“Walking the Walk” of Public Service Motivation: Public Employees and Charitable Gifts of Time, Blood, and Money

Abstract

Research generally indicates that public employees “talk the talk,” but do they also “walk the walk” of the public service motive (PSM)? Are public service employees more likely than others to engage in public service activities? The behavioral implications of PSM are addressed by studying the involvement in charitable activities of public, nonprofit, and private workers. Using data from the 2002 General Social Survey, multivariate logistic regression models are estimated to examine self-reported gifts of time, blood, and money to charitable organizations. It is found that government employees are more likely to volunteer for charity and to donate blood than for-profit employees are. Additionally, nonprofit workers are also more likely than their for-profit counterparts to volunteer. However, no difference is found among public service and private employees in terms of individual philanthropy. These findings generally lend support for the hypothesis that PSM is more prominent in public service than in private organizations, especially as it pertains to government personnel.

A decade ago Robert Behn (1995) identified learning how to motivate employees as one of the “big” questions of public management. Yet typical treatments of motivation in the academic literature are of limited utility for understanding much behavior in public organizations (DiIulio 1994; Perry 2000). Built on the assumption of a utility-maximizing bureaucrat, this rational choice (or principal–agent) approach to motivation may help explain why bureaucrats “shirk, subvert, and steal on the job,” but it cannot explain why they “strive… support… and sacrifice on the job” (DiIulio 1994, 281). In short, such an application of rational choice theory cannot explain prosocial behavior in public organizations (Perry 2000).

A theory of public service motivation (PSM) has been offered as an alternative. Consistent with conventional wisdom in public administration that government employment is a calling, PSM assumes bureaucrats are characterized by an ethic to serve the public. They act out of a commitment to the common good, rather than mere self-interest. Hence, they are motivated by different rewards than are those who do not answer the call. The existence of PSM has significant implications for job choice, job performance, and organizational effectiveness (Perry and Wise 1990; Rainey 1982; Romzek 1990); for the redress of declining social capital (Brewer 2003); and for the legitimacy of bureaucratic discretion.

In recent years a significant amount of research has examined the topic of public service motivation. The primary focus of this research has been on identifying the nature of PSM and asking if it is indeed characteristic of public employees. Generally, the findings of recent research lead to the conclusion that government workers do possess attitudes consistent with such an ethic and that this public service ethic may be fairly widespread among public employees (Brewer and Selden 1998; Crewson 1997).

While public employees may “talk the talk,” do they also “walk the walk” of public service? Less scholarly attention has been paid to the behavioral implications of PSM. Therefore, the research question addressed in this analysis is: Are public service employees more likely than others to engage in public service activities? To answer this question, self-reported contributions of time, blood, and money to charities are studied. Using data from the 2002 “General Social Survey,” logistic regression models are estimated to examine the relationship between public employment and self-reported charitable activities.

UNDERSTANDING PUBLIC SERVICE MOTIVATION

Employment in the public sector often has been portrayed as a calling, a sense of duty, rather than a job (Pattakos 2004; Perry 1996; Staats 1988). For instance, Frederickson (1997) refers to “the calling of the public service” as being at the heart of the “spirit of public administration.” Striking a similar chord, Gawthrop refers to public service being based on “the notion of duty as a love or an intense inner commitment to a cause that extends beyond the exigencies of the moment” (1998, 74).

Individuals who respond to this call are portrayed as being different from those who do not. They are “public servants” who are committed to the public good and characterized by an ethic built on benevolence, a life in service of others, and a desire to affect the community. From this viewpoint public bureaucrats are “principled agents” who do much of the heavy work of governing (Brehm and Gates 1997; DiIulio 1994). Pattakos illustrates the essence of the public service motive through the following comment made to him by a federal government employee: “I am attracted to the ideal of public service. It is fundamental to my own personal values and provides an opportunity for me to live these ideals… . I come from parents who believed that helping others was the best way to spend your life … service to humanity is the best work of life” (2004, 111). The commitment to serve is further articulated by a prison physician who was held hostage during a federal prison riot: “So you see, this is what you get from lousy government bureaucrats, most of whom make less than $30,000 a year—loyalty to each other, selflessness in the line of duty, and a dedication to protect the public they serve” (DiIulio 1994, 278).

