Early Impacts of Room to Grow: A Multifaceted Intervention Supporting Parents and Children Age Zero to Three

Abstract

Children experiencing poverty or low incomes fare worse than their more advantaged peers on a host of developmental and educational outcomes. Interventions have focused on strengthening parenting in families with young children, when supports appear to be most critical. But most parenting programs for low-income families fail to address parents’ economic needs, which almost always take precedence relative to broader educational or developmental goals. In this article, we describe the early results of a multifaceted intervention aimed at supporting parents, infants, and toddlers in the first three years of life. The Room to Grow program provides parents, primarily mothers, with support from a clinical social worker, connections to community referrals, and up to $10,000 in material support for the baby in the form of in-kind assistance such as clothes, books, toys, strollers, and other necessities. The current study examines proximal outcomes of the intervention after one year using a randomized controlled trial evaluation design. The study finds that early impacts on proximal outcomes are uniformly positive, especially with regards to the presence of books and developmental goods in the home, developmentally-oriented parenting outcomes, and reduced stress and aggravation in the domain of parenting.

1. Introduction

Children growing up in low-income households fare worse than their more advantaged peers across a wide spectrum of developmental and educational outcomes from birth to early childhood and beyond (Duncan & Brooks-Gunn, 1997; Duncan, Morris, & Rodriguez, 2011; Chaudry & Wimer, 2016). Based on 2018 data from the United States Census Bureau (Semega, Kollar, Creamer, & Mohanty, 2019), 16.2% of children live in poverty and 6.9% live in deep poverty (income under 50% of the poverty line). Children newborn to age five are the poorest, with 3.5 million being poor in 2018, and 1.6 million of those young children living in deep poverty. Black and Hispanic children continue to suffer disproportionately from poverty and low incomes (Semega et al., 2019). Racism, segregation, and social inequality contribute to family poverty and disparities in children’s health and development through the lifespan (Bailey et al., 2017).

The relationship between family poverty and child development seems to be most critical in a child’s earliest years, when brain development is most rapid and when children spend most of their time with their parents and families (Blair & Raver, 2016). Parents play a crucial role in laying the groundwork for their children’s healthy cognitive and socio-emotional development (Johnson, Riis & Noble, 2016). However, poverty can affect families’ ability to support the child’s development by increasing stress, affecting parents’ well-being, and reducing their human capital (Finegood & Blair, 2017).

Cumulative evidence confirms the importance of intervening early, beginning in the prenatal period, to support families facing poverty and enhance children’s positive development (Gassman-Pines & Yoshikawa, 2006; Minkovitz, O’Neill, & Duggan, 2016; Love, Chazan-Cohen, Raikes, & Brooks-Gunn, 2013). Interventions that focus solely on families won’t be able to change larger social inequalities and structural racism that contribute to perpetuating poverty and health disparities. Nonetheless, evidence from the past decade has documented the effectiveness of home visiting programs (Olds, 2010) and parenting programs in enhancing families’ ability to support children living in poverty (see Morris et al., 2017 for a review). Yet, many programs commonly face high attrition rates of 50% or higher (Gomby, 2000), which compromises the interventions’ potential benefits. In addition, while most zero-to-three programs focus on improving parenting and parent-child relationships, they unintentionally ignore parent’s economic and material needs, which can compromise both program participation and success. Yet we know from the welfare reform experiments of the 1990s and more recent local experiments (e.g., New Hope (Huston et al., 2005) that young children’s development is improved when their families’ financial situation improves (Hill & Morris, 2008).

Moreover, formal federal programming for low-income families with children newborn to age 3 continues to be scarce (Love, Chazan-Cohen, Raikes & Brooks-Gunn, 2013). For example, Early Head Start (EHS), a federally funded program that provides in-home and center-based services to pregnant women and their children in the first three years of life with the aim of promoting positive parenting practices and connecting families to outside services, only serves approximately 3% of eligible infants and toddlers. While sparse community-based programs that support families living in poverty with young infants exist, gold-standard evidence of the impact of such interventions is still limited. Further evidence is needed to better understand the cumulative impact of family support interventions that focuses on families’ strength and supports engagement with community services in this critical window of development (Love, Chazan-Cohen, Raikes & Brooks-Gunn, 2013).

In this study, we present initial experimental results on the efficacy of Room to Grow (RtG), a research-informed program for parents of young children aged zero to three. RtG is novel in its combination of one-on-one sessions at RtG’s family centers (site-based rather than in-home) with a clinical social worker every three months that provide structured parent coaching, community referrals/connections, and concrete financial assistance in the form of material goods (e.g. books, toys, clothing, baby equipment) that total approximately $10,000 over the three-year long program. In contrast to other programs aimed at supporting low-income families focusing on only parenting and family well-being, Room to Grow also embeds the provision of concrete material assistance to enhance the effectiveness of parenting support and community referrals to low-income families by freeing up scarce resources and facilitating the materials to engage in stimulating interactions. In addition, it is based on a strengths-based approach, personalized programming, respect for cultural differences, and reciprocal relationships with clinicians which have been considered essential components to build trust and participation necessary for program success (Henshaw et al., 2011; Ginn et al., 2017).

