading. Based on the group of indices as a whole, then, we can conclude that the model had acceptable fit. Table 2 shows the results of the path model. As shown in Fig. 1, the direct effect of mindsets on charitable giving for the simple model with no mediation pathways was negative and significant (β = -0.17, s.e. = 0.06, p-value = 0.002), suggesting that individuals with a fixed mindset are less likely to donate to a homeless charity (supporting H1). When we add our hypothesized mediator to the model, the significance of the relationship between mindsets and charitable giving disappears (β = -0.08, s.e. = 0.04, p-value = 0.07), suggesting a significant mediating effect.

Table 2 Study 1: Bootstrap analysis of the direct and indirect effects
Full size table
Fig. 1
figure 1

Study 1: Proposed Model and Regression estimates

Full size image

The relationship between mindsets and perceived donation efficacy is negative and significant (β = -0.19, s.e. = 0.05, p-value < 0.001), suggesting that individuals with fixed mindsets are more likely to believe that their donation money is less effective in inducing change (supporting H2a). The relationship between perceived donation efficacy and charitable giving is positive and significant (β = 0.65, s.e.= 0.05, p-value < 0.001), suggesting that the more individuals believe that their money would be effective in inducing change, the more likely they are to donate (supporting H2b). Finally, the mediated pathway from mindsets on charitable giving through perceived donation efficacy is significant (β = -0.12, LLCI = -0.21, ULCI = -0.04), confirming that perceived donation efficacy mediates the relationship between mindsets and charitable giving (supporting H2c).

5.3 Discussion

In support of H1, and consistent with previous research on charitable giving to other causes (Khalil et al., 2020), individuals with a more fixed (vs. growth) mindset indicated they were less likely to donate to homelessness charities. Further, Study 1 supported a mediated pathway between mindsets and charitable giving through perceived donation efficacy, supporting H2a-c and consistent with previous work (Bekkers & Wiepking, 2011; Carroll & Kacharsky, 2019; Hsieh & Yucel-Aybat 2018). In other words, those with a more fixed mindset tend to believe their donation is less likely to be efficient in driving change, so they intend to donate less to homelessness charities.

6 Study 2

Study 2 sought to replicate and build upon the relationships supported in Study 1 to establish a consistent effect across samples and gain a richer understanding of the foundational relationship. To do so, we drew upon research showing that both cognitive and affective variables influenced consumer preferences (Muncy, 1986; Zajonc & Markus, 1982). As such, in addition to replicating the results from Study 1 (H1-H2), Study 2 considers key constructs from the donation literature, including cognitive variables (perceived cause controllability and attribution) and positive and negative affective variables (sympathy and blame, respectively). Below we discuss relevant literature on these variables, their associated hypotheses, and propose the full hypothesized model.

6.1 Perceived Controllability

Perceived controllability may refer to either belief about control of a condition from a preventative (onset) or curative (offset) standpoint (Brickman et al., 1982). The former refers to acquiring a condition, while the latter refers to managing a condition once it occurs (Burnette et al., 2017). The current research defines perceived controllability from a curative, offset standpoint (Hsieh & Yucel-Aybat, 2018). Much research on perceived controllability is vaguely defined, so there is limited research that directly examines the impact of mindsets on curative, offset controllability. However, supporting the idea that those with a growth (vs. fixed) mindset believe an unfavorable outcome can be changed, a study by Song et al., (2020) found that a growth (vs. fixed) mindset in students was associated with higher perceived controllability of ability when they did poorly on an exam. Burnette et al., (2017) also found a positive relationship between a growth weight mindset and perceived controllability for obesity. In addition, we believe these variables will serially mediate the proposed relationship between mindsets and donation intentions, as Lee et al., (2020) found a relationship between perceived controllability and perceived donation efficacy. We thus hypothesize the following:

H3a

A more fixed mindset will be associated with lower perceived controllability of homelessness.

H3b

Perceived controllability of homelessness will be positively associated with donation efficacy.

H3c

Perceived controllability of homelessness and donation efficacy will serially mediate the relationship between mindsets and donation intentions.

