Characteristics of Teen Families Accessing a Supportive Housing Program

There is a need for strength-based research that considers the complex realities of teen families. The purpose of the current study is to provide a springboard for this research by describing the characteristics of teen parents and their children involved in a supportive housing program. We used a community-based participatory research approach with a descriptive design. Teen parent participants completed self-report questionnaires about their relationship with their children, resilience, self-esteem, and parenting attitudes. In addition, the children of teen parents completed standardized assessments of their development. A total of 21 parents (18 mothers and 3 fathers) and 20 children participated. Results indicate that self-esteem and resilience are areas of difficulty for participants. In addition, most parents fell into the medium risk range with respect to the parenting attitudes measured. Of note, few parents fell into the highrisk range on most scales, suggesting that most participants have the foundation for successful parenting across the areas measured. Results also show that the majority of participants are demonstrating typical attachment, discipline practices, involvement, parenting confidence, and relational frustration. An area of particular strength for participants is their involvement with their children. It also appears overall that the children of teen parents are developing on a fairly typical trajectory that is reflective of the general population. This study reinforces the heterogeneity of teen families, with teen parents and their children showing different areas of strengths and challenges across the domains measured. Based on our findings, we discuss implications for research, policy, and practice.


Introduction
The health and wellness of children and families has long been a focus of research and social policy. Evidence that the foundations for lifelong health and development are laid in early childhood (i.e., birth to age 6) has resulted in a specific focus on the early years as a time of significant opportunity and risk (Shonkoff & Phillips, 2000). There has been a corresponding focus on the instrumental role of parents in contributing to their children's development, and an interest in families headed by teen parents (Reisch et al., 2010). There is extensive research to suggest that teen parents face substantial disadvantages that can result in poor social, economic, and health outcomes for themselves and their children (see, e.g., Mollborn & Dennis, 2012;Slomski Long, 2009;Smith et al., 2013). However, an exclusive emphasis on the deficits, risks, and challenges of teen parenting has not resulted in a complete understanding of how teen families function, or the most effective supports that teen families require in order to be successful. Therefore, research that takes a strength-based position and considers the complex realities of teen families is needed. The purpose of the current study is to provide a starting point for this research by describing the characteristics of teen parents and their children involved in a supportive housing program. To begin, the state of the research on teen families is reviewed, followed by the overarching study context and purpose.

