“I’ll Give You the World”: Socioeconomic Differences in Parental Support of Adult Children

Abstract

Research has shown that parents with higher socioeconomic status provide more resources to their children during childhood and adolescence. The authors asked whether similar effects associated with parental socioeconomic position are extended to adult children. Middle-aged parents (N = 633) from the Family Exchanges Study reported support they provided to their grown children and coresidence with grown children (N = 1,384). Parents with higher income provided more emotional and material support to the average children. Grown children of parents with less education were more likely to coreside with them. Parental resources (e.g., being married) and demands (e.g., family size) explained these patterns. Of interest is that lower income parents provided more total support to all children (except total financial support). Lower income families may experience a double jeopardy; each grown child receives less support on average, but parents exert greater efforts providing more total support to all their children.

Since the start of the 21st century, income inequalities have increased dramatically throughout the world (Marshall, 2014; Piketty & Saez, 2014), and family life has become increasingly disparate by socioeconomic status (SES). Higher SES parents typically invest more time and material resources in young children than lower SES parents do (Conger, Conger, & Martin, 2010; Kornrich & Furstenberg, 2013). Theorists have suggested that parent–child ties also may serve as a vehicle for transmission of inequalities via differential parental resources in adulthood (Swartz, 2009). Indeed, parents provide vital support as young people transition to adulthood (Fingerman, Cheng, Wesselmann, et al., 2012; Furstenberg, 2010; Johnson, 2013). Parents who are better off financially possess a greater capacity to provide material support to their progeny (Swartz, Kim, Uno, Mortimer, & O’Brien, 2011). Similarly, more well-educated parents have better access to information and advice for young adults to gain opportunities for the future. Yet it is not clear whether lower SES parents compensate for lack of material support with other types of assistance, such as practical help. Parental tangible and intangible support may be pivotal in determining a grown child’s own SES and accomplishments (Johnson, 2013; Swartz, 2009). In this study we examined support that parents from different socioeconomic positions provide grown children.

Consistent with prior research, we looked at parental income and education as indicators of resources (Conger et al., 2010; Henretta, Grundy, & Harris, 2002; Johnson, 2013). We examined several types of support and considered the following metrics: (a) support the average grown child in a family receives and (b) the total support a parent provides to all grown children. Total support presents a novel metric. Because of the implications of offspring receiving parental support, research has focused on the amount of support each child receives. Nonetheless, even in families in which support to each child is relatively low, parents with multiple children who need help may offer considerable resources across all children. The interplay of what parents provide and what offspring receive may shed new light on SES differences in parental support.

Socioeconomic Differences in Intergenerational Support

Parents may differentially invest in grown children as a function of SES background. Studies in the United States and Britain have found that parents who have more education, or who are better off financially, provide more financial assistance to adult children (Grundy, 2005; Henretta et al., 2002; Henretta, Wolf, van Voorhis, & Soldo, 2012; Schoeni & Ross, 2005). Yet much of the data addressing this issue are more than 20 years old. For example, using data from the 1980s, Eggebeen and Hogan (1990) reported that parents in poverty provided considerably less financial support, advice, and child care (for children of their grown children) than parents who were better off; only 17% of parents in poverty had engaged in support of a grown child in the past month. Nevertheless, in the 1980s, even upper SES parents did not provide a great deal of support to grown children often; only 45% of upper SES parents reported that they provided support to a child at least once a month (Eggebeen & Hogan, 1990). Similarly, Henretta and colleagues (2002) reported SES differences using data from the late 1980s in the United Kingdom and the early 1990s in the United States and found that parents who were better educated and had higher incomes were more likely to provide financial support and help with chores, but lower and middle SES parents provided more child care.

Parental support of grown children has increased dramatically from the late 1980s into the 21st century (Fingerman, Cheng, Tighe, Birditt, & Zarit, 2012). A majority of parents report providing several types of support (e.g., advice, financial, practical) to a grown child several times a month or more often (Arnett & Schwab, 2013; Fingerman, Cheng, Wesselmann, et al., 2012). Indeed, the transition to adulthood has become prolonged, with upper SES offspring delaying marriage and pursuing education longer, and lower SES offspring often facing uncertain job prospects and childbirth without a spouse (Furstenberg, 2010, 2011). As such, inequities in parental support may have an even greater impact today than in the past. In fact, a study using data from 2001 found that more highly educated parents provided more money for their grown children’s college expenses than less educated parents (Henretta et al., 2012), a key issue in the 21st century given that costs prohibit some young people from pursuing education.

Explanatory Model

The multidimensional intergenerational support model (Birditt & Fingerman, 2012; Fingerman, Sechrist, & Birditt, 2012) provides a framework within which to examine reasons underlying SES disparities in parental support. This model addresses three interrelated factors: (a) providers’ resources, (b) recipient’s needs, and (c) family context.

Social support is partially a function of what a provider can give. A variety of factors favor upper SES parents with more resources to provide adult children. Constraints on time, money, and emotional energy partially determine what a lower SES parent is able to provide a child (Conger et al., 2010).

Several theories also speak to the recipients’ needs in determining intergenerational support. The altruism model in economics (Silverstein, 2006) and contingency theory in sociology posit that offspring increase support to aging parents who incur health problems in late life (Eggebeen & Davey, 1998; Remle, 2011). Likewise, support to grown children may reflect needs. Suitor, Pillemer, and Sechrist (2006) found that aging mothers were more likely to help grown children who were unmarried or experiencing health problems; Fingerman, Miller, Birditt, and Zarit (2009) found a similar pattern of parental support when young adults incurred life problems. Offspring of parents from different SES backgrounds may occupy statuses (e.g., student or parent of young children) that elicit different types of support.

The family context, in particular family size, also may limit the amount of support a particular family member receives. For example, grown children in larger families typically receive less support than grown children in smaller families (Fingerman et al., 2009). However, lower SES parents may provide more total support because of the number of offspring who require support. Prior studies have suggested that middle-aged adults attempt to expand resources when faced with demands from many family members (Fingerman et al., 2011; Grundy & Henretta, 2006). Similarly, lower SES parents may stretch available time to assist a great number of grown children who need help.

In the current study we considered how provider resources, offspring needs, and family context are intertwined with regard to SES differences in parental support of grown children. Moreover, this study is unique because we considered the total support parents provide to all offspring as well as the support they provide to each child in the family.

Types of Support

We also asked whether different types of support vary by parental SES, including material (e.g., money, coresidence) and intangible support (e.g., advice, companionship, technology, child care). Prior research has clearly established that financial support varies as a function of SES (Henretta et al., 2002; Johnson, 2013; Schoeni & Ross, 2005). We considered practical support because research in the 20th century has found SES differences in parental assistance with chores or practical tasks (Grundy, 2005; Henretta et al., 2002).

Child care for grandchildren may also serve as a form of support. A study conducted in Britain found a negative association between education and provision of child care for grandchildren (Henretta et al., 2002); less educated adults provided more child care for their grandchildren. In this study, we expected to find lower SES parents providing more child care.

Parents may also allow grown children to reside with them as a form of support (Aquilino, 2006). Upper SES young adults may coreside with parents (at least part of the year) if they pursue education (Aquilino, 2006; Fry, 2013). By contrast, in lower SES families, offspring may not pursue education and may experience difficulties securing employment or jobs with sufficient income to establish homes of their own (Furstenberg, 2011). National data in the United States indicate that offspring who have less education are more likely to reside with parents after age 25 (Fry, 2013). Thus, we expected lower SES parents to be more likely to offer offspring coresidence.

