Participation in EHS improves outcomes related to nurturing and responsive child-parent relationships, with some positive impacts identified in other domains as well. The current evidence base draws primarily from the Early Head Start Research and Evaluation Project. Future research is needed that draws from more diverse EHS samples, on the pathways within the EHS program that lead to positive impacts, and on how state action can support EHS, leading to improved child and family outcomes.
The research discussed here meets our standards of evidence for being methodologically strong and allowing for causal inference, unless otherwise noted. Each strong causal study reviewed has been assigned a letter, and a complete list of causal studies can be found at the end of this review, along with more details about our standards of evidence and review method. The findings from each strong causal study reviewed align with one of our eight policy goals from Table 1. The Evidence of Effectiveness table below displays the findings associated with EHS (beneficial, null,iii or detrimental) for each of the strong studies (A through V) in the causal studies reference list, as well as our conclusions about the overall impact on each studied policy goal. The assessment of the overall impact for each studied policy goal weighs the timing of publication and relative strength of each study, as well as the size and direction of all measured indicators.
Due to the complexity of summarizing the many subgroup findings for EHS, Table 2 includes only those studies that assessed outcomes at child age 2; findings for other age groups are discussed in the text following the table. Patterns of impacts are generally consistent across the prenatal-to-3 period, although impacts may vary by indicator at different child ages. Some strong causal studies are excluded from Table 2 if they did not assess outcomes at child age 2. For studies that assessed outcomes at multiple ages, Table 2 reports impacts for age 2 only. Studies excluded from this table for any reason are indicated by an asterisk in the list of strong causal studies.
Several studies examine the impact on an indicator for multiple subgroups, and results may differ by subgroup, making it difficult to categorize the indicator impact as clearly beneficial, null, or detrimental. In these cases, a study is categorized as beneficial if at least one subgroup outcome was positive, regardless of null impacts for other subgroups. A study is categorized as null if all impacts for all subgroups are null. If impacts were detrimental for at least one subgroup, the study is categorized as both detrimental and either beneficial or null and is marked with an asterisk. Indicators that are measured similarly in different studies have been grouped together in Table 2; additional details on indicator definitions can be found in the text following the table. Several studies of EHS use the same data set and examine the same indicators. For ease of presentation, only one publication (study I) is included in Table 2 as representative of these findings.iv
Table 2: Evidence of Effectiveness for Early Head Start by Policy Goal
|Policy Goal||Indicator||Beneficial Impacts||Null Impacts||Detrimental Impacts||Overall Impact on Goal|
|Access to Needed Services||Safety Net Program Participation||Ir*, d||Io, p||Ir*||Trending^ Mixed|
|Parents’ Ability to Work||Ever Employed||Io, p, r, d||Trending Mixed|
|Average Hours Per Week in Employment||Io, p, r, d*||Id*|
|Participation in Education/Job Training||Io, p, r, d|
|Average Hours Per Week in Education/Job Training||Io, p, r, d|
|Sufficient Household Resources||Percentage of Families with Incomes Above the Federal Poverty Line||Io, p, r, d||Trending Null|
|Perceived Family Resources||Io|
|Parental Health and Emotional Wellbeing||Parenting Distress/Stress||Io, p, r||Id||Trending Mixed|
|Maternal Depression||Io, p, r, d|
|Health Status||Io, p, r, d|
|Nurturing and Responsive Child-Parent Relationships||Home Learning Environment and Behaviors||Io, p, r, d||Mostly Positive|
|Family Routines||Io||Ip, r, d|
|Parent-Child Interactions||Io, p, r, d, V|
|Child Development and Child Rearing Knowledge||Io, p, r, d|
|Emotional Responsivity||Id||Io, p, r|
|Maternal Verbal-Social Skills||Io, p, r, d|
|Discipline and Physical Punishment||Io, p, r, d||Q|
|Family Conflict||Io, d||Ip, r|
|Absence of Punitive Interactions||Id||Io, p, r|
|Parent-Child Dysfunctional Interaction||Ir||Io, p, d|
|Child Abuse Risk Factors||M|
|Nurturing and Responsive Child Care in Safe Settings||Participation in Good-Quality Child Care||K||Trending Positive|
|Optimal Child Health and Development||Aggressive Behavior||Io, r||Ip, d,Q||Mixed|
|Behavior During Play||Ip, r, d*||Io,V||Id*|
|Emotional Regulation||Io, p, r, d,M|
|Orientation/Engagement||Io, p, r, d|
|Language and Vocabulary Skills||Io, p, r, d, R*, V|
|Cognitive Development/Developmental Functioning||Io, r, d||Ip,P|
|Parent-Reported Child Health Status||Io|
* For safety net program participation, beneficial impacts were found at child age 2 for Hispanic EHS families and detrimental impacts were found for Black EHS families, relative to their respective control group counterparts. For average hours per week in employment, null impacts were found for low- and medium-risk subgroups at child age 2, but were detrimental for the high-risk subgroup. For behavior during play, a small, detrimental impact was found on engagement during parent-child play for the low-risk EHS group, but a small, beneficial impact was found for the medium-risk EHS group at child age 2, relative to their control group counterparts. Additional details follow in the relevant policy goal discussion.
