Praxis+Assignment

=Praxis Assignment=

The Demand for Volunteer Labor: A Study of Hospital Volunteers
====This quantitative article looks at the relation between cost and demand for volunteer labor. It mostly talks about volunteers in hospitals but I feel like the article also gives a good look at many corporations looking for volunteer labor.====

Handy, Femida, and Narasimhan Srinivasan. "The Demand for Volunteer Labor: A Study of Hospital Volunteers." //SAGE Journals Online//. SAGE. Web. 19 Feb. 2011. .

Femida Handy University of Pennsylvania and York University Narasimhan Srinivasan University of Connecticut The authors challenge the assumption that organizations are willing to use all the volunteer labor available to them. Rather, they are influenced by the costs incurred of utilizing volunteer labor. This article provides a modest first look at the demand for volunteers by nonprofit institutions. Specifically, the article presents an economic analysis of the demand of volunteer labor by hospitals in the Toronto area and examines some of the factors that may determine the hospitals’ willingness to use volunteer labor. Using data generated from 28 hospitals in Toronto, which use a total of more than 2 million volunteer hr per year, the authors show that the quantity of volunteer hours demanded is a decreasing function of their costs. Other factors such as productivity, output, and labor market institutions also influence the demand for volunteers. Keywords: hospital volunteers; demand for labor; nonprofits; Canada An important avenue of resource allocation in the Canadian economy is private philanthropy. In 2000, gifts of money and time amounting to $4.9 billion and 1.1 billion hours, respectively, were given in private donations to nonprofits in Canada. This is a significant share of resources transferred voluntarily by Canadians. Twenty-seven percent of all adult Canadians volunteer, and the economic value of their time donations is valued at more than $14 billion, which represents 1.4% of Canada’s GDPand far exceeds the monetary donations (Hall, McKeown, & Roberts, 2001; Statistics Canada, 2004b). More than 161,000 nonprofit and voluntary organizations in Canada employ volunteers in the production of a variety of goods and services. Nine out of 10 volunteers (93%) are engaged in the production of services in these organizations (Statistics Canada, 2004a). In a national survey of volunteers, more than half (57%) of all volunteers reported that they helped to organize or supervise activities, 41% served on working committees and governing boards, 40% were involved in canvassing or fund-raising, 30% provided consulting or administrative work, and somewhat less than 30% were involved in educating, lobbying, coaching, providing care, delivering food, driving, performing maintenance, and so forth on behalf of an organization (Hall et al., 2001). Although there are many studies that examine individual decisions to supply volunteer labor, there is a paucity of literature on organizational demand for such labor. This research addresses this gap in the literature and seeks to delineate factors that influence an organization’s demand for volunteer labor. We study a particular set of organizations—hospitals—that have a long tradition of using both paid and volunteer labor. Specifically, we examine hospitals in Toronto, which are increasingly relying on volunteer labor to enhance the quality of health care provided. Traditionally, hospital volunteers consisted of hospital auxiliaries that were mainly composed of society ladies and spouses of physicians and who ran gift shops and helped with fund-raising. More recently, however, hospitals have been recruiting large numbers of volunteers from all walks of life for these and many other types of services.1 These trends in the use of volunteers by hospitals have required hospitals to move from ad hoc management of volunteers to engaging professional managers to train, screen, and manage volunteers. Although this professionalization of volunteer management was, in part, a response to the growing number of volunteers, it was also a response to the increasing vulnerability of hospitals to liability issues. In turn, this professionalization has increased the costs (per volunteer) that hospitals face in using volunteer labor.2 Using data generated from interviews with the CEOs of 28 hospitals in the Toronto area and data on their volunteer programs, we examine some factors that may influence the demand for volunteers within hospitals. We proceed as follows: The next section discusses the literature on volunteer labor. This is followed by a section that considers factors that may determine the extent of the employment of volunteer labor in hospitals. In this part, we also consider how CEOs’ attitudes toward their volunteers may be viewed as proxies for some of the factors that affect demand for volunteer labor. The subsequent section presents our research methodology and is followed by a report of the findings of our research. In the final section, we analyze our results and offer concluding remarks. LITERATURE REVIEW OF VOLUNTEER LABOR When constructing the supply function of unpaid labor, economists deviate from the traditional models used for paid labor, which are often modeled as a function of wages, nonwage income, and available hours. Because volunteer labor receives a zero wage, it must be viewed differently (Freeman, 1997). There exist several models of the supply of volunteer labor, each with some predictive value. First is a consumption model in which volunteering is a utility-yielding activity—the individual receives satisfaction from the very act of volunteering. This individual maximizes utility subject to a time and budget constraint (Andreoni, 1990; Menchik&Weisbrod, 1987; Prouteau&Wolff, 2004; Segal & Weisbrod, 2002). Second is an investment model in which volunteering is undertaken by an individual to enhance future income potential. Here, the individual invests volunteer hours out of available leisure hours to maximize future earnings (Katz & Rosenberg, 2004; Menchik & Weisbrod, 1987). Other models include viewing the output of volunteers as a public good, where the volunteer receives satisfaction from the output produced by his or her volunteering efforts (Schiff, 1990; Unger, 1991). In addition to theoretical work, there is a large empirical literature on the supply of volunteer labor. Much of this literature has focused on identifying socioeconomic and personal characteristics that are likely to predict volunteering among the general population using the above models (Carlin, 2001; Freeman, 1997; Menchik & Weisbrod, 1987; Proteau & Wolff, 2004; Smith, 1994; Vaillancourt, 1994; Vaillancourt & Payette, 1986; Van Dijk & Boin, 1993; Wolff, Weisbrod, & Bird, 1993). Large empirical surveys profile typical volunteers, examine the motives for volunteering, specify where volunteers tend to work, and focus on the characterization of volunteers based on their socioeconomic demographics (Davis Smith, 1998; Hall et al., 2001; Hodgkinson & Weitzman, 1992; Independent Sector, 2001). In all cases of modeling the volunteer labor supply, the implicit assumption is that organizations are willing to use all the volunteer labor that is offered (quantity and quality) for each type of volunteer job they establish. In other words, the assumption is made that demand for (volunteer) labor is infinite when the wage rate of labor is zero.3 However, although volunteers do not impose direct wage costs, there are other costs that they do impose. The equating of a zero wage rate with zero costs is not, in general, a realistic assumption, and it ignores the reality that faces organizations that use volunteer labor (Steinberg, 1990). An organization’s nonwage costs of employing volunteers in terms of day today operating costs such as recruitment, screening, training, managing, and providing office space, materials, and so on are significant. Indeed, such costs are particularly important within the context of hospitals, where screening and training are essential (Handy & Srinivasan, 2004). In our interviews with hospital CEOs, we found that volunteers are subject to careful screening, orientation, training, and often work side by side with paid health professionals. Although hospitals minimize risks (and their liability) by carefully prescribing what volunteers may or may not do, issues of privacy, contact with minors, and other sensitive issues generally require that volunteers be carefully screened and well trained. Hospitals in the Toronto area are involved in a continuous recruiting of volunteers. However, they do not accept all volunteers who apply. In some hospitals, waiting lists exist for certain volunteer jobs, whereas in a few there are shortages, especially of skilled volunteers. Many reported that they cannot expand their volunteer base due to the lack of resources to manage the increasing number of volunteers. Individuals looking for specific volunteer positions, specific hours, or those who could not make a commitment for a specified minimum number of hours find it difficult and sometimes impossible to obtain volunteering work (Karom Group of Companies, 2001; LaPerriere, 1998). As one CEO told us, “We will have to rethink our policy in accepting short-term volunteers; the turnover is very costly.” Thus, organizations clearly view volunteers as imposing a cost on them despite the fact that the wage rate of volunteers is zero. Emanuele (1996) found that nonprofit organizations in the United States seem to be choosing the amount of volunteer labor they use in accordance with an implicit downward sloping demand curve for volunteer labor that is consistent over time.Due to data limitations, she was not able to uncover specifications for the demand curve for labor. Emanuele’s findings lend credence to the fact that organizations using volunteer labor do not accept all volunteer labor that is supplied and associate costs with its use. DEMAND FOR VOLUNTEER LABOR IN HOSPITALS To fully understand the amount of volunteer labor being utilized, it is necessary to understand the supply and demand curves for volunteer labor. Equations of volunteer labor supply are based on utility-maximizing individuals’ decisions on how to allocate their time. The individual’s decision to provide hours of volunteer labor includes the opportunity costs facing the individual, the after-tax price of charitable contributions (considered as a substitute for volunteering for the individual), wealth, and attitudes to volunteering. The latter are usually proxied by socioeconomic characteristics such as age, gender, education, and religiosity. Due to differing expectations and objectives as well as histories, geography, and culture, different organizations attract and utilize different types of volunteer labor. For example, the ownership of the organization, or the subsector of the economy in which the organization is located, can lead to a sorting out of volunteer labor (Segal & Weisbrod, 2002). To minimize these differences, we focus on volunteer labor in a relatively homogeneous sector, namely, publicly subsidized nonprofit hospitals in the Toronto area, which face similar political and cultural environments. In an era of increasing demand for health care and a dwindling of public support, publicly subsidized nonprofit hospitals in Toronto have increasingly turned to private donations of money and time to augment their resources. Such donations fund capital expenditures and new services as well as maintain existing services. Some hospitals rely on significant amounts of volunteer labor to produce many of their services. In principle, it is necessary to know the organization’s objective function in order to determine the derived demand for volunteer labor. The demand curve for volunteer labor should be derived from the objective function of the organization, utilizing volunteer labor as one of the inputs in production. Organizations are faced with a choice of how much to produce and also how much to use of each of the various inputs of production. In other words, the demand function for one input, that is, hours of volunteer labor, will depend on its productivity, its price, and other available substitutes. Several non economic factors will also influence the decisions of organizations to use volunteer labor. However, as is well recognized, the objective function of a nonprofit is not well understood (Steinberg, 1990). It is, of course, possible to focus on a “sub objective function” as done by Schiff (1990), who assumed that this function is the net value of volunteers. However, such objective functions are simplistic. For example, assume that hospitals wish to maximize the amount of health care they provide.4 To do this they need funding. Hence, a sub objective function may be to maximize revenue (including donations of time). However, that does not imply maximizing the net value of volunteers. First, volunteering may crowd out other donations by the private sector, by substitutions of time for money donations. Moreover, volunteering and money donations may crowd out government support (Handy & Webb, 2003). Conceptually, various sub objective functions could theoretically be collapsed into a single measure such as the monetary equivalent of the present value of patient welfare, but it is difficult in practice. In