1.1. An integrated framework of poverty’s effects on families and children

Room to Grow is based on a large body of research supporting two main models related to poverty and family functioning. These models have shed light on the mechanisms by which family poverty affect children’s cognitive and emotional development. The stress-induced family processes model outlines the relationship between low income, parental stress, compromised parental relationships, and child development (Elder, 1974; McLoyd, 1998). Low income can exacerbate parental psychological distress and increase the risk of developing mental health disorders such as depression, which have been shown to reduce parents’ capacity to engage in warm, responsive and stimulating interactions with children (Yoshikawa, Aber & Beardslee, 2012). Poverty and parental stress can also affect parents’ sense of confidence (Coleman & Karraker, 1998; Raikes & Thompson, 2005) and their belief in their ability to effectively perform parenting-related tasks (Farkas & Valdés, 2010; Jones & Prinz, 2005), and have a positive influence on their children’s development. In turn, low parental confidence has also been associated with less sensitive, warm, and responsive parenting, and increased use of harsh discipline and hostile behaviors towards their child (Sanders & Woolley, 2005; Teti & Candelaria, 2012). The effects that stress has on these family processes has detrimental effects on the children’s brain development (Luby et al., 2013; Dufford, Kim, & Evans, 2020 ), behavior and mental health (Lupien, McEwen, Gunnar & Heim, 2009), and academic achievement (Tamis-LeMonda et al., 2017; Gilkerson & Richards, 2018; Hair et al., 2015; Romeo et al., 2018), thus perpetuating inequalities.

In parallel, the parent investment framework posits that parents’ inability to invest in stimulating materials, schools, healthcare, housing, and time in engaging with their children hampers children’s development, particularly cognitive development and academic achievement (Francesconi & Heckman, 2016). Indeed, families’ material hardship, or the inability to meet basic needs such as housing, medical care, bills, and utilities, has been linked to negative child outcomes (Neckerman et al., 2016; Wimer, Nam, Waldfogel, & Fox, 2016). Poverty is also associated with reduced investments in developmental inputs such as toys, books, and high-quality early care and education (Yeung, Linver & Brooks-Gunn, 2002). Particularly, parents’ limited access to toys and books has been linked to reduced capacity to engage in cognitively stimulating activities, which in turn affects children’s cognitive development and achievement (Zuckerman, 2009). It is important to note that parents’ capacity to invest in their children is in embedded in larger social inequalities and discrimination. Families facing poverty, and particularly Black and minority families, live in poorer neighborhoods, with higher rates on community violence, attend lower-quality schools, and receive health care at lower-quality hospitals which also can impact children’s health and development (Evans & English, 2002; Evans et al., 2005).

Despite the fact that these two models are sometimes presented separately, a growing body of research has shown the interconnection between their different components. For example, Yeung, Linver, and Brooks-Gunn (2002) showed that while parental investments in cognitively stimulating materials and activities mediated the effect of income on the child cognitive outcomes, and maternal emotional stress mediated the effect of income on child behavioral problems, parental investments and household environment also influenced maternal well-being, demonstrating the interconnection of both models. These findings, together with more recent evidence (Gard et al., 2017) suggest that programs that integrate strategies to reduce financial strain and support healthy family processes (e.g., enhancing self-confidence and capacity to cope with stress, improving parents’ mood, and enhancing supportive parenting practices), may be the most successful at improving child development across multiple domain of functioning.

1.2. Programs focusing on family processes.

Based on the family process model, most early childhood programs have focused on supporting the quality of the home environment in which low-income children grow up as a means of buffering the effects of poverty on young children (Evans & Kim, 2013). Some interventions follow a targeted approach by focusing on improving different aspects of parenting quality - sensitivity, warmth, and cognitive stimulation (i.e. reading, talking, and singing to the child) which promote secure parent-child relationships and provide the foundation for the child to develop cognitively, socially, and emotionally (Chang et al., 2009; Dexter & Stacks, 2014). What these interventions have in common is that they target specific parenting behaviors and practices using role model, practice, and video support, and are intense and short in duration (e.g. Playing and Learning Strategies (PALS), Parent–Child Interaction Therapy (PCIT), The Incredible Years Program, Triple P). Such interventions have shown small to moderate effects increasing parental sensitivity, warmth, and language stimulation, and child socio-emotional skills (Landry, Smith, Swank, & Guttentag, 2008), reducing child disruptive behaviors (Lieneman et al., 2017), and coercive discipline, and increasing self-efficacy (Gridley, Hutchings, & Baker-Henningham, 2015; Perrin et al., 2014; Gross et al., 2003). Such findings provide evidence about the importance of enhancing supportive parenting practices directly.

At the other end, programs with a broad approach support multiple components of parent and child-wellbeing in a more holistic way, an example being home visiting programs, which are the most common services available to families of low socioeconomic status. Home visiting programs such Nurse Family Partnership, Healthy Families America, or Parents as Teachers have the commons goals of working with parents to (1) identify family strengths, needs, and interests; (2) inform and support parents in areas of identified needs and interests; and (3) involve families in services available in the community by providing referrals to community service providers. This last component of many home visiting programs is also known as case management. Low income families tend to experience a range of barriers to access and use of social services and health care, leaving some of their needs unmeet. For instance, low-income families tend to underutilize childcare subsides due to lack of knowledge, hassles, and restrictions to access the subsidy system (Shlay, Weinraub, Harmon, & Tran 2004). Similar barriers have been identified to accessing early intervention services. In addition, many low-income parents may experience limited opportunities to access adequate jobs (Brooks & Buckner, 1996), stable housing (Brisson & Covert, 2015; Park, Fertig, & Metraux, 2014), or high quality health care (Anderson et al., 2006). Therefore, embedding case management into early childhood programs aims to help low income parents navigate an often fragmented and complex system to get access, and use services that already exist in their communities.

