One of the most important factors influencing child and adolescent development is the quality of parenting and parent-child relationships (Sanders & Turner, 2018). Positive parenting practices include parental monitoring, warmth, sensitivity, communication, encouragement, and support (Perkins et al., 2013; Walker, 2021). Research suggests that when parenting occurs under favorable conditions, for example, within safe, well-resourced, low-crime neighborhoods that support parenting practices, children and youth benefit in several key developmental areas (Perkins et al., 2013; Sanders & Turner, 2018). These areas include healthy brain development, language and communication, social-emotional development, positive peer relationships, academic success, and physical and mental health (Perkins et al., 2013; Sanders & Turner, 2018). When parenting occurs under conditions that incite chronic stressors due to aspects of daily living related but not limited to income, poverty, inadequate neighborhood infrastructure (e.g., housing, schools, health services), precarious physical environments (e.g., access to clean air, clean water, lead free), and perilous social environments (e.g., crime and safety), children and youth’s health and development are adversely affected (Perkins et al., 2013; Sanders & Turner, 2018). These chronic stressors that parents experience impact parenting practices and styles that facilitate healthy development by reducing parent’s capacity to offer warm and responsive interactions that support well-being (Chaudry & Wimer, 2016).

Given the unique social and environmental conditions faced by families living in public housing, compounded by the social problems and violence that are prevalent in that setting, there is a need to accurately assess the relationship between parenting practices and child and adolescent outcomes in a way that is culturally sensitive and reflects the unique parenting practices of maternal and paternal caregivers. Doing this effectively demands that researchers use valid and reliable measurement tools. We contribute to research on youth’s perception of parenting practices within the context of public housing by examining the validity of the Parental Attitudes Measure (PAM) (Lamborn et al., 1991). We specifically examine the scales’ performance when youth report on both maternal and paternal parenting practices, respectively.

The Context of Urban Public Housing

Social context plays a significant role in adolescent welfare (Ross & Mirowsky, 2009). Specifically, evidence points to the influence of housing and neighborhood characteristics on adolescent appraisal of their social environment, efficacious beliefs, and depressive symptoms (Coley et al., 2013; Perry et al., 2015). This is particularly relevant in understanding the day-to-day functioning of urban African American adolescents living in public housing in the United States. Public housing provides a unique social context for youth and families in the U.S. As Rothstein (2017) observed, African American households tend to be overrepresented in public housing, many of which are located in under-resourced, racially segregated neighborhoods due to a legacy of racial bias and discrimination in the U.S. as well as housing policies that favored White families and households. This legacy created what Jargowsky (2015) refers as a durable architecture of segregation and others have described as a complex and reinforcing bundle of public policies, which included the systematic clearing and dispersal of middle-class African American neighborhoods, construction of intentionally segregated public housing projects, redlining of majority African American neighborhoods to deny home loans and ownership opportunities to African American families. Other outcomes include various blockbusting activities to invoke fear among White homeowners and induce white-flight to suburban communities all coalesced to spatially segregate and isolate African American families into marginalized, low-income neighborhoods throughout every city in the U.S. (Gordon, 2008; Jargowsky, 2015; Massey & Denton, 2003; Sharkey, 2013). As a result, the housing and neighborhoods in which African American families were confined experienced substantial systematic disinvestment across multiple decades and created socio-ecological risk factors that jeopardized family stability and parent capacity. Thus, families living in public housing occupy neighborhoods with poverty rates much higher than families participating in other affordable housing programs in the U.S. (National Low Income Housing Coalition, 2012). These public housing neighborhoods are often epicenters of drug markets and, consequently, ground zero for the war on drugs, which are marked by the violence that drug involvement tends to leave in its wake (Alexander & West, 2012; Nebbitt, 2015; Nebbitt et al., 2014). Despite attempts to address housing quality issues and deconcentrate poverty through mixed-income development (e.g., HOPE VI and Choice Neighborhoods) and the use of housing choice vouchers (HCV) to move low-income families into the private housing market, public housing remains a significant housing policy in the U.S. Indeed, approximately 3000 local housing authorities operate public housing across 7,000 properties totaling nearly 1 million public housing units. An estimated 1.6 million residents currently reside and are served by public housing in the U.S. Accordingly, the annual budget for public housing totals over $8 billion and has increased in recent years (U.S. Department of Housing and Urban Development, n.d.). Additionally, the shift in housing policy to private market solutions means that public housing increasingly serves a large proportion of hard-to-house families that face substantial barriers to housing and family stability. These challenges are particularly acute for African American families who make up the majority of public housing residents (43%) (Fischer et al., 2021). We focus specifically on youth in public housing because of their unique characteristics. Additionally, despite known associations between housing and neighborhood conditions on parenting practices (Kim et al., 2022; Office of Policy Development and Research, 2014; Pinderhughes et al., 2007), we address gaps in current research to better understand effects on parenting within the unique context of public housing. Indeed, in many large U.S. cities, the social context described above exacerbates the chances of adverse community exposure for African American adolescents as they negotiate familial, social, and contextual challenges.

Parenting Practices Within the Context of Public Housing

At its best, public housing provides for a safe and stable platform for the receipt of needed programs and services that promote family stability and upward socioeconomic mobility (Fischer et al., 2021). At its worst, unhealthy housing conditions, locations in under-resourced environments, and strict policies have the potential to exacerbate challenges facing low-income families and undermine the intended benefits of public housing (Fischer et al., 2021). Within this context, positive parenting practices and behaviors are a critical social determinant of health for children (Walker, 2021). Indeed, organizations like the World Health Organization consistently recognize parenting as a vital resource for healthy youth development (World Health Organization, 2022). Parenting behaviors also mitigate the effects of other social determinants, including adverse neighborhood environments on youth adjustment when parents limit exposure to neighborhood and community conditions that may jeopardize child health and safety (Nebbitt, 2015; Walker, 2021).

