Program Evaluation on Early Head Start

Early Head Start Relationships: Association with Program Outcomes

James Elicker

Human Development and Family Studies, Purdue University

Xiaoli Wen

Early Childhood Education, National College of Education, National Louis University

Kyong-Ah Kwon

Department of Early Childhood Education, Georgia State University

Jill B. Sprague

Human Development and Family Studies, Purdue University

Research Findings: Interpersonal relationships among staff caregivers, parents, and children have been recommended as essential aspects of early childhood intervention. This study explored the

associations of these relationships with program outcomes for children and parents in 3 Early Head

Start programs. A total of 71 children (8–35 months, M ¼ 20), their parents, and 33 program caregivers participated. The results showed that caregiver–child relationships were moderately

positive, secure, and interactive and improved in quality over 6 months, whereas caregiver–parent

relationships were generally positive and temporally stable. Caregiver–child relationships were more

positive for girls, younger children, and those in home-visiting programs. Caregiver–parent relation-

ships were more positive when parents had higher education levels and when staff had more years of

experience, had more positive work environments, or had attained a Child Development Associate

credential or associate’s level of education rather than a 4-year academic degree. Hierarchical linear

modeling analysis suggested that the quality of the caregiver–parent relationship was a stronger

predictor of both child and parent outcomes than was the quality of the caregiver–child relationship.

There were also moderation effects: Stronger associations of caregiver–parent relationships with

observed positive parenting were seen in parents with lower education levels and when program

caregivers had higher levels of education. Practice or Policy: The results support the importance of caregiver–family relationships in early intervention programs and suggest that staff need to be

prepared to build relationships with children and families in individualized ways. Limitations of this

study and implications for program improvements and future research are discussed.

Early Head Start is a federally funded community-based program for low-income families with

infants and toddlers and pregnant women, with goals to enhance child development and promote

healthy family functioning (Early Head Start National Resource Center, 2008). A guiding

Correspondence regarding this article should be addressed to James Elicker, PhD, Department of Human Development

& Family Studies, Purdue University, Fowler Memorial House, 1200 West State Street, West Lafayette, IN 47906-2055.

Early Education and Development, 24: 491–516

Copyright # 2013 Taylor & Francis Group, LLC

ISSN: 1040-9289 print/1556-6935 online

DOI: 10.1080/10409289.2012.695519

principle of Early Head Start is the importance of building positive relationships between program

caregivers and participating families in order to support positive change. Early Head Start advocates

that strong positive staff–family relationships that continue over time are key elements of a

high-quality program. More generally, in the literature, building positive staff–client relationships

has been recognized as an evidence-based best practice in early intervention programs targeting

high-risk young children and their families (e.g., Klass, 2003; Raikes & Edwards, 2009; Turnbull,

Turbiville, & Turnbull, 2000). These relationships include those among the child, the family, and

program caregivers, recognizing the parent–child bond as the child’s most significant relationship

(Lally & Keith, 1997).

Although building positive relationships is emphasized in the Early Head Start Performance

Standards (U.S. Department of Health and Human Services, 2002), in the National Association

for the Education of Young Children’s Developmentally Appropriate Practice in Early Child- hood Programs (Copple & Bredekamp, 2009), and in the goals of many early intervention programs, few studies have actually examined the association between dyadic interpersonal

program relationships and outcomes for children and parents. In this study we explored the qual-

ity of the interpersonal relationships among program caregivers, parents, and children in three

midwestern Early Head Start programs, with the goal of better understanding program relationships

and contributing data useful for improving services in Early Head Start and other relationship-based

early intervention programs. First, this study was intended to contribute new data about how infants’

and toddlers’ development proceeds in the context of relationships in Early Head Start. Second,

the study focused on describing program–family relationships: their nature and variation, how

they change over time, how they vary with program caregivers’ and families’ characteristics, and

whether they are associated with parent and child outcomes. These issues are critical for Early Head

Start, given that interpersonal relationships are viewed as a key component of program quality.

However, they have not been a primary focus of many studies, including the large-scale national

Early Head Start Research and Evaluation Project (Love et al., 2002).

MOTHER–CHILD ATTACHMENT RELATIONSHIPS: A FOUNDATION FOR EARLY INTERVENTION

The early experience of a majority of young children today encompasses a network of inter-

personal relationships, both within and outside the family. The child development and infant

mental health literatures have converged on a relationship perspective to conceptualize early

social-emotional development (e.g., Cicchetti, Toth, & Lynch, 1995; Elicker & Fortner-Wood,

1995; Lyons-Ruth & Zeanah, 1993). This relationship-focused perspective is based on modern

attachment theory (Bowlby, 1973; Sroufe, 1983), which suggests that infants’ everyday inter-

actions with their primary caregivers result in strong emotional relational bonds. Depending

on the sensitivity and consistency of caregiving over time, attachment bonds vary in the amount

and quality of felt security conveyed to the infant. Another attachment postulate is that primary

relationships are mentally represented by even very young children. These representations, or

internal working models, subsequently guide children’s behavior in interpersonal relationships,

including relationships outside of the family, such as relationships with peers and teachers

(Elicker, Englund, & Sroufe, 1992). In this way, primary caregiving relationships, including

those with important nonparental caregivers, are expected to have an impact on children’s devel-

oping personalities, social competence, and other capacities.

492 ELICKER, WEN, KWON, SPRAGUE

Researchers using an attachment-relationships perspective have focused attention primarily

on the mother–infant relationship (e.g., Garcia Coll & Meyer, 1993; van IJzendoorn & Sagi,

1999). Many studies across several cultures have found consistent and strong associations

between mother–infant attachment security and children’s later social-emotional and cognitive

functioning (e.g., Crittenden & Claussen, 2000; Shonkoff & Phillips, 2000). Therefore, promot-

ing the parent–child attachment relationship has become a focus of many early intervention pro-

grams supporting children from high-risk backgrounds (Emde, Korfmacher, & Kubicek, 2000).

Guided by attachment theory and mother–child research, birth-to-3 practitioners have come to

view their relationships with children as focal aspects of successful early interventions and child

care programs, given the extensive amount of time that they have spent with the children (e.g.,

Ahnert, Pinquart, & Lamb, 2006; Howes, 1999; Lally et al., 2003).

NONPARENTAL CAREGIVER–CHILD RELATIONSHIPS: A GROWING RESEARCH INTEREST

A growing body of research supports a general hypothesis that infants’ and toddlers’ relationships

with nonparental caregivers in early care and education settings affect socioemotional and cognitive

development in ways similar to, though perhaps not as strongly as, relationships with parents. Recent

research in infant and toddler child care has shown that teacher–child interactions and relationship

quality are linked both concurrently and predicatively with children’s social-emotional, language,

and cognitive outcomes (e.g., Hausfather, Toharia, LaRoche, & Engelsmann, 1997; National Insti-

tute of Child Health and Human Development Early Child Care Research Network [NICHD

ECCRN], 2000a, 2000b). Research underscores the key role of teacher sensitivity, responsive-

ness, and positive teacher–child relationships in supporting child development (e.g., Early

et al., 2007; Howes et al., 2008; LoCasale-Crouch et al., 2007; Mashburn, 2008). However, many

studies of teacher–child relationships have used only global assessments that have focused on tea-

chers’ relationships with the whole child care group, such as the Infant=Toddler Environment Rating Scale–Revised (Harms, Cryer, & Clifford, 2003) and the Caregiver Interaction Scale

(CIS; Arnett, 1989) rather than the relational dynamics within specific caregiver–child pairs.

It cannot be determined from studies using classroom-level relationship assessments whether

children’s outcomes are specifically associated with the quality of dyadic teacher–child relation- ships. Among the more focused measures for assessing dyadic relationships are those for

caregiver–child attachment security (e.g., using the Strange Situation, Ainsworth, Blehar,

Waters, & Wall, 1978; or the Attachment Q-Set, Waters, 1995) and composite summaries of

caregiver–child interactions (e.g., the Observational Record of the Caregiving Environment,

NICHD ECCRN, 1996). There is evidence that caregiver–child relationships, when assessed

using these dyadic measures, are associated with aspects of children’s cognitive, language,

and social-emotional functioning and growth. For example, Howes and colleagues found signi-

ficant concurrent associations between attachment security with child care providers and pre-

schoolers’ social competence (Howes, 1997; Howes & Smith, 1995). However, the strongest

evidence to date for the developmental influence of child care relationships comes from the

NICHD Study of Early Care, which found that cumulative positive dyadic caregiving during

the first 3 years was significantly associated with children’s school readiness, expressive

language, and receptive language at 3 years (NICHD ECCRN, 2000b). Based on best practice

EARLY HEAD START RELATIONSHIPS 493

recommendations, theoretical predictions, and the evidence summarized, we hypothesized that

dyadic relationship quality between Early Head Start program caregivers and infants and

toddlers would be concurrently associated with the children’s social and cognitive development.

PROGRAM CAREGIVER–PARENT RELATIONSHIPS: A NEGLECTED ARENA

Compared with the research on caregiver–child relationships discussed previously, even less

research attention has been focused to date on relationships that develop between parents and

professionals (the adult relationships) in the context of early childhood programs. This is despite

the fact that birth-to-3 practitioners often identify relationships with parents as central to the

success of early intervention, as we discovered in preliminary case study interviews with Early

Head Start program caregivers (Elicker, Magaňa, & Sketchley, 2000).

Belsky’s (1984) determinants of parenting model suggests that relationships and social support

provided to mothers by spouses or other adults result in more positive, responsive parenting of the

infant and ultimately in more positive child development outcomes. In the early intervention field,

relationships between program caregivers and parents have gained increasing attention as a factor

promoting healthy parent–child attachment (Hans & Korfmacher, 2002). These programs typi-

cally aim to influence a mother’s daily interactions with her infant and=or her internal working model of attachment through not only providing modeling and information about parenting but

also fostering a supportive, relationship between the mother and the program staff (Emde et al.,

2000). These supportive adult relationships are believed to provide emotional security to the

parent, which promotes more positive ways of thinking and feeling about self, others, and rela-

tionships (e.g., Egeland & Erickson, 2003; Heinicke, Fineman, Ponce, & Guthrie, 2001; Juffer,

Bakermans-Kranenburg, & van IJzendoorn, 2007). Relationship-based interventions have been

successful in promoting positive changes in parental sensitivity, parenting behaviors, and

parent–child attachment relationships (Egeland, Weinfield, Bosquet, & Cheng, 2000).

