Material Assistance: Who Is Helped by Nonprofits?
Guo, Baorong, Social Work Research
Although a vast amount of literature on the characteristics of public assistance recipients exists, little is known about the characteristics of the clientele of nonprofits that provide material assistance. This study examines the question of who receives material assistance from nonprofits. Three panels of the Survey of Income and Program Participation (1996, 2001, and 2004) adult well-being module were used to see if the use of nonprofit assistance is associated with individual and household characteristics, type of material hardship, and the use of other sources of assistance. Results show that poverty status, education, area of residence, and public program participation have a significant association with nonprofit material assistance, regardless of the type of material hardship and the use of other sources of assistance. These associations stayed relatively stable during the years of observation. It is interesting that households headed by an individual with college education were more likely than those headed by an individual without a high school diploma to receive material assistance from nonprofits. These findings may help nonprofits develop appropriate assistance programs and reach target populations.
KEY WORDS: low-income households; material assistance; material hardship; nonprofits; social service
Material hardship is a condition of failure to meet the "minimum levels of very basic goods and services, such as food, housing, and medical needs" (Beverly, 2001, p. 143). In the United States, one in every 10 people lives in material hardship (Beverly, 2001). To help those with material hardship, government at various levels and nonprofit groups provide a wide range of material assistance. During the recent economic downturn, the need for material assistance rose and an increasing number of people turned to nonprofits for assistance (Ducey, 2009). Despite the important role of nonprofits in social assistance historically and currently (for example, Allard, 2008; Lipsky & Smith, 1989; Young, 1999), little is known about the clientele of the nonprofit sector (Allard, 2008). Two reasons may account for this. First, nonprofit assistance is still small in size and scope relative to public assistance, although this is changing as the partnership between government and nonprofits grows (Salamon, 1995; Young, 1999). Second, as noted by Allard, the local nature of service providers poses a challenge to research on nonprofit groups, and this is partly reflected by the lack of nationally representative data on the use of nonprofit assistance by individuals.
Despite how little we know about the clientele of the nonprofit sector, a huge gap can be seen in the extent of material hardship and the availability of material assistance to the needy. Because of this, nonprofits are expected to play a larger role in social assistance to the needy (Alexander, Nank, & Stivers, 1999; Salamon, 1995). For this to happen, we need to look closely at who receives material assistance from nonprofits and what factors are associated with the use of nonprofit assistance. An inquiry into these questions will not only benefit nonprofits as they try to develop appropriate assistance programs and reach target populations, but also inform policymakers of the similarities and differences between the clientele of the nonprofit sector and that of the public sector. The Survey of Income and Program Participation (SIPP), which incorporates questions regarding nonprofit assistance, makes it possible to closely study the clientele of nonprofit assistance. Three panels of the SIPP (1996, 2001, and 2004) were used by the current study to examine the characteristics of nonprofits' clients and the determinants of the use of nonprofit assistance by families with material hardship.
As is widely recognized, nonprofit assistance is distinct from other sources of assistance, such as government, family, and friends. Nonprofits are private institutions that act in the public interest (Salamon, 1999). Nonprofit assistance is different from government assistance because it clearly reflects collective and voluntary action in response to citizens' needs in a way independent of government effort. Although the independence of nonprofits is frequently doubted as they receive more or less government support, there is wide agreement that nonprofits are "institutionally separate from the government" (Salamon, 1999). Nonprofit assistance is also different from family and friend assistance because it is formalized and institutionalized rather than based on personal networks.
The question of who uses nonprofit assistance can be examined from a microeconomic perspective on consumption. From this perspective, individuals and families try to maximize their well-being at different stages of life, and the use of nonprofit assistance can be considered such an effort in difficult times. Material hardship often occurs when households experience poverty or negative income shocks. The presence of material hardship indicates that people have to constrain their basic consumption due to economic pressure (Schoeni & Ross, 2005). To maintain consumption and maximize well-being, some families may pursue material assistance from family members and friends, nonprofit groups, or public programs (for example, the Supplemental Nutritional Assistance Program [SNAP]). The choice of different sources of assistance is, therefore, an economic decision that compares the cost and benefit of assistance. This decision takes into account a number of factors, such as benefit level, stigma, program eligibility, and administrative barriers. According to this theoretical model, some low-income families may decide to take nonprofit assistance to maximize their consumption utility when there are barriers to public assistance or assistance from family and friends or when other sources of aid are insufficient. Barriers may include, but are not limited to, the unavailability of assistance (for example, lack of resources in social networks and ineligibility for public programs), a lack of knowledge about public programs, and the stigma and transaction costs of public assistance. For example, a report (U.S. Department of Agriculture, 2008) showed that only two-thirds of those eligible for SNAP (formerly known as the Food Stamp Program) actually received public food assistance, suggesting that many people may have misinformation about the eligibility rules. Also, given the transient nature of material hardship (Ribar & Hamrick, 2003), public assistance may be viewed as less worthy than other alternatives due to the long and complex application process. Instead, people may turn to nonprofit assistance located in their communities. Sometimes, nonprofit assistance is preferred over public assistance because the latter is often associated with stigma (Moffitt, 1983). All of this seems to point to the substitution effect of nonprofit assistance. That is, families receive nonprofit assistance because they are less likely to obtain assistance from other sources (for example, government, family, and friends). Families receiving nonprofit assistance may be demographically different from those receiving assistance from other sources. A study showed that a small proportion of people actually use both private and public assistance (Wu & Eamon, 2007).
