ACF OPRE: National Survey of Child and Adolescent Well-Being (NSCAW)

6. Current Caregiver Characteristics, the Living Environment, and Caregiver Functioning

This chapter presents information about the current caregivers for children whose families have been investigated for maltreatment. Current caregivers are primarily mothers, although they also include some fathers, relatives, and—for about one-ninth of the children in the child welfare population—out-of-home caregivers. (As discussed in Chapter 3, in-home caregivers are primarily biological parents but also include relatives, other-than-kinship foster care providers, adoptive parents, and step parents.) Caregiver characteristics, including demographic information, employment status, income, and poverty level, are compared across CWS settings.

A brief section is presented on out-of-home caregivers including group home caregiver characteristics and foster parent years of experience. Group home caregivers are described separately from other out-of-home caregivers because of their unique status as individuals employed to care for children living in out-of-home care. Group home caregivers are also the caregivers who are caring for many of the older and more troubled children.

Caregiver characteristics, and their expression through their relationships with children, have implications for the safety, permanency, and well-being of children. Many caregivers of CWS-involved children are living at or below the poverty level and also are young, single, and undereducated—factors associated with poverty (McLoyd, Ceballo, & Mangelsdorf, 1997). Living in poverty is associated with developmental disadvantages to children. Stressful life events, often related to financial problems, are associated with ineffective discipline and conduct problems in children and youths (Liaw & Brooks-Gunn, 1994; Stern, Smith, & Jang, 1999). Poverty is also associated with lower academic achievement, although children living in poor families with more supportive home environments (e.g., maternal warmth/acceptance, organization of the home, and opportunities for learning and cultural experiences) show less diminishment of academic achievement (Dubow & Ippolito, 1994).

Other aspects of children's living environment also are likely to be strongly related to their safety, permanency, and well-being. This chapter presents findings about the home environments of all children in NSCAW—whether in in-home or out-of-home care. Several sections in this chapter describe caregiver and household characteristics that have been associated with child maltreatment, including the number of adults and children in the household and the relationship of the child to adults in the household.

NSCAW is the first national study to collect detailed information about the home environment, using the HOME-SF, for children living in foster care and children with a report of maltreatment who remain in the home. Developmental scholars have raised serious questions about the capacity of both the biological and foster homes of children who have become involved with CWS to provide developmentally appropriate and nurturing services (Beckwith et al., 1999; Greenwalt, Sclare, & Portes, 1998; Orme and Buehler, 2001).

Although public policy implicitly assumes that out-of-home environments are safer and more supportive of children's development, little direct measurement of these environments (and the characteristics of their caregivers) is available. A review of the characteristics of foster parents and the living environments in which children are raised (Orme & Buehler, 2001) revealed the limits of available research, indicating that “it is startling how little is known how foster parents and families vary on … important dimensions” (p. 15).

6.1 Caregiver Demographic Characteristics

Before describing details of caregivers' functioning and the living environment, we first view caregivers more generally. The demographic characteristics of in-home caregivers, relative caregivers, foster caregivers, and group home caregivers are presented. The extent to which caregivers' race/ethnicity matches the race/ethnicity of the child is described for the out-of-home caregivers. Data from the U.S. Census and National Survey of America's Families (NSAF) are used to learn how caregivers involved with CWS compare with those in other national surveys of the general population.

6.1.1 Caregiver Age

Caregivers of children who become involved with CWS vary widely in age. Much of this variation is associated with whether the caregiver is the in-home caregiver (usually the mother) or an out-of-home caregiver. The average age of the caregiver is approximately 34 years, with over half of all caregivers between the ages of 25 and 44 years. There is a significant association between service setting and caregiver age. Out-of-home caregivers are far older. Over half of the out-of-home caregivers are 45 years or older, compared with less than 10% of in-home caregivers. These older out-of-home caregivers are predominantly foster and kinship caregivers. About two-thirds of group home caregivers and permanent caregivers are 34 years and younger (Table 6-1).

6.1.2 Caregiver Gender

The majority of caregivers are female (90%), although this is somewhat less true among older caregivers. Significantly fewer caregivers aged 24 and younger are male (3.6%) compared with every other age group (25 to 34 years: 10.9%; 35 to 44 years: 11.4%; 45 to 54 years: 12.0%; 55 years and older: 16.3%). Caregiver gender does not vary significantly by service setting or caregiver race/ethnicity (Table 6-1).

6.1.3 Caregiver Race/Ethnicity

The majority of caregivers are either White, non-Hispanic (51%) or African American, non-Hispanic (26%). Almost one-fifth of caregivers are Hispanic (17%), and 7% of caregivers are of other races/ethnicities. The race/ethnicity of caregivers is similar across CWS settings (Table 6-1). For out-of-home placements, correspondence between the race and ethnicity of foster caregivers and children in their care has long been of interest to CWS policymakers and program administrators. Because of this, and because the race/ethnicity of children remaining in the home almost always matches the race/ethnicity of their caregiver, in this section we focus more on foster caregivers.23

Table 6-1. Current Caregiver Demographics by Service Setting
Characteristics		TOTAL	Setting
			In-Home			Out-of-Home
			No Services	Services	TOTAL In-Home	Foster Care	Kinship Foster Care	Group Care	TOTAL Out-of-Home^
			Percent/ (SE)
Race/ Ethnicity	African American	25.5 (2.7)	23.5 (2.9)	28.4 (3.4)	24.8 (2.8)	24.0 (2.8)	29.6 (4.0)	51.9 (12.2)	30.9 (3.6)
	White	51.4 (2.7)	51.2 (4.3)	51.2 (3.8)	51.2 (3.8)	51.2 (3.8)	56.8 (4.7)	41.3 (11.4)	53.9 (5.2)
	Hispanic	16.3 (3.3)	17.6 (4.0)	15.4 (3.0)	17.0 (3.5)	13.2 (7.0)	9.2 (2.2)	---	10.4 (3.5)
	Other	6.8 (1.0)	7.8 (1.3)	5.0 (0.8)	7.0 (1.1)	5.8 (1.6)	4.4 (1.1)	---	4.8 (0.9)
Age a	≤24	17.8 (1.2)	19.3 (1.5)	19.7 (1.8)	19.4 (1.3)	1.2 (0.4)	2.4 (1.1)	10.8 (4.2)	4.7 (1.6)
	25-34	41.0 (1.3)	45.9 (1.8)	40.1 (1.7)	44.3 (1.4)	13.5 (2.3)	8.0 (1.7)	56.8 (10.7)	15.0 (1.9)
	35-44	27.9 (1.3)	26.5 (1.7)	31.4 (2.6)	27.8 (1.5)	38.8 (3.9)	19.7 (2.9)	10.4 (3.6)	28.6 (2.6)
	45-54	9.1 (1.0)	6.9 (1.1)	6.7 (1.1)	6.9 b (0.9)	25.2 (2.3)	34.9 (3.5)	7.4 (3.5)	26.4 (2.0)
	≥55	4.3 (0.6)	1.4 (0.5)	2.0 (0.6)	1.6 (0.4)	21.4 (3.5)	35.1 (5.0)	---	25.4 (3.1)
Marital Status b	Married	32.4 (1.4)	29.8 (1.7)	28.2 (2.0)	29.4 (1.4)	66.1 (6.7)	50.4 (5.5)	72.9 (8.3)	56.4 (4.3)
	Separated	15.6 (1.2)	16.9 (1.6)	15.9 (1.8)	16.6 (1.3)	8.4 (5.5)	6.6 (1.5)	---	7.5 (2.1)
	Divorced	20.9 (1.3)	22.0 (1.7)	19.9 (1.7)	21.4 (1.4)	11.8 (2.1)	20.6 (4.4)	---	16.4 (2.5)
	Widowed	3.1 (0.7)	2.1 (0.6)	3.0 (1.0)	2.4 (0.5)	5.2 (1.4)	14.2 (6.3)	---	8.7 (2.8)
	Never Married	28.1 (1.7)	29.2 (2.1)	33.1 (3.2)	30.2 (1.8)	8.6 (2.6)	8.2 (2.0)	21.4 (7.1)	11.0 (1.6)
Highest Degree c	No degree	28.7 (1.5)	27.8 (1.9)	34.5 d (2.6)	29.6 (1.7)	7.9 e (1.5)	36.1 (5.3)	0	21.4 (2.8)
	HS Diploma GED	44.2 (1.2)	46.0 (1.6)	40.7 (2.4)	44.5 (1.3)	53.7 (4.6)	35.0 (4.2)	18.0 (10.6)	41.9 (2.9)
	Associates or VocTech	19.4 (1.2)	19.8 (1.7)	18.6 (1.7)	19.5 (1.3)	26.0 (4.5)	17.3 (4.4)	5.7 (2.5)	19.0 (2.3)
	Bachelor's or Higher	5.0 (0.8)	4.0 (0.9)	2.8 (0.5)	3.7 (0.7)	10.0 (1.6)	9.5 (3.4)	76.0 (10.7)	15.6 (1.8)
	Other	2.6 (0.5)	2.4 (0.6)	3.5 (0.8)	2.7 (0.5)	2.4 (0.6)	2.0 (0.9)	0.3 (0.3)	2.1 (0.5)
Employment Status f	Full-time	42.2 (1.5)	43.3 g (2.1)	38.0 h (2.3)	41.9 (1.7)	42.2 i (5.1)	37.9 j (4.6)	99.5 (0.4)	44.4 (3.0)
	Part-time	10.5 (0.9)	11.0 (1.5)	10.2 (1.2)	10.8 (1.1)	12.3 (2.4)	4.8 (1.0)	---	8.3 (1.4)
	Unemployed	10.6 (0.9)	10.3 (1.0)	14.6 (1.4)	11.5 (0.9)	1.7 (1.0)	3.7 (1.6)	---	4.0 (1.1)
	Do not work	29.3 (1.7)	27.9 k (1.8)	29.0 l (2.7)	28.2 (1.8)	37.0 (3.2)	49.4 (5.1)	0.2 (0.2)	38.4 (2.6)
	Other	7.3 (1.0)	7.4 (1.3)	8.3 (1.3)	7.7 (1.1)	6.7 (1.6)	4.2 (1.5)	0.3 (0.4)	4.9 (1.0)
TOTAL		100	65.5 (1.6)	24.0 (1.5)	88.8 (1.2)	4.5 (0.6)	5.2 (0.6)	0.9 (0.2)	11.2 (1.2)