Parenting Practices and Youth Outcomes

It is well documented that parenting practices such as level of involvement, supervision, and nurturing behaviors play a role in child development and well-being (Baumrind, 1991; Cripps & Zyromski, 2009; Steinberg et al., 1991). Research has shown that levels of parent involvement can have direct benefits on self-efficacy, self-esteem, self-evaluation, and peer relations (Cripps & Zyromski, 2009; Noel et al., 2021). For example, a study on youth in secondary school in Uganda found that parenting styles combined (e.g., authoritarian, permissive, authoritative) yielded a high and positive significant relationship with self-efficacy (r = 0.620) (Noel et al., 2021). A study by (Nebbitt, 2009) found that among a sample of African American adolescent males from urban public housing, maternal support and supervision were positively and significantly associated with self-efficacy (r = 0.449 and 0.333, respectively). Similarly, a study by Doyle and colleagues (2017) found that paternal caregivers with high levels of parental encouragement and monitoring were associated with statistically significant higher levels of self-efficacy among African American youth living in urban public housing. Further evidence links parental behavior with youth depressive symptomatology (Estreet et al., 2018; Ge et al., 1994, 1996; Willoughby & Hamza, 2011; Yap et al., 2014). Namely, supportive and warm parenting practices have been related to decreased depression (Hashimoto et al., 2011; Piko & Balázs, 2012; Smokowski et al., 2015; Tian et al., 2013). Moreover, parental supervision has been shown to buffer mental health risk among African American youth living in urban communities (Estreet et al., 2018) and positive parenting practices of both maternal and paternal caregivers are associated with higher self-efficacy, and lower depressive symptoms (Nebbitt et al., 2021; Nebbitt et al., 2013; Doyle et al., 2017; Noel et al., 2021).

While there is growing evidence focused on parenting practices in childhood and adolescence and its long lasting effects on self-efficacy, mental health, and other important outcomes for children and youth (Boer & Tranent, 2013; Driscoll et al., 2008; Noel et al., 2021), research on African American families in public housing specifically, overwhelmingly focuses on single-female-headed households with children due to their over-representation in public housing (Nebbitt et al., 2013). Consequently, little attention has been directed to the role of fathers and other father figures on the well-being of children and youth among high-risk families (see Doyle et al., 2015; Estreet et al., 2018; Teel et al., 2016; Nebbitt, 2009; Nebbitt et al., 2013 for exceptions). A substantial proportion of parenting studies do not include fathers or father figures when examining children’s outcomes, and when they are included, measures of paternal parenting practices are often derived from assessments of maternal parenting practices (Cabrera et al., 2018). The importance of both maternal and paternal caregivers in the wellbeing of children and youth cannot be overstated. Valid and reliable measurement reflecting how maternal and paternal parenting practices are similar, different, complementary, or additive in the study of youth outcomes are critical in building a more robust body of research on supportive parenting practices for African American youth living in public housing (Cabrera et al., 2018).

The Parental Attitudes Measure (PAM)

The Parental Attitudes Measure (PAM; Lamborn et al., 1991) was created in response to calls for research that emphasized the interactive effects of different dimensions of parental behavior, namely the parental warmth and acceptance with parental control or strictness. Before the development of the PAM, studies focused on single categorization and prototypic parenting styles that included authoritative (e.g., firm and supportive), authoritarian (e.g., firm and directive but less supportive), and permissive (e.g., low control) styles of parenting (Baumrind, 1971; Baumrind & Black, 1967; Lamborn et al., 1991). Lamborn and colleagues were compelled to deconstruct permissiveness into two additional subcategories including indulgent permissiveness (low control with foundations in trust and democracy) and neglectful permissiveness (disengagement from responsibilities of child rearing) to reflect the variability in reasons for parental laxity. In doing so, they aimed to examine the interactive effects of parental warmth and strictness on adolescent outcomes (Lamborn et al. 1991). The PAM was developed using many items on parenting practices from existing measures (e.g., Dornbusch et al., 1985; Patterson & Stouthamer-Loeber, 1984). Lamborn and colleagues sampled from schools in Wisconsin and California that represented a diverse group of high school youth in terms of ethnicity, family structure, and community type (rural, urban, suburban) to examine the reliability and validity of a parental attitudes measure (PAM), focusing their study on adolescent’s self-report of parenting practices which centered their experiences (Lamborn et al., 1991). They averaged PAM scores of both parents in two-parent households and used scores from mothers in single-parent households. Although three factors emerged from their exploratory factor analysis, Lamborn and colleagues focused on two factors which were parental acceptance/involvement and strictness/supervision, which is consistently with Maccoby and Martin’s (1983) framework. The diverse sample produced acceptable reliabilities for each dimension (α = 0.72 and α = 0.76, respectively; Lamborn et al., 1991). Findings revealed that parental acceptance and involvement appeared to be key contributors to psychological well-being and positive self-concept, while strictness and supervision were suggested to deter behavioral problems (Lamborn et al., 1991). Further, they suggest that the combination of strictness and acceptance may have a positive effect on achievement (Lamborn et al., 1991).

While the scale performed well with the diverse sample, it is important to note that the sample of African American youth only comprised 9% of youth in the study, and the community of urban public housing was not explored. Even so, the PAM has been used extensively with African American youth. Five parental monitoring questions from the PAM were used in a study of 205 5th to 8th grade African American urban youth and had an acceptable reliability (α = 0.69) (Richards et al., 2004). This study found age and sex differences in parental monitoring with younger adolescents and girls were found to be less monitored than older adolescents and boys. The scale also performed well with African American early adolescents aged 10 to 15 from two hospital emergency departments in Baltimore, Maryland. Within this population, the parental encouragement subscale and the parental supervision subscales performed acceptably (α = 0.75 and α = 0.89, respectively), with each subscale having statistically significant low to moderate associations with aspects of future orientation, including commitment to learning, goals, and hope (Hilley et al., 2019).