Belsky’s parenting model and the success of relationship-based, parent-focused interventions

suggest a hypothesis that supportive Early Head Start staff–parent relationships will result in more

positive parenting and thus better developmental outcomes for both parents and children. How-

ever, only a few studies to date have empirically tested the associations between staff–parent

relationships and parent or child program outcomes. Data from the Nurse Home Visiting program

in Memphis demonstrated that mothers’ perceptions of empathy from program staff were signifi-

cantly associated with empathic attitudes the mothers had toward their children (Korfmacher,

Kitzman, & Olds, 1998). An evaluation of the University of California at Los Angeles Family

Development Project, a 2-year intervention for first-time mothers, revealed that mothers’ trust

in and ability to work with the program staff had a significant association with the mothers’ respon-

siveness to their children’s needs (Heinicke et al., 2000). In the current study, we hypothesized that

supportive program caregiver–parent relationships would be positively associated with both child

social and cognitive development and positive parenting in Early Head Start programs.

In summary, effective infant-toddler interventions can be viewed not only in terms of how

well they support children’s development and positive parenting but also in terms of how well

they build supportive staff–child and staff–parent relationships factors (Bernstein, 2002). Thus,

the intervention program is conceived as system of interdependent interpersonal relationships,

including children, parents, and program staff, producing an overall relational climate and

494 ELICKER, WEN, KWON, SPRAGUE

constituting a key intervention process (Bertacchi, 1996; Egeland et al., 2000; Emde et al., 2000;

Heinicke et al., 2000).

FACTORS IMPACTING PROGRAM RELATIONSHIPS

Which child, family, and program caregiver characteristics are associated with relationships

between program caregivers, parents, and children? Not many studies have specifically looked

into this question, which would inform intentional practice targeting specific participant groups.

There is some evidence that program staff characteristics are related to staff members’ ability

to connect with families. For example, caregivers with higher education or more early childhood

training provide higher quality care to young children (e.g., Webster-Stratton, Reid, &

Hammond, 2001; Whitebook, Sakai, & Howes, 2004). It is also hypothesized that ongoing train-

ing and relationship-based program support is crucial to service providers’ ability to engage and

retain families in early childhood intervention program and to build strong partnerships with

families (Jorde-Bloom, 2004; Parlakian, 2001; Wasik & Bryant, 2001). One study explicitly

compared the differences between nurse and paraprofessional home visitors (who did not have

advanced education and training) in terms of their program implementation and participant

outcomes in the Nurse Family Partnership program (Korfmacher, O’Brien, Hiatt, & Olds,

1999). The results showed that nurses and paraprofessionals did not differ in how participating

families rated the quality of their helping relationship. However, families visited by parapro-

fessionals tended to have less contact with the program and to drop out sooner. In the current

study, we hypothesized that program caregivers with higher education levels, more professional

training, and more program-related support would develop higher quality relationships with

Early Head Start children and parents.

Much evidence has shown that family risk (e.g., socioeconomic disadvantage, low education,

minority group status, high stress, and single parenthood) is among the factors that keep parents

from being involved (and building partnerships) and gaining benefits from intervention pro-

grams (e.g., Halpern, 2000; Robinson et al., 2002). Parents’ psychological characteristics

(e.g., depression) have also been related to families’ engagement in early intervention services.

For example, one study demonstrated that in a home visitation program, mothers with insecure

attachment relationship histories were less likely to engage with home visitors (Korfmacher,

Adam, Ogawa, & Egeland, 1997). Parents who had better interpersonal relationship skills more

readily formed collaborative relationships with intervention program staff (e.g., Brookes, Ispa,

Summers, Thornburg, & Lane, 2006). Other studies have shown similar patterns in how parti-

cipants’ psychological resources are associated with the way they use program services and

interact with staff (e.g., Florian, Mikulincer, & Bucholtz, 1995; Spieker, Solchany, McKenna,

DeKlyen, & Barnard, 2000). In the current study, we hypothesized that Early Head Start parents

with higher education levels and lower depressive symptoms would develop higher quality

relationships with both their own children and the Early Head Start program caregivers.

Finally, individual child characteristics might impact relationships with Early Head Start

caregivers. It has been reported that boys tend to demonstrate more negative interactions and rela-

tionships with mothers and teachers and generally have less optimal early school outcomes than

girls (Pianta & Walsh, 1996). There is also evidence that a child’s age may affect caregiver–child

relationship quality. For example, previous studies found that children who entered child care at

EARLY HEAD START RELATIONSHIPS 495

younger ages were more likely to have stable or secure relationships with caregivers than children

who entered child care at preschool age (Elicker, Fortner-Wood, & Noppe, 1999; Howes & Hamil-

ton, 1992). Similarly, the mother–child relationship is also affected by child age. One study

showed that mothers tended to display a lower level of warm responsiveness and a higher level

of restrictiveness as their infant approached toddlerhood (Smith, Landry, & Swank, 2000).

INTERACTIONS BETWEEN FAMILY AND PROGRAM FACTORS

These family and program factors not only may impact program relationships but could interact

with program relationships in predicting program outcomes. One of the challenges in evaluating

interventions with families at risk is the complexity of how family characteristics and program fac-

tors interact to produce program outcomes (Berlin, O’Neal, & Brooks-Gunn, 2003). In some pre-

vious intervention studies, family and program factors were considered as two separate contexts

rather than as interacting to produce program effects (Wen, Korfmacher, Hans, & Henson,

2010). In fact, there is some evidence that family and program factors (e.g., program relationships)

interact in rather complicated ways to predict program outcomes. For example, a prenatal and post-

partum support program with African American teen mothers found that the more program contact

mothers with limited vocabulary skills (a family factor) had with the home visitor (a program fac-

tor), the more likely they were to have positive birth experiences (Wen et al., 2010). However, very

few studies have investigated the interactions between family characteristics (e.g., maternal edu- cation) and program context variables (e.g., program–family relationships). A goal of the current

study was to test for possible moderating effects of family and program caregiver characteristics on

associations between Early Head Start relationships and child and parent program outcomes.

RESEARCH QUESTIONS

The present exploratory study followed Early Head Start families and program caregivers over

a 6-month period to address the following research questions: (a) What is the nature of relation-

ships among Early Head Start program caregivers, children, and families, and how do these

relationships change over time? (b) What are the characteristics of the program, caregivers,

and families that are associated with Early Head Start relationship quality? (c) Are qualities

of these Early Head Start relationships associated with child and parent outcomes? (d) Do

characteristics of program caregivers and families moderate the associations between Early Head

Start relationships and child and parent program outcomes?

METHOD

Participants

A total of 71 parent–child dyads and 32 caregivers from three Early Head Start programs in

midwestern communities participated. Of the sample children, 52% were in the Early Head Start home visitation program, which provided weekly home visits and targeted both parents and

child, and the rest were in an Early Head Start full-time center-based child care program that

496 ELICKER, WEN, KWON, SPRAGUE

mainly worked with the child. The average age of the child participants was 20 months, ranging

from 8 to 35 months. This age range is representative of the Early Head Start population. A total

of 42% of the children were boys. The majority of the children (69%) were White, 14% were African American, and 10% were Hispanic. Children had been enrolled in Early Head Start for an average of 14 months (SD ¼ 7.46) at the time of the study’s initial assessments.

Of the participating parents (mostly mothers), 84% had finished high school or had some college education, 58% were married or living with a partner, 48% had a job at the time the study began (average of 29 work hours per week), and 37% were in school or job training. Parents’ average age was 25 years, ranging from 17 to 40 years. All families were low income and lived

below the federal poverty income level and were therefore eligible for Early Head Start services.

Of the Early Head Start staff participants, 63% had a bachelor’s degree or higher, and 72% had majored in early childhood education or child development or had taken relevant courses. The

majority of the staff (79%) was White. On average, the caregivers had worked with the program for 2.4 years (SD ¼ 1.48, range ¼ 3 months to 6 years). The sample size of caregivers was evenly distributed across the three participating Early Head Start programs (ns ¼ 10, 10, and 13). On average, each caregiver worked with two children in the sample (34% worked with one child, 34% worked with two children, and 32% worked with more than two).

The participants were followed over a 6-month time period. For the follow-up assessments, 49

parent–child dyads (70% of the original sample; 45% were in the home visitation program) remained available. The reasons for attrition involved participants moving out of the program service area,

dropping out the program, or having difficulty scheduling follow-up assessments before the child

graduated from the program. Many Early Head Start families’ living situations were in flux. Lack

of voicemail, changing phone numbers, relocating residence out of the area, and irregular work

schedules were among factors that prevented follow-up visits. The attrition rate found in this sample

is typical among studies that involve a Head Start population (e.g., Love et al., 2002). An attrition

analysis based on chi-square and t tests showed that in comparison with the original sample, the Time 2 sample had a lower proportion of parents without a high school diploma (10% vs. 16%) and a lower proportion of parents living with a partner or spouse (55% vs. 58%). Otherwise, the Time 1 and Time 2 samples were not statistically different in terms of other demographic characteristics.

Procedure

Participants were recruited by solicitation at parent meetings and by distribution of flyers within the

three Early Head Start programs. The eligibility criteria were that the child had to be between 8 and

30 months old and have been enrolled in the program for at least 1 month. The study involved four

assessments, two at the initial time and two at the 6-month follow-up. Of the initial assessments, one

was conducted with the child and parent at home, and the other was conducted with the child and

Early Head Start program caregiver in the classroom or the child’s home (if the child was in the home

visitation program). The initial home assessment that involved the parent and child took about 2 hr

and was conducted during a time when the child was well rested and fed. The child was assessed

using a standardized developmental measure, the Mullen Scales of Early Learning, and a semistruc-

tured parent–child play interaction session was videotaped for 15 min. In addition, the parent was

asked to complete a survey regarding family demographic information, child development, depress-

ive symptoms, and the parent’s current relationship with the child’s primary Early Head Start care-

giver. For the second assessment, which involved the Early Head Start caregiver and the child, their

EARLY HEAD START RELATIONSHIPS 497

interactions were observed for 2 hr in the classroom (if the child was in a center-based program) or

during a home visit (if the child was in the home-based program). In addition, the Early Head Start

caregiver completed a survey describing his or her professional backgrounds, the child’s develop-

ment, and his or her relationship with the parent. The same measures and procedures were repeated

in the parent–child and caregiver–child 6-month follow-up assessments.