From a different perspective, low-income ramifies may take nonprofit assistance due to the insufficiency of assistance from other sources. In other words, nonprofit assistance and public assistance are supplements for each other rather than substitutes, and families that receive public assistance may also pursue nonprofit assistance to maximize their consumption. Indeed, nonprofit assistance and public assistance are more than supplements because receipt of assistance from both sources may reflect a complementary government--nonprofit relation. That is, nonprofits are engaged in social assistance not only through provision of additional resources, but also through a partnership with government. Consequently, those who receive public and nonprofit assistance may have somewhat similar demographic characteristics. According to Young (1999), government has inherent limits in its ability to meet diverse social needs, and the nonprofit sector often faces resource constraints. As a result of the inevitable inadequacy of both sectors, a partnership has emerged in which the government provides financing and the nonprofits deliver services. Just as government has increasingly relied on purchase-of-service contracts with nonprofits, nonprofits have become increasingly dependent on government funding for delivery of services to the public. The growing partnership between government and nonprofits since the 1960s precisely reflects this complementary perspective (Young, 1999). That levels of nonprofit antipoverty activity are higher in wealthier areas (meaning more resources for nonprofits) where government contribution is also higher is evidence for this complement framework (Joassart-Marcelli & Wolch, 2003).
Both substitution and complementarity effects of nonprofit assistance are supported to some extent by the empirical literature. Consistent with the assumption of substitution of nonprofit assistance in place of government assistance, research has shown that factors commonly associated with the use of public assistance (such as age, education, marital status, and employment) are not associated with the use of private assistance. In Neustrom, Deseran, and Moore's (2001) study using the Louisiana Survey of Families and Households, a number of factors (including age, education, employment status, rural/urban residence, family structure, marital status, Aid to Families with Dependent Children participation, number of children in a household, and race) were examined as they related to sources of informal assistance, but none of the variables yielded consistent results. Similarly, Wu and Eamon (2007) did not identify any significant factors related to private assistance. Although not explicitly supporting substitution effects, Wu and Eamon noted that only 7% of their sample received both private and public assistance. These findings clearly show that the clientele of private assistance is different from that of public assistance.
As mentioned previously, the complementary perspective finds its support in Joassart-Marcelli and Wolch (2003). In addition, Sommerfeld and Reisch (2003) observed growing budget sizes and service expansion in over half of the study nonprofits after welfare reform. Allard (2008) also noted that recipients of nonprofit assistance are more likely to be women who are primary caregivers in poor families, suggesting that nonprofit assistance is an important supporting element for welfare recipients. These findings support the complementary effects.
Based on the exiting literature, the current study examines the characteristics of the clientele of nonprofit assistance. It contributes to the literature by using national representative data to specifically examine nonprofit assistance, a unique category of social assistance. To specifically focus on nonprofit assistance, this study uses data from the SIPP to examine the following three questions: (1) Among low-income households that encountered material hardship, who received material assistance from nonprofits? (2) Did the clientele of nonprofit material assistance change in the decade after the mid-1990s? and (3) Were recipients of nonprofit material assistance different from those who did not receive any form of assistance? The SIPP data provides a representative sample of Americans who are low income and have material hardship. In addition, the three panels of the SIPP (1996, 2001, and 2004) allow for a close examination of the trends in clients of nonprofit assistance.
DATA AND METHOD
The SIPP is designed to collect information from the U.S. civilian, noninstitutionalized population on income, labor force participation, program participation, and demographic characteristics. The survey uses a continuous series of nation panels; the survey runs four years for most of the panels after 1984.
In addition to a set of core questions, the SIPP also includes various topical modules (for example, child care, child support, and adult well-being) that are assigned to particular interviewing waves of the survey. The adult well-being module was chosen for this study. This module (after the 1996 panel) includes questions about sources of material assistance specified as government, nonprofits/charity, family, friends, and other. The adult well-being module was used in 1998 for the 1996 panel, in 2003 for the 2001 panel, and in 2005 for the 2004 panel.
A sample of low-income households with material hardship was chosen on the basis of the following criteria. First, given that the SIPP collects data from all household members age 15 years and older, only observations of household heads (household reference individuals) were selected because household characteristics are shared by each household member. Second, households with an income at 200% of the federal poverty line or below were selected because they were more likely to experience material hardship than those with a higher income (Iceland & Bauman, 2007). Third, only households that had experienced material hardship were chosen because questions about material assistance were asked on the basis of responses to questions about material hardship (see the Measures section). Finally, non-family households (n = 139) were excluded. Application of these sample selection criteria resulted in a sample of 6,222 households in the pooled data of the three panels.
Material Hardship. Measures of material hardship were taken from the SIPP adult well-being module, which includes a set of questions about household material hardship experiences. For example, it asks, "Did the household pay the full amount of the rent or mortgage in the past 12 months?" Other forms of material hardship include failure to pay utility bills, utilities cut off, telephone disconnection, lack of medical care, lack of dental care, and eviction. Households that had a positive response on any of these items were considered to have material hardship. For each of these material hardship questions, if a positive response was given, there was a follow-up question asking "When (you/your household) had this problem, did any person or organization help?" One or more choices were expected from the following six categories: (1) a family member or relative; (2) a friend, neighbor, or other nonrelative; (3) a department of social services; (4) a church or nonprofit group; (5) other, and (6) no assistance. This question regarding sources of material assistance was not asked of those who did not report any material hardship. Given this skip pattern in the survey, the sample was limited to households with material hardship because the study is focused on sources of material assistance.
Dependent Variable. The dependent variable "nonprofit assistance" is a dummy variable that indicated whether a household received any form of nonprofit material assistance. Households were coded 1 if they received material assistance from nonprofits in the past 12 months to address one or more of the seven types of material hardship described previously; otherwise they were coded 0.
Independent Variables. There were four categories of independent variables in this study. The variable "demographic/household characteristics" includes race (black, white, other), gender (1 = female; 0 = male); family types (1 = married couple; 0 = otherwise), age, poverty status (0 = under the poverty line; 1 = otherwise), education (less than high school, high school, and more than high school), region (1=rural; 0 = urban), homeownership (1 = homeowner; 0 = renter), employment (1 = employed; 0 = unemployed), number of children, participation in one or more public programs (including Temporary Assistance for Needy Families, Food Stamps, Medicaid, and Social Security). The variable types of material hardship refers to the seven types of material hardship noted previously, each indicated by a dummy variable. Three dummy variables indicated other sources of material assistance, namely, family, friends, and government. For instance, households that reported "yes" to the third category of sources of material assistance were considered recipients of government assistance. Assistance from family and assistance from friends were measured in a similar way. Government material assistance was defined differently than public program participation because government material assistance aims to directly address material hardship, whereas these public programs are designed to directly address poverty through income support and indirectly address material hardship. Also, government material assistance is mostly temporary support from local governments, whereas support from major public programs (provided by the federal and state governments) is on a relatively regular basis. Finally, this study included three panel variables that indicated the 1996, 2001, and 2004 panels, respectively.