Research on Teen Families
The rate of births to mothers aged 15-19 reached a record low of 21 births per 1000 females in the US and 10 births per 1000 females in Canada in 2016 (World Bank, 2018). Although teen birth rates differ based on such factors as geographical location and ethnicity, there has been an overall downward trend in teen pregnancy and birth rates nearly every year since 1991 (Hamilton & Mathews, 2016). Despite this decline, teen pregnancy and parenting remain a focus of research, social policy, and public concern. To illustrate, a recent Google search using the term "teen pregnancy" yielded 322,000,000 results. Similarly, a search of any academic database will produce a mountain of recent evidence regarding the impact of early childbearing on the lives of teen parents and their children (e.g., Ruedinger & Cox, 2012;SmithBattle et al., 2021;Wall-Wieler et al., 2016).
Government officials and policymakers have framed teen pregnancy and childbearing as an epidemic, a social problem, and a source of blame for weakening family structures (Furstenberg, 2007). In line with this discourse, researchers have typically taken a deficit-based approach to understanding teen families, emphasizing the consequences, problems, and risks of teen pregnancy and parenthood (Collins, 2010), and focusing more on preventing teen pregnancy than on effective supports for teen families. Researchers have overwhelmingly demonstrated that teen pregnancy and parenthood are associated with poor social, economic, and health conditions for both parents and their children (SmithBattle, 2018a). As a result, some researchers have concluded that preventing teen pregnancy could result in sweeping, positive changes to society as a whole. For example, it has been asserted that "reducing births to teenagers will improve the well-being of children, adolescents, families, and communities. Fewer teenage births will lower taxpayers' burden and benefit national and state economies" (Barnet et al., 2010, p. 375). Importantly, however, researchers examining the health, economic, and social outcomes of teen parents and their children have often failed to distinguish between the impacts of the economic and social disadvantages associated with teen parenting and the impacts of teen parenting alone (Lawlor & Shaw, 2002;Mayers et al., 2008). There is growing recognition from researchers that the risk for negative outcomes experienced by teen families may be conferred more from poverty and other social determinants of health than parental age (Thompson, 2016). It has also been argued that, although concern over the public and private costs of teenage childbearing has a basis in reality, the representation of these costs has been hyperbolic (Furstenberg, 2007;SmithBattle, 2018b). This is in part because much of the early research on teen families neglected to consider significant background differences between teen and older parents, such as living in disadvantaged neighbourhoods, that could account for differences in health, social, and economic outcomes between teen and older parents (SmithBattle, 2018b). When accounting for these background differences, the effects of parental age decrease substantially (Diaz & Fiel, 2016).
Although there is a dearth of longitudinal research with teen families, in a seminal longitudinal study, Furstenberg (1976) found that, on five-year follow-up, teen mothers fared worse than their non-parenting classmates in terms of educational attainment, employment, financial status, and life satisfaction. However, at seventeen and thirty-year follow-ups, teen mothers' circumstances had improved substantially across the outcomes examined, and disparities were far smaller than would be predicted by prior research (Furstenberg et al., 1987;Furstenberg, 2003). In addition, women who were teen mothers perceived their lives and wellbeing as having improved significantly between early adulthood and middle age, whereas women who delayed childbearing felt less well off in many ways than they did before they formed families (Furstenberg, 2003). Overall, research suggests that teen families may experience a delayed developmental curve, with an overlay of persistent environmental factors, such as poverty and stigma, that are not specific to parents' age. Overlooking these complexities can result in largely unhelpful solutions that do not address the circumstances of teenage families (Furstenberg, 2007).
In addition, the evidence base is weak with respect to key characteristics of teen parents that we know contribute to healthy functioning and child development in families headed by older parents. In particular, there are well-established correlations between healthy child development and parent characteristics such as empathy (Stern et al., 2015), knowledge of child development (Sonnenschein et al., 2014), attitudes toward discipline practices (Wang & Kenny, 2014), and parenting confidence (Winter et al., 2012). However, the literature offers little with respect to how these characteristics are embodied by teen parents. There is also a lack of research that examines the strengths of teen families, including the development of quality teen parent-child relationships, as well as the resilience of teen families (Reisch et al., 2010). This is critical given the potential influence of these areas on successful parenting, and ultimately, healthy child development outcomes. Therefore, an enhanced understanding of these characteristics in teen families is needed, and in part can be achieved through research that takes a strength-based perspective that also considers the complexity of teen parents' lived realities.

Study Context
The current paper draws on data from a larger study that aimed to (1) develop a supportive housing program model for teen families, and (2) investigate the impacts of the program on teen parents and their children (see Tremblay et al., 2018;Tremblay et al., 2020). The project was carried out through a partnership between the Terra Centre, Brentwood Community Development Group, and researchers from the Community-University Partnership for the Study of Children, Youth, and Families (CUP) at the University of Alberta. Terra Centre (Terra) is a non-profit organization that has been supporting teen parents in Edmonton for more than 40 years, with a mission of empowering teen parents to succeed. Brentwood Community Development Group (Brentwood) was formed in 1977 with the aim of building supportive communities by providing affordable housing to individuals and families. Terra and Brentwood partnered in 2014 to offer safe, secure, and affordable housing to teen parents and their children in combination with wraparound supports. The supportive housing program takes place in a 207-unit townhouse site owned and managed by Brentwood in a neighbourhood in Edmonton, Alberta. Shortly after forming their partnership, Executive Directors from the two agencies approached researchers from the University of Alberta to develop and study a supportive housing program model for teen families.
The Successful Families Program was formed with a long-term vision for the children of teen parents to achieve their potential and become valued adults who contribute to society. Within the program, Brentwood acts as the landlord, providing subsidies as well as a house located across the street from participants' homes that has been converted to office and programming space. Terra provides support staff, employing three full-time housing staff with the program in addition to a full-time housing manager. Through group activities, in-home visitations, and community activities, staff provide collaborative, individualized, strength-based services to empower participants to maintain their housing, reduce social isolation, engage with their community, and successfully raise their children in alignment with their goals. Participants are required to have the financial resources and capacity to live independently, and therefore undergo a screening process before being accepted into the program.