summary, it is almost impossible to define a hospital’s objective function, or even its sub objective function. Despite this, for most reasonable objective functions, it is possible to derive testable predictions of the demand for volunteer labor by assuming that the hospital pursues its goals in an efficient manner. For example, the choice of using an additional hour of labor as an input should be made if and only if the value of the additional output (marginal rate of productivity [MRP]) fromthis hour is equal to the price paid for this hour. For overall efficiency, this should be true for all inputs of production, and at equilibrium these input ratios of MRP to price should be equal for all inputs.5 For example, hospitals will eschew volunteer labor as its price increases (the costs per hour of volunteer labor incurred by the organization) and turn to substitute inputs with lower input ratios (of MRP to price) such as minimum wage labor. Thus, we expect the utilization of volunteer labor to be positively influenced by its productivity and negatively influenced by its costs. A hospital will also be mindful of the environmental constraints under which it operates, such as the existence of labor contracts and its obligations to the community as reflected in its mission statements. Thus, the factors influencing the demand for volunteer labor likely include the cost and productivity of volunteer labor, the total output generated, and non economic factors such as organization culture and organizational constraints. We now examine each of these factors and then suggest ways of measuring them. Costs Per Hour of Volunteer Labor As explained above, efficiency requires that labor as an input is hired until the value added by the last hour is offset by the cost of that hour. Hence, even if, for institutional reasons, this equality is not strictly maintained, it is clear that the amount of labor used is dependent on the level of wages: The demand function for labor shows a negative relationship between the wage levels and the amount of labor demanded. In the case of volunteers, although the explicit wages are zero, zero wages does not imply that there are zero costs to the employer. As discussed earlier, volunteers impose costs of recruiting, screening, training, managing, and recognition. These costs are likely to be substantial. A reasonable proxy for such costs in a hospital is the volunteer administration budget, which includes the salaries and office costs of managers whose task is to coordinate and manage volunteer resources in the hospital. We argue that such costs per volunteer hour affect the organization’s decision regarding the amount of volunteer labor to use. When these costs increase, hospitals are likely to use less volunteer labor: We hypothesize a negative influence of costs per volunteer hour on the number of volunteer hours used by the hospital. Scale-Output As output increases, the organization’s demand for inputs will increase; thus, the number of volunteer hours an organization uses will have a direct relationship to its output. We use the size of the hospital as captured by the number of beds as a proxy for output. This is a reasonable measure of the output of a hospital because the number of patients in the hospital at any one time is a good proxy for the output produced by the hospital. Also, given that all the hospitals in our sample receive their revenue from the same insurance source, that is, the Government of Ontario, the number-of-beds measure may be a reasonable proxy for relative output. Trade-Off for Time Donations Money donations allow the CEO to purchase other inputs for health care production, whereas time donations can only be used to produce services provided by volunteers. Thus, we focus on the trade-off CEOs are willing to make between time donations and money donations, to indicate what they perceive to be the relative advantage of using volunteer labor in the production of health care. This acts as a proxy for the MRP ratios. Assume a CEO can choose a monetary donation instead of 1 hr of volunteer time, everything else being equal. We use a hypothetical question to elicit the minimum money donation that the CEO would be willing to forgo for 1 hr of volunteering. This is an indicator of the CEO’s dollar valuations of an hour of volunteer time as an input in production. To put it differently, the value put on an hour of volunteer time is the hospital’s opportunity cost of giving up that hour. If, for example, a volunteer hour can be replaced by paid labor at the cost of $12.50, then any trade-off value above $12.50 suggests that the CEO intrinsically values the work of a volunteer in excess of what paid labor (a substitute) can produce in terms of the hospital’s direct output. Higher valuation of the volunteer hours would suggest that volunteer labor produces benefits to the hospital beyond its labor. These may include the goodwill generated by their presence, their service as ambassadors to the general public, enhancement of community relations, and so on. For example, if the CEO places the trade-off value at $50.00, she or he is signaling that the value of the production of 1 hr of volunteer labor is 4 times greater than the value of the output produced by 4 hr of a substitute input, that is, an hour of paid labor (at $12.50). Thus, we hypothesize that hospitals where CEOs provide higher trade-off values on volunteering hours will have greater utilization of volunteer labor than the hospitals where CEOs provide lower values. Productivity and CEO Satisfaction In the case of volunteer labor whose input is often in the softer services, such as making patients and their families comfortable and other nonmedical jobs, there is no single quantitative measure of productivity. Although some outputs of volunteer labor such as running gift shops, providing information, answering telephones, and assisting with clerical tasks may be measured using market wages,6 it is difficult to measure this without detailed data on the number of volunteer hours assigned to each task. Furthermore, such measures would assume that in the absence of volunteer labor, paid labor would be hired to replace volunteer labor; however, it is not clear that in the absence of volunteer labor all of the services would be offered, thus making replacement value at market wages inappropriate as productivity measures. How, then, should the productivity of volunteer labor be evaluated? We suspect that in practice, the value of such labor is judged subjectively. If so, it may be proxied by how the CEO of the hospital perceives volunteers as fulfilling the mission of the hospital. We therefore suggest that the utilization of volunteer labor will be influenced by its productivity as perceived by the CEO.7 Because use of volunteer labor requires the hospital to expend real resources, the ultimate decision on allocation of resources rests with the CEO. The CEO’s evaluation of the output produced by volunteers will therefore affect the hospital’s demand for volunteer labor. Demand for Volunteer Services and Mission Statements Volunteer labor cannot be used in direct medical interventions. They can, however, be used to produce many soft services that focus on the comfort and emotional well-being of the patients and on reducing patient and family anxiety (Handy & Srinivasan, 2004). The extent to which a particular hospital values soft services can be gleaned by examining the hospital’s mission or vision statements. These statements almost invariably make mention of the quality of nonmedical care, that is, care that can be produced by volunteers. The following hospital mission statement is illustrative: “We believe that compassion, caring and technical excellences are equally important.”8 In view of this, Govekar and Govekar (2002) suggest that to ascertain the demand for volunteer labor, it is necessary to examine the institution’s mission statement to see if the statement reflects the kinds of services that volunteer labor produces. Such a public declaration in the mission statement of the kind of services produced by the hospital—services that are generally provided by volunteers— signals that volunteer labor is an essential input in the production of the services provided. Thus, we posit that mission statements of hospitals that explicitly mention that the care provided by the hospital includes the services generally provided by volunteers, or explicitly mention volunteers in the delivery of health care, are likely to reflect a high demand for volunteer labor. Organizational Constraints: Labor Contracts Often, paid workers view volunteers negatively because volunteersmaybe viewed as replacement for paid labor. In cases where labor contracts exist to prevent volunteer labor substituting for paid labor, any perceived substitution between paid and unpaid work can lead to friction in unionized environments and is subject to grievance (Macduff, 1997; Zahnd, 1997). In these cases, a hospitalmaybe constrained fromusing the desired levels of volunteer labor9 both directly, as a result of numerical constraints on the number of volunteers producing a given service, and indirectly in order to achieve industrial peace. Furthermore, due to hospital regulations (to minimize their liabilities) volunteers cannot provide any service that is of a medical nature or done by medical professional staff. Although some of the hospitals in our sample did not have labor contracts, the contracts of those that did had an explicit clause stating that volunteers may not perform work done by paid staff except in those areas that were run by volunteers before 1986 (Canadian Union of Public Employees [CUPE], 2000).10 We therefore hypothesize that hospitals with collective labor contracts will use less volunteer labor that those not subject to such constraints. To sum up the above discussion, Table 1 identifies the factors influencing demand and provides the expected direction of the relationship between the dependent variable—hours of volunteer labor utilized by the hospital—and the six independent variables. In later sections, we examine the data to determine whether the six factors mentioned above explain the use of volunteer labor by hospitals. Before doing so, we describe in the next section the methodology used in obtaining the data. METHOD Data for this study were obtained from28 hospitals in the greater metropolitan area of Toronto, Canada. The geographically restricted sample was dictated by cost considerations. Although this area contains 57 hospitals, we limited our sample to the 31 hospital sites that use at least 100 volunteers, have at least one paid staff responsible for volunteer administration, and have data available on their volunteer programs. For this research, the sample of CEOs we wish to survey is 29; this number reflects the case of multiple hospital sites being run by a single CEO: Because we conducted only one interview with the CEO in charge, we ended up with a sample size of 29. Due to its protocol, one hospital was unable to allow us to access required data and was therefore not included in the data set. Our sample was therefore reduced to 28. We sent a letter inviting the hospital CEOs to participate in this study. We then telephoned to schedule a meeting. Face-to-face interviews with CEOs lasted 30 to 45 min. Unfortunately, the SARS (severe acute respiratory syndrome, a highly infectious virus) outbreak in 2003 led to the closing down of all major hospitals in the Toronto area for a considerable period of time during this study. Promised interviews were cancelled and could be rescheduled only after the SARS threat had subsided. These interviews were conducted either in person or by phone and lasted between 35 and 40 min. In a few cases when the CEO was not available, generally due to the SARS epidemic, we interviewed an individual designated by the CEO. Although designates’ titles varied from chief financial officer to vice president of human resources, these individuals usually worked closely with their CEOs, were involved in key decision making, and were often responsible for overseeing volunteer resources in an administrative role. Hence, they were perfectly suitable proxies for the CEOs. We refer to all our interviewees as CEOs. We used a structured questionnaire, with several open-ended questions. The questions covered a variety of topics ranging from overall satisfaction with the volunteer program to the trade-off of donations of time for monetary donations, to detailed and specific questions on how CEOs made budget decisions for their hospitals’ volunteer programs. In many of the interviews, the open-ended questions generated detailed responses and thereby provided a richer understanding of the informational basis for decision making. All interviewees were assured of confidentiality; the quantitative findings that follow are reported in the aggregate, and quotes are not identified by name or hospital affiliation. FINDINGS The hospitals in our sample had an average of 545 beds each. Nearly two thirds of the hospitals described themselves as acute and general hospitals; the rest categorized themselves as providing long-term rehabilitation or psychiatric care. Due to recent mergers, some hospitals in our sample operated at more than one location, sometimes with separate volunteer programs and distinct sets of volunteers. An