Statistical analysis began with a descriptive summary of the demographic and socioeconomic characteristics of households in the pooled sample and the individual panels. Then, for the first question--that is, "What factors are associated with the use of nonprofit assistance?"--rare events logistic regression was used. In this study, the use of nonprofit material assistance was a rare event in low-income households: only 4.73% of low-income households in the pooled sample received nonprofit assistance. When popular statistical procedures such as logistic regression are used to study rare events in finite samples, they can sharply underestimate the probability of rare events (King & Zeng, 2001). To provide corrected estimation, King and Zeng extended analytical approximations provided by McCullagh and Nelder (1989) and used weighted least squares to estimate the bias in parameter estimation. Then a correction was made with this bias being removed from estimation. According to King and Zeng, the correction is most effective when the sample size is under a few thousand and the occurrence of the events is less than 5% in the sample, both of which apply in the current study.
Rare event logistic regression was run to include the independent variables hierarchically. The base model included only the panel variables and the head of household's characteristics (that is, age, gender, race, education, marital status, and employment) and household characteristics (that is, number of children, poverty status, region, and public program participation). For the next step, the seven variables that indicated different types of material hardship were added to see if the use of nonprofit assistance was particularly associated with certain types of material hardship. Finally, other sources of material assistance--family, friends, and government--were entered to examine whether the use of nonprofit assistance was associated with any other sources of material assistance.
To examine the second question--that is, "Does the clientele of nonprofit material assistance change over time?"--the Chow test was conducted on the pooled sample to see whether the interaction terms of the panel variables and other independent variables had significant results. In other words, it examined whether the effects of these independent variables on the use of nonprofit assistance changed over the three observation periods. For example, if households in poverty were more likely to receive nonprofit assistance in the second or third panel, we would expect to see a significant positive regression coefficient for the interaction terms of poverty status and indicators of the second or third panel. Finally, households that received nonprofit assistance were compared with those that did not receive any assistance. All the statistics reported herein were weighted using the household weight variable provided by the SIPP.
The characteristics of low-income households in the pooled sample and in each panel are summarized in Table 1. Consistent with national statistics (for example, U.S. Census Bureau, 2009), this low-income sample had more women than men. The average age of household heads in each panel is around 40 years. Slightly below 50% of this low-income sample live under the federal poverty line. The sample shows an even distribution of three levels of education: less than high school, high school, and above high school. Nearly half of respondents were married, and the rest were single men (7.1%) or women (46%). On average, each household has 1.8 children. Slightly more than one-third of respondents are homeowners. Although over 60% of respondents have a job, over 40% of the households participated in one or more public programs, such as TANF, food stamps, Medicaid, and Supplemental Security Income.
The extent of material hardship in low-income households varied. The most common material hardship was failure to pay utility bills, followed by not seeing a doctor as needed and failure to pay rent or mortgage. Eviction due to failure to pay rent or mortgage occurred least frequently.
Regarding material assistance, family was clearly the first and foremost source of assistance, followed by government assistance, nonprofit assistance, and friend support. Despite the prevailing view that nonprofits play an important role in helping poor families, less than 5% of the households that experienced material hardship were helped by nonprofit organizations. Consistent in the three panels, about 77% of households with material hardship did not receive any form of material assistance.
Model 1: Characteristics of Recipients of Nonprofit Assistance
The association between characteristics of households/heads and the use of nonprofit material assistance is presented in model 1 of Table 2. First, the positive regression coefficients of the two panel variables indicate an increasing probability of receiving material assistance from nonprofit organizations over time; however, only the second panel variable (the 2001 panel) was statistically significant at the .05 level. Second, different from previous studies (for example, Neustrom et al., 2001; Wu & Eamon, 2007), several demographic/socioeconomic variables, including poverty status, household head's education, residence region, and public program participation, were significant in the model. Households that were in poverty, were headed by individuals with more education, were in rural areas, and participated in public programs were more likely to use nonprofit assistance. As indicated in Table 3, for a typical case in the sample (defined in Table 1), the predicted probability of receiving nonprofit material assistance for households with and without poverty was 6.45% and 4.64%, respectively; households living in rural areas had a probability 1.3% higher than those urban households; and public program participants were 3% more likely than nonparticipants to use nonprofit assistance. An interesting finding is that, holding other variables constant, households headed by college graduates were significantly more likely than those headed by individuals without a high school diploma to obtain nonprofit assistance; the difference in the predicated probability was nearly 2%. Given that the use of nonprofit assistance is a rare event and the predicted probability is very small for all households in general, this difference suggests that the probability of using nonprofit assistance for college graduates was 1.6 times that of those without a high school diploma. The results of model 1 show the chance for a disadvantaged household (for example, a rural household headed by a black single mother, who does not own a home, has not graduated from high school, is unemployed, and participates in public programs) to receive nonprofit material assistance was nearly 8%.
Model 2: Types of Material Hardship and Nonprofit Assistance
Model 2 in Table 2 included a set of dummy variables to examine the probability of using nonprofit assistance in relation to the type of material hardship (eviction for failure to pay is excluded due to its extremely low frequency in the sample). Three housing-related material hardships, including failure to pay rent/mortgage, failure to pay utility bills, and utility disconnection, were significantly related to the use of nonprofit assistance. In other words, households with one or more of these material hardships had a greater probability of receiving nonprofit assistance. As indicated in Table 3, when everything else was the same, households failing to pay their rent or mortgage were twice as likely to receive nonprofit assistance as those without this material hardship. A similar probability ratio was found for those who failed to pay utility bills versus those who did not.