Purpose
Within the broader project goal of investigating the impacts of the Successful Families program on teen families, the purpose of the current descriptive study is to describe the characteristics of teen parent program participants and their children. More specifically, three research questions are addressed: (1) What is the relationship quality of teen parents and their children who are accessing a supportive housing program? (2) What are the parenting attitudes, resilience, and self-esteem of teen parents accessing the supportive housing program? and (3) How are the children of teen families involved in a supportive housing program developing across domains? Given the limited extant information in these areas, we aim to contribute to the literature on the wellbeing of teen parents and development of their children in order to inform service delivery and provide a springboard for future research.

Approach and Design
A community-based participatory research approach (CBPR; Israel et al., 2003;Minkler & Wallerstein, 2003) was used in the overarching project. Equitable community-academic partnerships are at the heart of CBPR along with collaboration, co-learning, mutual benefit, and a focus on issues of local importance. In line with a CBPR approach, the researchers and community partners from Terra and Brentwood made decisions together regarding the research questions and methods, and worked collaboratively to recruit participants, collect and analyze data, and mobilize knowledge. More specifically, program staff and researchers reviewed results together and worked collaboratively to consider how to shape programming based on results of the study. The researchers and program staff also worked together to mobilize knowledge about the study by copresenting at conferences and at a meeting of the agency's non-profit board. Under the umbrella of the CBPR approach used in the overarching project, a descriptive design was used for the study presented in this paper.

Participants
All parent participants were invited to complete self-report questionnaires about their relationships with their children, resilience, self-esteem, and parenting attitudes. To recruit participants, the research team attended community events such as a community barbeque and summer events at the park, and held a research information evening, which participants could attend to sign up for the project. In addition, staff recruited participants directly, and facilitated contact and appointment booking between the researchers and participants. Importantly, program staff needed to trust the researchers to interact sensitively and appropriately with participants. Once participants witnessed the development of this trust such that program staff could "vouch" for the researchers, participants similarly began to demonstrate trust in the researchers by agreeing to take part in the study and sharing their information with the researchers.
Of the 40 families in the program, 21 parents (18 mothers and 3 fathers) and 20 children participated. All three fathers were partners of teen mothers who also participated in the study. Table 1 depicts demographic information for participants. It should be noted that, for the purpose of this study, "teen parenting" is defined as parents' age when their child was born and not parents' age at the time of participation in the study.
Two parents did not complete two of the tools (AAPI-2 and CD-RISC) as they did not have time to complete the tools in the researchers' presence, chose to take the tools home to complete, and did not return them. Additionally, participants provided consent for their children to participate in child development assessments.

Data Collection
One of the researchers (MT) collected data from participants in person. Participants had the option to complete the self-report questionnaires and child development assessments at the Terra house (situated directly across from the families' housing and in which the Terra housing staff are based) or at their own homes. All but three participants chose to complete questionnaires and assessments at the Terra house. An event was also held where participants baked cookies with staff members while their children completed assessments. Participants received gift cards for their participation and were provided with feedback reports from the child development assessments. The researcher who explained feedback reports to parents (MT) was a student in a school and clinical child psychology doctoral program and was supervised by two of the study authors who were also Registered Psychologists (RG and JP). Four self-report and two standardized child development tools (depending on the age of the child) were used and are described below.

Parent Self-Report Tools
Rosenberg Self-Esteem Scale The Rosenberg Self-Esteem Scale (Rosenberg, 1989) was used to measure participants' attitudes of rejection or approval toward themselves. The questionnaire has ten items rated on a five-point Likert scale. This tool has strong psychometric properties and is one of the most widely used measures of self-esteem in North America (Sinclair et al., 2010).

Connor-Davidson Resilience Scale (CD-RISC)
The CD-RISC is a 25-item questionnaire and was used to measure parents' resilience, defined as the capacity to effectively cope and adapt in the face of adversity (Connor & Davidson, 2003). Each item is rated on a five-point scale, with higher scores reflecting greater resilience. The CD-RISC has strong psychometric properties (Connor & Davidson, 2003;Davidson & Connor, 2016). We chose to use the CD-RISC because of its validation in the general population as well as clinical samples and diverse cultures.
Behavior Assessment Scale for Children -Third Edition Parenting Relationship Questionnaire (BASC-3 PRQ) The BASC-3 PRQ (Kamphaus & Reynolds, 2015) is designed to capture a parent's perspective on the parent-child relationship for parents of children aged 2-18. Each item is rated on a five-point Likert scale from Strongly Agree to Strongly Disagree. There are seven BASC-3 PRQ scales, described in the results section of this paper. Normative scores are based on the child's age and parent's gender. The BASC-3 PRQ has strong construct, content, and criterion-related validity, internal consistency, and test-retest reliability (Kamphaus & Reynolds, 2015).