^ Includes children in “other” out-of-home placement settings (back)

a Caregiver age differs significantly by service setting (X2 =81.3, p<.001) (back)

b Caregiver marital status differs significantly by service setting (X2=133.1, p<.001) (back)

c Caregiver education level differs significantly by service setting (X2=91.0, p<.001) (back)

d Education level of in-home caregivers receiving child welfare services and non-kinship foster caregivers is different (X2=34.4, p<.001) (back)

e Education level of non-kinship and kinship caregivers is different (X2=17.6, p<.01) (back)

f Caregiver employment status differs significantly by service setting (X2=114.4, p<.001) (back)

g Employment status of in-home caregivers not receiving child welfare services and group home caregivers is different (X2=22.7, p<.001) (back)

h Employment status of in-home caregivers receiving child welfare services and group home caregivers is different (X2=22.9, p<.001) (back)

i Employment status of non-kinship caregivers and group home caregivers is different (X2=26.6, p<.001) (back)

j Employment status of kinship caregivers and group home caregivers is different (X2=20.0, p<.001) (back)

k Employment status of in-home caregivers not receiving child welfare services and kinship caregivers is different (X2=18.3, p<.01) (back)

Race/ethnicity of the child and nonkinship foster caregivers generally match. Approximately two-thirds of African American children have an African American foster parent and most White children (92%) have a White foster parent. Only 42% of Hispanic children have a Hispanic foster parent and slightly less than one-third of other race/ethnicity children are living with other race/ethnicity caregivers (Table 6-2). This match of other race/ethnicity children and caregivers may, however, be an overestimate because this category comprises children who are a variety of races/ethnicities (e.g., Native Indian/Alaskan, Asian/Hawaiian/ Pacific Islander). Among all the children in foster care or kinship care, 78% live with a caregiver with shared racial/ethnic identity.

Table 6-2. Nonkinship Foster Care: A Comparison of the Child to Caregiver Race/Ethnicity ^
Race/Ethnicity of Child	Race/Ethnicity of Current Caregiver
Race/Ethnicity of Child	African American Percent (SE)	White Percent (SE)	Hispanic Percent (SE)	Other Percent (SE)
African American	65.5 (6.0)	16.0 (4.7)	13.4 (7.2)	5.1 (2.9)
White	3.3 (1.2)	92.4 (2.0)	2.9 (1.3)	2.4 (1.1)
Hispanic	3.6 (1.6)	48.5 (20.5)	42.0 (21.0)	2.7 (2.0)
Other	4.7 (2.2)	42.4 (9.4)	9.1 (5.5)	31.4 (7.9)

^ Bold numbers indicate that the caregiver is the same race/ethnicity as the child. Children in group care and other care are excluded. (back)

We also examined the converse—the proportion of foster caregivers of each racial/ethnic group who care for a child with the same ethnicity. While 88% of African American foster caregivers care for African American children, 68% of White foster parents care for White children, and slightly under 50% of Hispanic foster parents care for Hispanic children. Children of other races/ethnicities live in the care of other race/ethnicity caregivers 43% of the time. Again, this seeming match of other race/ethnicity caregivers and children may be deceptive (Table 6-3).

Table 6-3. Nonkinship Foster Care: A Comparison of the Caregiver to Child Race/Ethnicity^
Race/ Ethnicity of Current Caregiver	Race/Ethnicity of Child
	Percent/ (SE)
	African American Percent (SE)	White Percent (SE)	Hispanic Percent (SE)	Other Percent (SE)
African American	88.4 (3.3)	3.3 (1.2)	3.6 (1.6)	4.7 (2.2)
White	11.7 (2.8)	68.3 (6.6)	13.8 (6.9)	6.3 (2.2)
Hispanic	38.8 (5.1)	8.5 (6.7)	47.7 (5.1)	5.4 (1.6)
Other	34.2 (15.3)	15.9 (7.3)	7.1 (4.5)	42.8 (14.9)

^ Bold numbers indicate that the caregiver is the same race/ethnicity as the child. Children in Group Care and other care are excluded. (back)

6.1.4 Marital Status

Approximately one-third of caregivers are married, one-third are separated or divorced, and one-third has never been married. A small proportion of caregivers are widowed (3%). Caregiver marital status differs significantly by service setting. A majority (56%) of out-of-home caregivers are married, whereas in-home caregivers' marital status is more evenly distributed among those married, divorced, and never married. In-home cases open to CWS and in-home cases with closed CWS cases have nearly matching percentages for each type of marital status (Table 6-1), with about 29% currently married, a slightly larger proportion never married, and the remainder married previously.

6.1.5 Education Level

Caregivers reported on the highest educational degree they had obtained. Almost one-third of caregivers have less than a high school diploma, and nearly half have a high school diploma. Almost 20% have an Associate or Vocational-Technical degree, and only 5% have a bachelor's degree or higher. An additional 3% have other types of degrees (e.g., RN). A significant association exists between education level and child setting. The highest proportion of caregivers with no high school degree are kinship caregivers (36%) and caregivers with in-home open CWS cases (35%). Kinship and in-home caregivers receiving child welfare services are more likely than foster caregivers to have no high school degree. Kinship caregivers have the most even distribution of educational achievement, with just over one-third having no degree, just over one-third having a high school diploma, and just less than one-third having an other, Associate, or Bachelor's degree or higher. Group home caregivers are significantly more educated than any other group, as all group home caregivers have at least a high school diploma or GED and 75% have a Bachelor's degree or higher (Table 6-1).

6.1.6 Employment Status

Almost half of all caregivers work full time outside the home, whereas approximately one-third do not work. The remaining caregivers are fairly evenly distributed in the proportion that work part time (11%), are unemployed (11%), and report other employment status (7%). A significant association exists between employment status and service setting. Group home caregivers are significantly different from caregivers in every other service setting. Group home caregivers are, by definition, employed and almost all work full time, while only about 40% of caregivers in the other types of service settings work full time. Additionally, kinship caregivers are significantly more likely to not work than in-home caregivers with open or closed CWS cases (Table 6-1).

6.1.7 Comparisons to Other National Studies

The 2000 U.S. Census and 1999 National Survey of America's Families (NSAF) data provide a basis for comparing caregivers of children involved with CWS to other national samples (Table 6-4). Caregivers involved with CWS are compared to other families on two variables: age and education. The Census defines family as “a householder and one or more other people living in the same household who are related to the householder by birth, marriage, or adoption” (U.S. Census Bureau, 2001). Consequently, not all Census families include children. This explains why 21% of all U.S. heads of all households are over age 54, in contrast to the NSCAW population, in which just 4% are over age 54. The Census definition of family may also explain why Census families are fairly evenly distributed across age categories.

Table 6-4. Caregiver Age: Comparison to 2000 U.S. Census and 1999 NSAF Data ^
Age	Total	Setting					CENSUS ^^	NSAF ^^^
		In-Home		Out-of-Home			CENSUS ^^	NSAF ^^^
		No Services	Services	Foster Home	Kinship Foster Care	Group Care	All Families	All house-holds	Child's Mother Lives Elsewhere
		Percent
≤24	18	19	20	1	2	11	14	7	4
25-34	41	46	40	13	8	57	14	34	21
35-44	28	26	31	39	20	10	16	43	38
45-54	9	7	7	25	35	7	13	14	24
≥55	4	1	2	21	35	15	21	2	13

^ Standard errors for NSCAW data are shown in Table 6-1 and are unavailable for NSAF and Census data. (back)

^^ The Census defines family as “a householder and one or more other people living in the same household who are related to the householder by birth, marriage, or adoption” (U.S. Census Bureau, 2001); consequently, not all Census families include children. (back)

^^^ NSAF defines family as anyone living in the same household related by birth, marriage, or adoption, any unrelated children, any unmarried partners, and anyone living in the household related to these unmarried partners. (back)

NSAF is a nationally representative survey of the characteristics of households with at least one member under age 65, including children, adults, and their families; therefore, it is more comparable to our sample of families than the Census data. The total NSCAW population is composed of caregivers younger than the entire sample of NSAF families (18% versus 7% under age 25). Caregiver age for NSAF households in which the mother lives elsewhere are most similar to foster and kinship caregiver families (http://newfederalism.urban.org/nsaf/cpuf, August 11, 2002).

NSCAW caregivers appear to have lower education levels than those in NSAF. Comparing caregiver education level to the NSAF sample, a greater proportion of caregivers of children involved with CWS have less than a high school education (29% versus 13%), while a smaller proportion have a Bachelor's degree or higher (5% versus 26%). The education level of all out-of-home caregivers most closely resembles NSAF households where the child's mother lives elsewhere (Table 6-5).

6.1.8 Caregiver Characteristics By In-Home Versus Out-of-Home Setting

In-home and out-of-home caregiver characteristics were compared using logistic regression to determine if bivariate relationships tested above were changed once the influences of caregiver characteristics of age, gender, race, marital status, education level, and employment status were controlled. The multivariate analysis generally confirmed bivariate analyses. Caregiver age, marital status, and education level differ significantly by in-home versus out-of-home settings. No significant associations exist for gender, race/ethnicity, or employment status (Table 6-6).

In-home caregivers are far younger than out-of-home caregivers. In-home caregivers have much greater odds than out-of-home caregivers of being under age 45 and significantly less odds of being aged 45 or older. With the older age groups as the reference groups, in-home caregivers are much more likely to be younger – they have over three times the odds (OR = 3.41, p < .001) of being aged 35 to 44 in contrast to being aged 45 to 54, and over 18 times the odds (OR = 18.48, p < .001) of being aged 35 to 44 years in contrast to being aged 55 years and older.

