Building Blocks for a Happy Life: Longitudinal Associations Between Early Life Income, Mentorship and Later Well-Being

Abstract

Longitudinal data from the Panel Study of Income Dynamics (PSID; N = 2996) were used to test hypotheses about the link between well-being and financial and social developmental resources. Results suggest that higher average family income from birth to age 18, and personal and professional mentoring received between 17 and 30, were positively associated with indicators of positive well-being and negatively related to negative indicators of well-being. Interactions between early life family income and mentoring during emerging adulthood were not significant predictors of any of the well-being outcomes. In all cases, the magnitudes of the coefficients became larger when simultaneously accounting for early life income, emerging adulthood mentoring, and their interactions – suggesting that financial and social resources in earlier life are independently linked to later life well-being. Findings highlight that mentoring received in emerging adulthood benefits downstream hedonic and eudemonic well-being, regardless of financial resources.

The financial and social resources available to people earlier in life may provide a foundation for well-being in adulthood (Moro-Egido et al., 2021; Ward & King, 2019). Early economic advantages, particularly in the first few years of life, are shown to be developmentally linked to downstream subjective well-being (Evans & Cassells, 2014; Gariepy et al., 2017), however, most research focuses heavily on how wealth may reduce negative indicators of well-being, rather than the promotion of flourishing. Likewise, social resources such as having a network of mentors have been shown to predict later subjective well-being and flourishing (Fruiht et al., 2021) – and are sometimes suggested to be “the great equalizer”, for those who grow up in less favorable socioeconomic environments. However, in making these claims, most prior work investigating the downstream benefits of mentoring has not rigorously measured or accounted for family financial resources simultaneously. Motivated by Bourdieu’s (1968) framework of economic, social, and cultural capital, this study investigates financial resources and mentoring, simultaneously, as potential determinants of both positive and negative indicators of subjective well-being and flourishing later in life.

Bourdieu’s (1986) framework of capital distinguishes between the different forms of resources that a person may acquire and utilize throughout their lifespan. The most intuitive, economic capital—or a resource that can easily be turned into money—is paralleled in this framework by social capital, which is capital that comes from one’s social network in the form of social support and social trust. Finally, cultural capital can be thought of in terms of experiences and resources like education, and exposure to cultural goods and ideas. All three forms of capital are thought to be predictive of physical and cognitive health as well as well-being (Qi et al., 2023; Sabatini et al., 2023), but the literature is generally more robust in the areas of social and economic capital (Pinxten & Lievens, 2014). In the present study, our longitudinal model examines social and economic capital as predictors of well-being, and accounts for interaction effects between these two forms of capital.

Economic Capital in Childhood and Downstream Well-being

Just as income is often linked to better well-being among adults (Killingsworth, 2021; Tan et al., 2020), financial resources in childhood are linked to better downstream outcomes. In the context of the United States (US), economic capital in childhood and adolescence demonstrate both a short and long-term impact on psychological well-being, cognitive performance, and physical health (Gariepy et al., 2017; Sobolewski & Amato, 2005). For instance, one longitudinal study from the US demonstrated that lower family income at age 9 was related to worse short-term memory performance, and greater experiences of helplessness, more internalizing and externalizing behaviors, as well as higher allostatic load both in childhood and emerging adulthood (Evans, 2016). Similarly, a nationally representative study of older adults in the US suggests that higher self-reported childhood SES was related to having better physical health, fewer functional limitations, as well as better cognitive performance and mental health in late adulthood (Luo & Waite, 2005). Research is also beginning to demonstrate a causal link between economic resources in early childhood and differences in infant neural development (Troller-Renfree et al., 2022) as well as downstream physical health (Braga et al., 2020; McInnis, 2023).

While a variety of research exists establishing the relationship between income in childhood and adolescence and well-being in adulthood in the US, a good deal of this work has methodological limitations. Specifically, well-being is generally operationalized using a largely deficit framework that emphasizes internalizing and externalizing behaviors, emotional regulation, and depressive symptoms (Evans & Cassells, 2014; Luo & Waite, 2005; Sobolewski & Amato, 2005) rather than more robust measures of well-being. In contrast, the current study operationalizes well-being as comprised of its two most empirically supported facets: hedonic and eudaimonic well-being. Broadly, hedonic well-being, which is commonly described as “subjective well-being” (Diener et al., 1999), is the measure of a person’s emotional (positive and negative affect) and cognitive states (satisfaction with life). Distinctly, eudaimonic well-being, or “flourishing,” is the measure of a person’s feeling that they are living life to their full potential in multiple domains such as relationships, fulfilment, purpose, and mastery (Ryff et al., 2016). While hedonic and eudaimonic well-being are sometimes lumped under a single umbrella, a strong body of research suggests that these two facets are distinct and associated with convergent and divergent factors (for review see Huta, 2017).

To date, very little research has aimed to understand the relationship between family income and well-being using the lenses of subjective well-being or flourishing. Furthermore, because positive and negative indicators of well-being are independent of one another (Diener & Emmons, 1984; Karademas, 2007) they may be driven by different childhood resources. Likewise, for studies to capture the full understanding of earlier economic and social resources on later life, considering positive and negative indicators of well-being separately is necessary to bring scientific rigor when examining the assets that contribute to overall well-being.