Addition of different types of material hardship did not change the effects of most demographic/ socioeconomic variables in the model. As indicated by the magnitude of their regression coefficients in Table 2, the influences of these variables on the dependent variable became slightly smaller in model 2. Household head's age became marginally significant at the .1 level, suggesting that households headed by younger adults were less likely to obtain nonprofit assistance.
Model 3: Other Sources of Material Assistance
Model 3 included three additional sources of material assistance--family, friends, and government. As it turns out, the likelihood of using nonprofit assistance did not differ between those with and those without family support. The regression coefficient of assistance from friends was marginally significant at the .1 level. The results indicate that recipients of government assistance were more likely to receive nonprofit assistance than those who did not receive government assistance.
Those variables that were significant in model 1 and model 2 remained significant in model 3, except poverty status was statistically significant only at the .1 level. As expected, controlling for other sources of material assistance reduced the magnitude of regression coefficients of formerly entered variables.
Model 4: Trends in Nonprofit Assistance
In addition to the rare events logistic regression, the Chow test was conducted to see whether the associations of these independent variables with the dependent variable changed over time. The main effects of independent variables and their interaction effects with the three panel variables are reported in Table 4. The main effects represent the role of these variables in predicting the probability of using nonprofit assistance in the first panel only (and therefore the significance level of the main effects cannot be compared with those in models 1 to 3). For this part of the analysis, the discussion focuses mainly on the interaction effects. As shown in Table 4, most interaction effects were not statistically significant, indicating that the associations of the independent variables with the dependent variable stayed relatively stable over time. But some patterns emerged when the regression coefficients of the interaction terms were closely examined. For example, as discussed previously, households headed by college graduates, on average, had a higher probability of using nonprofit assistance. However, the difference in the likelihood of obtaining nonprofit assistance between those with and without college education became smaller over the years. In contrast, the probability of receiving nonprofit assistance for those living in rural areas and public program participants became higher.
There were only two variables--household head's age and material assistance from friends--that showed significant interaction with the panel variables. Specifically, compared with the first panel, households headed by younger adults were more likely to receive nonprofit assistance, and households with assistance from friends were less likely to receive nonprofit assistance in the third panel.
Models 1 through 4 can be considered contrasts between households that received and that did not receive nonprofit assistance. Therefore, Table 5 provides a comparison of households that received nonprofit assistance and those that did not receive any source of assistance. There are two reasons why households with material hardship did not receive any help. On the positive side, their material hardship may have been minor and transient, and these households were thus able to manage through difficult times without any help. On the negative side, some households, although in need of help, did not receive any assistance because of reasons such as lack of transportation or lack of knowledge about where to obtain help. The question arises of whether nonprofit assistance recipients were more or less disadvantaged than those without any assistance. The two groups differed significantly on such variables as household head's gender and education, poverty status, and public program participation. Households that received nonprofit assistance were, in general, more disadvantaged than those that did not receive any assistance. The two groups did not differ from each other in their need for medical and dental services, but recipients of nonprofit assistance were more likely to have experienced an eviction, utility shutoff, or failure to pay rent/mortgage and utility bills than those who did not receive nonprofit assistance.
This study contributes to the literature by focusing on material assistance provided by nonprofits to low-income households with material hardship. In addition, it examined whether the association between the independent variables and the use of nonprofit assistance is time variant. Consistent with previous studies (for example, Guo, 2010; Wu & Eamon, 2007), the results showed a low proportion of households received nonprofit assistance, and this stayed relatively stable in the years of observation.
In contrast to previous studies that indicate the substitution effects of public and private assistance (for example, Neustrom et al., 2001; Wu & Eamon, 2007), the findings of this study suggest the supplement and complement effects of public and private assistance because public program participation was a significant predictor of the use of nonprofit material assistance and participants in public programs were more likely than nonparticipants to receive nonprofit assistance. This, on the one hand, may suggest that low-income households with material hardship need assistance in addition to public assistance. Nonprofit organizations may consider targeting these households because public program participation precisely reveals their socioeconomic disadvantages. On the other hand, in an era that calls for stronger collaboration between nonprofits and the public sector in funding and service delivery, public programs are not only sources of assistance, but also gateways to other resources (such as nonprofit assistance). It is important to see if those who are in need but do not receive any public assistance have access to other assistance and resources. Undoubtedly, disadvantaged populations will benefit from the joint efforts of nonprofits and the government.
Several other demographic/socioeconomic variables (such as poverty status, education, and residence region) were found to have a significant association with the use of nonprofit material assistance. These significant indicators might reflect some characteristics of the clientele of nonprofit assistance, although most variables in the model were not significant. First, as expected, nonprofit assistance targets the most disadvantaged population, such as those riving in poverty, those living in rural areas, and those participating in public programs. Despite the relative homogeneity in income (as a result of the low-income criterion for sample selection), households with those disadvantaged characteristics were still more likely to obtain nonprofit assistance, and the performance of these variables was not affected regardless of the type of material hardship and the presence or absence of other sources of material assistance.
Second, there was an urban/rural divide in nonprofit assistance; people living in urban areas were less likely to receive material assistance from nonprofits. This was unexpected because there are fewer nonprofits in rural areas than in urban areas (Snavely & Tracy, 2003). It is suspected that this result is needs driven because the poverty rates are higher in rural areas. According to the U.S. Department of Agriculture Economic Research Services (2004), 14.2% of the non-metro population lives in poverty; this is two percentage points higher than the metro population. Although the finding shows a greater likelihood for rural households to receive nonprofit material assistance, as reported in Swierzewski (2007), rural nonprofits are significantly disadvantaged compared with their urban counterparts in obtainment of grants to support organizational operations and service programs.
Third, both Table 2 and Table 5 indicate that, in this low-income sample, college graduates were more likely to use nonprofit assistance than those with less education. This is in contrast to the public assistance literature, in which most participants have been found to be those with low levels of education (for example, Blank & Ruggles, 1996). Explanations for this can be quite different. People with more education may have a greater awareness of the existence of nonprofit assistance or need short-term assistance only and therefore intentionally avoid the long and complex application for public assistance. Another explanation, however, considers the potential sample selection bias. That is, the college graduates in this low-income sample are probably a very particular group of college graduates who have significant health or mental health issues, making them different from those college graduates not in the sample. It is not clear which explanation is true. Therefore, future research is necessary to closely examine whether education is a determinant of nonprofit assistance and whether it actually differentiates the client base of nonprofit assistance and public assistance.