Table 6-5. Education Level of Primary Caregiver: Comparison of NSCAW Data and 1999 NSAF Data ^
		In-Home			Out-of-Home				NSAF^^^
	Total	Setting							NSAF^^^
		No Services	Services	Total In-Home	Foster Care	Kinship Foster Care	Group Care	TOTAL Out-of-Home^^	All House-holds	Child's Mother Lives Elsewhere
	Percent
No degree	29	28	35	30	8	36	---	21	13	17
HS Diploma/ GED	44	46	41	44	54	35	18	42	44	48
Associate's or VocTech	19	20	19	19	26	17	6	19	17	19
Bachelor's or Higher	5	4	3	4	10	10	76	16	26	16
Other	3	2	4	3	2	2	---	2	0	0

^ Standard errors for NSCAW data are shown in Table 6-1 and are unavailable for NSAF and Census data. (back)

A significant association also exists between marital status and in-home versus out-of-home setting. In-home caregivers have significantly greater odds than out-of-home caregivers of being separated, divorced, or never married. Out-of-home caregivers have significantly greater odds of being married.

Table 6-6. Logistic Regression Modeling In-home versus Out-of-Home Care
Characteristics		OR	95% CI
Age	35 - 44	(reference group)
	≤ 24	6.91 *	3.23, 14.80
	25 - 34	3.18 *	2.18, 4.62
	45 - 54	.29 *	.19, .44
	≥ 55	.05 *	.03, .10
Gender	Female	(reference group)
Gender	Male	1.19	.61, 2.32
Race/Ethnicity	White	(reference group)
	African American	.76	.50, 1.41
	Hispanic	1.40	.73, 2.68
	Other	1.78	1.04, 3.05
Marital Status	Married	(reference group)
	Separated	4.85 *	2.32, 10.13
	Divorced	3.87 *	2.39, 6.27
	Widowed	1.47	.62, 3.47
	Never Married	2.80 *	1.95, 4.02
Highest Degree	High School Diploma/GED	(reference group)
	No degree	1.51	1.00, 2.30
	Associates or VocTech	1.23	.79, 1.91
	Bachelor's or Higher	.34 *	.18, .63
	Other	1.61	.67, 3.89
Employment Status	Full-time	(reference group)
	Part-time	1.45	.71, 2.95
	Unemployed	1.41	.73, 2.75
	Do not work	1.03	.68, 1.55
	Other	2.12	1.09, 4.12

Cox and Snell pseudo-R2 is .19

* p<.001

A significant association exists for caregiver education and in-home versus out-of-home setting as well. On average, in-home caregivers have a significantly lower level of education than out-of-home caregivers. In-home caregivers are more likely to have a high school diploma or GED than a Bachelor's degree or higher. In-home caregivers have only one-third the odds (OR = .34, p < .001) of out-of-home caregivers of having a Bachelor's degree or higher.

When in-home caregivers are compared with children receiving and not receiving child welfare services (not shown in table), few differences are observed. In-home caregivers of children not receiving child welfare services tend to have greater odds of being aged 25 to 34 years (OR = 1.44, p ≤ .01) and working full time (OR = 1.55, p < .01). These findings provide evidence that many caregivers involved with CWS are facing challenges that can make childrearing extremely challenging, particularly for children with the kinds and degree of problems described in Chapter 5. Caregivers of children remaining in the home are younger, less educated, and more likely to be single than out-of-home caregivers—characteristics that may elevate risks for child maltreatment (e.g., Gringlas &Weintraub, 1995, Sedlak & Broadhurst, 1996) if steps are not taken to assist families. The findings that these conditions exist at such a high rate in the homes of families with allegations of child abuse and neglect is not surprising, but the high level of these risk factors in foster family homes is less well-known.

6.2 Out-of-Home Caregivers

Group home caregiver characteristics and foster parent years of experience are presented in this section. Group home caregivers are different from all other types of caregivers; therefore, a description of group caregivers is provided separately. Only group home caregivers are employed for the sole purpose of caring for the children under their supervision.

The number of years of experience of foster parents was examined separately for nonkinship and kinship foster care and group care to identify any differences between the three types of out-of-home placements. Kinship caregivers tend to have the least experience with caring for a CWS-involved child.

6.2.1 Group Home Caregiver Characteristics

Although relatively few children are residing in group care during the first months after CWS involvement, group care is a more common place to reside for older children and for children who remain in CWS for a longer time (U.S. Department of Health and Human Services, Administration for Children and Families, in press; Webster, 2001). Nationally, according to the Adoption and Foster Care Analysis Reporting System (AFCARS)-derived information, about 8% of children in CWS supervised out-of-home care were residing in group care in 2001; this proportion rises to more than 40% for children aged 16 and older (Wulczyn, 2001). Yet there has not been a national study of group care for nearly two decades (Dore, Young, & Pappenfort, 1984). Although children were sampled and not group homes, and only one group home staff member was interviewed for each child, NSCAW offers unique information about group care providers.

Most group home caregivers are either African American (52%) or White (41%), with the remaining proportion being either Hispanic or other race/ethnicity. Group home caregivers are generally older than in-home caregivers and younger than nonkinship and kinship foster caregivers, with most group home caregivers being between the ages of 24 and 45. Group home caregivers are usually married (73%) or never married (21%). All have at least a high school diploma and most have a Bachelor's degree or higher, which most likely is due to employment requirements.

6.2.2 Foster Parent Years of Experience

Caregivers have an average of 4.5 years of experience as foster caregivers, though this varies widely depending on the age of the caregiver and whether they are foster caregivers or kinship caregivers (Table 6-7). In general, older caregivers have more experience as foster parents, and foster caregivers have more experience than kinship caregivers. There are no differences in foster parent experience by the race/ethnicity of the caregiver. Multivariate analysis confirmed that the CWS setting and caregiver's age have a significant relationship to the length of foster caregivers experience.

6.2.3 Discussion of Out-of-Home Caregivers Characteristics and Experience

Group home caregivers are distinguishable from other types of caregivers in several ways. Most notably, they are the only category of out-of-home caregivers who are hired as employees of a facility that provides a home for children who cannot remain in their own home. Also, group home caregivers are the largest group of caregivers with a Bachelor's degree or higher. Group home caregivers are somewhat older than in-home caregivers and may have more experience caring for children, but youths living in group homes are older and have more problems than children in other CWS settings.

Kinship caregivers tend to have the least experience of out-of-home caregivers in caring for CWS-involved children. Kinship caregivers generally have less familiarity with CWS and may need more information about resources to assist them in caring for children.

6.3 Household Characteristics

The characteristics of children's living environments play a role in their well-being. Children benefit when their basic needs for food, shelter, and adult attention are met. Households in which there is financial strain (i.e., inadequate income to cover the needs of all household members) pose a number of risks to children. The common co-occurrence of poverty and child maltreatment has long been recognized (Giovannoni & Billingsley, 1970). In fact, in the eyes of some scholars, CWS are primarily a response to poverty (Pelton, 1989). Caregivers who do not have sufficient income to pay for basic necessities for their children are likely to feel overwhelmed and anxious. Severe levels of poverty can result in family homelessness, child malnutrition, and inadequate medical care for children.

The number of household members, including the total number of adults and children in the home, has been posited as a risk factor for child maltreatment. The Third National Incidence Study of Child Abuse and Neglect (NIS) found that large family size is a risk factor for child maltreatment. Specifically, NIS reported that incidence rates of educational and physical neglect were higher for children in large families (those with four or more children) than for “only” children and children in families with two to three children (Sedlak & Broadhurst, 1996).

Table 6-7. Length of Foster Caregiver Experience (In Years) by Age and Race
Characteristics		TOTAL	Foster Care	Kinship Care	Group
Characteristics		Mean/ (SE)
Caregiver Age	<25 years	1.3 b c d (0.7)	---	0.3 (0.09)	3.0 (1.4)
	25-34 years	2.0 e f g (0.3)	1.6 (0.2)	1.3 (0.5)	3.2 (1.1)
	35-44 years	3.7 h (0.6)	4.7 i (0.7)	1.2 (0.3)	11.0 j (2.7)
	45-54 years	4.3 (0.4)	6.8 k (0.8)	2.6 (0.5)	5.2 l (0.9)
	55+ years	7.0 (1.0)	9.1 (1.6)	5.0 (1.0)	---
Caregiver Race/ Ethnicity	African American	5.2 (0.7)	5.6 (0.8)	4.2 (1.0)	7.1 (3.4)
	White	4.2 (0.4)	6.1 m (0.5)	2.5 (0.6)	4.9 (1.2)
	Hispanic	3.4 (0.5)	4.5 n (0.6)	1.4 (0.4)	---
	Other	5.8 (1.5)	6.6 (2.6)	4.1 (1.4)	---
TOTAL		4.5 (0.3)	5.8 a (0.4)	3.0 (0.4)	6.4 (1.7)

Note: 31% of foster and kinship caregivers had 6 months or less experience.

a Caregivers of children in foster care had more years' experience than caregivers of children in kinship care (t = 4.9, p ≤ .001). (back)

b Caregivers between 35 and 44 years old had more years' experience than caregivers under 25 years old (t = 2.8, p ≤ .01). (back)

c Caregivers between 45 and 54 years old had more years' experience than caregivers under 25 years old (t = 3.5, p ≤ .001). (back)

d Caregivers 55 years and older had more years' experience than caregivers under 25 years old (t = 6.8, p ≤ .001). (back)

e Caregivers between 35 and 44 years old had more years' experience than caregivers between 25 and 34 years (t = 3.0, p ≤ .01). (back)

f Caregivers between 45 and 54 years old had more years' experience than caregivers between 25 and 34 years (t = 4.4, p ≤ .001). (back)

g Caregivers 55 years old and over had more years' experience than caregivers between 25 and 34 years (t = 5.5, p ≤ .001). (back)

h Caregivers 55 years old and over had more years' experience than caregivers between 35 and 44 years (t = 3.0, p ≤ .01). (back)

i Foster caregivers between 35 and 44 years old had more years' experience than kinship caregivers between 35 and 44 years (t = 4.6, p ≤ .001). (back)

j Group caregivers between 35 and 44 years old had more years' experience than kinship caregivers (t = 3.7, p ≤ .001). (back)

k Foster caregivers between 45 and 54 years old had more years' experience than kinship caregivers between 45 and 54 years (t = 4.6, p ≤ .001). (back)

l Group caregivers between 45 and 54 years old had more years' experience than kinship caregivers (t = 2.6, p ≤ .01). (back)

m White foster caregivers had more years' experience than White kinship caregivers (t = 4.6, p ≤ .001). (back)

n Hispanic foster caregivers had more years' experience than Hispanic kinship caregivers (t = 3.5, p ≤ .001). (back)

Information in this section describes household characteristics for children. The total household income and level of poverty are considered across child welfare service settings and compared to national statistics. The relationship between the total number of children in the household and child setting is also presented. In addition, children's relationships to adults in the household are considered, as well as the extent of children's out-of-home placement with their siblings.