Some visiting programs, such Nurse Family Partnership, a program led by nurses, have showed positive impacts across multiple child and adult domains, including prenatal health improvement, increased partner relationship stability, reduced use of welfare, reduced childhood injuries, increased inter-birth intervals, increased maternal employment and earnings, and improved child language, cognitive, and academic functioning (Olds, 2010). However, other evidence also shows the limitations of visiting programs. A recent evaluation of the Mother and Infant Home Visiting Program Evaluation (MIHOPE), a mandated RCT of the abovementioned home visiting programs, showed that while home visiting had positive effects on some outcomes related to parenting skills, child maltreatment, and child development, no significant effects were found on more distal outcomes such maternal and infant mental health, family economic self-sufficiency, and child health, underscoring some potential limitations of such programs (Lee et al., 2019; Michalopoulous et al, 2019). Meta-analyses of home visiting programs find the same inconsistencies; while evaluations of a model program in one region may be very positive, the same program model may struggle to sustain these core outcomes across multiple regions (Howard & Brooks-Gunn, 2009). The impact of case management has also been questioned. While there is some evidence about the benefits of case management programs for infant health care and screening, for children with special care needs, and maternal prenatal care (Rosenbaum et al., 2009), The National Impact Evaluation of the Comprehensive Child Development Program (CCDP) provided only limited evidence of positive effects. Combined case management and home visiting to connect low-income children and their parents with health, education, and social services demonstrated no impacts (Goodson et al., 2000). While case management may benefit families within particular contexts of high need and vulnerability, its results may be insufficient if services offered are inadequate or if families do not perceive a need for offered services. Thus, case management must be tailored to the specific needs and motivations of parents.

Beyond home visiting programs, there are additional parenting supports available for low-income families, many of which are provided by larger community-based organizations. For example, ‘baby college’ is a parenting support initiative within a larger multi-sector partnership model serving families in Harlem (Harlem children’s zone; HCZ). However, there is a gap in rigorous evaluation of such community based supplementary programs, and some studies suggest that outcomes from broader community initiatives are comparatively weak relative to its core services (Dobbie & Fryer, 2011). Furthermore, community based initiatives such as HCZ rely on a network of partners to provide services, leading to difficulties with both scaling and replicating success (Bower & Rossi, 2019; Pendall & Hendey, 2016).

1.3. A need for financial and material support

To be most effective when working with low-income parents, programs may need to address not only parenting knowledge, quality and confidence, but also the daily material hardships parents experiencing poverty struggle with alongside raising their newborns. Based on the parental investment framework, higher income and increased parent’s capacity to invest in children can lead to meaningful improvements in children’s outcomes. Social policies in the U.S. such the Earned Income Tax Credit (EITC) and Child Tax Credit (CTC) are designed to boost children and families out of poverty (Marr et al. 2015) and have been found to improve maternal health (Evans and Garthwaite, 2014), infant health (Hoynes, Miller, and Simon, 2015), and academic achievement for young children in low-income households (Dahl and Lochner, 2012). However, these policies are targeted to parents with a requisite amount of earnings, and so they fail to support children whose parents are unable to find regular work (Moffitt 2015). In-kind federal programs that subsidize certain household expenses, such food, health expenses, housing, or childcare, also contribute to reducing poverty. Yet, many low-income families, particularly those in single-parent households experiencing job instability, may experience substantial unmet economic needs and face the lack of income necessary to buy essential goods, including baby items such clothes, diapers, and learning and enrichment materials such books and toys (Kaushal, Gao, & Waldfogel, 2007). All of this is aside from the burden and stress it can cause parents to be unable to cover essential good for the baby. Moreover, scarcity of books or toys can limit verbal interactions, placing the child at higher risk for adverse developmental outcomes (Gazso, McDaniel, & Waldron, 2016; Rowe, 2018). Provision of child-oriented resources may represent an avenue to promote supportive parent-child interactions within family-focused interventions (Zuckerman & Khandekar, 2010). In addition, despite qualitative studies suggest that low-income families mobilize resources within their families and communities to find baby goods, providing families with such items could contribute to reduced financial stress (Massengale, Erausquin, & Old, 2017).

Direct material support is still absent from most family support programs potentially diminishing parents’ capacity to act on the knowledge, skills and confidence that they already have and acquire from well-designed interventions. Room to Grow aims to address this gap in the early childhood programing field, informed by the family process model and acknowledging the importance of parents’ investment in providing stimulating environments and interactions.

2. The Room to Grow Program

Room to Grow (RtG) is a three-year strengths-based parenting program designed to ensure that parents will have the resources they need by offering three key supports over the program’s three years: (1) structured parent coaching, (2) material goods, such as books, toys, clothing, and baby equipment, and (3) community referrals. The central concept of the RtG program is that parent coaching and referrals will be more effective if families are also receiving substantial material goods and similarly that the material goods will be more beneficial if they are coupled with parenting education and community connections.

The program, founded by clinical social worker Julie Burns, focused on establishing positive relationships between parents and trained social workers, emphasizing trust and authentic relationships (Howe et al., 2018). Such relationships are thought to improve retention in the program across three years, ensure that parenting advice is well-received and implemented, and that community referrals are accessed by parents. Finally, the Room to Grow approach is grounded in a two-generation model, which emphasizes meeting the needs of both parents and children jointly.

Families are referred to Room to Grow during an expectant parent’s third trimester of pregnancy by community members, including RtG program graduates, and a partner network of hospitals, prenatal clinics, and community-based organizations working with low-income families. In addition, potential program participants can apply directly to the program. Expectant families or their referrer complete a simple online application form, and then enrollment is determined by RtG based on eligibility guidelines and caseload availability. Eligibility criteria are as follows: 1) families are able to begin the program before their baby is born; 2) parents demonstrate financial need, typically based on eligibility for government benefits; and 3) they express an interest in and are committed to participating in the three-year-long program. Demonstration of financial need is initially indicated from the online application form, which asks about participation in a number of means-tested social programs.

When parents enroll, they participate in a joint agreement with the program that confirms their commitment to participating in the program for the full three years and outlines the support they can expect to receive. Appointments occur at RtG’s offices. Families have flexibility to reschedule their visits if needed within a three-week window of the original appointment. If two visits in a row are missed in the first year of the program, or if three visits are missed in total, the family is closed out of the program.