Broadly speaking, PSM is “an individual's predisposition to respond to motives grounded primarily or uniquely in public institutions and organizations” (Perry 1996, 6). This motive has rational, normative, and affective bases (Perry and Wise 1990). In his effort to devise a general measure, Perry (1996) further identified four empirical components of the PSM construct: attraction to policy making, commitment to the public interest, compassion, and self-sacrifice.

Employing Q-sort methodology, Brewer, Selden, and Facer (2000) find that there is not just one conception of this public service ethic. Instead, they conclude that it means different things to different people as they identify four different conceptions: Samaritans, communitarians, patriots, and humanitarians. Common to all of these are a commitment to serving the public, making a difference, and support for social equity. Unlike Perry, Brewer, Selden, and Facer find that an interest in politics or policy making is not a characteristic of any of these conceptions of PSM.

Furthermore, monetary rewards are not considered to be significant motivators for this ethic (Brewer and Selden 1998; Brewer, Selden, and Facer 2000). For instance, volunteerism has been found to actually decrease in the presence of small monetary rewards, because extrinsic rewards “crowd out” the intrinsic rewards of the activity (Deci, Koestner, and Ryan 1999; Frey and Goette 1999; Frey and Jegen 2001; Frey and Oberholzer-Gee 1997).

Thus, PSM is a multidimensional concept that characterizes the motivations of individuals who engage in prosocial behavior (Brewer and Selden 1998; Perry 1996; Rainey 1982). A commitment to the public interest, service to others, and self-sacrifice underlie an understanding of PSM.

While it has been developed in the context of public administration, PSM is an individual, not a sector-specific, concept (Brewer and Selden 1998). However, it is more likely to characterize individuals in public rather than private organizations. One reason is that by virtue of their missions public organizations are more likely to provide individuals with an opportunity to engage in public service. Thus, individuals who are attracted to public service are likely to self-select into public organizations (Perry 2000).

Another reason relates to the level of goal clarity and specificity in organizations, as well as the presence of clear reward–performance expectations. A rational choice approach to motivation is based on an assumption of utility-maximizing behavior that occurs in the presence of well-defined goals and clear reward–performance expectations (Shamir 1991). However, public organizations are more likely to be faced with vague and conflicting goals established by external actors, responsibilities that are difficult to measure and monitor, and managerial constraints on the amount and use of monetary and other tangible rewards (Allison 1983; Brehm and Gates 1997; Perry 2000; Perry and Porter 1982; Wilson 1989). Under these conditions PSM provides a more useful basis for understanding employee motivation (Perry 2000).

Beyond the public–private distinction, the behavior of individuals in nonprofit organizations is more likely to be characterized by PSM than that of those in business. This is because nonprofits are likely to have a strong public service mission and to have environments that more closely resemble the messy nature of public organizations. For instance, nonprofit organizations are faced with external principals and constraints, multiple stakeholders who often possess different interests, and providing services that are difficult to measure and monitor (Corder 2001; Douglas 1987). Therefore, it is hypothesized that PSM more likely characterizes employees of public service organizations (public and nonprofit) than it does workers in the private sector.

One approach to testing the existence of PSM has been to operationalize it as a valuing of intrinsic job rewards above extrinsic ones (Crewson 1997; Houston 2000). Intrinsic rewards are those that are derived from the satisfaction an individual receives from performing a task (e.g., sense of accomplishment, feeling of self-worth). In contrast, extrinsic rewards are those offered to an employee by someone else (e.g., pay raise, promotion, status, and prestige). To address the existence of public service motivation in this way, research has asked the question: Do public employees rank intrinsic rewards (as compared to extrinsic rewards) higher in importance than private sector employees do? The most common answer is yes.

For instance, it has been found that public employees place less of an emphasis on higher pay as a motivator (Jurkiewicz, Massey, and Brown 1998; Kilpatrick, Cummings, and Jennings 1964; Rainey 1982; Schuster 1974; Wittmer 1991) and more emphasis on service to society and the importance of meaningful work (Crewson 1997; Houston 2000; Rainey 1982; Wittmer 1991). Research is not conclusive on this point, however, as several studies have not found differences between public and private employees consistent with the public service motive (Baldwin 1987; Gabris and Simo 1995; Maidani 1991).