6.3.1 Total Household Income

Studies have consistently shown that families who are receiving CWS are likely to be receiving Temporary Assistance for Needy Families (TANF) and other income-tested federal service programs (e.g., U.S. Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation, 2000). NIS also found that families with incomes of less than $15,000 are four times more likely to be identified for child maltreatment than families on the whole (Sedlak and Broadhurst, 1996). Yet detailed descriptive data about family income of the families coming to the attention of CWS have not been available.

Approximately one-quarter of households who have been investigated for child maltreatment had a total household income under $10,000 (Table 6-8). A high proportion of households had a total income under $25,000 (65%). There was a significant association between service setting and income level (X2= 56.5, p ≤ .001). Although out-of-home caregiver households were more likely to have higher incomes than households with children living at home, and foster homes were significantly more likely to have higher incomes than kinship care homes (X2= 17.6, p < .01), substantial numbers of children in out-of-home placement were living in very poor households. There were no differences in household income among households with children living at home who received services and households that did not receive services (Table 6-8).

In-home and out-of-home caregivers of children involved with CWS have generally lower income levels than Census families (Table 6-9). Half of families in the general population have an income level at or above $50,000, compared with only 10% of those caring for children involved with CWS. The proportion of families caring for children involved with the CWS who live on less than $25,000 per year is high (almost 70%). Given the sizable families living in many of these households, financial resources are seriously strained for many caregivers involved with CWS.

While the Census definition of family does not necessarily include children, it is apparent that higher proportions of caregivers who have been investigated for child maltreatment are falling in the lower income brackets compared to households in the 2000 Census. Even if it were possible to examine only families with dependent children using the Census data, income levels might be more comparable, but it is unlikely that the large difference would be changed drastically.

6.3.2 Poverty Level

To determine the extent to which the household incomes reported above equate to levels of poverty, we followed the procedures used by the U.S. Census Bureau (Dalaker, 2000), which include both the family's income level and the number of adults and children in the household. According to the Census Bureau, the average threshold indicating poverty ranges from $11,239 for a two-member household to $35,060 for a household with nine or more members. We collected information about income levels in $5,000 increments that ranged from 0 to $5,000 per year to over $50,000 per year. The midpoint of each increment was chosen to indicate the household's income. Households with an income “over $50,000” were all assigned an income of $75,000 for the purposes of calculating poverty. This choice was based on information from NSAF that indicated that twice as many families had incomes greater than or equal to 300% of poverty than had incomes of 200% to 300% of the poverty level (Urban Institute, 2002). Children living in group homes or “other” types of out-of-home care are excluded from the following analyses.

Table 6-8. Total Household Income
			In-Home			Out-of-Home
		TOTAL	Setting
			No Services	Services	TOTAL In-Home	Foster Care	Kinship Foster Care	Group Care	TOTAL Out-of-Home^
		Percent/ (SE)
Household Income	<$10,000	24.3 (1.3)	24.8 (1.6)	30.1 (2.7)	26.3 (1.4)	5.0 (3.0)	6.5 (1.9)	---	7.5 (1.7)
	$10,000 - $14,999	18.3 (1.3)	20.0 (1.8)	16.9 (1.5)	19.1 (1.3)	4.6 (1.6)	15.9 (4.1)	---	10.8 (2.3)
	$15,000 - $24,999	23.1 (1.2)	23.2 (1.7)	23.5 (1.8)	23.3 (1.3)	13.9 (2.3)	29.6 (4.2)	---	21.6 (2.5)
	$25,000 - $34,999	13.7 (1.0)	14.3 (1.3)	11.4 (1.2)	13.5 (1.1)	21.3 (4.7)	11.7 (2.6)	---	15.7 (2.5)
	$35,000 - $49,999	10.4 (1.0)	9.3 (1.1)	8.7 (1.3)	9.1 (4.1)	23.7 (2.9)	21.3 (5.6)	---	21.0 (3.5)
	$50,000 and over	10.2 (1.1)	8.4 (1.2)	9.4 (1.9)	8.7 (1.2)	31.7 (4.2)	14.9 (3.5)	---	23.5 (3.0)
TOTAL		100	100	100	100	100	100	---	100

^ Includes “group care” and “other” out-of-home placement types. (back)

Table 6-9. Total Household Income Compared with 2000 Census Data
		In-Home		Out-of-Home
	Total	Setting^					Census^^ Families
		No Services	Services	Foster Care	Kinship Foster Care	Group Care
	Percent
< $10,000	24	25	30	5	7	---	6
$10,000 to $14,999	18	20	17	5	16	---	5
$15,000 to $24,999	23	23	24	14	30	---	11
$25,000 to $34,999	14	14	11	21	12	---	12
$35,000 to $49,999	10	9	9	24	21	---	17
$50,000 +	10	8	9	32	15	---	50

^ Standard errors are shown in Table 6-1. (back)

More than half of all NSCAW households have an income below the federal poverty threshold, and more than one in five have an income at less than 50% of the poverty level. More than 80% of study families whose children remained at home have incomes below 200% of the poverty level (Table 6-10). The proportion below the poverty threshold varies substantially by service setting: households with children living at home are more likely to be below the poverty threshold than foster or kinship care households. There is no significant difference in income level between the households with children living at home who received services and those who did not receive services.

Table 6-10. Proportion of Households by Poverty Level
Poverty Level	TOTAL	Setting
		In-Home			Out-of-Home
		No Services	Service	TOTAL In-Home	Foster Care	Kinship Foster Care	TOTAL Out-of-Home^
	Percent / (SE)
<50%	20.6 (1.4)	21.2 a (1.6)	24.7 (2.7)	22.1 (1.4)	6.3 b (3.1)	5.0 (1.4)	5.6 (1.5)
50% to <100%	29.4 (1.5)	30.5 (2.0)	29.4 (2.0)	30.2 (1.6)	10.1 (2.3)	34.9 (4.9)	23.4 (2.9)
100% to <150%	17.7 (1.0)	17.8 (1.4)	18.7 (1.6)	18.0 (1.1)	19.0 (2.6)	11.5 (3.2)	15.0 (1.8)
150% to <200%	13.1 (0.7)	13.5 (0.9)	10.8 (1.0)	12.8 (0.7)	18.3 (4.7)	15.1 (3.2)	16.6 (2.8)
200% and over	19.2 (1.4)	17.1 (1.8)	16.5 (1.7)	16.9 (1.5)	46.2 (3.6)	33.5 (5.3)	39.4 (3.3)
TOTAL	100	100	100	100	100	100	100

^ Note: Group homes and “other” types of out-of-home care were excluded. (back)

a Households with children living at home are more likely to be below the poverty threshold than foster or kinship care households (X2= 93.9, p < .001). (back)

b Households of foster children have higher incomes than households of children in kinship care (X2= 14.2, p < .01). (back)

Many foster and kinship households have incomes below the poverty threshold as well (16% and 40%, respectively). Only 39% of out-of-home care provider families have incomes that place them beyond 200% of the poverty level. Households of foster children are less likely to be below 100% of the poverty level and more likely to be above 200% of the poverty level than households of children in kinship care (Table 6-10).

Comparisons to NSAF and Census data indicate that families who have had involvement with the CWS are much more likely to be below the poverty level than those in the general population. Among all families—in-home, foster, and kinship—those involved with the CWS are five times more likely to have income at only 50% of the poverty level than families in the general population, and nearly three times more likely to have very low income than families in NSAF (Table 6-11).

Table 6-11. Proportion of Households by Poverty Level, NSAF and NSCAW^
Poverty Level	NSCAW	NSAF	Census
<50%	21	7	4
50 to <100%	29	10	7
100 to <150%	18	11	9
150 to <200%	13	11	9
200% and over	19	60	71

^ Standard errors for NSCAW data are provided in Table 6-10. (back)

To better understand the relationship between child poverty and the type of out-of-home setting, we conducted a regression analysis on type of foster care placement (foster versus kinship care). These findings confirm the bivariate finding that children in foster care are substantially more likely to live in households with incomes above the poverty line than children in kinship foster care (OR = 3.64, p < .001). (It was not possible to compare the likelihood of placement and the association with household poverty level because we lack household income information for families from which children were removed.)

6.3.3 Total Number of Household Members

Overall, children involved with the CWS live in households with an average of 4.4 household members (including the study child) and a maximum total number of 11 household members. The average for children who are involved with the CWS but who are living at home is also 4.4, which is significantly less (p < .001) than the 4.9 household members for children living in out-of-home care (excluding children in group care). The greater number of household members in out-of-home care, however, seems to be accounted for in nonkinship foster homes, as the average number of household members in these homes is 5.5. This average number of children in nonkinship foster homes is significantly more (p < .001) than the average in kinship foster care homes and for children remaining at home, which is 4.4 household members (Table 6-12).

Overall, as well as within the subpopulation of children living at home, older children are significantly more likely to live in larger households. One possible explanation is that older children are more likely than younger children to have younger siblings in the home with them. The fact that the significant difference between household sizes of older and younger children does not extend to children in out-of-home care would support this explanation. The only significant difference with regard to average household size based on race/ethnicity is between Hispanic and White children. Both overall and within the subpopulation of children remaining at home, Hispanic children live in households with more members than White children.

