SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Theory and hypotheses
  5. Data and methodology
  6. Analysis and results
  7. Discussion
  8. Acknowledgement
  9. References
  10. Biographies

This paper examines the mechanisms for giving by investigating the psychological and values differences between men and women's motivations for giving. We explored two of the eight mechanisms for giving developed by Bekkers and Wiepking as a framework for why people give—principle of care and empathic concern. Are there differences in these motives for giving by gender, and can these differences in values and the psychological benefits that people receive when making donations explain gender differences in charitable giving? Are women more likely to give and give more than men because of their higher levels of empathic concern and principle of care? We used two US national data sets to test our hypotheses. Our results for both data sets indicate significant differences in motives by gender, as well as differences in the probability of giving and amount given by gender, even after controlling for empathic concern and principle of care measures. Our findings are discussed in terms of the importance of viewing charitable giving through a gender lens as well as practical implications for practitioners. Copyright © 2011 John Wiley & Sons, Ltd.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Theory and hypotheses
  5. Data and methodology
  6. Analysis and results
  7. Discussion
  8. Acknowledgement
  9. References
  10. Biographies

Women's philanthropy has been shaped to a significant extent by their shifting economic position and social roles. Income and education, in particular, are strong predictors of giving, and women have made notable gains in both over the past three decades. More US women—about 60 per cent—are in the labor force today compared with 40 per cent four decades ago (Bureau of Labor Statistics, 2011). Women are entering the global labor force in record numbers as well. 1.2 billion of the world's 2.9 billion workers are women, representing a worldwide increase of about 200 million women employed in the past 10 years. In 2006, 47.9 per cent of working women worldwide were in wage and salaried employment compared with 42.9 per cent 10 years earlier. Furthermore, in six out of nine global regions, female employment-to-population ratios increased over the last 10 years (International Labour Organization, 2007).

Similarly, the proportion of working women with a college degree in the USA roughly tripled from 1970 to 2008: 36 per cent of women in the workforce held college degrees in 2008, compared with 11 per cent in 1970 (Bureau of Labor Statistics, 2010). Globally, more women are gaining access to education as well, but equality in education is still far from the reality in some regions of the world (International Labour Organization, 2008).

Women today are also earning more than ever before, although gender imbalances persist. In 1979, US women working full time earned 62 per cent of what men did; in 2008, women's earnings were 80 per cent of men's (Bureau of Labor Statistics, 2010). Internationally, women also typically earn less than men, although women's share of professional and managerial jobs is slowly increasing worldwide (International Labour Organization, 2004). The proportion of US wives earning more than their husbands also has grown. In two decades' time (1987 to 2007), the percentage of working wives who earn more than their working husbands grew by 8 per cent to a total of 26 per cent (Fry & Cohn, 2010).

Changes in marital status also play important roles. We know, for example, that married couples give more and are more likely to give than singles (Mesch et al., 2006) (Rooney et al., 2005) in large part because married people tend to be more connected with social networks, which is linked to philanthropic giving. Marriage in the USA, however, is becoming a less dominant lifestyle. As of 2009, the number of unmarried and single Americans comprised 43 per cent of the US population 18 and older—53 per cent of that group were women (U.S. Census Bureau, 2010).

A recent study conducted at the Women's Philanthropy Institute at the Center on Philanthropy at Indiana University found significant differences in philanthropic giving between single-headed male and female households (1) across income levels; (2) by marital status; and (3) across charitable subsectors—as to the likelihood of giving as well as the dollar amount given—controlling for other factors that affect giving (Mesch, 2010). In general, this report finds that female-headed households are more likely to give and give more to charity than male-headed households across all charitable subsectors and income levels. Other research supports these findings as well (Mesch et al., 2006) (Piper & Schnepf, 2008) (Simmons & Emanuele, 2007).

Why is this occurring? The purpose of this paper is to delve deeper into the mechanisms for giving by examining the psychological and values differences between men and women's motivations for giving. Are there gender differences in these motives, and can these differences in the benefits that people receive when making donations explain gender differences in charitable giving? We start by reviewing the literature and theory on donor motivations for giving—particularly focusing on gender differences—to inform our research questions. We then describe the data sets we use to test our hypotheses, present our findings, and conclude with practical applications for fundraising practice.