A reason for the inconsistent findings in research on job motivators is the frequent use of limited research designs. Most of these studies employ surveys with samples that were no larger than 350 respondents (e.g., Baldwin 1987; Gabris and Simo 1995; Jurkiewicz, Massey, and Brown 1998; Maidani 1991; Wittmer 1991). In these instances, individuals surveyed came from only a handful of organizations, typically in the same locale. Additionally, comparisons of public and private sector employee responses are assessed with only bivariate statistics and, in a couple of cases, multivariate regression models with a limited set of control variables.

Crewson (1997) and Houston (2000) are notable exceptions in that both analyze survey data sets with large national probability samples and estimate multivariate models with more complete sets of control variables. It is worth noting that both Crewson and Houston conclude that motivational differences exist between public and private workers that are consistent with PSM.

Beyond job reward motivators, PSM suggests that public employees are more likely than private employees to possess attitudes that are “other directed.” Consistent with this expectation, Brewer (2003) finds that public employees score higher on attitudinal items related to social trust, altruism, equality, tolerance, and humanitarianism. Other studies have found public employees to possess more altruistic attitudes than private sector workers (Rainey 1997), be more supportive of democratic values (Blair and Garand 1995), and possess a higher sense of civic duty (Conway 2000). Based on these findings Brewer concludes that public servants “are motivated by a strong desire to perform public, community, and social service” (2003, 20).

It appears that public employees do indeed “talk the talk” of public service, but the PSM construct implies that they also “walk the walk” of public service. The behavioral implications of PSM must be studied to determine its utility for understanding motivation in public service organizations. It is hypothesized that public employees will be more civically active than employees of business.

One form of civic activity is involvement in the political process. Studies of political behavior have consistently found that public employees are more likely to vote in elections than others are (Garand, Parkhurst, and Seoud 1991a; Watson 1997; Wolfinger and Rosenstone 1980). This difference is even more pronounced among state and local government employees (Garand, Parkhurst, and Seoud 1991b). While increased political behavior is suggested by the public service motive, involvement in political activities such as voting may be a function of self-interest. A public employee may vote for candidates supportive of policies that will ensure that his or her livelihood is protected. Thus, examining the political activities of public employees provides a weak test of PSM unless researchers can rule out self-interest.

A more substantial test of the public service motive relates to civic involvement beyond politics, motivation for which is not easily explained by self-interest. This is the approach Brewer (2003) takes by looking at the general civic involvement of public employees. Using data from the American National Election Study he constructs an index of civic participation based on the number of civic groups an individual is involved in, belongs to, pays dues to, engages in activities with, and discusses politics in. After controlling for other relevant explanatory factors Brewer finds that public employees are more civically active as they perform more than one-third more civic activities than other citizens. This leads him to the conclusion that “public servants appear to be catalysts for the building of social capital in society at large” (2003, 5).

Additionally, Brewer and Selden (1998) examine whistle-blowing among federal employees as an act consistent with the public service ethic. Whistle-blowers are high performers who possess high levels of job commitment and satisfaction, yet place themselves at risk to further the public interest. As expected, it was found that whistle-blowers are more likely to possess PSM-related attitudes than individuals who observe but do not report inappropriate acts.

The analysis presented below extends the focus on the behavioral implications of PSM by examining civic participation in a similar vein as Brewer (2003) does. The PSM construct suggests that public employees will engage in behavior consistent with community-oriented and altruistic motives. By studying membership and participation in civic organizations Brewer (2003) examines behavior that is community oriented. The present study complements this research by examining charitable acts, which are consistent with altruistic or other-directed motives. In particular, attention will be paid to volunteering time to charitable organizations, donating blood, and making monetary contributions to charities. Additionally, the behavior of nonprofit employees is examined separately from that of government employees.

Each of these three charitable activities embodies the essence of the public service motive in terms of public interest, service to others, and self-sacrifice. While self-interest may motivate some to engage in charitable acts (e.g., to make contacts, enhance social status, gain expertise), a substantial volume of research indicates this behavior is largely accounted for by altruistic motives (Piliavin and Charng 1990). For this reason, an examination of charitable activities on the part of public service employees provides an appropriate test of PSM.