^Trending indicates that the evidence is from fewer than two strong causal studies or multiple studies that include only one location, author, or data set.
Note on Table 2: In addition to overall impacts, study I assesses subgroup impacts for three categories of subgroups (program model approach, race and ethnicity, and level of demographic risk). The following notations are used to differentiate impacts for different groups. Io indicates overall study population effects. Ip indicates effects by program model approach subgroups (center-based, home-based, and mixed approach EHS programs). Ir indicates effects by race and ethnicity subgroups (White children, Black children, and Hispanic children and families). Id indicates effects by level of demographic risk subgroups, based on five risk factors (low, medium, and high). Study R assesses differences by child gender.
The majority of research on the impacts of EHS is derived from the Early Head Start Research and Evaluation Project (EHSRE), a large-scale, randomized controlled trial (RCT) of EHS conducted early in EHS program history when the program was first being implemented.v The EHSRE study was conducted at 17 sites nationally, which each offered either center-based, home-based, or mixed-approach EHS programs. If programs offered mixed-approach EHS, families at that site were served by either home-based or center-based services, but individual families did not receive both types of services simultaneously. Families in the EHSRE study were recruited to participating sites from 1996 to 1998, randomized to either the control group or the EHS treatment group, and enrolled either prenatally or before the child was 12 months old. Depending on when the family enrolled, children and families received different amounts of program services at each assessment period (by child age).34,I
Results from the EHSRE study demonstrate that EHS participation had positive impacts on the type and amount of services families received: “the estimated program impacts on the receipt of key program services and core child development services (home visits and center-based child care) were significant, large, and broad-based during the combined follow up period” (p. 188).J This finding suggests that EHS worked as intended, as families participating in the program received more of the core and comprehensive services EHS strives to provide. As described below, EHS has demonstrated positive impacts on child, parent, and family outcomes; however, these favorable impacts emerge in the context of many more null findings in each of these areas. Due to the large number of indicators tested in the main EHSRE impact studies and publications, beneficial or detrimental results are discussed in detail and null results are discussed more broadly below; however, it is important to keep the context of null findings in mind when interpreting overall findings. Studies examining sustained impacts of EHS over time also found mixed results. One study examining the impact of formal early care and education experiences from birth to age 5 found that children who participated in EHS and subsequently participated in formal early care and education programs between ages 3 and 5 had positive outcomes in “social-emotional, vocabulary, parenting, and home environment” domains at pre-K entry (p. 127).D However, across policy goals, the grade 5 follow-up study found only null impacts on child, parent, and family outcome measures; subgroup analyses from this study are generally excluded from the discussion that follows.T
Subgroup analyses of EHSRE study data also allow for the examination of how EHS participation impacts different groups of children, such as those categorized by race/ethnicity, by demographic risk factors,vi and by program approach. These analyses control for a large number of potential confounding factors but should still be interpreted with a degree of caution because participants were not randomized based on child and family characteristics or to specific program approaches. In the EHSRE study, programs used three approaches to serve children and families: center-based programs, which served children directly through center-based early care and learning services; home-based programs, which served families through home visits and focused on improving parenting skills and parent-child relationships; and mixed-approach programs, which served some families through the center-based model and other families through the home-based model.vii,I Study authors recommend that differences in impacts by program approach between the EHS and control groups be interpreted as “the effectiveness of program approach for programs that adopted that approach, given their community contexts and eligible populations” (p. 95).E Although the generalizability of the subgroup analyses is limited, understanding how EHS affects diverse groups of children and families is critical for understanding program effectiveness, and these analyses suggest that program impacts associated with EHS participation do vary by child and family characteristics, as well as program model.
Access to Needed Services
Limited evidence exists on the impact of EHS on access to needed services beyond the receipt of core and comprehensive services provided as a part of EHS program models. At child ages 2 and 3, the EHSRE study found null impacts overall for indicators of safety net program participation, including ever receiving welfare,viii ever receiving cash assistance (Aid to Families with Dependent Children (AFDC) or TANF), ever receiving food stamps, and the total benefit amount received for each.I,J
Access to Needed Services: Subgroup Findings by Race, Ethnicity, Demographic Characteristics, and Program Approach
Largely consistent with overall impacts on access to needed services, analyses of ESHRE study subgroup impacts at child ages 2 and 3 found mostly null results for subgroups by race, ethnicity, demographic risk level, and program approach. Statistically significant findings for indicators of access to needed services were limited and mixed.I,J At child age 2, relative to their respective control group counterparts not enrolled in EHS, a greater share (15.3 percentage points higher) of high demographic risk families participating in EHS reported ever receiving AFDC/TANF benefits, but a smaller percentage (9.3 percentage points lower) of Black families participating in EHS reported ever receiving food stamps.I At child age 3, compared to their respective control group counterparts, Hispanic families participating in EHS were 12.3 percentage points more likely to have reported receiving AFDC/TANF benefits, but Black families participating in EHS reported lower AFDC/TANF benefits (by $775) and total welfare benefits (by $1,522).J
Among demographic risk groups, moderate demographic risk families participating in EHS reported more food stamp benefits (by $676), and a greater percentage of high demographic risk families participating in EHS reported ever receiving AFDC/TANF (by 10 percentage points). A higher share of families that report receiving benefits and higher total benefit amounts are considered beneficial in this case, because EHSRE study authors noted that programs may have sought to improve families’ self-sufficiency in the short-term by assisting families in accessing benefits for which they were eligible while working on other self-sufficiency goals for the longer term (e.g., education or employment).