Adult Adolescent Parenting Inventory -Second Edition (AAPI-2)
The AAPI-2 (Bavolek & Keene, 2010) is designed to assess the parenting and child rearing attitudes of adult and adolescent parent and pre-parent populations. This 40-item inventory is rated on a five-point Likert scale from Strongly Agree to Strongly Disagree. The AAPI-2 provides a risk index for five specific parenting and child rearing behaviours, described in the results section of this paper. We chose to use the AAPI-2 because of its strong psychometric properties (Bavolek & Keene, 2016) and design for use with parents as young as age 13. Normative data for the AAPI-2 were established with adult and teen parents from 53 different agencies in 23 different US states.

Bayley Scales of Infant and Toddler Development -Third Edition (Bayley-III)
The Bayley-III (Bayley, 2006) is an individually administered assessment of developmental functioning for children between 1 and 42 months of age. The Bayley-III is used to identify developmental delays, assist in intervention planning, and elevate understanding of a child's strengths and challenges in five developmental domains that comprise separate scales (Piñon, 2010).
The Bayley-III Cognitive Scale (91 items) assesses sensorimotor development, exploration and manipulation of objects, object relations, concept formation, and memory. The Language Scale assesses receptive communication (49 items) and expressive communication (48 items) separately, and the Motor Scale assesses fine motor skills (66 items) and gross motor skills (72 items) separately. Each of the Bayley-III Cognitive, Language, and Motor Scales are administered by an examiner who interacts directly with the child, whereas information for the Social-Emotional (35 items) and Adaptive Behavior Scales (41 items) are gathered through a questionnaire completed by the child's parent or primary caregiver.
Normative data for the Bayley-III are representative of the US population with respect to children's race/ethnicity, sex, parent education level, and geographic region (Bayley, 2006). The Bayley-III has strong internal consistency, inter-rater and test-retest reliability, as well as construct, content, and criterion-related validity (Bayley, 2006). We chose to use the Bayley-III because of its strong psychometric properties, coverage of multiple developmental domains, and engaging, play-based format.

NEPSY-II
The NEPSY-II (Korkman et al., 2007) is an individually administered assessment of neurocognitive functioning for children between 4 and 18 years of age. Eight NEPSY-II subtests across four domains were used in the current study. Tasks in the language domain measure how well a child understands and uses words and sentences to communicate with others. In the memory domain, tasks measure how a child takes in, stores, and remembers information. The sensorimotor domain reflects how well a child can control hand movements, and the visuospatial domain reflects how well a child sees and arranges visual information.
This tool has strong internal consistency, inter-rater and test-retest reliability, as well as construct, content, and criterion-related validity (Korkman et al., 2007). The NEPSY-II normative data are representative of the US population with respect to children's race/ethnicity, geographic location, and parent education (Brooks et al., 2009). The NEPSY-II was used with children who were older than 42 months of age, and who were too old to be assessed with the Bayley-III.

Data Analysis
All quantitative data were entered into a Microsoft Excel database. Raw data for the Bayley-III, NEPSY-II, AAPI-2, and BASC-3 PRQ were converted to standard scores using the published norms available for each tool, and descriptive statistics were calculated using the standard scores for these tools. For each of these tools, the number of participants who fell into descriptive categories established by the tool developers (e.g., average, above average, below average) is presented in the results section that follows. For the Rosenberg Self-Esteem Scale and CD-RISC, descriptive statistics were calculated using raw scores because norms and standard score conversions have not been developed. For these tools, descriptive categories (e.g., average, below average, above average) have also not been developed, and results are therefore presented as average raw scores.