CORRELATES OF CHARITABLE BEHAVIOR

Public service motivation offers an explanation for the giving spirit. It holds that public service employees are more likely than others to engage in charitable activities. A rigorous test of this hypothesis requires controlling for other individual attributes known to correlate with charitable behavior. The research literature on volunteerism, blood donation, and individual philanthropy provides guidance for the development of a multivariate model that will serve as the basis for properly testing the PSM thesis.

With the growth of the nonprofit sector a substantial research literature has developed examining the gift of time. Among individual attributes, socioeconomic status is a strong determinant of this behavior (Goss 1999). Higher levels of education, as well as income, have consistently been found to correlate with higher levels of volunteering (Reed and Selbee 2001; Tiehen 2000; Villancourt 1994; Wilson and Musick 1997). Although it is less frequently examined, individuals with more prestigious occupations also exhibit greater tendencies to donate their time to charitable organizations (Reed and Selbee 2001). It has been theorized that membership in these “dominant status” segments of society provides one with greater social status and prestige that is likely to be reinforced or enhanced through volunteer activities (Goss 1999; Lemon, Palisi, and Jacobson 1972; Smith 1983). These individuals also are likely to have the resources that permit them to engage in charitable acts.

Although not a “dominant status” group, women are more likely to volunteer than men (Caro and Bass 1995; Villancourt 1994). Similarly, while the above thesis would indicate that whites are more likely to volunteer than minorities, the relationship between race and volunteerism has generated mixed results (Bobo and Gilliam 1990; Palisi and Korn 1989; Sundeen 1992; Wilson and Musick 1997).

Another common thesis holds that the more embedded one is in the community, the more likely one is to volunteer (Chambré 1987; Goss 1999; Smith 1994). This social connectedness thesis suggests that individuals who are married and own their home are more likely to be settled into their community and feel a responsibility for their neighbors. Similarly, as individuals age their propensity to volunteer increases (Reed and Selbee 2001; Tiehen 2000). However, this relationship may be curvilinear, as some research suggests that volunteering begins to decline after the age of fifty-five (Chambré 1987; Clary, Snyder, and Stukas 1996). This decline may be a function of increasing poor health and other factors that pose barriers for the elderly to civically participate.

Social connectedness also is influenced by family composition and community size. The presence of school-age children in the household increases an individual's exposure to volunteer activities (Reed and Selbee 2000; Tiehen 2000; Wilson and Musick 1997). Also, smaller communities are likely to promote a greater sense of identity and belonging, thereby increasing a sense of community in general and making it easier for individuals to get involved (Sundeen 1992; Sundeen and Siegel 1987; Villancourt 1994).

Finally, religiosity must be considered in any list of sociodemographic correlates of volunteerism. Religiosity influences the gift of time in two ways. First, affiliation with a church exposes individuals to social networks that draw people into volunteering activities (Becker and Dhingra 2001). In this way religiosity can be thought of as an additional factor that influences social connectedness. Second, going beyond social connectedness, religiosity influences the internalization of values consistent with volunteerism. Churches promote cultural norms of benevolence, charity, and service to others (Eckel and Grossman 2004; Wuthnow 1991). Historically religious organizations were instrumental in addressing the unmet needs of a community (Putnam 2000).

Therefore, the greater an individual's level of religiosity, the more likely an individual is to volunteer (Greeley 1997; Lam 2002; Putnam 2000). For instance, Gill (1999) found that those who volunteer are more likely to be weekly attendees of church than nonvolunteers. Other scholars have similarly found churchgoing to be positively correlated with a greater propensity to volunteer (Becker and Dhingra 2001; Chambré 1987; Putnam 2000; Wilson and Musick 1997; Wuthnow 1991). Another aspect of religiosity, denominational membership, also influences volunteering patterns (Curtis, Baer, and Grabb 2001; Hoge et al. 1998; Wilson and Janoski 1995). Lam (2002) looks beyond churchgoing and membership to focus on other aspects of religious life. He concludes that the frequency of prayer and religious reading is positively associated with giving the gift of time. However, the increased volunteering (and even philanthropy) that results from religiosity is generally confined to related religious organizations and does not translate into giving more time to secular organizations (Cnaan, Katernakis, and Wineburg 1993; Gerard 1985; Putnam 2000; Sundeen 1992; Wuthnow 1991).