Parents’ Ability to Work
Evidence of the impact of EHS on parents’ ability to work is also derived from the EHSRE study. Among the overall study population, no statistically significant impacts were found on employment indicators for parents’ ability to work at child ages 2, 3, or 5,I,J,S or employment or education indicators at the grade 5 follow-up.T Among outcomes related to education and training at child ages 2 and 3, a greater percentage of EHS parents reported being in school or in job training programs as compared to the control groupI,J (effect sizesix were 0.09 and 0.16,S respectivelyx). Parents participating in EHS also reported 1.1 and 1.2 average hours more per week in education or training at child ages 2 and 3, respectively, as compared to their control group counterparts.I,J
Parents’ Ability to Work: Subgroup Findings by Race, Ethnicity, Demographic Characteristics, and Program Approach
In line with overall analyses, the impacts of EHS on employment indicators were null for most subgroups and at most child ages (2, 3 and 5).E,I,J,N,S At child age 2, one detrimental impact was found: higher-risk families participating in EHS reported 3.4 fewer average hours per week in employment than their similarly at-risk control group counterparts.I At child age 3, Black parents and parents in mixed approach EHS programs were more likely to report ever being employed (effect sizes 0.23N and 0.16,E respectively).J
Similar to findings for the overall study population in the EHSRE study, positive impacts on education and training were more common. The following subgroups participating in EHS were more likely to report ever being in education or training programs relative to their respective control group counterparts not enrolled in EHS:I,J White families at child age 3 and Hispanic families at child ages 2 and 3 (effect sizes range from 0.19 to 0.28),N moderate-risk families at child ages 2 and 3 (effect size at age 3 0.21),N,xi parents participating in home-based and mixed-approach EHS programs at child ages 2 and 3 (effect sizes at age 3 range from 0.14 to 0.25).E,xii At child ages 2 and 3, Black and Hispanic families participating in EHS and parents participating in home-based EHS programs reported a higher number of average hours per week in education and training relative to their respective control group counterparts (impacts ranged from 1.3 to 2.4 average hours per week).I,J One study that examined the effects of EHS participation by the level of program implementation of the HSPPS found no impacts of full implementation of these standards on parent participation in employment or education/job training at program end or kindergarten entry. This study found that in incompletely implemented programs, EHS parents in home-based programs at program end were more likely to be in education or job training programs compared to parents in the control group (effect size 0.43).G
Sufficient Household Resources
The EHSRE study measured two indicators of sufficient household resources: the share of families with incomes above the federal poverty line and perceived family resources (overall population only) and found no statistically significant impacts at any child age or for any subgroups.I,J,T,xiii
Parental Health and Emotional Wellbeing
Evidence on the impact of EHS on parent health and emotional wellbeing comes from the analyses of the EHSRE study data. The main impact evaluation reportsI,J identified only a few positive impacts: at child age 2, EHS parents had lower distress associated with parenting, and at child age 5, parents reported a lower number of depressive symptoms compared to their control group counterparts, but effect sizes were small (-0.11 and -0.10, respectively).S No impacts were found on parent health status at child age 2 or 3.I,J
Beyond the main EHSRE study reports, a study using EHSRE data examining how the impacts of EHS affect later maternal depression found that the EHS impacts on child outcomes (aggression, developmental functioning) and family outcomes (parenting distress, spanking the child) at ages 2 and 3 explained the connection between EHS participation and maternal depression at age 5.C A study of one EHSRE site implementing an infant mental health (IMH)-based EHS program examined the effects of the IMH-based EHS program two to four years after program completion and found small to moderate impacts on indicators of parent emotional wellbeing.L Three positive impacts on measures of coping were found: EHS parents had higher levels of empowerment on two aspects of the Psychological Empowerment Scale (attitudes and skills/knowledge, effect sizes 0.59 and 0.46 averaged across ages 5 and 7, respectively), and higher levels of perceived mastery (age 7, effect size 0.41) relative to the control group. However, EHS families were less likely to seek support from neighbors relative to their control group counterparts (effect size 0.35), a detrimental finding. With regard to healthy family relationships, the EHS program group had higher scores on healthy functioning and lower scores on unhealthy functioning (effect sizes 0.51 and 0.46, respectively) than control group families. The study found no impact on parenting stress.
