Maternal Functioning, Time, and Money: The World of Work and Welfare

Abstract

Numerous studies have assessed families’ employment and financial stability following welfare reform. Yet little research has addressed whether welfare and work transitions are linked with other changes in family functioning. Using a representative sample of approximately 2,000 low-income urban families from the Three-City Study, analyses assessed whether mothers’ welfare and employment experiences over a two-year period following welfare reform were related to changes in family well-being. Lagged regression models controlling for family characteristics and earlier levels of functioning found that moving into employment and stable employment (of 30 hours or more per week) were linked to substantial increases in income and improvements in mothers’ psychological well-being. Movements into employment also were associated with declines in financial strain and food insecurity. Sustained or initiated welfare receipt was related to relative declines in income, physical health, and psychological well-being, but also to improved access to medical care. In contrast, mothers’ welfare and work experiences showed very limited relations to changes in the quality of parenting or of children’s home environments. These patterns were similar for families with young children and those with adolescent children. Results suggest that parenting behaviors are more resistant to change than are maternal emotional and economic functioning.

Vociferous welfare reform debates in the 1990s reflected arguments concerning how moving low-income mothers out of welfare dependency and into stable employment would affect women’s well-being and the family environments provided to children. Some viewpoints held that leaving dependency and moving into self-sufficiency would help boost the financial resources of families and enhance economic security. Moreover, the experiences of employment— becoming economically self-sufficient, gaining new skills and social connections, and creating a new work identity— would improve low-income women’s self-esteem and psychological well-being. Finally, it was hoped that these changes would enhance the family environments and parenting provided to children raised in low-income families, affording greater structure, support, and resources provided to children. Over the longer term, proponents argued that such positive changes would advance poor children’s development. Alternative arguments predicted just the reverse— that low-income mothers, faced with lost welfare benefits and lacking the skills or opportunity to gain and sustain stable employment, could suffer diminished financial stability, increased psychological stress, and declines in the quality of parenting and environments provided to their children (see DeParle, 2004 and Haskins, 2001, for overviews of the arguments concerning welfare reform). Although recent scholarship has found rather minimal evidence of substantial changes in children’s health and development in the years immediately following welfare reform (Besharov, 2003; Chase-Lansdale et al., 2003; Kalil, Dunifon, & Danziger, 2001; Morris, 2002), many questions remain concerning whether the intervening processes noted above have responded to changes in low-income women’s employment and welfare behaviors induced by welfare reform.

Both conceptual models and previous empirical work provide hypotheses regarding how welfare and employment experiences might influence families’ economic security, maternal well-being, and parenting practices. Economic theory states that welfare and employment experiences could alter both the time and economic resources that parents provide to their children, but that these two factors would likely operate in counteracting ways (Becker, 1981). For instance, entrances into employment would be expected to increase parents’ income and hence their ability to provide resources to their children, but also would be expected to decrease the amount of time that parents spend with their children.

Psychological models of family processes also provide insights into how welfare and work changes may influence parental and family functioning. Psychological theory proposes that changes in income could alter the economic stability of families and hence influence parents’ psychological well-being and the quality of parenting they provide to their children (Chase-Lansdale & Pittman, 2002; Conger et al., 1992; McLoyd, 1998). For example, a loss of welfare benefits could increase financial strain, increase feelings of anxiety or depression, and prompt a decline in mothers’ consistent and supportive parenting behaviors. A host of research studies have provided support for various aspects of this model (Conger et al., 1992; McLoyd, Jayaratne, Ceballo, & Borquez, 1994; Yeung, Linver, & Brooks-Gunn, 2002). However, much of this research focuses on income as the primary exogenous variable, rather than considering the source of income. Research that specifically addresses employment and welfare as sources of income provides a less consistent story regarding whether moving women off welfare or into employment leads to more positive or more negative family processes and resources for children. Below we briefly review empirical evidence regarding these conceptual models and hypotheses.

Welfare and employment experiences and changes in income and financial strain

Beginning in the early 1990s, a number of experimental evaluations of state welfare reform programsfound that welfare-to-work transitions did not necessarily lead to substantial net increases in economic resources (see Slack et al., this volume). Longitudinal survey research has also found inconsistent results concerning whether moving off welfare and into employment increases families’ economic resources. For example, the Women’s Employment Study (WES), a longitudinal study following Michigan welfare recipients, has found that greater work effort among former and current welfare recipients is predictive of increases in income as well as declines in subjective experiences of economic strain, such as food insecurity, inability to afford medical care, and other types of material hardship (Danziger, Corcoran, Danziger, & Heflin, 2000). But other studies assessing the economic effects of welfare reform using national survey data have found mixed results, as income increased for some women but decreased for others as welfare receipt declined and employment increased (Bennett, Lu, & Song, 2002; Moffitt, 2002). More recently, some argue that more disadvantaged and less job-ready mothers are being moved off welfare, sometimes involuntarily (e.g., through sanctions), leading to increased financial hardship (Kalil, Seefeldt, & Wang, 2002) and possibly heightened psychological stress and instability (Cancian & Meyer, 2004; Lohman, Pittman, Coley, & Chase-Lansdale, 2004). In addition, national data from the late 1990s indicated significant declines in use of food stamps among families leaving welfare, leading to concerns over families’ food security (though these declines leveled off beginning in 2000 following policy changes to the Food Stamp program, see Zedlewski & Rader, 2005).

These mixed results suggest the need for continued study of low-income families’ economic resources following welfare reform. Since both access to in-kind supports (e.g., food stamps, Medicare) and costs (e.g., transportation, child care) change as families connect or disconnect from the welfare system and from employment, it is imperative to assess both families’ cash income as well as their experiences of economic insecurity in other areas, such as food, health, and more general financial hardship.

Mothers’ health and well-being

A second area of interest is whether welfare and employment experiences are related to low-income mothers’ psychological and physical well-being. Considerable research on maternal employment has found that employed mothers have better mental health and self-esteem than nonemployed mothers (Hoffman & Youngblade, 1999; McLoyd, 1998), although much of this work assessed primarily middle-class mothers who did not experience welfare reform. Studies of low-income mothers have also found significant relationships between employment and enhanced physical and psychological health, but it is important to acknowledge that this relationship may be bidirectional or endogenous. Results from the WES (e.g., Corcoran, Danziger, & Tolman, 2004) and others (Ensminger, 1995) suggest that poor physical and psychological health may contribute to welfare dependency and hamper low-income women’s efforts to sustain employment. One recent investigation (Raver, 2003) found that increased work effort over a two-year period predicted small increases in low-income mothers’ depressive symptoms. In contrast, the experimental literature has found no consistent significant effects of the welfare reform experiments on maternal depression (Crosby, Dowsett, Gennetian, & Huston, 2005; Huston, et al., 2005) or other aspects of parents’ mental health, such as self-esteem and mastery (see Morris et al., 2001 for an overview). Additional research and analytic techniques are needed to further tease apart the temporal and causal relationships between welfare and employment and psychological well-being among women.

