Recipient Organization
AUBURN UNIVERSITY
108 M. WHITE SMITH HALL
AUBURN,AL 36849
Performing Department
Human Development and Family Studies
Non Technical Summary
In 2017, approximately 20.4% of children (n = 213,513) in Alabama experienced two or more adverse childhood experiences (ACEs), including frequent socioeconomic hardship, parental incarceration, family violence, neighborhood violence, living with someone who was mentally ill or suicidal, living with someone who had a substance abuse problem, or racial bias (National Survey of Children's Health [NSCH], 2018). Studies documented the exposure to adversity in childhood as a strong risk factor for short- and long- term health outcomes, including poor health, smoking, heavy alcohol use, diabetes, obesity, anxiety and depression, low life-satisfaction, illicit drug use, and violence (Hudges et al., 2017; Huang et al., 2015; Petruccelli, Davis, & Berman, 2019).Between 2008 and 2014, 25% of children in Alabama reported living in families with incomes below the federal poverty line, and 15% of children were residents of high or concentrated poverty areas, defined as areas with higher crime rates, physical and mental health issues, and unemployment rates (U.S. Census Bureau's American Community Survey, 2008- 2014 5-year estimates). Disadvantaged neighborhoods are also often home to racial and ethnic minorities, low-income parents, single-mother households, and unemployed individuals (Cronin & Gran, 2018; Haas, Berg, Schmidt-Sane, Korbin, & Spilsbury, 2018; Mehra, Boyd, & Ickovics, 2017). Research suggests that children living in disadvantaged neighborhoods with fewer resources are at greater risk for malnutrition and death, abuse and neglect, and poorer developmental outcomes (Coulton, Crampton, Irwin, Spilsbury, & Korbin, 2007; Jimenez et al., 2019).Unfortunately, Alabama ranks as the state with the fifth highest child physical abuse rate, and it has a poverty rate well above the national average. In 2018, 12,060 children in Alabama were victims of at least one incident of child maltreatment, such as physical abuse (52%), neglect (42%), emotional abuse (<.5%), medical neglect (1%), and sexual abuse (16%) (National Child Abuse and Neglect Data System Child File, FFY 2000-2018). In view of the deleterious effects of physical punishment on childhood development, the proposed study considers the disciplinary practices that fathers employ to address children's behavioral difficulties. With increasing emphasis on the importance of paternal involvement in reducing risk factors during childhood and in promoting positive developmental outcomes in children at the earliest stages of the life cycle, the study aims to extend understanding of how different family structural-contextual and proximal processes are linked to father engagement and childhood outcomes across ethnic/cultural groups and socioeconomic backgrounds in diverse neighborhoods. Unlike other studies that approach family and childhood difficulties from a purely psychopathological or deficit perspective, the proposed work focuses on protective factors (neighborhood resources, paternal warmth and involvement) that may mediate how difficulties associated with fathers' mental health and harsh discipline ultimately impact childhood development. Should these and other protective factors emerge as successful mediators of childhood social and academic difficulties, they can be more fully implemented in future efforts aimed at enhancing the quality of father-child relationships and parenting in families in challenging home and neighborhood environments. Both community leaders (religious, elders) and mental health agencies that work with high-risk families can potentially use different elements of the results from the proposed analyses to shape existing practices.
Animal Health Component
(N/A)
Research Effort Categories
Basic
100%
Applied
(N/A)
Developmental
(N/A)
Goals / Objectives
Informed by process models of paternal engagement, family stress theory, risk and resilience perspectives and the bio-ecological systems model that focuses on contextual factors and intersystem relationships, the main goal of the proposed project is to provide an understanding of the complex interplay between neighborhood risk, paternal mental health, parenting, and children's socioemotional, cognitive, and language skills across racial/ethnic groups.The objectives of the proposed research are:(a) to create a neighborhood disadvantage index by utilizing a broad range of neighborhood factors, including neighborhood sociodemographic characteristics, crime index, overweight and obesity rates, prevalence of child maltreatment, area-level food environment factors and food access, supplemental nutrition assistance program participation and benefit levels, park and recreation access, availability of early childhood education programs and supportive services, and availability of local resources and health care services;(b) to examine neighborhood, family, and child-level factors that are associated with the level and quality of engagement of fathers with young children among resident and nonresident fathers across racial/ethnic groups;(c) to examine how fathers' mental health and different dimensions of fathers' engagement affect children's socio-emotional, cognitive, and behavioral outcomes, whether these associations differ across family structural arrangements, socioeconomic status, and neighborhood risk; and(d) to identify high-risk neighborhoods and underserved areas of Alabama and disseminate the results of this research to programs and services to make resources available to those who could benefit most.