6.2 Attribution

Causal attributions refer to whether people perceive events or behaviors to derive from internal causes such as personality vs. external causes and other situational characteristics (Heider, 2013). There is robust literature linking mindsets with attributions. Importantly, research suggests that having a fixed (vs. growth) mindset may lead people to make more dispositional (vs. situational) attributions for an outcome (Hong et al., 1999). In other words, believing people/ability cannot change may lead people to look internally at what they perceive as a constant, while having a growth mindset may lead people to focus on variables in the external environment that could contribute to an outcome. Indeed, people with fixed (vs. growth) mindsets tend to view negative social behaviors and academic outcomes as reflecting an enduring disposition or ability (Dweck et al., 1995). Therefore, we hypothesize this will occur for attributions of homelessness:

H4

A more fixed mindset will be associated with a stronger internal attribution (personality) for homelessness.

Research has also linked attribution to subsequent emotions and giving behaviors toward those in need. Hsieh & Yucel-Aybat (2018) suggested that those with a fixed mindset donated less because they made personality (internal) attributions for their social condition. Lee et al., (2020) manipulated perceived attribution and found that external attributions led politically liberal participants to respond more favorably to charity appeals. We thus expect that internal attributions will have a negative impact on donations as they are closely linked with emotions felt toward victims. Weiner’s attribution-emotion-action theory also suggests that attributions may prompt both emotions toward those in need and behavioral responses (Weiner, 1995). Early researchers Kleck (1969) and Katz (1981) indicated that people might simultaneously feel positive and negative emotions towards stigmatized individuals. We thus investigate the impact of two other-oriented emotions linking attributions and charitable giving - one positive (sympathy) and one negative (blame).

6.3 Blame

We believe a negative feeling of blame will be associated with a lower desire for charitable giving. According to the culpable control model (Alicke, 2000), an attribution of personal responsibility is the main factor in ascribing feelings of blame. In support of this, DePalma et al., (1999) and Lee et al., (2014) found that when others perceive victims to be responsible for their own social condition, they are less inclined to help the victims. Further, Zagefka et al., (2011) found greater donations to natural disasters than human disasters because people blamed victims of natural disasters to a lesser extent. Others have also found a negative relationship between blame and prosocial behaviors (Chang, 2011), including charitable giving (Hsieh & Yucel-Aybat, 2018).

Because fixed mindset individuals should be more likely to make an internal (e.g., personality) attribution for a homeless person’s social condition and perceive homeless people as having an unchanging personality (Dweck et al., 1995), they should be more likely to blame the homeless for their condition. Indeed, Hsieh & Yucel-Aybat (2018) found a relationship between a fixed mindset and blame for controllable health conditions. We thus hypothesize that:

H5a

A stronger internal attribution (personality) for homelessness will be positively associated with feelings of blame.

H5b

Stronger feelings of blame will be negatively associated with donation intentions.

H5c

A stronger internal attribution (personality) for homelessness and blame will serially mediate the relationship between mindsets and donation intentions.

6.4 Sympathy

Sympathy is a positively-oriented emotion defined as feelings of “concern for another’s welfare” (Eisenberg et al., 2002) when others are suffering (Wispe, 1986). Sympathy has been viewed as one of the essential components of prosocial behavior (Bagozzi & Moore, 1994) and is positively associated with charitable giving (Conlin & Bauer, 2021; Pham & Septianto, 2020).

Research suggests that having a growth personality mindset may lead people to make more situational, external attributions for an outcome (Hong et al., 1999). Thus, they may also be more likely to view an ambiguous social condition, such as homelessness, as emanating from circumstances beyond the affected person’s control. This tendency would likely generate sympathy since homeless individuals would not be seen as causing their misfortune (Schumann et al., 2014). In contrast, fixed mindset individuals tend to view people as having a core moral character which is endemic to their nature, and see them as deserving of punishment (Dweck et al., 1995). Thus, we believe that fixed mindset people will view homeless individuals more harshly and be less sympathetic to them. These expectations lead to the following hypotheses:

H6a

An internal attribution will mediate the relationship between mindsets and sympathy for homeless people.

H6b

Increased sympathy for homeless people will be positively associated with intentions to donate to a homelessness charity.

H6c

A stronger internal attribution (personality) for homelessness and sympathy will serially mediate the relationship between mindsets and donation intentions.

Figure 2 below summarizes our theoretical framework, showcasing the parallel serial mediation where the relationship between mindsets and charitable giving is mediated by perceived controllability and perceived donation efficacy on one pathway, and attribution and both positive (sympathy) and negative (blame) affect on the other pathways.