The Interplay of Economic Capital with Social and Cultural Capital

While higher family income is correlated with better physical and psychological health downstream, it likely predicts these outcomes both directly and indirectly through its impact on other forms of capital. Higher family income can provide greater access to better nutrition and health care to promote physical health, setting the stage for health disparities that persist and amplify in adulthood (Braveman & Barclay, 2009). Familial wealth also affords access to cultural capital via enrichment activities (Tuchman & Pillow, 2018) and higher education (Luo & Waite, 2005; Sobolewski & Amato, 2005) to promote cognitive function and psychological well-being. There is also cumulative risk associated with poverty in the form of added stressors, such as housing problems, family turmoil, and exposure to violence, as well as the quality of the parental marital relationship and the parent-child relationship. These secondary impacts of financial instability have been shown to mediate the relationship between family income and psychological well-being (Evans & Cassells, 2014; Sobolewski & Amato, 2005). However, among adults, social capital may be one particularly salient resource in buffering against the impacts of economic hardship (Moro-Egido et al., 2021; Sarracino & Piekałkiewicz, 2020). Therefore, non-financial resources that buffer a child against the challenges of being raised in a lower-income household and provide access to additional assets may also be critical in promoting more positive outcomes despite financial hardship.

Mentoring and Social Capital as Predictors of Well-Being

Mentoring, or when an experienced adult takes an active role in supporting the personal or professional development of a less experienced person (Thomson et al., 2014) early in life has been shown to predict a variety of strengths and resources in adulthood and may be an asset that helps buffer lower income youth against adversity (Hurd et al., 2014; Zimmerman et al., 2002). The benefits of mentoring youth and adolescents are vast and diverse (Van Dam et al., 2018). Mentors support their proteges by providing not only socio-emotional support, but also support for cognitive and identity development (Rhodes et al., 2006; Miranda-Chan et al. 2016). These different approaches to mentorship may benefit youth differently, depending on mentors’ experiences and goals. Still, research consistently demonstrates that mentoring in both socio-economic and cognitive domains during adolescence and emerging adulthood are associated with higher educational attainment and better psychological well-being, as measured by increases in optimism and self-efficacy, and fewer depressive symptoms (Crisp & Cruz, 2009; Miranda-Chan et al., 2016). Furthermore, there is increasing evidence for the long-term benefits of mentoring for well-being later in life (Boeder et al., 2021; Hagler & Rhodes, 2018; McDonald & Lambert, 2014). Adolescence and emerging adulthood, a phase of development that straddles adolescence and adulthood (Arnett, 2000), may be particularly salient times for mentorship. In these years, young people begin to explore adult roles and set the stage for their future identities and careers (Arnett, 2000). Having a more experienced adult to help scaffold this process and provide additional support and resources may be particularly beneficial as young people demand more autonomy from their parents and desire to set their own course (Larson et al., 1996).

To better understand the benefits of mentoring, some mentoring research separates the impact of strong social ties (e.g., relatives and fictive kin) from weaker social ties (e.g., teachers and community members; Hagler & Rhodes, 2018; Le et al., 2021; Raposa et al., 2018). This allows researchers to distinguish the differences between support that comes from a young person’s existing social network, who may share many demographic characteristics with a protege, from more distal mentors that may be able to share social capital and help with skill building that a protege may lack around navigating higher education (Raposa et al., 2018). Generally, findings suggest that young people with mentors from school or the larger community see more benefit from their mentorship (Erickson et al., 2009; Fruiht & Chan, 2018; Hagler & Rhodes, 2018) in terms of academic success. However, family members and other close ties may be particularly well-suited to provide more emotional support and help a young person navigate identity development (Albright et al., 2017). In sum, the literature suggests that both familial and non-familial mentorship can have a real and lasting impact on well-being, and that mentoring, like social capital, may be particularly beneficial for less-resourced youth (Erickson et al., 2009; Miranda-Chan et al., 2016; Hagler & Rhodes, 2018).

Given the general finding that mentorship – regardless of the social role of the mentor or specific functions that a mentor may serve – can be an asset to development, career mentoring research increasingly investigates the constellations of support that surround a protege (Higgins & Kram, 2001). The developmental networks framework, for instance, describes a developmental network in terms of its structure (e.g., the number of mentors and diversity of roles from which mentors come), and content (types of support received; Dobrow et al., 2012). In the domains of professional and academic mentoring, this framework has been used to capture the social capital transmitted through developmental relationships more cohesively than investigating a single source of mentoring (Dobrow et al., 2012).

While higher SES youth are generally more likely to report having a mentor, particularly from outside of their family (Fruiht et al., 2021; Gowdy et al., 2020; Raposa et al., 2018) relatively little research has investigated the differences in developmental network characteristics of people from different backgrounds. However, access to social capital via networks of mentoring support may be another mechanism by which family income predicts downstream success and well-being, as coming from a family with more economic capital may make it more likely that a young person has the opportunity to connect with mentors. By looking both at the simultaneous effects of networks of mentoring support and income as well as the interaction between the two, the present study aims to better understand if discrepancies in access to mentoring intensifies a gap in well-being driven by family income, or if mentoring can offset any effects of lower income on well-being.

The Present Study

Given the benefits and protective effects of capital for young people in the US, the present study aims to investigate the relationship between family income and mentorship in youth and emerging adulthood with downstream well-being. Using a national sample of the US, drawn from the Panel Study of Income Dynamics, we consider average family income from birth to age 18 as reported by participants’ parents/guardians during their youth, as well as a retrospective self-report of mentoring received between ages 17 and 30, to predict a wide variety of well-being outcomes in adulthood. Rather than a single indicator of mentoring tied to a specific mentor role (e.g., familial, school-based, etc.) or mentoring function (instrumental, relational, cognitive, etc.), mentoring is captured using an indicator of the general size and breadth of a respondents’ developmental network, capturing both familial and non-familial mentors who provided support in the domains of career development and relationships. This allows us to capture mentoring as a source of social capital in emerging adulthood more broadly than using a single mentor nomination. Building upon prior work suggesting that family income buffers young people against deficit indicators of well-being, we use an inclusive measure of well-being that captures positive and negative affective well-being, satisfaction with life, as well as eudaimonic well-being.

Hypothesis 1:

Household income over the first 18 years of life is positively associated with satisfaction with life, flourishing, and positive affect in adulthood, and negatively associated with negative affect in adulthood.

