Early Head Start (EHS) is a two-generation program for low-income pregnant women and children up to age 3. EHS provides a wide range of services including child development services, child care, parenting education, case management, health care and referrals, and family support. The core service programs offer are home-based (with service delivered through weekly home visits and monthly group socializations), center-based (with child care services offered along with other activities), and family child care (with families served through family child care homes). Programs can provide one or more of these services to enrolled families based on its determination of the best mix of services to meet their needs. Home- and center-based services are by far the most prevalent at both the program and family levels (Vogel et al. 2011). Program quality in EHS is driven by adherence to the Head Start Program Performance Standards (ACF 1996). 

 The EHS Family and Child Experiences Survey or Baby FACES is the first national study of the program providing information that no studies have to date. It includes a nationally representative sample of 89 programs and families of nearly 1,000 children in two birth cohorts (newborns and 1- year-olds) and provides longitudinal data on program operations/management, service quality, characteristics of families and staff, service receipt, and child development. For the first time, program staff provide weekly reports on the services that families were offered and received throughout their enrollment along with reasons for missed services. Children and families are followed annually until age 3 or they leave the program. Key research questions include describing EHS, its staff and services, and understanding what services families and children in EHS receive while enrolled.

The paper focuses on providing an overview of Baby FACES (which competes its final year of data collection in spring 2012) and basic descriptive findings about programs and the quality and amount of services offered. It also presents multilevel models to examine the relationship between program, staff, child, and family characteristics and Early Head Start participation (defined as length of program enrollment, number of home visits and center days, and staff ratings of family engagement). For each model, level-1 represents the within-program (child- and staff-level) relationship between child, family, and staff characteristics and the Early Head Start participation measure of interest. Level-2 describes variation between programs and allows exploration of how program characteristics are related to participation outcomes. Intraclass correlations indicate the proportion of total variance in participation that is due to differences between children in different programs. By comparing estimates of the intraclass correlations before and after including program-level predictors in our models, we can examine whether and which program-level characteristics contribute to the variation in services.

 Findings presented in this paper have relevance for national policy and planning activities and technical assistance efforts as well as local self-assessment and improvement.