Theory and hypotheses

  1. Top of page
  2. Abstract
  3. Introduction
  4. Theory and hypotheses
  5. Data and methodology
  6. Analysis and results
  7. Discussion
  8. Acknowledgement
  9. References
  10. Biographies

The framework for this study uses (Bekkers & Wiepking, 2011) literature review of the eight mechanisms that drive charitable giving. In their review, they elucidate the empirical literature as to why people donate money to charitable organizations according to mechanisms, which include: (1) awareness of need; (2) solicitation; (3) costs and benefits; (4) altruism; (5) reputation; (6) psychological benefits; (7) values; and (8) efficacy. We focus on two of these mechanisms—the psychological benefits, which “refers to the intangible benefits that donors bestow on themselves as a result of donating” such as altruism and empathy (p. 15), as well as the values mechanism, which includes the attitudes and values of donors such as prosocial values or moral principle of care. We selected these two mechanisms because previous research (described next) has found significant dispositional differences, particularly between men and women. Empathic concern is the tendency to experience concerned, sympathetic, or compassionate reactive outcomes in response to the needs of others—and the principle of care is the endorsement of a moral principle that one should help others in need (Wilhelm & Bekkers, 2010). Empathic concern is a psychological mechanism that drives giving and is exemplified by motives labeled as “warm glow” or “empathic joy” (Andreoni, 1989) (Batson & Shaw, 1991). Principle of care may be viewed as a values mechanism “endorsed by donors [in making] charitable giving more or less attractive” ((Bekkers & Wiepking, 2011), p. 18). Individuals who espouse these values are more likely to give and are motivated to make the world a better place (Bekkers and Wipeking, 2011).

Using data from the General Social Survey, (Wilhelm & Bekkers, 2010) investigate the relative strength of these dispositional constructs as correlates of helping behavior and find a strong and consistent relationship between the principle of care and (to a lesser extent) empathic concern with many types of helping behavior. More specifically, their findings reveal a significant positive relationship between both empathic concern and principle of care with giving money to charity—although, upon further analysis, they find that principle of care mediates the empathy-helping relationship.

Gender differences in motives for giving

A substantial literature exists as to gender differences in altruism, prosocial, empathy, and other motives for helping behavior from the economic, sociology, and psychology disciplines. In general, this literature finds women to be more selfless, empathetic, and generous than men (Andreoni and Vesterland, 2001) (Cox & Deck, 2006) (Croson & Buchan, 1999) (Eagly & Crowley, 1986) (Eagly & Koenig, 2006) (Eckel & Grossman, 1998) (Eisenberg & Lennon, 1983) (Erdle et al., 1992) (Hoffman, 1977) (Jolliffe & Farrington, 2005) (Mills et al., 1989) (Piliavin & Unger, 1985) (Skoe et al., 2002).

The literature on charitable giving finds significant gender differences as well (Bekkers, 2007) (Bennett, 2003) (Croson et al., 2009) (Einolf, 2010) (Marx, 2000) (Mesch et al., 2006) (Piper & Schnepf, 2008) (Rooney et al., 2007) (Simmons & Emanuele, 2007) (Van Slyke & Brooks, 2005). Although much of this research “controls for the donor's sex, it does not look more closely at how there might be differences in giving and volunteering between the sexes” ((Simmons & Emanuele, 2007), p. 536). Possible explanations for gender differences in donor motivations for giving to charity are: (1) differences in gender roles and the ways in which women have been socialized as the caregivers of their families (Gilligan, 1982); (2) women view philanthropy as a way to show their caring and express their moral beliefs (Newman, 1995)—whereas, men give because of social roles such as status and social expectations (Eagly & Steffen, 1984) (Skoe et al., 2002); (3) women experience emotions more strongly than men (Harshman and Paivio, 1987); and (4) women are more egalitarian and engage in reciprocal behavior whereas men are more competitive (Croson & Gneezy, 2009).

Empirical research finds differences in motives as well. (Einolf, 2010) found women to score higher than men on measures of empathy, moral obligation, prosocial identity, and religiosity. Similarly, (Simmons & Emanuele, 2007) found that, on average, women donate more of both money and time and conclude that “altruism is a major contributing factor” (p. 547) where “society places more expectations on women to be altruistic and to act in an altruistic manner” (p. 546). (Kottasz, 2004) sample of young high-earning professional men and women found men to be more motivated by egoistic than altruistic motives when donating to charity, whereas, women were more interested in obtaining personal recognition for their donations. Other researchers have found differences between men and women in religious motives (Van Slyke & Brooks, 2005), personal values (Bennett, 2003), and social norms (Croson et al., 2009).

This paper expands on previous research by specifically examining two motives for giving based on (Bekkers & Wiepking, 2011) framework of the eight mechanisms that drive charitable giving and (Wilhelm & Bekkers, 2010) research on empathic concern and a moral principle to care about others. “Although there has been much research on volunteering and charitable giving in general, there has been surprisingly little research on how and why participation in these activities varies by gender” (Einolf, 2010, p. 2). Given the literature on the prominent role gender plays in various types of helping and prosocial behavior, we need to examine this more fully in research on charitable giving (Einolf, 2010). This study uses two national data sets from the USA to examine how gender differences in motivations affect giving.