Overall, this study adopted a short-term longitudinal design to examine the nature and impact of

Early Head Start relationships. Although the participants had an average of 1 year of program

exposure at the beginning of the study, it was unclear whether their relationships with program care-

givers had stabilized or were still in the process of growing. There is a limited literature to guide

research regarding relationship development, in terms of formation and maintenance. Our working

hypothesis was that relationship building is a continuous process that may fluctuate over time,

especially in the context of interventions with children and families experiencing high-level life stress

and challenges. Considering documented evidence of Early Head Start participants’ overall low level

of program involvement (e.g., Love et al., 2002), we hypothesized that the relationships between pro-

gram caregivers and families would improve over the course of the study for participants who

remained in the program. In addition, a 6-month time period is a reasonable timeframe in which to

measure changes in infants, toddlers, and parents that might be related to their program experiences.

Measures

Relationship Quality

The quality of relationships between Early Head Start caregivers and children was assessed using

three observational measures in the classroom setting or during the home visit, and all of the

measures were appropriate for children aged 0–3. The quality of the relationship between Early

Head Start caregivers and parents was assessed using a parent and caregiver self-report scale.

Program caregiver–child relationships: attachment security. The Safe and Secure Scale (Booth, Kelly, Spieker, & Zuckerman, 2003), a 15-item, 9-point scale derived from the Attach-

ment Q-Set (Waters, 1995), was used to describe the focus child’s secure base behavior toward

the Early Head Start caregivers (example item: ‘‘If child care provider reassures him by saying

‘It won’t hurt you,’ child will approach or play with things that initially made him cautious or

afraid’’: 1 ¼ very unlike this child, 5 ¼ neither like nor unlike this child, 9 ¼ most like this child). The 15 scale items focus on the child’s ability to seek and receive positive attention, feel safe and

protected, receive support for exploration, receive consolation when distressed, and seek and

accept assistance. The scale has established reliability (a ¼ .81) and was positively correlated with proximal measures of child care quality (Booth et al., 2003). The interobserver reliability

for the current study was established at 80% exact agreement.

Program caregiver–child relationships: level of involvement. The child’s interactive involvement with his or her caregiver was rated using a slightly modified version of the Adult–Child

Involvement Rating Scale (Howes & Stewart, 1987). For each 20-s interval during eight equally

spaced, 15-min observation sessions, the highest level of interactive involvement between the child

and caregiver was scored. Scale points range from 0 to 6 (0 ¼ absent, adult not present; 1 ¼ ignores, ignores, the adult ignored the child; 2 ¼ routine, if the caregiver touched the child for changing or other routine caregiving but made no verbal responses to child; 3 ¼ minimal, if the caregiver touched the

498 ELICKER, WEN, KWON, SPRAGUE

child only for necessary discipline, to answer direct requests for help, or to give verbal directives with

no reply encouraged; 4 ¼ social, if the caregiver answered the child’s verbal bids but did not elaborate or used some unnecessary positive physical contact; 5 ¼ elaborative, if the caregiver engaged in some positive physical gestures, acknowledged the child’s statements and responded but did not restate the

child’s statement, sat with the child during play; and 6 ¼ intense, if the caregiver hugged or held the child, restated the child’s statements, engaged the child in conversation, or played interactively with

the child). Summary scores were calculated as the mean level over the 2-hr observation period. The

scale has adequate test–retest reliability and is correlated with other relevant child care variables

(Howes & Smith, 1995). For the current study, an 80% exact agreement was attained between the observers during two consecutive live observations.

Program caregiver–child relationships: positive caregiving. The CIS (Arnett, 1989) was adapted to assess the interaction quality between Early Head Start caregivers and children. The scale

consisted of 26 items (e.g., ‘‘speaks warmly to the child,’’ ‘‘listens attentively when the children

speak to her’’) rated on a 4-point Likert scale (1 ¼ not at all, 2 ¼ somewhat, 3 ¼ quite a bit, and 4 ¼ very much). The measure consists of four subscales: positive interactions (10 items; warm and developmentally appropriate behavior), punitiveness (8 items; hostility, harshness, and use of threat),

detachment (4 items; low involvement and disinterest), and permissiveness (4 items). The permissive-

ness subscale was dropped because it measures caregiving style rather than the nature of caregivers’

interactions with children. The scale has high internal consistency (a ¼ .81; Jaeger & Funk, 2001) and was found to predict child language development and attachment security (Whitebook, Howes, &

Phillips, 1989). For the current study sample, the internal consistency reliability was slightly lower

(a ¼ .63). The mean score of the 22 items in the three subscales of positive interactions, punitiveness (reversed), and detachment (reversed) was computed for analyses. Table 1 presents intercorrelations

TABLE 1

Correlations Among the Early Head Start Relationships

Variable 1 2 3 4 5

Time 1 (N ¼ 71) Caregiver–child relationship

1. Attachment security — .39�� .55�� .08 .03 2. Child–caregiver involvement — .53�� �.08 .02 3. Positive caregiving — .09 .02

Caregiver–parent relationship

4. Parent-rated relationship — .28�

5. Caregiver-rated relationship —

Time 2 (N ¼ 49) Caregiver–child relationship

1. Attachment security — .49�� .74�� .17 .25 2. Child–caregiver involvement — .49�� �.02 .08 3. Positive caregiving — .05 .14

Caregiver–parent relationship

4. Parent-rated relationship — .15

5. Caregiver-rated relationship —

�p < .05. ��p < .01.

EARLY HEAD START RELATIONSHIPS 499

among these measures and suggests that the three caregiver–child relationship measures were signifi-

cantly correlated at both Time 1 and Time 2 (rs ¼ .39–.74).

Program caregiver–parent relationships. The Parent–Caregiver Relationship Scale (Elicker, Noppe, Noppe, & Fortner-Wood, 1997) is a measure with parallel forms that both

parents and Early Head Start program caregivers completed independently to assess their percep-

tions of relationship quality with each other. Items in the scale assess the factors level of

trust=confidence (e.g., ‘‘The caregiver is someone I can rely on’’), collaboration (e.g., ‘‘We talk about problems right away’’), affiliation (e.g., ‘‘I am interested in the caregiver’s personal life’’),

and caring (e.g., ‘‘This parent is a caring person’’). The scale consists of 35 items scored on

a 5-point Likert scale (1 ¼ strongly disagree to 5 ¼ strongly agree). For the current sample, the internal consistency reliability was .97 for the parent report and .96 for the program caregiver

report. Table 1 shows that correlations between parent and staff ratings were modest.

Child and Parent Outcomes

Early Head Start children’s social competence and cognitive abilities were assessed using

a naturalistic observation scale, parent and caregiver reports, and standardized assessments by

trained examiners, and the outcome measures aimed to look at the whole child’s development.

In addition, Early Head Start parents’ interactive parenting behaviors in a videorecorded home

play session were coded.

Child outcome: object play. A child’s level of play with objects, considered an indicator of cognitive development, was rated using the scale developed by Rubenstein and Howes (1979)

during a 2-hr observation in the classroom or during a home visit. This 5-point scale rates the

complexity of the child’s object play from oral contact and passive holding to active mani-

pulation and exploration of the object’s unique properties for creative uses (1 ¼ oral contact, 2 ¼ passive, 3 ¼ active, 4 ¼ exploitative, and 5 ¼ creative). For each 20-s interval during eight equally spaced, 15-min observation sessions, the highest level of play complexity was coded.

The mean level of object play over the 2-hr observation period was calculated. The scale has

adequate test–retest reliability and predicted other child development outcomes (Howes &

Smith, 1995). An interobserver reliability criterion of 80% exact agreement during two con- secutive live observations was met by all observers prior to data collection.

Child outcome: social competence. Both the parents and Early Head Start caregivers reported on the child’s emerging social competence and behavioral problems using the Brief

Infant–Toddler Social and Emotional Assessment (BITSEA; Briggs-Gowan & Carter, 2002).

The BITSEA includes 60 items, and each item is scored on a 3-point scale (0 ¼ not true=rarely, arely, 1 ¼ somewhat true=sometimes, and 2 ¼ very true=often). The scale is composed of two subscales—problem behaviors (49 items, such as hits, bites, or kicks the parent) and social competence behaviors (11 items, such as paying attention for a long time)—and has established reliability (range ¼ .66–.89) and validity (Briggs-Gowan, Carter, Skuban, & Horwitz, 2001). In the current study, a composite score was created by combining social competence and reversed

problem behavior item raw scores into a total score representing overall socioemotional com-

petence (with a higher score representing greater competence). Internal consistency reliabilities

for parent and caregiver reports were .80 and .69, respectively. Because parent and caregiver

500 ELICKER, WEN, KWON, SPRAGUE

reports were not highly correlated at either time point (rs ¼ .48 and .19 at Time 1 and Time 2, respectively), scores were computed separately for parent and caregiver BITSEA social-

emotional adjustment.

Child outcome: cognitive skills. The Mullen Scales of Early Learning, American Guidance Services edition (Mullen, 1995) is an individually administered developmental test

that can be used with children from birth to 68 months. The participating children were admini-

strated the test during both initial and follow-up assessments. The assessment consists of four

cognitive scales (visual reception, receptive language, expressive language, and fine motor)

plus one gross motor scale. Scores on the four cognitive scales are combined to yield an Early

Learning Composite score, which, according to Mullen (1995) is an indicator of general intel-

lectual competence, with a mean of 100 and a standard deviation of 15. The internal consistency

reliabilities for the five subscales were reported to range from .75 to .83, and the interrater

reliabilities ranged from .91 to .99 for age groups between 1 and 44 months. The concurrent

validity of the measure has been supported by its correlations with other early development

measures (e.g., the Bayley Mental Development Index, rs ¼ .53–.59; the Preschool Language Assessment Auditory Comprehension, r ¼ .85; Mullen, 1995). Parent outcome: responsive parenting behaviors. Six responsive parenting behaviors

were coded from the videotaped parent–child play interactions in the home setting using the

Three Bag Coding Scales (NICHD ECCRN, 1999). sensitivity (how the parent observes and

responds to the child’s cues), intrusiveness (the degree to which the parent controls the child,

rather than respecting his or her perspective), stimulation of cognitive development (the parent’s

effortful teaching to enhance perceptual, cognitive, and linguistic development), positive regard

(parent’s expression of love, respect, and admiration for the child), negative regard (parent’s

expression of discontent with, anger toward, disapproval of, and rejection of the child), and

detachment (parent’s lack of awareness of, attention to, and engagement with the child). Each

behavior was rated on a 7-point scale (1 ¼ a very low incidence of the behavior, 7 ¼ a very high incidence of the behavior). The interrater reliability on each scale was established at 90% exact agreement or better. Because all six parenting behaviors were positively intercorrelated at both

data collection points (bivariate rs ranging from .36 to .83), a composite score representing overall responsive parenting behaviors was computed for both Time 1 and Time 2 by summing

the six individual behavior scores (negative parenting behavior scores were reversed). For the

current sample, the internal consistency for the six parenting behaviors was .91.