Hypothesis 2:

Mentoring in emerging adulthood is positively associated with satisfaction with life, flourishing, and positive affect in adulthood, and negatively associated with negative affect in adulthood.

Hypothesis 3:

There is a significant interaction between early life family income and mentoring received in emerging adulthood such that among those with lower early life family income mentoring received in emerging adulthood is a more powerful predictor of well-being in adulthood than it is for those with higher early life family income.

Method

Participants

The current study leveraged data from the Panel Study of Income Dynamics (PSID) to examine the link between well-being and earlier life income and mentoring resources. The PSID is a nationally representative survey of the Unites States, which began in 1968 with almost 5,000 families. Since then, the survey has followed original sample members and their descendants, consistently collecting a rich set of data on family demographic characteristics and income. Prior to 1997, the survey was conducted annually, and became biennial thereafter. In later years, the PSID has collected additional information through supplementary surveys. The 2014 PSID Childhood Retrospective Circumstances Study (PSID-CRCS) data was collected between May 2015 and January 2015 from English speaking PSID heads of households or their spouses who were 19 years or older. The 2016 PSID Well-being and Daily Life Supplement (PSID-WB) data was collected between March and December of 2016 from English speaking PSID heads of households or their spouses who were 30 years or older.

This study uses the PSID core data from 1968 to 2015 to obtain information on childhood family income and composition as well as demographic covariates from current PSID heads of household or spouses who were formerly children in PSID households. That is, childhood family income and family composition were based on parents reported income during participants’ childhood. Participants reported retrospectively on their mentoring relationships in emerging adulthood in the Childhood Retrospective Circumstances Study (PSID-CRCS). All well-being indicators were obtained from the Well-being and Daily Life Supplement (PSID-WB) capturing well-being at the time of the survey. Therefore, our final sample includes only individuals who turned 30 by the time the PSID-WB was conducted and those who grew up in a PSID household. To preserve the representativeness of the sample, the PSID constructs various weighting variables which adjust for attrition as well as other sampling decisions that might affect the representativeness of the data. For all our analyses, we use the cross-sectional weights which are published in the PSID-WB which contains our outcomes of interest.

The final participant sample (N = 2996) was 53% female, 84% married, with an average age of 48.01 years (SD = 10.67) and lived in households during childhood with on average 4.83 family members (SD =1.61). In line with the demographics of the larger PSID core sample, participants were 85% white and 14% Black; the final 1% of the sample was comprised of participants of other races. In constant 2021 dollars, the sample’s family income averaged $93,382 US (SD = $65,899), with males making up 85% of the head of households. Participants received mentoring from an average of 1.71 individuals (SD = 1.57). Table 1 contains further details of the final sample.

Table 1:

Descriptive Statistics

	Mean	Standard Deviation
Family income between birth and age 15	$93,382.56	$65,899.02
Number of mentoring relationship types	1.71	1.57
Work mentor: family	0.51	0.50
Work mentor: non-family	0.58	0.49
Relationship mentor: family	0.46	0.50
Relationship mentor: non-family	0.49	0.50
Individual Characteristics
Age	48.01	10.67
Male	0.47	0.50
White	0.85	0.36
Parental Characteristics
Number of family members between birth and age 18	4.83	1.61
Male headed household between birth and age 18	0.85	0.29
Married parents between birth and age 18	0.84	0.30
Outcomes
Satisfaction with Aspects of life	2.70	0.41
Life Satisfaction	3.69	0.95
Flourishing	4.17	0.64
Positive affect	3.50	0.74
Positive experienced well-being	3.65	0.87
Negative affect	1.80	0.76
Negative experienced well-being	1.66	0.68
Tiredness or pain	2.25	0.96
Number of individuals	2,996

Measures

Satisfaction with Aspects of Life

Ten items captured participants’ satisfaction with the aspects and conditions of their life overall. This scale, adapted from Campbell and colleagues (1976) measure in the Quality of American Life survey, asks participants to self-report their overall level of satisfaction with specific conditions of their life including tangible conditions (job, city, financial circumstance, hobbies), relationships (family life, romantic relationship, friendships), as well as their overall health and faith. Items were rated on a 5-point Likert scale from 1 (completely satisfied) to 5 (not at all satisfied). Scores for all items were averaged to create a composite score that measures satisfaction with aspects of life and captures the overall quality of life. Cronbach’s alpha for this scale was .865, suggesting a high level of reliability among the items.

Life Satisfaction

The Satisfaction with Life Scale (SWLS; Diener et al., 1985), was used to capture life satisfaction. Items on the SWLS were assessed on a 5-point Likert scale and scores for items were averaged to create a composite score. When rated on a 5-point Likert scale ranging from 1 (Strongly disagree) to 5 (Strongly Agree) it shows good internal consistency (α = .89; Kobau et al., 2010) and convergent and discriminant validity with measures of meaning in life, self-determination, affect, and other measures of life satisfaction (Kobau et al., 2010). The SWLS is generally administered on a 7-point Likert scale, which also shows good internal consistency (Diener et al., 1985; α = .87) and temporal stability over two months (α = .82; Diener et al., 1985) as well as five years (Fujita & Diener, 2005). The 7-point version also shows discriminant validity from depression and distress measures (Diener et al., 1985), construct validity with other measures of overall well-being and esteem (Diener et al., 1985) and convergent validity with other-report measures of well-being (Schneider & Schimmack, 2010). In the present study, Cronbach’s alpha for this scale was .886, which is consistent with previous studies.