Fig. 2
figure 2

Homelessness Charitable Giving Theoretical Framework

Full size image

6.5 Method

A similar procedure to the one used in Study 1 was used for Study 2. However, we attempted to solve the smaller sample issue found in Study 1 by collecting data on a larger sample of subjects. Four hundred ninety-eight participants passed all screening measures and were retained for analysis. Of those participants, 55% were male, and 45% were female. Ages ranged from 24 to 77 years (M = 39.05, SD = 11.97). Participants completed the same mindset (M = 3.55, SD = 1.48, α = 0.93), perceived donation efficacy (M = 5.39, SD = 1.17, α = 0.90), and charitable donation items (M = 5.30, SD = 1.56, α = 0.88) as in Study 1. They also completed the measures regarding perceived controllability, attribution, sympathy, and blame below.

Attribution was measured using four items that assessed the extent to which potential donors believe homelessness is due to internal personality. Participants indicated levels of agreement with statements such as “As much as I hate to admit it, one’s personality may contribute to being homeless” using a 7-point Likert format (1 = strongly disagree to 7 = strongly agree). Higher numbers indicate greater internal attribution for homelessness (M = 3.74, SD = 1.47, α = 0.84). Perceived controllability of homelessness was measured using an item adapted from Hsieh & Yucel-Aybat (2018) that assessed the extent to which potential donors believed that homeless status could change. Participants indicated levels of agreement with the statement “I believe homelessness can be controlled once it occurs” using a 7-point Likert format (1 = strongly disagree to 7 = strongly agree). Higher numbers indicate greater perceived social condition controllability (M = 5.30, SD = 1.37). Sympathy was measured using two items adapted from Lee (2009) that assessed sympathy for homeless people. Participants indicated levels of agreement with statements such as “I feel sorry for homeless people” using a 7-point Likert format (1 = strongly disagree to 7 = strongly agree). Higher numbers indicate greater sympathy (M = 5.82, SD = 1.15, α = 0.82). Blame was measured with an item adapted from Hsieh & Yucel-Aybat (2018) that assessed the extent to which donors assign blame to the victim for their social condition. Participants indicated levels of agreement with the statement “Someone who is homeless is careless” using a 7-point Likert format (1 = strongly disagree to 7 = strongly agree). Higher numbers indicate greater blame (M = 3.62, SD = 1.80). Additional demographic items, including gender, ability to donate, and the extent to which Covid-19 influenced their perceptions of homelessness, were included as control variables. We report the descriptive statistics and the correlation analysis in Table 3.

Table 3 Study 2: Mean, standard deviation, and correlation
Full size table

6.6 Results

As in Study 1, our hypothesized structural model was tested using Amos maximum likelihood method. Indirect effects were also tested using the bootstrap procedure with 2,000 bootstrap samples (Fritz & Mackinnon, 2007). The model had a good fit across all fit indices (CFI = 0.964, RMSEA = 0.065, SRMR = 0.051), and 51% of the variance in charitable giving is explained by the relations in the path model. Table 4 shows the results of the path model. As shown in Fig. 3, the direct effect of mindsets on charitable giving for the simple model with no mediation pathways was negative and significant (β = -0.13, s.e. = 0.04, p-value < 0.001), replicating the main effect in Study 1 and supporting H1. When we add our proposed mediator to the model, the significance of the relationship between mindsets and charitable giving disappears (β = 0.02, s.e. = 0.03, p-value = 0.61), suggesting a significant parallel serial mediation effect. Study 2 also replicated the simple mediating pathway between mindsets and charitable giving through perceived donation efficacy, supporting H2a, H2b, and H2c.

Table 4 Bootstrap effect of the direct and indirect effects
Full size table
Fig. 3
figure 3

Study 2: Regression Estimates

Full size image

The relationship between mindsets and perceived controllability was negative and significant (β= -0.14, s.e.=0.04, p = .005), suggesting that individuals with a fixed mindset believe that homelessness is not controllable (supporting H3a). The relationship between perceived controllability and perceived donation efficacy was positive and significant (β = 0.22, s.e.= 0.03, p < .001), suggesting that individuals who believe that homelessness is controllable tend to think that their donations will be more effective (supporting H3b). As previously stated, the relationship between perceived donation efficacy and charitable giving is positive and significant (H2b; β = 0.57, s.e.= 0.05, p < .001). Furthermore, the serial mediated pathway from mindsets to charitable giving through perceived controllability and perceived donation efficacy is significant (β= -0.02, LLCI= -0.04, ULCI= -0.01), supporting H3c. Thus, individuals with a fixed mindset believe that homelessness is less controllable and believe their donations will be less effective in inducing change, resulting in less donation.