Across studies that use this measure with African American youth living in public housing, the measure performed acceptably with parental supervision (α = 0.76) and parental encouragement (α = 0.87 to 0.88) for youth aged 13 to 19 (Nebbitt & Lombe, 2008; Nebbitt, 2009; Nebbit et al., 2011; Nebbitt et al., 2014; Doyle et al., 2017; Yu et al., 2012). The scale also performs well when looking specifically at maternal and paternal supervision (α = 0.84 to 0.92 and α = 0.76 to 0.93 respectively) and maternal and paternal encouragement (α = 0.88 to 0.92 and α = 0.83 to 95, respectively) among African American youth ranging from ages 11 to 21 years old living in public housing across studies (Nebbitt, 2009; Nebbitt et al., 2013; Foell et al., 2021). The studies focusing on African American youth living in public housing found statistically significant associations between maternal, paternal, and general parental attitudes and youth outcomes. For example, statistically significant associations between maternal support and self-efficacy among African American adolescent females (r = 0.268; Nebbitt & Lombe, 2008), between paternal support and both exposure to delinquent peer and delinquent behaviors (r = −0.19 and −0.24, respectively; Nebbit et al., 2011); and between parenting behaviors and depressive symptoms (r = −0.27; Foell et al., 2021).

Aside from the scale development by Lamborn and colleagues (1991), the PAM has only been validated for use among Turkish youth (Yılmaz, 2000). Within this population, both the test-rest reliability and Cronbach’s alpha were acceptable or very good for the encouragement subscale (0.82 and 0.70 respectively) and the supervision subscale (α = 0.88 and α = 0.69 respectively) (Yılmaz, 2000).

While the PAM has been used in research across diverse populations of African American youth, it has yet to be validated within the context of public housing. Validating a this measure is particularly significant considering the crucial role of parenting as a protective factor against social determinants prevalent within public housing, that may compromise child health and safety. Previous studies also have not looked at the psychometric properties of this scale by exploring maternal and paternal attitudes specifically. To fill this gap, we examined how the Parental Attitudes Measure (PAM; Lamborn et al., 1991) performed with a sample of African American youth living in urban public housing. Specifically, we examined if there were differences in how the measure performed when youth were asked about maternal and paternal caregivers, respectively. This study makes a signficiant contribution to research and practice about the scale’s validity within the context of public housing and the scale’s performance with maternal and paternal caregivers.

Theoretical Orientation

The Relational and Resilience Theory of Ethnic Family Systems (R&RTEFS) is a theory that has evolved from sociological research on family stress and coping that examines family disruption and variability in life changes (McCubbin & McCubbin, 2013). While theoretical perspectives have traditionally focused on predominantly white, middle class family systems, R&RTEFS highlights the limitations of examining families of ethnic diversity within a Western nuclear family perspective and recognizes that ethnically diverse family systems have depth, values, and beliefs that guide their functioning within the ecological context by which they live (McCubbin & McCubbin, 2013). The R&RTEFS is guided by a conceptual framework with three core elements: family schema, family patterns of functioning (POF), and family relational well-being (McCubbin & McCubbin, 2013). Family patterns of functioning is particularly relevant for our study in that POF facilitates the family’s development and maintenance of supportive relationships (McCubbin & McCubbin, 2013). Patterns of functioning may include relationships maintenance (e.g., supportive and encouraging relationships) and system maintenance (e.g., control and supervision). These POF are essential considerations for examining parental attitudes among African American youth living in public housing to develop appropriate interventions, provide support, and suggest policy change to address barriers and issues impacting this population. Furthermore, this theoretical orientation underscores that African American families living in urban public housing have strengths and values that guide positive relationships despite being faced with stressors that are magnified as a result of the ecological context in which they live.

Purpose

Our study advances two goals. The primary goal is to produce psychometric data (i.e., factor loadings and internal consistency) on an adapted version of the Parental Attitude Measure (PAM) that focuses on the attitudes of maternal and paternal parenting practices among African American youth living in public housing. The second goal is to evaluate the empirical and theoretical meaningfulness of the adapted PAM within this understudied population, test the construct validity, and examine mean differences between maternal and paternal PAM scores. Therefore, this study examines the following questions: (1a) Do the underlying dimensions of the PAM persist among youth living in public housing reporting on attitudes of both paternal and maternal caregivers, respectively? (1b) What is the internal consistency of the adapted PAM among youth living in public housing who report on attitudes of both paternal and maternal caregivers? (2) Does the measure retain a good model fit when youth report on both paternal and maternal caregivers? (3) What is the strength and direction of the association between the paternal and maternal PAM measure with theoretically related constructs (e.g., depression and self-efficacy)? (4) Are there differences in youth attitudes of paternal and maternal parenting practices across cities and by gender? Using data collected from 660 African American youth living in six urban public housing developments across three cities, question one will be answered using exploratory factor analyses (EFA), and question two will be answered using confirmatory factor analyses (CFA). Question three will examine construct validity using Pearson correlations, and question four will be answered using univariate and bivariate statistics.

Method

Research Setting

The study was conducted in family-only public housing locations in three large cities. The sample for this paper is composed of adolescents living in six public housing developments located in New York City (Queens), North Philadelphia, and Washington DC. Given the purpose of the parent study (see Nebbitt, 2015) to contribute to knowledge related to youth’s mental and behavioral health within the context of means-tested urban family public housing developments, elderly and disabled housing, HOPE VI redeveloped neighborhoods, and Section 8 residents were excluded from the study. Section 8 residents were excluded from the study given the possibility for recipients of this voucher program to rent privately owned residences (Williams Shanks et al., 2022). While HOPE VI’s revitalization efforts have yielded benefits to public housing (Coley et al., 2023; Nebbitt, 2015), it is noteworthy that 70% of the U.S. public housing stock remains untouched by these initiatives (Nebbitt, 2015). Consequently, this study excluded HOPE VI redevelopments and emphasizes an examination of non-transformed public housing developments is centered in this study of low-income urban families. Descriptions of each research site are provided below.