We examined several types of intangible support. Emotional support may reflect affection in the parent–child tie, but it may not differ by parental SES. Other forms of intangible support may differ by SES. For example, parents with higher education have access to more information and may be better situated to provide advice. They also may spend more time socializing with their grown children. Researchers have linked companionship (e.g., socializing with the other party, lending a listening ear to let a person talk about his or her day) to better psychological and physical health (Rook, August, & Sorkin, 2011). Upper SES parents may embrace cultural norms to spend time with children (Hulbert, 2003; Sayer, Bianchi, & Robinson, 2004), and these patterns may persist after children are grown. Finally, we included technical assistance. Technologies and electronics are increasingly important in daily life (Cotten, McCullough, & Adams, 2012). In current cohorts, young adults often are savvier than their parents when it comes to technology (Lefkowitz, Vukman, & Loken, 2012). Nonetheless, well-educated parents may have pockets of knowledge about technologies. Less educated parents may have less to offer in this domain.

Parental Resources

Resources play a key role in SES differences in support. For example, parents with higher incomes have a greater ability to provide money and materials to grown children. Upper SES adults are more likely to have salaried jobs that require additional hours outside a typical workday (Settersten, 2007), and work overload may preclude upper SES parents from engaging in child care for grandchildren. Yet upper SES adults may have flexibility in their work hours (Bianchi & Milkie, 2010) and may use paid assistance for their own household tasks (e.g., a house cleaner, lawn service). Thus, upper SES parents may have more time to assist their grown children with other tasks, despite working long hours.

Constraints in job schedules may preclude lower SES adults from being involved with grown children (Bianchi & Milkie, 2010). Despite little flexibility in their work hours, however, married working-class mothers and fathers with young children sometimes work different shifts (i.e., one parent works evenings) to cover child care (Barnett & Gareis, 2007). Similarly, some lower SES middle-aged parents may juggle schedules and work shift hours to care for grandchildren while their grown children work (Henretta et al., 2002).

Financial stability also affords emotional energy for providing support (Conger et al., 2010). Unfortunately, lower SES parents are more likely to experience financial and other problems of their own (Everson, Maty, Lynch, & Kaplan, 2002; Prawitz, Kalkowski, & Cohart, 2013). Financial crises are associated with a diminished ability to parent well in early life (Conger et al., 2010), and the same may be true for parenting adult offspring.

In addition, upper SES parents may have social resources that facilitate support to grown children. For example, parents who are married to one another have more tangible and intangible resources to share. The intergenerational stake hypothesis specifies that parents invest in grown children because they view their offspring as a legacy for the future (Giarrusso, Feng, & Bengtson, 2005). Married parents share an investment in their children that fosters support. The presence of a spouse also may mitigate a parent’s own emotional and practical needs, allowing the parent to support grown children. Indeed, studies have found that married parents are more likely to provide money, help with chores, and support to grown children than parents who are divorced, remarried, or single (Aquilino, 2005; Henretta et al., 2012; Kalmijn, 2013). Lower SES adults are more likely to never marry or to divorce than are higher SES adults (Cherlin, 2010; Trail & Karney, 2012).

Family Context of Support

Family context also may help explain why offspring support differs by parental SES. Lower SES parents tend to have a greater number of children (U.S. Census Bureau, 2012). Resource depletion theory suggests children who grow up in larger families receive fewer parental resources because of competing demands from other children in the family (Strohschein, Gauthier, Campbell, & Kleparchuk, 2008). Likewise, in adulthood grown children in larger families receive less parental support and less funding for college than the average child in smaller families (Fingerman et al., 2009; Henretta et al., 2012).

Younger children in the family may also influence parental support of grown children. In the United States, the age of majority and legal decision is 18, often accompanied by graduation from high school and increased independence. Children under age 18 place legitimate demands on parents’ time (Nomaguchi, 2012; Nomaguchi & Milkie, 2003; Umberson, Pudrovska, & Reczek, 2010).

Finally, we considered caregiving for an aging parent as a competing demand. Lower SES families are likely to have parents who experience earlier health declines that require family caregiving (Schoeni, Freedman, & Martin, 2008). Caregiving for an aging parent decreases support to adult children in that family (Fingerman et al., 2011; Grundy & Henretta, 2006).

Offspring Factors That Elicit Support

Offspring from different SES backgrounds also have different needs for support. In the 21st century, upper SES young adults experience a prolonged transition to adulthood associated with continuing parental assistance (Furstenberg, 2011). For example, student status is associated with parental support. Upper SES parents offer financial resources to pay for their children’s higher education (Henretta et al., 2012) and other types of support during college (Attias-Donfut & Wolff, 2000; Fingerman et al., 2009; Fingerman, Cheng, Wesselmann, et al., 2012).

Offspring age also may be a factor in explaining SES differences in parental support. Younger adult offspring receive more support than older adult offspring (Fingerman et al., 2009; Hartnett, Furstenberg, Fingerman, & Birditt, 2013). Upper SES parents typically have children at older ages than lower SES parents (Martinez, Daniels, & Chandra, 2012). As such, upper SES parents may offer their grown children more support because on average, their grown children are younger.

Patterns regarding parental SES, offspring statuses, and parental support are not straightforward. Although youth and student status favor upper SES offspring, other statuses may favor lower SES children. For example, married offspring are less likely to receive parental support than their unmarried counterparts (Sarkisian & Gerstel, 2008). Marriage has become increasingly the jurisdiction of the upper SES in the 21st century (Trail & Karney, 2012). Offspring who divorce or who experience physical or emotional health problems also receive more parental support (Suitor et al., 2006); young adults from lower SES families may be more likely to encounter such life problems (Furstenberg, 2011).

Lower SES offspring also may struggle to find employment or end up in jobs with irregular hours and few benefits. Moreover, in lower SES families grown children have children of their own at younger ages. Offspring who have children are likely to receive more parental support (Bucx, van Wel, & Knijn, 2012). Lower SES grown children may have these children as single parents, rather than with a spouse (Gibson-Davis & Rackin, 2014), and their parents may help with these children.

Total Support to All Children in the Family

Most of the parental support literature is predicated on support to one child, or the average child, in a family, without considering the other children in that family (Johnson, 2013; Silverstein, Conroy, Wang, Giarrusso, & Bengtson, 2002; Swartz et al., 2011). This approach makes sense because grown children’s well-being has been linked to parental support (Fingerman, Cheng, Wesselmann, et al., 2012; Johnson, 2013). Nonetheless, patterns for what parents provide in total may not be the same as patterns for what each offspring receives. Because parents from lower socioeconomic backgrounds typically have more children (U.S. Census Bureau, 2012), they may give more support overall across all children. In addition, according to contingency theory, parents may attempt to respond to the needs of all their grown children despite finite resources, and lower SES offspring may have greater total needs.

Indeed, studies that have examined multiple generations have found that middle-aged parents who face demands from parents and grown children attempt to stretch the amount of support they provide (i.e., resource expansion), but each family member receives less support (Attias-Donfut & Wolff, 2000; Fingerman et al., 2011; Grundy & Henretta, 2006). In this study, we asked whether lower SES parents provide more total support than upper SES parents (although the average offspring of a lower SES parent may not receive as much support).