average of 700 individuals volunteered at each site, with an average contribution of more than 71,000 hr per hospital per year in 2002. Our findings are organized as follows: In subsection A, we examine how CEOs obtain information on the volunteer programs. We assume that such information will eventually affect the allocation of resources for volunteer programs and their perception of the contribution of their volunteers. Second, in subsection B we carry out a content analysis of mission statements. In subsection C, we estimate the effect of various factors on the extent to which Table 1. Factors Influencing Demand for Hospital Volunteers Note. MRP = marginal rate of productivity. hospitals use volunteer labor. This is based on a simple regression, where the dependent variable is the number of volunteer hours utilized by the hospital. A. THE VIEW FROM THE TOP: CEO PERCEPTIONS OF VOLUNTEER DEMAND AND SUPPLY Because CEOs have the final word on resource allocation, we were interested in assessing how CEOs receive information on their volunteer resources and thereby make decisions about volunteer labor. In particular, we wished to measure how informed the CEOs were about the contribution made by service volunteers in their hospitals—independent of any occasional interaction with volunteers. We asked the CEOs what reports they received on the activities and the performance of their volunteers and who provided them with these reports. The reports varied by hospital and ranged from reports made by directors of volunteer programs reporting on hours provided by volunteers to more general information gathered through requests for budgets and reports of fund-raising or honoring volunteers. We grouped our results into (a) quantitative reports, such as volunteer hours donated and/or the number of volunteers and programs, and (b) qualitative reports, which included receiving information at events, description of volunteer services, training or recruitment sessions, and dealing with problems or issues arising. Two of the CEOs received only quantitative reports, and three received only qualitative reports. The other 23 CEOs received both qualitative and quantitative reports. In addition, all CEOs received informal feedback, such as comments from staff, patients, and families and unsolicited letters. This suggests that the majority of CEOs had good reporting mechanisms in place. Without exception, all CEOs believed that their volunteer programs are cost-effective. It was their view that indirect and direct costs such as those incurred in recruiting, training, and managing volunteers are easily outweighed by the benefits delivered by the volunteers. One comment stating that “four hundred volunteers managed by a staff of three is a good management ratio” was indicative of the general tenor of the comments made by the CEOs. It was not evident that any of the CEOs had performed any explicit cost benefit analysis of the use of volunteers, though they may have had discussions with volunteer managers regarding their budgets. In fact, the CEOs tended to view these budgets as measures of the costs of having a volunteer program. Nevertheless, CEOs generally saw volunteers as providing more than labor at a low cost. They also viewed volunteers as playing a public relations role and as highly important links to their communities. Previous research has indicated that many of the hospitals in the Toronto area would like to increase their volunteer base but find it difficult to do so due to the lack of resources available to manage new volunteers (Handy & Srinivasan, 2004). In view of this, we asked the CEOs of the hospitals what constraints they face in expanding their volunteer base. The majority (slightly more than 50%) indicated the lack of resources to be able to deal with larger number of volunteers. These included resources to recruit, supervise, and train volunteers as well as physical resources such as office space and other facilities. A little less than a third of the CEOs indicated that the demand for certain positions could not be met by the available volunteers and that as a result, they were not able to expand volunteer services. Only a few of them stated that the supply of volunteers was limited, indicating that this was one of the major reasons for their inability to expand. Others, especially those in teaching hospitals (which might be viewed by volunteers as being more prestigious), had an excess of volunteers. Other constraints indicated by the CEOs included potential labor issues with existing union contracts and retaining a staff-to-volunteer ratio balance. More than 85% of the CEOs recognized that over the years there have been significant changes in the nature of volunteers and volunteering. The demographics are changing (such as more students, more males, and a greater cultural diversity), and the turnover rate is increasing. We asked the CEOs to share the major challenges that the hospital faces regarding their volunteer programs. Nearly three quarters suggested that their hospitals face a lack of resources and structure to support their volunteer programs adequately. Two thirds suggested that recruitment of volunteers to fill certain types of positions is becoming difficult, as volunteers are either unable or unwilling to take on specified tasks. This is exacerbated by the competition from other institutions for volunteers. Short-term volunteers were seen as a drain on resources (in terms of recruitment and training), and one hospital CEO suggested that the acceptance of such volunteers should be reconsidered. Consider how satisfied CEOs are with their volunteer resources. High satisfaction will, ceteris paribus, imply that CEOs will want to expand their volunteer base. We asked CEOs to rate their satisfaction with the volunteer contribution on a scale of 1 to 10 (1 = lowest satisfaction, 10=highest satisfaction) The responses averaged 8.7, with a standard deviation of 1.3; this level of CEO satisfaction suggests that CEOs are receiving good reports from the surveys they conduct and the informal feedback they receive from patients and families, as well as their own observations. B. DEMAND FOR VOLUNTEER LABOR We examined each hospital’s mission or vision statement or statement of goals, available on its Web site, to see whether it included in the provision of health care the kinds of services generally produced with volunteer labor input. All 28 hospital mission statements referred to having a goal to provide excellence in health care that they saw as comprising more than services provided by medical professionals. They all emphasized the holistic nature of health care by stressing the need to ensure that health care included compassionate care, spiritual well-being, and so forth. In different ways, they stressed that health care was not simply provided by doctors and nurses but was a function of a whole “team” of workers. Some explicitly mentioned volunteers in the provision of health care services, while others alluded to the team but did not specify the team members. That all 28 mission statements directly or indirectly related quality of care with including services provided by volunteers did not enable us to discriminate between the hospital using the mission statement as an explanatory variable directly. More detailed content analysis of the mission statements showed fewer than half (12 of 28) of the hospitals we studied mentioned volunteers or volunteer programs explicitly in their mission statements. A few made it a point to recognize the contribution of volunteers and pledged to make it an integral part of their health care provision. However, most hospitals (25 of 28) mentioned volunteers and the need for volunteers in achieving their mission in other printed and electronic literatures. Due to the enormous variability in how volunteers were mentioned, the nature of mission statements, and published and electronic literatures available from the various institutions, it was, therefore, not possible to construct a meaningful quantitative index for this measure to include in our statistical analysis. We also asked CEOs how much of a money donation they were willing to trade off for an hour of volunteer time. This was not an easy question for CEOs to ponder because it is a question “outside the box.” To help respondents, we asked whether they were willing to accept $5.00 or an hour of volunteer time. We raised the dollar amount in increments of 5 until the CEO chose a monetary donation over 1 hr of volunteer time. Several CEOs still found this question difficult and either declined to answer or gave very high values (exceeding $1,000) that we omitted in our calculations as “protest” answers. Most of the other values ranged from $15 to $50, with an average of $25.90. This somewhat overestimates the value of volunteer time, which has been estimated to have an average replacement value of $17.57 for hospitals (Handy& Srinivasan, 2004). As CEOs could have easily substituted volunteer labor (with paid labor or other inputs) using the money donation received, the amount indicated by the CEO is a reasonable proxy for how the CEOs valued the productivity of volunteer labor as an input into the production of health care in their hospital. C. DETERMINANTS OF DEMAND FOR VOLUNTEERS From the findings above and our discussion on the determinants of volunteer demand, we model the demand for volunteer hours to be a function of five independent variables: 1. cost/hour: costs per volunteer hour (total costs divided by the number of volunteer hours); 2. trade-off value: the productivity of 1 hr of volunteer labor for the trade-off dollar figure for 1 hr of volunteer labor, as indicated by the CEO; 3. CEO satisfaction: CEO satisfaction (on a scale of 1-10, 1 = not satisfied with volunteers and 10 = highly satisfied with volunteers); 4. beds: number of beds at the hospital; and 5. union: the existence of a labor union (a dummy variable with 0 and 1, where 1 indicates the existence of a labor contract that constrains volunteer hours and 0 indicates no labor contract or one that does constrain volunteer labor). As mentioned above, we used only five of the six variables suggested in Table 1; we omitted mission statements because all 28 mission statements spoke of providing services that are generally done by volunteers. In addition, as a result of the difficulty in quantifying the appearance of “volunteers” in the mission statements due to the extreme heterogeneity in the types and nature of these statements, we could not include it in our analysis. Our data of 28 observations represent 15,284 hospital beds, the combined size of the hospitals that use an aggregate of 2,003,292 volunteer hours per year, that is, an average of more than 71,000 hr per year per hospital. Using the data collected on Variables 1-5 above, we ran a linear regression to estimate the effects of these variables on the demand for volunteer hours. The linear regression produces an R2 of 0.73. The coefficients are all statistically significant at p less than .05 levels with the exception of CEO satisfaction, as seen in Table 2. In the linear regression analysis, we find that four of the five determinants of demand for volunteers are significant, and that three of these are significant in the expected direction.11 It is important to note, however, that although we describe our regression as a demand function, our analysis must be viewed as preliminary because what is actually observed is the equilibrium result of the interplay of demand and supply. In other words, there exists what is known in econometrics as an identification problem. Nonetheless, we feel that our interpretation of the regression as a demand function is reasonable because such a large proportion of the hospitals felt that there was no shortage of potential volunteers. The first determinant of the demand for volunteer labor that we analyze is the cost per volunteer hour that is incurred by the hospital. This includes all the various costs of recruiting, screening, training, managing, and retaining. As expected, we find that the cost of volunteer labor has a negative effect on the quantity of volunteers demanded. And because the marginal cost per volunteer hour is small, this highly significant negative relationship between volunteer hours and the cost per volunteer hour suggests that the demand for volunteer labor is very sensitive to costs: The demand curve for volunteer labor is the traditional downward sloping curve. As expected, we find a direct and significant relationship between the volunteer hours utilized and the trade-off value given by the CEO for an hour of volunteer labor.12 This suggests that there is greater demand for volunteer labor in those hospitals where the CEO views volunteer labor as more productive. With respect to CEO satisfaction, we found a positive effect, as expected, but this was not statistically significant. This may, perhaps, be explained by the small amount of variability that is observed in this variable. As expected, we find that the scale effect on the number of volunteer hours used, that is, the number of beds in a hospital, has a positive and significant effect on the hospital’s demand for volunteer hours. The existence of a constraining labor contract significantly influenced the use of volunteer hours, but in the opposite direction than expected. We find