In addition, the study investigated the association between various types of material hardship and nonprofit assistance. Limited by the types of material hardship recorded in the SIPP, the study mainly focused on hardships related to shelter, utility, and health services. The results suggest that households experiencing hardships related to shelter and utility were more likely to obtain nonprofit assistance. This is suspected to be a need-driven response because, as shown in Table 1, more households experienced shelter and utility hardships than health hardships. It is also possible that the nonprofit sector is better at providing sheltering and related assistance because it has developed service capacity in this regard over the long history of social services for the needy.
The relationships between nonprofit assistance and other sources of assistance were explored in the study. The use of nonprofit assistance was significantly related to the use of material assistance from government and friends (at the .1 level in model 3) but was not related to family assistance in any of the models. The significant relationship between government material assistance and nonprofit assistance resembles the dynamics between public program participation and nonprofit assistance. This may reflect the association of severe material hardship with greater needs for collectively provided resources. It is less clear, however, why the use of support from friends was associated with nonprofit material assistance. It is likely that households with material hardship need additional help because support from their informal networks (that is, friends) is far from enough. Due to data constraints, households' help-seeking behaviors could not be further examined in the current study. It is suspected that the use of assistance might be related to the magnitude of material hardship, amount of assistance, and even the temporal order in which different sources have been sought. Nonprofit service organizations need to understand the help-seeking behaviors of households and develop appropriate services to meet their increasing requests for help with basic needs (Curtis & Copeland, 2003; Reisch & Sommerfeld, 2003).
The findings of this study should be taken with caution. First, as discussed, there may have been a sample selection bias in the study, which could complicate the interpretation of the effect of education on the use of nonprofit assistance. Second, the estimation might be biased to some extent due to the operational definition of nonprofit assistance used by this study. The measure of nonprofit assistance specifically refers to the seven types of material assistance discussed in the Measurement section. It does not include food and clothing assistance and other assistance provided by nonprofits. Third, in the SIPP, questions about the sources of material assistance combine charitable nonprofits and churches in the same category, not allowing for a differentiation between churches, faith-based organizations, and secular nonprofit organizations. It is not clear how the results of the study are affected by these data constraints. Finally, it is likely that material assistance provided by some nonprofits is a result of contracting out by government agencies. If so, it is not possible to make a clear distinction between nonprofit assistance and public assistance.
Using the longitudinal information from a national representative sample of low-income households, this study examined factors associated with nonprofit material assistance. As the results indicate, nonprofit material assistance has reached its target population--those households that are socially and economically disadvantaged--but there is still more room for the expansion of nonprofit assistance. The findings of this study provide useful information for social workers in many nonprofit organizations to understand the client base of nonprofit assistance and develop material assistance programs. Nonprofits may reallocate resources with a greater emphasis on material assistance for low-income households. Previous studies have found that far fewer nonprofits than expected are primarily engaged in providing material assistance (Salamon, 1992; Sosin, 1986). According to Salamon's (1992) estimation, less than 20% of nonprofit human service agencies provide material assistance and over 70% do not actually focus on poor or low-income clients (Salamon, 1992). From outside nonprofits, new resources need to be explored and developed to expand nonprofits' capacity in material assistance. Nonprofit managers, social work practitioners, and policy advocates need to work together to build a stronger nonprofit sector that readily responds to social needs outside the public safety net.
The Research Notes column presents concise reports on the results of a study, discussions of methodological issues, solutions to problems, and development of ideas or insights. Send your manuscript as a Word document through the online portal at http:// swr.msubmit.net (initial, one-time registration is required).
Alexander, J., Nank, R., & Stivers, C., (1999). Implications of welfare reform: Do nonprofit survival strategies threaten civil society? Nonprofit and Voluntary Sector Quarterly, 28, 452-475.
Allard, S. W. (2008). Helping hands for the working poor: The role of nonprofits in today's safety net [Working Paper]. Retrieved from http://depts.washington.edu/wcpc/ sites/default/files/papers/Allard%2520WCPC% 2520Conference%2520Paper%25208-25-08_0.pdf
Beverly, S. G. (2001). Material hardship in the United States: Evidence from the Survey of Income and Program Participation. Social Work Research, 25, 143-151.
Blank, R. M., & Ruggles, P. (1996). When do women use Aid to Families with Dependent Children and food stamps? The dynamics of eligibility versus participation. Journal of Human Resources, 31, 57-89.
Curtis, K., & Copeland, I. (2003). The impact of welfare reform on nonprofits and people they serve in Delaware. Washington, DC: Aspen Institute, Nonprofit Sector Research Fund.
Ducey, L. (2009, September). Nonprofits scramble as Census reports rising poverty rates. Phoenix Business Journal.
Guo, B. (2010). Beyond the public safety net: The role of nonprofits in addressing material hardship of low-income households. Nonprofit and Voluntary Sector Quarterly, 39, 784-801.
Iceland, J., & Bauman, K.J. (2007). Income poverty and material hardship: How strong is the association? Journal of Socio-Economics, 36, 376-396.
Joassart-Marcelli, P., & Wolch, J. R. (2003). The intrametropolitan geography of poverty and the nonprofit sector in southern California. Nonprofit and Voluntary Sector Quarterly, 32, 70-96.
King, G., & Zeng, L. (2001). Logistic regression in rare events data. Political Analysis, 9, 137-163.
Lipsky, M., & Smith, S. R. (1989). Nonprofit organizations, government, and the welfare state. Political Science Quarterly, 104, 625-648.
McCullagh, P., & Nelder, J. A. (1989). Generalized linear models (2nd ed.). New York: Chapman and Hall.
Moffitt, R. (1983). An economic model of welfare stigma. American Economic Review, 73, 1023-1035.