Washington DC

City one includes two low-rise barrack-style housing developments. Two hundred and twenty-three residents lived in site one consisting of 58 low-rise barrack-style buildings with 234 units occupying four city square blocks. For hundred and twenty-six residents lived in site two, which consisted of 108 low-rise barrack-style buildings housing 432 apartments and occupying five square city blocks. Ninety-eight percent of the families were African American. Fifty-eight percent of the residents were under the age of 18. In 2006, the median household income was approximately $10,200 (Office of Resident Services, 2006).

New York City

City 2 included two housing sites. The first site comprised 96 six-story high-rise housing developments, which hosted 3142 apartments. This site occupied approximately 10 city square blocks and was home to slightly under 3000 families. In 2006, the median household income was about $20,000. Minority families (African American 60% and Latino 33%) represented over 90% of the households in this site. Seventy-five percent of the population were under the age of 18, with 60% being between the ages of 10 – 18. The second site in City 2 included 31 six- and seven-story buildings. The location was home to slightly over 2100 families. This site occupied approximately seven city square blocks. In 2006, the median household income was slightly over $22,000. Minority families (African American 45% and Latino 36%) represented over 80% of the families in this site. Fifty-three percent of the population is under the age of 18, with 16% being between the ages of 10–18 (Department of Resident Service, 2007).

Philadelphia

City three included two housing sites. The first site consisted of 43 buildings containing 535 units. The second development housed two 17-story high-rise buildings containing 499 units. The two sites had approximately 2230 residents, 63% under the age of 21. Ninety-eight percent of the residents were African American. The 2007 median household income in the two housing sites was approximately $7500.

Sample Recruitment and Data Collection Procedures

The study was conducted with the full approval of the local housing authorities in each city. Recruitment consisted of flyers and announcements at local community centers. Specifically, members of the research team and employees at community-based organizations posted flyers in community centers and in agencies around each public housing. In addition, recruitment cards were distributed to youth living in the target location. Flyers and recruitment cards included a brief overview of the study, the date and location for data collection, and contact information for the principal investigator (PI) and research assistant (RA). The department of recreation and a social service agency in or near each study site provided space for data collection. The Institutional Review Board at Howard University approved the study’s protocol (IRB 06-SW-05).

The following criteria were used for inclusion into the study: being an African American youth (aged 11–20) and residence in the targeted housing sites. The exclusion criterion was the inability to demonstrate the capacity to give informed consent as determined by the Capacity-to-Consent Screen (Zayas et al., 2005). These criteria were used to screen out youth who could not demonstrate the capacity to give informed assent or consent. Only youth who already provided parental consent and youth assent were screened for capacity to consent.

Potential study participants who fulfilled the inclusion criteria met at the designated data collection sites. During this meeting, the PI and RAs provided a thorough overview of the study and explained the rights of human participants in research. Potential participants were provided parental consent and youth assent forms. Assenting youth 11–17 who obtained parental consent and youth aged 18 to 20 years old who provided informed consent received a Capacity-to-Consent Screen (Zayas et al., 2005). Only youth who scored eight or above on the Capacity-to-Consent Screen participated in the study. Two males (ages 13 and 14) in City 2 scored below the cut-off point and were dropped from the study. Youth who demonstrated the capacity to give informed consent met in groups of 10 to 15 youth to complete the survey. The RA read all survey questions and responses aloud, and youth circled their responses. The survey took approximately 40 min to complete. Youth received a $15 gift certificate and a snack for their time and participation. A detailed discussion of the methodology (e.g., eligibility, recruitment, interview schedule, measurement) and study findings for the larger project can be found in (Nebbitt, 2015).

Measures

Parental Attitudes Measure

Parenting practices were measured using the Parental Attitudes Measure (PAM) (Lamborn et al., 1991). The measure was modified to assess attitudes of paternal caregiving practices (i.e., fathers’ and father figures’) separately from attitudes of maternal caregiving practices (i.e., mothers’ and mother figures’), as opposed to a prompt of “parent” with no indication of which parent the youth is reporting on in the prompts. Respondents were provided five options to describe the person they consider their mother or father. These included: (1) birth mother or father, (2) a relative (e.g., aunt, uncle, grandparent), (3) an unrelated adult (e.g., foster parent, stepparent), (4) none of those people I live with are considered to be my mother or father, or (5) I live alone. In this study, options four and five were combined into one “other” category.

The PAM assesses two domains of parenting behaviors from the perspectives of youth. These domains include parental supervision and parental encouragement (Lamborn et al., 1991; (Nebbitt, 2015). Seven items assessed parental encouragement. Examples of these items included: “I can count on the person I consider my mother / father to help me out, if I have some kind of problem” and “The person I consider my mother / father keeps pushing me to think for myself.” Responses on this scale range from “Never” (1) to “Always” (4). Five items assessed parental supervision. Examples of scale items include: “The person I consider my mother / father knows where I am after school” or “The person I consider my mother / father knows what I do with my free time,” Responses on this scale ranged from “Don’t Know” (1) to “Know Exactly (4).” Higher scores indicate higher parental encouragement and higher parental supervision.