While this review of literature has focused on volunteering behavior, these same sociodemographic variables will be used to explain donations of blood and money. This is justified on the basis that volunteering and making financial contributions to charitable organizations are complementary activities (Lee, Piliavin, and Call 1999; Menchik and Weisbrod 1987; Putnam 2000). Individuals who volunteer are more philanthropic than nonvolunteers (Independent Sector 2002; Piliavin and Charng 1990; Reddy 1980). For instance, a study of charitable giving and volunteering among Canadians estimates that a core of “29% of adults in 2000 [accounted] for 85% of total volunteer hours, 78% of charitable dollars donated, and 71% of civic participation” (Reed and Selbee 2001, 765). As a result, sociodemographic correlates of volunteering (e.g., education, income, religiosity) are similarly related to charitable giving (Barry 1996; Eckel and Grossman 2004; Independent Sector 2002; Jackson et al. 1995). One notable exception is that charitable giving does not decline in later years but continues to increase throughout one's life (Independent Sector 2002; Jackson et al. 1995).

Blood donation, on the other hand, is viewed as a charitable act that is quite different from volunteering and philanthropy (Lee, Piliavin, and Call 1999). Individuals engage in giving “the gift of life” far less frequently than they volunteer or make a charitable contribution.¹ A nationwide survey of 4,216 adults found that in the year 2000, 88 percent of respondents made a charitable contribution and 44 percent volunteered (Independent Sector 2002). In contrast, only 4–6 percent of the eligible population donates blood in any given year (Lee, Piliavin, and Call 1999). Beyond being less frequent, blood donation is different in that the donor must address anxiety, fear, and pain (Piliavin and Callero 1991) and not everyone is eligible to donate (Titmuss 1971).

The typical profile of the blood donor is one who is white, married, and male; possesses a college education; is in his thirties or forties; and has a white-collar job (Bettinghaus and Milkovich 1975; Oswalt 1977; Piliavin and Callero 1991; Titmuss 1971). The relationship between age and blood donation is likely to be curvilinear, as individuals have lower rates of donation when they approach the age of fifty and are even less likely to donate in their sixties (Pindyck et al. 1987).

In sum, sociodemographic attributes are correlated with volunteerism, philanthropy, and blood donation. A proper test of the implications of PSM for these charitable acts necessitates controlling for these attributes. Based on PSM it is hypothesized that public service employees will volunteer, donate money, and give blood more than private sector employees, after controlling for other sociodemographic variables. In this way, the estimated models provide a rigorous test of the behavioral implications of PSM.

DATA AND METHODS

The data for this project were taken from the 2002 General Social Survey (Davis, Smith, and Marsden 2003). Key variables used in the analysis below came from modules added to the 2002 survey and were administered to 1,796 respondents—or 65 percent of the entire sample.

Measures of participation in public service activities were developed from questions that asked how often during the past twelve months the respondent volunteered for a charity, donated blood, and gave money to a charity. Ordinal response categories were collapsed to create dichotomous variables indicating that an individual either had or had not done each of these activities during the past year.²

Two measurement schemes were used to classify respondents based on their sector of employment. First, a dummy variable was created to represent whether or not a respondent is employed in a “government” organization.³ Second, an alternative classification of the sector in which a respondent is employed was created from the survey question: “Do you work for a private company, a nonprofit organization, or the government or a government agency?” Two binary variables were created from responses to this question representing whether a respondent works for a government agency or a nonprofit organization.⁴ The advantage of this second scheme for measuring employment sector is that it permits comparing the behavior of nonprofit employees with both public and private sector workers.

Additional dummy variables were created to represent whether a respondent is female, white, or married. Age is represented in both level and squared forms to permit testing for a curvilinear effect on charitable activity. Occupational prestige is an estimation of the social standing of an occupation as revised by Nakao, Hodge, and Treas (1990). In the present sample it ranges in value from seventeen to eighty-six and has a mean of 43.86. Last, the number of children seventeen years of age and younger in the household and the log of community size (population) are additional continuous variables entered in the models.