Neustrom, A., Deseran, F. A., & Moore, D. (2001). Sources and types of formal and informal assistance: Patterns among Louisiana's welfare populations. Paper presented at the annual workshop of the National Association for Welfare Research and Statistics, Baltimore. Retrieved from http://www.nawrs.org/Baltimore/Papers/ w2b3.pdf
Reisch, M., & Sommerfeld, D. (2003). Welfare reform and the future of nonprofit organizations. Nonprofit Management & Leadership, 14, 19-46.
Ribar, D., & Hamrick, K. (2003). Dynamics of poverty and food insufficiency: Food assistance and nutrition research (Report No. 36). Washington, DC: U.S. Department of Agriculture.
Salamon, L. M. (1992). Social services. In C. T. Clotfelter (Ed.), Who benefits from the nonprofit sector? (pp. 134-173). Chicago: University of Chicago Press.
Salamon, L. M. (1995). Partners in public service: Government-nonprofit relations in the modern welfare state. Baltimore: Johns Hopkins University Press.
Salamon, L. M. (1999). America's nonprofit sector: A primer (2nd ed.). New York: Foundation Center.
Schoeni, R. F., & Ross, K. E. (2005). Material assistance received from families during the transition to adulthood. In F. Furstenberg & R. Settersten (Eds.), On the frontier of adulthood: Theory, research, and public policy (pp. 396-416). Chicago: University of Chicago Press.
Snavely, K., & Tracy, M. B. (2003). Collaboration among rural nonprofit organizations. Nonprofit Management & Leadership, 11, 145-165.
Sommerfeld, D., & Reisch, M. (2003). The "other America" after welfare reform: A view from the nonprofit sector. Journal of Poverty, 7, 69-95.
Sosin, M. (1986). Private benefits: Material assistance in the private sector. Orlando, FL: Academic Press.
Swierzewski, R. (2007). Rural philanthropy: Building dialogue from within. Washington, DC: National Committee for Responsive Philanthropy. Retrieved from http://www.ncrp.org/files/NCRP2007-Rural_Philanthropy_REVISEDEDITION_FINAL-Lowres.pdf
U.S. Census Bureau. (2009). Income, poverty, and health insurance coverage in the United States: 2008. Retrieved from http://www.census.gov/prod/2009pubs/ p60-236.pdf
U.S. Department of Agriculture. (2008). Reaching those in need: State food stamp participation rates in 2006. Washington, DC: Author.
U.S. Department of Agriculture Economic Research Services. (2004). Rural poverty at a glance. Retrieved from http://www.ers.usda.gov/publications/rdrr100/ rdrr100.pdf
Wu, C. F., & Eamon, M. K. (2007). Public and private sources of assistance for low-income households. Journal of Sociology & Social Welfare 34, 121-149.
Young, D. (1999). Complementary, supplementary, or adversarial? A theoretical and historical examination of nonprofit-government relations in the United States. In E. Boris & E. Steuerle (Eds.), Nonprofits and government: Collaboration and conflict. Washington, DC: Urban Institute Press.
Baorong Guo, PhD, is associate professor, School of Social Work, University of Missouri-St. Louis, 1 University Boulevard, St. Louis, MO 63121; e-mail: firstname.lastname@example.org.
Original manuscript received August 3, 2009
Final revision received June 24, 2010
Accepted July 6, 2010
Advance Access Publication September 17, 2012
Table 1: Sample Characteristics (Weighted Frequencies and Percentages) 1996 Panel (a) Pooled Sample (n = 1,868; Characteristic (N = 6,222) 30.1%) Demographic/household characteristic Race Black 1,599.37 (25.71) 503.42 (26.95) White 4,292.13 (68.98) 1,275.97 (68.31) Other 330.50 (5.31) 88.60 (4.74) Female 4,143.56 (66.60) 1,179.83 (63.16) Age (years) (c) 39.65 (12.94) 39.07 (12.71) Poverty 2,764.08 (44.42) 779.42 (41.72) Education Less than high 1,664.09 (26.96) 603.36 (33.19) school High school 2,176.98 (35.27) 645.18 (35.49) Above high school 2,330.92 (37.77) 569.46 (31.32) Family type 2,921.44 (46.95) 869.64 (46.55) (married couple) Region (rural) 1,516.60 (24.37) 425-51 (22.78) Number of children 1.76 (1.38) 1.85 (1.39) (c) Homeowners 2,345.92 (37.70) 699.67 (37.46) Employment 3,851.69 (61.90) 1,209.51 (64.75) Public program 2,838.17 (45.62) 781.22 (41.82) participation Type of material hardship Did not pay rent or 2,294.52 (36.88) 660.19 (35.34) mortgage Evicted for failure 105.45 (1.69) 32.04 (1.72) to pay rent or mortgage Did not pay gas, 3,642.42 (58.54) 1,088.12 (58.25) oil or electricity bill Lost gas, oil, or 665.70 (10.70) 181.52 (9.72) electricity for failure to pay Telephone 1,787.99 (28.74) 521.84 (27.94) disconnected for failure to pay Needed to see doctor 2,147.14 (34.51) 614.00 (32.87) or go to hospital but did not Needed to see 2,613.86 (42.01) 755.19 (40.43) dentist but did not Material assistance by source (d) Nonprofit 294.55 (4.73) 73.56 (3.94) Family 801.58 (12.88) 210.33 (11.26) Friends 200.29 (3.22) 52.70 (2.82) Government 400.16 (6.43) 96.12 (5.15) No support 4,755.99 (76.43) 1,489.98 (79.48) 2001 Panel (a) 2004 Panel (a) Typical (n = 1,582; (n = 2,772; Case Characteristic 31.3%) 38.6%) (b) Demographic/household characteristic Race Black 394.03 (24.91) 703.54 (25.38) 1 White 1,113.61 (70.39) 1,895.54 (25.38) 0 Other 74.37 (4.70) 173.26 (6.25) 0 Female 1,018.26 (64.37) 1,970.41 (71.08) 1 Age (years) (c) 39.82 (12.82) 39.96 (13.19) 39.65 Poverty 675.72 (42.71) 1,328.26 (47.92) 1 Education Less than high 