Parenting

Finally, a host of studies have assessed links between employment and welfare experiences and parents’ provision of stimulating, supportive, or harsh parenting to children, with a mixed set of findings. Data from the experimental assessments have found no consistent effects of welfare reform programs on aspects of parenting including warmth, harshness, and cognitive stimulation (see Huston et al., 2005; Morris, Huston, Duncan, Crosby, & Bos, 2001). Data from the WES indicate that movements into employment while retaining welfare benefits predicted improvements in the quality of parenting (Dunifon, Kalil, & Danziger, 2003; see also Brooks-Gunn, Klebanov, Smith, & Lee, 2001; Smith, Brooks-Gunn, Kohen, & McCarton, 2001; Raver, 2003). On the other hand, movements off welfare— whether combined with entries into stable employment or not— were not related to parenting practices, controlling for mothers’ financial and psychological functioning.

Mixed findings are also apparent regarding the home environments provided to children, as indicated by measures of resources and cognitive stimulation in the home. In pre-welfare-reform studies, some research found that movements off welfare and into employment predicted improved home environments for children (Brooks-Gunn et al., 2001; Smith et al., 2001), whereas other research found no significant link (Wilson, Ellwood, & Brooks-Gunn, 1995). One possible explanation for the mixed effects of welfare and employment on parenting and the home environment is that the employment contexts experienced by low-income, low-skilled women often involve low wages, unstimulating conditions, and other stressful experiences which detract from, rather than enhance, parents’ ability to provide supportive home environments for their children (Menaghan & Parcel, 1995; Raver, 2003).

Two additional parenting constructs are important to assess: family routines and parents’ time spent with children. Research on poverty suggests that welfare dependency and lack of employment may contribute to inconsistent and less structured family routines (Wilson, 1996). On the other hand, high hours of employment may decrease parents’ time with their children (Becker, 1981) and hence their ability to routinely engage in family activities. Very limited research has empirically assessed these claims. The New Hope study found no effect of this work program on warm and structured parenting, a measure which included the regularity of routines (Huston et al., 2005). In a demographic overview, Bianchi (2000) found that dramatic increases in maternal employment over the past decades have not brought with them substantial decreases in the time mothers spend with their children. Because employed mothers tend to spend less time in leisure and other non-child activities, they compensate for time away from their children due to work (Chase-Lansdale et al., 2003).

Effects of child age

Parenting practices and other family processes may also differ according to children’s age and developmental needs. Thus, child age may serve as a moderator of the link between welfare and work experiences and family processes. For example, Chase-Lansdale et al. (2003) reported that low-income mothers’ movements into employment following welfare reform predicted decreases in time spent with young children, but little change in time spent with adolescents, who are in school and other activities much of the day regardless of parental work behavior. As another example, young children’s need for alternate care arrangements may increase the financial demands for families moving into employment, hence exacerbating links between new employment and financial strain for families of younger versus older children.

Research goals

In this investigation we attempt to address the inconsistencies and overcome some of the limitations in extant literature assessing whether welfare and work experiences induced by welfare reform are related to changes in family well-being. Using a large, representative sample of low-income families with children, we follow women’s welfare and work experiences over a two-year period from 1999 to 2001. Controlling for a host of mother, child, and family characteristics, as well as earlier levels of well-being, we assess whether stability and change in mothers’ employment and welfare experiences are linked to changes in economic stability, maternal well-being, and the parenting and home environments provided to children. We also assess whether these links may differ for families with younger versus older children. Through these analyses, we assess the numerous constructs identified by economic (Becker, 1981) and psychological (Chase-Lansdale & Pittman, 2002; Conger et al., 1992; McLoyd, 1998) models to be primary mechanisms through which parental economic activities affect children’s experiences and development.

Method

Sampling and data collection

Data are drawn from waves one and two of Welfare, Children, and Families: A Three-City Study, a longitudinal, multimethod study of the well-being of children and families in the wake of federal welfare reform. The Three-City Study includes, among other components, a household-based, stratified, random-sample survey with over 2,400 low-income children and their mothers in low-income neighborhoods in Boston, Chicago, and San Antonio. In 1999, over 40,000 households in randomly selected low-income neighborhoods (93% of block groups selected for sampling had a 20% or higher poverty rate) were screened, with a screening response rate of 90%. In selected families with household incomes of 200% or less of the poverty line and a child between the ages of 0 to 4 or 10 to 14 years, interviewers randomly selected one focal child per family and invited the focal child and his or her primary female caregiver (termed “mother” as nearly 90% of caregivers were biological mothers of the focal child) to participate. Eighty-three percent of selected families agreed to participate, resulting in an overall response rate of 74%. A second wave of interviews was completed with 88% of these families 16 months later, on average, in 2001. For further sampling details see Winston et al. (1999). To put the study into historical context, the data were collected approximately three to five years after the passage of the Personal Responsibility and Work Opportunity Reconciliation Act in 1996, and five to ten years after many states had implemented waivers to alter their welfare policies.

In each family, focal children and mothers participated in separate in-home interviews. Sections of the interviews which covered particularly sensitive topics (e.g., psychological distress) were conducted using Audio Computer-Assisted Self-Interviewing (ACASI), which has been shown to increase the validity of reporting on sensitive topics (Turner et al., 1998). Interviews were translated (and verified with back-translations) into Spanish, and this version was used by approximately 12% of mothers who reported their current primary language as Spanish. All respondents were paid for their participation in the study.

This paper focuses on mothers who participated in both waves of the survey and had valid data on the central variables of interest (weighted n = 1,974). Attrition analyses found that families who were excluded from analyses had lower incomes and more medical hardship, and were more likely to be on welfare and less likely to be employed 30 hours per week or more in the first wave than were included families. Excluded families were also less likely to be African American or to report English as their native language than were included families. In addition, however, excluded families reported less time apart from their children and more regular family routines than families included in analyses. Sample probability weights adjust (through the inverse of the sampling probability) for sample clustering, stratification, and nonresponse. Using these weights makes the sample representative of mothers of children and adolescents in low-income families living in low-income neighborhoods in the three cities.