Participant Characteristics

Participant demographic characteristics and program information were collected through the

parent and caregiver surveys. In addition, parental depressive symptoms were assessed using the

Center for Epidemiological Studies–Depression scale (Radloff, 1977), a 20-item, 4-point scale

(example items: ‘‘I thought my life had been a failure’’, or ‘‘I felt fearful’’: 0 ¼ rarely or none of the time, 3 ¼ all of the time). Higher scores reflect greater depressive symptoms.

Early Head Start program caregivers also completed the Early Childhood Work Environment

Survey (Jorde-Bloom, 1988), designed for early childhood care administrators, teachers, or

support staff for assessing dimensions of organizational climate (e.g., peer cohesion, support

for professional and personal growth, and the presence of facilitative leadership). The survey

EARLY HEAD START RELATIONSHIPS 501

consists of 75 items, with 10 items rated on 5-point Likert scales and the rest rated on dichot-

omous scales (yes vs. no). The scale has adequate reliability and validity and has been used in

several studies assessing early childhood program work environments (e.g., Jorde-Bloom, 1999).

RESULTS

This study explored questions regarding the nature of Early Head Start relationships, how they

change over time, their association with short-term child and parent program outcomes,

how family and program characteristics are associated with relationships, and how these charac-

teristics moderate the associations between Early Head Start relationships and program

outcomes. The following sections are organized to present results answering these questions.

What is the Nature of Early Head Start Relationships, and How Do They Change over Time?

Table 2 summarizes the data on relationship quality among Early Head Start caregivers and part-

icipants at the two assessment points, disaggregated by program type (center-based vs.

home-based programs). Attachment security between children and their program caregivers

was observed at a moderate level of quality (between 5 and 6 on the 9-point scale) at both Time

1 and Time 2 and in both programs, with a relatively large degree of variance. Similarly, mean

levels of children’s interactive involvement with caregivers were low to moderate (between

TABLE 2

Early Head Start Relationships and Outcomes: Descriptive Statistics

Center-Based program Home-Based program

Time 1 (N ¼ 34) Time 2 (N ¼ 27) Time 1 (N ¼ 37) Time 2 (N ¼ 22)

Variable M SD M SD M SD M SD

Caregiver–child relationship

Attachment security a

5.11 0.79 6.03 1.04 5.16 0.87 5.75 1.19

Child–caregiver involvement b

2.52 0.87 3.31 0.71 3.27 0.86 3.59 0.47

Positive caregiving c

3.10 0.27 3.49 0.30 3.28 0.24 3.44 0.36

Caregiver–parent relationship d

Parent-rated relationship 4.22 0.45 4.17 0.72 4.01 0.76 4.00 0.66

Caregiver-rated relationship 4.11 0.57 3.95 0.55 3.94 0.62 3.82 0.65

Program outcomes

Object play 2.40 0.51 2.70 0.50 2.55 0.50 2.71 0.59

Early learning (Mullen) 104.59 15.13 103.52 16.01 100.05 16.92 94.45 17.96

Parent-rated BITSEA 0.99 7.34 4.33 7.57 �1.15 11.23 �2.49 12.63 Caregiver-rated BITSEA 4.25 9.89 2.82 12.38 2.47 12.23 2.29 12.40

Parenting behavior 32.12 4.42 33.52 3.23 29.67 6.05 30.59 6.11

Note. BITSEA ¼ Brief Infant–Toddler Social and Emotional Assessment. a Attachment security was assessed on a 9-point scale.

b Child–caregiver involvement was assessed on 6-point scale.

c Positive caregiving was assessed on 4-point scale. d Caregiver–parent relationship was assessed on a 5-point scale.

502 ELICKER, WEN, KWON, SPRAGUE

‘‘routine’’ and ‘‘minimal’’ at Time 1 and between ‘‘minimal’’ and ‘‘social’’ at Time 2). The

program caregivers displayed overall a high level of positive caregiving toward the group, as

measured by the CIS scale at both assessment points. Overall, the results suggested moderately

positive relationship quality between Early Head Start caregivers and children. As for adult

relationship quality, both program caregivers and parents rated their relationships quite posi-

tively (means were greater than 4 on the 5-point scale for parents and approaching 4 for care-

giver ratings), although program caregivers’ perceptions were slightly less positive than parents’.

A multivariate general linear model test was performed to examine whether the Early Head

Start relationships changed over the observed 6-month time period. The test (Roy’s largest root)

showed that there was an overall positive change in the Early Head Start relationships from the

Time 1 to Time 2 assessments, F(5, 43) ¼ 5.99, p ¼ .00. The univariate statistics indicated that the three relationship measures that assessed Early Head Start caregivers and children had statistically

significant positive changes over time (ps < .01); however, the adult relationship ratings from program caregivers and parents did not change significantly across this 6-month period (p > .24).

What Are the Characteristics of the Program, Caregivers, and Families that Are Associated with Early Head Start Relationship Quality?

The associations between participant characteristics and Early Head Start relationships were

examined using the full sample at the Time 1 assessment, allowing for a more extensive analysis

of these factors. The program and participant characteristics included in this analysis were child

age, gender, and ethnicity; parent education and depression; program caregiver work experience,

education, and perceptions of the work environment; and program service delivery model

(center-based child care vs. home visitation). Pearson correlations were conducted with con-

tinuous family and program characteristic variables (i.e., child age, parent depression, program

caregiver work experience, and perceptions of the work environment). Analysis of variance was

conducted with the discrete variables (i.e., child gender, ethnicity, parent education, program

caregiver education, and program service delivery model).

Analyses of variance showed that girls had more secure relationships than boys with their

center-based caregivers and home visitors (Safe and Secure Scale; Ms ¼ 5.30 and 4.90 for girls and boys, respectively), F(1, 69) ¼ 4.26, p < .05. Children in the home-based program were observed as showing a higher level of interactive involvement with program caregivers than

children in the center-based program (Ms ¼ 3.27 and 2.52, respectively), F(1, 69) ¼ 13.48, p < .01. And caregivers in the home-based program displayed a higher level of positive caregiving as measured by the CIS than those in the center-based program (Ms ¼ 3.28 and 3.10, respectively), F(1, 69) ¼ 8.93, p < .01. Child ethnicity and caregiver education level were not significantly associated with any of the three caregiver–child relationship measures.

As for the caregiver–parent relationships, program caregivers reported more positive relation-

ships with parents who had higher levels of education (Ms ¼ 4.13 and 3.62 for parents who had finished high school and parents who had not, respectively), F(1, 69) ¼ 7.53, p < .01. However, parents reported better relationship quality with program caregivers who had lower education levels (Ms ¼ 4.31 vs. 3.99 for caregivers who had an associate’s degree and ones who had a bachelor’s degree or higher, respectively), F(1, 69) ¼ 4.66, p < .05. Adult relationship quality was not different for the two program service delivery models.

EARLY HEAD START RELATIONSHIPS 503

Table 3 presents correlations among the continuous measures of participant characteristics

and the Early Head Start relationships. Younger children tended to display higher attachment

security and experienced more positive interactions with program caregivers (as measured by

the CIS). Caregivers who had more years of work experience demonstrated higher levels of

interactive involvement with children; those who had more positive perceptions about their Early

Head Start program work environment reported more positive relationships with the parents.

Parental depressive symptoms were not significantly related to any of the relationship measures.

Are Early Head Start Relationships Associated with Child and Parent Outcomes?

Table 2 presents descriptive statistics for the child and parent program outcome measures at

the two assessment points, disaggregated by the program type (center-based vs. home-based

program). The paired samples t tests showed that children’s object play (t ¼ �2.41, p ¼ .02) and parenting behaviors (t ¼ �2.90, p ¼ .006) had significant gains over the 6-month period, whereas the other three child outcomes did not show significant changes.

The associations between Early Head Start relationships and program outcomes were

analyzed through hierarchical linear modeling (HLM; Raudenbush & Bryk, 2002) because of

the nested nature of the data. The sample children were nested within three different centers,

two different program models (center-based vs. home-based programs), and 32 Early Head Start

caregivers. However, multilevel modeling could not be performed at the level of the center or

program type because these higher group-level samples were too small. They were, however,

included as dummy-coded covariates. The study involved 32 program caregivers, and on

average, each caregiver worked with two children in the sample (34% worked with one child, 34% worked with two children, and 32% worked with more than two). Therefore, it was appro- priate to conduct multilevel analysis at the caregiver level. The methodology literature suggested

that even though this Level 2 caregiver sample was relatively small, the analysis would allow for

precise estimations (Maas & Hox, 2005; Snijders & Bosker, 1999). However, because of the

overall small sample in this study and the degree of data attrition over time, the HLM analysis

could be conducted only with the Time 1 full sample (n ¼ 71), not with the Time 2 sample (n ¼ 49). These analysis decisions were guided by general principles of sample size for HLM

TABLE 3

Correlations Between Participant Characteristics and Early Head Start Relationships (N ¼ 71)

Variable Child age

Caregiver work

experience

Caregiver-Rated work

environment Parent depression

Caregiver–child relationship

Attachment security �.32�� .11 �.12 �.22 Child–caregiver involvement �.03 .30� �.06 .10 Positive caregiving �.25� .03 �.07 �.001

Caregiver–parent relationship

Parent-rated relationship �.10 .27 .08 �.15 Caregiver-rated relationship .06 .04 .44�� �.23

�p < .05. ��p < .01.