Flourishing

Flourishing was measured with the Flourishing Scale (FS; Diener et al., 2010). The scale consists of eight questions, capturing eight aspects of psychological well-being including meaning and purpose, supportive and rewarding relationships, engagement and interest, contribution to the well-being of others, competency, self-acceptance, optimism, and being respected. Items were assessed on a 5-point Likert scale ranging from 1 (Strongly disagree) to 5 (Strongly agree) and scores were averaged to create a composite score. The validated scale, which uses a 7-point Likert, demonstrates a one factor structure (α = .86) and acceptable temporal stability over one month (r = .71; Diener et al., 2010). It demonstrates construct validity with various other measures of well-being and self-determination (Diener et al., 2010; Hone et al., 2014) and discriminant validity with the Center for Epidemiological Studies Depression Scale (Hone et al., 2014). In the present study, Cronbach’s alpha for this scale was .885.

Positive and Negative Affect

Mroczek and Kolarz’ Positive and Negative Affect Scale (PANAS; Mroczek & Kolarz, 1998) measured affect experienced across the previous 30 days from when the interview was conducted. The 12-item PANAS uses a 5-point scale where responses ranged from 1 (None of the time) to 5 (All of the time). Sample positive emotions include “cheerful,” “satisfied, “calm,” and sample negative emotions are “nervous,” “hopeless,” and “worthless.” The PANAS demonstrates a two-factor structure with positive emotions (α = .91) and negative emotions (α = .85) loading onto separate factors (Joshanloo, 2017; Kobau et al., 2010). The scales correlate with life satisfaction, neuroticism, extraversion, esteem, and life satisfaction in the expected directions (Joshanloo, 2017). Scores for the six positive emotion items and six negative emotion items were averaged separately to create a composite score for each. Cronbach’s alphas for positive and negative affect in this sample were .931 and .886, respectively.

Experienced Well-being

Daily experienced well-being was measured using a 14-item inventory adapted from the Day Reconstruction methodology developed by Kahneman et al. (2004). This measure used 5 items to capture positive emotional experiences (calm, happy, enthusiastic, content, interested), 7 items about negative emotional experiences (angry, frustrated, sad, stressed, lonely, worried, bored), and 2 items about experiencing tiredness and pain. Responses ranged from 1 (all of the time) to 5 (none of the time). A separate composite was created for each of the three subscales (positive well-being, negative well-being, tiredness and pain) by averaging scores from each subscale. The measure was developed to capture fleeting emotions while performing specific tasks throughout the day, akin to a retrospective experience sampling methodology. However, in the present study participants were asked to report how much of the previous day they experienced each emotion. A slightly longer adaptation of this measure used in the 2010 American Time Use Survey demonstrated validity via a robust relationship with self-reported health as well as good internal reliability (Lee et al., 2016). Cronbach’s alpha for positive experienced well-being was .906, for negative experienced well-being was .883, and for tiredness and pain was .603.

Family Income

We capture family income by calculating average pre-tax family income between birth and age 18. Participants in the present study were PSID descendants who had been children in PSID households prior to becoming PSID household heads/spouses themselves. Therefore, to construct average pre-tax family income, we utilized income reports from participants’ parents/guardians during their childhood. Mean childhood family income was calculated from partially available data for those with some missing waves of income data. We sum the observed family income at all ages between birth and age 18 and divide it by the number of times income was observed for each individual. For more than 90% of our sample we observe income at least 8 times. Only 1% of our sample had their income observed only once or twice. Before constructing average pre-tax family income, we convert annual family income to constant 2021 dollars to account for changes in family income that might result from inflation and would confound the estimated effects. Average income throughout childhood provides a good measure of permanent income which is arguably the most relevant type of income for these later life outcomes (Curtis et al., 2001; McInnis, 2023).

Mentoring

The PSID-CRCS asks four questions relating to mentoring in emerging adulthood that were composited to capture mentoring. First, respondents indicated whether between the ages of 17 and 30 a family member (other than someone who raised them) “provided [them] with positive support or mentoring that helped [them] succeed in [their] interpersonal relationships, such as marriage or a marriage-like relationship” (Beier et al., 2000), as well as whether a family member had “provided [them] with positive support or mentoring that helped [them] succeed in [their] work life”. These same two items were then repeated regarding mentorship from an adult outside of their family. The most common mentor type reported was a non-family mentor for work life (58%) and the least common was a family member mentoring for relationships (46%). We sum these four dichotomous variables to construct a single continuous variable which measures the number of different types of mentors reported ranging from 0 (no mentors) to 4 (both a family and non-family mentor for both relationships and work life).

Covariates

Consistent with the previous literature we control for several factors that are correlated with well-being such as marital status (married or unmarried), age, sex, and race. Race is divided into two groups – white and non-whites – because most of our sample is either white or black, and only a very small proportion identify with a different racial group. Given that we consider the longer-term impacts of income during childhood, we consider additional factors that might impact long-term well-being such as whether the individual grew up in a male headed household and the family size. These time varying characterizations, such as whether the individual resided in a male headed household and the family size as a child, are constructed as the proportion of their childhood spent in a male headed household and the average family size throughout childhood. These demographic variables were pulled from the 1968–2015 PSID core sample dataset.

Analytical Plan

For our analyses, we estimate multivariate linear regressions, with various well-being measures as our dependent variables. All our models include controls for various confounders which are described in table 1. Our key independent variables are average income between birth and age 18 and availability of mentoring relationships that the individual had as an emerging adult. We estimate three separate models as listed below.

$${well}_i^{>30}={\beta }_0+{\beta }_1{Inc}_i^{0-18}+{\beta }_2{X}_i+{\beta }_3{W}_i^{0-18}+{\mathrm{\pounds }}_i$$ (1)

$${well}_i^{>30}={\alpha }_0+{\alpha }_1{Ment}_i^{17-30}+{\alpha }_2{X}_i+{\alpha }_3{W}_i^{0-18}+{\mathrm{€}}_i$$ (2)

$${well}_i^{>30}={\gamma }_0+{\gamma }_1{Inc}_i^{0-18}+{\gamma }_2{Ment}_i^{17-30}+{\gamma }_3{Inc}_i^{0-18}\mathrm{*}{Ment}_i^{17-30}+{\gamma }_2{X}_i+{\beta }_3{W}_i^{0-18}+{\mathrm{Э}}_i$$ (3)

where i represents individual, and the superscripts denote the age range over which the variables are measured. $wel{l_i}^{>30}$ is one of our dependent variables measuring well-being. $In{c_i}^{0-18}$ represents average family income between birth and age 18, measured in constant 2021 dollars. $Men{t_i}^{17-30}$ represents the number of mentoring types received between ages 17 and 30. $X_i$ represents individual level control variables for marital status, age, sex and race (white and non-white). ${W_i}^{0-18}$ denotes control variables for the proportion of time between birth and age 18 that the child had a male head of household and the average number of family members living together between birth and age 18. The various epsilons represent the error terms or the unobserved factors that impact well-being.