Visits are approximately two hours and occur every three months, starting with a prenatal visit, for a total of 13 visits across the three years of the program. Between in-person visits, families stay in close touch with their RtG social worker through phone calls, texts, and emails, in order to build on progress and address new questions or ongoing concerns.

Room to Grow relies on its own curriculum, which is informed by evidence-based practices for enhancing family functioning, as well as early experiences with program implementation. For example, guidance around age-appropriate feeding practices is based in part on the work of Gartner et al. (2005), and guidance around bedtime routines is based in part on the work of Mindell et al. (2009). RtG has developed a curriculum handbook to guide training of its social workers. Each visit has a standardized component, with the first half of the visit informed by asking questions about topics such as maternal depressive symptoms, child development milestones and home stability, and the second half of the visit focused on the provision of items in the RtG “Baby Boutique” and continued conversation of parenting and child development topics.

As noted above, the first half of each quarterly two-hour visit begins with one-on-one interactions between the parent and their clinical social worker. During this time, the parent is offered parenting education, coaching, support, and guidance. The content is meant to be tailored based on parents’ specific needs and circumstances, and informed by scheduled assessments that screen for depressive symptoms, parenting efficacy, child development, home environment stability, and other topics. Although every visit is tailored and adapted to each family as described above, the structure of each visit covers the following topic areas outlined in the curriculum, all of which are adjusted for relevance and children’s developmental stage: (1) welcome and updates; (2) health and safety (e.g. feeding and routines, sleep safety, and medical and dental health); (3) child development (e.g. language, socio-emotional, and motor skills); (4) parent-child relationship (e.g. parental mental health, attachment, and positive discipline); (5) family system stability (e.g. family planning and home environment); and (6) goal setting. As noted above, contact with families often occurs between sessions by phone or other means. In these cases, clinicians document the date and time as well as any clinical notes in the family’s record in RtG’s database.

Prior to each session, clinicians are encouraged to: (1) review all curriculum content pertinent to the scheduled visit; (2) review the assessment schedule for the scheduled visit; (3) review the material goods likely to be disseminated and confirm availability of items; (4) review assessment data from the prior visit; (5) review any clinical notes and material from prior visits and/or interim phone/email contact; (6) note referrals made in order to determine if referrals were accessed and were helpful; (7) seek any necessary supervision, from peers or program leaders; and (8) seek any external information from community partners as needed. During each session, clinicians monitor conversations for opportunities to provide parents with community referrals. Referrals are provided either in response to a clear request from parents or at the clinician’s discretion based on the conversations that unfold during the visit. The most common referrals are to Early Intervention, mental health services, child care, dental assistance, and legal assistance, though referrals also occur for many other emergent issues stemming from the first half of parents’ visits.

In the second half of each visit, parents move to Room to Grow’s Baby Boutique, where clinicians and the parents jointly select new or nearly new baby items appropriate to the child’s developmental stage. These items include: clothing, books, toys, equipment (e.g., baby monitors, bassinets, playards, strollers, bathtubs), and accessories (e.g., bottles, breastfeeding supplies, cups, dishes, utensils). RtG’s material goods are either new or nearly new. The majority of these material goods are donated by families who choose to recycle the items their own babies have outgrown and by baby companies who manufacture these items. Corporate and community partners also frequently host collection drives for books, toys, and clothing, or purchase new items for the organization. RtG also purchases a very small amount of items directly to supplement high-need items (e.g. Spanish language books). RtG maintains a distribution guide organized by visit to assist in the list of items and quantities recommended for each visit. In the period covered in this paper (i.e., the prenatal through 9-month visit), a typical family will have received approximately 40 books, 100 clothing items, a stroller, and numerous other items such as a baby monitor, breast feeding supplies, sippy cups, etc. Distributed items are logged in RtG’s database system. While in the Baby Boutique, clinicians make links between selected items and developmental milestones discussed during the first half of the session, model the introduction and use of toys, guide and encourage parents’ attempts to engage their child with the toys and materials, and restate the importance of daily parent-child interactions such as play and joint book reading. The average value of the material goods provided is estimated at approximately $10,000 per family over the three years of the program.

Embedded throughout the curriculum is an emphasis on data collection and screening tools that are used to inform the parenting coaching, support, and guidance. The first layer of this data collection involves family history data collected at each visit. Clinicians record information about the family’s housing, home environment, discipline practices, and reading at home, among other information, within the Salesforce database. This allows clinicians to efficiently understand what is happening in families’ lives early on in each visit. At the end of every session, clinicians engage in a goal-setting process with the family. This includes a discussion of long-term goals (those achievable within the three years) and short-term goals (those achievable within the next three months before the next visit). Clinicians and parents identify goals that are specific, measurable, and realistic, and discuss methods by which to achieve them. Progress on these goals is monitored at each quarterly visit.

3. Study and Methods

The current study aimed to examine the impacts of Room to Grow on proximal outcomes after one year of Room to Grow programming, after the families’ 9-month program visit. The research reported in this manuscript has been approved by Columbia University’s Human Research Protection Office, under protocol number AAAR1340. The proximal outcomes examined include parenting (increased parental sense of competence, increased cognitive stimulation, decreased aggravation in parenting, decreased parental stress), maternal mental health (reduced depressive symptoms), prevalence of goods for the babies (increases in the number of books, toys, etc.), and reduced financial distress related to providing goods for the baby. The study is a randomized controlled trial (RCT), with families randomized equally into either a treatment group that receives RtG or a control group that does not. Our hypothesis is that access to RtG will increase positive outcomes on the proximal outcome measures. The study was reviewed and approved by the Columbia University Institutional Review Board.