Self-Efficacy

Self-efficacy was used to examine convergent construct validity with the adapted PAM. As such, self-efficacy was expected to have a positive relationship with the PAM. Self-efficacy was measured using the General Self-Efficacy (GSE) scale, a self-report measure of a person’s belief in their competence to cope with a broad range of stressful or challenging demands (Schwarzer & Jerusalem, 1995). The GSE scale has been validated for use with youth and adults from age 12 across 25 countries, and in 30 languages (Lei et al., 2020). The GSE includes 10 items and has a response format ranging from Not true at all = 1 to True all the time = 4. Examples of questions in this scale include: “I can always manage to solve difficult problems if I try hard enough” and “I can solve most problems if I invest the necessary effort.” Total scores were summed, with higher scores indicating higher self-efficacy. Internal consistency for the GSE scale generally ranges from 0.76 to 0.90 (Schwarzer & Jerusalem, 1995). The PAM has demonstrated acceptable reliability in a sample of African American youth living in public housing (α = 0.69) (Nebbitt, 2015). The reliability of the scale with the present study sample was 0.94.

Depression

Depression was also used to examine divergent construct validity with the adapted PAM and was expected to have a negative relationship with the PAM. Depression was measured using the Center for Epidemiological Studies - Depression (CES-D) scale (Radloff, 1977). The CES-D is a 20-item measure that asks how often respondents have experienced symptoms associated with depression over the past week. The response options ranged from 0–3, with Less than 1 day = 0, 1 to 2 days = 1, 3 to 4 days = 2, and 5 to 7 days = 3. Examples of questions in this scale include: “I was bothered by things that usually don’t bother me” and “I thought my life had been a failure.” Total scores were summed, with higher scores indicating higher depressive symptoms. The CES-D has high internal consistency, with Cronbach’s α scores ranging from 0.85 with psychiatric patients and 0.90 across the general population (Radloff, 1977), and has been validated for use with African American youth living in public housing, generating a reliability of 0.90 (Lu et al., 2017). The reliability of the scale for this sample was 0.91.

Analytic Procedures

Prior to conducting our analyses, we evaluated the fitness of our data by examining histograms, skewness, and kurtosis. Skewness among the sample ranged from −0.95 to −0.236 and kurtosis ranged from −1.27 to −0.499. Both skewness and kurtosis statistics values fell within the satisfactory range between −1 and 1 and −2 and 2 respectively (Tabachnick & Fidell, 2014). Univariate tests (e.g., percentages, means, ranges, and standard deviation) were conducted to describe the sample. Bivariate and multivariate tests (Kruskal-Wallis, Wilcoxon Signed Rank Test, and post-hoc Dunn’s test) were used to identify differences across sample characteristics including city, gender, and who they considered to be their mothers and fathers and who they reported living with.

An exploratory factor analysis (EFA) was conducted using principal components analysis (PCA) with varimax rotation. Eigenvalues greater than 1 and scree plots above the elbow of the curve were used to determine how many factors were extracted (Costello & Osborne, 2005). In addition to eigenvalues and scree plots, acceptable factor loading for items was set at > 0.3, with a minimum of three items loading above 0.3 to retain each factor (Costello & Osborne, 2005). Finally, the Kaiser-Meyer-Olkin measure of sampling adequacy (KMO) and Bartlett’s Test of Sphericity were conducted to ensure suitability of the data for structure retention (Izquierdo et al., 2014).

Confirmatory factor analyses (CFAs) were then conducted to assess model fit. We used robust weighted least squares (WLSMV) as the estimator because it is the recommended estimator for ordinal data and makes fewer assumptions than maximum likelihood (Bowen & Masa, 2015; Muthén & Muthén, 2017). A nonsignificant chi-square test indicates that the covariance structure of the data fits the covariance structure of the hypothesized model. Additional fit indices were also examined to provide a rigorous assessment of model fit. These criteria for model fit include generally accepted standards for fit indices, including RMSEA < 0.06; SRMR < 0.08; CFI > 0.90; and X2 > 0.05 (Boomsma, 2000; Kline, 2005).

Following the CFA with maternal and paternal parenting attitudes, construct validity was examined by calculating Pearson correlations. We tested convergent construct validity between the maternal and paternal PAM with self-efficacy and divergent construct validity with depression. Analyses were conducted using SPSS version 26 (IBM Corp., 2019) and Mplus version 8.4. (Muthén & Muthén, 2017).

Results

Sample Demographics and Univariate Descriptives of Key Study Variables

Sample characteristics are presented in Table 1. The age of respondents ranged from 11 to 20 years (M = 15.39; SD = 2.4). Fifty-two percent of the sample identified as male, and 47.4% identified as female. Seventy-nine percent of participants considered their birth mother as the person in their household who they consider to be their mother, while 15% of participants considered a relative (e.g., aunt, grandmother) or an unrelated adult (e.g., foster parent or stepparent) in their household as the person they consider to be their mother. The remaining 5.2% of participants were categorized as “other,” which includes the option that they do not consider any of the people they live with as their mother or the option that they live alone. Fifty-seven percent of participants considered their birth father as the person in their household whom they consider to be their father, while 22.8% of participants considered a relative (e.g., uncle, grandfather) or an unrelated adult (e.g., foster parent or stepparent) in their household as the person they consider to be their father. The remaining 19.3% of participants were categorized as ‘other’ with the definition described above. Seventy-seven percent of participants responded “yes” to the presence of their mother or a female living in the household, while 54.1% responded yes to the presence of their father or a male living in the household.

Table 1 Sociodemographic Characteristics of African American Adolescents (N = 660)
Full size table

Bivariate and Multivariate Analyses of Maternal and Paternal PAM items

Mean scores and standard deviation for items on both the maternal and paternal parental attitudes measure are presented in Table 2. A Wilcoxon Signed Rank Test was conducted to test differences between the mean of items for maternal and paternal parenting attitudes across participants. Statistically significant differences were found for each item, with scores for maternal attitudes being statistically significantly higher than paternal attitudes.