The first model includes average income along with the covariates, but excludes mentoring. The second model includes mentoring along with the covariates but excludes income. The third model includes income, mentoring and an interaction between mentoring and income, as well as the covariates. If average income between birth and age 18 is correlated with mentoring between age 17 and 30, then the estimated impacts from both model 1 and model 2 would be biased. If income is positively correlated with mentoring, then the estimated effects would be too large, while if income and mentoring are negatively correlated then the estimated effects would be too small or biased toward zero.

Results

Our final sample of 2,996 adults comprise only those 2015 PSID heads of households or spouses who were children in PSID families prior to being heads/spouses, participated in both the PSID-CRCS and PSID-WB, and who had valid data on our target variables. Approximately 7% of cases were excluded listwise because of missing data on key variables.

A series of multiple regression models (Table 2) were estimated to identify the unique contributions of earlier life family income from birth to 18 and mentoring received in adolescence/emerging adulthood on later adulthood life well-being indices. Table 2 demonstrates each series tested: Model 1) the hypothesized unique developmental pathway between family income on well-being, excluding mentoring received; Model 2) the hypothesized developmental pathways between mentoring received in adolescence/emerging adulthood on well-being, excluding family income; Model 3) includes all three variables–family income, mentoring, and the interaction terms between mentoring and family income on later life well-being.

Table 2.

Family Household Income and Mentoring Predicting Well-Being (N = 2996)

	Model 1	Model 2	Model 3
Adult Well-Being Outcome	Unstandardized B coefficient (SE)
Positive
Satisfaction with Aspects of Life
Early Life Income (ELI)	.007 (.001)	---	.002 (.001)
Mentoring Received (MR)	---	.014 (.010)	.021 (.012)*
ELI x MR	---	---	ns
Life Satisfaction
Early Life Income (ELI)	.012 (.004)***	---	.013 (.005)**
Mentoring Received (MR)	---	.057 (.011)***	.062 (.027)**
ELI x MR	---	---	ns
Flourishing
Early Life Income (ELI)	.005 (.002)**	---	.006 (.004)
Mentoring Received (MR)	---	.061 (.007)***	.064 (.015)***
ELI x MR	---	---	ns
Positive Affect
Early Life Income (ELI)	−.001 (.003)	---	.000 (.003)
Mentoring Received (MR)	---	.046 (.009)***	.053 (.017)***
ELI x MR	---	---	ns
Positive Well-Being
Early Life Income (ELI)	−.002 (.003)	---	−.001 (.004)
Mentoring Received (MR)	---	.035 (.009)**	.042 (.020)*
ELI x MR	---	---	ns
Negative
Negative Affect
Early Life Income (ELI)	−.006 (.003)***	---	−.006 (.004)
Mentoring Received (MR)	---	−.016 (.009)*	−.016 (.016)
ELI x MR	---	---	ns
Negative Well-Being
Early Life Income (ELI)	−.005 (.003)*	---	−.008 (.003)**
Mentoring Received (MR)	---	−.012 (.009)	−.032 (.018)*
ELI x MR	---	---	ns
Tiredness/Pain
Early Life Income (ELI)	−.017 (.004)***	---	−.019 (.005)***
Mentoring Received (MR)	---	−.013 (.014)	−.030 (.023)
ELI x MR	---	---	ns

Notes. Covariates for each model include age, gender, race/ethnicity, marital status, average family income size, and head of the household gender; Model 1: Regression of unique contribution of earlier life family income (from birth to age 18) without mentoring received on well-being indices, adjusted for covariates; Model 2: Regression of unique contribution of mentoring received in adolescence/emerging adulthood on well-being indices without family income, adjusted for covariates; Model 3: Regression of main effects of earlier life family income (from birth to age 18) and mentoring received on well-being indices, and their interaction effects (ELIxMR), adjusted for covariates;

^***p <.001;

^**p <.01;

^*p <.05; ns = not significant

Hierarchical regression results are presented in Table 2. In support of hypothesis 1, results from model 1 demonstrate that higher early-life family income is significantly associated with better well-being outcomes in adulthood when not accounting for mentoring received. Specifically, in five out of eight well-being indices, greater family income is significantly associated with well-being. That is, more life satisfaction (B = .012, SD = .004), p < .001), greater flourishing (B = .005, SD = .002), p < .01) and fewer reports of negative affect (B = −.006, SD = .003), p < .001), less negative well-being (B = −.005, SD = .003), p < .05), and lower tiredness or pain (B = −.017, SD = .004), p < .001).

In parallel, hierarchical regression results in Table 2, from Model 2, demonstrate that more mentoring received in earlier life is significantly associated with more well-being when not accounting for early life family income, thus supporting hypothesis 2. Specifically, having more mentors is significantly associated with five out of eight well-being indices including more life satisfaction (B = .057, SD = .011), p < .001), greater flourishing (B = .061, SD = .007), p < .001), more reports of positive affect (B = .046, SD = .009), p < .001) and positive well-being (B = .035, SD = .009), p < .01) and fewer reports of negative affect (B = −.005, SD = .003), p < .05).