Prior research has shown that even for financial assistance, lower SES parents provide a comparable proportion of income to their grown children (Kornrich & Furstenberg, 2013); that is, despite limited discretionary funds lower SES parents provide their grown children with 10% of their income, a proportion similar to that provided by upper SES parents (Kornrich & Furstenberg, 2013). With regard to less tangible forms of support, lower SES parents may exert more total effort across multiple children. Few studies have considered these issues.

Other Factors Associated With Support and the Present Study

We also controlled for factors associated with parental support of grown children. For example, mothers usually provide more support than fathers (Rossi & Rossi, 1990). Similarly, daughters typically receive more tangible and intangible social support than sons (Suitor et al., 2006), but sons are more likely to coreside with parents than are daughters (Fry, 2013). Racial differences have been observed in parental support. Although findings are complex, White families tend to provide more support to grown children (but Black families may provide more to aging parents; Fingerman et al., 2011; Sarkisian & Gerstel, 2004; Suitor, Sechrist, & Pillemer, 2007). We controlled for these factors in this study.

In sum, in this study we examined support parents provided to grown children associated with parental SES. We considered support provided to the average grown child in a family as follows.

• SES differences: We expected upper SES parents to provide more advice, financial, practical, and technical support, and lower SES parents to provide coresidence and child care for grandchildren. We did not specify hypotheses regarding emotional support or companionship.

• Resources: We expected parental resources in the form of fewer work hours, fewer financial problems, and being married to be associated with greater parental support.

• Demands and statuses eliciting support: With regard to demands, we considered demands from other grown children (i.e., family size) and caregiving for the parent’s parent. We also expected factors eliciting support—offspring’s age, student status, and parental status—to be associated with more support.

• Total support: Initially, we provide a novel perspective by asking about total support. We hypothesized that lower SES parents provide more total support, even if their average child receives less support than the average child in upper SES families. We also asked whether family size, marital status, and offspring statuses accounted for total support provided.

Method

Participants included 302 men and 331 women age 40–60 (M = 50.60, SD = 4.99 years) from the Family Exchanges Study who resided in the Philadelphia Metropolitan Statistical Area (including suburbs and outlying rural areas) and had at least one child over age 18 (Fingerman et al., 2011). We identified potential participants via purchased lists of phone numbers and addresses from Genesys Corporation. When recruitment occurred in 2007–2008, approximately 93% of adults aged 45 to 65 had landlines (Blumberg & Luke, 2007, 2009) and could be captured by these lists; we also supplemented with random-digit dialing within geographic area codes. We oversampled in Philadelphia County, areas with a high density of racial minorities, and lower SES neighborhoods to obtain a diverse sample (37% of participants identified as ethnic or racial minority members). The response rate was 75%. Data were collected in 2008. Descriptive information about the sample is provided in Table 1.

Table 1.

Demographic Characteristics of Parents and Offspring

Variable	Parent (n = 633)	Offspring (n = 1,384)
	M(SD) or proportion	M(SD) or proportion
Age	50.60(4.99)	25.19(5.80)
Household incomea	4.40(1.45)	3.77(1.61)
Years of educationb	14.18(2.02)	13.73(1.92)
Healthc	3.48(1.07)	4.26(0.93)
Miles from parent		188.66(613.36)
Life problems		0.91(1.20)
Number of children	2.82(1.46)
Number of working hours	44.98(14.41)
Age of the youngest offspring	22.32(4.49)
Total support to all childrend	158.62(147.24)
Average support to each childe		4.32/1.56
Male	.48	.52
Racial minority	.37
Coresides with parent		.23
Coresides with any offspring	.37
Has financial problem	.15
Has children under age 18	.40	.27
Caregiving for aging parent	.24
Work status
Employed full time	.65	.51
Employed part time	.11	.18
Student		.19
Otherf	.24	.12
Marital status
Married for the first time	.63	.17
Remarried	.07	.01
Single/never married	.07	.71
Unmarried/otherg	.22	.11

^aHousehold income in 2007: 1 = less than $10,000, 2 = $10,001–$25,000, 3 = $25,001–$40,000, 4 = $40,001–$75,000, 5 = $75,001–$100,000, and 6 = more than $100,000; numbers are based on reports from 592 offspring participants for offspring household income.

^bBecause 19% of young adults were students at the time of the study, their ultimate educational attainment may be higher than reported here.

^c1 = poor, 2 = fair, 3 = good, 4 = very good, 5 = excellent.

^dNumber of days out of 365 provided support to each child, summed across all children in the family.

^e1 = less than once a year or not at all, 2 = once a year, 3 = a few times a year, 4 = monthly, 5 = a few times a month, 6 = weekly, 7 = a few times a week, 8 = daily.

^fRetired, disabled, unemployed, or homemaker.

^gWidowed, divorced or separated, and cohabiting.

Procedure

The surveys utilized computer-assisted telephone interviews. Interviews lasted approximately 1 hour. Participants responded to questions about their own background (e.g., age, education, household income, and marital status).

Participants answered questions about each of their grown children. Participants had 1,785 children, of whom 1,384 were over age 18 (M = 25.19, SD = 5.80). Offspring ranged in age from 18 to 46 years; 90% of offspring were young adults age 18–33. Because the upper age range of offspring extended into midlife, we examined offspring age 18–33 in the models and then repeated the models with the full age range of offspring. The pattern of findings was identical, and therefore we included the full age range of offspring.

Parental socioeconomic background

To assess parental SES, we used the following three variables: (a) household income for 2007, (b) years of education, and (c) education level by degree (e.g., some high school, high school degree, some college, college degree, and postgraduate education). The average household income was similar, but education was higher than in the population of the greater Philadelphia Metropolitan Statistical Area (U.S. Census Bureau, 2008). Correlations for these variables were as follows: (a) income and years of education, r = .46; (b) income and education by degree, r = .45; and (c) years of education and education by degree, r = .97, p < .001. Because years of education and degree were redundant, we examined income and years of education.

Support

Participants completed the Intergenerational Support Index (ISI; Fingerman et al., 2009) rating the frequency with which they provided each child with seven types of support, ranging from 1 (once a year or less often) to 8 (daily): (a) financial support, (b) practical support, (c) advice, (d) companionship, (e) technical support, (f) listening to talk about daily life, and (g) emotional support. This measure was developed on the basis of theory regarding social support, including tangible and intangible support and companionship (Antonucci, 2001; Vaux, 1988; Vaux & Harrison, 1985). Previously, the ISI has not included technical support, but we did so here because technical support may differ by parental SES. We calculated mean scores of the seven items (α = .88) and refer to that index as overall support to each child. Participants responded in regard to support they personally provided. (For a complete list of social support questions, see the Appendix.)

In addition, we asked about the amount of money parents gave each child, using a bracketing technique to generate an ordinal variable (e.g., Johnson, 2013; Juster & Willis, 2006). We asked, “Including education, but not shared housing or shared food, have you given [child’s name] financial help totaling $100 or more in the past 12 months?” If “yes,” parents indicated whether they provided at least $500 to that child, and likewise for $1000. Responses were classified: 1 = “less than $100,” 2 = “$101 to $500,” 3 = “$501 to $1,000,” and 4 = “more than $1,000” (see Appendix).