that the existence of labor contracts does not reduce the demand for volunteer hours. This result suggests that it may be necessary to look more closely at the relations between management and labor and not simply at the existence of a labor contract. The nature of the relationship will depend on the way labor contracts can impinge on the use of volunteer labor. It is possible that if there is a clear demarcation on what volunteer labor can and cannot do with respect to paid labor, then the existence of the labor contractmay not impinge negatively on the demand for volunteer hours within this delineated work domain. Rather than deter the use of volunteer labor, labor contracts, which remove uncertainty, may be conducive to its use. It is also likely that hospitals with labor contracts work harder to utilize volunteer labor, as they pay higher wages to unionized labor.13 The significant positive correlation may also suggest that hospitals that have made peace with existing labor unions regarding volunteer labor and their presence can increase volunteer labor without being afraid of creating tensions within hospitals. And those hospitals without labor contracts may be more careful in increasing their demand for volunteer labor with the fear of creating problems among their paid staff and creation of unions. Given that the majority of the hospitals in our sample have labor contracts in place, there exists a culture of labor contracts in which the hospitals, employed workers, and volunteers have learned to coexist. Table 2. Regression Coefficients VOLHOURS = volunteer hours. CONCLUSIONS Many scholars have decried the paucity of research on the demand side for volunteer labor. This article takes a first, albeit modest, step in this direction by way of an empirical examination of the determinants of the demand for volunteer labor. The approach taken uses the economics literature on the demand for labor and the conventional wisdom regarding volunteers. In addition, the existing literature on the supply of volunteer labor is used to derive insights on how to proceed with the demand side of the picture. The study offers a rudimentary demand function for volunteer labor from an organizational perspective. It uses the perspective of CEOs in 28 nonprofit hospitals in the Toronto area, data on costs of volunteer hours, and organizational constraints to derive demand for volunteer labor. Recognizing the near impossibility of specifying an objective function or a production function for a nonprofit hospital, the model nonetheless identifies some of the key factors that are likely to influence the demand for volunteer labor. Because CEOs make resource allocation decisions within hospitals, we assume that the CEO’s perspectives would influence the demand for volunteer hours based on his or her perception of the value and productivity of volunteer labor. After isolating several factors likely to influence the use of volunteer labor in a hospital environment, we focused on five independent variables and ran a linear regression of these variables on volunteer hours. This regression explains 73% of the variation for the demand for volunteer hours. Our findings provide some support to the expected direction of the effects of costs, output, and productivity on the use of volunteer hours. Our findings indicate that the use of volunteer labor by hospitals is negatively related to the costs per volunteer hour and positively related to measures of productivity and output. We also found confirmation for the notion that organizational constraints matter. It is also interesting to note that the relationship between the presence of a union and the use of volunteer labor was significant, but not in the expected direction. Whereas our original view was that the presence of a labor union deters the use of volunteers, the opposite appears to be true. Further work on this issue is needed, in which information on the nature of the working relationship between hospital management and the labor union is explicitly examined. Due to the limitations of the size of our sample, further research on a larger number of homogeneous organizations and in different sub sectors is required. The elusive problem of specifying objective functions for organizations must be confronted to help specify a theoretically based demand function for volunteer labor. In the case of hospitals, patients (consumers) cannot choose the services of volunteers without consuming other services provided by paid staff. These services are bundled with other services, and we had to make simplifying assumptions on output measures. Future research could be based on organizations with less complex objective functions. In summary, the results appear to point to the existence of a downward sloping demand curve for volunteer labor: The demand for volunteer labor is not infinite. This has theoretical as well as practical policy implications. Policies that promote volunteering and thereby increase the supply of volunteers do not necessarily help nonprofits lower their costs (by using larger available amounts of volunteer labor). Indeed, to the extent that the use of volunteers is demand rather than supply constrained, policies should focus on facilitating the incorporation of volunteer labor rather than increasing its supply. Make-work projects designed to absorb the available supply of volunteer labor will distort the goals and efficiency of organizations. Funding arrangements for organizations using a large number of volunteers should be targeted to help organizations use professional management techniques. This will help to reduce costs, increase efficiency, and reduce liabilities that may be attendant with the use of volunteer labor. In addition, effective management will help organizations use available volunteer labor in meaningful ways that will provide benefits not only to organizations utilizing them but also to volunteers. Finally, as mentioned above, although our article focuses on the demand side of volunteer labor and delineates the determinants, the supply of volunteers and its interaction with demand remain important issues for research. Notes 1. Auxiliaries have been disbanded in many hospitals or merged with volunteer departments. In some instances where they coexist with volunteer departments tensions exist between the auxiliaries and volunteer departments (Atkinson, 1997; Handy & Srinivasan, 2004). 2. It seems very likely that beyond a certain number of volunteers, the cost associated with volunteers is convex. 3. The assumption is that the demand for volunteer labor exceeds the supply of volunteer labor at a wage rate of zero. 4. It is difficult, if not impossible, to distill even this seemingly simple objective function into an operative measure. How does one measure health care? How is health care aggregated across individuals? Which group of individuals enters a specific hospital’s objective function? 5. MRPi / pi = 1 for all inputs i of production in the long run. 6. Of course, measuring productivity is complex, even in well-defined work such as, for example, a volunteer receptionist: the number of people greeted, the diameter of the smile produced by the volunteer, and the amount of eye contact all enter into productivity. 7. A comprehensive measure of productivity should also include the perceptions of patients and their families. However, due to the ethical protocols in hospitals, eliciting such information from patients and their families is not permissible. 8. Retrieved June 17, 2003, from Lake Ridge Health Corporation, www.lakeridgehealth. on.ca/get/vision.htm. 9. Many other types of organizations have restrictions on the use of volunteers for work done by paid staff; for example, labor contracts by the Canadian Union of Public Employees (CUPE) with school boards specify the limitations of the use of parent volunteers in schools and annually review the use of volunteers (retrieved June 13, 2004, from http://www.sd61.bc.ca/hr/pdf/ volunteer. PDF). 10. Section 11.02 of the Central Hospital Agreement—a result of a grievance filed by CUPE (CUPE, 2000; Handy, Mound, & Vaccaro, 2004). 11. Furthermore, a bi variate correlation shows no significant correlation among the five independent variables. Multicollinearty was not an issue. 12.Afew CEOs were unable to answer this question; we used the mean as a substitute for these missing values in our regression. 13. We thank one of our anonymous referees for suggesting this point. References Andreoni, J. (1990). Impure altruism and donations to public goods: A theory of warm-glow giving. Economic Journal, 100, 464-477. Atkinson, A. (1997). How, now, health care auxiliaries? Canadian Journal of Volunteer Resources Management, 4(2), 2-4. Canadian Union of Public Employees (CUPE). (2000, March). CUPE: The facts on volunteers. Toronto: Author, Ontario Research Branch. Carlin, P. S. (2001). Evidence of volunteer labor supply of married women. Southern Economic Journal,64, 801-824. Davis Smith, J. (1998). The 1997 national survey of volunteering. London: National Centre for Volunteering. Emanuele, R. (1996). Is there a downward sloping curve demand curve for volunteer labor? Annals of Public and Cooperative Economics, 67, 193-208. Freeman, R. (1997).Working for nothing: The supply of volunteer labor. Journal of Labor Economics, 15, 140-166. Govekar, P. L.,& Govekar,M.A. (2002). Using economic theory and research to better understand volunteer behavior. Nonprofit Management and Leadership, 13(1), 33-48. Hall, M., McKeown, L., & Roberts, K. (2001). Caring Canadians, involved Canadians: Highlights from the 2000 national survey of giving, volunteering and participating (Catalogue no. 71-542-XPE). Ottawa: Statistics Canada. Handy, F., Mound, R.,& Vaccaro, L. M. (2004). Promising practices for volunteer administration in hospitals. Toronto: Canadian Centre for Philanthropy. Handy, F.,& Srinivasan ,N. (2004). Improving quality while reducing costs? An economic evaluation of the net benefits of hospital volunteers. Nonprofit and Voluntary Sector Quarterly, 33, 28-54. Handy, F.,& Webb, N. (2003).A theoretical model of the effects of public funding on savings decisions by nonprofit service providers. Annals of Public and Cooperative Economics, 74(2), 1-22. Hodgkinson, V. A.,& Weitzman, M. S. (1992). Giving and volunteering in the United States. Washington ,DC: Independent Sector. Independent Sector. (2001). Giving and volunteering in the United States. Retrieved from http:// www.independentsector.org/programs/research/ Karom Group of Companies. (2001). Report on volunteerism within member hospitals of the Ontario Hospital Association. Toronto, Canada: Author. Katz, E., & Rosenberg, J. (2004). An economic theory of volunteering (Working paper). DeKalb: University of Northern Illinois. LaPerriere, B. (1998). Volunteerism in the Canadian health sector. Ottawa: Volunteer Canada. Macduff, N. (1997). Solving the hazards of unions and volunteer relations in government organizations. The Journal of Volunteer Administration, 14(1), 34-39. Menchik, P. L.,& Weisbrod, B. A. (1987).Volunteer labor supply. Journal of Public Economics, 32(2), 159-183. Proteau, L., &Wolff, F. C. (2004). Volunteer work and labor supply. Unpublished paper, Faculté des Sciences Economiques, Université de Nantes, France. Schiff, J. (1990). Charitable giving and government policy: An economic analysis. New York: Greenwood. Segal, L. M., &Weisbrod, B. A. (2002). Volunteer labor sorting across industries. Journal of Policy Analysis and Management, 21(3), 427-447. Smith, D. H. (1994). Determinants of voluntary association participation and volunteering :A literature review. Nonprofit and Voluntary Sector Quarterly, 23, 243-263. Statistics Canada. (2004a). Cornerstones of community: Highlights of the national survey of nonprofit and voluntary organizations (Catalogue no. 61-533-XPE). Ottawa: Author. Statistics Canada. (2004b). Non-profit institutions and volunteering: Economic contribution. Ottawa: Author. Retrieved from http://www.statcan.ca/Daily/English/040920/d040920c.htm Steinberg, R. (1990). Labor economics and the nonprofit sector: A literature review. Nonprofit and Voluntary Sector Quarterly, 19(2), 151-169. Unger, L. S. (1991). Altruism as a motivation to volunteer. Journal of Economic Psychology, 12(1), 71- 100. Vaillancourt, F. (1994). To volunteer or not: Canada, 1987. Canadian Journal of Economics, 27, 813- 825. Vaillancourt, F.,& Payette, M. (1986). The supply of volunteer work: The case of Canada. Journal of Voluntary Action Research, 15(4), 45-56. Van Dijk, J., & Boin, R. (1993). Volunteer labor supply in the Netherlands. De Economist, 141, 402- 418. Wolff, N., Weisbrod, B., & Bird, E. J. (1993). The supply of volunteer labor: The case of hospitals. Nonprofit Management and Leadership, 4(1), 23-45. Zahnd, L. (1997). Volunteer staff relationships in a unionized environment. Journal of Volunteer Resources Management, 6(1), 8-9.

Journal of Social Issues, Vol. 58, No. 3, 2002, pp. 447--467

One part of this study would be very helpful. This study looks at the relationship between different variables such as race and religion and their relationship to volunteerism.
Penner, Louis. "Dispositional and Organizational Influences on Sustained Volunteerism: An Interactionist Perspective - Penner - 2002 - Journal of Social Issues." //Wiley Online Library//. Web. 22 Feb. 2011. .