Testing the hypothesis related to volunteer activity implicit in PSM initially is performed in a bivariate analysis using “crosstabs” and chi-square test statistics. To control for the influence of sociodemographic variables found to be significant in previous research on volunteerism, blood donation, and individual philanthropy, multivariate logistic regression models are also estimated.

FINDINGS

Of the 1,796 total respondents who were asked questions that are the basis of this analysis, missing data on one or more questions reduce the final sample used to 1,301 for the bivariate analysis and 1,142 for the multivariate analysis. As mentioned above, categorizing respondents based on their sector of employment is accomplished in two ways. First, using an existing General Social Survey variable, 17 percent (or 223) of the respondents are employed in government and the remaining 83 percent (or 1,078) are not. Second, using another variable that has a three-sector breakdown of employment results in 17 percent (or 153) of respondents classified as government workers, 7 percent (or 61) as nonprofit employees, and 76 percent (or 679) as employed in private companies. While this second classification allows distinguishing nonprofit from private personnel, its use has the drawback of reducing the number of available cases for analysis. To take advantage of the larger available sample for the first employment classification variable and the separate nonprofit and for-profit categories of the second classification scheme, both will be used in separate models.

In terms of the frequency of the three charitable activities, the most common is donating money to charity, which was reported by 78 percent of respondents (see figure 1). Slightly less than half (45 percent) indicated that they had volunteered their time for a charity during the past twelve months. The least common activity is donating blood, something that was reported by only 15.6 percent of respondents. Last of all, the remaining 17 percent engaged in none of these activities over the past year.

If an individual engages in one of these activities, she or he is likely to engage in another. This is especially true among volunteers. Nearly all volunteers made a charitable monetary donation, a finding consistent with previous research that indicates volunteers are more philanthropic than nonvolunteers (Independent Sector 2002; Piliavin and Charng 1990; Reddy 1980).

The bivariate analysis reported in table 1 indicates that public employees are more likely than other citizens to report engaging in each of these three activities. However, the largest of these differences is only 12 percent (volunteering time), and the smallest is 8 percent (blood donation). While these differences between public employees and others in terms of charitable behavior are statistically significant, they are not as substantively impressive.

When respondents are grouped by employment sector the differences are more substantial (see table 2). Individuals employed by government or a government agency are more likely than those employed in a private company to volunteer time, donate blood, and donate money. Nonprofit employees are also more likely to have performed each charitable act than are private employees. Furthermore, nonprofit workers are more likely to have volunteered than even public employees. Overall, it appears that nonprofit workers are more similar to public than private employees. These findings are consistent with the hypothesis derived from PSM that public employees (or public service employees in general) are more likely to engage in charitable acts than private employees.

While consistent with PSM, the bivariate analysis offers a relatively weak test for the public service ethic. The remainder of the analysis reports the results of multivariate logistic regressions. Two models are estimated for each charitable activity. The first model in each table differentiates between government employees and others, while the second distinguishes among three categories of employment (government, nonprofit, and private).

Table 3 reports logistic regressions for giving the gift of time to a charitable organization. In both models the public employee variable is positive and statistically significant. The predicted probabilities calculated based on model #1 indicate that government employees have a 51 percent probability of volunteering compared to others who have a 43 percent probability of doing so.⁵ Thus, being a government employee increases the probability of donating time by about 8 percent, although this effect is statistically significant at only the 0.10 level.⁶ The difference between public and private employees may be suppressed in model #1, however, because the “other” category comprises more than just employees of for-profit organizations. A better test of the differences between public service and private personnel is provided by model #2, where the base group is private employees and separate nonprofit and government employee variables are included. In this second model both public service parameter estimates are statistically significant. The predicted probabilities based on model #2 indicate that private employees have a 34 percent probability of volunteering as compared to 46 and 52 percent for government and nonprofit employees, respectively. Therefore, being a government employee increases the probability of volunteering over private workers by 12 percent, whereas being a nonprofit employee increases this probability by 18 percent. The difference in the propensity to volunteer between those in the nonprofit and public sectors is not statistically significant.⁷