The relationship between mindsets and attribution was positive and significant (β = 0.30, s.e. = 0.04, p < .001), suggesting that individuals with a more fixed mindset tend to believe that homelessness is due to internal causes (supporting H4). The relationship between attribution and blame was positive and significant (β = 0.61, s.e. = 0.42, p < .001), suggesting that individuals who believe homelessness is due to internal causes tend to blame individuals for their social condition (supporting H5a). The relationship between blame and charitable giving is negative and significant (β = -0.06, s.e. = 0.03, p = .05), suggesting that individuals blaming individuals for their social condition donate less (supporting H5b). Furthermore, the serial mediated pathway from mindsets to charitable giving through attribution and blame is significant (β= -0.01, LLCI= -0.02, ULCI= -0.01), supporting H5c. Thus, individuals with a fixed mindset tend to believe homelessness is due to the internal characteristics of the homeless, blaming homeless people for their social condition and resulting in less donation. The relationship between attribution and sympathy was negative and significant (β = -0.17, s.e. = 0.03, p < .001), suggesting that individuals who believe homelessness is due to internal causes tend to be less sympathetic with individuals facing the social condition (supporting H6a). The relationship between sympathy and charitable giving is positive and significant (β = 0.21, s.e. = 0.05, p < .001), suggesting that the greater the sympathy towards individuals facing homelessness, the higher the likelihood to donate (supporting H6b). Furthermore, the serial mediated pathway from mindsets to charitable giving through attribution and sympathy is significant (β= -0.01, LLCI= -0.02, ULCI= -0.04), supporting H6c. Thus, individuals with a fixed mindset tend to believe homelessness is due to the internal characteristics of the homeless and are less sympathetic towards homeless people for their social condition, resulting in lower donations.

6.7 Discussion

Study 2 replicated the findings of Study 1, showing that individuals with a more fixed (vs. growth) mindset were less likely to donate money to a homelessness charity (H1). Study 2 also built on Study 1 by showing how mindsets influence charitable giving via parallel serial pathways involving cognitive and affective mediators.

On the first pathway, the direct effect of mindsets on charitable giving is explained through cognitive variables of perceived controllability (H3a-c) and perceived donation efficacy (H2a-c). Our study shows that individuals with a fixed vs. growth mindset are less likely to believe that homelessness is controllable. Our finding is similar to previous work linking mindsets with perceived controllability (Song et al., 2020; Burnette et al., 2017). In other words, individuals with a fixed (vs. growth) mindset perceive homelessness as less controllable and are thus more likely to believe that their donations will be less effective in inducing any change, resulting in less donation.

On the second pathway, the direct effect of mindsets on charitable giving is explained through the cognitive variable of attribution (H4) followed by both positive (sympathy, H5a-c) and negative (blame, H6a-c) affect. In other words, individuals with a fixed vs. growth mindset are more likely to believe that people are homeless because of some internal characteristic that they possess. This internal attribution leads individuals with a fixed (vs. growth) mindset to blame homeless people more and to have less sympathy towards them, resulting in less donation.

7 General Discussion

This research sought to determine the impact of chronic mindsets on charitable giving to a social cause - the nature of which may be ambiguous to donors - and the mechanisms by which this impact occurs. We hypothesized that personality mindsets produce natural differences in perceptions of the cause and controllability of a social condition, homelessness. After a pretest confirmed that homelessness is perceived as ambiguous in controllability, data from two studies supported all hypotheses. We showed that a more fixed mindset is associated with lower charitable giving intentions (Studies 1 and 2) through multiple pathways involving cognitive (Studies 1 and 2) and affective mediators (Study 2). This paper reinforces the relevance of mindsets in the area of charitable giving and builds on previous research in marketing, psychology, and nonprofit dynamics to explain charitable giving.