Louis A. Penner∗ University of South Florida Community service often involves sustained prosocial actions by individuals. This article focuses on one kind of such actions, volunteerism. Volunteerism involves long-term, planned, prosocial behaviors that benefit strangers, and usually occur in an organizational setting. A selective review of the literature on the correlates of volunteerism is presented. One part of the review concerns the relationship between dispositional variables and volunteerism; it includes newdata from an online survey that show significant relationships among personality traits, religiosity, and volunteer activities. The other part concerns how organizational variables, alone and in combination with dispositional variables, are related to volunteerism. A theoretical model of the causes of sustained volunteerism is presented and the practical implications of this model are discussed. Among social psychologists, there is a long history of interest in when and why people act prosocially (Schroeder, Penner, Dovidio, & Piliavin, 1995). Until relatively recently, research on prosocial behavior focused primarily on a very specific kind of prosocial action—bystanders intervening to provide immediate and short-term help to a physically distressed stranger. In the last few years, however, more attention has been given to prosocial behaviors that continue for an extended period of time—sustained prosocial actions. There are a number of different kinds of behaviors that might be classified as sustained prosocial actions (e.g., working as a firefighter, caring for a chronically-ill loved one), but this article is primarily concerned with volunteerism. Because volunteerism can mean different things to different people, in the first section of the article I define the term, discuss its most salient attributes, and consider the differences and similarities between volunteerism and other kinds ∗Correspondence concerning this article should be sent to Louis A. Penner, Department of Psychology, University of South Florida, Tampa, Florida 33620 [e-mail: penner@chuma1.cas.usf.edu]. 447 C 2002 The Society for the Psychological Study of Social Issues 448 Penner of prosocial behaviors. In the next section, I use an interactionist perspective or framework to discuss the variables associated with volunteerism. I begin this section with a selective review of published research on the dispositional correlates of volunteerism and present some new data from an on-line survey that provide information about how dispositional variables are related to volunteerism. Then, I turn to the question of how organizational variables, alone and in combination with dispositional variables, are related to volunteerism. In the final portion of the article, I present a conceptual model of direct and indirect influences on sustained volunteerism and discuss some practical implications of the model. Volunteerism Volunteerism can be defined as long-term, planned, prosocial behaviors that benefit strangers and occur within an organizational setting. Based on this definition, volunteerism has four salient attributes: longevity, planfulness, nonobligatory helping, and an organizational context. Each of these is briefly discussed below. Longevity. Volunteering is usually a relatively long-term behavior. For example, a recent national survey of volunteerism in the United States (Independent Sector, 1999) found that almost 50 percent of the people who volunteer do so on a regular rather than a one-time basis. Another recent survey of volunteers found that more than 90 percent of them wanted to engage in long-term volunteer activities (VolunteerMatch, personal communication, September 15, 2001). And longitudinal studies of volunteers have found that once people begin to work regularly as a volunteer, a large percentage of them continue this activity for several years (Omoto & Snyder, 1995; Penner & Finkelstein, 1998; Penner & Fritzsche, 1993). Planfulness. Volunteering is typically a thoughtful and planned action. On first inspection, data from national surveys of volunteers would seem to contradict this statement. For example, in its national survey the Independent Sector (1999) found that about 90 percent of the people asked to volunteer agree to do so. However, it seems unlikely that requests to become a volunteer are directed at a random group of people or that people impulsively agree to become a volunteer at the moment they are asked to volunteer. It seems much more probable that the targets of these requests have previously indicated some interest in becoming a volunteer and are already, for whatever reason, favorably disposed toward this activity. Further, thework of Davis et al. (1999) suggests that before people actually agree to volunteer, there is some thoughtful consideration of both the costs and benefits of engaging in this action. This decision process can be contrasted with the one that usually precedes helping in emergencies. In such instances, the helping decision is made very quickly, without much (or sometimes no) conscious thought, Sustained Volunteerism 449 and is greatly influenced by the salient characteristics of the particular situation that confronts the potential helper (Dovidio & Penner, 2001). This is not to suggest that volunteering is totally immune to situational forces. For example, in the first few days after the September 11 attacks on the World Trade Center and the Pentagon, the number of people who contacted one on-line service to volunteer for different charities almost tripled; more people offered their services as volunteers than at any other time in the service’s three year history (VolunteerMatch, personal communication, September 15 2001). But it seems reasonable to argue that these people’s behavior was still much more thoughtful, planned, and deliberate than bystander interventions in emergencies. The events that inspired the volunteering occurred at some distance from where the modal volunteer lived; and volunteering required locating the on-line service, selecting an organization that needed volunteers, and providing personal information so the organization could later contact the volunteer. Also, the events of September 11 produced increases in volunteering for all the organizations listed by this service. For example, while organizations that provided emergency services showed the largest increases, there were also substantial increases in the number of people who wanted to volunteer for organizations that provided services for animals, children, gays and lesbians, seniors, and numerous other target groups (VolunteerMatch, personal communication, September 15 2001). This suggests that many of these people had thought about volunteering for a certain kind of organization well before the day of the attacks. Nonobligatory helping. Because the recipients of a volunteer’s beneficence are either strangers or an organization that serves these individuals, the volunteer is not motivated by a sense of personal obligation to a particular person (Omoto & Snyder, 1995). Omoto and Snyder characterize this kind of prosocial behavior as “nonobligated helping.” By contrast, when helping is directed at a close friend or relative, it typically results from a prior, personal, and reciprocal relationship between the helper and the recipient; thus, there is some implicit or explicit personal obligation to help (Dovidio & Penner, 2001; Omoto & Snyder, 1995; Penner & Finkelstein, 1998). Organizational context. Finally, volunteerism is far more likely than other kinds of helping to take place within an organizational setting. There are certainly individuals who, on their own, engage in sustained, nonobligated helping of virtual strangers (see Colby & Damon, 1992). However, most volunteers (perhaps as high as 85 percent) work as part of an organization (Independent Sector, 1999). Thus, organizational variables are far more important in volunteerism than in one-to-one, interpersonal kinds of helping. My research on variables that might affect volunteerism has been guided by this conceptualization of its most salient attributes. For example, the fact that volunteerism involves long-term, planned helping led me to devote substantial attention 450 Penner to dispositional variables and their relationship to volunteerism. This is because dispositional variables are more likely to manifest themselves in enduring behaviors than in transitory ones, such as bystander interventions. Similarly, the fact that volunteerism is likely to occur in an organizational setting has led me (and other researchers) to pay attention to the organizational variables that might influence it. A selective review of research on these two classes of variables and volunteerism follows. In the interest of clarity, the dispositional and the organizational correlates of volunteerism are discussed separately. However, my approach to volunteerism is firmly imbedded in an interactionist perspective. Specifically, two assumptions are made about the variables discussed below. First, neither dispositional nor organizational variables can, by themselves, provide a full explanation of why people initially decide to volunteer and then continue to volunteer over an extended period of time. Second, the two classes of variables affect one another and interact to affect volunteerism. That is, the influence of many organizational variables on volunteerism may be moderated and/or mediated by dispositional variables and vice versa. This point is discussed in more detail shortly. Dispositional Variables and Volunteerism In the present context, “dispositional variables” will be used as a generic term for several different enduring attributes of individuals. These include things such as their personal beliefs and values, personality traits, and motives. The notion that dispositional variables, especially personality traits, are related to prosocial behaviors has not always enjoyed wide acceptance among helping researchers. Indeed, in one of the first comprehensive monographs on helping, Piliavin,Dovidio, Gaertner, and Clark (1981) concluded that the search for a prosocial personality had been “futile.” One reason for this conclusion was that, at that time, most helping studies concerned bystander interventions in emergencies. In such circumstances situational demands are often so strong that they may suppress the influence of dispositional variables on helping decisions (Epstein, 1979). However, there may have been another reason for the dismal findings concerning the personality correlates of prosocial actions. Most researchers did not, in fact, search for the “prosocial personality”; rather they studied how a very specific personality trait related to a very specific kind of helping. When significant findings were obtained, attempts to “replicate” them often involved a quite different kind of helping. Most of these replications failed, but not because personality is unrelated to helping, but rather because the salient characteristics of the criterion measure had changed. Thus, perhaps what the null results really showed is that one relatively specific personality trait is unlikely to be related to a wide range of helping behaviors (Penner, Escarraz, & Ellis, 1983). This line of reasoning led my students and me to search for the personality characteristics that form the core of a “prosocial personality” (Penner, Fritzsche, Craiger, & Freifeld, 1995). Sustained Volunteerism 451 The Prosocial Personality Penner et al. (1995) began their search for the prosocial personality with the identification of personality traits that had been found to correlate with some kind of prosocial behavior in at least two published studies. Then, the list of traits was reduced by excluding a trait if: (1) there was no coherent theoretical explanation of why it correlated with prosocial or helpful actions; or (2) it did not correlate with other, independent measures of prosocial tendencies. Following this, a factor analysis was performed on the remaining measures. Both quantitative and theoretical considerations led to a two-factor solution. The first factor was called Other-oriented Empathy; it appears to primarily concern prosocial thoughts and feelings. People who score high on this factor are empathetic and feel responsibility and concern for the welfare of others. The second factor was called Helpfulness; it appears to concern prosocial actions. High scorers on this factor have a history of being helpful and are unlikely to experience self-oriented discomfort in response to others’ distress (Penner et al., 1995). In passing, it should be mentioned that this empirically derived “description” of the prosocial personality is quite similar to Oliner and Oliner’s (1988) description of the personality traits of Christians who rescued Jews during the Holocaust and to Colby and Damon’s (1992) summary of the personal characteristics of the 23 lifelong altruists they studied. The instrument that measures the prosocial personality is called the Prosocial Personality Battery (PSB). Scores on the two factors of the PSB correlate from .25 to .50 depending on the sample (Penner et al., 1995; Rioux & Penner, 2001). Other-oriented Empathy and Helpfulness appear to be not only empirically distinct, but conceptually distinct as well. For example, scores on the Other-oriented Empathy dimension strongly correlate with measures of personality attributes such as agreeableness and nurturance, but scores on the Helpfulness dimension do not. Conversely, scores on the Helpfulness dimension correlate strongly with measures of dominance and assertiveness, but scores on the Other-oriented Empathy dimension do not (Penner et al., 1995). Also, whereas scores on the Other-oriented Empathy dimension correlate with affective and cognitive responses to distress in another person, scores on the Helpfulness dimension do not (Penner & Fritzsche, 1993). Despite these differences, scores on both dimensions of the prosocial personality do correlate with prosocial behaviors. Among the behaviors that have been found to correlate with one or both of them are: speed of response in simulated emergencies, the frequency of mundane, everyday acts of helping over a month, frequency of helping co-workers, willingness to mentor co-workers, and willingness to serve as an organ donor (see Allen, 1999; Cicognani, 1999; Dovidio & Penner, 2001; Borman, Penner, Allen, & Motowidlo, 2001). Here, however, the behavior of primary interest is volunteerism. 