Table 6-12. Total Number of Household Members (Including Study Child)
			In-Home			Out-of-Home
		TOTAL	Setting
			No Services	Services	TOTAL In-Home	Foster Care	Kinship Foster Care	TOTAL Out-of-Home^
		Mean / (SE)
Age	0-2	4.3 a (0.1)	4.1 (0.1)	4.2 (0.1)	4.1 c d (0.1)	5.7 (0.3)	4.9 (0.3)	5.3 (0.2)
	3-5	4.3 b (0.1)	4.2 (0.1)	4.3 (0.1)	4.3 (0.1)	5.2 (0.5)	4.2 (0.5)	4.5 (0.4)
	6-10	4.5 (0.1)	4.4 (0.1)	4.6 (0.1)	4.4 (0.1)	5.6 (0.3)	4.1 (0.2)	4.9 (0.2)
	11+	4.7 (0.1)	4.7 (0.2)	4.7 (0.1)	4.7 (0.1)	5.1 (0.3)	4.5 (0.3)	4.8 (0.2)
Race/ Ethnicity	African American	4.4 (0.1)	4.3 (0.1)	4.5 (0.1)	4.4 (0.1)	5.2 (0.2)	4.2 (0.2)	4.8 (0.2)
	White	4.3 e (0.1)	4.2 (0.1)	4.3 (0.1)	4.2 f (0.1)	5.9 (0.2)	4.3 (0.2)	4.9 (0.2)
	Hispanic	4.8 (0.1)	4.7 (0.2)	4.8 (0.2)	4.7 (0.1)	5.3 (0.4)	5.7 (0.3)	5.4 (0.3)
	Other	4.6 (0.2)	4.7 (0.3)	4.4 (0.3)	4.6 (0.2)	5.4 (0.3)	3.8 (0.4)	5.0 (0.4)
TOTAL		4.4 (0.1)	4.4 (0.1)	4.5 (0.1)	4.4 g (0.1)	5.5 h (0.2)	4.4 (0.1)	4.9 (0.1)

^ Includes children in “other” out-of-home placement settings, but not children in group care. (back)

a Children aged 11 years and older live in households with more members than children 0 to 2 years of age (t=-2.7, p<.01). (back)

b Children aged 11 years and older live in households with more members than children 3 to 5 years of age (t=-2.8, p<.01). (back)

c Children aged 6 to 10 years remaining at home live in households with more members than children 0 to 2 years of age remaining at home (t=-3.1, p<.01). (back)

d Children age 11 years and older remaining at home live in households with more members than children 0 to 2 years of age remaining at home (t=-3.6, p<.001). (back)

e Hispanic children live in households with more members than White children (t=-3.4, p≤.001). (back)

f Hispanic children remaining at home live in households with more members than White children remaining at home (t=-3.0, p<.01). (back)

g Children in out-of-home placements live in households with more members than children remaining at home (t=-5.9, p<.001). (back)

h Children in nonkinship foster care live in households with more members than children in kinship foster care (t=5.2, p<.001). (back)

6.3.4 Number of Children in the Household

Overall, children involved with CWS live in households with an average of 2.5 children under the age of 18 (including the study child).24 (Note that the questionnaire program allowed for a maximum of 10 in the household roster, but few families reached the maximum.) As might be expected, similar patterns of significance emerge with regard to the number of children in the household, as was seen with the total number of household members. The average number of children in households where the child remains at home closely approximates the average number of children overall and is significantly less than the average number of children in out-of-home care (p<.001). Again, this significant difference is accounted for largely in nonkinship foster homes, which have a significantly higher average number of children than do kinship care homes (p<.001) (Table 6-13).

In addition to Hispanic children living in households with more children than White children, both overall and within the in-home subpopulation (similar to the pattern of Hispanic children living in larger households overall), African American children also live in households with more children than White children. Again, this holds true overall, as well as for children remaining at home, though not for children living in out-of-home placement.

The finding that older children are significantly more likely than younger children to live in households with higher average numbers of children, both overall and within the in-home subpopulation, supports the theory discussed earlier that these children have had greater opportunity to see younger siblings being brought into the household. For children in out-of-home care, however, this trend is reversed, as the homes in which infants are placed have significantly more children than those in which children aged 11 years and older are placed. (Yet infants are not statistically more likely to be placed with siblings.) Perhaps older children are placed in homes with fewer children so that behavior problems and other issues they may be facing can be more readily addressed. At the same time, younger children also need substantial attention to their needs.

6.3.5 Number of Children in the Household as a Risk Factor

To examine the association between family size and service receipt for children in the NSCAW study population, a logistic regression was run that modeled receipt of services among children who are living at home and utilized number of children in the household as an independent variable, while controlling for child age, gender, and race/ethnicity. Although not significant at the p<.01 level, the model did reveal a trend indicating that children remaining at home and living in households with four or more children (including the study child) were more likely than those remaining at home and living in households with two or three children to receive services from CWS (p<.05)—similar to the NIS results insofar as both studies indicate that families with four or more children appear to need greater levels of service. No difference exists between children living in households with four or more children and “only” children.

Older children and African American and Hispanic children more frequently experience the risk resulting from having several children living in the home. In making age comparisons of households with four or more children, youths aged 11 years and older are significantly more likely than children aged 0 to 2 and 3 to 5 to live in a home with four or more children: 30% versus 15% and 17%, respectively (X2=23.6, p<.001; X2=14.7, p<.001). Significantly more 6- to 10-year-olds than 0- to 2-year-olds live in households with four or more children: 25% versus 15% (X2=12.4, p<.001). Significantly more African American (27%) and Hispanic (28%) children than White children (17%) live in households with four or more children (X2=9.9, p<.01 and X2=7.9, p<.01, respectively).

Table 6-13. Number of Children in the Household (Including Study Child)
			In-Home			Out-of-Home
		TOTAL	Setting
			No Services	Services	TOTAL In-Home	Foster Care	Kinship Foster Care	TOTAL Out-of-Home^
		Mean / (SE)
Age	0-2	2.3 e f (0.1)	2.1 (0.1)	2.2 (0.1)	2.1 k l (0.1)	3.7 (0.3)	2.9 (0.2)	3.3 n (0.2)
	3-5	2.4 g (0.1)	2.4 (0.1)	2.4 (0.1)	2.4 j m (0.1)	3.3 (0.4)	2.4 (0.3)	2.6 (0.2)
	6-10	2.6 (0.1)	2.6 (0.1)	2.8 (0.1)	2.6 (0.1)	3.7 (0.2)	2.3 (0.2)	3.0 (0.2)
	11+	2.7 (0.1)	2.8 (0.1)	2.8 (0.2)	2.8 (0.1)	3.0 (0.3)	2.6 (0.2)	2.7 (0.2)
Race/ Ethnicity	African American	2.7 c (0.1)	2.6 (0.1)	2.7 (0.1)	2.6 h (0.1)	3.4 (0.2)	2.6 (0.2)	3.0 (0.1)
	White	2.4 d (0.1)	2.3 (0.1)	2.4 (0.1)	2.3 i (0.1)	3.7 (0.2)	2.4 (0.2)	2.8 (0.1)
	Hispanic	2.7 (0.1)	2.6 (0.1)	3.0 (0.1)	2.7 (0.1)	3.2 (0.4)	3.6 (0.2)	3.2 (0.3)
	Other	2.6 (0.2)	2.7 (0.2)	2.3 (0.2)	2.6 (0.2)	3.6 (0.4)	1.7 (0.3)	3.0 (0.4)
TOTAL		2.5 (0.04)	2.5 (0.1)	2.6 (0.1)	2.5 a (0.04)	3.5 b (0.2)	2.6 (0.1)	3.0 (0.1)

^ Includes children in “other” out-of-home placement settings, but not children in group care. (back)

a Children in out-of-home placements live in households with more children than children remaining at home (t=-5.4, p<.001). (back)

b Children in nonkinship foster care live in households with more children than children in kinship foster care (t=4.8, p<.001). (back)

c African American children live in households with more children than White children (t=2.6, p≤.01). (back)

d Hispanic children live in households with more children than White children (t=-3.0, p<.01). (back)

e Children aged 6 to 10 years live in households with more children than children 0 to 2 years of age (t=-4.6, p<.001). (back)

f Children aged 11 years and older live in households with more children than children 0 to 2 years of age (t=-4.5, p<.001). (back)

g Children aged 11 years and older live in households with more children than children aged 3 to 5 years (t=-3.7, p =.001). (back)

h African American children remaining at home live in households with more children than White children remaining at home (t=2.8, p<.01). (back)

i Hispanic children remaining at home live in households with more children than White children remaining at home (t=-3.0, p<.01). (back)

j Children aged 3 to 5 years remaining at home live in households with more children than children 0 to 2 years of age remaining at home (t=-2.7, p<.01). (back)

k Children aged 6 to 10 years remaining at home live in households with more children than children 0 to 2 years of age remaining at home (t=-6.8, p<.001). (back)

l Children aged 11 years and older remaining at home live in households with more children than children 0 to 2 years of age remaining at home (t=-5.7, p<.001). (back)

m Children aged 11 years and older remaining at home live in households with more children than children 3 to 5 years of age remaining at home (t=-3.1, p<.01). (back)

n Children 0 to 2 years of age in out-of-home placements live in households with more children than children aged 11 years and older in out-of-home placements (t=2.8, p<.01). (back)

6.3.6 Number of Adults in the Household

Overall, children involved with CWS live in households with an average number of 1.9 adults.25 Overall, African American children live in households with fewer adults than do White or Hispanic children. This is also true within the subpopulation of children in out-of-home care (p≤ .001). Coupled with the finding that they live in households with more children than do White children, it appears that African American children may be more prone to having less adult supervision and support than other children, based solely on household composition. To test the difference, and significance thereof, of the combination of number of children and number of adults in the household between various racial/ethnic groups, a child-to-adult ratio was computed. Overall, the mean child-to-adult ratio for White children (1.4) was significantly lower than the ratio for both African American children (1.9; t=4.8, p<.001) and Hispanic children (1.7; t=-2.9, p<.01). This significant difference is mimicked within the subpopulation of children remaining at home and, for the African American children versus White children comparison, within the subpopulation of children in out-of-home care.

The only other significant difference with regard to the number of adults in the household is between children aged 6 to 10 years and children 0 to 2 years of age. Overall, as well as within the subpopulation of children remaining at home, children aged 6 to 10 years are living with fewer adults than are children 0 to 2 years of age (Table 6-14).