Parental Health and Emotional Wellbeing: Subgroup Findings by Race, Ethnicity, Demographic Characteristics, and Program Approach
Analyses from the EHSRE study data suggest similar patterns of mixed parent health and emotional wellbeing impacts for subgroups. No significant impacts were found for any subgroups at child ages 2 and 3 for parent health status.I,J Among subgroups of families by race and ethnicity, at child age 2, Black EHS parents had lower distress associated with parenting,I and at child age 5, Black EHS families reported a lower number of depressive symptoms and had a lower likelihood of reporting someone in the household with alcohol or drug problems, compared to their Black control group counterparts (effect sizes ranged from 0.21 to 0.30).N At the grade 5 follow-up, Black families formerly enrolled in EHS reported less use of alcohol compared to Black families in the control group (effect size -0.34).T Among demographic risk subgroups, parents in moderate-risk families participating in EHS were less likely to experience distress associated with parenting at child ages 2N,xiv and 3J and reported fewer depressive symptoms at child age 5.N Effect sizes for these outcomes ranged from -0.20 to -0.27.N By program approach, significant impacts were found for families in mixed-approach EHS programs at age 2I and home-based EHS programs at age 3J (effect sizes were -0.23 and -0.13,E respectively).
Nurturing and Responsive Child-Parent Relationships
Evidence of the impact of EHS on nurturing and responsive child-parent relationships is derived primarily from the EHSRE study and data, with the greatest number of positive effects seen at child age 2.I,J,S Despite mostly positive findings, a number of null results were found across ages; the EHSRE study examined more than a dozen different indicators within this policy goal at various ages and only statistically significant results are discussed in detail below. No statistically significant impacts on indicators of child-parent relationships at the grade 5 follow-up were found.T
In the EHSRE study, positive impacts of EHS participation were seen for a number of outcomes in the areas of home environments, parent behaviors and interactions with children, and family routines. Positive impacts included: more supportiveness of the home environment for language and literacy (ages 2 and 3), a greater percentage of parents reading daily to the child (ages 2, 3, and 5) and at bedtime (ages 2 and 3), more parent supportiveness during play (ages 2 and 3), less parent detachment during play (age 2), a higher percentage of parents reporting a regular bedtime (age 2), more teaching activities (age 2 and 5), and parents attending meetings or open houses (age 5).I,J,S,xv Effect sizes for significant findings were small, ranging from 0.09 to 0.19.S At child age 2, EHS participation was also linked to slightly higher scores (by 0.1 points) of parents on the Knowledge of Infant Development Inventory.I Null impacts were found for emotional responsivity, warmth, and maternal verbal-social skills at child ages 2 and 3.I,J
Compared to the control group, a lower percentage of parents with children in EHS reported spanking their children (ages 2 and 3,J effect sizes -0.11 and -0.13, respectively).S A higher percentage of parents with children in EHS suggested positive responses (i.e., talk and explain, prevent or distract) to hypothetical discipline strategies at age 2I and a lower percentage of suggested negative responses (i.e., physical punishment, threaten or command) in discipline situations at age 3 (differences ranged from 3.5 to 6.1 percentage points) compared to control group parents, both beneficial outcomes.J At child age 2, parents with children in EHS reported slightly less family conflict (effect size -0.09)S than families in the control group, but this effect was not sustained at age 3.I,J Null impacts were found for the absence of punitive interactions, parent-child dysfunctional interactions, and a number of positive safety practices.I,J
Several studies beyond the main EHSRE impact reports also examined indicators of child-parent relationships. A study of two EHSRE sites examining how EHS may impact mothers with mental health risks found that the impacts of EHS on maternal hostility, maternal sensitivity, and child interaction with a parent during play were concentrated among mothers who were depressed and mothers who were both depressed and reported insecure relationship attitudes.O One study of a small sample at a single EHSRE site found that EHS participation buffered the negative impact of child abuse risk factors on positive parenting regardxvi and children’s emotional regulation.M An RCT that began in 2010 of Educare, described as a hybrid between EHS and a model demonstration program (with enhanced requirements beyond the HSPPS), found more parent-positive parent-child interactionsxvii among families enrolled in the program relative to the control group (effect size 0.42), but null impacts on negative parent interactions.V
Nurturing and Responsive Child-Parent Relationships: Subgroup Findings by Race, Ethnicity, Demographic Characteristics, and Program Approach
Subgroup child-parent relationship impacts of EHS participation by race and ethnicity are generally mixed (positive and null), with the majority of positive impacts by group seen among Black families.I,J,N Compared to Black parents in the control group, Black parents participating in EHS had more supportive home environments for language and literacy (child ages 2xviii and 3), a greater share of parents reporting reading at bedtime (age 2), and lower reported parent-child dysfunctional interaction (age 2). Results also suggested less severe discipline strategies (ages 2 and 3), more supportiveness during play (ages 3 and 5), a higher likelihood to have a regular bedtime (age 3), more warmth (age 3), and more children’s books in the home (age 5), relative to Black control group parents (effect sizes range from 0.19 to 0.40N).I,J At the grade 5 follow-up, the only significant impact on child-parent relationships by race/ethnicity were found among Black families: Black families formerly enrolled in EHS were more involved in school (effect size 0.37).T