Hypotheses

We predict that women will manifest stronger motives for giving, be more likely to give, and give more to charity than men. The following hypotheses are based on several different disciplines. The economics literature provides strong evidence indicating that women are more altruistic and generous, and more likely to reciprocate than men (e.g., (Buchan et al., 2008)) (Chaudhuri & Gangadharan, 2007) (Croson & Buchan, 1999) (Dufwenberg & Muren, 2006) (Eckel & Grossman, 1998) (Kamas et al., 2008) (Schwieren & Sutter, 2008) (Simmons & Emanuele, 2007). Similarly, the psychology literature reveals gender differences in giving, helping, empathy, and altruistic behavior (Eisenberg & Lennon, 1983) (Erdle et al., 1992) (Hoffman, 1977) (Piliavin & Charng, 1990). As such, we hypothesize the following:

H 1. Women will score significantly higher than men on measures of empathic concern and principle of care.

Previous literature also indicates that these motives affect giving to charity. In general, we see a strong and positive relationship between altruism, empathy, prosocial values, and philanthropy, and that these values are positively related to charitable giving (Bekkers & Wiepking, 2011) (Bennett, 2003) (Wilhelm & Bekkers, 2010). Research on charitable giving from the philanthropic literature suggests that women are more charitable and give more to charity than men (Einolf, 2010) (Mesch et al., 2006) (Mesch, 2010) (Piper & Schnepf, 2008) (Rooney et al., 2007)). As such, we predict that women will be more likely to give and give more to charity as a result of their stronger motives of empathic concern and principle of care. We hypothesize the following:

H 2. Women will be significantly more likely to give to charity than men.

H 3. Women will give a significantly higher amount to charity than men.

There may be additional motivational mechanisms between men and women that drive giving behavior. (Simmons & Emanuele, 2007) study addresses “not only how a person's sex influences their giving patterns, but what portion of these differences may be explained by observable differences between men and women, and which are the result of an unobservable variable, which might reflect the effect of a difference in altruism” (p. 537). They find that “women tend to give more than would be predicted for them by explanatory variables” (p. 546). That is, there may be “forces encouraging women to donate time and money that are not explained by a standard wage equation and may be reasonably attributed to differences in altruism” between men and women, “that imply that women bring an extra willingness to give” (p. 547). As such, we hypothesize the following:

H 4. Women will be significantly more likely to give and give more to charity than men, even when controlling for empathic concern and principle of care.

Data and methodology

  1. Top of page
  2. Abstract
  3. Introduction
  4. Theory and hypotheses
  5. Data and methodology
  6. Analysis and results
  7. Discussion
  8. Acknowledgement
  9. References
  10. Biographies

Sample

To test our hypotheses, we used two US national data sets. We selected these two data sets because they include the principle of care and empathic concern questions used by (Wilhelm & Bekkers, 2010), and we wanted to test our hypotheses using different years. The first data set is two cross-sectional waves (2002 and 2004) of the General Social Survey (GSS). Because the other data set, described next, studied people 25 years and older who are working; for the GSS, we limited the analysis on the giving decisions to the respondents 25 years and older and who reported paid employment, either full time or part time. With these limitations by age and employment status, the sample size for the two waves of the GSS is 1405, after excluding the cases who did not answer all of the questions needed for the giving to charity analyses.

The second data set, Knowledge Networks (KN), is from a study conducted by the Center on Philanthropy in the fall of 2008, regarding individual giving and volunteering in the prior 12 months. The survey response rate was 71 per cent. With weighting, it is a nationally representative sample of the working population 25 years and older. The survey was fielded by KN, drawing from its membership of over 50 000 individuals. KN is a survey consulting firm that has built a representative, random sample of US households (http://www.knowledgenetworks.com). Surveys are fielded online to qualified panel members who are recruited through online and other means in order to assure a random sample. Households without a computer are provided technology to use when responding to surveys. For the giving to charity analyses, we used 4554 cases of 6000 collected, but the sample size after applying the weights to the data is 4414. Those not analyzed did not answer all of the questions needed for this analysis.