504 ELICKER, WEN, KWON, SPRAGUE

analysis (Tabachnick & Fidell, 2006). Because of the sample size limitations of the study, it was

necessary to reduce the number of variables included in the analysis. Given the fact that the

Early Head Start caregiver–child relationship measures were significantly correlated (see

Table 1), a composite score (computed by standardizing each measure and then summing the

weighted units) was used to represent overall caregiver–child relationship quality. Similarly,

a composite score for overall caregiver–parent relationship quality was computed by averaging

the total scores of the parents’ and Early Head Start caregivers’ ratings. Although the parent and

caregiver ratings were only modestly correlated, this composite score took into account the

perspectives of both parties. The resulting caregiver–child and caregiver–parent relationship

composite scores were not significantly correlated, r(71) ¼ .04, p > .05. Two-level HLM analyses were performed to predict each child and parent outcome. Analyses

were conducted in two steps. The first step was to examine unconditional models that did

not include any covariates, only the intercept. Next two-level contextual models were estimated

that included the child and family covariates and relationship variables as Level 1 predictors and

caregiver-level covariates as Level 2 predictors. The Level 1 covariates included the caregiver–

child and caregiver–parent relationship variables, child age, child gender, parent education,

parent depression, and program type (center-based child care vs. home visitation; dummy

coded). These predictors were significantly correlated with one or more outcomes in the prelimi-

nary analysis. The Level 2 covariates included the program caregivers’ work experience and

education level. Preliminary analyses suggested weak correlations among all of the covariates.

Only the caregiver–child relationship composite was inversely and modestly correlated with

child age (r ¼ �.25). Therefore, concern about collinearity was allayed. The models were computed in HLM 6.06 and estimated using full maximum likelihood. Both

Level 1 and Level 2 covariates were grand-mean-centered (given the small n per caregiver). The Level 2 intercept was modeled as randomly varying. The contextual model was as follows:

The Level 1 model was:

Yij ¼ b0j þ b �

1j ðPROGRAMijÞ þ b �

2j ðP EDUijÞ þ b �3j ðDEPRESSIONijÞ þ b

� 4j ðCHILDSEXijÞ þ b

� 5j ðCHILDAGEijÞ

þ b �6j ðSC RELATIONijÞ þ b �

7j ðSP RELATIONijÞ þ rij The Level 2 model was:

b0j ¼ c00 þ c �01 ðCAREGIVER EXPERIENCEjÞ þ c �

02 ðCAREGIVER EDUjÞ þ u0j b1j ¼ c10 b2j ¼ c20 b3j ¼ c30 b4j ¼ c40 b5j ¼ c50 b6j ¼ c60 b7j ¼ c70

The two Level 2 covariates were included to predict the intercept from the Level 1 model.

EARLY HEAD START RELATIONSHIPS 505

The unconditional model indicated that the proportion of variance in child=parent outcomes between Early Head Start caregivers (intraclass correlation coefficient).ranged from about 1% to 27% (object play, 27%; early learning, 6%; parent-rated social competence, 2%; caregiver- rated social competence, 1%; and parenting behaviors, 18%). Therefore, a larger proportion of variance resided within children and parents.

Results of the contextual HLM model are presented in Table 4. For the statistically significant

coefficients, effect sizes were calculated as ðB�XSDXÞ=SDY, where SDY represents the variance term for the intercept at the level at which X centers in the model. Effect sizes are, therefore,

interpreted as expected change in standard deviation units in Y that is associated with

a standard deviation change in X. The caregiver–child relationship quality composite was not

related to any of the child or parent outcomes at the initial assessment after child, parent, and

TABLE 4

Association Between Early Head Start Relationships and Program Outcomes (N ¼ 71)

Variable

Object

play

Early

learning

Parent-Rated

BITSEA

Caregiver-Rated

BITSEA

Parenting

behaviors

Fixed effects, coefficient (SE)

Intercept (c00) 2.50 ���

(0.06)

102.12��

(1.97)

�0.06 (1.03)

3.07�

(1.29)

30.58��

(0.76)

Caregiver work experience (c01) �0.02 (0.05)

�1.44 (1.48)

�1.28 (0.79)

0.12

(0.99)

�0.22 (0.54)

Caregiver education: bachelor’s

or higher (c02) a

�0.11 (0.16)

�0.42 (5.05)

�1.56 (2.68)

4.46

(3.35)

0.66

(1.87)

Program type: home-based

program (c10) b

0.05

(0.16)

�0.63 (5.02)

�1.21 (2.69)

�1.07 (3.36)

�1.09 (1.78)

Parent education: high school

or higher (c20) c

0.03

(0.15)

9.39

(5.31)

�0.28 (3.04)

5.34

(3.80)

2.08

(1.39)

Parent depression (c30) 0.003 (0.01)

0.02

(0.23)

�0.11 (0.13)

�0.02 (0.16)

0.11

(0.07)

Child gender: female (c40) d �0.23�

(0.11)

3.36

(3.88)

3.24

(2.20)

5.08

(2.75)

2.36�

(1.07)

Child age (c50) 0.03 ��

(0.01)

�0.09 (0.27)

0.25

(0.15)

0.10

(0.19)

0.23��

(0.08)

Caregiver–child relationship

composite (c60) 0.04

(0.03)

�0.40 (0.88)

0.37

(0.49)

�0.61 (0.61)

�0.55 (0.29)

Caregiver–parent relationship

composite (c70) �0.25� (0.12)

11.97��

(4.09)

7.86��

(2.28)

5.27

(2.85)

4.73��

(1.16)

Random effects (variance

components)

Level 1 variance (r) 0.17 220.29 73.81 115.45 13.88

Intercept (u0), variance [v 2 , p] 0.04

[39.12, .10]

16.95

[33.61, .25]

0.09

[21.16, >.50]

0.14

[21.21, <.05]

10.12

[65.53, <.001]

Note. BITSEA ¼ Brief Infant–Toddler Social and Emotional Assessment. a Reference category is high school, Child Development Associate credential, or associate’s degree.

b Reference category is center-based program.

c Reference category is no high school diploma or general equivalency diploma. d Reference category is male.

�p < .05. ��p < .01. ���p < .001.

506 ELICKER, WEN, KWON, SPRAGUE

program characteristics were controlled. However, caregiver–parent relationship quality was sig-

nificantly associated with three child outcomes—object play (effect size ¼ �0.73, p ¼ .04), early learning composite (effect size ¼ 0.03, p ¼ .01), and parent-rated social competence (BITSEA; effect size ¼ 0.05, p ¼ .006)—although the association with children’s level of object play was in an unexpected direction. Caregiver–parent relationship quality was also significantly

associated with observed positive parenting behaviors (effect size ¼ 0.17, p < .001). Girls displayed lower levels of object play compared with boys (effect size ¼ �0.67, p ¼ .04), but girls received more positive parenting from their parents than did boys (effect size ¼ 0.08, p ¼ .04). Child age was positively associated with level of object play (effect size ¼ 1.32, p ¼ .002) and parenting behaviors (effect size ¼ 0.12, p ¼ .007).

Do Characteristics of Program Caregivers and Families Moderate the Links Between Early Head Start Relationships and Child and Parent Program Outcomes?

Given the limited sample size, only three family and caregiver characteristics of interest were

included as potential moderators: child gender, parent education level, and Early Head Start

caregiver education level. Full data collected at the initial assessment were again used for this

analysis. HLM analyses were run by including one outcome as the dependent variable,

a moderator variable (dummy coded), a relationship composite variable (grand mean centered),

and an interaction term between the moderator and the relationship variable (also grand mean

centered) as the independent variables. Child gender and parent education were included as

Level 1 variables, and Early Head Start caregiver education was included as a Level 2 variable.

If the interaction term is significant, it indicates that the family or program caregiver charac-

teristic variable moderates the association between the relationship quality and the outcome vari-

able. To further explore a significant moderation effect, the association between the relationship

quality variable and the outcome measure was examined within two subsample groups formed

based on the moderators (i.e., with and without high school diploma based on parent education;

girls and boys based on child gender). This analytic strategy for examining and interpreting

moderation effects is based on recommendations by Aiken and West (1991). Graphic plots of

the moderation effects and the simple slope computation (predicting effects of the relationship

variable) for each subgroup were constructed using Mod-Graph software (Jose, 2008).

Two significant moderation effects were found, and they are displayed in Figures 1 and 2,

where the x-axis represents the caregiver–parent relationship quality composite and the y-axis

represents the level of observed positive parenting behaviors (total score). As shown in

Figure 1, the association between caregiver–parent relationship quality and observed responsive

parenting behaviors was moderated by parent education. The interaction between parent edu-

cation and the caregiver–parent relationship was found to be significant (b ¼ �5.56, p ¼ .03; effect size ¼ �0.05). A stronger association between observed positive parenting behaviors and caregiver–parent relationship quality was seen in parents who did not finish high school.

This association was weaker among parents who had a high school diploma or more education.

Another significant interaction was found with Early Head Start caregivers’ education level

and the quality of the parent–caregiver relationships (b ¼ 7.68, p ¼ .006; effect size ¼ 0.57). As indicated in Figure 2, Early Head Start caregivers’ education level moderated the association

between the quality of parent–caregiver relationships and observed positive parenting behavior,

such that families who worked with program caregivers holding a bachelor’s degree or higher

EARLY HEAD START RELATIONSHIPS 507

FIGURE 1 Moderation effect by parent education. Parent education moderates the association between Early Head Start

caregiver–parent relationships and the observed responsive parenting behaviors.

FIGURE 2 Moderation effect by program caregiver education. Early Head Start caregivers’ education moderates the

association between caregiver–parent relationships and the observed responsive parenting behaviors.

508 ELICKER, WEN, KWON, SPRAGUE

displayed more positive parenting when their relationships with the Early Head Start caregiver or

home visitor were more positive. However, this association was not observed among families

working with program caregivers with lower degrees. Child gender did not show any moderating

effects on associations between the Early Head Start relationships and child or parent outcomes.

DISCUSSION

This study of the interpersonal relationships among parents, children, and program caregivers in

three midwestern Early Head Start programs contributes new insights, in terms of both theory

and application. Though the sample was relatively small, including 71 child–parent–caregiver

triads, the study included intensive observations and assessments over a 6-month period. Using

relatively new measures, we described the nature and variation of relationships children and

parents had with Early Head Start program staff. We identified factors that were associated with

more or less positive Early Head Start program relationships, and we found that positive and

supportive relationships, especially those between parents and program caregivers, predicted

positive outcomes for both parents and children.