Fewer statistically significant impacts were seen among White and Hispanic families at any child age. Among White families participating in EHS, a lower percentage of parents reported spanking their child in the previous week and less family conflict was reported at child age 2 relative to White control group families (effect sizes ranging from -0.18 to -0.22 N).I White parents participating in EHS also reported less intrusiveness during play and suggested less severe discipline strategies, relative to their control group counterparts at child age 2.I Two detrimental impacts were identified among White families who participated in EHS: at child age 3, higher parent-child dysfunctional interaction was reported,J and children in these families were more likely to have witnessed violence at age 5 than their control group counterparts (effect size 0.21).N Among Hispanic families, a greater share of EHS parents read daily to their children at ages 2, 3, and 5 (effect sizes ranging from 0.23 to 0.27),N reported higher scores on the Knowledge of Infant Development Inventory, and reported reading at bedtime at child age 2 relative to Hispanic control group parents.I,J
By level of demographic risk, both low- and high-risk families saw few positive impacts on child-parent relationships as a result of EHS participation; evidence of the effectiveness of EHS was generally found among the moderate-risk group. Positive parenting and home environment outcomes identified among moderate-risk EHS families relative to their control group counterparts included: a supportive home environment for language and literacy (age 2 and 5), reading daily (age 2 and 3), parent detachment during play (age 3), a regular bedtime (age 2), and teaching activities (ages 2 and 3).I,J Effect sizes for these outcomes ranged from 0.18 to 0.36.N The main EHSRE impact evaluation reports examined a greater number of indicators of child-parent relationships and also found positive impacts for parents’ knowledge of infant development (age 2), reading frequency (age 2), emotional responsivity (age 2), parent supportiveness during play (ages 2 and 3), and reading at bedtime (age 3).I,J
By program approach, most positive outcomes in the EHSRE studies were seen for mixed-approach programs, and these effects were concentrated at child ages 2 and 3. For example, families participating in mixed-approach EHS programs saw a range of positive impacts, including more supportive environments for language and literacy (age 2), more teaching activities (age 2), higher reading frequencies (age 2), more parents reading daily (ages 2 and 3), a smaller share of parents reporting spanking (ages 2 and 3), more parent supportiveness (ages 2 and 3) and less detachment during play (age 3), more parents reading at bedtime (age 3), more parents suggesting positive discipline strategies (age 3), and more parents reporting attending meetings or open houses (age 5).E,I,J Effect sizes were still small, ranging from 0.16 to 0.28.E Mixed approach programs may have had the greatest impact because multiple program options allowed families to opt into the EHS approach they preferred (i.e., either home-based or center-based EHS).
In contrast, center-based programs demonstrated very few positive impacts across the range of child-parent relationship indicators at different child ages.E,I,J Similarly, few statistically significant impacts were found for home-based EHS programs, despite a theory of change that suggests home-based programs should demonstrate positive impacts specifically in child-parent relationships, given the focus on parent education and training. Positive impacts for parents in home-based EHS programs relative to the control group were seen for the following indicators: supportive home environment for language and literature (ages 2 and 5), parent detachment (age 2) and parent supportiveness (age 3) during play, the share of parents who read at bedtime (age 2) and daily (age 5), teaching activities age 5), and having at least 26 children’s books in the home (age 5).E,I,J Consistent with other findings, effect sizes were relatively small, ranging from 0.10 to 0.16.E,xix
Three studies outside of the main EHSRE publications have also examined home-based EHS program impacts; however, all studies used data from ESHRE sites. One study of home-based EHSRE study sites examining the impacts of full or incomplete implementation of HSPPS on indicators of child-parent relationships found that parents in EHS in fully implemented programs were less likely to report spanking their children at program end, provided more supportive environments for language and literacy, engaged in more teaching activities, and were more likely to read daily at pre-K entry compared to their control group counterparts (effect sizes ranged from 0.20 to 0.26).G A study of one EHSRE site found participation in EHS was linked to higher attachment security when children were 18 months old.P Another study of the same EHSRE site found that EHS had no impact on physical punishment (spanking) at 36-months,Q which is inconsistent with the overall impact evaluation findings; this finding may be due to the assessment of spanking in only one EHS home-based site sample, rather than the full sample.
Nurturing and Responsive Child Care in Safe Settings
Only one EHSRE study included in this review assessed the impact of EHS participation on the quality of care children received among children participating in center-based or mixed-approach EHS programs. The study found that EHS participation increased the “percentage of children who were in good-quality center care” at all ages (p. 80).K At child ages 14 and 24 months, the share of children participating in EHS in good-qualityxx center-based care (versus low-quality or no center-based care) was approximately three times the control group, and at child age 36 months, EHS children were still more likely to be in good-quality care, although differences from the control group were smaller at this age (12.0 percentage points). Positive impacts were seen at both center-based and mixed-approach EHS programs, but the effects on these measures of quality were larger at center-based sites.