For offspring who had children of their own, we asked whether parents provided child care to these grandchildren and, if so, how often (also rated 1 = once a year or less often to 8 = daily). Seventy-three parents (11.5%) reported providing child care for their grandchildren (the children of 94 grown children). This assessment was considered as an outcome in analyses. Furthermore, participants indicated whether they were raising a grandchild in a surrogate parental role. Only 26 participants (4%) endorsed this item. Participants raising grandchildren had lower income (M = 3.81, or $25,001–$40,000), but we did not further analyze this variable.

Finally, parents indicated whether each grown child had resided in the same household (other than vacations or visits) for more than 3 weeks in the past year. This question was based on prior studies assessing intergenerational coresidence (Sweet, Bumpass, & Call, 1988).

Parent resources

Participants indicated how many hours they spend on paid work activities (including commuting). Most participants (75.6%) were employed, and they worked 33.50 hours (SD = 23.23) on average. We considered employment hours for pay as a continuous variable, treating participants not employed for pay as “zero” hours; the recoded variable was highly correlated with work status (r = .83 between hours working and employed or not for pay).

We asked whether participants had experienced financial problems in the past 12 months (1 = yes and 0 = no); 98 participants (15.5%) reported serious financial problems. We coded marital status as 1 (married to other parent) and 0 (divorced or never married to other parent).

Family context

Parents reported their total number of children. On average, they had 2.82 children (SD = 1.46), including 2.19 children over age 18 (SD = 1.24, range: 1–10). We considered the number of children in the family as a possible competing demand on parental resources. Moreover, children under age 18 in the home may place legitimate demands on parental time. Because number of children under 18 was correlated with total number of children, we coded the presence of children under age 18 dichotomously (1 = children under age 18 present in the home, 0 = no children under age 18).

We assessed caregiving for an aging parent (1 = caregiving for a parent and 0 = not caregiving for a parent). Participants answered five questions regarding parents’ potential disability in activities of daily living (e.g., with bathing, housework, transportation). Caregiving was defined as helping a parent with a disability with these tasks of daily living several times a month; 23.9% of the sample indicated they did such caregiving.

Offspring statuses and situations eliciting support

We considered student status as a factor that elicits support (1 = student, 0 = not student). We considered offspring age, marital status (1 = married, 0 = not married), and parental status (1 = parent, 0 = not parent). We also included 10 items assessing life problems children had experienced in the past 2 years (e.g., divorce, health problems, victim of a crime, loss of a job). Consistent with prior studies, we used the total number of problems (Fingerman et al., 2009; Greenfield & Marks, 2006).

Covariates

Covariates included parent gender (1 = father and 0 = mother), parental age, offspring gender (1 = son and 0 = daughter), and parent racial minority status (1 = racial minority and 0 = non-Hispanic White; parent and offspring racial minority status were nearly identical).

Analytic Strategy

Preliminary analyses examined bivariate correlations between variables. Then, to test hypotheses regarding how much the average grown child receives as a function of his or her parents’ SES, we estimated multilevel models using the PROC MIXED procedure in SAS. These models take into account the nested structure of the data. Participants reported on each of their grown children. Multilevel models handle correlated error in responses for multiple children as well as unequal numbers of children nested within each parent.

We estimated analyses for overall support to each child using the ISI as well as each type of support separately (e.g., emotional, practical, advice) as outcomes (eight models). We also considered the amount of money parents provided to each child (one model). We considered child care with models estimated two ways, including (a) only those offspring who had children of their own and (b) all offspring, coding offspring who did not have children with frequency as 1 (less than once a year or not at all); the pattern of findings was the same in both models. Because coresidence was measured as a binary variable, we ran multilevel logistic regression models using the PROC GLIMMIX procedure in SAS for coresidence. In all models, we included the following control variables: parent gender, parent age, parent racial minority status, and offspring gender.

The independent variables were parental income and years of education. We initially included the interaction of income and education to ask whether parents who are both well educated and have a high income are particularly likely to support their children. The interaction term was not significant in any models. Thus, we considered only the main effects of income and education entered as main effect terms in the models.

If the pattern of findings was the same when education and income were both presented in the model as when they were entered in separate models, we included both indicators of SES in the same model (e.g., the multilevel models predicting average support to each child). The pattern of findings differed when income and education were entered simultaneously in the regressions examining total support, however. Therefore, in the regressions, we presented findings for income and education in separate models.

We considered three types of explanatory variables: (a) parental resources (e.g., parents’ work hours and marital status), (b) family context (e.g., family size, children under age 18, parental caregiving), and (c) offspring statuses eliciting support (e.g., offspring student, marital, and parental status, offspring age). To assess the explanatory variables we tested indirect effects via bootstrapping (Preacher & Hayes, 2008). Bootstrapping is a nonparametric method; estimates are derived through many repetitions of resampling (Kenny, 2013). The indirect effect is computed by multiplying Path A (i.e., effects of SES indicators on explanatory variables) × Path B (i.e., effects of explanatory variables on support) and generating a sampling distribution empirically. With the distribution, a confidence interval and a p value can be determined; that is, whether the indirect effect is significantly different from zero. The coefficients for the indirect effect indicate the magnitude of reduction of the effects of the independent variable on the outcome. Here, we calculated SES effects on parental support through explanatory variables (i.e., indirect effects) using 10,000 bootstrapped samples with Mplus. This bootstrapping technique does not require a significant effect between the initial predictor (i.e., education, income) and the outcome (i.e., support), to detect an indirect effect through another variable.

A second goal of the study was to examine the total amount of support parents of differing socioeconomic backgrounds provided to all their children combined. In this study, participants rated the frequency of support given to each child using a standard approach from the literature with categorical temporal increments that ranged from daily to once a year or less often (Fingerman et al., 2011; Rossi & Rossi, 1990; Silverstein et al., 2002). To generate a summed total for support across children in the family, we converted the rating scale to units of 365 days—that is, once a year = 1, once a month = 12, once a week = 52, and so forth—to generate a number of “support days” parents provided per year. Then, we summed ratings for all children in the family. These totals are imprecise measures because a parent might spend an entire day (scored as 1 day) helping a child move or 15 minutes in that day listening to the child discuss a problem (scored as 1 day). Moreover, a parent could help two children on the same day and get scored for that day twice. Nonetheless, these totals provide an estimate of a parent’s support across all grown children in the family. We refer to this index as total support in the family.

We used linear regressions at the participant level to examine total support. We looked at total support for the combined ISI and each type of support separately. Parental income and education served as the independent variables. Parental gender, age, and racial minority status served as control variables. We generated a categorical control variable for offspring gender that incorporated the composition of gender in the family (i.e., all daughters, all sons, mix of sons and daughters) treating “all daughters” as the comparison category.

We considered parental marital status, family size, and presence of children under age 18 as parent-level explanatory variables for significant SES differences in the total support provided in the family. The majority of parents (68.1%) had more than one child, and we considered the total number of children occupying different statuses as explanatory variables (e.g., total number of grown children who were students, total number of grown children who were not married). We also included the age of the youngest child as an index of grown children’s ages in these models. We used the bootstrapping technique described previously to test the indirect effects of the explanatory variables on total support.