3.1. Study Recruitment

Late-term pregnant mothers are typically referred to Room to Grow from outside service providers, most of which are hospital staff across New York City and a range of non-profits. In addition, potential program participants can self-refer into the program through Room to Grow’s online application portal. To generate our study sample, we proceeded as follows. First, Room to Grow staff prescreened referrals to ensure that they were eligible for the program, primarily by confirming through intake forms that prospective parents demonstrate low-income status. Eligible referrals were then passed to Columbia staff for scheduling and randomization. Research staff called prospective Room to Grow parents to explain the study and the randomization process. If the mother agreed to participate in the study, she was randomized into either the treatment or control group prior to the home visit. If a mother did not wish to participate in the study, the referral was handed back to Room to Grow, where she was entered into a lottery with a 50/50 chance of being admitted into the program—thus ensuring that study participation did not impact a prospective parent’s chance of receiving Room to Grow services. Because the program receives many more eligible referrals than available slots in the program, randomized assignment to treatment or control group is ethically sound. Roughly 95 percent of mothers who received the call from Columbia staff agreed to participate in the study. The research team maintained separate randomization lists for referred parents by clinician language (as sometimes program slots were only available for English-only parents if Spanish speaking clinicians’ caseloads were full).

3.2. Baseline Study Procedures

Columbia research staff conducted the baseline visit in study participants’ homes, though participants were offered the option of meeting at the Columbia offices or another quiet location if the mother preferred. These visits occurred prior to the baby’s birth in all cases. The vast majority, 86%, of baseline visits were conducted in-home. Treatment and control group study participants were given an incentive of $50 cash for their participation, and control group participants also received a gift basket filled with approximately $75-$100 worth of developmentally-appropriate gifts for their impending child. Approximately four business days after the home visit, Columbia staff informed control group participants that they were not selected for the Room to Grow Program; participants in the treatment group received follow-up from Room to Grow. Experienced interviewers conducted the interviews in either English or Spanish. Interviews in Spanish were conducted by bilingual native Spanish speakers. Consent for follow-up was also obtained during the baseline visit. Baseline surveys collected information on family demographics and economic characteristics, as well as baseline measures of future proximal outcome measures when appropriate. Surveys did not collect information that would provide evidence of risk of harm to the mother or baby, however, in the case that such information became apparent, investigators had approved protocols for reporting such risks within the study’s human subjects protection protocol. A general resource guide was also available for research staff to provide to subjects in cases of reported distress or parenting/financial needs (www.findhelp.org). In total, 320 baseline surveys were completed. However, one mother’s survey was struck from the sample because of an error in randomization, and two additional mothers were removed from the sample when study staff learned that their babies had passed away. In total, the analytic sample is therefore 317 mothers, which includes 158 in the treatment group and 159 in the control group. We recognize that the modest sample size in the RCT limits statistical power, but we also note that with the existent sample size we are able to detect statistically significant effects larger than 0.2 in effect size, and marginally significant effects larger than 0.15 in effect size.

3.3. Year One Follow Up Procedures

Year one follow-up efforts began approximately 10.5 months after the babies’ projected due dates. We label this year one both because the babies in the study were approaching age one and because the mothers in the study had participated in Room to Grow for approximately one year (prenatal visit, 3-month, 6-month, and 9-month visits). Information from baseline surveys was utilized to determine when a child would likely turn age 10.5 months, and after this date research staff began calling to schedule follow-up surveys. Research staff initiated contact with treatment and control group mothers and invited them to participate in a 40-minute phone survey with a $35 incentive. Retention in the year one survey was high, with 282 of the 317 original mothers completing a survey (89%). This included 140 control group mothers (88%) and 142 treatment group mothers (90%). So not only was retention high, but retention was also balanced between the treatment and control groups.

The year one survey contained ten primary outcome measures across four domains, all of which are considered proximal outcomes given the early nature of the study relative to completion of the full program and because these outcomes are those that are hypothesized to lead to future specific parent and child development outcomes following full program participation. The four domains are: goods in the home; parenting; mental health; and financial distress. It is important to note that goods in the home such as toys and books might be thought of as program outputs rather than outcomes, but we include them as outcomes since these are primary targets of the intervention. In some cases we shortened scales by removing specific items, as described below. This was done in order to keep the survey instrument manageable in length given the limited time we could expect to keep mothers in the study on the phone. Items deleted were based on loadings from prior studies, and/or if the language in the items was deemed too complex to administer within phone surveys. The specific measures that we used are as follows. Except where otherwise noted, a higher score indicates a more favorable outcome.

Goods in the home

(1). Toys and Games in the Home:

Study participants are asked if they have certain types of developmentally appropriate toys and games in the categories of first infant toys (soft toys or soft cloth, mirror, rattle or squeeze toy, plastic or wooden toys that fit on a ring), activity/manipulative toys (blocks/stackable toys, musical toys, busy box/activity center, shape sorter), and imagination toys (bath toys: boat, duck, car or something to push, toy telephone or radio with dials, stuffed animal, doll). Respondents can answer yes/no for each (1=yes, 0=no), yielding a total score ranging from 0 to 13.

(2). Books in the Home:

Study participants are asked how many children’s books they have in the home.

Parenting

(3). Parental Sense of Competence:

The Parenting Sense of Competence Scale (PSOC; Gibaud-Wallston & Wandersman, 1978) measures parental competence on two dimensions: Satisfaction and Efficacy. It is originally a 17-item Likert-scale questionnaire (on a 6-point scale ranging from strongly agree [1] to strongly disagree [6]), with nine questions under Satisfaction and seven under Efficacy. Satisfaction section examines the parents’ anxiety, motivation and frustration, while the Efficacy section looks at the parents’ competence, capability levels, and problem-solving abilities in their parental role. We use five items from the scale, and condense response options to a four-point scale ranging from strongly agree (1) to strongly disagree (4). The five items are: (a) The problems of taking care of a child are easy to solve once you know how your actions affect your child, an understanding I have acquired; (b) Being a parent is manageable, and any problems are easily solved. (c) If anyone can find the answer to what is troubling my child, I am the one. (d) I honestly believe I have all the skills necessary to be a good mother to my child. (e) Being a good mother is a reward in itself. The Cronbach’s alpha for these five items is 0.53.