Control variables

Demographic factors found to influence both the likelihood of giving as well as the amount given to charity are included in the analysis (see (Bekkers & Wiepking, 2007), for review of the literature on these variables). These include the following variables: (1) marital status; (2) race and ethnicity; (3) income; (4) education; (5) geographic region of residence; (6) living in a metro area; (7) religious affiliation; (8) frequency of religious attendance; (9) political party affiliation; (10) number of children in the household; and (11) age and a quadratic of age.1

Table 1 shows the descriptive statistics of the sample and variables used in the study for each data set. In general, both the GSS and KN samples are comparable. For GSS, 1165 households or 83 per cent of the sample gave to charity, whereas for KN (weights applied), 3354 households or 76 per cent of the sample gave to charity. The mean amount given to charity for KN for the entire sample (weights applied) is $762.98 (SD $3704.31). Exceptions to the comparability of the two datasets include: (1) a smaller percentage of those with a high school education or less and a greater percentage of those with some college; (2) a lower percentage of those reporting income less than $35 000: and (3) a greater percentage of those living in a metro area in the KN dataset.

Table 1. Descriptive statistics of respondents by gender
 General Social Survey, 2002 and 2004 Restricted to age >24 and workingKnowledge networks (weights applied)
# of RsAllMaleFemale# of RsAllMaleFemale
N =140514057276784414441423822032
Gender  52%48%  54%46%
Gave to charity116583%80%86%335476%71%82%
Amount given     $762.98 ($3704.31)$826.90 ($2395.68)$688.07 ($4803.47)
Married75354%58%49%245056%60%50%
# of child in household 0.64 (1.10)0.54 (1.01)0.74 (1.18) 0.46 (0.88)0.45 (0.86)0.48 (0.92)
White106276%77%74%313671%75%67%
Black17813%9%16%46411%8%14%
Hispanic1098%9%7%52812%10%14%
Other564%4%4%2866%7%6%
High school or less82759%60%58%159436%37%36%
Some college1269%9%9%124428%26%31%
BA or higher45232%32%32%157636%38%33%
Region = Northeast24818%19%16%84419%20%19%
Region = Midwest35325%23%27%106224%24%24%
Region = South51236%38%35%160336%36%37%
Region = West29221%20%21%90521%21%20%
MSA = Metro49435%33%37%372384%84%85%
Catholic33724%25%23%122828%29%26%
Other Christian78956%52%60%284765%63%66%
Jewish211%2%1%982%2%2%
Religion — Other25818%21%15%2415%6%5%
High religious attender32923%20%27%146233%30%36%
Moderate religious attender33424%20%27%57113%11%16%
Low religious attender51337%41%32%176240%42%38%
Never attend22916%19%14%61814%17%10%
Less than $35 00046233%28%39%72016%13%21%
$35 000–$75 00053938%40%37%218850%50%49%
More than $75 00040429%33%25%150634%38%30%
Republican41630%32%27%195344%48%40%
Democrat47834%29%39%230552%48%57%
Other political party51136%39%33%1564%4%3%

Analysis and results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Theory and hypotheses
  5. Data and methodology
  6. Analysis and results
  7. Discussion
  8. Acknowledgement
  9. References
  10. Biographies

The analysis consisted of several steps. The first was to create empathic concern and principle of care scales, using a factor analysis following (Wilhelm & Bekkers, 2010). We then assigned empathic concern and principle of care scale values to each respondent in each dataset, and we tested to see if women scored higher on the two motivation scales than men using the control variables in Table 1. We then examined the probability of giving and the amount given also using the control variables found in Table 1.

We conducted two sets of regression analysis on giving behavior. The first addressed probability of giving and used both GSS and KN datasets. In the second regression, the dependent variable for giving was the natural logarithm of the amount given to charity and was possible only for the KN data. The total giving for the KN data set is the sum of the responses provided by respondents answering a series of “area” giving questions, following the model of the Center on Philanthropy Panel Study, which is part of the Panel Study of Income Dynamics. From the separate “area” amounts that respondents provided (i.e., giving to religion, education, help meet people's basic needs, etc.), a total giving amount was calculated. We used linear probability models for the probability of giving analysis and the Tobit model for the amount given analysis.

Empathic concern and principle of care scales: factor analysis

The empathic concern and the principle of care scales are found in both the GSS and KN datasets. For GSS, the empathic concern questions used a five-point scale from “Does not describe very well” to “Describes very well,” whereas the principle of care questions used a five-point scale from “Strongly agree” to “Strongly disagree.” For KN, the empathic concern questions used a five-point scale from “Does not describe me at all” to “Describes me very well,” whereas the principle of care questions used a five-point scale of “Strongly disagree” to “Strongly agree.” The unweighted sample size used for the factor analysis to create the empathic concern scale was 2698 for GSS and 5848 (5838 weighted) for KN, whereas the unweighted sample size used to create the principle of care scale was 2702 for GSS and 5930 (5913 weighted) for KN.