7.1 Theoretical Implications

The current research makes several important contributions. First, it extends the literature on mindsets and charitable giving research to examine a social cause (homelessness), the perceived controllability of which may be ambiguous. While past related research has mainly focused on manipulating mindsets in messages (Hsieh & Yucel-Aybat, 2018), our findings suggest that donors’ chronic, measured mindsets play a role in donation intentions. By demonstrating a direct relationship between chronic mindsets and charitable giving, we provide a more foundational model that may be useful for understanding the relationship between mindsets and charitable giving.

Second, this research extends prior research by identifying new constructs and pathways through which a fixed mindset may lower donation intentions. The first pathway is driven by homelessness’s perceived controllability (curative) and its impact on donation efficacy. The second pathway is guided by causal attribution and its impact on positive and negative affect toward those in need. The significant pathway involving the positive feeling of sympathy is noteworthy, as prior relevant research has focused only on negative feelings such as blame (Hsieh & Yucel-Aybat, 2018). Further, the dual-valence examination in the current research brings a more comprehensive understanding of the role of emotions in charitable giving.

Third, we provide greater clarity on the role of “cause controllability” in mindsets and charitable giving research by distinguishing between cognitions of attribution (origins of the condition) and controllability (changeability of the condition). Interestingly, attribution and perceived controllability were not significantly related (r = .10, p = ns), suggesting that these cognitions about the cause are separate and drove distinct pathways. Further, the regression coefficients in the final model suggest that the stronger pathway was that involving perceived controllability and perceived donation efficacy (vs. attribution and emotions) – the more cognitively driven of the two pathways identified. In other words, potential donors may be thinking more with their heads than their hearts when donating to homelessness charities.

The current research may also provide insight into the overall impact of mindsets on judgments and helping behaviors. Despite a perspective that those with a fixed mindset often make internal attributions for outcomes and find unfavorable characteristics punishable (Dweck, 1995), some literature has found a positive link between a growth mindset and unfavorable impacts on constructs such as blame (Ryazanov & Christenfeld, 2018) and body shaming (Burnette et al., 2017; Hoyt et al., 2017). In light of the current research, these other findings may have occurred because people assumed internal attribution yielded victim choice and responsibility. Instead, a growth personality mindset in our study was perceived as less internally caused and as more manageable, both of which were favorable for donation intentions.

7.2 Practical Implications

Our findings point to the need for nonprofits to consider donors’ individual differences, including mindsets (Mainardes et al., 2016; Maleki & Hosseini, 2020). To increase the presence of a growth mindset among potential donors, nonprofits may implement efforts to reach donors who are chronically more growth-oriented and communicate a growth mindset in their messages. Nonprofits may also focus messaging around the two pathways. First, messages may highlight curative/offset controllability (i.e., people’s social condition status can change) and the efficacy of donations, specifying how affected individuals can be helped via donations and showing how these contributions can make a difference in their lives. Second, nonprofits can orient the tone of advertising and fundraising communications to create a feeling of sympathy among donors rather than reduce perceived blame since sympathy had a larger impact on donation intentions than blame.

7.3 Limitations and Future Research

As with any research, this research is not without limitations, at which future research efforts may be aimed. First, this paper examined the impact of chronic, measured mindsets of personality on donation intentions via cognitive and affective process variables. While the data are strong, correlation research cannot determine causal relationships among constructs. Future research may experimentally examine this relationship. Future experimental research may also wish to examine whether message framing constructs from one or both pathways (e.g., perceived donation efficacy, sympathy, attribution, etc.) could augment or mute the impact of donors’ chronic mindset. Second, the operationalization of donation in our study was donation intentions. Future research may examine whether these relationships hold for donation behavior. In addition, future research may examine whether the current model generalizes to other helping behaviors, such as volunteering, and to other social conditions beyond homelessness.

7.4 Conclusions

The current research provides insight into charitable giving behaviors by considering how donors’ beliefs about the controllability of personality can influence their donation intentions to a social cause whose controllability is ambiguous. Two studies found support that a fixed (vs. growth) mindset was associated with lower donation intentions. Parallel serial mediation analyses revealed important cognitive and emotional mechanisms that explain this relationship, with donation efficacy as the most significant proximal mediator, followed by the positive emotion of sympathy and the negative feeling of blame. These results extend previous literature and have actionable insights for nonprofit organizations that wish to maximize contributions to their causes.