Results

Bivariate Associations Involving Parental SES

We first examined bivariate associations for parental income and years of education with explanatory variables. Compared to parents with lower income or less education, higher SES parents worked longer hours (r = .26, p < .001 for income, r = .13, p < .01 for education) and were less likely to report a serious financial problem in the past year (r = −.39, p < .001 for income, r = −.15, p < .001 for education). Higher SES parents were more likely to be married (r = .49, p < .001 for income, r = .11, p < .01 for education). They were less likely to provide care for a parent with a disability (r = −.18, p < .001 for income, r = −.13, p < .001 for education) and had smaller families (r = −.10, p < .05 for income, r = −.12, p < .001 for education). Higher SES parents were more likely to have children who were students (r = .21, p < .001 for both income and education), less likely to have children who were parents themselves (r = −.23, p < .001 for income, r = −.18, p < .001 for education), and their grown children were younger (r = −.18, p < .001 for income, r = −.15, p < .001 for education). Offspring problems were not correlated with parental SES.

For control variables, parental SES indicators (e.g., income and education) were not correlated with parental gender, age, or offspring gender. Racial minority parents had lower income (r = −.44, p < .001) and fewer years of education (r = −.17, p < .001) than non-Hispanic White parents.

Parental SES and Support to Each Child

We asked whether support (the combined ISI and each type of support) varied by parental income and education. Multilevel models examined support provided to the average child within each family.

Multilevel models for overall support were significant for income, with parents of higher income providing more frequent overall support to the average offspring (see Table 2). For specific types of support, parents with higher income and better education provided more frequent financial support and also provided a greater amount of money. Parents with higher income also listened to their children talk about daily life and provided more frequent technical help. Coresidence was significantly associated with parental education; parents with less education were more likely to have offspring who resided with them.

Table 2.

Multilevel Models for Effects of Parental Income and Education on Support to Each Child and Coresidence

Model variables	Overall supporta		Technical supporta		Listening to talk about daily lifea		Financial supporta		Amount of financial supportb		Coresidencec
	B	SE	B	SE	B	SE	B	SE	B	SE	B	SE
Fixed effects
Intercept	8.17***	0.60	3.11***	0.79	8.65***	0.78	8.65***	0.84	3.55***	0.44)	4.28***	0.94
Parent incomed	0.09*	0.04	0.11*	0.06	0.15**	0.05	0.24***	0.06	0.24***	0.03)	0.08	0.07
Parent years of education	0.02	0.03	0.04	0.04	0.00	0.04	0.11**	0.04	0.09***	0.02)	−0.13**	0.04
Control variables
Parent gendere	−0.22*	0.10	0.79***	0.13	−0.75***	0.13	−0.07	0.14	0.09	0.07)	−0.25	0.16
Parent age	−0.08***	0.01	−0.05***	0.01	−0.05***	0.01	−0.14***	0.01	−0.06***	0.01)	−0.08***	0.02
Parent minority statusf	−0.21	0.12	0.43**	0.15	−0.52***	0.15	−0.16	0.16	−0.20*	0.08)	−0.39*	0.18
Offspring gendere	−0.28***	0.07	−0.06	0.07	−0.42***	0.10	−0.13	0.10	−0.10	0.06)	0.37*	0.15
Random effects
Intercept variance	0.81***	0.09	1.99***	0.16	1.22***	0.14	1.68***	0.18	0.36***	0.05)
Residual variance	1.37***	0.07	1.12***	0.06	2.54***	0.13	2.48***	0.13	0.87***	0.04)	0.43*	0.18
−2 log likelihood	4,655.8		4,820.8		5,408.9		5,483.1		3,926.3		5,342.0

Note. Six models showing significant effects of parental income and/or education are presented here. There were no significant differences by income and/or education for emotional support, practical support, companionship, advice, and child care.

^aCoded 1 = less than once a year or not at all, 2 = once a year, 3 = a few times a year, 4 = monthly, 5 = a few times a month, 6 = weekly, 7 = a few times a week, and 8 = daily.

^bCoded 1 = less than $100, 2 = $101–$500, 3 = $501–$1,000, and 4 = more than $1,000.

^c1 = living in the same household and 0 = living in separate households.

^dHousehold income in 2007; 1 = less than $10,000, 2 = $10,001–$25,000, 3 = $25,001–$40,000, 4 = $40,001–$75,000, 5 = $75,001–$100,000, and 6 = more than $100,000.

^e1 = male and 0 = female.

^f1 = racial minority and 0 = non-Hispanic White; a multilevel logistic model was estimated.

^*p < .05.

^**p < .01.

^***p < .001.

There were no significant differences by parental SES in provision of practical support, emotional support, advice, companionship, or in parental child care (Supplementary Table 1 on the Journal of Marriage and Family [JMF] website; http://onlinelibrary.wiley.com/journal/10.1111/(ISSN)1741–3737).

Explanatory Factors Associated With Parental Support to Each Child

Parental resources, family context, and offspring statuses

To explain SES differences in support given to each child, we considered parental resources (i.e., hours at work, financial problems, marital status), family competition for support (i.e., family size, presence of children under age 18, caregiving for an aging parent), and offspring situations eliciting support (i.e., student, marital, and parental status, age and life problems). We estimated models for the overall ISI and for each type of support significantly associated with SES in prior analyses (i.e., Table 2), examining indirect effects of SES via these variables on parental support using bootstrapping (Preacher & Hayes, 2008).

As can be seen in Table 3, three indicators of parental resources and competing demands were significantly associated with overall support to the average grown child in a family: (a) marital status (married parents gave more), (b) smaller family size, and (c) having a child under age 18. Furthermore, parental marital status contributed to the indirect effect of income on overall support, listening to talk about daily life, and coresidence and to the indirect effect of education on overall support and coresidence. In addition, hours parents worked contributed to an indirect effect of income for listening to talk about daily life. For financial and technical support, parental explanatory variables did not contribute to indirect effects of income or education.

Table 3.

Multilevel Models With Parental Resources and Offspring Statuses as Explanatory Variables for Socioeconomic Differences in Support and Coresidence