The results of the GSS factor analysis for the seven empathic concern questions reveal one factor with an eigenvalue of 2.06. No other eigenvalues were above 1.0. The factor loadings range from 0.42 to 0.66, and the scale reliability coefficient is 0.74. The KN factor analysis results are similar. For the empathic concern questions, the factor analysis reveals one factor with an eigenvalue larger than 1.0, being 2.61. The factor loadings range from 0.47 to 0.74, and the scale reliability coefficient is 0.81. The questions used for the empathic concern scale are the following:

I often have tender, concerned feelings for people less fortunate than me.

Sometimes I don't feel very sorry for other people when they are having problems (reverse coded).

Other people's misfortunes do not usually disturb me a great deal (reverse coded)

When I see someone being taken advantage of, I feel kind of protective toward them.

When I see someone being treated unfairly, I sometimes don't feel very much pity for them (reverse coded).

I am often quite touched by things that I see happen.

I would describe myself as a pretty soft-hearted person.

Following (Wilhelm & Bekkers, 2010), we used three questions for the principle of care. In the factor analysis for GSS, there was one factor with an eigenvalue of 0.81. The factor loadings range from 0.38 to 0.58, and the scale reliability coefficient is 0.54. For KN, the principle of care factor analysis reveals one factor with an eigenvalue larger than 1.0, being 1.23, and the factor loadings range from 0.48 to 0.72. The scale reliability coefficient is 0.71. The principle of care questions are the following:

People should be willing to help others who are less fortunate (reverse coded in GSS).

Personally assisting people in trouble is very important to me (reverse coded in GSS).

These days people need to look after themselves and not overly worry about others (reverse coded in KN).

In both data sets, women scored higher than men on both scales, but the distribution covered the full range for both genders, with some women scoring near the lowest possible value and some men scoring at or near the highest possible. Table 2 shows the descriptive statistics for the empathic concern and principle of care scales by gender and data set. When the scales were used as dependent or independent variables in the subsequent analyses, they were standardized to have a mean of zero and a standard deviation of one.

Table 2. Summary statistics for empathic concern and principle of care scales
 GSS (scale 1–5)KN (scale 1–5) (Weights applied)
AllMenWomenAllMenWomen
  1. Standard deviations are in parentheses.

N =1405727678441423822032
Empathic concern3.98 (0.68)3.82 (0.67)4.16 (0.65)3.72 (0.68)3.56 (0.65)3.91 (0.67)
Principle of care3.84 (0.63)3.76 (0.61)3.92 (0.63)3.63 (0.70)3.52 (0.71)3.75 (0.66)

We then used ordinary least squares (OLS) to examine gender differences in the scale values (standardized), using the control variables described earlier. Our results for both data sets indicate significant differences in motivations by gender, as shown in Table 3. After controls, males scored significantly lower on both the empathic concern and principle of care scales than women (−0.456 standard deviations, p < 0.001 [empathic concern] and −0.272 standard deviations, p < 0.001 [principle of care] for KN, and −0.459 standard deviations, p < 0.001 [empathic concern] and −0.221 standard deviations, p < 0.001 [principle of care] for GSS). Thus, we find strong support for our first hypothesis—women score significantly higher than men on measures of empathic concern and principle of care.

Table 3. Comparison of empathic concern and principle of care scale values
 OLS results
Dependent variableEmpathic concern scalePrinciple of care scale
  • Note: The coefficient on male (β1) can be interpreted as: being a male is associated with a decrease in the empathic concern/principle of care scale values of β1 standard deviations, compared with females. The controls include all variables in the descriptive statistics in Table 1. The scale variables are standardized to have a mean of zero and a standard deviation of one.

  • *

    p < 0.05

  • **

    p < 0.01

  • ***

    p < 0.001

Data setKNGSSKNGSS
Independent variable = male−0.456***−0.459***−0.272***−0.221***
R-squared0.1410.0980.1260.082

Giving results

To test Hypotheses 2, 3, and 4, we use a linear probability model and Tobit model. We test these hypotheses using four different models—Model 1 includes gender only; Model 2 includes gender plus empathic concern; Model 3 includes gender plus empathic concern and principle of care; Model 4 is the full model including all control variables. With the GSS sample, we could examine only the probability of giving. With KN, we examined the probability of giving as well as the amount given.