452 Penner Volunteerism. The research on the prosocial personality strongly suggests that its two dimensions are related to various aspects of volunteer behavior. For example, Penner and Fritzsche (1993) found that scores on both dimensions distinguished volunteers at a homeless shelter from a matched group of non-volunteers. Further, within the volunteer sample, the short-term and long-term volunteers differed in their scores on both dimensions. In another study, Penner and Finkelstein (1998) administered the PSB to volunteers at an AIDS service organization. Five and eleven months later they measured the level of general volunteer activities and the amount of time the volunteers spent with someone who was HIV-positive or had AIDS. Among male volunteers, Other-oriented Empathy (but not Helpfulness) correlated significantly with subsequent levels of general volunteer activities and the amount of time the volunteers personally spent with someone who was HIV positive or had AIDS. Additionally, in this and other studies scores on the Helpfulness dimension significantly correlated with the number of service organizations for which volunteers worked (Little, 1994; Sibicky, Mader, Redshaw, & Cheadle, 1994; Penner & Fritzsche, 1993). These studies contained relatively small samples of volunteers who worked in a restricted number of service organizations. More recently, I was able to use the Internet to collect data from a much larger sample of volunteers working in a wide variety of service organizations in the United States. In May of 1999 USA Weekend, a Sunday supplement magazine that is carried by 560 American newspapers, contained an article about “altruism” (Paul, 1999). A portion of the article discussed research on the prosocial personality and invited readers to visit the supplement’s website, “USA WEEKEND ONLINE,” and complete a “test” that measured how prosocial they were. Readers who went to the website were linked to a copy of the PSB, along with some questions about their demographic characteristics (e.g., age, income, education, gender, ethnicity) and religious beliefs (whether they were affiliated with a specific religion, and howreligious they were). They were also asked if they had they volunteered in the last year. If they had, they provided information about: the number of charities for which they volunteered, the nature of their primary charity, how much time they spent working for that organization, and their tenure as a volunteer for that group. In order to insure anonymity, responses to the questionnaire were sent to a file on the USA Weekend server and then the data, stripped of all identifiers, were forwarded to me in a spreadsheet format. More than 1100 people completed the survey. About 76 percent of them reported having worked as a volunteer during the previous 12 months. (These people were classified as “active volunteers.”) The respondents were overwhelmingly of European ancestry (90 percent) and predominantly female (77 percent); about 48 percent had completed at least some college and the same percentage had a total family income of $40,000 or more. About 60 percent self-identified as Protestant or Catholic; another 25 percent said they belonged to other religions; and the Sustained Volunteerism 453 remaining 15 percent said they were not members of any organized religion. Overall, 45 percent of the respondents described themselves as “very” or “extremely” religious (henceforth, this variable will be called religiosity). Seventy-six percent of the respondents indicated that they had volunteered in the last year. (This was substantially higher than the 55 percent volunteer rate found in Independent Sector’s [1999] national survey of volunteering in the United States.) It would be hard to claim that this was a representative sample of American volunteers and non-volunteers. In addition to the obvious biases introduced by who reads USA Weekend and who has access to the Internet, there was probably a large self-selection bias in who opted to respond to the survey. For example, it seems reasonable to argue that the over-representation of active volunteers among the total respondents represents some sort of self-affirmation of their beliefs that they were prosocial individuals. Thus, while the demographic profile of volunteers obtained in this study (i.e., predominately, wealthy, well-educated women of European ancestry) was quite similar to the profile found in another recent survey of volunteers (VolunteerMatch, personal communication, September 15, 2001), these data cannot be used to provide population estimates of the incidence of volunteerism in United States or the proportion of volunteers with a particular demographic attribute. However, a number of survey researchers now argue that even though a sample may be biased from a random sampling perspective, the patterns of correlations obtained from such a sample usually closely approximate those obtained from an unbiased sample (see Brehm, 1993; Dillman, 2000; Krosnick, 1999). Therefore the data from this on-line survey can be used to study the dispositional correlates of volunteerism. The data analysis was conducted in two stages. In the first, the key question was: What variables distinguished those respondents who were active volunteers from those who were not? In the second stage, only the active volunteers were considered and the key question was: What variables distinguished the more active volunteers from the less active ones? The latter set of analyses presented a much more difficult predictive task because most of the active volunteers had a long history of substantial involvement in volunteer activities and, thus, there was a substantial restriction of range with regard to the criterion variables (i.e., number of organizations, etc.). For example, about 68 percent of the volunteers reported that they worked for multiple charities; more than 70 percent said they spent at least a fewhours every week as a volunteer for their primary charity; and a majority (about 68 percent) reported they had worked at the charity for at least two years. Nonetheless, the latter analyses permitted a much more fine-grained examination of the correlates of volunteerism. That is, they enabled us to identify the variables associated with different levels of activity among a group of active volunteers. In the interest of brevity and clarity, only a small portion of the findings from this survey will be discussed and specific results will be presented only when they 454 Penner speak directly to the dispositional correlates of altruism.1 Because of the extremely large sample size, alphawas set at .01; all statistics reported belowwere significant at that level or beyond. Volunteers versus Non-volunteers Active volunteers and non-volunteers did not differ with respect to age, education, gender, or income. However, as expected, volunteers scored significantly higher than non-volunteers on both the Other-oriented Empathy and Helpfulness dimensions of the prosocial personality, ts (1084) = 7.06 and 5.75, respectively. Turning to religion, itwas found that people who belonged to an organized religion were more likely to be volunteers (80 percent) than people who did not belong (62 percent), χ2(2) = 21.29; and, relative to non-volunteers, volunteers scored higher on the religiosity measure (i.e., how religious they were), t (1084) = 7.50. To eliminate the possibility that the associations between the religion-related questions and volunteering might have been due to the individuals who volunteered for religious organizations (about 22 percent of the sample), these people were excluded and the data reanalyzed. Even with these people excluded, all the significant findings described above remained significant. Correlates of Volunteer Activities Table 1 presents the intercorrelations among the three aspects of volunteer activities—number of organizations worked for, length of service at primary charity, and amount of time spent as a volunteer at that charity—and the correlations between these measures and the demographic and dispositional variables noted above. The first thing that can be seen from this table is that the three activities were substantially intercorrelated (all rs > .40). This should be kept in mind as I discuss the correlates of each kind of activity. With regard to the demographic correlates of volunteer activities, age was significantly and positively associated with number of organizations and length of time spent working for that organization (rs=.15 and .24, respectively); education was significantly and positively associated with all three activities (rs = .17, .24, and .10, respectively); and income was significantly and positively correlated with number of organizations (r = .11). Gender (coded as a dummy variable) was not correlated with any of the volunteer activities. The religiosity measure was significantly correlated with all three measures of volunteer activities. The stronger people said their religious beliefs were, the more organizations theyworked for (r = .23), the longer their tenure as a volunteer (r = .24), and the more time they spent working as a volunteer (r = .16). (The 1A complete set of all the analyses is available from the author. Table 1. Correlates of Volunteer Activities Number of Volunteer Volunteer Other-Oriented Organizations Length Time Age Education Gender Income Religiosity Empathy Number of Organizations Volunteer Length 0.51 Volunteer Time 0.41 0.41 Age 0.15 0.24 0.05 Education 0.17 0.24 0.10 0.38 Gender 0.02 0.01 −0.02 0.03 −0.01 Income 0.11 0.07 −0.03 0.02 0.21 −0.02 Religiosity 0.23 0.24 0.16 0.06 0.07 0.03 0.02 Other-Oriented Empathy 0.24 0.16 0.11 0.10 0.13 0.14 0.00 0.17 Helpfulness 0.25 0.18 0.16 0.26 0.12 0.08 0.03 0.09 0.53 Note. n = 847; Gender was coded as: 1 = male, 2 = female. r s > .10, p < .01. r s > .12, p < .001. r s > .15, p < .0001. 456 Penner same pattern was obtained when people who said they belonged to a religion were compared to non-belongers in χ2 analyses.) These relationships remained even when respondents who worked for religious organizations were excluded from the analyses. Other-oriented Empathy was significantly and positively correlated with all three activities (rs = .24, .16, .11, respectively); and the same was true for the relationships between Helpfulness and these activities (rs = .25, .18, .16, respectively). The final question asked of the active volunteer data concerned whether religious beliefs and the prosocial personality dimensions would explain variance in volunteer activities that was not explained by demographic characteristics. To answer this question, three hierarchical multiple regressions were conducted in which each of the volunteer activities was regressed onto: age, education, gender, income, religiosity, Other-oriented Empathy, and Helpfulness. In all the regressions, the four demographic variables (i.e., age, education, etc.) were entered as a block at the first step and the three dispositional variables (i.e., religiosity, Otheroriented Empathy, etc.) were entered as a block at the second step. The critical question was whether there would be a significant change in R2 when each block was added to the equation.2 When number of volunteer organizations was regressed onto the predictor variables, the overall R2(.14) was significant, F(7, 839)=19.37, p < .001. Both the demographic and dispositional blocks of variables added significant amounts of variance accounted for to the equation. Specifically theR2 (i.e., change in R2) for the demographic variables was .045, F(4, 842)=9.94; theR2 for the dispositional variables was .094, F(3, 839)=30.41. In the regression involving length of time as a volunteer the R2(.14) was also significant, F(7, 839)=20.06. Again both the demographic and dispositional blocks of variables added significant explained variance to the equation—R2 demographic variables .085, F(4, 842)=19.45; R2 dispositional variables .058, F(3, 839)=19.07, p< .001. Finally, although the R2 for amount of time spent volunteering was much smaller than the R2s for the other two activity measures (.06), it was significant F(7, 839)=7.71, p< .001. However, only the block of dispositional variables produced a significant R2, F(3, 839)=12.75. A brief discussion of these findings would seem worthwhile. Education was positively correlated with all three measures of volunteer activities. This finding was consistent with other studies of the demographic correlates of volunteering (e.g., Piliavin & Callero, 1991; Statistics Canada, 2001). Some have speculated that the reason for this relationship is that better educated people have the kind of jobs that allow them more time to devote to their volunteer activities 2 In the interest of brevity and clarity, the analyses that included religious affiliation, a categorical variable with five levels, are not discussed in this article. These are available from the author. Sustained Volunteerism 457 (Schroeder et al., 1995). But others have suggested another explanation. For example, on the basis of data from several case studies Bellah, Madsen, Sullivan, Swidler, and Tipton (1985) argued that people from upper social economic classes (e.g., better educated people) may be more willing to volunteer because this provides them with a way to give some additional meaning to their lives. That is, they need something beyond their jobs to make them feel fulfilled. Let us nowturn to the association between religiosity and volunteerism, which was not a primary focus of this study, but the findings are interesting. Religiosity was positively associated with all three kinds of volunteer activities. This finding has also been obtained in a recent national survey of volunteers in Canada (Statistics Canada, 2001). Further, in the present study religiosity produced the strongest associations with volunteer activities. Itwould be premature to conclude from these findings that being religious is invariably positively associated with volunteerism. In fact, there is other evidence that different kinds of religious motivations and beliefs may moderate when and for whom religious people offer their services as volunteers (Jackson & Esses, 1997). However, these results do suggest that one should include some measures of religiosity in any comprehensive examination of the causes of volunteerism. The two dimensions of the prosocial personality were significantly associated with all three aspects of volunteer activities. This finding is quite consistent with earlier work by Penner and his associates on the personality correlates of volunteerism. (Although, interestingly, this is the first study in which the relationship between Other-oriented Empathy and the number of organizations for which a person volunteers was as strong as the relationship for Helpfulness; see Penner et al., 1995; Penner & Finkelstein, 1998) The findings are also consistent with the results from other studies that have examined longterm, voluntary prosocial behaviors among paid employees of large organizations (e.g., Allen, 1999; Facteau, Allen, Facteau, Bordas, & Tears, 