For the Early Head Start children, we found that caregiver–child relationships varied but were

generally positive in terms of attachment security, interactive involvement, and positive care-

giving and that these relationships became increasingly positive over the 6-month period of

our observations. This suggests that given time, Early Head Start caregivers will develop more

positive and secure relationships with children. Such a continuity of care hypothesis is consistent

with current recommendations and some previous research emphasizing more time and conti-

nuity as key factors for nonparental caregivers to develop supportive relationships with young

children (e.g., Elicker et al., 1999; Howes & Hamilton, 1992; Lally, 2009).

Caregiver–parent relationships were uniformly positive in these Early Head Start programs

based on both parties’ reports. Parents’ perceptions were somewhat more positive, with less

variation, than ratings by program caregivers, a finding consistent with previous research

(Elicker et al., 1997; Green, McAllister, & Tarte, 2004; Korfmacher, Green, Spellmann, &

Thornburg, 2007; Roggman, Boyce, Cook, & Jump, 2001; Wen & Elicker, 2012). Although

Early Head Start caregivers’ evaluations tended to be positive, their relatively lower ratings

could be due to the fact that they hold higher standards for what an optimal relationship with

parents is like. Caregivers’ relationship assessments are made in the context of work with many

parents, therefore their evaluations may be more discerning and more variable across parents,

whereas parents have limited experience in relationships with program caregivers. These are

possibilities that should be explored in future research. Given the generally high ratings, it

was not surprising to find that relationships did not increase in quality over time, probably

reflecting a ceiling effect. Future research may want to use measures of caregiver–parent

relationships that capture a broader range of variation in relationship perceptions.

What characteristics of children, parents, or program caregivers predicted positive relation-

ships? We found that Early Head Start caregivers were more likely to have positive, secure rela-

tionships with girls and with younger children (infants) compared with boys and older children

(toddlers). It may be that boys and toddlers present challenges for caregivers striving to develop

positive relationships with them, as boys may be more active and less verbal, and toddlers are

more likely than infants to attempt to gain independence and autonomy, sometimes reacting with

EARLY HEAD START RELATIONSHIPS 509

resistance (e.g., Smith et al., 2000). Girls have been observed in other research to have more

secure attachment relationships with child care providers than boys (Ahnert et al., 2006). These

findings are also consistent with available evidence about parent–toddler relationships (e.g.,

Schoppe-Sullivan et al., 2006) and suggest a need for increased attention in Early Head Start

staff training to the process of building positive relationships with toddlers and with boys.

Children in the home-based program showed higher levels of interactive involvement with their

Early Head Start caregivers, and home-based caregivers displayed more positive caregiving than the

center-based caregivers. This is probably related to the fact that in home visiting, the program

caregiver focuses on one individual child and the parent intensively, usually for 3 hr one time

per week, and therefore is better able to spend quality time with the child and be attentive to the

child’s needs. In classroom settings, caregivers work with a small group all day every day, typically

four to eight children, including the focal child that we were observing. Therefore, this difference

in observed interactive involvement between the center-based and home-based samples (less

frequent=more intense vs. more frequent=less intense) is probably due to the distinctive context for caregiver–child interaction that each of these program service delivery models presents.

Caregivers who rated their program work environment as more supportive, and those who

had attained a Child Development Associate credential or associate’s degree, tended to have

more positive relationships with parents than those who reported more negative views of their

work environment or those with higher levels of education. These findings support the impor-

tance of a positive and supportive work environment for caregivers in relationship-focused

programs like Early Head Start (Bertacchi, 1996; Jorde-Bloom, 2004). The results also suggest

that there may be special challenges for caregivers with more advanced levels of education to

develop trust, open communication, and feelings of collaboration with intervention families. It

could be beneficial in future research to examine more closely how and whether staff education

level makes a difference in the interactions and relationships that develop between program care-

givers and parents. Better educated, sometimes less experienced caregivers, when working with

parents with lower incomes and education levels, may feel unprepared, threatened, or mistrustful

in their interactions with those parents. However, caregivers with more similar backgrounds to

the parents’ backgrounds might be more comfortable balancing power and trust in their relation-

ships with parents. In fact, similarity in background with families has been a rationale for hiring

community members as paraprofessionals in some early intervention and family support

programs (Behnke & Hans, 2002; Hans & Korfmacher, 2002). We do not think these results

necessarily imply that advanced education needs to be an impediment for program caregivers

in establishing positive relationships with families. In fact, the results also showed that relation-

ship quality with parents was most strongly associated with positive parenting outcomes when

program caregivers had higher education levels. An important task for future inquiry should be

to determine how to provide preservice education, in-service training, and technical support for

professional caregivers to enable them to utilize their knowledge to effectively understand and

support parents and children in Early Head Start and similar programs serving families at risk.

Contrary to our hypothesis, the quality of caregiver–child relationships did not predict current

or future levels of children’s cognitive or social development. However, we did find that the

quality of caregiver–parent relationships was associated with concurrent measures of children’s

object play, social competence, early learning, and positive parenting. There are a number of

plausible explanations for these links. Supportive staff relationships with parents, with a focus

on the child’s development, may be a key part of an effective intervention for positive change

510 ELICKER, WEN, KWON, SPRAGUE

in both parents and children, a hypothesis that has been supported by results of randomized

studies (e.g., Heinicke et al., 2000; Roggman, Boyce, & Cook, 2009). The findings of the current

study contribute further evidence, though we cannot draw causal conclusions, given the correla-

tional design of the study.

The obtained inverse association between the quality of program caregiver–parent relation-

ships and children’s complexity of object play was counter to our expectations. A possibility that

bears further investigation is that children whose parents are more positively engaged with pro-

gram caregivers (and perhaps other adults) are also socially oriented and less likely to engage in

more complex play with objects at this age and in this observation context. Our finding that there

were no significant associations between the quality of caregiver–child relationships and child

outcomes was also unexpected and not consistent with previous research (e.g., NICHD ECCRN,

2000b). It is possible that this particular age (around 20 months) and this study’s short, 6-month

observation period, coupled with the relatively small sample, did not enable us to detect

influences of program caregiver–child relationships that may be modest, compared with the par-

enting influences. Our results did show that caregiver–child relationships improved over time.

However, our longitudinal analysis focused on the effects of initial relationship quality (using

the Time 1 sample), primarily because of sample size limitations. More extended longitudinal

studies with larger samples and repeated measures of both relationships and child outcomes

would enhance understanding of the influences of program caregiver–child relationships. It

may also be that normal developmental changes or other influences (especially family influ-

ences) at this age overshadow the influences of program relationships on children’s development

(e.g., Downer & Pianta, 2006; NICHD ECCRN & Duncan, 2003).

The discovery of family and caregiver characteristics that were moderating influences help

us better understand potential associations between relationships and program outcomes. One

moderation effect suggests that parents with lower education levels are more likely to exhibit

positive parenting in a play session with their child when they have more positive relationships

with their child’s program caregivers, whereas the parenting of those with higher education

levels was less strongly linked to the quality of this adult relationship. A plausible interpretation

is that supportive relationships with program caregivers are especially important in supporting

positive parenting in parents with lower education levels. Relationships with a caring and

supportive professional may provide both emotional security for the parent and a model for

how to support the child. Although this study design was not experimental, this finding is also

consistent with previous research showing that families with higher risk levels benefit most from

early intervention programs (Bradley, Burchinal, & Casey, 2001; Robinson & Emde, 2004).

However, although program caregivers with higher levels of education (bachelor’s degree or

higher) tended to have less positive relationships with Early Head Start parents, we found that when these more educated caregivers did establish positive relationships with parents, the quality of the relationships was significantly associated with observed positive parenting behaviors. This

association between relationship quality and parenting outcome was significantly weaker for

caregivers with lower education levels. One possible conclusion from this pattern of results is

that building positive relationships with parents might be an important and necessary step for

higher educated staff to help families achieve more favorable program outcomes. Overall,

our findings suggest that program caregivers with varying levels of experience or education do

productively engage families and do promote positive program outcomes for children and families,

perhaps in distinctive ways. Also, individualized approaches to in-service training and other

EARLY HEAD START RELATIONSHIPS 511

program supports may be needed for program caregivers with different backgrounds, so that

each one is able to engage families, develop positive relationships, and use those relationships

to support positive changes in children and parents.

As with many intensive observational field studies of early childhood programs with high-

risk clients, limitations stemmed from a smaller sample size and the attrition of research parti-

cipants over time. The use of larger samples and additional strategies to retain study participants

will be beneficial in future research examining program relationships with Early Head Start

families. Finally, although multilevel modeling analyses were conducted to address the issue

of nested data, future research should strive to examine interpersonal relationships within more

homogeneous program contexts or in samples of sufficient size to tease out all of the patterns of

nested groups.

This study breaks new ground with its focus on the assessment and description of Early Head

Start caregiver relationships with both children and parents over time, using multiple measures

and methods to assess relationships and program outcomes. The findings challenge researchers

and practitioners to consider variations in program relationships and the fact that these relation-

ships can be an important aspect of the intervention. The results reinforce the findings in

previous research that it may be challenging to develop supportive relationships with parents

who have lower levels of education. However, positive and supportive relationships with

program caregivers may support more positive outcomes for those who are at risk for negative

parenting or less optimal child development outcomes. Likewise, these results suggest that it

may be more challenging for caregivers to develop relationships with boys and toddlers than

with girls and infants, so these issues may also be important to consider as staff are guided to

build secure, positive relationships with children.

The study of relationships in early intervention programs is a complex matter. The relational

experiences that young children, parents, and professionals have when they come together in

a program like Early Head Start are rich and variable. Relationship constructs such as attachment

security have been tremendously helpful in understanding and supporting early development.

However, researchers will certainly need more differentiated relationship constructs and ways

of assessing them in the future study and improvement of relationship-focused early intervention

programs like Early Head Start.

ACKNOWLEDGMENTS

This research was supported by a Head Start-University Partnership grant (#90YD0111=03) from the U.S. Department of Health and Human Services, Office of Planning, Research, and

Evaluation. The authors are especially grateful to the children, parents, and staff who

participated from the Early Head Start partner programs in Lafayette, Indiana (Bauer Family

Resources), Kokomo, Indiana (Bona Vista Programs), and Marion, Indiana (Carey Services).

REFERENCES

Ahnert, L., Pinquart, M., & Lamb, M. E. (2006). Security of children’s relationships with nonparental care providers:

A meta-analysis. Child Development, 74, 664–679. doi:10.1111/j.1467-8624.2006.00896.x

Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting interactions. Newbury Park, CA: Sage.