The study also found that EHS participation had large impacts on child-caregiver interactions: at child ages 24 and 36 months, a greater share of children participating in EHS at center-based sites experienced high levels of caregiver talk in center care compared to children in the control group; positive impacts were also seen for children in care at a center in mixed-approach sites, but the effects were not as consistent.xxi
The quality center-based child care experiences among EHS participants were associated with positive child outcomes; however, these impacts cannot be interpreted as causal due to limitations of study design. Furthermore, some caution is needed when generalizing overall impacts of EHS participation on the quality of care children received; this study only examined quality levels in center-based providers among a subsample of the EHSRE study population and did not examine quality in home-based providers or for other aspects of the EHS program.
Optimal Child Health and Development
The EHSRE study examined more than 20 different indicators of child wellbeing at a number of ages in the overall study sample. The EHSRE study found small, positive impacts of EHS participation across a range of social-emotional, cognitive, language, and health outcomes at ages 2, 3, and 5 demonstrating evidence of effectiveness; however, null results were also found across a range of outcomes at different ages.I,J,S No sustained positive child impacts overall were found at the grade 5 follow-up of the EHSRE study.T Beyond the main EHSRE study findings, several other studies have used RCTs (including those that focus on specific EHSRE study sites) or quasi-experimental designs to assess the impact of EHS on child outcomes. The results of these studies and the EHSRE study findings are discussed below.
Among the overall study sample in the EHSRE, at child age 2, EHS program participants demonstrated less aggressive behavior relative to the control group.I At age 3, EHS children demonstrated less aggressive behavior, xxii lower negativity toward their parents during play, higher engagement during play, and greater sustained attention with objects during play, relative to the control group.J At age 5, EHS children also demonstrated fewer social and behavioral problems and more positive approaches to learning.S Effect sizes for all statistically significant outcomes were small (ranging from 0.10 to 0.18); findings for all other outcomes were null.S,xxiiiAlthough results for individual measures were not always sustained over time, these findings suggest positive social-emotional outcomes associated with EHS participation during the program eligibility period, as well as shortly after the end of the program.
Two studies using data from individual EHSRE sites found null impacts on child social-emotional outcomes, including aggressive behaviorQ and child emotion regulation.M The RCT of Educare found that participation in Educare resulted in fewer parent-reported problem behaviors (effect size -0.28), but null impacts were found on child behavior during play and social-emotional competence when children were an average of 2 years old at one year after randomization into the program.V One study of center-based EHSRE sites also found evidence that “participating in center-based EHS services may help mitigate the impact of [family] conflict on children’s aggressive behavior” (p. 952), suggesting that EHS may help buffer negative impacts of family conflict.U These mixed results are consistent with the findings of the EHSRE study.
Cognitive and Language Development
At child age 2, the EHSRE study found that EHS children had larger vocabularies and higher developmental functioning relative to the control group.xxiv At age 3, positive impacts for developmental functioning were sustained, and positive impacts were found on a measure of English language receptive vocabulary. At age 5, the only statistically significant finding was that EHS children had greater Spanish language receptive vocabulary than their control group counterparts. Effect sizes for cognitive and language outcomes were again generally small (ranging from 0.11 to 0.26).S A separate study of one EHSRE site found mixed impacts of participation in EHS on cognitive development: impacts were null at child age 2 but positive at child age 3 (a small impact of a 0.19 point increase).P Moderately-sized positive impacts were also found in the Educare evaluation: children participating in Educare had higher expressive and receptive English, but not Spanish, language skills relative to the control group (effect sizes 0.36 and 0.58, respectively).V
Many of the indicators of child health examined in the EHSRE study represent the receipt of services that EHS programs promote by design, including visits to doctors and dentists for routine and needed care, receipt of screenings, and general receipt of health services. Because these indicators focus on services families should receive by nature of their participation in EHS, they are not a focus in this section; however, null impacts of EHS participation were found for many of these indicators at child ages 2 and 3.I,J
Among health outcomes assessed in the EHSRE study, EHS children were more likely to have received any immunizations at age 2 (effect size 0.09)S and were less likely to ever have been hospitalized for an accident or injury in their third year (1.3 percentage points less).J No other health outcomes were significant at ages 2, 3, or 5 (e.g., emergency room visits, parent-reported child health status).I,J,S Two studies of EHS in North Carolina found positive impacts of EHS participation on oral health, including an increased likelihood of receiving dental care (overall dental care use and preventative care, with a combined adjusted odds ratio of approximately 2.5)A and lower odds of reporting negative oral health-related quality-of-life impacts, although the effect size of the latter outcome was small (odds ratio 0.7).B
Few studies examine child safety and welfare impacts as a result of EHS participation. One study using a subsample of the EHSRE study population found that EHS participation reduced the likelihood of a child welfare encounter between ages 5 and 9 (adjusted odd ratio 0.64) but not at other child ages.F EHS participation also had a positive impact on the length of time between first and second child welfare encounters: “children in the control group were 2.7 times more likely [to experience] a second child welfare encounter earlier than children in the EHS program group” (p. 131).F