2000; Midili & Penner, 1995; Penner, Midili,&Kegelmeyer, 1997; Negrao, 1997; Rioux&Penner, 2001; Tillman, 1998). In all of these studies, significant, positive associations have been found between the two dimensions of the prosocial personality and selfreports of sustained prosocial actions. And in two of the three studies where peer reports of prosocial behaviors were also obtained, significant positive relationships were found. Thus, the relationship between the prosocial personality and sustained prosocial actions in an organizational setting is probably not restricted to unpaid volunteers. Motives Before I turn to the research on organizational variables, I want to briefly discuss the role of motives in volunteerism. This is because the model presented in Figure 1 gives a prominent place to motives as causes of volunteerism. The 458 Penner discussion here is necessarily brief—but a much more detailed treatment of motives has been provided by Clary et al. (1998). The theoretical rationale for research on the role of motives in volunteerism comes from Snyder’s functional approach to prosocial behaviors, which focuses on the function or purpose served by such behaviors (see Clary & Snyder, 1991; Snyder, 1993). This approach is predicated on the notion that much of human behavior is motivated by specific goals or needs. Thus, if one wants to understand why a person has engaged in some behavior, one needs to identify the purpose or need served by that behavior. In the case of volunteering, people engage in this behavior, at least in part, because it serves one or more of their goals and needs. There is a substantial body of work (e.g., Clary et al., 1998; Omoto & Snyder, 1995; Penner&Finkelstein, 1998) that suggests personal motives play an important role in volunteerism. For example, in the longitudinal study described earlier, Penner and Finkelstein (1998) also measured the motives of the AIDS volunteers. They found that among the male volunteers “value expressive” motives, measured at the beginning of the study, correlated significantly with subsequent levels of both general volunteer activities and the amount of time a volunteer spent with someone who was HIV-positive, or had AIDS. Clary and Orenstein (1991) and Davis, Hall, and Meyer (2001) have obtained similar results in studies conducted in several different kinds of volunteer organizations. And Rioux and Penner (2001) have found that motives also play a significant role in long-term, voluntary prosocial behaviors among paid employees of large organizations. Organizational Variables and Volunteerism As noted earlier, once a person has made the decision to volunteer, volunteerism usually occurs in an organizational context. Thus, it is necessary to discuss the organizational variables that are most likely to influence a volunteer’s behavior in this context. A review of the theoretical and empirical literature suggests that two kinds of organizational variables should have an impact on volunteerism. They are: (1) an individual member’s perceptions of and feelings about the way he or she is treated by the organization and (2) the organization’s reputation and personnel practices. A few studies have examined how perceptions and feelings affect volunteerism. For example, Omoto and Snyder (1995) found that satisfaction with the organization was significantly associated with length of tenure as a volunteer; and Penner and Finkelstein (1998) and Davis, Hall, and Meyer (2001) found that organizational satisfaction was associated with the amount of time spent working as a volunteer. Further, Penner and Finkelstein (1998) and Grube and Piliavin (2000) both found a significant positive relationship between organizational commitment and the amount of time people reported working for a service organization. These findings are consistent with research by industrial and organizational psychologists on the correlates of sustained, voluntary prosocial actions among Sustained Volunteerism 459 paid employees of organizations. For example, Organ and Ryan (1995) conducted a large meta-analysis of the correlates of these kinds of behaviors and found that “job attitudes” (i.e., job satisfaction with the job, perceived organizational fairness, organizational commitment, and perceived leadership supportiveness) consistently correlated with self- and peer-reports of prosocial actions directed at individuals and the organizations themselves. (See also Borman et al., 2001; Midili & Penner, 1995; Rioux & Penner, 2001.) Turning to reputation and practices, I am aware of only one study that has directly addressed how an organization’s reputation affects volunteer activities. Grube and Piliavin (2000) reported that ratings of the prestige of an organization were positively associated with number of hours worked for the organization and negatively associated with intent to leave it. I am not aware of any studies that have investigated the impact of personnel practices on volunteers’ behavior, but the industrial and organizational psychology literature would suggest they are important. For example, Graham (in press) argued that if companies want to increase voluntary prosocial actions among their employees, they need to design jobs that are highly motivating and interesting, and that provide feedback to the job occupant. Skarlicki and Latham (1996) provided some direct evidence that changes in organizational practices can affect employees’ inclinations to act prosocially. They manipulated the level of organizational justice displayed by officers of a union and then subsequently measured prosocial behaviors among union members. They found that such behaviors occurred significantly more often among members whose officers had received the organizational justice training and that this relationship was directly linked to perceived organizational justice among the members. This suggests that an organization that treats its workers fairly can reasonably expect an increase in voluntary prosocial actions among its employees. This should be true whether the organization is for profit and the employees are paid or the organization is a charity and the “employees” are volunteers. Thus, volunteers who are satisfied with their job, committed to the organization, have positive affect while on the job, and believe they are being treated fairly should display a higher level of volunteer activity. Interactions Between Dispositions and Organizational Factors Although the dispositional and organizational correlates of volunteerism were presented separately, it must be reemphasized that these two classes of variables are not independent of one another. Consistent with the interactionist theme presented earlier, they influence one another and the resultant interactions between them influence sustained prosocial actions. A few examples serve to illustrate this point. Consider, first, the relationship between job attitudes and the dimensions of the prosocial personality. Midili and Penner (1995) found that paid workers who scored high on Other-oriented Empathy also reported high levels of job satisfaction, 460 Penner perceived more organizational justice, and had a more positive mood on the job. (The last finding was replicated by Rioux and Penner [2001].) This suggests that the prosocial personality may affect sustained prosocial actions both directly and indirectly, through its influence on the job-related thoughts and feelings. The relationships among motives and organizational and personality variables provide another example of why one needs to take an interactionist perspective on the causes of sustained prosocial actions. Among both volunteers and paid workers, the strength of prosocial motives is associated with job attitudes (e.g., job satisfaction, perceived organizational justice, organizational commitment) and the two dimensions of the prosocial personality (Forde, 2000; Omoto & Snyder, 1995; Penner & Finkelstein, 1998; Rioux & Penner, 2001). At this point it is impossible to decipher the causal links among these variables, but the findings underscore the point that it would be unwise to talk about the impact of motives on volunteerism independently of their relationship with the other correlates of this behavior. This position is implicitly and explicitly reflected in the model of sustained volunteerism that is discussed below. Sustained Volunteerism: A Conceptual Model Figure 1 presents a conceptual model of the causes of sustained volunteerism. Because this is a conceptual or structural model, I will not discuss how the latent variables shown in the figure would be measured or operationalized. However, measures for all the variables in the model do exist. The model is based on my own work and the work of other researchers who study volunteerism (e.g., Grube & Piliavin, 2000; Lee, Piliavin, & Call, 1999; Clary et al., 1998). It should not be viewed as a definitive statement on the causes of sustained volunteerism, but rather as a working model that will hopefully be of heuristic value to others interested in this and related kinds of sustained prosocial behaviors.3 The model is organized temporally and begins with the Decision to Volunteer, the point at which the person makes a commitment to become a volunteer. The data on volunteering in the days following the September 11 attacks strongly suggest that Situational Factors (e.g., historical events) can have an impact on a person’s Decision to Volunteer. However, the model assumes that Situational Factors are less influential causes of this decision than are the variables discussed below. Therefore, the path from Situational Factors to Decision to Volunteer is represented by a broken line. A much more potent determinant of the Decision to Volunteer is Volunteer Social Pressure, which is a potential volunteer’s subjective perceptions of how significant others feel about him/her becoming a volunteer and his/her motivation 3 Penner et al. (1997) proposed a very similar model of the causes of sustained prosocial behaviors among paid employees of organizations. Demographic Characteristics Personal Beliefs and Values Volunteer- Related Motives Organizational Attributes & Practices Relationship with the Organization Initial Volunteerism Sustained Volunteerism TIME Volunteer Role Identity Volunteer Social Pressure Decision to Volunteer Prosocial Personality Situational Factors Fig. 1. The Causes of Sustained Volunteerism. The figure represents a conceptual model of the direct and indirect influences on sustained volunteerism. The stronger causal relationships are represented by solid lines; the weaker ones by dashed lines. 462 Penner to comply with these feelings. Several studies have found that before people decide to volunteer they are exposed to both explicit and implicit kinds of social pressures. The greater these pressures, the more likely the person is to volunteer (Grube & Piliavin, 2000; Independent Sector, 1999; Piliavin & Callero, 1991). Thus, the second (and stronger) causal path in the model is from Volunteer Social Pressure to Decision to Volunteer. However, as suggested earlier, it does not seem likely that the targets of social pressure to volunteer are randomly selected; some people are more likely to be asked to volunteer than others. Similarly, it does not seem likely that all people respond affirmatively to implicit or explicit pressures to volunteer; some people are more likely to agree than others. Therefore, the model identifies some additional direct and indirect causes of the Decision to Volunteer. Each of these is described below. The model proposes that one demographic variable, three dispositional variables, and one organizational variable are related to the Decision to Volunteer. The first of these, Demographic Characteristics, is made up of things such as age, income, education, etc. (Strictly speaking, this is known as a “composite” variable [Bollen & Lennox, 1991].) The three dispositional latent variables are: Personal Beliefs and Values, which involves religious beliefs and other yet unspecified values and beliefs related to prosocial tendencies; Prosocial Personality, which concerns personality traits associated with prosocial thoughts, feelings, and behaviors; and Volunteer-Related Motives, which concerns the motives that underlie volunteering (see Clary et al., 1998). The organizational latent variable that influences the Decision to Volunteer is Organizational Attributes and Practices, which, as discussed earlier, involves an organization’s reputation, values, and practices. The model posits that the dispositional variables directly influence both the likelihood that a person will be the target of social pressure to become a volunteer and the decision to volunteer itself. However, Organizational Attributes and Practices influences only the Decision to Volunteer. That is, the model proposes that, because of their attributes and practices, some organizations are more likely to attract certain volunteers than others. Once the decision to become a volunteer is made, then the question becomes: What factors are responsible for differences in Initial Volunteerism? Initial Volunteerism is the amount of time and effort a person expends during the early stages of his/her tenure as a volunteer. The research presented in this article and elsewhere strongly indicates that differences in levels of Initial Volunteerism covary with differences in Demographic Characteristics, Personal Beliefs and Values, Prosocial Personality, Volunteer-Related Motives, Organizational Attributes and Practices and one other organizational variable, Relationship with the Organization. Relationship with the Organization involves the kinds of job attitudes presented earlier (e.g., job satisfaction, organizational commitment, etc.). Note, that in the interest of simplicity and clarity, the figure does not show any bidirectional links between Sustained Volunteerism 463 the dispositional variables and the organizational variables. However, it is a core assumption of the model that there are reciprocal influences within and among the different classes of variables. Thus, each of the causal variable’s impact on Initial Volunteerism is both direct and indirect. The next path in the model is from Initial Volunteerism to Volunteer Role Identity. Volunteer Role Identity is a concept developed primarily by Piliavin and her associates (Grube & Piliavin, 2000; Piliavin, Grube, & Callero, this issue) and concerns the extent to which a person identifies with and internalizes the role of being a volunteer; that is, the extent to which this role and the relationships associated with it become part of a person’s self-concept. According to Grube and Piliavin (2000), a particular role identity is shaped by the behavioral expectations of others who interact with the