512 ELICKER, WEN, KWON, SPRAGUE

Ainsworth, M. D. S., Blehar, M. C., Waters, E., & Wall, S. (1978). Patterns of attachment: A psychological study of the strange situation. Hillsdale, NJ: Erlbaum.

Arnett, J. (1989). Caregivers in day-care centers: Does training matter? Journal of Applied Developmental Psychology,

10, 541–552. doi:10.1016/0193-3973(89)90026-9

Behnke, E. F., & Hans, S. L. (2002). Becoming a doula. Zero to Three, 23, 9–13. Belsky, J. (1984). The determinants of parenting: A process model. Child Development, 55(1), 83–96.

Berlin, L. J., O’Neal, C. R., & Brooks-Gunn, J. (2003). Early childhood intervention research initiatives. In

J. Brooks-Gunn, A. S. Fuligni, & L. J. Berlin (Eds.), Early child development in the 21st century (pp. 65–90). New York, NY: Teachers College Press.

Bernstein, V. (2002). Strengthening families through strengthening relationships: Supporting the parent-child relationship

through home visiting. Newsletter of the Infant Mental Health Promotion Project, 35, 1–5.

Bertacchi, J. (1996). Relationship-based organizations. Zero to Three, 17, 1–7. Booth, C. L., Kelly, J. F., Spieker, S. J., & Zuckerman, T. G. (2003). Toddlers’ attachment security to child-care providers:

The Safe and Secure Scale. Early Education & Development, 14, 83–100. doi:10.1207/s15566935eed1401_6

Bowlby, J. (1973). Separation: Anxiety and anger. New York, NY: Basic Books.

Bradley, R. H., Burchinal, M. R., & Casey, P. H. (2001). Early intervention: The moderating role of the home

environment. Applied Development Science, 5, 2–8. doi:10.1207=S1532480XADS0501 1

Briggs-Gowan, M. J., & Carter, A. S. (2002). Brief Infant–Toddler Social and Emotional Assessment (BITSEA) manual,

version 2.0. New Haven, CT: Yale University. Briggs-Gowan, M. J., Carter, A. S., Skuban, E. M., & Horwitz, S. M. (2001). Prevalence of social-emotional and

behavioral problems in a community sample of 1- and 2-year-old children. Journal of the American Academy of Child

and Adolescent Psychiatry, 40, 811–819. doi:10.1097/00004583-200107000-00016

Brookes, S., Ispa, J. M., Summers, J. A., Thornburg, K. R., & Lane, V. (2006). Building successful home visitor-mother

relationships and reaching program goals: A qualitative look at contributing factors. Early Childhood Research

Quarterly, 21, 25–45. doi:10.1016/j.ecresq.2006.01.005

Cicchetti, D., Toth, S. L., & Lynch, M. (1995). Bowlby’s dream comes full circle: The application of attachment theory

to risk and psychopathology. Advances in Clinical Child Psychology, 17, 1–75. Copple, C., & Bredekamp, S. (2009). Developmentally appropriate practice in early childhood programs. Washington,

DC: National Association for the Education of Young Children.

Crittenden, P. M., & Claussen, A. H. (2000). The organization of attachment relationships: Maturation, culture, and context. New York, NY: Cambridge University Press.

Downer, J. T., & Pianta, R. C. (2006). Academic and cognitive functioning in first grade: Associations with earlier home and

child care predictors and with concurrent home and classroom experiences. School Psychology Review, 35(1), 11–30.

Early, D. M., Maxwell, K. L., Burchinal, M., Bender, R. H., Ebanks, C., Henry, G. T., . . . Zill, N. (2007). Teachers’ education, classroom quality, and young children’s academic skills: Results from seven studies of preschool

programs. Child Development, 78, 558–580. doi:10.1111/j.1467-8624.2007.01014.x

Early Head Start National Resource Center. (2008). What is Early Head Start? Retrieved from http://www.ehsnrc.org/

AboutUs/ehs.htm

Egeland, B., & Erickson, M. F. (2003). Lessons from STEEP: Linking theory, research, and practice for the well-being

of infants and parents. In A. J. Sameroff, S. C. McDonough, & K. L. Rosenblum (Eds.), Treating parent-infant

relationship problems: Strategies for intervention, (pp. 213–242). New York, NY: Guilford Press.

Egeland, B., Weinfield, N. S., Bosquet, M., & Cheng, V. K. (2000). Remembering, repeating, and working through:

Lessons from attachment-based interventions. In J. D. Osofsky, & H. E. Fitzgerald (Eds.), WAIMH handbook of

infant mental health: Vol. 4. Infant mental health in groups at high risk (pp. 35–89). New York, NY: Wiley.

Elicker, J., Englund, M., & Sroufe, A. L. (1992). Predicting peer competence and peer relationships in childhood from

early parent-child relationships. In R. Parke, & G. Ladd (Eds.), Family-peer relations: Modes of linkage, (pp. 77–106).

Hillsdale, NJ: Erlbaum.

Elicker, J., & Fortner-Wood, C. (1995). Research in review: Adult-child relationships in early childhood settings. Young

Children, 51(1), 69–78. Elicker, J., Fortner-Wood, C., & Noppe, I. C. (1999). The context of infant attachment in family child care. Journal of

Applied Developmental Psychology, 20, 319–336.

Elicker, J., Magaňa, L., & Sketchley, K. (2000). Relationship work in Early Head Start programs: How do staff describe

it?. Unpublished technical report, Purdue University, West Lafayette, IN.

EARLY HEAD START RELATIONSHIPS 513

Elicker, J., Noppe, I. C., Noppe, L. D., & Fortner-Wood, C. (1997). The Parent–Caregiver Relationship Scale: Rounding

out the relationship system in infant child care. Early Education & Development, 8, 83–100. doi:10.1207/

s15566935eed0801_7

Emde, R. N., Korfmacher, J., & Kubicek, L. F. (2000). Toward a theory of early relationship-based intervention. In J. D.

Osofsky, & H. E. Fitzgerald (Eds.), WAIMH handbook of infant mental health: Vol. 2. Early intervention, evaluation, and assessment, (pp. 3–32). New York, NY: Wiley.

Florian, V., Mikulincer, M., & Bucholtz, I. (1995). Effects of adult attachment style on the perception and search for

social support. Journal of Psychology, 129, 665–676. Garcia Coll, C. T., & Meyer, E. C. (1993). The sociocultural context of infant development. In C. H. Zeanah (Ed.),

Handbook of infant mental health (1st ed., pp. 56–72). New York, NY: Guilford Press.

Green, B. L., McAllister, C. A., & Tarte, J. (2004). The strengths-based practices inventory: A measure of

strengths-based practices for social service programs. Families in Society, 85, 326–335. Halpern, R. (2000). Early childhood intervention for low-income children and families. In J. P. Shonkoff, & S. J. Meisels

(Eds.), Handbook of early childhood intervention (2nd ed., pp. 361–386). New York, NY: Cambridge University

Press.

Hans, S., & Korfmacher, J. (2002). The professional development of paraprofessionals. Zero To Three, 23(2), 4–8. Harms, T., Cryer, D., & Clifford, R. M. (2003). Infant=Toddler Environment Rating Scale: Revised edition. New York,

NY: Teachers College Press.

Hausfather, A., Toharia, A., LaRoche, C., & Engelsmann, F. (1997). Effects of age of entry, day-care quality, and family

characteristics on preschool behavior. Journal of Child Psychology and Psychiatry, 38, 441–448. doi:10.1111/

j.1469-7610.1997.tb01529.x

Heinicke, C. M., Fineman, N. R., Ponce, V. A., & Guthrie, D. (2001). Relation-based intervention with at-risk mothers:

Outcome in the second year of life. Infant Mental Health Journal, 22, 431–462. doi:10.1002/imhj.1010 Heinicke, C. M., Goorsky, M., Moscov, S., Dudley, K., Gordon, J., Schneider, C., & Guthrie, D. (2000).

Relationship-based intervention with at-risk mothers: Factors affecting variations in outcome. Infant Mental Health

Journal, 21(3), 133–155. doi:10.1002/1097-0355(200007) 21:3<133::AID-IMHJ1>3.0.CO;2-P

Howes, C. (1997). Teacher sensitivity, children’s attachment, and play with peers. Early Education & Development, 8, 41–49. doi:10.1207/s15566935eed0801_4

Howes, C. (1999). Attachment relationships in the context of multiple caregivers. In J. Cassidy, & P. R. Shaver (Eds.),

Handbook of attachment: Theory, research, and clinical applications (pp. 671–687). New York, NY: Guilford Press. Howes, C., Burchinal, M., Pianta, R., Bryant, D., Early, D., Clifford, R., & Barbarin, O. (2008). Ready to learn?

Children’s pre-academic achievement in pre-kindergarten programs. Early Childhood Research Quarterly, 23,

27–50. doi:10.1016=j.ecresq.2007.05.002

Howes, C., & Hamilton, C. E. (1992). Children’s relationships with child care teachers: Stability and concordance with

parental attachments. Child Development, 63, 867–878. doi:10.1111/j.1467-8624.1992.tb01667.x

Howes, C., & Smith, E. (1995). Relations among child care quality, teacher behavior, children’s play activities, emotional

security, and cognitive activity in child care. Early Childhood Research Quarterly, 10, 381–404. doi:10.1016/0885-

2006(95)90013-6

Howes, C., & Stewart, P. (1987). Child’s play with adults, toys, and peers: An examination of family and child-care

influences. Developmental Psychology, 23, 423–430.

Jaeger, E., & Funk, S. (2001). The Philadelphia child care quality study: An examination of quality in selected early

education and care settings. Philadelphia: United Way of Southeastern Pennsylvania. Jorde-Bloom, P. (1988). Factors influencing overall job satisfaction and organizational commitment in early childhood

work environments. Journal of Research in Childhood Education, 3(2), 107–122.

Jorde-Bloom, P. (1999). Using climate assessment to improve the quality of work life in early childhood programs.

Advances in Early Education and Day Care, 10, 115–146. doi:10.1016/S0270-4021(01)80006-9

Jorde-Bloom, P. (2004). Teacher job satisfaction: A framework for analysis. Early Childhood Research Quarterly, 1(2),

167–183.