Table 2 Descriptive statistics of maternal and paternal parental attitudes measure items and differences in responses
Full size table

Mean scores and standard deviations for items on both the maternal and paternal parental attitudes measure were also examined by gender and by city. The Wilcoxon Rank Sum Test was used to examine differences by gender and the Kruskal-Wallis Test was used to examine differences by city. Statistically significant differences by gender emerged for every item of the maternal PAM except item 5: “When my parent wants me to do something they explain why” and item 9 “My parent knows how I spend my money.” Contrarily, statistically significant differences by gender emerged for only four items of the paternal PAM, including item 4 “My parent helps me out with my school-work if there is something I don’t understand,” item 5 “When my parent wants me to do something they explain why,” item 9 “My parent knows how I spend my money,” and item 12 “my parents know what I do with my free time.” Results for gender are presented in Table 3.

Table 3 Descriptive statistics of maternal and paternal parental attitudes measure items by gender and differences in responses
Full size table

When examining differences by city, statistically significant differences were found for 2 items on the maternal PAM and 1 item on the paternal PAM. For the maternal PAM, these differences emerged for item 1 “I can count on my parent to help me out if I have some kind of problem” (χ2 = 8.71, p < 0.01) and item 6 “When you get a bad grade in school how often does your parent encourage you” (χ2 = 9.16, p < 0.01). A Dunn test was used to examine which cities were significantly different on these items. For item 1, differences emerged between Washington DC and New York City (p < 0.01). For item 6, differences emerged between Washington DC and New York City (p < 0.01) as well as between New York City and Philadelphia (p < 0.05). For the Paternal PAM, these differences only emerged for item 6 with differences found between Washington DC and New York City (p < 0.01). Results by city are presented in Table 4.

Table 4 Descriptive statistics of maternal and paternal parental attitudes measure items and differences in responses by host city
Full size table

Exploratory Factor Analysis

The results of the exploratory factor analysis indicated that a two-factor structure was recommended for both maternal and paternal PAM based on eigenvalues greater than or equal to 1.0 (Kaiser, 1960). For the questions focused on maternal parenting attitudes, factor 1 eigenvalue was 5.534 and factor 2 eigenvalue was 1.532. For the questions focused on paternal parenting attitudes, Factor 1 eigenvalue was 7.125 and Factor 2 eigenvalue was 1.547. Factor 1 was named encouragement and factor 2 was named supervision. All items loaded above 0.3 with excellent KMO and Barlett’s Test of Sphericity values for both maternal [KMO = 0.914; Bartlett’s Test 3219.967(66) p = 0.00] and paternal [KMO = 0.945; Bartlett’s Test = 5649.614(66) p = 0.00]. The cumulative variance was 58.822% for mothers and 73.570% for fathers. Factor loadings for the total sample and by gender are displayed in Table 5.

Table 5 Factor loadings for the 12 maternal and paternal PAM items for the total sample and by gender
Full size table

The internal consistency of the adapted PAM was very good. For items related to maternal parenting attitudes, Factor 1 had a reliability of 0.86 and Factor 2 had a reliability of 0.84. For items related to paternal parenting attitudes, Factor 1 had a reliability of 0.93 and Factor 2 had a reliability of 0.92 (DeVellis, 2016). The total reliability for the maternal and paternal PAM was also examined by gender and by research site. For maternal PAM, the reliability for the total measure was 0.88 among female respondents and 0.89 among male respondents. For the paternal PAM, the reliability for the total measure was 0.94 among female respondents and 0.93 among male respondents. The measure had good reliability across cities as well with a reliability of 0.91 in Philadelphia and Washington, DC, and a reliability of 0.86 in New York City.

Confirmatory Factor Analysis

CFA results indicated that with the exception of chi-square, the two-factor model had acceptable fit for mothers χ2(53) = 234.791, p < 0.00; RMSEA = 0.075, CI [0.065, 0.085]; CFI = 0.97; SRMR = 0.04; and fathers (χ2(530 = 198.519), p < 0.00; RMSEA = 0.086, CI [0.058, 0.078]; CFI = 0.99, SRMR = 0.02) (see Table 6).

Table 6 Confirmatory factor analysis model fit indices for the statistics of maternal and paternal parenting attitudes measure (PAM)
Full size table

Construct Validity

Construct validity was evaluated by determining whether the Maternal and Paternal Parental Attitudes Measure subscales assessed in the EFA and CFA were significantly correlated with self-efficacy and depression. Previous literature (Nebbitt, 2009; Doyle et al., 2017; Noel et al., 2021) indicate that self-efficacy should correlate positively with parental attitudes and depression should correlate negatively (Lombe et al., 2021; Estreet et al., 2018; Ge et al., 1994, 1996; Willoughby & Hamza, 2011; Yap et al., 2014). Self-efficacy was positively and statistically significantly correlated with maternal encouragement (r = 0.31; p < 0.001), paternal encouragement (r = 0.08; p < 0.05), and maternal supervision (r = 0.22; p < 0.0001) but not with paternal supervision (r = −0.03; p > 0.05). Depression was negatively and statistically significantly correlated with maternal encouragement (r = 0.16; p < 0.001) and maternal supervision (r = 0.15; p < 0.001) but was not statistically significantly correlated with paternal encouragement (r = −0.04; p > 0.05) and paternal supervision (r = −0.03; p > 0.05).

Two post hoc analyses were conducted to examine the correlations under different conditions. The first post hoc analysis considered whether the maternal and paternal caregivers that participants reported on were their birth mothers and fathers or non-birth mothers and fathers. Correlations with maternal encouragement and supervision remained significant with self-efficacy and depression among participants reporting on birth mothers, however, there was no significant relationship detected between maternal encouragement and depression (r = −0.15; p = 0.09) nor maternal supervision and self-efficacy among participants responding about non-birth mothers (r = 0.07; p = 0.42). Similarly, correlations with paternal encouragement and supervision remained the same as the overall model with significance only emerging with paternal encouragement and self-efficacy (r = 0.14, p = 0.02), however, there were no significant relationships detected between paternal encouragement or supervision with self-efficacy and depression when examined among participants responding about non-birth fathers.