(4). Chaos in Home:

The Confusion, Hubbub, and Order Scale (CHAOS; Matheny, Washs, Ludwig, & Philips, 1995)) is a questionnaire filled out by parents that is designed to assess the level of confusion and disorganization in the child’s home environment. The statements are scored using a 4-point scoring system. The questionnaire originally consists of 15 statements, to each of which a parent or caregiver assigns a number between 1 and 4 that correspond to the following: 1 = Very much like your own home; 2 = Somewhat like your own home; 3 = A little bit like your own home; 4 = Not at all like your own home. We use two items (6 and 10) from the CHAOS scale. The two items we use are: (a) You can’t hear yourself think in our home; and (b) It’s a real zoo in our home. The Cronbach’s alpha for these two items is 0.83. A higher score on this measure indicates a less favorable outcome.

(5). Cognitively Stimulating Activities:

The STIMQ (Mendelsohn et al., 1999) is a measure of cognitive stimulation in the home consisting of 4 subscales: Availability of Learning Materials, Reading, Parental Involvement in Developmental Advance, and Parental Verbal Responsivity. We use items from the Parental Verbal Responsivity subscale. A higher score reflects greater stimulation in home (favorable). These ask about how often the parent does the following in a typical week with the child: (a) Read books (b) tell stories; (c) sing songs; and (d) play peek-a-boo. These items have also been asked in the national Fragile Families and Child Wellbeing Study (https://fragilefamilies.princeton.edu/) Answer categories range from not at all to every day on a four-point scale and items are averaged. The Cronbach’s alpha for these items is 0.60.

(6). Aggravation in Parenting:

Aggravation in Parenting is a measurement of stress experienced by parents associated with caring for children. This scale was created using items from the Parenting Stress Index and the Child-Rearing Scale. The Aggravation in Parenting Scale is a 9-item scale, taken from the JOBS (Job Opportunities and Basic Skills Training Program) Child Outcomes Study (Moore et al., 1995), and is also found in the Child Development Supplement of the Panel Study of Income Dynamics. The scale measures the amount of parenting stress brought on by changes in employment, income or other factors in the parent’s life. We use a short-form of the scale consisting of 6 out of the 9 items in the scale, following Yu & Singh (2012): (a) Being a parent is harder than I thought it would be; (b) I feel trapped by my responsibilities as a parent; (c) I find that taking care of my child(ren) is much more work than pleasure (d) I often feel tired, worn out, or exhausted from raising a family; (e) [child] seems harder to care for than most children; and (f) I often feel angry with [child]. Response categories range from strongly agree to strongly disagree on a four-point scale. The Cronbach’s alpha for these six items is 0.64. A higher score indicates a less favorable outcome.

Mental health

(7). Depressive Symptoms:

The Center for Epidemiological Studies-Depression (CES-D), is originally a 20-item measure that asks caregivers to rate how often over the past week they experienced symptoms associated with depression, such as restless sleep, poor appetite, and feeling lonely. We use a shorter version (7-item) that has been used in several previous studies and shows good reliability (Levine, 2013). The seven items are: (a) I did not feel like eating; my appetite was poor; (b) I had trouble keeping my mind on what I was doing; (c) I felt depressed; (d) I felt that everything I did was an effort; (e) My sleep was restless; (f) I felt sad; and (g) I could not “get going.” Answer categories are: (1) Rarely (less than one day), (2) Sometimes (1–2 days), (3) Occasionally (3–4 days), (4) Most or all of the time (5–7 days). The Cronbach’s alpha for these seven items is 0.84. A higher score indicates a less favorable outcome.

(8). Perceived Stress:

The Perceived Stress Scale (PSS; Cohen, Kamarck, & Mermelstein, 1983) is a widely used instrument for measuring the perception of general stress. This 14-item scale measures how stressful or uncontrollable participants find their lives. Respondents rate the frequency of their feelings and thoughts related to events and situations that occurred in the last month. Following FFCWS, we use 4 items from this scale: (a) In the past month, how often have you felt unable to control the important things in your life? (b) In the past month, how often have you felt confident about your ability to handle personal problems? (c) In the past month, how often have you felt that things were going your way? (d) In the past month, how often have you felt that difficulties were piling up so high that you could not overcome them? Answer categories are: (1) Never, (2) Almost never, (3) Sometimes, (4) Fairly often, (5) Very often. The Cronbach’s alpha for these four items is 0.73. A higher score indicates a less favorable outcome.

Financial worry

(9). Worry about Baby Expenses:

Study participants are asked to rate (on a scale of 1–10) how worried they are about having enough money to cover their baby’s expenses with 1 being “Not at all worried” and 10 being “Very worried”. A higher score indicates a less favorable outcome.