Parent Characteristics
Successful Families participants' self-esteem was measured with the Rosenberg Self-Esteem Scale. On this self-report tool, participants' average score was 20.8 out of a possible 40 points (n =19), with scores ranging between 15 and 28. Descriptive categories have not been established for this tool (e.g., what constitutes low versus high self-esteem). However, as a reference point, in a study of 18-19-year-old female Canadian high school students the average score on the Rosenberg Self-Esteem Scale was 29.04, and a score below 21 was deemed "very low self-esteem" (Bagley et al., 1997). Eight out of 19 Successful Families participants' scores were below 21.
Participants' self-reported resilience was measured with the CD-RISC. As with the Rosenberg Self-Esteem Scale, descriptive categories have not been established for this tool. Participants' average score on the CD-RISC was 62.23 (n = 19), with scores ranging between 30 and 90. As a point of comparison, the mean score for the general US population was 80.7 in the original validation study for the CD-RISC (Connor & Davidson, 2003). Results therefore indicate that teen parent participants fell well below the average with respect to their self-reported resilience scores.
The AAPI-2 was used to assess parenting and child rearing attitudes. Nineteen participants completed the AAPI-2. Results are organized into five scales (Oppressing Power and Independence, Role Reversal, Corporal Punishment, Lack of Empathic Awareness, Inappropriate Parental Expectations) that serve as the basis for assessing attitudes known to contribute to child abuse and neglect, as well as levels of risk (low, medium, high) for abusive and neglectful parenting practices. Figure 1.0 shows the number of participants who scored in each classification range for the five AAPI-2 scales.
Most participants scored in the medium risk range across all five AAPI-2 scales. Thus, most participants appear to be at medium risk for inappropriate parental expectations, a lack of empathic awareness (i.e., experiencing their children's needs and desires as irritating and overwhelming and as coming into conflict with a parent's own needs), belief in corporal punishment, role reversal (i.e., the tendency to reverse parent and child roles such that children are expected to be sensitive to and responsible for their parents' wellbeing and parents look to their children for care and comfort), and oppressing power and independence (reflecting the attitude that obedience and complete compliance to parental authority should be demanded).

Parent-child Relationships
The BASC-3 PRQ was used to assess participants' perspectives on their relationships with their children. BASC-3 PRQ scores are classified into three ranges, consisting of average (reflecting a typical parent-child relationship), below average (reflecting the presence of potential or developing relationship problems that should be monitored), and lower extreme (reflecting significant relationship problems for which intervention may be warranted).
Twelve program participants completed the BASC-3 PRQ. A lower number of participants completed the BASC-3 PRQ compared to other self-report tools because the BASC-3 PRQ can only be administered to parents with children ages 2 and older. Figure 1.1 depicts the number of participants who scored in each classification range on the five BASC-3 PRQ scales. Results show that the majority of participants who completed the BASC-3 PRQ are demonstrating typical attachment, discipline practices, involvement, parenting confidence, and relational frustration. An area of particular strength for participants is their involvement with their children. Parenting confidence and relational frustration represent potential areas for improvement for some participants.

Number of Participants in each Classification Range on the AAPI-2
The Bayley-III was used to assess the developmental functioning of Successful Families participants' children between the ages of 1 and 42 months. A total of 13 children completed the Bayley-III. Two children refused to complete the language subtests of the Bayley-III due to fatigue; thus, language data were only collected from 11 participants. In addition, four parents chose to take the social-emotional and adaptive skills questionnaires home to complete but did not return them; therefore, social-emotional and adaptive skills data were only collected for 9 participants.
There are five developmental domains measured by the Bayley-III, consisting of Cognitive, Language, Motor, Social-Emotional, and Adaptive Skills. Across each of these developmental domains, mean scores were in the average range.

Number of Participants in each Classification Range on the BASC-3 PRQ
The NEPSY-II was used to assess the neurocognitive functioning of Successful Families participants' children ages 4 and older given that the Bayley-III can only be used with children up to age 42 months. A total of five children completed the NEPSY-II, although two children refused to complete the speeded naming and word generation subtests. Four domains are measured by the NEPSY-II, consisting of Language, Memory and Learning, Sensorimotor, and Visuospatial. Figures 1.3 and 1.4 depict eight subtests within these domains with respect to the number of participants in each classification range. There was variation among scores on the NEPSY-II. More specifically, participants showed the most difficulty on a language task that required them to name body parts and a memory task that required repetition of sentences. Participants showed relative strengths on a visuospatial task as well as a language task that required rapid naming of colors and shapes. In general, most participants appear to be following a fairly typical developmental trajectory, with areas of strength and weakness that are reflected in the general population.