Measures

Welfare and employment variables

The central independent variables of interest assess mothers’ welfare and employment experiences across the two interview waves. At each interview, mothers used a calendar format to report on their welfare receipt, employment status, and employment hours over the previous 2 years (or since the previous interview). At each wave, mothers were coded as “on welfare” if they reported welfare receipt in the majority of the past 11 months including at least 2 of the previous 3 months. The dichotomous on or off welfare variables were then combined across the two waves into a set of exclusive categories: Stable Welfare (on welfare at both waves), Into Welfare (off welfare wave 1; on welfare wave 2), Off Welfare (on welfare wave 1; off welfare wave 2), and No Welfare (off welfare both waves). Similarly, at each wave mothers were coded as “employed” if they reported paid employment for 30 hours or more per week in the majority of the past 11 months including at least 2 of the past 3 months. This definition of employment focuses on essentially full-time employment. Combining data from the two waves created the exclusive categories of Stable Employment, Into Employment, Out of Employment, and No Employment. In contrast to simple point-in-time assessments of welfare and employment, these definitions account for both recency and stability, assessing families’ predominant experiences over the 11 months preceding each interview (corresponding with the shortest time period between the two waves for survey families). Alternative definitions of employment (i.e., 20 hours per week and 40 hours per week) were also tested in parallel analyses, with robust results across definitions (not shown).

Economic well-being

Three sets of dependent variables were considered in the realms of economic well-being, maternal functioning, and parenting practices. All were measured identically at wave 1 and wave 2, and all were reported by mothers. Four variables assessed families’ economic well-being. Total household monthly income was calculated by ascertaining the income received for each individual in the household from a variety of sources (e.g., employment, welfare, food stamps, etc.). The income across household members was then summed.

Three composites assessed families’ experiences of economic hardship in various realms, with higher scores indicating greater hardship. A 6-item financial strain index (Coley & Chase-Lansdale, 2000) assessed whether mothers experienced difficulty paying bills and affording necessities (e.g., “Does your household have enough money to afford the kind of housing, food and clothing you feel you should have?” 1= definitely no to 4 = definitely yes; α₁ = .72, α₂ =.75). Medical hardship was assessed with two questions concerning whether the mother or the focal child was unable to receive medical care in the past 12 months because they could not afford it (0 = no; 1 = yes). Total scores ranged from 0 to 2. A reduced version of the U.S. Department of Agriculture food security measure (Nord, Andrews, & Carlson, 2004) asked whether the mother, other adults in the household, or the focal child had missed or reduced the size of meals due to financial hardship (8 items in total, e.g., “At any time in the past 12 months, did you cut the size of any of (child’s) meals because there wasn’t enough money for food?”). Items were combined into three dichotomous variables (0 = no hunger; 1 = hunger) for the child, mother, and all adults, then summed to produce a total household food insecurity score which ranged from 0 to 3.

Mothers’ health and well-being

Four variables assessed mothers’ psychological and physical functioning. Mothers completed the Rosenberg Self-Esteem Scale (Rosenberg, 1986), which includes 10 items measuring both positive (e.g., “On the whole, I am satisfied with myself”) and negative (e.g., “I feel I don’t have much to be proud of”) aspects of self-concept on a 4-point Likert scale (1 = strongly disagree; 4 = strongly agree). Total scores indicate more positive self-concept (α₁ = .76; α₂ = .77). Psychological distress was assessed with the 18-item version of the Brief Symptom Inventory (BSI 18, Derogatis, 2000), which asks how much respondents had been distressed or bothered by symptoms in the past 7 days using a 5-point scale ranging from “not at all” (0) to “extremely” (4). The BSI 18 contains three subscales: depression (e.g., “During the past 7 days, how much were you distressed or bothered by feeling hopeless about the future?” α₁ =.85; α₂ =.87), anxiety (e.g., “During the past 7 days, how much were you distressed or bothered by suddenly feeling scared for no reason?” α₁ =.84; α₂ = .84), and somatization (e.g., “During the past 7 days, how much were you distressed or bothered by nausea or upset stomach?” α₁ = .80; α₂ = .77). To measure respondents’ physical health, mothers were asked to rate their health (“In general, how is your health?”) on a 5-point scale of 1 (excellent) to 5 (poor).

Parenting

The final set of dependent variables tap into central aspects of parenting and the quality of the home environment provided to the focal child. The total time the mother was apart from the child, in hours, was assessed using a time diary from the day before the interview. Family routines, a measure of the regularity of strength-promoting family activities, were assessed through items selected from the Family Routines Inventory (Jensen, James, Boyce, & Hartnett, 1983). Four items were rated using a Likert scale (e.g., “Family eats dinner/supper at the same time each night.” 1 = almost never to 4 = always; α_T1 = .60; α_T2 = .65).

The level of cognitive stimulation in the home was assessed using mother report and interviewer observation items from the Home Observation for Measurement of the Environment-Short Form (HOME-SF, Center for Human Resource Research, 1993; revised from the original HOME, Caldwell & Bradley, 1984). Each item on the HOME–SF is scored dichotomously to indicate the presence or absence of a developmentally supportive aspect in the child’s home environment. For developmental appropriateness, the number and wording of the items differed depending upon child age (9 items for infants/toddlers, e.g., “How often do you read stories to child?”; 14 items for younger children, e.g., “Have you or another family member helped (child) to learn numbers?”; and 13 items for older children, e.g., “Does your family encourage (child) to start and keep doing hobbies?”). Scores were summed, age-standardized, and transformed into standard scores. The short form of the HOME has been found to have adequate validity in low-income and ethnically diverse samples (Bradley, Corwyn, Pipes McAdoo, & Garcia Coll, 2001).

Parenting quality was assessed using items from the Raising Children Checklist (Shumow, Vandell, & Posner, 1998). Mothers reported on 17 items concerning parenting behaviors directed toward the focal child (1= definitely true to 4 = definitely false), with some items varying by child age (e.g., for younger children, “I expect (child) to obey me without any questions asked.” For older children, “I don’t talk with (child) very much.” For all ages “I try to explain the reasons for the rules I make.”). Items were reversed when appropriate, averaged, and age-standardized to create a total parenting quality score, which reflects responsive and firm parenting practices (α_T1 = .70; α_T2 = .72 for younger children; α_T1 = .69; α_T2 = .71 for adolescents).