3.4. Analytic Approach

Because this is an RCT, we begin by providing baseline information about the treatment and control groups and the balance between them, using t-tests to test for significant differences. Two-tailed significance tests (t-tests for continuous measures, and chi-square tests for categorical measures) are used to test for differences between the treatment and control groups on baseline demographics and levels of disadvantage. We then turn to intent-to-treat estimates of the impact of Room to Grow on those assigned to it one year post-assignment. We estimate these both unadjusted (i.e. the simple difference in mean outcomes between the treatment and control groups) and adjusted for baseline characteristics, applying one-tailed tests of significance because our hypothesis is that the effect of RtG will be positive. The baseline characteristics are all of the indicators in Table 1, which we describe below. We focus on results that are significant at p<.05 or less, but also discuss those that approach significance at p<.10 or less. To calculate effect sizes, we calculate a Cohen’s d statistic, whereby differences in the outcome means between the two groups are divided by the pooled baseline standard deviations in each outcome. We follow convention and consider statistics around .20, .50, and .80 to represent small, medium, and large effects, respectively (Ferguson, 2016).

Table 1.

Sample Descriptives and Treatment/Control Group Balance

Baseline Demographics	Total Sample (n=317)	Control (n=159)	Treatment (n=158)	p
Race				0.358
White non-Hispanic	4%	3%	6%
Black non-Hispanic	36%	39%	32%
Hispanic	53%	52%	54%
Other non-Hispanic	7%	6%	8%
Age				0.698
25 years or under	44%	43%	45%
over 25 years	56%	57%	55%
Foreign born	42%	43%	42%	0.858
Education				0.833
Less than HS	29%	31%	28%
HS grad/GED	31%	30%	32%
Some College or More	40%	40%	41%
Live with parents	23%	19%	27%	0.102
Live with spouse/partner	31%	31%	30%	0.838
First child	45%	43%	47%	0.402
Health status fair or poor	16%	19%	13%	0.178
Baseline Disadvantage	Total Sample (n=317)	Control (n=159)	Treatment (n=158)	p
Poverty	70%	69%	71%	0.741
Any material hardship	72%	72%	72%	0.972
Live in shelter	25%	28%	22%	0.205

4. Results

We begin by presenting baseline descriptive statistics on the analytic sample, both overall and by treatment and control group (Table 1). Overall, we see that the sample of mothers is highly disadvantaged and/or comes from demographic groups that traditionally face higher rates of disadvantage. Nearly 90 percent of the sample is Black or Hispanic, and nearly half of the mothers are under the age of 25. Two in five are foreign born, and three in five have only a high school degree or less. Nearly a quarter were living with their parents in their third trimester, and a bit less than one third were living with a spouse or partner. For just under half the sample, their unborn child would be their first child. Sixteen percent of the sample reported being in fair or poor health at baseline.

Roughly 70 percent of the sample was living in poverty at baseline and most of the rest had incomes between 100 and 200 percent of the official poverty line (not shown), which is usually considered low-income. Material hardships were substantial across the sample, with fully 72% of the sample reporting a hardship across the five domains that we asked about at baseline. Lastly, one in four mothers reported living in a shelter at the time of their baseline survey.

With regards to balance between the treatment and control groups, we see that the two groups are very well balanced, and there are no statistically significant differences between them (Table 1). Nevertheless, as we move to estimates of early program impacts, we control for these demographics to assess whether any non-significant differences may be affecting estimates of treatment effects.

Table 2 presents our core estimates of early treatment effects of the Room to Grow program. These are measured after the babies in the study were 10.5 months old, or after four quarterly visits in the program (prenatal, 3-month, 6-month, and 9-month visits). We begin by describing effects on goods in the home. For both toys and goods and also number of books in the home, the early effects are both moderate to large and statistically significant. Treatment group mothers report on average one more toys or goods than control group mothers and nearly 14 more books in the home as their children near one years old. Adjusting for baseline demographics does little to change the effect sizes (0.39 for toys and goods; 0.51 for books in the home) or statistical significance levels (p < .001 for both).

Table 2.

Intent-to-treat results, unadjusted and adjusted for demographics.

	Treatment	Control	Unadjusted Effect Size	Unadjusted P-Value*	Adjusted Effect Size	Adjusted P-Value
Goods in the Home
Toys and Goods (0–13; SD: 2.73)	9.4	8.2	0.44***	0.000	0.39***	0.000
Books in Home (0–140; SD: 26.9)	34.1	20.2	0.52***	0.000	0.51***	0.000
Parenting
Parental Sense of Competence (6–15; SD: 1.78)	13.2	12.6	0.34**	0.006	0.32**	0.004
Chaos in Home (0–3; SD: 0.81)	0.51	0.65	0.17†	0.080	0.18†	0.063
Cognitively Stimulating Activities (1–4; SD: 0.66)	3.15	3.01	0.21*	0.039	0.18†	0.068
Aggravation in Parenting (0–2.5; SD: 0.52)	0.72	0.85	0.25*	0.021	0.28**	0.010
Mental Health
Depressive Symptoms (0–20; SD: 4.64)	4.17	4.74	0.12	0.153	0.12	0.159
Perceived Stress (0–16; SD: 3.27)	4.72	5.3	0.18	0.068	0.17†	0.078
Financial Worries
Worry Enough Money for Baby (1–10; SD: 3.2)	5.18	5.33	0.05	0.350	0.04	0.379

^*p-value determined through a one-tailed t-test; adjusted effect sizes are based on models controlling for all covariates listed in Table 1.

^***p < .001

^**p < .01

^*p < .05

^†p < .10.

Moving to our proximal outcomes related to parenting, we again see promising results. The strongest significant effect that we observe is on parental sense of competence, with an effect size of over 0.30 (p < .01), regardless of demographic controls. For chaos in the home and engagement in cognitively stimulating activities, effect sizes are closer to 0.20, and approach statistical significance (p < .07). Effects on aggravation in parenting are statistically significant after adjusting for demographic controls (effect size = 0.28; p = .01). Noticeably, both effects trend in favor of the treatment group, consistent with the estimated effects on other outcomes.