person in the context of that role, and the self-attributions that result from the person consistently engaging in behaviors associated with that role (also see Piliavin et al., this issue). Consistent with this theorizing, the present model posits that a person’s experiences during the Initial Volunteerism will shape his/her Volunteer Role Identity. A high and involving level of volunteer activity will likely produce a strong volunteer role identity. And it is a person’s Volunteer Role Identity that is the direct and proximal cause of Sustained Volunteerism, the amount of volunteer activity a person engages in after he or she has been a volunteer for some significant period of time. The link between Initial Volunteerism and Volunteer Role Identity is directly supported by the work of Grube and Piliavin (2000), and less directly by Penner and Finkelstein (1998). Findings from Piliavin and Callero (1991) and Penner and Finkelstein would appear to support the causal path between Volunteer Role Identity and Sustained Volunteerism. Finally, the model proposes that in addition to their mediated relationships with Sustained Volunteerism (through their influence on Initial Volunteerism) the other variables (e.g., Prosocial Personality, Relationship with the Organization, etc.) have some direct influence on this sustained prosocial action. However, this influence is less than their influence on Initial Volunteerism and, of course, less than the influence of Volunteer Role Identity on Sustained Volunteerism. (These weaker relationships are also indicated by broken lines.) The primary reason the model proposes that the relationships will be weaker is that as volunteers develop a Volunteer Role Identity, dispositional, and organizational variables should become less important causes of Sustained Volunteerism. Instead, the most potent direct causes of Sustained Volunteerism are people’s perceptions of themselves and the roles they occupy (i.e., their Volunteer Role Identity). Theory and Practice The present approach to sustained volunteerism was predicated on the assumption that it is more likely than other prosocial actions to be influenced by dispositional and organizational variables. The immediate goal of this article was 464 Penner to identify these variables and present a conceptual model of how they independently and collectively affect sustained volunteerism. The model presented in Figure 1 can be empirically tested and if it is substantially correct, it provides some fairly straightforward suggestions as to howservice organizations might attract and retain volunteers. A few of these are considered briefly here. The findings on the role of motives in the decision to volunteer suggest that service organizations interested in recruiting new volunteers might benefit by identifying the things that would motivate a certain target group to volunteer and then highlight these motives in their recruiting appeals directed at this target group. There already are data that indirectly support this suggestion (see Clary et al., 1998). Of equal—if not more—importance is what service organizations might do to retain volunteers. That is, it can be argued that if service organizations face a personnel problem, it is not a shortage of people who want to volunteer. Instead, it is attrition among people in the early stages of their tenure with the organization (VolunteerMatch, personal communication, September 15, 2001). In this regard, nonprofit service organizations may be aided by the work of industrial and organizational psychologists who study prosocial behaviors among paid employees. As noted earlier, job attitudes directly affect a worker’s proclivity to engage in such behaviors (Organ & Ryan, 1995). Although some portion of the differences in job attitudes may be due to dispositional factors, a more direct and powerful cause of differences is how a person is treated by the organization. It seems reasonable to argue that the same principles would apply in the case of unpaid volunteers. That is, the better they are treated by the service organization, the greater their initial levels of volunteerism will be. Virtue may be its “own reward,” but intelligent and progressive management practices would not hurt either (Graham, in press; Skarlicki & Latham, 1996). One should not assume that just because a person is motivated by altruistic concerns that his or her initial level of volunteer service would be unaffected by attitudes toward the service organization. Thus, service organizations must do more than simply recruit volunteers; they must work to maximize the volunteers’ involvement with the organization. If the initial level of volunteering can be maintained, a volunteer role identity should develop. Once this identity has emerged, the organization has a volunteer who should remain a long-term and active contributor (Lee et al., 1999). Service organizations may want to also turn to some basic social psychological theories for other ways in which they could foster a volunteer role identity. For example, the research on the justification of effort and cognitive dissonance (e.g., Cooper, 1980) might be of direct value to a service organization looking for ways to keep good volunteers. Greatly simplified, this research indicates that, all other things being equal, working hard for something makes a person like it more (Gerard & Mathewson, 1966). This would suggest that immediately getting a new volunteer involved in reasonable organizational activities should engender more positive attitudes toward the organization. By the same token, dissonance theory Sustained Volunteerism 465 (and common sense) suggest that the worst personnel mistake a charity can make is to have no tasks for a new recruit to do. These few examples suggest some of the ways in which social psychologists and other behavioral scientists might become valuable resources for service organizations. That is, they could provide theory-based suggestions that might increase sustained volunteerism. This would be of immeasurable benefit to specific organizations and the general public as well. Thus, the study of volunteerism provides us with another chance to prove the wisdom of Kurt Lewin’s (1951) claim that nothing is so practical as a good theory. References Allen, T. D. (1999, April). Mentoring others: Mentor dispositions and desired prot´eg´ee characteristics. Paper presented at 14th Annual Meeting of Society of Industrial and Organizational Psychology. Bellah, R., Madsen, R., Sullivan, W. M., Swidler, A., & Tipton, S. M. (1985). Habits of the heart: Individualism and commitment in American life. Berkeley, CA: University of California Press. Bollen, K., & Lennox, R. (1991). Conventional wisdom on measurement: A structural equation perspective. Psychological Bulletin, 110, 305–314. Borman,W. C., Penner,L. A., Allen,T. D.,&Motowidlo, S. (2001). Personality predictors of citizenship performance. International Journal of Selection and Assessment, 9, 52–69. Brehm, J. (1993). The phantom respondents. Ann Arbor, MI: University of Michigan Press. Cicognani, E. (1999). Chi diventer`a donatore di organi? Uno studio sulle determinanti dei comportamenti di salute “altruistici,” Psicologia della Salute, 3/4, 112–133. Clary, E. G.,&Orenstein, L. (1991). The amount and effectiveness of help: The relationship of motives and abilities to helping behavior. Personality and Social Psychology Bulletin, 17, 58–64. Clary, E. G., & Snyder, M. (1991). A functional analysis of altruism and prosocial behavior: The case of altruism. In M. Clark (Ed.), Prosocial behavior (pp. 119–148). Newbury Park, CA: Sage. Clary, E. G., Snyder, M., Ridge, R., Copeland, J., Haugen, J., & Miene, P. (1998). Understanding and assessing the motivations of volunteers: A functional approach. Journal of Personality and Social Psychology, 74, 1516–1530. Colby, A., & Damon,W. (1992). Some do care: Contemporary lives of moral commitment. New York: The Free Press. Cooper, J. (1980). Reducing fears and increasing attentiveness: The role of dissonance reduction. Journal of Experimental Social Psychology, 47, 452–460. Davis, M. H., Hall, J. A.,&Meyer,M. (2001). The first year: Influences on the satisfaction, involvement, and persistence of new community volunteers. Unpublished manuscript. Davis, M. H., Mitchell, K. V., Hall, J. A., Lothert, J., Snapp, T., & Meyer, M. (1999). Empathy, expectations, and situational preferences: Personality influences on the decision to volunteer helping behaviors. Journal of Personality, 67, 469–503. Dillman, D. (2000). Mail and Internet surveys: The tailored design method (2nd ed.). New York: John Wiley. Dovidio, J. F., & Penner, L. A. (2001). Helping and altruism. In M. Brewer & M. Hewstone (Eds.), Blackwell international handbook of social psychology: Interpersonal processes (pp. 162–195). Cambridge, MA: Blackwell. Epstein, S. (1979). The stability of behavior: I. On predicting most of the people much of the time. Journal of Personality and Social Psychology, 37, 1097–1126. Facteau, J. D., Allen, T. D., Facteau, C. L., Bordas, R. M.,&Tears, R. S. (2000). Structured interviewing for OCBs: Construct validity, faking, and the effects of question type. Paper presented at the 15th Annual Meeting of the Society for Industrial and Organizational Psychology,NewOrleans, LA. Forde, D. S. (2000). Correlates of organizational behavior motives. Unpublished honors thesis. 466 Penner Gerard, H., & Mathewson, G. C. (1966). The effects of severity of initiation on liking for a group: A replication. Journal of Personality and Social Psychology, 2, 278–287. Graham, J. (in press). Promoting civic virtue organizational citizenship behavior: Contemporary questions rooted in classical quandaries. In W. C. Borman & S. J. Motowidlo (Eds.) Human resources management review. Grube, J.,&Piliavin, J. A. (2000). Role identity, organizational experiences, and volunteer experiences. Personality and Social Psychology Bulletin, 26, 1108–1120. Independent Sector. (1999). Giving and volunteering in the United States 1999: Executive summary. Washington, DC: Author. Jackson, L. M., & Esses, V. M. (1997). Of scripture and ascription: The relation between religious fundamentalism and intergroup helping. Personality and Social Psychology Bulletin, 23, 893–906. Krosnick, J. (1999). Survey research. Annual Review of Psychology, 50, 537–558. Lee, L., Piliavin, J. A., & Call, V. (1999). Giving time, money, and blood: Similarities and differences. Social Psychology Quarterly, 62, 276–290. Lewin, K. (1951). Formulation and progress in psychology. In D. Cartwright (Ed.), Field theory in psychology: Selected theoretical papers. New York: Harper. Little, S. A. (1994). Altruism in college volunteers: Relationships to prosocial personality, constructive thinking, and parenting variables. Unpublished doctoral dissertation. University of Rhode Island, Kingston, RI. Midili, A., & Penner, L. A. (1995). Dispositional and environmental influences on organizational citizenship behavior. American Psychological Association, New York, NY. Negrao, M. (1997). On good Samaritans and villains: An investigation of the bright and dark side of altruism in organizations. Unpublished manuscript. Oliner, S., & Oliner, P. (1988). The altruistic personality: Rescuers of Jews in Nazi Europe. New York: The Free Press. Omoto, A., & Snyder, M. (1995). Sustained helping without obligation: Motivation, longevity of service, and perceived attitude change among AIDS volunteers. Journal of Personality and Social Psychology, 68, 671–686. Organ, D. W., & Ryan, K. (1995). A meta-analytic review of attitudinal and dispositional predictors of organizational citizenship behavior. Personnel Psychology, 48, 775–802. Paul, A. M. (1999, July 25). Born to be good. USA Weekend On-Line. Penner, L. A., Escarraz, J., & Ellis, B. B. (1983). Sociopathy and helping: Looking out for number one. Academic Psychology Bulletin, 5, 195–209. Penner, L. A., & Finkelstein, M. A. (1998). Dispositional and structural determinants of volunteerism. Journal of Personality and Social Psychology, 74, 525–537. Penner, L. A., & Fritzsche, B. A. (1993). Measuring the prosocial personality: Four construct validity studies. American Psychological Association, Toronto, Canada. Penner, L. A., Fritzsche, B. A., Craiger, J. P., & Freifeld, T. R. (1995). Measuring the prosocial personality. In J. Butcher & C. D. Spielberger (Eds.), Advances in personality assessment (Vol. 10). Hillsdale, NJ: Lawrence Erlbaum. Penner, L. A., Midili, A. R., & Kegelmeyer, J. (1997). Beyond job attitudes: A personality and social psychology perspective on the causes of organizational citizenship behavior. Human Performance, 10, 111–132. Piliavin, J. A., & Callero, P. (1991). Giving blood: The development of an altruistic identity. Baltimore: Johns Hopkins University Press. Piliavin, J. A., Dovidio, J. F., Gaertner, S. L., & Clark, R. D. III (1981). Emergency intervention. New York: Academic Press. Rioux, S., & Penner, L. A. (2001). The causes of organizational citizenship behavior: A motivational analysis. Journal of Applied Psychology, 86, 1306–1314. Schroeder, D. A., Penner, L. A., Dovidio, J. F., & Piliavin, J. A. (1995). The psychology of helping and altruism. New York: McGraw-Hill. Sibicky, M., Mader, D., Redshaw, I.,&Cheadle, B. (1994). Measuring the motivation to volunteer. Paper presented at the annual meeting of the Midwestern Psychological Association, Chicago, Illinois. Skarlicki, D. P., & Latham, G. P. (1996). Increasing citizenship behavior within a labor union: A test of organizational justice theory. Journal of Applied Psychology, 81, 161–169. Sustained Volunteerism 467 Snyder, M. (1993). Basic research and practical problems: The promise of a functional personality and social psychology. Personality and Social Psychology Bulletin, 19, 251–264. Statistics Canada. (2001). Canadians, caring and involved: Highlights from the 2000 national survey of giving, volunteering, and participating. Ottawa, Canada: Minister of Industry. Tillman, P. (1998). In search of moderators of the relationship between antecedents of organizational citizenship behavior and organizational citizenship behavior: The case of motives. Unpublished master’s thesis, University of South Florida, Tampa, FL. LOUIS A. PENNER is a Professor of Psychology at the University of South Florida. He received his PhD in Social Psychology from Michigan State University in 1969. His research focuses primarily on organized forms of prosocial behavior among volunteers and paid employees. This research examines both the dispositional variables and the organizational policies and procedures that might be associated with such behaviors. As part of this effort, he has developed measures of prosocial personality characteristics and the motives that underlie prosocial actions.