Demographic characteristics

Because numerous demographic and human capital characteristics often select people into welfare and employment experiences and may mask the relationship between welfare or employment and family processes, a variety of characteristics were used as covariates in the multivariate analyses. For characteristics that vary over time, both the wave 1 level of the variable as well as the change between waves 1 and 2 were included. The focal child’s sex was coded as male = 1, female = 0. Child age at wave 1 and change in age by wave 2 were coded in years. Race/ethnicity was coded into three categories, Hispanic of any race (omitted), non-Hispanic black, and non-Hispanic white/other. Mothers’ age was coded in years, and marital status through a dummy variable denoting married at wave 1, with additional dummy variables denoting movement into or out of marriage by wave 2. Mothers who were never married constituted the reference group. Mothers’ education was coded into 4 categories with dummy variables indicating a high school degree, a technical degree, or college educated or above, with less than high school omitted. Change in education dummy variables indicated whether the mother achieved a high school degree or greater than high school education by the second wave. Additional dummy variables indicated whether the mother was the biological mother of the focal child, and whether the primary caregiver changed by wave 2. Mothers also reported whether English was their first language, and their city of residence was coded as Boston or Chicago, with San Antonio as the omitted category. The number of minor children in the household was coded as a continuous variable, as was change in the number of minors by wave 2.

Analysis plan

Ordinary least squares (OLS) regression analyses with robust standard errors to adjust for sample clustering (at the block selection level) were used to estimate how mothers’ welfare and employment experiences over waves 1 and 2 of the survey were associated with changes in economic well-being, maternal functioning, and parenting. The base regression model is the following:

$$\begin{array}{l}{\text{Outcome}}_{2\text{i}}={\text{B}}_0+{\text{B}}_1{\text{OntoWel}}_{12\text{i}}+{\text{B}}_2{\text{OffWel}}_{12\text{i}}+{\text{B}}_3{\text{StableWel}}_{12\text{i}}+{\text{B}}_4{\text{IntoEmp}}_{12\text{i}}+\\ {\text{B}}_5{\text{OutOfEmp}}_{12\text{i}}+{\text{B}}_6{\text{StableEmp}}_{12\text{i}}+{\text{B}}_7{\text{Outcome}}_{1\text{i}}+{\text{B}}_8{\text{ControlVariables}}_{1\text{i}}+\\ {\text{B}}_9{\text{VaryingControlVariables}}_{12\text{i}}+{\epsilon }_i\end{array}$$

The first three independent variables in the model are those measuring welfare transitions and stability, and the coefficients on each depict the effect of that welfare pattern in comparison to the omitted category of being off welfare at both waves. Similarly, the next three variables assess the effect of each employment pattern in comparison to the omitted category of being not employed at both waves. Not on welfare over both waves and not employed over both waves were chosen as the omitted groups because they represent the largest subgroups (see Table 3). Adjusted Wald posthoc tests were used to assess slope differences between the welfare patterns included in the model (e.g., stable welfare versus off welfare) and between employment patterns (e.g., into employment versus out of employment). Controlling for the outcome from wave 1, this lagged model controls for unmeasured factors that have time-invariant effects on the family contexts under consideration (Cain, 1975), and hence provides less biased estimates of effects of welfare and employment patterns on changes in family processes over time (Kessler & Greenberg, 1981). In addition, the inclusion of child, mother, and family demographic variables helps to control for a number of characteristics known to be associated with selection into particular patterns of employment and welfare receipt and with family processes, thus decreasing concerns over spurious findings. As noted, these control variables included both time-invariant (e.g., child gender) and time-varying (e.g., maternal marital status) variables. For the latter, both the wave 1 level and the change from wave 1 to wave 2 were included in order to capture individual and family demographics over both waves of data. Although a number of important child and mother characteristics are included as covariates, the models cannot control for unmeasured characteristics that might be correlated with welfare or employment transitions as well as with changes in family processes. It also is important to note that the models cannot determine the ordering of effects. For example, a change in mothers’ psychological distress may precipitate a change in employment status. Hence, results should be interpreted as correlational rather than causative, with the acknowledgment that multiple interpretations are plausible.

Table 3.

Descriptives of family process variables by welfare and employment groups

	No welfare				Onto welfare				Off welfare				Stable welfare
	Wave 1 Mean	SD	Wave 2 Mean	SD	Wave 1 Mean	SD	Wave 2 Mean	SD	Wave 1 Mean	SD	Wave 2 Mean	SD	Wave 1 Mean	SD	Wave 2 Mean	SD
Income ($)	1211.44	974.23	1937.37	1385.87	950.29	777.66	1180.19	973.45	1047.87	686.39	1429.98	965.72	935.24	548.16	1146.52	811.29
Fin. strain	−.13	.66	−.08	.69.	.02	.64	−.06	.73	−.01	.70	−.23	.68	.0346	.63	−.07	.66
Med. hard.	.21	.49	.18.	.43	.32	.67	.03	.18	.17	.47	.19	.47	.07	.27	.04	.23
Food insec.	.24	.63	.21	.62	.25	.56	.19	.59	.21	.63	.19	.62	.31	.71	.16	.57
Self-esteem	43.37	6.52	44.04	6.71	41.66	8.65	42.73	7.48	43.04	6.06	43.26	6.52	41.61	8.00	42.37	7.15
Depression	2.29	3.23	2.08	3.22	3.39	4.08	3.63	4.13	2.68	3.54	2.56	3.34	4.26	4.77	4.21	5.39
Anxiety	1.71	2.81	1.48	2.65	2.26	3.19	2.72	4.13	2.00	3.43	1.96	2.90	3.29	4.06	3.28	4.14
Somatization	1.86	2.81	1.66	2.54	2.08	2.51	2.29	3.12	2.08	3.01	2.30	2.63	3.18	3.62	3.46	4.06
Poor health	2.69	1.13	2.65	1.07	2.48	1.06	2.89	1.02	2.54	1.10	2.47	1.12	2.89	1.23	3.04	1.20
Hours apart	5.58	5.27	7.02	5.46	3.66	5.64	4.74	4.87	4.50	5.66	6.86	5.46	3.23	4.49	5.72	5.64
F. routines	2.75	.65	2.74	.68	2.89	.73	2.90	.78	2.79	.65	2.77	.65	2.94	.74	2.88	.74
HOME	101.16	14.32	100.67	15.54	102.40	11.28	96.81	16.35	99.95	13.71	99.50	13.47	96.64	16.18	97.66	14.42
Parenting	3.45	.33	1.55	.33	3.32	.38	1.61	.40	3.51	.28	1.52	.31	3.34	.35	1.70	.36
	No employment				Into employment				Out of employment				Stable employment
	Wave 1 Mean	SD	Wave 2 Mean	SD	Wave 1 Mean	SD	Wave 2 Mean	SD	Wave 1 Mean	SD	Wave 2 Mean	SD	Wave 1 Mean	SD	Wave 2 Mean	SD
Income ($)	1072.68	775.24	1371.64	1120.34	1044.08	908.09	2138.93	1397.61	998.12	971.54	1531.51	994.24	1559.60	1032.89	2334.80	1383.37
Fin. strain	−.04	.68	−.05	.67	−.22	.64	−.27	.69	−.04	.71	.09	.75	−.18	.63	−.12	.67
Med. hard.	.20	.50	.16	.41	.14	.43	.14	.41	.11	.32	.15	.39	.14	.37	.10	.39
Food insec.	.30	.70	.20	.65	.18	.52	.08	.39	.25	.60	.39	.87	.15	.47	.23	.63
Self-esteem	42.54	7.14	42.87	7.15	41.92	7.14	44.73	6.00	44.20	5.93	43.78	6.51	44.99	5.57	45.62	5.34
Depression	2.84	4.00	2.89	4.33	2.68	3.43	2.01	3.08	2.57	3.43	2.94	3.32	2.36	2.97	1.93	2.56
Anxiety	2.21	3.50	2.14	3.52	2.14	3.26	1.94	2.80	1.95	2.89	1.80	3.03	1.46	2.10	1.07	1.76
Somatization	2.39	3.30	2.35	3.30	1.87	2.66	1.95	2.52	2.00	2.95	1.97	2.62	1.48	2.14	1.43	2.14
Poor health	2.83	1.18	2.84	1.11	2.38	1.06	2.46	1.03	2.78	1.02	2.78	1.00	2.33	.94	2.41	1.09
Hours apart	4.00	5.10	5.94	5.50	4.51	4.82	7.19	4.98	6.89	4.59	6.58	6.03	7.88	5.19	8.61	5.53
F. routines	2.82	.68	2.80	.72	2.82	.71	2.79	.65	2.77	.70	2.76	.64	2.74	.65	2.71	.68
HOME	99.33	14.59	99.31	14.74	100.84	13.70	100.91	14.23	101.45	13.02	102.01	14.18	102.40	16.18	98.99	18.06
Parenting	3.43	.31	1.60	.33	3.41	.36	1.46	.28	3.38	.43	1.66	.36	3.50	.32	1.55	.37