6.3.7 Relationship of Adults in the Household to the Child

In addition to the number of adults in the household, the relationship of these adults to the child is likely to affect development and well-being. The current household compositions of children living at home are presented in Table 6-15. About one-third of the children living at home are living with both parents (or a parent and a step parent), and for the vast majority of these children (83%) there are no other adults in the household. On the other hand, whereas over half of the children living at home are living with their mother (or stepmother) without a father (or stepfather), there is at least one other adult present in 42% of these households. This is also true for children living with their father (or stepfather) only; that is, there is at least one other adult in 45% of these households.

In many cases when parents are unable to care for their children, this responsibility falls to the grandparents. For 3% of the children involved with CWS who are living at home, there are no parents living there and one or more grandparent was the head of the household at the time of the interview (Table 6-15). Yet, overall, there is one or more grandparent living in the household for 13% of the children living at home. Of these households with a grandparent, 82% (11% of all the households) are three-generation households (i.e., a parent [biological, adoptive, or step] is also present).

Recent studies examined the relationship of household composition to child maltreatment and found some evidence for the hypothesis that having an unrelated male adult in the household presents a maltreatment risk. A recent study of maltreatment injury deaths concluded that having unrelated male adults in the household puts children at increased risk for maltreatment injury death (Stiffman et al., 2002). Another follow-up study of extremely premature infants identified “unmarried cohabitation” as a risk factor for child maltreatment (Strathearn et al., 2001).

Table 6-14. Number of Adults in the Household
			In-Home			Out-of-Home
		TOTAL	Setting
			No Services	Services	TOTAL In-Home	Foster Care	Kinship Foster Care	TOTAL Out-of-Home^
		Mean / (SE)
Age	0-2	2.0 c (0.1)	2.0 (0.1)	2.0 (0.1)	2.0 d (0.1)	2.0 (0.1)	2.0 (0.2)	2.0 (0.1)
	3-5	1.9 (0.1)	1.9 (0.1)	1.9 (0.1)	1.9 (0.1)	1.9 (0.2)	1.8 (0.3)	1.9 (0.2)
	6-10	1.8 (0.04)	1.9 (0.1)	1.8 (0.1)	1.8 (0.1)	1.9 (0.1)	1.8 (0.1)	1.9 (0.1)
	11+	1.9 (0.1)	1.9 (0.1)	1.9 (0.1)	1.9 (0.1)	2.2 (0.2)	1.9 (0.2)	2.1 (0.1)
Race/ Ethnicity	African American	1.8 a b (0.1)	1.8 (0.1)	1.8 (0.1)	1.8 (0.1)	1.8 (0.1)	1.7 (0.1)	1.7 e f (0.1)
	White	1.9 (0.03)	1.9 (0.04)	1.9 (0.1)	1.9 (0.04)	2.2 (0.1)	1.9 (0.1)	2.1 (0.1)
	Hispanic	2.0 (0.1)	2.1 (0.2)	1.8 (0.1)	2.0 (0.1)	2.1 (0.1)	2.1 (0.1)	2.2 (0.1)
	Other	2.0 (0.1)	1.9 (0.1)	2.1 (0.4)	2.0 (0.1)	1.8 (0.1)	2.0 (0.3)	2.0 (0.2)
TOTAL		1.9 (0.03)	1.9 (0.1)	1.9 (0.1)	1.9 (0.03)	2.0 (0.1)	1.9 (0.1)	2.0 (0.1)

^ Includes children in “other” out-of-home placement settings, but not children in group care. (back)

a African American children live in households with fewer adults than White children (t=-2.5, p≤ .01). (back)

b African American children live in households with fewer adults than Hispanic children (t=-2.7, p<.01). (back)

c Children aged 6 to 10 years live in households with fewer adults than children 0 to 2 years of age (t=2.8, p<.01). (back)

d Children aged 6 to 10 years remaining at home live in households with fewer adults than children 0 to 2 years of age remaining at home (t=2.6, p≤ .01). (back)

e African American children in out-of-home care live in households with fewer adults than White children in out-of-home care (t=-3.3, p≤ .001). (back)

f African American children in out-of-home care live in households with fewer adults than Hispanic children in out-of-home care (t=-3.9, p<.001). (back)

Although our analysis is not able to determine a relationship between risk of maltreatment and household composition—because household composition data was collected well after the maltreatment allegation was made—we endeavored to describe patterns of caregiving with attention to possible associations to child maltreatment risk.

We looked at cases of children living at home with their biological or adoptive mother as the primary caregiver. In 12% of these cases, the mother indicates being unmarried and living with a boyfriend or partner who is not the child's father. An additional 13% indicate some other male unrelated to the child (including stepfather) living in the household. In total, approximately one-quarter of children involved with CWS and living at home with their mother are also living with a male unrelated to them.26 Logistic regression analyses, which also controlled for child age, race/ethnicity, and gender, were run to examine whether either of these situations (i.e., mother's boyfriend or the broader category of any unrelated male in the household) is associated with receipt of CWS. Results showed no association between either of these situations and receipt of ongoing services for children remaining at home.27

Table 6-15. Adults in the Household of Children Living at Home
Adults in Household			Percent of Total^
Mother a and/or Father b present	Mother and Father	and no other adults	28
		and other related adult(s) only	5
		and other unrelated adult(s) c only	1
		and other related and unrelated adults	.2
		total	34
	Mother only	and no other adults	33
		and other related adult(s) only	12
		and other unrelated adult(s) only	10
		and other related and unrelated adults	1
		total	56
	Father only	and no other adults	3
		and other related adult(s) only	1
		and other unrelated adult(s) only	1
		and other related and unrelated adults	.4
		total	6
	Total		96
No Mother or Father present d	Grandmother and/or Grandfather	and no other adults	1
		and other related adult(s) only	1
		and other unrelated adult(s) only	.2
		and other related and unrelated adults	.3
		total	3
	Aunt and/or Uncle	and no other adults	1
		and other related adult(s) only	.1
		and other unrelated adult(s) only	.1
		and other related and unrelated adults	0
		total	1
	Total		4
TOTAL			100

^ Section and column totals may be affected by rounding. (back)

a Includes biological mother, adoptive mother, and stepmother. (back)

b Includes biological father, adoptive father, and stepfather. (back)

c Includes adult step and foster siblings, babysitter, and other non-related adults. (back)

d These are presented in a hierarchy, such that those in the “grandmother and/or grandfather” category may also have aunts and/or uncles in the household. (back)

6.3.8 Siblings in Out-of-Home Care

When more than one child must be taken into custody by a child welfare agency and placed in an out-of-home care situation, the preferred practice is to place the children together, if at all possible. Keeping siblings together has benefits, including increasing the comfort level for the children, as well as making it easier for the child welfare worker to provide services to the children and their family. According to information from the household rosters collected during the current caregiver interviews, 41% of children in out-of-home care are living with one or more full, half, step, or adoptive sibling. (Note that this is a proportion of all children in out-of-home care, not just those with siblings, as that information is not available through the household roster data.) The proportion of children in nonkinship foster care who live with one or more siblings is not significantly different from the proportion of children in kinship foster care who live with one or more siblings.28 The approximately one-third to one-half of children in nonkinship and kinship foster care who live with one or more sibling is, however, significantly greater than the 4% of children in group care who live with one or more siblings, X2=15.4, p<.01 (Table 6-16).

6.3.9 Discussion of Household Characteristics

Table 6-16. Proportion of Children in Out-of-Home Care Living with One or More Sibling ^
Percent (SE)	Proportion Living with One or More Sibling
Total ^^	41.0 (3.1)
Foster Care	36.3 (3.9)
Kinship Foster Care	51.0 (5.4)
Group Care	3.7 (3.7)

^ Includes full, half, step and adoptive siblings. (back)

^^ Includes children in “other” out-of-home placement settings. (back)

Many children involved with CWS live in impoverished households. Children remaining in the home are more likely to live in poverty compared to children placed outside the home. The substantial proportion of children living below the poverty line is not limited to children living in their biological families—many of the children in kinship foster care and foster care are experiencing economic deprivation. Total household incomes are small, with many families living on less than $25,000 per year. A better indicator of family financial circumstances is poverty level—children involved with the child welfare system are five times more likely to have income at only 50% of the poverty level compared to families in the general population. Although the many families that become involved with CWS are not poor, exposure to economic deprivation does represent a risk for future maltreatment and a developmental threat to children.

Overall, the number of children living in a household investigated for child maltreatment is not exceedingly high, but there are a sizable proportion of youths living in homes with four or more children—especially among children living in out-of-home care. Older children are more likely to be living in homes with more children, as are Hispanic and African American children. Families of these children who are caring for four or more children are likely to have greater service and income support needs. This group includes many children living in foster care. In summary, a large proportion of families involved with the child welfare system are living at or below the poverty level. The consequent lack of financial resources and related stress makes adequate care of children a substantial challenge. Many biological and foster families who have a large number of children in the household face an additional challenge.

6.4 Children's Living Environment

Almost nothing is known about the home environments of biological families who retain the care of their children without ongoing services, following an investigation by CWS. This study provides a unique opportunity to gain a detailed description of in-home and out-of-home environments. To understand the caregiving environment for CWS-involved children, HOME-SF scores were computed separately for each of three age groups: less than 3 years old, 3 to 5 years old, and 6 to 10 years old. The HOME-SF score provides information about the caregiving environment and is also a reasonably good predictor of future performance on academic and social indicators (Bradley et al., 2001; Moore, Halle & Mariner, 2000). Cognitive stimulation, emotional support, and physical environment subscale scores were also determined. The total HOME-SF score comprises the cognitive stimulation and emotional support scores. The scores are calculated by summing the number of questions answered yes. Therefore, in each of the four scales, a higher score indicates the presence of more positive characteristics in the home environment. The administration of the HOME-SF is slightly different than what was used in the NLSY, but we believe that the comparability is quite high. These slight differences include changing the NLSY references from “Mother” to “Parent/Guardian” and placing the reference period (e.g., in the past week) at the beginning of the question rather than at the end.

Note that for the analyses in this section, HOME-SF scores are examined for children living at home, in nonkinship or kinship foster care only. Children in group homes and “other” out-of-home settings were excluded because they may not have lived in that setting for very long, and it is not clear what conclusions can be drawn from analysis of these environments.