Discussion
The purpose of this study was to describe teen parents' perspectives on their relationship with their children, resilience, self-esteem, and parenting attitudes, and to measure the development of teen parents' children involved in the Successful Families supportive housing program. The data utilized from the current study is part of a larger research study (see Tremblay et al., 2018;Tremblay et al., 2020). Although the sample size for this study is small, thereby limiting the generalizations that can be made based on the data, this study makes a contribution to the limited literature in this area by providing insights relevant to practice and laying a foundation for future research; in particular, how we can build predictive or correlational research designs to measure the characteristics of teen parents and their children over time.

Parent Characteristics
Results indicate that self-esteem is an area of difficulty for participants. This is important because self-esteem has implications for the functioning of teen parents and their children. For example, self-esteem has been found to be a significant predictor of parental behaviours in response to infant distress (Leerkes & Crockenberg, 2002), mediate the effects of daily stressors on depression symptoms (Hall et al., 1996;Orth et al., 2009), and contribute to positive social behaviour (Mann et al., 2004). Although researchers have not widely explored teen parents' selfesteem, this population faces a number of challenges, including higher rates of mental health difficulties, social isolation, and adverse childhood experiences, that may inhibit the development of healthy self-esteem (Cox et al., 2008). These realities, combined with the ongoing stigma that many teen parents face, are consistent with the parents in this study struggling with self-esteem. This is an important area for service providers to consider in working with and offering resources for teen parents, and may be a priority given the implications for self-esteem with respect to overall wellbeing. It is also important to keep in mind that the Rosenberg Self-Esteem Scale was not normed with teen parents. Given that the reference group for the tool is the general US population rather than teen parents, it is not possible to compare the self-esteem of participants in this study with other teen parents, and this is true for the measurement of resilience in our sample as well.
Self-esteem and resilience are highly correlated, with self-reported self-esteem being a predictor of self-reported resilience (Balgiu, 2017). Given participants' relatively low mean score on the Rosenberg Self-Esteem Scale, the low mean score on the resilience tool (CD-RISC) may be expected, and represents an additional important area to target in programming for teen parents. Importantly, many participants came to the Successful Families program with limited housing options, and therefore from difficult living situationsfor example, co-habiting with a partner or parents with whom relationships were unhealthy and often abusive. The CD-RISC measures selfperceived resilience, and it is therefore conceivable that participants in the midst of leaving challenging circumstances may not perceive themselves as able to overcome challenges, thereby deflating their resilience scores. The Successful Families program works from a strength-based philosophy, helping participants recognize their potential for overcoming challenges, and it is possible that scores may increase after participants spend time engaged in the program.
With respect to parenting attitudes, most participants scored in the medium risk range across AAPI-2 scales. Results suggest that many participants harbor the attitude that obedience and complete compliance to parental authority should be demanded. It is possible that, for some participants, awareness of the stereotypes surrounding teen parenting (for example, their children being disobedient, frequently "acting up") may lead to fears about their children's behaviour, and therefore lead parents to lean toward oppressing power and independence rather than risk their children displaying disobedience. Results also suggest that participants may experience their children's needs and desires as overwhelming, and that their children's needs may come into direct conflict with their own needs. Given that teen parents are navigating their own developmental processes and needs, it is understandable that they may require additional support to develop a high level of empathic awareness for their children's needs. Results also reflect that most participants are in the medium risk range for a belief in corporal punishment. The rationale for using corporal punishment is often to teach children right from wrong, and parents who value the use of corporal punishment might benefit from education regarding the potential risks of corporal punishment and benefits of positive reinforcement for shaping children's behaviour (Bavolek & Keene, 2010). In this vein, most participants are at medium risk for inappropriate parental expectations, which may stem from an inaccurate perception of children's skills and abilities. Parents may therefore benefit from education regarding the needs and capabilities of children at various stages of growth and development.
Overall, most parents fell into the medium risk range with respect to the parenting attitudes measured by the AAPI-2. Of note, few parents fell into the high-risk range on most scales, suggesting that most participants have the foundation for successful parenting across the areas measured by the AAPI-2. At the same time, results suggest that most participants could benefit from an enhanced understanding of how their children's needs may conflict with their own, the benefits of positive reinforcement, and their children's typical developmental needs, and to boost their empathic awareness. Given that it is possible to provide education and support to augment these areas, these results are promising.