Next we turn to indicators of mothers’ mental health and financial worries. Here we see no significant effects on maternal depressive symptoms (though treatment group mothers evince lower absolute scores on the scale). Treatment effects on perceived stress approach statistical significance and favor treatment group mothers (effect size = 0.17; p < .08). For financial worries about being able to provide for one’s child, estimated effect sizes are small and statistically insignificant.

We note that our estimates are “intent to treat” (ITT), as they include in the treatment group all the families randomly assigned to Room, regardless of their actual participation. While take up of the program was high, 12.7% of mothers who were randomly offered a place did not participate. Adjusting our estimated effects for this non-participation would suggest that “treatment-on-the-treated” effects for those participating in the program would be roughly 14.5% larger than the ITT estimates reported here.

5. Discussion

The randomized controlled trial described here shows the early effects of a novel program for families with low incomes that combines parenting supports and referrals with direct material supports totaling up to $10,000 across children’s first three years of life. The Room to Grow program seeks to provide parents with education, coaching, and knowledge to help them meet their children’s needs and their goals for their children’s development, but couples this with community referrals and direct financial support in the form of in-kind goods that should allow them to focus less on day-to-day needs and more on their children’s long-term futures.

This article documents the initial findings from the RCT. We first demonstrate the high levels of disadvantage faced by the mothers in our sample, and further demonstrate that treatment and control group mothers are well balanced on observable demographic and financial baseline characteristics. Early estimates of program impacts on proximal outcomes at the end of one year of treatment are promising, even after controlling for baseline characteristics. There are clear, medium-to-large, and positive impacts on toys, goods, and books in the home, as we would expect with a program that directly provides them to families (indeed these might be considered outputs rather than outcomes of the program). We also find significant early effects of the program on parents’ sense of competence, cognitively stimulating activities (marginally statistically significant after adjustments for demographics), and reduced aggravation in parenting. There is suggestive, though not statistically significant, evidence of potential small impacts on reduced chaos in the home and perceived stress. Though early, we see this as preliminary evidence of the Room to Grow program’s promise in achieving its intended proximal outcomes of improving access to cognitively stimulating materials in the home and positive parenting techniques. The question remains whether these effects will be sustained and indeed whether Room to Grow will have larger impacts on parenting after three years of program participation.

Our study does entail some limitations. Our sample size is modest, limiting our ability to detect small and moderate sized program effects. Despite this limitation, we are able to detect effect sizes above 0.20 and we observe a number of such effects both with and without controlling for baseline characteristics. Our results also show a clear pattern in favor of treatment group mothers across all outcomes included in the study. Given the limited sample size, however, we are not able to test for treatment effects by subgroups. For example, we might wish to know whether the program was working better for mothers who entered the program with greater levels of disadvantage or mothers who entered in relatively better off circumstances. Our limited sample size precludes such analyses. While our sample size is relatively modest, we do note that attrition between baseline and age one follow up is minimal, and only negligibly different between our treatment and control groups. Fully 89 percent of those who completed a baseline survey also completed an age one follow-up survey, and this entailed 90 percent of treatment group mothers and 88 percent of control group mothers. This gives us more confidence in our early estimates of treatment group impacts that we describe here.

We are currently reliant on phone survey data to document proximal outcomes and early treatment effects. In addition, given the small-scale nature of the RCT and the small research staff involved, we were unable to keep research staff blinded to randomization status. Gold standard measures of parenting and child development necessitate in-person assessments, coded by blinded and independent research staff. We will be collecting such measures after children graduate from the Room to Grow at age 3. Phone surveys also create limits on the amount of time available to speak with mothers of newborns, who are already time-challenged parenting their young children. This necessitated shortening some scales that would ideally contain numerous other subitems. However, we have made these decisions based on their use in prior studies, and believe their reliability is sufficient for analyses.

The results described here thus entail proximal outcomes hypothesized to be consequential in moving the needle on family and child outcomes known to be important in children’s later lives. As such, the results reported here are encouraging but preliminary. As we collect additional data from the mothers and children in the Room to Grow RCT, we will learn a great deal more about the potential promise of this novel program that combines parenting and financial support for mothers and children in need. As with any RCT, understanding the reasons for the patterns of results we see is an important topic. While a full implementation evaluation is beyond the scope of the current paper, understanding implementation issues and how parents experience the program and its services are important topics for future research and are part of the long-term evaluation of the program.

The findings reported here add to the literature on programs aimed at parents and children in children’s formative early years. This literature has focused on some of the most prevalent programmatic strategies, such as home visiting, center-based group parenting programs, and income support programs aimed at alleviating financial hardship (Love, Chazan-Cohen, Raikes & Brooks-Gunn, 2013). While specific interventions in each of these areas have shown promise, the overall evidence base shows that programs aimed solely on providing parents with skills and education suffer from high attrition and inconsistency in achieving positive results when implemented at scale (Beasley et al., 2018). The novelty of the Room to Grow program examined here is its combination of direct material support for mothers and their babies alongside structured parenting programming, and community referrals. As we have noted throughout, the results presented here are early on in the course of the program, approximately 1/3^rd of the way through a three-year program. It is too early to determine whether Room to Grow makes a meaningful and lasting contribution to enhancing important outcomes for children and their families. As such, it would be inappropriate to speculate on the policy implications of our findings, though if they held up and persisted it would indicate that more attention may need to be made to helping parents financially in addition to traditional knowledge and education-based strategies. It would also lead to a need to understand whether the novel approach could be scaled to other cities and other populations of poor mothers. Despite the early nature of the results, they are encouraging, warranting further study as families progress through the completion of the program.

Room to Grow is a novel parenting intervention combining parenting support with direct financial support

Early findings from a randomized controlled trial find promising impacts on parenting outcomes

Program led to significant increases in books and toys/goods for the child

Program led to significant increases in parents’ competence and reduced aggravation in parenting