Model variables	Overall supporta		Technical supporta		Listening to talk about daily lifea		Financial supporta		Amount of financial supportb		Coresidencec
	B	SE	B	SE	B	SE	B	SE	B	SE	B	SE
Fixed effects
Intercept	8.94***	0.63	3.28***	0.90	9.82***	0.83	8.17***	0.86	3.58***	0.46	4.65***	1.18
Parent incomed	0.00	0.05	0.05	0.07	0.02	0.06	0.12	0.06	0.18***	0.03	−0.01	0.09
Parent years of education	−0.05	0.03	−0.02	0.04	−0.06	0.04	−0.02	0.04	0.04	0.02	−0.18***	0.05
Explanatory variables
Parent hours of working	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00
Parent financial problemse	0.04	0.14	0.07	0.21	0.18	0.19	0.08	0.19	−0.06	0.10	0.22	0.26
Parent marital statusf	0.34**	0.11	0.05	0.16	0.35*	0.15	0.42**	0.15	0.11	0.08	0.74***	0.22
Parent number of children	−0.16***	0.03	−0.07	0.05	−0.18***	0.05	−0.18***	0.05	−0.10***	0.03	−0.03	0.06
Parent child under age 18g	0.01	0.12	0.10	0.17	−0.07	0.15	0.24	0.16	0.00	0.09	−0.07	0.21
Parent care for aging parenth	−0.13	0.11	−0.36*	0.16	−0.27	0.15	0.02	0.15	−0.06	0.08	−0.16	0.21
Offspring student statusi	0.09	0.11	0.10	0.12	−0.07	0.15	1.13***	0.14	0.39***	0.08	−0.94***	0.22
Offspring marital statusj	−0.26*	0.12	−0.31*	0.12	−0.06	0.17	−0.77***	0.16	−0.49***	0.09	−2.55***	0.62
Offspring parent statusk	0.14	0.10	−0.06	0.11	0.04	0.15	0.09	0.14	−0.05	0.08	−0.30	0.26
Offspring age	−0.11***	0.01	−0.04***	0.01	−0.11***	0.01	−0.13***	0.01	−0.05***	0.01	−0.14***	0.03
Offspring life problems	0.07*	0.03	0.02	0.04	−0.01	0.04	0.18***	0.04	0.11***	0.02	0.03	0.07
Control variables
Parent genderl	−0.30**	0.10	0.70***	0.14	−0.81***	0.13	−0.29*	0.13	−0.01	0.07	−0.47**	0.18
Parent age	−0.01	0.01	0.00	0.02	0.01	0.02	−0.02	0.02	−0.01	0.01	0.00	0.02
Parent minority statusm	0.16	0.11	0.60***	0.16	−0.15	0.15	0.37*	0.15	0.04	0.08	0.00	0.21
Offspring genderl	−0.33***	0.07	−0.14	0.08	−0.45***	0.10	−0.23*	0.09	−0.16**	0.06	0.29	0.16
Random effects
Intercept variance	0.68***	0.08	2.06***	0.16	0.97***	0.13	1.31***	0.15	0.27***	0.04	—	—
Residual variance	1.06***	0.06	0.98***	0.06	2.22***	0.13	1.83***	0.11	0.71***	0.04	0.62**	0.22
−2 log likelihood	3,866.2		4,248.1		4,623.5		4,546.7		3,243.4		5,786.5

^aCoded 1 = less than once a year or not at all, 2 = once a year, 3 = a few times a year, 4 = monthly, 5 = a few times a month, 6 = weekly, 7 = a few times a week, and 8 = daily.

^bCoded 1 = less than $100, 2 = $101–$500, 3 = $501–$1,000, and 4 = more than $1,000.

^c1 = living in the same household and 0 = living in the separate households; a multilevel logistic model was estimated.

^dHousehold income in 2007; 1 = less than $10,000, 2 = $10,001–$25,000, 3 = $25,001–$40,000, 4 = $40,001–$75,000, 5 = $75,001–$100,000, and 6 = more than $100,000.

^e1 = have financial problems, 0 = no financial problems.

^f1 = remarried or married, 0 = not married.

^g1 = have children under age 18, 0 = no children under age 18.

^h1 = caregiving for aging parent, 0 = not caregiving for aging parent.

ⁱ1 = student, 0 = not a student.

^j1 = remarried or married, 0 = not married.

^k1 = have children, 0 = no children.

^l1 = male, 0 = female.

^m1 = racial minority, 0 = non-Hispanic White.

^*p < .05.

^**p < .01.

^***p < .001.

Regarding offspring statuses that may elicit support, offspring student status, marital status, age, and life problems were associated with different types of support to the average child. Moreover, bootstrapping techniques revealed that student status contributed to an indirect effect of income and education on overall support, financial support, and listening to talk about daily life. Offspring marital status, parent status, and age also explained the effects of education on overall support, technical support, listening to talk about daily life, financial support, and coresidence.

Parental SES and Total Support to All Children

Next, we estimated regressions using total support parents provided to all children in the family. These outcomes were calculated from the metric of number of days out of 365 days a year that parents provided support to each child, summed across all children within the family. We estimated 10 regressions for total support: the overall ISI, each type of support, child care, and total number of coresident children over age 18. We included three parent control variables (age, gender, and racial minority status) and two dichotomous codes for offspring gender in the family (i.e., all sons and mix of sons and daughters, with all daughters as a comparison category).

When we entered income and education simultaneously, the indicators of SES were significant in only two regressions (total days giving financial support and total days of companionship). When we entered these SES indicators separately, however, income was significantly associated with total frequency of financial support, emotional support, practical support, advice, and companionship. The totals for listening to children talk about their day and providing technical support were not associated with income. When education was entered into the regression separately, education was associated with total support of the same types, with the exception of advice. For reasons of parsimony, Table 4 includes findings from income predicting total support in the family, and Supplementary Table 2 on the JMF website presents parallel findings for education.

Table 4.

Regressions for Effects of Parental Income on Total Support to All Children

	Financial support		Emotional support		Practical support		Advice		Companionship
	B	SE	B	SE	B	SE	B	SE	B	SE
Intercept	372.79***	81.15	606.75***	126.20	393.14***	74.90	578.79***	93.54	379.95***	73.74
Parent incomea	16.61**	5.67	−18.80*	8.81	−11.83*	5.23	−14.08*	6.53	−24.48***	5.15
Control variables
Parent genderb	−32.93*	14.78	−104.47***	22.99	−29.42*	13.64	−39.63*	17.04	−2.33	13.43
Parent age	−6.15***	1.52	−4.35	2.36	−4.58***	1.40	−7.13***	1.75	−3.68**	1.38
Parent minority statusc	−11.38	17.00	0.47	26.44	−4.89	15.69	36.84	19.60	1.65	15.45
All sonsd	−18.67	19.76	−78.87*	30.74	−19.45	18.24	−33.62	22.78	−9.00	17.96
Mix of sons and daughtersd	16.97	18.81	94.31***	29.26	51.82**	17.36	79.12***	21.68	69.94***	17.09
F	5.79***		11.63***		6.45***		11.02***		10.38***
Adjusted R2	.05		.10		.05		.09		.09

^aHousehold income in 2007; 1 = less than $10,000, 2 = $10,001–$25,000, 3 = $25,001–$40,000, 4 = $40,001–$75,000, 5 = $75,001–$100,000, and 6 = more than $100,000.

^b1 = male, 0 = female.

^c1 = racial minority, 0 = non-Hispanic White.

^dComparison category = all daughters.

^*p < .05.

^**p < .01.

^***p < .001.

The data in Table 4 show that higher income was associated with total frequency of financial support, but lower income was associated with the total frequency of providing emotional support, practical support, advice, and companionships across all offspring.

Resources and demands for total support

We considered explanatory variables for significant SES differences in total support: parent marital status, family size, and presence of children under age 18. We also included offspring statuses in these models as explanatory variables using totals: the total number of offspring who were students, who were married (i.e., expecting a negative effect because non-married offspring receive more support), who were parents, and age of youngest grown child. We estimated these models for each type of support associated with SES (i.e., significant findings in Table 5 for income and Supplementary Table 3 on the JMF website for education).

Table 5.