Probability of giving

The dependent variable is giving to charity (Yes/No) for both the GSS and KN samples. Heteroskedastic robust standard errors were used in all of the models investigating the probability of giving because the linear probability model always results in heteroskedasticity. As seen in Table 4, in Model 1, gender is statistically significant in both datasets indicating that men are less likely to give to charity than women ( β = −0.056, p < 0.01 in GSS and β = −0.116 , p < 0.001 in KN). In Model 2, empathic concern and gender are significant in both datasets. Once again, men are less likely to give than women (β = −0.041, p < 0.05 in GSS and β = −0.056, p < 0.05 in KN). In Model 3, empathic concern, principle of care, and gender are significant in the KN data, but only principle of care and gender are significant for the GSS data. In both datasets, men are again less likely to give to charity than women (β = −0.040, p < 0.05 in GSS and β = −0.057, p < 0.05 in KN). In the full model, Model 4 with all controls, gender and principle of care are significant in both models; however, empathic concern remains significant only in the KN data set.

Table 4. Linear probability models for giving to charity
 GSSKN
 Model 1Model 2Model 3Model 4Model 1Model 2Model 3Model 4
  • Note: Standard errors in parentheses. The scale variables are standardized to have a mean of zero and a standard deviation of one. Heteroskedastic robust standard errors were used in all models.

  • *

    p < 0.05,

  • **

    p < 0.01,

  • ***

    p < 0.001

Male−0.056** (0.020)−0.041* (0.020)−0.040* (0.020)−0.057** (0.019)−0.116*** (0.028)−0.056* (0.028)−0.057* (0.027)−0.081** (0.026)
Empathic concern 0.031** (0.011)−0.006 (0.012)−0.006 (0.012) 0.117*** (0.013)0.055** (0.017)0.058*** (0.016)
Principle of care  0.074*** (0.012)0.049*** (0.011)  0.090*** (0.017)0.075*** (0.017)
Married   0.010 (0.022)   0.032 (0.030)
White   0.037 (0.050)   0.079 (0.064)
Black   0.012 (0.061)   −0.021 (0.076)
Hispanic   −0.038 (0.064)   −0.075 (0.075)
Income more than $75 000   0.194*** (0.029)   0.108* (0.044)
Income $35 000–$75 000   0.177*** (0.026)   0.053 (0.042)
Some college   0.097*** (0.028)   0.092** (0.033)
BA or higher   0.082*** (0.020)   0.122*** (0.035)
Northeast   −0.008 (0.028)   −0.032 (0.039)
Midwest   0.056* (0.025)   0.034 (0.031)
West   0.060* (0.026)   0.038 (0.035)
MSA = metro   −0.031 (0.020)   0.021 (0.036)
Catholic   0.048 (0.034)   0.032 (0.065)
Other Christian   0.020 (0.032)   −0.042 (0.065)
Jewish   0.112** (0.040)   0.067 (0.074)
High religious attender   0.116** (0.037)   0.073 (0.045)
Moderate religious attender   0.122*** (0.036)   0.059 (0.057)
Low religious attender   0.064 (0.033)   0.013 (0.045)
Republican   0.032 (0.023)   0.076 (0.090)
Democrat   0.034 (0.023)   0.041 (0.090)
Number of children   −0.023* (0.010)   −0.005 (0.013)
Age   0.008 (0.006)   0.011 (0.007)
Age squared   −0.000 (0.000)   −0.000 (0.000)
_cons0.858*** (0.013)0.851*** (0.014)0.850*** (0.014)0.285* (0.138)0.822*** (0.019)0.803*** (0.018)0.810*** (0.018)0.250 (0.212)
N14051405140514054554455445544554
N (Weighted)    4414441444144414
R20.0060.0120.0410.2020.0180.0890.1150.189

These results support Hypotheses 2 and 4. The results of the full model indicate that men are significantly less likely to give to charity than women (β = −0.057, p < 0.01 in the GSS and

β = −0.081, p < 0.01 in the KN). Hence, for the full model, being male is associated with an average decrease in the probability of giving of 5.7 per cent in GSS and 8.1 per cent in KN.

Amount given

To test Hypothesis 3 and 4, we use the Tobit model, with the natural log of giving as the dependent variable, because the natural logarithms of dollar amounts more closely follow the normal distribution curve. Using KN, we can look at amount given and its relationship to both the empathic concern and principle of care scales. Table 5 presents the marginal effects, conditional on giving, for each independent variable, holding the other independent variables in the specification fixed at their mean values. The marginal effects for the binary variables are for a discrete change from zero to one, whereas the marginal effects for the other variables are calculated at their mean values. The results of the full model indicate that: (1) both principle of care and empathic concern positively affect the amount that givers give

Table 5. Tobit marginal effects for amount of giving (KN data)
 Model 1Model 2Model 3Model 4
  • Note: Marginal effects; standard errors in parentheses. (d) for discrete change of dummy variable from 0 to 1. The scale variables are standardized to have a mean of zero and a standard deviation of one.