This section begins with a brief discussion of the items that constitute the HOME-SF and the differences between children living at home and children living in nonkinship or kinship foster care. Next is a discussion of the differences in the total HOME-SF score for all children (birth to 10 years old) by placement type and the differences in each score by age group. The section continues with a discussion of differences in-home-SF scores by child race/ethnicity. The relationship between HOME-SF scores and caregiver age and race/ethnicity is then considered. Brief discussions of scores on a punitiveness subscale and analysis of the neighborhood context follows. Finally, the section ends with our conclusions.

6.4.1 HOME-SF Items

Among children birth to 2 years, children living out of the home (nonkinship or kinship foster care) were more likely to experience a number of positive aspects of the home environment than children living at home (Table 6-17). For example, children birth to 2 years living out of the home received more physical affection and were less likely to be restricted from exploring by their caregivers than children living at home. In addition, children birth to 2 years living out of the home were more likely than children living at home to have caregivers who kept the child in close view during the interview, who spoke with a distinct and audible voice, and who conversed freely with the interviewer. Finally, children from birth to 2 years in out-of-home care were more likely to have a safe play environment than children living at home.

Table 6-17. The HOME-SF Observational Items, Proportion of Questions Answered Yes, 0-2 Years Old
	TOTAL	Total In-Home	Total Out-of-Home
	Setting
	Percent / (SE)
M/G respond positively to praise of child	97.7 (0.7)	97.3 a (0.9)	99.7 (0.2)
M/G spoke to child 2 times or more	94.0 (1.4)	93.4 (1.6)	96.2 (1.5)
M/G tells child name of object or person	66.1 (3.3)	67.1 (3.5)	61.0 (5.2)
M/G responds verbally to child's speech	88.0 (2.1)	88.1 (2.2)	87.1 (3.3)
M/G spontaneously praises child at least twice	83.1 (2.9)	82.8 (3.2)	84.7 (5.5)
M/G caressed, kissed or hugged child at least once	92.7 (1.3)	91.7 b (1.5)	97.5 (1.5)
M/G does not shout at child	56.7 (4.1)	55.7 (4.5)	61.6 (6.7)
M/G does not express annoyance or hostility toward child	56.1 (3.9)	55.5 (4.2)	59.4 (7.1)
M/G does not scold or criticize child during visit	54.1 (3.8)	54.2 (4.0)	53.3 (6.5)
M/G slapped or spanked child	6.5 (1.6)	6.4 (1.7)	6.8 (5.1)
M/G physically restricted child more than 3 times	25.0 (2.8)	27.2 c (3.1)	13.4 (3.3)
M/G provided toys or interesting activities for child	76.3 (2.4)	75.1 (2.7)	82.5 (3.6)
M/G kept child in view/looked at child often	94.6 (1.1)	93.8 d (1.3)	98.6 (0.4)
M/G's speech is distinct and audible	93.9 (1.1)	92.9 e (1.3)	98.7 (0.5)
M/G initiates verbal exchanges with visitor	91.9 (1.4)	92.1 (1.3)	90.8 (4.9)
M/G converses freely and easily	93.2 (1.4)	92.0 f (1.7)	99.6 (0.3)
M/G's voice conveys positive feelings toward child	97.9 (0.6)	97.8 (0.7)	98.4 (1.4)
At least 10 books present and visible	49.6 (3.7)	48.9 (3.8)	53.3 (7.3)
Child's play environment is safe	90.4 (1.4)	89.1 g (1.6)	97.0 (1.0)

Note: M/G = mother/guardian. Children in group homes and other out-of-home placements were excluded. (back)

a M/G caring for children in out-of-home care are more likely to respond positively to praise of child than M/Gs caring for children living at home (X2=6.8, p<.01). (back)

b M/G caring for children living in out-of-home care are more likely to caress, kiss or hug their child than M/Gs caring for children living at home (X2=7.0, p<.01). (back)

c M/G caring for children living in out-of-home care are less likely to restrict their child's actions than M/Gs caring for children living at home (X2=7.1, p<.01). (back)

d M/G caring for children living in out-of-home care are more likely to keep the child in view or look at her often than M/Gs caring for children living at home (X2=7.8, p<.01). (back)

e M/G caring for children living in out-of-home care are more likely to speak with a distinct and audible voice than M/Gs caring for children living at home (X2=12.1, p<.001). (back)

f M/G caring for children living in out-of-home care are more likely to converse freely and easily than M/Gs caring for children living at home (X2=12.0, p<.001). (back)

g Children living in out-of-home care are more likely to have a safe play environment than children living at home (X2=9.7, p<.01). (back)

Based on caregiver responses, children aged 2 and under who live at home are more likely to have three or more books of their own, to have at least one push toy, and to have been spanked two or more times in the past week than children who live out of the home (Table 6-18).

Several differences exist in the home environments of children at home and in out-of-home care among 3- to 5-year-olds (Table 6-19). Children living out of the home were more likely than children living at home to live in an aesthetically pleasing neighborhood, have the rooms visible to the interviewer judged to be reasonably clean, and have the interviewer introduced by name. Children living in out-of-home care were less likely than children living at home to live in a dark or monotonous home and less likely to be scolded more than once.

Only one difference among 3- to 5-year-old children emerged from the caregiver responses. Children living in out-of-home care were less likely to be spanked frequently than children living at home (Table 6-20).

Table 6-18. The HOME-SF Interviewer Items, Proportion of Questions Answered Yes, 0-2 Years Old
	TOTAL	Total In-Home	Total Out-of-Home
	Setting
	Percent / (SE)
Child gets out of house 4 or more times per week	72.3 (2.2)	72.4 (2.4)	72.0 (3.3)
Child has 3 or more books of his/her own	80.0 (1.9)	84.5 a (1.8)	57.0 (5.0)
Reads stories to child 3 or more times per week	61.1 (2.3)	62.4 (2.5)	54.3 (6.3)
Caregiver takes child to grocery store 2 or more times per week	38.7 (2.2)	38.3 (2.5)	40.5 (5.6)
Child has at least 1 cuddly toy	98.8 (0.3)	98.8 (0.4)	98.5 (0.8)
Child has at least 1 push toy	83.3 (1.6)	85.5 b (1.6)	71.9 (33)
Caregiver believes parents should spend time teaching child	93.9 (1.1)	93.8 (1.2)	94.3 (1.7)
Child eats meal with mother and father/father-figure 1 or more times per day	49.6 (2.6)	48.9 (2.8)	53.8 (4.9)
Talks to child often or always while working	85.4 (1.9)	85.6 (2.2)	84.6 (3.0)
Spanked child 2 or more times in the past week	17.1 (2.2)	19.5 c (2.6)	4.4 (2.0)

Note: M/G is mother/guardian. Children in group homes and other out-of-home placements were excluded.

a Children living at home are more likely to have 3 or more books of their own than children living in out-of-home care (X2=13.9, p≤ .001). (back)

b Children living at home are more likely to have at least one push toy than children living in out-of-home care (X2=8.7, p<.01). (back)

c Caregivers of children living at home are more likely to have been spanked 2 or more times in the past week than children living in out-of-home care (X2=11.8, p<.001). (back)

Table 6-19. The HOME-SF Observational Items, Proportion of Questions Answered Yes, 3-5 Years Old
	TOTAL	Total In-Home	Total Out-of-Home
	TOTAL	Setting
	Percent / (SE)
M/G spoke to child 2 times or more	94.3 (1.4)	94.2 (1.6)	94.8 (3.0)
M/G responds verbally to child's speech	97.0 (1.1)	97.0 (1.2)	97.4 (2.5)
M/G spontaneously praises child at least twice	80.1 (3.4)	79.6 (3.7)	88.4 (4.7)
M/G caressed, kissed or hugged child at least once	77.2 (3.7)	77.4 (4.0)	75.1 (7.7)
M/G does not scold or criticize child during visit more than once	52.0 (4.9)	50.9 a (5.1)	73.1 (4.9)
M/G slapped or spanked child	3.0 (0.7)	3.0 (0.7)	3.5 (1.8)
M/G physically restricted child	5.4 (1.1)	5.4 (1.1)	7.3 (5.5)
M/G's voice conveys positive feelings toward child	95.4 (1.4)	95.4 (1.5)	95.7 (2.5)
At least 10 books present and visible	52.8 (5.1)	51.3 (5.2)	73.3 (9.9)
Child's play environment is safe	86.5 (2.1)	86.2 (2.2)	90.7 (4.7)
M/G usually answered child's questions/requests verbally	97.4 (1.1)	97.4 (1.2)	98.1 (1.5)
M/G helps child demonstrate some achievement	60.0 (4.1)	60.3 (4.3)	55.3 (10.5)
M/G introduces interviewer to child by name	56.2 (4.4)	54.8 b (4.5)	79.4 (6.0)
M/G used correct grammar and pronunciation	91.6 (1.7)	91.3 (1.8)	97.1 (2.4)
M/G used complex sentence structure and vocabulary	77.4 (3.2)	76.5 (3.5)	92.7 (2.8)
Child's artwork is displayed in the house	49.0 (4.8)	47.9 (5.0)	65.3 (11.6)
Building appears safe	92.7 (1.6)	92.6 (1.7)	94.2 (3.8)
Neighborhood is aesthetically pleasing	80.7 (2.7)	79.8 c (2.8)	96.4 (1.6)
100 sq/ft of living space per person	83.8 (3.1)	83.4 (3.1)	90.4 (4.3)
Interior of home is dark or perceptually monotonous	17.5 (2.5)	18.3 d (2.6)	3.7 (2.6)
All visible rooms are reasonably clean	83.3 (2.6)	82.6 e (2.7)	95.6 (2.7)
All visible rooms are minimally cluttered	74.2 (3.4)	73.9 (3.5)	80.2 (5.9)
Rooms are not overcrowded with furniture	60.8 (4.3)	60.6 (4.3)	66.1 (7.9)

Note: M/G = mother/guardian. Children in Group Homes and other out-of-home placements were excluded.