Regressions With Explanatory Variables for Effects of Parental Income on Total Support to All Children

	Financial support		Emotional support		Practical support		Advice		Companionship
	B	SE	B	SE	B	SE	B	SE	B	SE
Intercept	248.69**	92.74	756.63***	145.29	411.15***	87.37	600.37***	109.31	390.88***	85.85
Parent incomea	−0.48	6.05	−30.45**	9.48	−23.26***	5.70	−22.12**	7.13	−31.14***	5.60
Explanatory variables
Parent marital statusb	50.63**	16.96	56.14*	26.58	60.67***	15.98	33.34	20.00	51.58**	15.70
Parent number of children	4.73	6.51	43.85***	10.19	12.00	6.13	29.26***	7.67	19.43**	6.02
Parent child under age 18c	7.52	18.64	−135.55***	29.20	−20.59	17.56	−66.00**	21.97	−37.91*	17.25
Number of student offspring	55.49***	13.00	−1.90	20.36	5.94	12.25	4.71	15.32	10.13	12.03
Number of married offspring	−28.69*	12.74	−25.64	19.95	−4.97	12.00	−20.90	15.01	−27.55*	11.79
Number of offspring who are parents	1.37	12.17	34.27	19.07	23.03*	11.47	19.37	14.34	42.22***	11.26
Age of the youngest offspring	−6.46**	2.24	−14.11***	3.51	−7.85***	2.11	−8.84***	2.64	−3.41	2.07
Control variables
Parent genderd	−42.20**	14.31	−109.42***	22.42	−31.84*	13.48	−45.18**	16.87	−2.94	13.25
Parent age	−0.46	1.86	−1.44	2.91	−1.70	1.75	−3.99	2.19	−3.12	1.72
Parent minority statuse	10.58	17.76	−5.82	27.83	−0.07	16.73	34.41	20.94	−12.50	16.44
All sonsf	−25.78	18.88	−80.78**	29.59	−21.55	17.79	−37.06	22.26	−7.90	17.48
Mix of sons and daughtersf	8.58	19.86	10.87	31.11	20.06	18.71	26.74	23.41	32.44	18.38
F	8.67***		10.75***		6.81***		8.83***		8.97***
Adjusted R2	.14		.17		.11		.14		.15

^aHousehold income in 2007; 1 = less than $10,000, 2 = $10,001–$25,000, 3 = $25,001–$40,000, 4 = $40,001–$75,000, 5 = $75,001–$100,000, and 6 = more than $100,000.

^b1 = remarried or married, 0 = not married.

^c1 = have children under age 18, 0 = no children under age 18.

^d1 = male, 0 = female.

^e1 = racial minority, 0 = non-Hispanic White.

^fComparison category = all daughters.

^*p < .05.

^**p < .01.

^***p < .001.

As can be seen in Table 5, explanatory variables were associated with income differences in total financial, emotional, and practical support as well as advice and companionship. Bootstrapping techniques indicated that indirect effects of income occurred through parental marital status for financial, practical, and emotional support and for advice and companionship. Similar patterns were evident for the age of the youngest adult child. Bootstrapping techniques also showed indirect effects of income on companionship through the number of children who were parents. There was an indirect effect for the number of offspring who were students and financial support. Analyses with explanatory variables and parental education manifested a similar pattern (see Supplementary Table 3 on the JMF website).

Discussion

As the gap between “haves” and “have-nots” widens in the United States, important questions arise regarding how these patterns may be transmitted across generations in families. Parental support may benefit young people as they establish a foothold in adulthood (Fingerman, Cheng, Wesselmann, et al., 2012; Johnson, 2013). In the current study we examined many types of intangible assistance and considered support to multiple children in the family. Moreover, we developed an estimate of total support across all children and found contradictory patterns regarding SES and (a) parental support to the average child and (b) total support parents provide across all their children.

Parental SES played a key role in support parents provided adult offspring. As in prior research (e.g., Eggebeen & Hogan, 1990; Grundy, 2005), parents who had higher incomes or better education gave more material resources to the average offspring than lower SES parents (e.g., Johnson, 2013; McGarry & Schoeni, 1997; Remle, 2011). By contrast, for the estimate of total support to all children, lower SES parents provided more of several types of intangible support (upper income parents provided more total financial support). Patterns of support to grown children may be detrimental in lower income families because each offspring receives less, but parents are taxed more providing intangible support to all offspring.

Income and Education Disparities in Parental Support to Offspring

Higher parental SES was associated with each offspring receiving more support, specifically, financial support and intangible support such as help with technologies and listening to talk about daily life. Prior studies have shown parents who have more income provide more money to grown children (e.g., Johnson, 2013; McGarry & Schoeni, 1997). In this study we found that upper SES parents gave financial support more frequently and in greater amounts than lower SES parents. Financial support may allow grown children to pursue education and invest in the future in concrete ways (Furstenberg, 2011; Remle, 2011).

Technical support occurred less frequently than other types of support (Fingerman et al., 2009), perhaps because people do not need help with computers or buying new electronics on a daily basis. Moreover, young adults may offer their parents technical support more often than parents offer them support (Cheng, Birditt, Zarit, & Fingerman, 2013). Nonetheless, SES disparities in technical support from parents may have repercussions for young adults because technologies are important in education, career advancement, and communication.

Upper SES parents also were more likely to listen to their grown children talk about the minutiae of daily life. Researchers have found that adolescents show better adjustment when parents listen to them (Agliata & Renk, 2008). Upper SES parents typically engage in more verbal exchanges with their children during childhood than do lower SES parents (Huttenlocher, Waterfall, Vasilyeva, Vevea, & Hedges, 2010; Rowe, 2008). The greater frequency of listening to grown children may represent continuity from these earlier patterns. Future studies might include observational methods that permit comparisons of conversations among parents and grown children of different SES backgrounds. Future studies also might consider the implications of parental listening for offspring well-being.

Coresidence was associated with SES but in the opposite direction of other types of support; parents with less education were more likely to coreside with a grown child. These findings are consistent with increasing costs of housing and difficulties obtaining employment in lower SES families (Fry, 2013).

Explanations for Disparities in Support

Parental and family context factors contributed to SES differences. Parents with more income and education were more likely to be married to their children’s other parent, and marital status helped explain SES disparities in overall support. Marriage to the child’s other parent may increase shared investment in the child as well as fostering more resources to distribute (Henretta et al., 2012; Kalmijn, 2013). Married parents share a generational stake in their grown children (Giarrusso et al., 2005) and also may support one another and thus have greater emotional reserves to assist grown children. Likewise, married parents may be more economically secure. Although many studies ask married parents to report their own support to grown children (e.g., Aquilino, 2005; Suitor et al., 2006), some studies have assessed support from the couple (e.g., Grundy, 2005), and other studies have found that grown children do not report the same amount of support from both parents (e.g., Johnson, 2013). Future research might ascertain whether some types of support stem from a couple (e.g., financial support), whereas other types of support stem from an individual parent (e.g., listening). Indeed, SES disparities in marital patterns of midlife parents may be more pronounced in future cohorts based on current SES differences in marriage and childrearing among young adults (Gibson-Davis & Rackin, 2014).

The grown child’s statuses also helped explain parental SES differences. Parents with higher income and better education were more likely to have grown children who were students. Consistent with prior studies, students received more parental support than nonstudents (Attias-Donfut & Wolff, 2000; Fingerman, Cheng, Tighe, et al., 2012). Student status contributed to findings regarding lower SES biases in coresidence. Students are less likely to live with parents during the academic year (Pryor et al., 2012). Lower SES grown children may forego higher education or attend a local institution such as a community college and thus coreside with parents (Pryor et al., 2012).

This study was limited because we did not include offspring income. Offspring who are students may obtain higher income in the future, but research that examines current and future projected income among offspring may shed further light on how offspring differences in income are associated with parental SES differences in support.