Furthermore, when accounting for both early life family income, mentoring received in adolescence/emerging adulthood, and the effects of their corresponding interactions in Table 2, Model 3, several coefficient magnitudes tended to increase. On average, participants with more mentors were observed – on six out of eight well-being indices – to have more aspects of life they were content with (B = .021, SD = .012), p < .05), more satisfaction with life (B = .062, SD = .027), p < .01), greater flourishing (B = .064, SD = .015), p < .001), more reports of positive affect (B = .053, SD = .017), p < .001) and positive well-being (B = .042, SD = .020, p < .05), and less negative well-being (B = −.032, SD = .018), p < .05). Likewise, magnitudes for coefficients for participants with more early life family income were significantly linked to greater satisfaction with life (B = .013, SD = .005), p < .01), less negative well-being (B = −.008, SD = .003), p < .01), and lower tiredness or pain (B = −.019, SD = .005), p < .001). However, contrary to hypothesis 3, no statistically significant interaction effects were observed for adult well-being outcomes (p > .05).

Lastly, we conducted a series of post hoc analyses to examine whether our findings for earlier family income and mentoring received were robust to the inclusion of individual labor income during adulthood (see Table 3). We consider this because family income earlier in life is expected to correlate with the individual’s income during adulthood. When we control for current individual labor income, childhood family income is no longer statistically significant for all longitudinal links to well-being (p > .05) – however, early life mentoring received remains relatively unchanged in being a significant predictor for later life well-being.

Table 3.

Current Income, Family Household Income, and Mentoring Received Predicting Well-Being (N = 2996)

Well-Being Outcome	Unstandardized B coefficient (SE)
Positive
Satisfaction with Aspects of Life
Current Income	.003 (.001)**
Early Life Income (ELI)	.001 (.001)
Mentoring Received (MR)	.020 (.012)*
Life Satisfaction
Current Income	.029 (.007)***
Early Life Income (ELI)	.008 (.005)
Mentoring Received (MR)	.061 (.022)***
Flourishing
Current Income	.018 (.004)***
Early Life Income (ELI)	.002 (.003)
Mentoring Received (MR)	.063 (.014)***
Positive Affect
Current Income	.014 (.003)***
Early Life Income (ELI)	−.002 (.003)
Mentoring Received (MR)	.053 (.017)***
Positive Well-Being
Current Income	.013 (.004)***
Early Life Income (ELI)	−.003 (.004)
Mentoring Received (MR)	.041 (.019)**
Negative
Negative Affect
Current Income	−.021 (.006)***
Early Life Income (ELI)	−.002 (.004)
Mentoring Received (MR)	−.015 (.017)
Negative Well-Being
Current Income	−.013 (.004)***
Early Life Income (ELI)	−.003 (.004)
Mentoring Received (MR)	.041 (.019)***
Tiredness/Pain
Current Income	−.011 (.004)***
Early Life Income (ELI)	−.006 (.003)
Mentoring Received (MR)	−.032 (.018)*

Notes. Covariates for each model include age, gender, race/ethnicity, marital status, average family income size, and head of the household gender; No three-way interactions were significant in all models; Regression of main effects of current income, earlier life family income, and mentoring received on well-being indices, adjusted for covariates;

^***p <.001;

^**p <.01;

^*p <.05.

Discussion

Childhood experiences can shape the trajectories of our adult lives. Findings from the present study align with the types of capital framework (Bourdieu, 1986) and demonstrate that economic and social capital in youth and emerging adulthood predict downstream well-being among adults in the United States. However, the pattern of findings suggests a slightly different impact of economic versus social capital on well-being. Economic capital may reduce the number or extent of the hindrances to well-being, whereas social capital may promote flourishing, positive emotion, and general life satisfaction. Specifically, findings show that low income in childhood is predictive of worse well-being in adulthood. This shows up in two positive well-being indicators (satisfaction with life and flourishing), but more consistently in negative well-being indicators such as pain and daily experienced negative emotions. In line with prior research (Miranda-Chan et al., 2016), mentoring in emerging adulthood also predicted well-being, including satisfaction with life and flourishing. However, mentoring more consistently predicted positive indicators such as positive emotional experiences than negative ones.

Contrary to our hypotheses, there were no significant interaction effects between mentoring and income, suggesting that mentoring does not buffer against the potential negative impacts of low income on well-being. However, some notable effects were larger for both mentoring and income when both were entered into the model together. That is, once we controlled for the amount of mentoring a participant reported, income actually predicted reductions in negative well-being more strongly. Similarly, after controlling for childhood income, more mentorship predicted positive well-being more strongly. This was likely because of the fact that mentoring and income were found to be negatively correlated, therefore biasing the effect sizes towards zero when only one was entered into the model.

Main Effects of Income and Mentoring on Well-being

In line with prior research on the downstream correlates of financial resources in childhood, and in support of hypothesis 1, average family income over the course of childhood was predictive of well-being in adulthood. Prior work has focused primarily on the negative impacts of low family income in childhood, demonstrating that lower childhood income is related to more internalizing and externalizing behaviors, functional limitations and poor physical health, as well as mental health and cognitive issues (Evans, 2016, Luo & Waite, 2005; Sobolewski & Amato, 2005). However, we noted that little research has highlighted any relationship between childhood family income and well-being using a lens of flourishing or subjective well-being. Our findings suggest that family income is predictive of both positive and negative indicators of well-being downstream including positive associations with satisfaction with life and flourishing, and negative associations with negative affect, negative experiences of well-being, and feeling tiredness and pain. However, associations between financial resources and positive affect or positive daily experiences were not significant. These results suggest that while economic capital in childhood may help reduce the daily struggles we face in adulthood or give us access to cultural capital by way of opportunities to pursue meaningful endeavors and support eudemonic well-being, they do not necessarily promote positive emotion.