454.87 (28.75) 575.43 (20.76) 0 school High school 565.96 (35.78) 961.86 (34.70) 0 Above high school 561.17 (35.47) 1,234.71 (44.54) 1 Family type 766.08 (48.43) 1,277.13 (46.07) 1 (married couple) Region (rural) 459.05 (29.02) 605.97 (21.86) 0 Number of children 1.770-38) 1.68 (1.37) 1.76 (c) Homeowners 587.11 (37.11) 1,063.79 (38.38) 0 Employment 958.60 (60.59) 1,683.87 (60.75) 1 Public program 711.65 (44.98) 1,360.68 (49.09) 0 participation Type of material hardship Did not pay rent or 589.54 (37.27) 1,046.73 (37.76) 0 mortgage Evicted for failure 29.10 (1.84) 43.28 (1.56) 0 to pay rent or mortgage Did not pay gas, 909.64 (57.50) 1,652.42 (59.61) 1 oil or electricity bill Lost gas, oil, or 181.04 (11.44) 301.08 (10.86) 0 electricity for failure to pay Telephone 475.90 (30.08) 783.67 (28.27) 0 disconnected for failure to pay Needed to see doctor 588.07 (37.17) 932.25 (33.63) 0 or go to hospital but did not Needed to see 670.39 (42.38) 1,190.51 (42.95) 0 dentist but did not Material assistance by source (d) Nonprofit 85.53 (5.41) 133.34 (4.81) Family 232.02 (14.67) 352.20 (12.71) 0 Friends 57.77 (3.65) 88.12 (3.18) 0 Government 113.18 (7.15) 189.86 (6.85) 0 No support 1,1943.53 (61.40) 2,073.76 (86.33) (a) The sample size of each panel is unweighted and the percentage is weighted. (b) This column defines a typical case in the sample. The value that each variable takes on is specified. (c) For the two continuous variables, age and number of children, the mean and standard deviation (in parentheses) are reported. (d) Percentages may add up to more than 100% if a household used multiple sources of assistance. Table 2: Results of Rare Events Logistic Regression (N = 6,222) Model 1 Characteristic Coefficient SE Intercept -3.94 *** 0.342 Panel 1996 panel (reference group) 2001 panel 0.336 * 0.169 2004 panel 0.144 0.154 Demographic/household characteristic Race White (reference group) Black 0.013 0.142 Other -0.099 0.248 Sex (female) 0.114 0.164 Age (years) 0.007 0.005 Poverty status (above poverty) -0.342 * 0.137 Education Below high school (reference group) High school 0.227 0.166 Above high school 0.504 ** 0.167 Family type (married couple) -0.205 0.152 Region (rural) 0.274 * 0.134 Homeownership (homeowner) -0.207 0.142 Number of children 0.049 0.044 Public program participation (yes) 0.548 *** 0.142 Employment (employed) 0.128 0.135 Type of material hardship Did not pay rent or mortgage Did not pay gas, oil or electricity bill Lost gas, oil, or electricity for failure to pay Telephone disconnected for failure to pay Needed to see doctor or go to hospital but did not Needed to see dentist but did not Material assistance by source Government Family Friends Model 2 Characteristic Coefficient SE Intercept -5.14 *** 0.381 Panel 1996 panel (reference group) 2001 panel 0.328 0.171 ([dagger]) 2004 panel 0.135 0.155 Demographic/household characteristic Race White (reference group) Black 0.030 0.144 Other -0.139 0.255 Sex (female) 0.084 0.166 Age (years) 0.009 0.005 ([dagger]) Poverty status (above poverty) -0.282 * 0.139 Education Below high school (reference group) High school 0.167 0.169 Above high school 0.420 * 0.167 Family type (married couple) -0.231 0.152 Region (rural) 0.362 ** 0.136 Homeownership (homeowner) -0.150 0.140 Number of children 0.032 0.045 Public program participation (yes) 0.421 ** 0.143 Employment (employed) 0.085 0.137 Type of material hardship Did not pay rent or mortgage 0.731 *** 0.127 Did not pay gas, oil or electricity 1.055 *** 0.167 bill Lost gas, oil, or electricity for 0.405 * 0.159 failure to pay Telephone disconnected for failure 0.057 0.129 to pay Needed to see doctor or go to 0.058 0.150 hospital but did not Needed to see dentist but did not 0.169 0.143 Material assistance by source Government Family Friends Model 3 Characteristic Coefficient SE Intercept -5.155 *** 0.386 Panel 1996 panel (reference group) 2001 panel 0.302 0.170 ([dagger]) 2004 panel 0.104 0.156 Demographic/household characteristic Race White (reference group) Black 0.033 0.145 Other -0.121 0.255 Sex (female) 0.075 0.166 Age (years) 0.010 0.006 ([dagger]) Poverty status (above poverty) -0.0255 0.140 ([dagger]) Education Below high school (reference group) High school 0.179 0.168 Above high school 0.439 * 0.169 Family type (married couple) -0.215 0.154 Region (rural) 0.352 * 0.137 Homeownership (homeowner) -0.141 0.140 Number of children 0.034 0.046 Public program participation (yes) 0.357 * 0.147 Employment (employed) 0.085 0.137 Type of material hardship Did not pay rent or mortgage 0.699 *** 0.130 Did not pay gas, oil or electricity 0.998 *** 0.172 bill Lost gas, oil, or electricity for 0.376 * 0.160 failure to pay Telephone disconnected for failure 0.060 0.134 to pay Needed to see doctor or go to 0.020 0.152 hospital but did not Needed to see dentist but did not 0.178 0.143 Material assistance by source Government 0.551 ** 0.172 Family 0.042 0.168 Friends 0.449 0.245 ([dagger]) ([dagger]) p < .10 * p < .05, ** p < .01, *** p < .0001. Table 3: Predicted Probability (%) Characteristic Model 1 Model 2 Model 3 Poverty Below poverty 6.45 5.20 5.16 Above poverty 4.64 3.90 4.00 Region Urban 4.65 3.94 3.99 Rural 6.03 5.52 5.56 Education Less than high school 2.88 2.64 2.65 High school 3.57 