Results

Descriptive statistics

Table 1 presents descriptive statistics on the sample. At the first interview, most of the families were poor, with an average monthly income of $1,122, putting the majority of families well below the federal poverty line. It is notable, however, that the average monthly household income increased by $562 across the two waves. In contrast, mother reports of various types of economic hardship showed little mean change across the waves. Measures of maternal well-being and parenting also showed little mean change, with the exception that on average, mothers increased their time apart from the focal child by nearly two hours per day. Although space constraints prohibit exploration of the descriptive results in depth, Table 2 further presents the family process variables across the four welfare and the four employment groups.

Table 1.

Descriptives of study variables.

	Wave 1 Mean	SD	Wave 2 Mean	SD
Child white/other	.06	.24
Child black	.44	.50
Chicago	.33	.47
Boston	.33	.47
Child male	.50	.50
Child age (years)	7.53	5.12	1.37	5.13
Mother age (years)	33.16	9.95	34.51	9.88
Mother married	.29	.46	.31	.46
Mother education High School	.49	.50	.55	.50
Mother education beyond High School	.10	.29	.14	.34
Mother education college degree or above	.03	.16	.03	.16
Biological mother	.89	.31
English is 1st language	.70	.46
Number of minors	3.14	1.61	3.08	1.62
Total household monthly income	1122.05	874.16	1684.04	1271.37
Financial strain	−.09	.66	−.08	.69
Medical hardship	.18	.47	.15	.40
Food insecurity	.25	.64	.20	.63
Mother self-esteem	42.93	6.90	43.56	6.83
Mother depression	2.75	3.70	2.59	3.82
Mother anxiety	2.06	3.21	1.92	3.16
Mother somatization	2.13	3.01	2.09	2.97
Mother poor physical health	2.69	1.15	2.70	1.11
Total hours apart from child	4.92	5.31	6.66	5.49
Family routines	2.79	.68	2.78	.69
HOME	100.34	14.51	99.75	15.18
Total parenting quality	.06	1.00	−.02	1.03

Table 2.

Cross tabulation of welfare and employment patterns

		No Welfare	Onto Welfare	Off Welfare	Stable Welfare
		63.1%	5.9%	14.4%	16.7%
No Employment	57.6%	30.6%	3.8%	8.8%	14.4%
Into Employment	19.8%	12.9%	1.1%	4.6%	1.2%
Out of Employment	7.2%	5.6%	0.6%	0.3%	0.8%
Stable Employment	15.4%	14.0%	0.4%	0.7%	0.3%

Note. Total percentages for exclusive welfare and employment patterns are presented in the first row and first column, respectively. Combinations of welfare and employment patterns are presented in the internal rows and columns. Employment is coded as working 30 hours or more per week.

Table 3 presents a descriptive portrait of mothers’ welfare and employment patterns. Sixty-three percent of mothers reported being stably off welfare over the two waves of the survey, a much higher proportion than in samples of families drawn from the welfare rolls. Seventeen percent of mothers received welfare at both points, while 14% transitioned off welfare. A small number of mothers, only 6%, reported moving onto welfare. Regarding employment of 30 hours per week or more, 58% of the sample was not working at either interview. On the other hand, 15% of mothers maintained stable employment, 20% of mothers moved into employment, and 7% of mothers moved out of employment over the nearly two-year period.

Table 3 also presents cross tabulations between the welfare and employment patterns, indicating the independence of these two income sources. The first notable result from this cross-tabulation is that nearly 1/3 of the sample (31%) were neither receiving welfare nor employed full-time at both time points. Fewer mothers remained off welfare and were stably in the labor force (14%), or were moving into (13 %) or out of (6 %) 30 hours or more per week employment. A second notable pattern is that within the subsample of mothers moving off welfare, only about a third moved into employment (5% of the total sample). Finally, sizable groups were not employed at either time point, but remained on welfare (14 %), or moved off welfare (9 %). In short, these results suggest both that welfare and employment are operating relatively independently in this sample, and that combining welfare and employment status into one coherent set of mutually exclusive categories (e.g., from welfare to employment, from not welfare to employment, etc.) is not feasible. Additional analyses (not shown) tested interactive multivariate models, and also supported modeling welfare and employment experiences independently.