These findings support past work suggesting that financial resources in childhood may reduce the risk of physical and mental health issues in adulthood, two key impairments to well-being, but also suggest that financial resources help set the stage for a satisfying and meaningful adult life. When we controlled statistically for income in adulthood in a post hoc analysis, we no longer saw a significant effect of childhood family income, supporting the conclusion that higher family income in childhood likely opens doors to experiences, opportunities, and resources that promote financial stability in adulthood, which also contributes to well-being (Luo & Waite, 2005; Sobolewski & Amato, 2005; Tuchman & Pillow, 2018). That is, childhood family income likely promotes well-being in large part indirectly, via its impact on cultural capital such as education, access to health care, career development, and other resources not only in childhood, but also in adulthood.

In support of hypothesis 2, findings suggest that mentoring in emerging adulthood is also predictive of well-being in adulthood. In line with prior research suggesting that mentoring in adolescence and emerging adulthood can predict better mental health (Miranda-Chan et al., 2016), life satisfaction, and flourishing (Boeder et al., 2021) in adulthood, results showed that the number of different types mentors reported was predictive of their downstream life satisfaction and flourishing. In addition to our findings that parallel the results for family income, mentoring also predicted positive well-being via more positive affect and emotional experiences.

Mentoring in the present study was operationalized with a score that represented support from adults for one’s work life/professional development or interpersonal relationships during emerging adulthood – a transitional critical period for setting career and relational developmental trajectories (Arnett, 2000; Miranda-Chan et al., 2016). Given these specific mentoring functions, our findings may point to the way that mentoring may help people find meaningful work and develop healthy relationships that support them in the transition to adulthood. These benefits to career and relationship development may also have an impact on emotional experiences in later life. It is also possible that less central mentoring functions in these relationships, such as socio-emotional support, support for identity development, or the sharing of social capital in emerging adulthood may promote more positive emotional experiences downstream. It should be noted, however, that these correlational findings may also be detecting a bidirectional relationship, as it is possible that people who are higher on positive emotional affectivity throughout their lives have more access to mentorship.

Interactions Between Income and Mentoring

In addition to these main effects, we also tested the hypothesis that there would be a significant interaction between mentoring and childhood family income such that mentoring would have a compensatory effect on well-being for lower income participants. Although research suggests that mentoring benefits both high- and low-income individuals, mentors can serve as a tool for upward social mobility among less resourced youth (Erickson et al., 2009; Zimmerman et al., 2002). Furthermore, mentoring has been shown to buffer youth with behavioral or substance use issues, or who live in lower-income or higher-crime neighborhoods that might put them at higher risk for such challenges, against adversity (Tolan et al., 2014). However, this effect did not hold up in a large-scale meta-analysis looking at risk as a moderator of the impact of mentorship (Van Dam et al., 2018). Similarly, our results did not support the hypothesis that mentoring was more predictive of well-being among participants who had lower family incomes.

Instead, somewhat surprisingly, we found that mentoring was inversely related to income. As a result, in comparison to the main effects of income and mentoring, effect sizes for both were somewhat larger when both variables were entered into the model simultaneously. That is, while mentoring predicted well-being for people while controlling for income levels, just as income predicted well-being regardless of mentorship, our analyses bore out that both effects were intensified somewhat when controlling for the other. This suggests that any benefit of mentoring on downstream well-being is likely not tied to its relationship with income, or vice versa. The finding that those with lower income actually receive more mentoring in emerging adulthood supports the finding of Gowdy and colleagues’ (2020) analysis of the PSID dataset that demonstrated that respondents who reported their families struggled financially were more likely to report being mentored, and expands upon it by using an objective measure of financial resources.

Prior research on youth mentoring consistently suggests that more resourced young people with better educated parents are more likely to have access to mentorship (Erickson et al., 2009; Fruiht et al., 2021; Raposa et al., 2018). However, bias in the literature against familial mentors who generally share demographic characteristics with a protege, leads to operational definitions of mentoring that center non-familial mentors in analyses, which are consistently demonstrated to be more common among higher SES youth. Conversely, kin- and fictive-kin mentors are more common among youth of color and less resourced youth (Liao & Sanchez, 2019). Furthermore, prior research generally measures mentoring as a dichotomous variable, rather than asking about multiple mentors across multiple domains of life, comparing access to mentorship and mentoring functions based on a single reported mentoring relationship. In the present study, however, participants were asked to report up to four mentors (two familial and two non-familial) reporting distinct mentoring functions (work and relationships). This methodology may explain the deviation in our findings from prior work. When participants are given the opportunity to report both familial and non-familial relationships and frame mentoring relationships as shaping not only work success but also relationship success, lower income people seem to have more of these assets in their lives. That is, this more expansive definition of mentoring may be tapping mentoring as a developmental asset afforded to less resourced youth quite differently than past research.

In prior work that parsed the childhood predictors of familial versus non-familial mentoring using the PSID dataset, results showed that higher parental educational attainment and report of family financial struggles both predicted reports of a non-familial mentor (Gowdy et al., 2020). Taken together, these findings may continue to highlight the importance of social capital (by way of education or mentorship) over financial resources as a pathway to well-being. Conversely, our findings may also suggest that people with less income elicited or perceived more support from their communities or are more apt to recognize the help they received along the way. That is, people from lower SES backgrounds may be more able to identify the mentors who supported their work and relationship development than higher SES participants, partially accounting for the unexpected inverse correlation between mentoring and income. In addition to the value of our findings in understanding well-being, the developmental asset of strong family and fictive-kin networks in marginalized communities and communities of color may be more apparent when mentoring is measured in this more inclusive way.