3.09 3.13 Above high school 4.64 3.94 3.96 Public program Participants 7.76 5.92 5.62 Nonparticipants 4.64 3.95 3.98 Did not pay rent/mortgage Yes 7.80 7.76 No 3.90 4.04 Did not pay gas, oil, or electricity bill Yes 3.91 4.01 No 1.40 1.52 Lost gas, oil, or electricity Yes 5.86 5.71 No 3.92 4.03 Government material assistance Yes 6.51 No 3.87 Friend's material assistance Yes 4.17 No 4.00 Note: The predicted probability reported in this table is for a typical case in the sample. The typical case is defined as a households that takes mean values on continuous independent variables and median values on categorical independent variables, except that the examined variable in the first column is undefined (see the last column of Table 1 for the values of each independent variable for a typical case). Table 4: Rare Event Logistic Regression (Interaction Effects) for Model 4 (N = 6,222) Interaction Characteristic Main-Effect Effect (Panel 2) Intercept -6.148 *** (0.823) Panel 1996 panel (reference group) 2001 panel 1.608 (1.077) 2004 panel 1.211 (1.002) Demographic/household characteristic Race White (reference group) Black -0.112 (0.296) 0.124 (0.396) Other 0.219 (0.603) -0.839 (0.974) Sex (female) -0.132 (0.326) 0.054 (0.436) Age (years) 0.036 ** (0.012) -0.032 * (0.016) Poverty status (above -0.346 (0.304) 0.377 (0.407) poverty) Education Below high school (reference group) High school -0.090 (0.330) 0.372 (0.433) Above high school 0.645 * (0.302) -0.495 (0.447) Family type (married couple) -0.140 (0.319) -0.120 (0.438) Region (rural) 0.088 (0.338) 0.107 (0.421) Homeownership (homeowner) -0.320 (0.327) 0.230 (0.417) Number of children 0.049 (0.102) 0.109 (0.127) Public program participation 0.215 (0.311) 0.057 (0.426) (0= no, 1 = yes) Employment (employed) 0.452 (0.296) -0.466 (0.405) Type of material hardship Did not pay rent or mortgage 0.585 * (0.271) 0.341 (0.369) Did not pay gas, oil or 0.736 * (0.329) -0.137 (0.440) electricity bill Lost gas, oil, or 0.820 ** (0.335) -0.720 (0.459) electricity for failure to pay Telephone disconnected for 0.406 (0.256) -0.414 (0.345) failure to pay Needed to see doctor or go 0.323 (0.318) -0.489 (0.432) to hospital but did not Needed to see dentist but -0.273 (0.316) 0.403 (0.413) did not Material assistance by source Government 0.809 ([dagger]) (0.423) 0.094 (0.514) Family -0.079 (0.376) 0.444 (0.479) Friends 1.28 ** (0.477) -0.745 (0.656) Interaction Characteristic Effect (Pane1 3) Intercept Panel 1996 panel (reference group) 2001 panel 2004 panel Demographic/household characteristic Race White (reference group) Black 0.299 (0.369) Other -0.206 (0.679) Sex (female) -0.150 (0.413) Age (years) 0.037 * (0.013) Poverty status (above -0.050 (0.364) poverty) Education Below high school (reference group) High school 0.329 (0.446) Above high school -0.094 (0.432) Family type (married couple) -0.081 (0.384) Region (rural) 0.503 (0.391) Homeownership (homeowner) 0.242 (0.382) Number of children -0.087 (0.126) Public program participation 0.241 (0.372) (0= no, 1 = yes) Employment (employed) -0.466 (0.382) Type of material hardship Did not pay rent or mortgage -0.018 (0.330) Did not pay gas, oil or 0.692 (0.437) electricity bill Lost gas, oil, or -0.462 (0.410) electricity for failure to pay Telephone disconnected for -0.534 (0.326) failure to pay Needed to see doctor or go -0.268 (0.388) to hospital but did not Needed to see dentist but 0.685 ([dagger]) did not (0.379) Material assistance by source Government -0.629 (0.492) Family -0.082 (0.454) Friends -1.32 * (0.603) ([dagger]) p < .10. * p < .05. ** p < .01. *** p < .0001. Table 5: Comparison of Households That Received Nonprofit Assistance and Those That Did Not Receive Assistance Households Households That Received That Did Not Nonprofit Receive Any Assistance Assistance Characteristic (n = 326) (n = 4,907) Race White 63.19 68.11 Black 30.67 25.56 Other 6.04 6.34 Female (a) 75.77 66.66 Household type (married couple ) (a) 34.39 47.92 Poverty status (above poverty) (a) 42.33 59.14 Education (b) Less than high school 21.30 26.30 High school 34.26 35.37 More than high school 44.44 38.33 Region (rural) 30.37 25.46 Homeownership (yes) 30.75 39.69 Number of children (M 1.88 1.73 Employment (yes) 58.28 61.91 Public program participation (a) 64.11 43.43 Did not pay rent or mortgage (a) 57.36 30.36 Evicted for failure to pay rent or mortgage (a) 4.29 1.30 Did not pay gas, oil, or electricity bill (a) 81.60 52.33 Lost gas, oil, or electricity for failure to pay (a) 22.09 7.95 Telephone disconnected for failure to pay (a) 36.20 26.39 Needed to see doctor or go to hospital but did not 32.82 34.95 Needed to see dentist but did not 42.02 43.81 Note: Percentages are reported in the table unless specified otherwise. (a) Difference between the two groups significant at the .001 level. (b) Difference between the two groups significant at the .05 level.…
Questia, a part of Gale, Cengage Learning. www.questia.com
Publication information: Article title: Material Assistance: Who Is Helped by Nonprofits?. Contributors: Guo, Baorong - Author. Journal title: Social Work Research. Volume: 36. Issue: 3 Publication date: September 2012. Page number: 167+. © 1999 National Association of Social Workers. COPYRIGHT 2012 Gale Group.