Parent-child Relationships
Results show that the majority of participants who completed the BASC-3 PRQ are demonstrating typical attachment, discipline practices, involvement, parenting confidence, and relational frustration. An area of particular strength for participants is their involvement with their children. Relational frustration and parenting confidence represent potential areas for improvement for some participants, and could be an important area to focus on in programming. In addition, some participants were particularly strong in certain areas, reflected by higher scores within and across tools, and it might be possible for these participants to act as peer mentors to participants who are struggling with certain aspects of parenting. Overall, results reflect that, for the 12 participants who completed the BASC-3 PRQ, teen parent-child relationships are typical, with evident areas of strength.
Depending on their age, the children of teen parents completed the Bayley-III or NEPSY-II to measure their development. It is not possible to directly compare the results of the Bayley-III and NEPSY-II because each of these tools measures different constructs. In addition, sample sizes for both tests were small. In general, however, it appears that the children of teen parents are developing on a fairly typical trajectory, with areas of strength and weakness. Although measured differently between tools, one area of weakness across the Bayley-III and NEPSY-II for this sample was in the language domain. This has implications for programming, as the Successful Families program could intentionally support parents to develop a language-rich environment and/or create programming conditions for co-learning, thereby bridging gaps that parents may also have. It is also worth noting that visual and motor areas were stronger across both tools. Therefore, these areas of strength and weakness may simply reflect different prioritization-that teen parents in this study were investing time in applied and visual activities for their children, for example.
Overall, the profile of the participants' children's developmental domains appears consistent with that of the general population, in that some children are above the expected level in certain areas, some children are below the expected level in certain areas, and most children are at the expected level in most areas. This is an important finding given the stigmas surrounding teen parents and the widespread assumption that the children of teen parents may lag behind their peers developmentally. Researchers examining neural plasticity have demonstrated an enhanced capacity for resilience during the early years when supports and intervention are in place (Shonkoff, 2011).

Implications for Policy and Practice
This study reinforces the heterogeneity of teen families, with teen parents and their children showing different areas of strengths and challenges across the domains measured. This suggests that service providers and policymakers should steer away from focusing on teen parents as inherently at risk and aligns with researchers who have recently suggested that teen families may face risks that are more related to the social determinants of health than parental age (see Diaz & Fiel, 2016;SmithBattle, 2013). Teen families living in challenging circumstances (e.g., unsafe housing) face risks for poor social, economic, and health outcomes, but so do other families living in unsafe housing who are headed by older parents. In addition, not all teen parents live in challenging circumstances, and those who do have varying levels of resources available to navigate their challenges.
Along these lines, it is well understood that teen parents and their children, like other families, live in ecological systems that contribute risks and protective factors to healthy development and functioning. For this reason, it may be most helpful to shift policy and practice toward focusing on how these systems (e.g., programs, communities, education systems) can prevent and address challenging circumstances, such as poverty, for teen families rather than focusing heavily on preventing teen pregnancy in the first place. Simply put, it is unreasonable to expect teen parents to disrupt what are, in many cases, intergenerational cycles of poverty, unstable family environments, and difficult childhood experiences without helping them access the means to do so. 180 A promising means for supporting teen families, as enacted by the Successful Families program, is the use of a strength-based approach which can aid in developing programs and policies that are effective and meaningful for children and families (Black & Hoeft, 2015;Ricks, 2016). Strengths are discovered through relationships, emphasizing the importance of a relational approach to working with teen families (Saleeby, 2008). This represents a shift away from the deficit focus that has infused research and public policy regarding teen families toward a focus on the resources, strengths, and assets of teen parents and their children as well as the provision of learning and support opportunities to bridge potential and strengths. Teen parents have wisdom and knowledge that can be critical to dealing with challenges, and the family unit itself can represent a source of strength and resources. Service providers' use of a strength-based approach can instill a focus on strengths in teen parents' own interactions with their children. A strengthbased perspective may also serve to disrupt pervasive stereotypes about teen families, thus impacting parents' experiences of stigma.