Optimal Child Health and Development Subgroup Analyses: Race, Ethnicity, Demographic Characteristics, and Program Approach
By race and ethnicity, the greatest number of positive impacts were seen among Black children in both social-emotional and cognitive domains.I,J,N During the program at child age 2 and at child age 3, Black children participating in EHS had less reported aggressive behavior than their control group counterparts. Black children participating in EHS also had less negativity toward their parent during play and higher engagement and sustained attention with objects during play at age 3 relative to Black children in the control group; at age 5, former EHS children had more positive approaches to learning and greater levels of attention while performing difficult tasks. Black children in EHS also had a greater vocabulary and higher developmental functioning at age 2, relative to their counterparts in the control group; at ages 3 and 5, these children also had larger receptive vocabularies; and at age 5, Black children who had participated in EHS were reported to have fewer speech problems than their control group counterparts. Effect sizes for impacts among Black children ranged from 0.19 to 0.41.N At the grade 5 follow-up, the only significant child impacts by race/ethnicity were found among Black children: Black children enrolled in EHS had fewer reported externalizing behavior and attention problems compared to Black children in the control group (effect sizes -0.26 and -0.22, respectively).T A study examining the pathways through which EHS impacted Black children found that, in addition to direct impacts on child outcomes at age 3, EHS indirectly affected sustained attention, engagement with the parent, and negativity toward the parent through parent supportiveness during play.H
Very few statistically significant outcomes were found for White or Hispanic children at any age. At age 2, White EHS participant children had less negativity to their parent during play, relative to their White control group counterparts,I,N and at age 5, White former EHS participants had more speech problems than their counterparts (effect sizes ranging from 0.20 to 0.27).N The identification of more speech problems may be a detrimental impact, but the interpretation depends on whether groups were screened similarly for these types of problems. If groups were not screened at similar rates or proportions, identifying more speech problems may be due to more screenings among a subgroup, rather than the identification of a detrimental impact. Unfortunately, rates of screening cannot be identified with the information reported in this study. Significant impacts were also limited among Hispanic children: at age 3, Hispanic former EHS children had better orientation/engagementJ and at age 5, Hispanic former EHS children had better emotion regulation, greater Spanish receptive vocabulary, and fewer speech problems than their Hispanic counterparts in the control group (effect sizes ranging from 0.25 to 0.34).N
Differential impacts of EHS participation were also found among low-, moderate-, and high-demographic risk groups. Few significant impacts were found among children in the low-riskxxv and high-risk groups at any age; EHS had the greatest impact among the moderate-risk group.I,J,NAmong the moderate-risk group, EHS participants had better language outcomes at age 2 and cognitive outcomes at ages 2 and 3,I, J relative to their control group counterparts (effect sizes ranging from 0.25 to 0.35);N few significant effects were seen in any child outcome domains at ages 3 and 5.J,N Beyond demographic risk, one study using both national and site-specific EHSRE study data found that EHS participation affected child vocabulary through different pathways for boys and girls in the context of parenting stress: EHS participation reduced the impact of parenting stress on girls’ vocabulary scores (among girls whose mothers had moderate or high parenting stress, effect sizes of 0.19 and 0.42, respectively) and was protective for boys’ vocabulary growth from the effects of parenting stress.R
By program approach, a small number of child social-emotional and cognitive outcomes were significant for EHS children in mixed-approach programs, relative to the control group at these sites, but impacts were generally null for home- or center-based EHS programs at any age. I,J,E In mixed-approach EHS programs, positive vocabulary impacts were found at ages 2 and 3 and among social-emotional indicators: children in EHS had lower aggressive behavior at age 2xxvi and fewer social behavioral problems at age 5, greater engagement of the parent during play at ages 2 and 3, and sustained attention with objects during play at age 3. Effect sizes were generally small, ranging from 0.16 to 0.29.E A study of home-based EHSRE study sites examining the impacts of full or incomplete implementation of HSPPS on child outcomes found mixed evidence of differential impacts by implementation status at program end and pre-K entry.G Compared to children in the control groups, children in fully-implemented home-based EHS had higher cognitive development; were less likely to visit the emergency room at program end; and had higher math skills, positive approaches to learning, and engagement during play at pre-K entry (effect sizes range from 0.19 to 0.33).xxvii
- An impact is considered statistically significant if p<0.05.
- Studies, E, N, and S also report findings from the Early Head Start Research and Evaluation Project. Study S reports overall impacts and studies E and N report subgroup impacts. All of these are covered in study I, so only study I is included in the table to avoid double counting results.
- For additional information on the Early Head Start Research and Evaluation Project, see: https://www.acf.hhs.gov/opre/research/project/early-head-start-research-and-evaluation-project-ehsre-1996-2010; results reported in this summary from the EHSRE study reports and related publications are statistically significant at the 0.05 level and do not include results significant at the trend level (p < 0.1). Throughout this document, the main publications of this project are referred to as “the EHSRE study.”