Furthermore, and consistent with the results of past studies (e.g., Hartnett et al., 2013), younger offspring were more likely to receive parental support. Other factors, such as offspring marital and parental status explained only some effects of income and education. The lack of explanatory power may partially reflect complexities in family life. Married offspring typically receive less support than unmarried offspring (Sarkisian & Gerstel, 2008). Parents of young children typically receive more support from their own parents (Bucx et al., 2012). In this study, most of the grown children who had children of their own were married. As such, the effects of the two variables may have been a wash. In samples with young single parents, findings may be different.

Total Support to All Children

The pattern for total support to all children differed from the pattern regarding support to each child. Estimates in this study suggest that lower SES parents provide more total support associated with time (e.g., practical support, advice, and companionship). Yet, consistent with the literature regarding financial support (e.g., Henretta et al., 2002; Johnson, 2013; McGarry & Schoeni, 1997), upper SES parents provide more total financial support. Thus, lower SES parents spend more time on total intangible support, but each child receives less.

Furthermore, and consistent with prior work, the average grown child in larger families received less support (Fingerman et al., 2009). Our findings for total support were consistent with resource expansion, however, whereby parents attempt to expand their resources to meet the needs of multiple family members (e.g., Attias-Donfut & Wolff, 2000; Fingerman et al., 2011; Grundy & Henretta, 2006). Family size accounted not only for what each child received (less in lower SES and larger families) but also in the greater total expenditure of intangible support from lower SES parents. Although some studies have explicitly asked aging parents about their preferences and support decisions among multiple children (e.g., Suitor et al., 2006), future research should pursue these issues among parents of different socioeconomic backgrounds. Perhaps lower SES parents are less likely to distinguish among grown children as they attempt to distribute support widely and thus provide considerable total support across children.

Suggestions for Future Research

Future research should pursue additional issues regarding SES differences in parental support. For example, measurement of SES was derived from only two indicators: parental education and income. Yet some prior research has shown that parental homeownership is associated with greater financial transfers to grown children (McGarry & Schoeni, 1997). Similarly, coding of occupational prestige in the United Kingdom and the United States has linked more prestigious jobs to greater financial transfers (Henretta et al., 2002). We detected the same phenomenon in this study, suggesting that parental income is robust in explaining transfers to offspring. Income disparities have increased dramatically in the United States over the past decade and may be a source of disparities in other SES indicators in future generations.

The study response rate was comparable to similar studies (e.g., Suitor et al., 2006), but was relatively high given the wide SES variability and high racial minority participation. Nonetheless, similar to prior studies of intergenerational ties (e.g., Rossi & Rossi, 1990; Silverstein et al., 2002; Suitor et al., 2006), the study was based in one geographic area. This area included rural residences, but it is not clear whether patterns of intergenerational transfers are distinct in urban and rural areas. Moreover, the most impoverished adults may be transient, and their place of residence may vary over time. Future research should examine geographic contexts associated with variability in SES.

The assessment of overall family support also provides only a rough estimate of time spent assisting grown children. Time use diaries could provide a more accurate estimate.

This study suggests that lower income parents may experience a double jeopardy in launching their children into adulthood because they spend more time giving support, but each of their children receives less support on average. Prior research has found that grown children experience benefits from parental support, including better adjustment and mental health (Fingerman, Cheng, Wesselmann, et al., 2012). Future studies might focus on grown children’s outcomes to ascertain whether lower SES offspring suffer because of a dearth of parental support or whether the support provided is sufficient.

Appendix.

Social Support Questions

Now, we’d like to know about the different kinds of help and support you provide to your to your child(ren) who (is/are) 18 or older. Please tell me the actual help you provide, not what you wish you provided or might provide under different circumstances.

1. Please think about financial support. Financial support involves giving money, loaning money, or helping them purchase goods, services, insurance, or education.

   How often do you provide child name with financial support (include holiday/birthday cash gifts)?

   • (1)Less than once a year or never
   • (2)Once a year
   • (3)A few times a year
   • (4)Monthly
   • (5)A few times a month
   • (6)Weekly
   • (7)A few times a week
   • (8)Daily

2. Now, think about the amount of financial support you have given to child name in the past 12 months—including any loans you have provided.

   • 2a)Did you give child name at least $500 in financial support in the past 12 months?
   • (1)Yes - Go to question 2b
   • (5)No - Go to question 2c
   • 2b)Did you give child name more than $1,000 in financial support in the past 12 months?
   • (1)Yes - Go to question 3
   • (5)No - Go to question 3
   • 2c)Did you give child name at least $100 in financial support in the past 12 months?
   • (1)Yes - Go to question 3
   • (5)No - Go to question 3

3. Now, please think about other practical assistance—for instance, fixing something around the house, running an errand, or providing a ride.

   How often do you provide child name with practical assistance? Please tell me the actual help you provide, not what you wish you provided or might provide under different circumstances.

   • (1)Less than once a year or never
   • (2)Once a year
   • (3)A few times a year
   • (4)Monthly
   • (5)A few times a month
   • (6)Weekly
   • (7)A few times a week
   • (8)Daily

4. Now, please consider advice you provide—that is, help with a decision or suggestions about things they could do.

   • How often do you give child name advice?

   • (1)Less than once a year or never
   • (2)Once a year
   • (3)A few times a year
   • (4)Monthly
   • (5)A few times a month
   • (6)Weekly
   • (7)A few times a week
   • (8)Daily

5. Please think about socializing—such as going out or doing activities together?

   • How often do you socialize with child name?

   • (1)Less than once a year or never
   • (2)Once a year
   • (3)A few times a year
   • (4)Monthly
   • (5)A few times a month
   • (6)Weekly
   • (7)A few times a week
   • (8)Daily

6. How about technological assistance—for instance, teaching them about a computer program, selecting electronic equipment, or how to use email?

   • How often do you provide child name with technological assistance?

   • (1)Less than once a year or never
   • (2)Once a year
   • (3)A few times a year
   • (4)Monthly
   • (5)A few times a month
   • (6)Weekly
   • (7)A few times a week
   • (8)Daily

7. How about talking about daily events—that is, talking with you about recent events or things that have happened in their lives, at work or with the family?

   • How often does child name talk with you about (his/her) daily life?

   • (1)Less than once a year or never
   • (2)Once a year
   • (3)A few times a year
   • (4)Monthly
   • (5)A few times a month
   • (6)Weekly
   • (7)A few times a week
   • (8)Daily

8. Please think about emotional support—Emotional support involves listening to someone’s concerns or being available when they are upset. By available we mean willing to listen, by phone, in person, or in any form.

   • How often do you provide child name with emotional support?

   • (1)Less than once a year or never
   • (2)Once a year
   • (3)A few times a year
   • (4)Monthly
   • (5)A few times a month
   • (6)Weekly
   • (7)A few times a week
   • (8)Daily

9. Do you provide child care for any of your grandchildren?

   • (1)Yes - Go to question 10
   • (5)No - Go to question 11

10. How often, if ever, do you provide childcare for child name’s children?

    • (1)Less than once a year or never
    • (2)Once a year
    • (3)A few times a year
    • (4)Monthly
    • (5)A few times a month
    • (6)Weekly
    • (7)A few times a week
    • (8)Daily

11. Other than vacations or visits, have you lived in the same household with (your adult child/any of your adult children) during the past 12 months? (If necessary read: Have you lived in the same household for more than 3 weeks in the past 12 months?

    • (1)Yes - If yes, ask 11a below.
    • (5)No
    • 11a.Which child(ren)? _____________________________________