The second post hoc analysis considered whether the maternal and paternal caregivers that participants reported on lived in the home. Correlations with maternal encouragement and supervision remained significant with self-efficacy (r = 0.32; p = 0.01) and depression (r = −0.15; p = 0.001) among participants reporting the person they consider to be their mother lives in the household. Correlations with paternal encouragement and supervision differed compared to the previous analyses. There was a significant correlation between paternal encouragement and self-efficacy (r = 0.20; p = 0.01) and depression (r = −0.13; p = 0.05) among participants reporting the person they consider to be their father lives in the household. There was a significant relationship between paternal supervision and depression (r = −0.15; p = 0.01), but no statistical significance was found with paternal supervision and self-efficacy among participants reporting the person they consider to be their father lives in the household.

Discussion

The purpose of the current study was to determine the psychometric properties of the parental attitudes measure (PAM) for maternal and paternal caregivers among African American youth living in urban public housing. Assessing parenting practices from the perspective of youth is critical. Previous research has indicated that parenting practice decisions are influenced by neighborhood characteristics, and that parents’ perceptions of their environments affect how they permit their children to move within these spaces (Jacoby et al., 2017). Because stressful environments shape parenting practices, it is likely that youth perceptions of parenting practices may subsequently be impacted as well. As such, it is critical to explore youth interpretations of parenting practices given their relevance in predicting youth outcomes (Smokowski et al., 2015). Our study may be the first to test the validity and reliability of the parental attitudes measure among African American youth living in urban public housing, a unique social context for youth in the U.S., and may be the first to validate the use of this measure with maternal and paternal caregivers, respectively. Examining parenting attitudes by maternal and paternal separately is critical to examining differences in maternal and paternal attitudes from the perspective of youth, especially so as we understand the construction of family units to be diverse (PEW Research Center, 2015). Our validation of the PAM indicates the scales’ strength with this sample of youth generally as well as by gender and across six urban public housing sites in three cities. The validation of the PAM also supports the theoretical orientation of the Relational and Resilience Theory of Ethnic Family Systems which recognizes that ethnically diverse family systems have depth, values and beliefs that guide their functioning within the ecological context in which they live (McCubbin & McCubbin, 2013).

EFA and CFA analyses for maternal attitudes and paternal attitudes were performed and results indicated that, consistent with the development of the measure (Lamborn et al., 1991), a 2-factor model best fit the data for this sample. The two factors represented parental encouragement (items 1–7) and parental supervision (items 8 - 12) and the measure can be said to be successfully validated for both maternal attitudes and paternal attitudes among African American youth living in urban public housing with all the fit indices examined indicating a good fit. The validation of this scale both supports the interpretations of findings from previous studies that used the scales for maternal and paternal caregivers (see Nebbitt, 2009; Yu et al., 2012; Foell et al., 2021) and supports the use of this measure within this population and context. This is important as the reliability and validity of this measure is critical for understanding and developing implications for addressing caregiver-child relationships that are faced with stressors that may come with living in communities of urban public housing.

Although the reliability for the two subscales were higher for paternal parenting attitudes than they were for maternal parenting attitudes, the reliability was still acceptable (DeVellis, 2016). Therefore, we can be confident that regardless of the paternal role of caregivers (maternal or paternal) we can reliably interpret the parenting attitudes of African American youth living in public housing. The measure’s high reliability and 2-factor structure was also found when examined across cities and by gender of study participants.

Our study also included a Wilcoxon ranked sum test to compare responses from maternal parenting attitudes and paternal parenting attitudes. Results indicated that the mean score on each item for the total sample was statistically significantly higher for mothers than for fathers in this sample. One explanation for this is that maternal–youth relationships may be characterized by greater communication and intimacy (Graziano et al., 2009; Noller & Callan, 1990). When examined by gender, there were significant differences between male and female respondents in all but two items on the maternal PAM and only significant differences between male and female respondents for four of the paternal PAM items. Previous research by Noller and Callan (1990) found that in a sample of 296 youth aged 13 to 17, male youth were found to be equally satisfied with communication with both parents, however, female youth were found to be more satisfied with conversations with mothers rather than fathers (Noller & Callan, 1990). Furthermore, it is also likely that respondents may be rating their live-in maternal caregiver higher as there was a higher percentage of mothers in the household than fathers within this sample and mothers are traditionally over-represented in public housing compared to fathers (Nebbitt et al., 2013). Since the PAM asks about the person youth consider to be their mother and father, we examined how the distribution of responses may change based on whether participants were answering about birth mothers and fathers and non-birth mothers and fathers. We found a consistently higher score for both maternal and paternal encouragement and supervision among male and female respondents reporting on parental attitudes of birthparents compared to non-birth parents (see Fig. 1).

Fig. 1
figure 1

Comparison of paternal and maternal parenting attitudes of birth and non-birth parents

Full size image

We also tested for convergent and divergent construct validity. For maternal parenting attitudes, the measure met the criteria for acceptable construct validity as depression was negatively correlated with both maternal encouragement and supervision and self-efficacy was positively correlated with maternal encouragement and supervision. This means that youth attitudes about maternal parenting practices (both encouragement and supervision) are associated with lower symptoms of depression and greater feelings of self-efficacy, which is consistent with Nebbitt et al., (2021) assessment of maternal caregiver typologies and its impact on internalizing behaviors.