- Race and ethnicity subgroups included Black, non-Hispanic; White, non-Hispanic, and Hispanic. Demographic risk in this study was measured by five factors: teenage parent status, single parent status, parent not employed or in school, parent receiving cash assistance, and parent has not completed high school (measured at program entry). In studies E, J, N, and S, children/families are considered low risk if they had 0 to 2 risk factors, moderate risk if they had 3 risk factors, and high risk if they had 4 or 5 risk factors. In study I, risk groupings are slightly different: low risk is defined as 0 to 1 risk factors, and moderate risk is defined as 2 to 3 risk factors. This creates some discrepancies between the findings in studies I and N, but results are reported according to low-, moderate-, and high-risk groups, as classified in the studies.
- The evaluation study took place during the early stages of national EHS implementation, therefore program approaches varied over time. Program approach is assessed as implemented in the fall of 1997. For additional details on changing program approaches over time, see Study I. Allowable EHS program approaches now vary from those in use at the time of the original EHSRE study design and program implementation; this is why program approach categories discussed here do not perfectly align with those discussed in the overview of EHS.
- Welfare receipt includes AFDC/TANF cash assistance, food stamps, general assistance, and SSI/SSA benefits.
- Effect sizes are reported from intent-to-treat analyses. Where positive and beneficial negative effect sizes are reported in a range, the absolute value of effect sizes is reported. When possible, effect sizes are reported from studies E, N, and S as these effect sizes are reported using Cohen’s d effect size measure. When no effect sizes are available from these studies, impacts are described using measures in studies I and J.
- For effect sizes reported as Cohen’s d, an effect size of 0.2 is considered small, 0.5 medium, and 0.8 large.
- Study I and N have conflicting findings for medium-risk families participating in education and training at age 2. Although study I finds a positive impact, the result reported in N is null.
- Studies E, I, and J have conflicting findings. At age 2, home-based impacts are only marginally statistically significant (therefore null by the 0.05 threshold) in E. At age 3, mixed-approach impacts are null in study J, but statistically significant in study E.
- Studies E, N, and S examine income using an annual income measure. The results are null at all child ages for all subgroups. Study G did find one positive subgroup impact: in fully-implemented EHS home-based programs, parents reported higher monthly income at kindergarten entry than control group parents (effect size 0.20).
- Studies I and N have conflicting findings. Study I reports a null result.
- Parent supportiveness during play was only statistically significant at the 0.05 level in study I (not study S). Parent detachment during play was only statistically significant at the 0.05 level in study S, not in study I.
- In this study, child abuse risk factors were assessed by a measure of mothers’ potential child abuse, defined as a score above a cutoff threshold (166) on the Abuse Scale of the Child Abuse Potential Inventory. Positive parenting regard is an element of parenting quality, defined by “expressions of love/praise, respect for and enjoyment of child” as coded from video interactions during the “Three Bag task” between the parent and child.
- Parent-positive parent-child interactions were measured using parent emotional supportiveness, parent stimulation of cognitive development, and parent detachment (reversed) scales from the Two Bags Task.
- The significant, positive impact for HOME support for language and literacy for Black families was identified in study N, but not study I (in the latter, it was marginally insignificant). The positive impact for White families for reading daily was found in study I, but not N; the impact for family conflict was found in study N, but not study I.
- Positive impacts for home supportiveness for language and literature and parent detachment during play (at child age 2) were reported only in study E, not study I.
- Good quality was defined by the study authors as a score equal to or greater than 5.0 on the ITERS or ECERS-R.
- The study authors defined good quality care as experiencing high levels of caregiver talk, assessed as scoring in the top quartile on the Child-Caregiver Observation System (C-COS). Three measures were assessed: incidents of any caregiver talk (at least 34), incidents of caregiver responding to child (at least 11), and incidents of caregiver initiating talk with child (at least 28). Differences were significant at all ages for all measures for children in center-based EHS sites, and impacts ranged from 5 percentage points to 24 percentage points. In mixed-approach sites, differences were statistically significant and beneficial for caregiver talk and caregiver initiating talk at 24 months, caregiver responses to the child at both ages. Beneficial impacts ranged from 4 to 17 percentage points. Detrimental impacts were found at 36 months for any caregiver talk and caregiver initiating talk (a smaller share of EHS children experienced high levels as compared to control group children).
- Aggressive behavior was statistically significant in study J, but not in study S (only marginally significant at the 0.1 level).
- Social-emotional wellbeing outcomes in the EHSRE studies were measured using the Child Behavior Checklist (Aggressive), Bayley Behavior Rating Scale, FACES Social Behavior Problems, and other measures discussed in text.
- Developmental functioning was measured by the Bayley Mental Development Index. Vocabulary outcomes were assessed using the MacArthur Communicative Development Inventories (CDI) at age 2 and the Peabody Picture Vocabulary Test at ages 3 and 5.
- At child age 2, a small, detrimental outcome was found for engagement during parent-child structured play for low-risk EHS families, relative to their control group counterparts.
- The positive impact at age 2 was found only in study E, but not in study I (where the impact was marginally significant at the 0.10 level).
- In incompletely implemented programs, children in EHS had higher engagement during play as compared to control group children (effect size 0.34). All other impacts were null at program end and pre-K for this group.