Prior to presenting the multivariate models, a final series of descriptive analyses were conducted to explore the nearly one third of the sample who reported neither full-time employment nor welfare receipt at both waves of the survey (data not shown). Within a low-income sample, this group is particularly interesting due to uncertainty about the sources of their support. Exploration of their employment experiences revealed that 22% worked for any number of hours in recent months, a much lower rate of employment than other mothers in the sample. The no welfare and no full-time work group also were more likely to be Hispanic and less likely to report English as their native language. They were also older and more likely to be married than other women. Total household income did not differ for these mothers, but they did report greater hardship (e.g., higher medical hardship and more food insecurity) and worse physical health. Yet they also had lower levels of psychological distress (depression and anxiety) than did other mothers in the sample. Finally, mothers who neither received welfare nor were employed across both waves spent more time with their child, but also had less stable family routines compared with other mothers in the sample. In short, it appears that the mothers who reported no individual, direct source of income from two primary income sources (employment and welfare) showed both resiliencies and weaknesses in comparison to other women in the sample.

Predicting changes in family economic well-being

Table 4 presents results from the lagged multivariate regression models predicting family economic well-being. In all models, time-varying and time-invariant child, mother, and family demographic variables were controlled as well as the wave 1 value of the dependent variable, in order to isolate relationships between maternal employment and welfare experiences and changes in family processes. It is also important to keep in mind that employment refers primarily to full-time employment, that is, 30 hours per week or more for the majority of the 11 months preceding each interview; similarly, welfare receipt refers to receipt in the majority of the 11 months preceding each interview. Regarding household income, movements into employment and stable employment predicted substantial increases in monthly household income in comparison to mothers who were stably not employed or who lost employment, with relative gains of over $800 per month for movements into employment and gains of over $600 per month for stable employment. Movements into employment also predicted relative declines in financial strain and food insecurity in comparison to most other employment patterns. No significant differences were found in families’ experiences of medical hardship in relation to mothers’ employment transitions.

Table 4.

Lagged OLS regression models predicting family economic Well-being

	Monthly income ($) Coef. (SE)	Financial strain Coef. (SE)	Medical hardship Coef. (SE)	Food insecurity Coef. (SE)
Into employment	824.64 (132.73) *** a	−.16 (.08)* a	−.05 (.05)	−.11 (.05)* a b
Out of employment	123.42 (172.63) a b	.13 (.11) a	.00 (.07)	.21 (.15) a
Stable employment	631.21 (152.96) *** b	−.05 (.06)	−.07 (.06)	.03 (.06) b
On to welfare	−432.49 (207.44) *	−.10 (.13)	−.17 (.04) *** a	−.02 (.06)
Off welfare	−360.96 (142.37) *	−.11 (.07) +b	−.03 (.05) a b	.08 (.06)
Stable welfare	−420.13 (139.90) **	.09 (.07) b	−.17 (.04) *** b	−.03 (.06)
F of Model	9.64***	12.16***	2.51***	2.60***
R2	.29	.27	.12	.15
N	1903	1894	1901	1898

Note:

⁺p < .10,

^*p <.05,

^**p <.01,

^***p <.001.

Employment groups are compared to the omitted category of no employment; welfare groups are compared to the omitted category of No Welfare. Within each column, groups with shared superscript letters are different from each other at the p < .05 level (or if followed by + at p < .10). Employment is coded as working 30 hours or more per week. All analyses controlled for the wave 1 value of the dependent variable, as well as city, mother language, mother age, child gender, race/ethnicity, child age and change in age, mother marital status and change in marital status, mother education and change in education, mother relationship to child and change in relationship, as well as number of minors in household and change in number of minors.

Mothers’ welfare transitions also predicted economic well-being. Like employment experiences, consistent results were found for household income. All experiences with welfare— moving onto welfare, staying on welfare, and moving off welfare— predicted relative declines in household income of approximately $400 per month in comparison to families who were stably off welfare over the two waves. Stable welfare receipt also predicted relative increases in financial strain, in comparison to families who moved off welfare whose financial strain declined. On the other hand, both stable welfare and movements onto welfare predicted relative declines in medical hardship in comparison to movements off or never receiving welfare, perhaps due to concomitant access to public health insurance programs. No differences in families’ food insecurity related to welfare experiences were detected.

In summary, these results suggest that movements into long-term full-time employment is linked with substantial increases in household income and declines in economic hardship. Sustained employment also helped to increase income, but did not predict improvements in financial hardship. On the other hand, families receiving welfare had particularly low levels of financial resources, experiencing income loss and relative increases in experiences of financial strain. New and sustained use of welfare, however, also predicted lower medical hardship.

Predicting changes in mothers’ health and well-being

Table 5 presents results for the regression models predicting measures of maternal psychological and physical health. Similar to the economic stability findings, new and continued full-time employment were linked to modest improvements in well-being. Specifically, movements into employment predicted improvements in self-esteem and declines in depression, in comparison to mothers moving out of or remaining out of employment. Similarly, stable employment predicted relative increases in self-esteem, decreases in anxiety, and minor decreases in depression and poor physical health in comparison to stable lack of employment. Employment experiences were not significantly related to mothers’ somatic symptoms.

Table 5.

Lagged OLS regression models predicting maternal functioning

	Self-esteem Coef. (SE)	Depression Coef. (SE)	Anxiety Coef. (SE)	Somatitization Coef. (SE)	Poor health Coef. (SE)
Into employment	2.16 (.83)** a+	−.85 (.23)*** a	−.01 (.22)	.11 (.25)	−.14 (.11)
Out of employment	.04 (1.01) a+	.09 (.46) a	−.29 (.24)	−.11 (.32)	−.05 (.13)
Stable employment	1.52 (.68)*	−.48 (.27)+	−.39 (.19)*	−.11 (.23)	−.22 (.12)+
On to welfare	−1.56 (1.15)	.75 (.51)	1.15 (.47)* a+	1.11 (.48)*	.35 (.26) b+
Off welfare	−.41 (.72)	.20 (.30)	.27 (.24) a+	.41 (.24)+	−.13 (.12) a b+
Stable welfare	.06 (.72)	.53 (.38)	.66 (.29)*	1.09 (.35)**	.27 (.12)* a
F of Model	7.97***	9.32***	11.25***	9.06***	10.44***
R2	.23	.43	.43	.30	.25
N	1892	1897	1898	1898	1903

Note:

⁺p < .10,

^*p <.05,

^**p <.01,

^***p <.001.

Employment groups are compared to the omitted category of no employment; welfare groups are compared to the omitted category of no welfare. Within each column, groups with shared superscript letters are different from each other at the p < .05 level (or if followed by + at p < .10). Employment is coded as working 30 hours or more per week. All analyses controlled for the wave 1 value of the dependent variable, as well as city, mother language, mother age, child gender, race/ethnicity, child age and change in age, mother marital status and change in marital status, mother education and change in education, mother relationship to child and change in relationship, as well as number of minors in household and change in number of minors.