Implications for Research
As with policy and practice, research with teen families would benefit from taking a strength-based approach rather than placing emphasis on the ways that teen families fall short in comparison to their counterparts. We found that using a strength-based approach to our research facilitated rapport with staff and participants, contributed to effective knowledge mobilization, and resulted in the rapid uptake of findings to inform programming.
In order to provide an accurate picture and contribute to an enhanced understanding of effective supports for this population, researchers must be well-prepared to deal with measurement challenges. We found that, although a correlational or experimental design could further add to our understanding of how to support teen families, the descriptive design of the current study was necessitated by the difficulty of collecting data from this population (see Tremblay et al., 2018). In particular, despite that this was a highly participatory study through which strong, trusting relationships with participants and buy-in from staff were established over a period of four years, collecting data from participants posed a significant difficulty as reflected in the high rate of cancellations and no-shows that we experienced. Elsewhere, we have detailed the challenges in engaging teen parents in research that has the potential to highlight their weaknesses (Tremblay et al., 2018). In many instances, the same teen parents who did not show up for child development assessment appointments willingly took part in our qualitative, arts-based methods of data collection, the latter over which they had more control and agency with respect to the information they chose to reveal. Allowing researchers to examine parenting qualities, attitudes, and relationships involves inherent risks, and these may be felt in particularly acute ways by teen parents who are simultaneously experiencing stigma and judgment. Along with the challenge of a small sample size, it was not possible to engage all families in completing tools immediately upon program entry, which would have been ideal from a measurement perspective. In addition, our study did not include a comparison group of teen parents who were not accessing the Successful Families program or a matched sample of young people who had not experienced teen pregnancy, and this is an important direction for future research. Moreover, our sample was heterogeneous in with respect to the ages of children; thus, we were required to use separate child development tools depending on children's ages, which further reduced our sample size. Our participatory approach aided in the development of trust, thereby facilitating participant recruitment and data collection. Strong community-university partnerships are critical for conducting research with teen parents, along with the involvement of policymakers, decision-makers, and other stakeholders in order to translate research results to practice.
The challenges of collecting sensitive data from teen parents are augmented by the inadequate availability of tools that are developed and normed with this population. This in turn limits researchers' ability to accurately reflect the realities of this population and therefore how best to support teen families. There is a lack of tools that use teen families as a reference sample, and this is problematic for the conclusions that we can draw. For example, given that teen parents are navigating different hurdles, such as their own identity development while also raising their children, we might expect teen parents to score differently than older parents on measures of parent-child relationship quality, and it may be inaccurate and harmful to equate differences in scores with negative implications. We were able to find and use one tool (the AAPI-2) that has been normed with teen parents, although use of this tool assumes a deficit based position by identifying those that are at risk for abuse and neglect. This raises questions around how service providers may be contributing to elevated risk with a deficit-oriented, risk focused approach. In addition, although the reference sample for the AAPI-2 includes teen parents, scores are not derived based on parent age. Each of these complexities pose challenges for research with this population and limits generalizability.
In addition, we gathered information from teen parents who were functioning sufficiently so as to qualify for acceptance into the Successful Families program, and these parents likely differ in important ways from teen parents who are not housed, and also from teen parents who have sufficient natural supports available to them such that they do not require access to housing and other services. Given the specific context of this study and the unique subset of teen parents who meet criteria for entry into the Successful Families program, the sample of teen parents who participated in this study is clearly not representative of teen parents in general.
Rather than producing generalizable results, however, this study serves as a reference point for baseline information on the characteristics of teen families. It is important to have information regarding the contexts in which teen parent characteristics differ. Moreover, as well as serving as a reference point for baseline information, this study makes a contribution by supporting a shift away from a deficit focus toward a strength-based perspective that takes into account the heterogeneous, complex realities faced by teen parents and their children. In addition, despite the breadth of extant research documenting the challenges of teen parenting, there is a distinct lack of research that examines the development of quality parent-child relationships between teen parents and their children, as well as the resilience of teen families (Reisch et al., 2010). Therefore, this study serves as a starting point for investigating these areas and provides unique information by describing constructs that we know little about with respect to teen families.
Overall, there is a need for future research to more clearly elucidate the strengths and resources of teen families, along with the supports that can most effectively further their success, a corresponding need for programs and practices to align with this research, and a need for sustainable, relevant policies to scaffold conditions for success for programs, systems, and teen families themselves.