Project Methods
1. To assess the spatial dependency between neighborhood socioeconomic and demographic characteristics, crime rates, child abuse and neglect rates, availability of early childhood education centers, local resources, and park and green spaces, and the quality of food environment, hot spot analysis using geographical information systems within the ArcGIS program will be used to identify clusters of statistically significant hot spots of each of neighborhood characteristics, The neighborhood socioeconomic characteristics will be geocoded in ArcGIS. Geocoded crime locations, geocoded ECE centers and park and green spaces, geographic and other geocoded information for the socioeconomic and demographic characteristics and neighborhoods will be used for modeling, and the sensitivity analysis will be conducted using cluster and outlier analysis within the ArcGIS program.2.To assess the associations between neighborhood risk (high crime rates, high child abuse and neglect rates, and low resources) and protective factors (local resources, quality of ECE centers, availability and accessibility of food environment, health care centers, and park and green spaces) and paternal involvement and discipline across racial/ethnic groups and family structures, neighborhood risk and protective factors will be created using spatially dependent functional analysis within the "adegenet" package in R program (Jombart, 2015). Neighborhood socioeconomic and demographic characteristics, crime index, the number of incidences of child abuse and neglect, the number and per capita density of health services, the number and per capita density of education and child development services, county- and census block-level food environments, the number and per capita density of local resources and park and green spaces will be subjected to spatial Principal Component Analysis (sPCA). Next, a geographically weighted regression will be conducted within the ArcGIS program. SPCA scores of neighborhood risk and protective factors will be entered as independent variables and centered individual level paternal involvement, and discipline will be entered as dependent variables to examine whether paternal involvement and discipline cluster spatially across different neighborhood factors.3.To assess the links between (a) paternal mental health, economic risk, and parenting differ across racial/ethnic groups, family structures, and neighborhood risk and protective factors, (b)neighborhood risk and protective factors and children's social, emotional, and cognitive skills across time, racial/ethnic groups, and family structures, and (c) mediating or moderating role of paternal mental health, economic risk, and parenting on the associations between neighborhood risk and protective factors and children's social, emotional, and cognitive skills across time, racial/ethnic groups, and family structures, a set of confirmatory factor analyses will be conducted to assess the psychometric properties of father and child constructs across resident and non-resident fathers, and racial/ethnic groups using Bayesian confirmatory factor analysis using the "blavaan" package in R program (Merkle & Rosseel, 2018). Model estimation will be performed using Monte Carlo Markov chain and the Gibbs sampler with 50,000 iterations, where the first 25,000 will be discarded as burn-in and the remaining 25,000 will be used to estimate the posterior distribution. Chain convergence will be monitored by using potential scale reduction factor (PSRF), in which PSRF less than 1.01 indicates model convergence, and trace and density, autocorrelation, and posterior predictive checking scatter plots will be evaluated. Non-informative priors will be used and model fit will be assessed by using posterior predictive p-value (ppp), a Bayesian variant of the root mean square error of approximation (BRMSEA; Hoofs, van de Schoot, Jansen, & Kant, 2018), and incremental fit indices including BCFI, BTLI, and BNFI. A ppp value around .10 (Cain & Zhang, 2018), a BRMSEA value smaller than .08, and higher BCFI, BTLI, and BNFI values indicate good model fit (Hoofs et al., 2018). Bayes Factor, widely available information criterion (WAIC), and leave-one-out cross-validation (LOO), will be used to compare the difference of the fit between models (Liang & Luo, 2019).Next, a series of Bayesian multi-level structural equation modeling with spatial random effects will be conducted within the Stan program (Stan development team, 2018) "RstanArm" package using multiple chains of Gibbs sampling. Model convergence and diagnostics, autocorrelation among parameter samples, and trace plot of sample chains will be evaluated using "rstantools" and "shinystan" packages (Stan Development Team, 2018). The first level will include father and child (paternal depression, paternal warmth, involvement, and discipline, economic risk, children's social, emotional, and cognitive skills), and control variables, and the second and third levels will include risk and protective factors across neighborhoods.