In contrast to the beneficial findings for full-time employment, welfare receipt was linked with poor maternal functioning. Both entrances onto welfare and sustained welfare receipt were related to increases in reports of anxiety, somatization, and poor physical health. In summary, results suggest that continued or new welfare receipt was related to increased psychological and physical health problems, while continued or new employment predicted improved psychological functioning.

Predicting changes in parenting

Table 6 presents results for parenting and home environment variables. With the exception of mothers’ time with children, these results show very limited significant relationships between mothers’ employment and welfare experiences and the resources and supports they provide to their children. As shown in the first column, stable full-time employment was related to a 2-hour increase in the time caregivers were apart from the focal child in comparison to mothers who moved out of or were stably out of employment. In contrast, movements onto welfare were related to a similarly sized decrease in the time caregivers and children were apart in comparison to all other welfare groups. And yet in the measures tapping into the quality of parenting and the home environment, almost no significant results emerged. Welfare experiences were not related to changes in family routines, the cognitive stimulation of the home environment, or the overall quality of parenting provided to children. For employment, stable employment predicted a relative decline in the quality of the home environment in comparison to movements into or out of employment, yet movements into employment predicted relative declines in the quality of parenting. Additional subgroup analyses focusing on adolescent reports of parental monitoring and harsh parenting also showed no significant results (results not shown).

Table 6.

Lagged OLS regression models predicting parenting

	Total time apart Coef. (SE)	Family routines Coef. (SE)	Home Coef. (SE)	Total parenting quality Coef. (SE)
Into employment	1.00 (.59)+	−.04 (.07)	.65 (1.55) a	−.15 (.12) a
Out of employment	.29 (.84) a	.01 (.10)	1.47 (2.32) b+	.16 (.12) a
Stable employment	2.19 (.60)*** a	−.03 (.08)	−3.03 (1.86) a b+	.02 (.11)
On to welfare	−2.29 (.66)** b c	.02 (.15)	−2.57 (.42)	−.07 (.20)
Off welfare	.35 (.59)c	.03 (.07)	.84 (1.66)	.08 (.11)
Stable welfare	−.20 (.59)b	.09 (.07)	−.53 (1.74)	.15 (.11)
F of Model	6.24***	8.16***	4.54***	8.32***
R2	.16	.23	.15	.25
N	1905	1887	1818	1864

Note:

⁺p < .10,

^*p <.05,

^**p <.01,

^***p <.001.

Employment groups are compared to the omitted category of no employment; welfare groups are compared to the omitted category of no welfare. Within each column, groups with shared superscript letters are different from each other at the p < .05 level (or if followed by + at p < .10). Employment is coded as working 30 hours or more per week. All analyses controlled for the wave 1 value of the dependent variable, as well as city, mother language, mother age, child gender, race/ethnicity, child age and change in age, mother marital status and change in marital status, mother education and change in education, mother relationship to child and change in relationship, as well as number of minors in household and change in number of minors.

Differences by child age

In a final set of analyses we assessed whether the aforementioned patterns of results were moderated by child age, that is, whether maternal welfare and employment experiences were differentially related to parenting, maternal functioning, or family economic stability in families with infant or pre-school-age focal children in comparison to families with adolescent focal children. Given the different developmental demands of very young versus early adolescent children, we hypothesized that some patterns might vary. Results of regression analyses which included interactions between maternal employment and welfare patterns and children’s age (coded as young versus early adolescent; results not shown) revealed a notable lack of age moderation effects. Over all the 13 family process variables assessed, significant interactions emerged at about the frequency expected due to chance. In sum, these results suggest that the links between mothers’ welfare and work experiences and family processes were similar across families with younger and older children.

Discussion

Decades of scholarship on poverty, welfare, and maternal employment led to divergent hypotheses concerning how welfare reform, and resultant changes in parents’ work and welfare behaviors, would affect families and children. As welfare reform was instituted in the late 1990s in combination with expanded work supports and substantial U.S. economic growth, welfare rolls plummeted and thousands of poor women moved into the labor force. And yet, little evidence has emerged of significant impacts on children’s development and well-being. This study sought to focus on the intervening processes through which mothers’ welfare and employment experiences might influence children, assessing whether, in the short term, these experiences predict changes in the proximal family processes indicated as important influences on children’s development. Our results support these claims for the proximal processes of economic stability, maternal well-being, and time with children, but not for the quality of parenting.

In summary, our results suggest that in the years following welfare reform, low-income mothers’ welfare and employment experiences were significant predictors of their families’ economic stability and of their own psychological functioning. Moreover, results were robust across numerous measures of economic and psychological well-being. Mothers who moved from unemployment into sustained employment of 30 hours per week or more reported increases in total monthly household income of nearly $800. Although other research has found that low-income mothers’ and welfare recipients’ economic gains from employment are often counteracted by losses in other in-kind and monetary benefits (e.g., Bloom & Michalopoulos, 2001), our results indicate that mothers moving into employment also reported declines in financial strain and food insecurity. Moreover, mothers who moved into employment reported feeling better about themselves. They showed relative increases in self-esteem and declines in depressive symptoms in comparison to mothers who remained out of the labor force.

Stable employment, that is full-time employment sustained over both waves of data collection, was also related to substantial increases in total household income as well as to improved self-esteem, lowered anxiety, and marginally lowered depression and poor physical health. Interestingly, stable employment was not related to changes in other aspects of economic hardship. Possibly these women were losing eligibility for other public and in-kind supports, and thus increased work income was offset by increased costs, leading to no substantive change in more experiential and subjective measures of economic hardship. It is important to reiterate that even these women, who sustained essentially full-time employment over a long time frame, were still disadvantaged, with low family incomes.

The effects of employment noted above were found when we controlled for mothers’ welfare experiences over the same time frame, as well as for other individual and family characteristics. Although women’s welfare and work experiences are often viewed as substitutes, this representative sample of low-income, urban families with children showed otherwise. Over 30% of the sample neither received welfare nor was employed full-time at both waves of the survey. Moreover, mothers moving off welfare were almost twice as likely to remain unemployed as they were to move into sustained employment. Mothers who accessed neither welfare nor employment showed some distinct characteristics from other women in the sample—they were more likely to be married and Hispanic, but did not differ in total household income, suggesting the likelihood of other regular sources of income (e.g., from spouses, other household members, or other social programs). Yet experientially, women with neither welfare nor employment showed both resiliencies (better psychological functioning) and weaknesses (greater physical health problems, medical hardship, and food insecurity). These descriptive results reiterate the importance of assessing families’ economic and household contexts in a coordinated and holistic fashion.