Jose, P. E. (2008). ModGraph-I: A programme to compute cell means for the graphical display of moderational analyses: The Internet version, Version 2.0. Retrieved from http://www.victoria.ac.nz/psyc/staff/paul-jose-files/

modgraph/modgraph.php

Juffer, F., Bakermans-Kranenburg, M. J., & van IJzendoorn, M. H. (2007). Promoting positive parenting: An attachment-

based intervention. New York, NY: Erlbaum.

514 ELICKER, WEN, KWON, SPRAGUE

Klass, C. S. (2003). The home visitor’s guidebook (2nd ed.). Baltimore, MD. Korfmacher, J., Adam, E., Ogawa, J., & Egeland, B. (1997). Adult attachment: Implications for the therapeutic process

in a home visitation intervention. Applied Developmental Science, 1, 43–52. doi:10.1207/s1532480xads0101_5

Korfmacher, J., Green, B., Spellmann, M., & Thornburg, K. R. (2007). The helping relationship and program

participation in early childhood home visiting. Infant Mental Health Journal, 28, 459–480. doi:10.1002/imhj. 20148

Korfmacher, J., Kitzman, H., & Olds, D. (1998). Interaction processes as predictors of program outcomes in a preventive

home visitation program. Journal of Community Psychology, 26(1), 49–64. doi:10.1177/105381510702900205 Korfmacher, J., O’Brien, R., Hiatt, S., & Olds, D. (1999). Differences in program implementation between nurses and

paraprofessionals in prenatal and infancy home visitation: A randomized trial. American Journal of Public Health,

89, 1847–1851.

Lally, J. R. (2009). The science and psychology of infant-toddler care: How an understanding of early learning has

transformed child care. Zero to Three, 30(2), 47–53.

Lally, J. R., Griffin, A., Fenichel, E., Segal, M., Szanton, E., & Weissbourd, B. (2003). Caring for infants in groups:

Developmentally appropriate practice. Washington, DC: Zero to Three.

Lally, R. J., & Keith, H. (1997). Early Head Start: The first two years. Zero to Three, 18(2), 3–8. LoCasale-Crouch, J., Konold, T., Pianta, R., Howes, C., Burchinal, M., Bryant, D., . . . Barbarin, O. (2007). Observed

classroom quality profiles in state-funded pre-kindergarten programs and associations with teacher, program, and

classroom characteristics. Early Childhood Research Quarterly, 22(1), 3–17. doi:10.1016/j.ecresq.2006.05.001 Love, J. M., Kisker, E. F., Ross, C. M., Schochet, P. Z., Brooks-Gunn, J., Paulsell, D., . . . Smith, C. B. (2002). Making

a difference in the lives of infants and toddlers and their families: The impacts of Early Head Start. Washington, DC:

U.S. Department of Health and Human Services.

Lyons-Ruth, K., & Zeanah, C. H. (1993). The family context of infant mental health: Infant development in the primary

caregiving relationship. In C. H. Zeanah (Ed.), Handbook of infant mental health (1st ed. pp. 14–37). New York, NY:

Guilford Press.

Maas, C. J., & Hox, J. J. (2005). Sufficient sample sizes for multilevel modeling. Methodology, 1(3), 86–92.

Mashburn, A. J. (2008). Quality of social and physical environments in preschools and children’s development of

academic, language, and literacy skills. Applied Developmental Science, 3, 732–749. doi:10.1080/10888690802199392

Mullen, E. M. (1995). Mullen scales of early learning. Circle Pines, MN: American Guidance Services.

National Institute of Child Health and Human Development Early Child Care Research Network. (1996). Characteristics

of infant child care: Factors contributing to positive caregiving. Early Childhood Research Quarterly, 11, 269–306.

doi:10.1016/S0885-2006(96)90009-5

National Institute of Child Health and Human Development Early Child Care Research Network. (2000a). Characteristics

and quality of child care for toddlers and preschoolers. Applied Developmental Science, 4(3), 116–135. doi:10.1207/ S1532480XADS0403_2

National Institute of Child Health and Human Development Early Child Care Research Network. (2000b). The relation of

child care to cognitive and language development. Child Development, 71, 960–980.

National Institute of Child Health and Human Development Early Child Care Research Network, & Duncan, G. J.

(2003). Modeling the impacts of child care quality on children’s preschool cognitive development. Child Develop-

ment, 74, 1454–1475. doi:10.1111/1467-8624.00617

NICHD Early Child Care Research Network. (1999). Childcare and mother-child interaction in the first three years of

life. Developmental Psychology, 35, 1399–1413. Parlakian, R. (2001). Look, listen, and learn: Reflective supervision and relationship-based work. Washington, DC:

Zero to Three.

Pianta, R. C., & Walsh, D. (1996). High-risk children in the schools: Constructing sustaining relationships. New York, NY: Routledge.

Radloff, L. S. (1977). The CES-D scale: A self-reported depression scale for research in the general population. Journal

of Applied Psychological Measurement, 1, 385–410. doi:10.1177/014662167700100306

Raikes, H. H., & Edwards, C. P. (2009). Extending the dance in infant and toddler caregiving: Enhancing attachment and relationships. Baltimore, MD: Brookes.

Raudenbush, S. W., & Bryk, A. S. (2002). Hierarchical linear models (2nd ed.). Thousand Oaks, CA.

Robinson, J. L., & Emde, R. N. (2004). Mental health moderates of Early Head Start on parenting and child development:

Maternal depression and relationship attitudes. Parenting: Science and Practice, 4(1), 73–94. doi:10.1037/a0015709

EARLY HEAD START RELATIONSHIPS 515

Robinson, J. L., Korfmacher, J., Green, S., Song, N., Soden, R., & Emde, R. N. (2002). Predicting program use and

acceptance by parents enrolled in Early Head Start. NHSA Dialog, 5(2), 311–324.

Roggman, L. A., Boyce, L. K., & Cook, G. A. (2009). Keeping kids on track: Impacts of a parenting-focused Early Head

Start program on attachment security and cognitive development. Early Education & Development, 20, 920–941.

Roggman, L. A., Boyce, L. K., Cook, G. A., & Jump, V. K. (2001). Inside home visits: A collaborative look at process

and quality. Early Childhood Research Quarterly, 16, 53–71. doi:10.1016/S0885-2006(01)00085-0

Rubenstein, J., & Howes, C. (1979). Caregiving and infant behavior in day care and homes. Developmental Psychology,

15, 1–24. Schoppe-Sullivan, S. J., Diener, M. L., Mangelsdorf, S. C., Brown, G. L., McHale, J. L., & Frosch, C. A. (2006).

Attachment and sensitivity in family context: The roles of parent and infant gender. Infant & Child Development,

15, 367–385. doi:10.1002/icd.449

Shonkoff, J. P., & Phillips, D. A. (Eds.). (2000). From neurons to neighborhoods: The science of early childhood development. Washington, DC: National Academies Press.

Smith, K. E., Landry, S. H., & Swank, P. R. (2000). The influence of early patterns of positive parenting on children’s

preschool outcomes. Early Education & Development, 11, 147–169.

Snijders, T. A., & Bosker, R. J. (1999). Multilevel analysis: An introduction to basic and advanced multilevel modeling. Thousand Oaks, CA: Sage.

Spieker, S. J., Solchany, J., McKenna, M., DeKlyen, M., & Barnard, K. B. (2000). The story of mothers who are difficult

to engage. In J. D. Osofsky, & H. I. Fitzgerald (Eds.), WAIMH handbook of infant mental health. Vol. 3: Parenting and child care (pp. 172–209). New York, NY: Wiley.

Sroufe, A. L. (1983). Infant-caregiver attachment and patterns of adaptation and competence in preschool: The roots

of maladaption and competence. In M. Perlmutter (Ed.), Minnesota Symposium in Child Psychology: Vol. 16.

Development and policy concerning children with special needs (pp. 41–81). Hillsdale, NJ: Erlbaum. Tabachnick, B. G., & Fidell, L. S. (2006). Using multivariate statistics. (5th ed.). Boston, MA.

Turnbull, A. P., Turbiville, V., & Turnbull, H. R. (2000). Evolution of family-professional partnership: Collective

empowerment as the model for the early twenty-first century. In J. P. Shonkoff, & S. J. Meisels (Eds.), Handbook

of early childhood intervention (2nd ed. pp. 630–650). New York, NY: Cambridge University Press. U.S. Department of Health and Human Services. (2002). Head Start program performance standards and other

regulations (45 CFR Parts 1301, 1302, 1303, 1304, and guidance, 1305, 1306, 1308, and guidance). Washington,

DC: U.S. Government Printing Office.

van IJzendoorn, M. H., & Sagi, A. (1999). Cross-cultural patterns of attachment: Universal and contextual dimensions.

In J. Cassidy, & P. R. Shaver (Eds.), Handbook of attachment: Theory, research, and clinical applications, (pp.

713–734). New York, NY: Guilford Press.

Wasik, B. H., & Bryant, D. M. (2001). Home visiting: Procedures for helping families (2nd ed.). Thousand Oaks, CA. Waters, E. (1995). The Attachment Q-Set (version 3.0). Monographs of the Society for Research in Child Development,

60(2–3, Serial No. 244), 234–246.

Webster-Stratton, C., Reid, M. J., & Hammond, M. (2001). Preventing conduct problems, promoting social competence:

A parent and teacher training partnership in Head Start. Journal of Clinical Child Psychology, 30(3), 283–302. Wen, X. (2005). Parent involvement in Early Head Start: Dimensions and associations with parenting outcomes.

(Unpublished doctoral dissertation). Purdue University, West Layfayette.

Wen, X., Korfmacher, J., Hans, S. L., & Henson, L. (2010). Young mothers’ involvement in a prenatal and postpartum

support program. Journal of Community Psychology, 38(2), 172–190. doi:10.1002/jcop.20358 Whitebook, M., Howes, C., & Phillips, D. (1989). Who cares? Child care teachers and the quality of care in America:

Final report of the national child care staffing study. Oakland, CA: Child Care Employee Project.

Whitebook, M., Sakai, L., & Howes, C. (2004). By a thread: How child care centers hold on to teachers, how teachers build lasting careers. Kalamazoo, MI: W. E. Upjohn Institute.

516 ELICKER, WEN, KWON, SPRAGUE

Copyright of Early Education & Development is the property of Taylor & Francis Ltd and its content may not

be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written

permission. However, users may print, download, or email articles for individual use.