  • *

    p < 0.05,

  • **

    p < 0.01,

  • ***

    p < 0.001

Male (d)−0.187*** (0.042)−0.008 (0.042)−0.013 (0.042)−0.117** (0.040)
Empathic concern 0.358*** (0.021)0.079** (0.028)0.112*** (0.027)
Principle of care  0.410*** (0.027)0.290*** (0.026)
Married (d)   0.115** (0.043)
White (d)   0.275*** (0.077)
Black (d)   −0.113(0.098)
Hispanic (d)   0.023(0.099)
Income more than $75 000 (d)   0.793*** (0.071)
Income $35 000–$75 000 (d)   0.378*** (0.057)
Some college (d)   0.241*** (0.051)
BA or higher (d)   0.666*** (0.054)
Northeast (d)   −0.025 (0.056)
Midwest (d)   0.129* (0.052)
West (d)   0.110* (0.056)
MSA = metro (d)   0.134* (0.053)
Catholic (d)   0.117 (0.096)
Other Christian (d)   0.077 (0.090)
Jewish (d)   0.394* (0.173)
High religious attender (d)   0.855*** (0.074)
Moderate religious attender (d)   0.451*** (0.086)
Low religious attender (d)   0.137* (0.063)
Republican (d)   0.513*** (0.117)
Democrat (d)   0.299** (0.113)
Number of children (d)   0.091*** (0.023)
Age   0.047*** (0.011)
Age squared   −0.000** (0.000)
N4554455445544554
N (Weighted)4414441444144414
Pseudo R20.0010.0140.0240.073

(care: inline image = 0.290; p < 0.001; empathy: inline image= 0.112; p < 0.001); and (2) being male is associated with a lower amount given among givers (inline image= −0.117; p < 0.01), after controlling for the scores on the two scales as well as on all of the other control variables. The marginal effect for the gender variable is that, conditional on giving, men gave approximately 12 per cent less on average than women.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Theory and hypotheses
  5. Data and methodology
  6. Analysis and results
  7. Discussion
  8. Acknowledgement
  9. References
  10. Biographies

Gender matters in philanthropy, and our study underscores the importance of seeking to better understand philanthropy through a gender lens. Our findings show that men and women not only differ in their motivations for giving, but that these motives affect giving to charity. Across two different US data sets, collected at different times, we find fairly consistent results for the importance of empathic concern and the principle of care mechanisms that influence giving. In general, we found support for all of our hypotheses: (1) that women score significantly higher than men on empathic concern and principle of care; (2) that these motives are positively and significantly related to giving for both men and women; and (3) that women, more than men, are more likely to give and give more, even when controlling for these motives as well as other factors that affect giving.

We also find support for (Wilhelm & Bekkers, 2010) finding of the relationship between empathic concern, principle of care, and helping behavior. Also using the GSS, they found that principle of care mediates the empathy-helping relationship; adding principle of care to the specifications resulted in empathic concern being insignificant. In our study, principle of care was significant and positively related to probability of giving across both samples in the full model. However, for GSS, when principle of care was added to both Models 3 and 4, empathic concern lost significance. For the KN sample, however, we found positive and significant results for both motives in Model 3 as well as the full model.

It is not clear as to why we see different results across data sets. One possibility is related to statistical power, because the sample size for KN was more than three times larger than the GSS dataset. Another possibility is that principle of care, does, indeed, mediate the effect of empathic concern on giving, as found by Wilhelm and Bekkers. In our study, adding principle of care to the KN specifications resulted in the coefficient of empathic concern being reduced by more than half when principle of care alone was added. Similar results were found in the full model that includes all of the controls. This provides some evidence that principle of care is an important mediating factor, although more research is needed to determine which construct is more influential. Future research should replicate these results using different samples.

Our findings are consistent with previous research that investigates dispositional motives for charitable giving by gender, where women were found to score higher than men on prosocial traits and sense of moral obligation (Einolf, 2010) and use different social norms that determine giving (Croson et al., 2009). Our study, however, tested only two of these motives. Given the paucity of research in this area, scholars seeking to better understand charitable giving across gender might want to include not only analysis of principle of care and empathy measures in further studies, but explore additional motives for helping behavior found in the literature—specifically using the framework provided by (Bekkers & Wiepking, 2011).