Limitations, Future Directions, and Implications

The findings of the present study speak to the potential impact of developmental assets on downstream well-being, however some caution should be used in interpreting these findings as a result of the limitations of our methodology. First, making causal claims about the impact of family income or mentoring on well-being would require a randomized control trial, or independent sources of identifying variation for family income or mentoring through natural experiments such as policy changes that affect income or mentoring independently. In the absence of such gold-standard methodologies, however, we utilized rigorous statistical analyses, and a large nationally representative sample to demonstrate compelling correlational findings. Particularly in the case of mentorship, it is possible that dispositionally happy young adults elicited more support and mentorship before going on to be generally happier adults, or that despite the two year gap between the collection of the PSID-WB and PSID-CRCS, happier adults were more likely to look back on their emerging adult years and recognize the impact of mentors on their development.

While this is somewhat less likely in the case of family income, these findings must also be approached with some caution as supplemental analyses demonstrate that nearly all of the variance in well-being explained by childhood family income can also be explained by participants’ downstream income in adulthood. That is, children from higher income families grow up to be adults with higher incomes – and generally higher income adults have higher levels of well-being. However, this is not to say that our finding around the relationship between childhood income and downstream well-being is not valuable. It demonstrates that financial resources in childhood likely contribute to thriving in adulthood in many possible ways (e.g., via reduced stress and conflict in the household, access to health care, education, and enrichment) all of which may have helped higher income participants to grow into adults with more financial resources, but also with more opportunities for flourishing and fewer adversities that may drive negative emotional experiences. These findings highlight the importance of economic supports for families to promote the best possible outcomes for youth and provides insights into the consequences of income inequality on well-being in the US.

Furthermore, the structure of the PSID gives us only a snapshot of participants’ well-being. Although income data is collected annually, the Well-Being and Daily Life Supplement captures well-being at just one time point. Therefore, it is possible that participants’ well-being at the time of this survey is reflective of world circumstances, developmental factors, or even just random variation. By controlling for age and utilizing a large sample we can hope to limit the impact of these factors, but still must acknowledge that a construct as complex as well-being across one’s adult life cannot be fully captured by a single time point. Furthermore, given that all PSID participants 30 or older were invited to report on their well-being, we can only ensure a two-year lag between when participants reported being mentored and when they reported on well-being. It is possible that the full benefit of experiencing mentoring in emerging adulthood had not yet been realized in all participants, given that some were just moving out of emerging adulthood at the time of the survey. However, a strength of this work is in our multidimensional and comprehensive operationalization of well-being that was able to capture overall well-being, and positive and negative daily emotional experiences. As a result, our findings were able to detect the differential effects of different forms of capital on positive versus negative well-being. These findings underscore the importance of considering well-being as a multidimensional and complex construct, and the importance of considering positive and negative aspects of well-being, as well as both eudemonic and hedonic aspects.

Similarly, our operationalization of mentoring was unique from that used in the majority of the current mentoring literature, which typically captures a single mentoring relationship. Participants were able to report up to four sources of mentorship and a sum score was calculated for each participant based on these reports of mentoring received (familial and non-familial, each in the domains of work and relationships). However, no additional data was collected about these mentoring relationships to ensure that a single individual was not counted as both a work and relationship mentor. Additionally, items did not capture other mentoring functions (e.g., emotional support, identity development), and did not assess the quality of the mentoring relationship (e.g., frequency of contact, relationship duration or closeness). Furthermore, like much of the literature on mentoring, mentoring in this manuscript was captured retrospectively. Participants were asked, in middle to late adulthood, to reflect on the mentoring that they experienced between the age of 17 and 30. Retrospective self-report is a common methodology in mentoring research, but may be impacted by participant bias, social desirability, and the imperfect nature of memory. Considering our findings about the inverse relationship between mentoring and childhood income using this operationalization, future research must continue to improve existing methodologies to understand the networks of support that promote positive development in the transition to adulthood. Additionally, future research should aim to better understand this relationship between mentoring and well-being by considering the specific mentoring functions that promote downstream well-being.

Finally, these findings suggest a need for more funding to support programming to promote mentorship for young people during the transition to adulthood (i.e., emerging adults). While mentorship did not seem to buffer against the hardships driven by lack of economic resources, our findings suggest that access to mentoring during this transition may help set the stage for a happier and more fulfilling adult life. Unlike youth mentoring programs, which match youth with a caring adult to promote psychosocial development (e.g., Big Brothers Big Sisters; Herrera et al., 2023), existing mentoring programs for emerging adults generally exist within the professional sphere, built into post-secondary education or early career development. Not surprisingly, relationship mentoring was generally less common than work mentoring in this sample. Religious- and faith-based organizations may provide structured pre-marital mentoring to support young adults as they embark upon marriage, however, there are otherwise very few opportunities for young adults to seek out relationship mentoring outside of naturally occurring relationships. Given the finding that access to more mentoring types predicts greater well-being downstream, these results speak to the potential of mentoring to support adults in their relationships, and the need for more opportunities for this type of mentoring in tandem with career mentoring in emerging adulthood. However, it is important to note that mentoring should be framed as an asset to help promote well-being, regardless of socioeconomic status, but may not mitigate the very real physical and psychological consequences of financial inequity.

Conclusions

Results from this work highlight the importance of economic and social capital in childhood and adolescence in understanding precursors to a happy and healthy adult life. In particular, this work shows us that childhood family income and mentoring in emerging adulthood have the potential to promote positive development and downstream well-being. As suggested by prior works, higher family income may buffer young people from stressors and adversity that lead to pain, malaise, and negative emotional experiences. Beyond the benefits of financial resources, our findings provide evidence for the value of social support and social capital in predicting well-being. Mentoring may set the stage for a bright and successful future through advice and guidance as a young person navigates career development and serious romantic relationships for the first time. While more research is needed to better understand the mechanisms by which mentoring promotes downstream well-being, the present findings highlight the longitudinal benefit of income and mentoring in promoting a happy and flourishing adult life.