A central feature of this study was that welfare and employment experiences were modeled separately but concurrently, with results showing that each had significant, independent relationships with changes in mothers’ economic and psychological well-being. Indeed, controlling for employment, mothers’ welfare experiences were also linked to changes in income and mothers’ psychological and physical functioning. All experiences with welfare over the two waves of the survey, i.e., stable receipt, as well as entrances into and exits from welfare, predicted relative declines in household income in comparison to families never on welfare. Two other patterns of findings in relation to mothers’ welfare experiences present an interesting dichotomy. First, families who used welfare reported less difficulty in accessing and affording medical care for themselves and their families, potentially related to public health insurance programs. Second, however, new and continued welfare receipt was also associated with worsening physical health as well as heightened symptoms of anxiety and somatization. In short, the same groups with increasing health concerns also reported less difficulty accessing medical care. These findings highlight continuing concerns over low-income families’ access to stable and quality medical services (Singer & Riff, 2001), and provide sobering evidence that mothers entering and remaining on welfare show significant detriments in their own health and well-being that have the potential for long-term negative impacts on families and children. Future research efforts are needed to assess how the medical services being accessed by welfare families can best meet the physical and psychological needs of this highly disadvantaged population.

The links between welfare and work and mothers’ well-being also raise issues of directionality and causality. As noted earlier, caution is warranted in interpreting these findings, and women’s psychological well-being and employment and welfare experiences may be particularly endogenous. That is, for example, high levels of depression, anxiety, or health problems may inhibit a woman’s entrance into stable employment and encourage continued welfare receipt. Similarly, the heightened stigma of welfare receipt in this era of sustained efforts to diminish the welfare rolls (Cherlin et al., 2000) may negatively influence recipients’ health. High levels of self-esteem may provide an important impetus that allows low-skilled women to sustain employment under conditions that are often demanding and provide little support or stimulation. On the other hand, pride and enhanced consistency and resources gained from employment may improve women’s self-esteem and psychological well-being, whereas the continued demands and looming time limits of welfare may lead to anxiety and distress. The lagged regression models used in the analyses, which control for a variety of characteristics as well as earlier levels of mothers’ health and well-being, help to control for related factors and issues of temporality. Yet the models do not specifically assess bidirectionality, nor can they determine causation. Further study with a variety of methods is needed to ascertain causal processes in the complex relationships between women’s well-being and their welfare and work experiences.

In contrast to the patterns of relationships with economic and psychological well-being, mothers’ welfare and work experiences were primarily unrelated to changes in the proximal environments and parenting provided to children. No significant relationships were found between mothers’ welfare experiences over a two-year period and changes in family routines, the home environment, or parenting quality provided to young children and early adolescents. Relationships between these parenting variables and maternal employment experiences were infrequent and inconsistent. Only mothers’ time with children showed significant links, with movements onto welfare predicting relative decreases, and stable employment predicting relative increases, in mothers’ time spent apart from their children. Without attention to the quality of that time, as well as to children’s experiences in nonparental care, however, it is probable that these changes hold little notable influence for children’s development. Moreover, economic theory suggests that changes in mothers’ time with children must be considered in concert with changes in income. In the Three-City Study, the benefits from increases in income due to employment may have been canceled out by decreases in the time that mothers spend with children (Chase-Lansdale et al., 2003). Similarly, increases in time with children due to welfare receipt may not have had much influence on child development because of declines in income.

As Chase-Lansdale and Pittman (2002) noted in a review of the welfare reform literature, parenting is a complex, personal, and multi-determined phenomenon. Given the host of influences on parenting (child and parental characteristics, personal and cultural norms and beliefs, as well as environmental constraints, supports, and experiences), and the difficulty that broad intervention and policy efforts have had in changing parental behavior (Brooks-Gunn, Berlin, & Fuligni, 2000), it is perhaps not surprising that parenting practices are relatively immune to substantial influence from distal sources such as welfare receipt or employment. Rather, attention to parents’ experiences or the characteristics and types of jobs gained by low-income parents may prove more illuminating in understanding effects on parenting behaviors. Extant scholarship suggests that parents’ experiences in employment, such as the stimulation and supportiveness of the work environment (Parcel & Menaghan, 1995), the stability of work schedules (Han, 2005; Presser, 2003), or the strain resulting from balancing work and family roles (Morris & Coley, 2004) may influence parenting practices and the home environment. Related research also notes that parents’ use of additional services, such as child care and after-school programs, may serve as an important conduit through which welfare reform and maternal employment experiences influence children (Gennetian et al., 2004; Huston et al., 2005). Finally, psychological models of family stress (Conger et al., 1992; McLoyd, 1998) suggest that the link between economic resources and children’s development is mediated through complex systems of interrelated family processes. Future research should seek to assess whether this model is replicated within populations affected by welfare reform.

In summary, this study suggests that women’s new and sustained employment experiences were significant correlates of enhanced economic stability and improved psychological well-being, but were not substantially related to the quality of the parenting and home environments provided to children. Welfare experiences, in contrast, were related to diminished economic stability (with the exception of access to health care) and declines in psychological and physical health, but also were unrelated to parenting and home environments. Importantly, these relationships were notably similar across families with younger versus older children. Focused on an approximately two-year period, the present study indicates that mothers’ welfare and work experiences are related to more distal, but not the most proximal, influences on children, hence providing some potential insight into the minimal research results concerning welfare reform’s effects on children’s development. Over a longer time frame, however, the more distal processes examined here may in turn predict changes in contexts and experiences more proximal to children’s lives, thereby having longer-term effects on children’s trajectories.

In closing, it is essential to keep in mind that the data presented here were collected under very favorable economic and social conditions, during years of a steady and significant economic expansion when jobs were plentiful and incomes were rising for all but the most severely disadvantaged. Indeed, the average income of the families in our sample increased by approximately 50% over the two waves of the survey. As the economy declined over the initial years of the twenty-first century, employment opportunities for workers at the lowest rungs became more constricted, and some states experienced increases in the welfare rolls. Moreover, as welfare reform continues to evolve and more families are affected by additional policy levers, such as time limits and restricted access to in-kind supports after leaving welfare, relationships among mothers’ welfare and work experiences and their economic hardships, personal well-being, and parenting practices may change. Distinctions between women who become firmly entrenched in the labor market and advance into more stable and better-paying jobs may show continued improvement. On the other hand, the multiple stressors, instabilities, and challenges of poverty may continue to impede the most disadvantaged families.