Several other factors emerge in both models to help explain self-reported volunteering behavior. Consistent with previous research, women are more likely than men to give time, with predicted probabilities of volunteering of about 43 and 34 percent, respectively. Additionally, higher socioeconomic status (education, income, occupational prestige) translates into a higher propensity to volunteer. Age emerges as statistically significant only in model #1, and contrary to previous research it is negatively related to the giving of time, suggesting that as one ages one becomes less likely to give of one's time.⁸ The presence of young children in the household is also significantly correlated with the donation of time in the hypothesized positive direction, but only in model #2. Finally, individuals who attend church at least once a week have a predicted probability of volunteering that is at least 22 percent higher than those who do not attend church as regularly.⁹ In sum, the profile of an individual most likely to volunteer for a charitable organization is a female who is employed in a public service organization (government or nonprofit), possesses high socioeconomic status, has children under the age of seventeen in the household, and attends church at least once per week.

Giving “the gift of life” is much more difficult to explain. The logistic regression models in table 4 do not perform nearly as well, as indicated by few statistically significant coefficients and poor goodness of fit indicators. With that said, PSM does help to explain this charitable act. The predicted probabilities calculated based on model #3 indicate that government employees have a 14 percent probability of donating blood compared to a 10 percent probability that other citizens will do so. Model #4 casts government employees in an even better light as these individuals have a 24 percent probability of giving blood compared to 15 and 19 percent for private and nonprofit employees, respectively. Even though the predicted probability of donating blood is higher for nonprofit versus for-profit personnel as predicted by PSM, this difference is not statistically significant. Beyond public employment, the only other factors that emerge as statistically significant correlates of blood donation in both models are age and religiosity. In both models, the probability of having donated blood in the past year initially increases with age but declines during midlife years (beginning about the age of thirty-six).¹⁰ Church attendance also influences giving “the gift of life,” as those who go to church at least once per week have a 17 percent probability of donating blood as compared to 10 percent for others, based on model #3. The difference between regular churchgoers and others is even larger in model #4, as the predicted probabilities of donating blood are 27 and 15 percent for these two groups, respectively. Thus, the profile of the likely blood donor is a female government employee in her mid thirties who attends church at least weekly.

While the bivariate analysis indicates that public service employees are more likely than others to donate money to charity, this pattern does not hold up in the multivariate analysis. Table 5 reports the two models that explain self-reported financial contributions to a charitable organization over the past twelve months. For both models #5 and #6 the public employee variable is in the expected direction, but neither of these are statistically significant. Not only is the nonprofit employee variable insignificant in model #6, it is in a direction contrary to PSM. Taken together, the performance of the public service employee variables in these two models indicates that PSM is not useful for explaining this charitable act. In contrast, sex, socioeconomic status, and age are important correlates of giving money. Women are more likely than men to have made a financial contribution to a charitable organization during the past year. The probability that women reported having donated money is 6–8 percent higher than it is for men.¹¹ Additionally, socioeconomic status and age are positively correlated with self-reports of this behavior.¹² Surprisingly, religiosity is unrelated to giving the gift of money, a finding contrary to much research on philanthropic behavior. In sum, the typical philanthropic individual is female, is highly educated, has a high income and high occupational prestige, and is above average in age.

While this approach to testing the PSM thesis has substantial merit, the reliance on self-reported charitable behavior obtained through personal interviews has limitations. It is reasonable to expect that in a personal interview there would be a social desirability bias in responses to questions about charitable acts. This bias would likely be most evident in reports of the frequency of engaging in these acts. By operationalizing charitable behavior merely as a yes/no dichotomy, social desirability bias may be less problematic. However, if individuals do feel the pressure to indicate that they volunteered, donated blood, or gave money when in fact they did not, the result likely would be to depress findings of a relationship between the sociodemographic variables and charitable behavior. This would in part explain the weak predictive value of the estimated models. Additionally, it is not problematic because there is no reason to think that public or nonprofit workers would be more biased in reporting their charitable activities than those in the private sector are.