(Simmons & Emanuele, 2007) conclude that gender differences in donation of money and time may be caused “by a difference in altruism between males and females” (p. 544), “indicating that women tend to give more than would be predicted for them by explanatory variables” (p. 546). We find support for these findings as well. Even after controlling for empathetic concern and principle of care, our results still find significant differences in the probability of giving and amount given to charity—suggesting that gender differences in giving to charity remains unexplained by demographic characteristics as well as empathic concern and principle of care motives that drive giving behavior—as found by Simmons and Emanuele. This certainly is a fruitful area of research that could be undertaken by many different disciplines.

Practical implications

Nonprofit organization fundraisers know well the importance of appealing to emotions. Our study supports the notion that women respond to emotional appeals more intensely and more frequently than men. What works for men may not work for women, too. Understanding women's motives for giving, and how they differ from men will allow fundraisers to better understand how to reach out to all of their donors—particularly the women's market—and institutions will be able to integrate working with women donors more fully into their comprehensive fundraising strategies.

The empirical studies provide substantive support for the concept that gender matters in philanthropy. Two barriers, which limit understanding of the value of viewing charitable giving through the gender lens have practical implications. Fundraisers who have previously elected not to work with female donors because of a perception that they take longer to cultivate as donors will benefit from considering the psychological benefits and values that connect female donors to giving. Once fundraisers understand the context in which philanthropy happens for men and women, and appreciate how men and women differ in their approaches to philanthropy, they will be able to better tailor their appeals and engagement with donors of both sexes. Then women will not report that they do not give because they are not asked.

Additionally, this study reinforces the need for nonprofits to focus on building new constituencies and the importance of becoming donor-centered. Research such as this gives the kind of insights needed to achieve this more relationship rather than transactional approach to the giving scenario. Our study finds that women are more likely to give and give more than men when we control for factors that typically influence giving. Practitioners would be well advised to make sure to include women in requests for donations as well as communications. We know what the potential is for changing the landscape for women's philanthropy, and we know the promise better understanding of this rich field offers.

Acknowledgement

  1. Top of page
  2. Abstract
  3. Introduction
  4. Theory and hypotheses
  5. Data and methodology
  6. Analysis and results
  7. Discussion
  8. Acknowledgement
  9. References
  10. Biographies

Authors acknowledge and thank the Center on Philanthropy at Indiana University for funding this research.

  • 1

    Marital status is defined as married or not (excluded category). Race/ethnicity is defined as white, non-Hispanic; black, non-Hispanic; Hispanic; and other race, non-Hispanic (excluded category). Income is defined as more than $75 000, $35 000–$75 000, and less than $35 000 (excluded category). Education is defined as high school or less (excluded category), some college, and BA degree or higher. Region is defined as Northeast, Midwest, West, and South (excluded category). Live in a metro or not (excluded category). For GSS, those living in an area with 50 000 residents or more are classified as living in a metro, following the US Census Bureau definition of an MSA. For KN, the MSA metro status was determined by the interviewers. Religious affiliation is defined as Catholic, Other Christian, Jewish, and Other religion including none (excluded category). Religious attendance is defined as High, Moderate, Low, and None (excluded category). For GSS, High is attend more than once a week or every week; Moderate is nearly every week, 2–3 times a month, and once a month; and Low is several times a year, once a year, and less than once a year. For KN, High is attend more than once a week or once a week; Moderate is once or twice a month; and Low is a few times a year and once a year or less. Political party affiliation is defined as Republican; Democrat; or Other including independents (excluded category).

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Theory and hypotheses
  5. Data and methodology
  6. Analysis and results
  7. Discussion
  8. Acknowledgement
  9. References
  10. Biographies

Biographies

  1. Top of page
  2. Abstract
  3. Introduction
  4. Theory and hypotheses
  5. Data and methodology
  6. Analysis and results
  7. Discussion
  8. Acknowledgement
  9. References
  10. Biographies
  • Debra Mesch is Professor of Public and Nonprofit Management in the School of Public and Environmental Affairs at IUPUI and Director of the Women's Philanthropy Institute at the Center on Philanthropy. She received her PhD and MBA at Kelley School of Business, Indiana University. Her research focuses on issues of women's philanthropy, volunteer motivation and management, executive compensation, and race and gender issues in giving and volunteering.

  • Melissa Brown advises nonprofit organizations in the USA and internationally, offering services for data analysis about charitable giving, nonprofit sector research, and project management. She worked for nearly 20 years at the Center on Philanthropy at Indiana University, most recently as managing editor of Giving USA, the yearbook of philanthropy published by Giving USA Foundation and Associate Director of Research. She is based in Indiana, USA.

  • Zachary Moore received his MA in Economics from IUPUI.

  • Amir Hayat received his MA in Economics from IUPUI and currently works as a research associate at the Center on Philanthropy at Indiana University.