Among paternal parenting attitudes, depression was not statistically significant with paternal encouragement and paternal supervision. However, like the maternal encouragement subscale, paternal encouragement was positively and statistically significantly correlated with self-efficacy. One explanation for such differences in the associations between depression and self-efficacy across maternal and paternal parenting attitudes is that gender differences in parenting styles may be present (Walker & McKinney, 2015). For example, in a family composed of an active and involved mother and father, it is possible that one parent may take on a warm and supportive style akin to an authoritative parenting style, while another may take on a more authoritarian or permissive parenting style which may range from controlling and rigid or passive and laissez faire, respectively (Baumrind, 1968). Another explanation for such differences may be based on the variability of paternal parenting typologies unaccounted for in this study. In a 2013 study by (Nebbitt et al., 2013) on African American families living in public housing, a latent class analysis revealed four typologies of paternal caregivers, finding that youth who perceive their paternal caregivers to have high encouragement and high supervision, scored higher on self-efficacy and lower on depressive symptoms. It is possible that the PAM may converge with lower depressive symptoms and high self-efficacy, only under specific conditions. Even so, across this sample, the correlation with paternal supervision was lower than previous studies. More can be explored about how the context of public housing, particularly its uniquely strict policies may shape paternal parenting practices.

The findings from tests of construct validity may be also understood in the context of previous literature that has found that the status of the parental caregiver (e.g., birth vs. non-birth) to be important in explaining internalized behaviors in their children (Nebbitt, 2009). Although just over half (57.9%) of the sample reported they considered their birth parent to be their father, there was still a large significant percentage of non-birth paternal figures in the lives of the youth in this sample (42.1%). This signals opportunities for future research that considers engaging non-birth fathers and improving efforts to facilitate strategies to strengthen their important role in minimizing internalizing symptoms of non-birth children and youth (Forehand et al., 2015). While birthmothers were represented in greater numbers in the sample, insignificant findings related to maternal engagement and depression and maternal supervision and self-efficacy among youth reporting on parenting attitudes of non-birth mothers warrant attention. Maternal encouragement among youth reporting on non-birth mothers did not have a significant association with a reduction in depression as suggested in the literature (Estreet et al., 2018) and maternal supervision among youth reporting on non-birth mothers did not have a significant association with an increase in self-efficacy which is also contrary to expectations based on the extant literature (Nebbitt et al., 2021; Nebbitt et al., 2013; (Doyle et al., 2017)). These findings signal for future research about the context of relationships with youth and birthmothers compared to relationships with non-birth mother figures. For example, the conditions by which a young person gains a mother figure who is not their birth mother (e.g., foster care, adoption, marriage, or other circumstances connecting youth to maternal kin caregivers) including age of the youth when this person filled this role can be a consideration to explore.

Finally, tests of construct validity also examined associations between maternal and paternal encouragement and supervision among youth who report that their mother or father live in the household, irrespective of whether that individual was their birth- mother or father. Tests confirmed convergent and divergent construct validity for all but the relationship between paternal supervision and self-efficacy. These findings underscore the significance of having a maternal or paternal figure in the household, challenging negative stereotypes surrounding African American fathers and emphasizing their role in their children’s lives as well as their importance in the household. Moreover, these results call for additional research into how the protective influences of maternal and paternal encouragement and supervision may be leveraged among non-residential mothers and fathers to promote higher self-efficacy and lower depressive symptoms among African American youth in public housing.

Implications

In research with minoritized populations, it is critical to ensure the use of valid and reliable measurement. How we understand parent-child relationships is especially important among African American families who continue to be disproportionately represented in hyper-segregated and marginalized low-income urban public housing neighborhoods (Nebbitt, 2015; Center on Budget and Policy Priorities, 2021; Rothstein, 2017). Continued efforts can be made to examine paternal caregiving relationships (for both birth and non-birth parents) from the perspective of youth and its influence on self-efficacy and depressive symptoms. Future research may confidently utilize this measure to investigate parental attitudes change by considering the composition of the family home in understanding differences in attitudes based on living or visitation arrangements between maternal and paternal caregivers. Future studies may also examine how parenting attitudes are influenced by additional contextual factors like the number of children in the family, parent mental health, and the supportive network of adults that may or may not be present for supporting caregivers. The inclusion of these factors will strengthen the recognition that family systems have variability in structure, stressors, and supports that guide their functioning within the ecological context in which they live.

Limitations

Our findings should be considered in light of a few limitations. First, the data come from a convenience and purposive sample of African American youth living in six public housing communities located across three U.S. cities in the northeastern and mid-Atlantic region of the United States that are situated within an urban context. This study was conducted in collaboration with willing, available, and accessible communities. As is common in studies that use convenience and purposive sampling, it is not likely that the sample is truly representative of the entire population. While African American youth make up a sizable proportion of the residents across these public housing communities in this study (98% in Washington DC and Philadelphia; 45–63% in New York City) it is also not possible to claim representativeness without random selection. While this study contributes to a gap in knowledge about this specific population living in this region, there are limits the generalizability of our findings to settings that fall outside of the context and scope of our study. Secondly, the youth included in our study were those who obtained parental consent and who responded to recruitment efforts (e.g., local community center efforts and flyers). Therefore, our results may be biased in that they reflect the attitudes of participants who were fortunate to have obtained permission from parents and caregivers to participate in a study that asked about relationships with caregivers. Finally, given the focus on African American youth, the findings may not be representative of other groups of youth who live in urban public housing (e.g., Latino youth and youth from immigrant families).

Conclusion

With 1.6 million residents currently residing in public housing communities in the U.S., a focus on low-income urban families remains an important area of research. The modified PAM was found to be a valid and reliable tool for examining maternal and paternal parenting attitudes among African American youth living in urban public housing in the northeast and mid-Atlantic regions of the United States. Findings indicate acceptable construct validity with maternal caregivers, particularly with birth mothers, but only partially with birth fathers and non-birth mother figures, suggesting differences in youth’s attitudes on parenting practices and its association with depression and self-efficacy. Acceptable construct validity among maternal and paternal caregivers who live in the home signal the importance of parental figures in the household. Future research should examine parental caregiving relationships between youth and birth and non-birth fathers and mothers and its influence on self-efficacy and depressive symptoms with consideration on the variability in the structures, stressors, and supports that guide parenting practices within environments that incite may chronic stressors due to aspects of daily living.