Furthermore, the analysis does not distinguish altruistic from nonaltruistic charitable giving. To the extent that nonaltruistic motives differ across the three sectors of employment, it is most likely to characterize for-profit employees who may feel greater work-related pressure to civically participate to promote their organization.¹³ The result would be to further diminish the observed differences in behavior. Additionally, public and nonprofit workers do not have a monopoly on altruistic behavior, as certainly many private sector workers perform charity from their hearts. The implication of these caveats is that the analysis above is an even more rigorous test of PSM than suggested earlier, thereby increasing confidence in the conclusions that are drawn about public service motivation.

CONCLUSION

The profession of public administration typically is viewed as a special calling, and those who answer the call are seen as being different from those who do not. This difference has been characterized as public service motivation, which is a commitment to the public interest, service to others, and self-sacrifice. While an individual-level construct, it is expected that PSM is more prominent in the public sector due to the environment and constraints that characterize these organizations. Furthermore, PSM should be more apparent among activities that require greater levels of self-sacrifice.

Research generally indicates that public employees “talk the talk,” but do they also “walk the walk” of public service? Following the lead of Brewer (2003), the above analysis examined the behavioral implications of PSM by studying self-reported giving of time, blood, and money to charity. Engaging in these charitable acts is consistent with the theory of PSM and is not sufficiently explained by a rational choice theory of motivation. The findings reported above indicate that public employees are more likely to volunteer for charity and donate blood. These findings lend support for the hypothesis that PSM is more prominent in public than private organizations.

This study went beyond the typical public–private comparison and examined the behavior of nonprofit employees. Due to similarities in the nature of the services provided and general operational environment, nonprofit organizations are more akin to public than private organizations. It was hypothesized that the public service motive would also characterize nonprofit employees. At least in terms of volunteerism, employees of nonprofit organizations also are more likely to be charitable than private sector workers are. The growing importance that nonprofit organizations are assuming in public service delivery highlights the necessity to consider the motives of these public service workers. Additional research focusing more fully on these members of the public service is needed before more definitive conclusions can be reached about the motivation of nonprofit workers.

A theory of PSM has significant implications for public administration. It suggests that motivational tools common in private management may not be as effective in public organizations. Private sector reward systems typically are built on the primacy of extrinsic rewards and may “crowd out” the intrinsic rewards that motivate public service employees. This may be one reason that pay-for-performance systems have had limited success in public organizations (Ingraham 1993; Kellough and Lu 1993). Instead, public sector incentive structures must provide an opportunity for employees to satisfy their public service motives.

PSM findings suggest that selecting individuals with norms consistent with PSM is one of the most important tools public managers have for enhancing organizational performance (Brehm and Gates 1997). How is this to be accomplished? First, managers must recognize that professional education provided in public administration and affairs programs is a valuable asset for potential employees. Active recruitment through established graduate programs may help to identify potential employees with the values consistent with public service who are in the nearby community. Second, increasing contact with future employees early in their academic careers by hosting interns or serving as classroom guest speakers may expose students to the importance of a career in public service. Third, it is vital to be aggressive in touting the successes of public programs and the rewards of public service, instead of passively watching only negative stories about the profession hit the front page of newspapers. Doing so may capture the attention of potential public service professionals and create career aspirations that otherwise would not be stimulated.

It suggests that social capital may be enhanced through the role that public bureaucrats play in governing (Brewer 2003). As civically active citizens, public employees are in a prime position to bring people together to initiate a dialogue on community issues. Engaging in charitable acts can increase trust among citizens and get more people involved in the community.

PSM also suggests that calls for public managers to serve as community leaders (Behn 1998) do not pose significant threats to democracy. Public officials are dedicated public servants concerned about the community and use the formal authority of their positions to serve the public interest. They are individuals who enhance democratic governance by being responsive, open to public input, and acting on what is in the best interests of the community. PSM thereby helps to legitimize the role public bureaucrats play in governing. These are a few of the behavioral implications of PSM that need to be examined more thoroughly.

In any event, it appears that public service motivation is evident in the charitable acts of public administrators. Public employees not only “talk the talk,” they also “walk the walk” of public service. These government officials have been given many labels (e.g., public employee, public administrator, government worker, bureaucrat). Perhaps “public servant” is most apropos.

References