a M/Gs caring for children living in out-of-home care are more likely to not scold the child more than once than mother/guardian's caring for children living in in-home care (X2=7.4, p<.01). (back)

b M/Gs caring for children living in out-of-home care are more likely to introduce the interviewer to the child by name than M/Gs caring for children living in in-home care (X2=7.3, p<.01). (back)

c Children living in out-of-home care are more likely to live in an aesthetically pleasing neighborhood than children living at home (X2=15.2, p<.001). (back)

d Children living in out-of-home care are less likely to live in a dark or perceptually monotonous home than children living at home (X2=10.4, p<.01). (back)

e Homes of children living in out-of-home care are more likely to have reasonably clean rooms than homes of children living at home (X2=7.7, p<.01). (back)

Table 6-20. The HOME-SF Interviewer Items, Proportion of Questions Answered Yes, 3-5 Years Old
	TOTAL	Total In-Home	Total Out-of-Home
	TOTAL	Setting
	Percent / (SE)
Child eats meal with mother and father/father-figure 1 or more times per day.	46.6 (3.5)	46.3 (3.5)	52.8 (13.8)
Child has at least 10 books	76.3 (3.4)	76.5 (3.6)	71.4 (8.1)
Family gets at least one magazine	55.7 (3.6)	54.4 (3.7)	78.2 (5.8)
Child has use of record or CD player or tape deck and 5 or more children's records, tapes or CDs	62.4 (3.0)	62.1 (3.2)	67.2 (7.8)
Child has some or a great deal of choice over what foods she/he eats	83.7 (2.2)	83.4 (2.3)	87.8 (3.2)
Television is on 4 hours or less per day	37.5 (2.8)	36.2 (3.1)	61.1 (9.3)
Family member takes child on outing 2 or more times per month	80.2 (2.9)	79.5 (3.1)	91.6 (3.5)
Family member has taken child to museum at least once in the past year	48.1 (3.8)	48.2 (4.0)	46.5 (10.8)
M/G spanked child less than 2 times in the last week	80.3 (2.8)	79.5 a (2.9)	93.8 (3.9)
M/G, other adult, or older child has helped child learn numbers at home	93.6 (1.7)	93.7 (1.8)	95.3 (2.5)
M/G, other adult, or older child has helped child learn the alphabet at home	90.6 (1.8)	90.8 (2.0)	86.6 (5.3)
M/G, other adult, or older child has helped child learn colors at home	91.9 (1.9)	91.8 (2.0)	94.1 (2.5)
M/G, other adult, or older child has helped child learn shapes and sizes at home	83.0 (2.2)	83.7 (2.3)	71.3 (7.8)

Note: M/G = mother/guardian. Children in group homes and other out-of-home placements were excluded.

a Caregivers of children living out-of-home are more likely to spank their children less than 2 times in the past week than caregivers of children living at home (X2= 9.2, p<.01). (back)

There are few differences between in-home and out-of-home placements among children between 6 and 10 years old. Those in out-of-home care are more likely than children who live at home to have the interviewer introduced by name and more likely to live in homes that have at least 100 square feet of space per person (Table 6-21).

Finally, children living in out-of-home care are more likely to be read to 3 or more times per week and to be expected to make their bed than children living at home (Table 6-22). In addition, families of children who live out of the home are more likely to receive a daily newspaper than families of children who live at home. Caregivers of children who live outside the home are more likely to do something other than sending the child to his/her room or grounding, spanking, talking, giving a chore, or ignoring the child in response to a child's temper tantrum.

6.4.2 An Overview of the HOME-SF Scores

Comparison of NSCAW HOME-SF scores with findings from the NLSY showed that CWS-involved children are more likely to have a low score and less likely to have a high score than children in the NLSY (Center for Human Resource Research, 2000) (Table 6-23). For instance, 26% of CWS-involved children from birth to 2 years had total HOME-SF scores between 0 and 12.9, compared with only 10% of young children in the NLSY. In addition, only 28% of the youngest CWS-involved children received high scores (16–18.9), though almost half of the youngest children in NLSY received high scores. These trends are also apparent among older children.

Table 6-21. The HOME-SF Observational Items, Proportion of Questions Answered Yes, 6-10 Years Old
	TOTAL	Total In-Home	Total Out-of-Home
	TOTAL	Setting
	Percent / (SE)
M/G respond positively to praise of child	93.1 (1.6)	92.8 (1.7)	96.8 (1.6)
M/G does not express annoyance or hostility toward child	55.7 (4.7)	53.8 (5.1)	75.4 (6.2)
M/G initiates verbal exchanges with visitor	93.4 (1.1)	93.1 (1.3)	96.7 (1.6)
M/G's voice conveys positive feelings toward child	94.9 (1.1)	94.6 (1.2)	97.3 (1.0)
Child's play environment is safe	86.3 (2.4)	86.1 (2.5)	87.4 (4.5)
M/G usually answered child's questions/requests verbally	95.5 (1.3)	95.2 (1.4)	98.4 (0.8)
M/G introduces interviewer to child by name	59.4 (3.5)	57.5 a (3.8)	79.0 (5.7)
M/G used complex sentence structure and vocabulary	84.2 (2.5)	83.4 (2.7)	92.4 (3.4)
Building appears safe	53.1 (4.4)	51.6 (4.8)	67.5 (5.4)
100 sq/ft of living space per person	83.2 (2.0)	82.3 b (2.1)	92.6 (2.3)
Interior of home is dark or perceptually monotonous	15.1 (1.7)	15.7 (1.8)	8.5 (2.8)
All visible rooms are reasonably clean	85.4 (2.1)	84.7 (2.4)	92.7 (2.8)
All visible rooms are minimally cluttered	71.5 (2.8)	71.0 (3.0)	76.3 (3.3)
Rooms are not overcrowded with furniture	57.0 (3.2)	55.9 (3.4)	68.5 (5.4)
M/G encouraged child to contribute to conversation	81.6 (2.4)	81.0 (2.5)	87.4 (4.6)
M/G conversed with child	93.9 (1.2)	93.5 (1.3)	97.7 (1.0)
M/G used some term of endearment for the child at least twice	61.8 (3.5)	60.8 (3.6)	70.8 (5.9)
M/G does not violate the rules of common courtesy	61.5 (4.2)	60.1 (4.5)	75.6 (5.8)
House has at least two pictures or other art on the walls	85.4 (2.3)	84.8 (2.4)	91.1 (4.0)
Child's room has a wall decoration appealing to children	78.9 (5.6)	77.8 (6.6)	88.8 (7.7)
House is not overly noisy	50.6 (3.8)	49.5 (4.1)	60.8 (6.2)

Note: M/G = mother/guardian. Children in group homes and other out-of-home placements were excluded.

a M/Gs of children living in out-of-home care are more likely to introduce the interviewer to the child by name than M/Gs of children living at home (X2=6.8, p ≤ .01). (back)

b Children living out of the home are more likely to have 100 sq/ft per person than children living at home (X2=8.1, p ≤ .01). (back)

Table 6-22. The HOME-SF Interviewer Items, Proportion of Questions Answered Yes, 6-10 Years Old
		TOTAL	Total In-Home	Total Out-of-Home
		TOTAL	Setting
		Percent / (SE)
M/G reads stories to child 3 or more times per week		40.1 (3.1)	38.8 a (3.2)	61.4 (5.3)
Child eats meal with mother and father/father-figure 1 or more times per day		34.2 (2.6)	33.7 (2.7)	39.0 (7.8)
Child has at least 10 books		83.1 (2.3)	83.9 (2.4)	74.5 (3.3)
M/G spanked child less than 2 times in the last week		92.8 (1.4)	92.6 (1.4)	94.2 (4.6)
Child is expected at least some of the time to make his/her own bed		80.5 (1.8)	79.3 b (2.1)	92.5 (1.8)
Clean his/her room		92.5 (1.9)	92.2 (2.1)	95.6 (1.4)
Clean up his/her spills		95.0 (1.2)	94.9 (1.3)	96.2 (1.1)
Bathe himself/herself		97.5 (0.7)	97.5 (0.7)	97.8 (0.8)
Pick up after himself/herself		97.3 (0.8)	97.2 (0.9)	98.5 (0.6)
Family has taken child to museum 1 or more times in the past year		60.7 (2.0)	60.3 (2.1)	64.8 (4.7)
Family member has taken child to musical or theatrical performance at least once in the past year		46.4 (2.8)	46.4 (3.0)	46.5 (6.1)
Whole family gets together with relatives or friends 2 or more times per month		45.4 (2.7)	44.1 (3.1)	58.0 (5.7)
M/G or father/father-figure discusses television programs with child		75.3 (1.7)	75.5 (1.7)	73.6 (6.9)
Child spends 4 or more hours with father/father-figure per week		52.7 (2.9)	53.7 (2.9)	42.9 (7.7)
Child spends time in outdoor activities with father/father-figure 1 or more times per week		54.3 (2.7)	54.6 (2.9)	51.0 (5.9)
If child said, "I hate you" or swore in a temper tantrum, caregiver would	Ground child	34.8 (2.7)	34.8 (2.8)	34.4 (5.7)
	Spank child	10.9 (1.8)	11.1 (1.9)	9.0 (4.9)
	Talk with child	70.9 (2.4)	70.6 (2.6)	73.7 (3.8)
	Give child a household chore	17.9 (2.7)	18.2 (3.0)	15.0 (3.5)
	Ignore child	10.0 (1.3)	9.9 (1.3)	11.9 (3.1)
	Send child to room	38.2 (2.7)	38.5 (2.8)	35.3 (5.3)
	Do something else	10.2 (2.0)	18.2 c (2.0)	39.8 (5.8)
Child reads for enjoyment several times or more per week		69.5 (2.3)	70.1 (2.5)	63.7 (4.5)
Child has use of a musical instrument at home		35.1 (2.6)	34.5 (2.8)	40.2 (6.0)
Family gets a daily newspaper		27.9 (2.1)	26.2 d (2.1)	45.5 (7.5)
Family encourages child to start or keep up hobbies 6-10 years		92.9 (1.6)	92.7 (1.7)	95.5 (1.3)
Child participates in arts or sports activities		39.9 (2.3)	40.2 (2.6)	37.4 (5.5)

Note: M/G = mother/guardian. Children in group homes and other out-of-home placements were excluded.

a M/Gs of children living out of the home are more likely to read stories to the child 3 or more times per week than M/Gs of children living at home (X2=9.4, p<.01). (back)

b Children living out of the home are more likely to be expected to make his/her own bed at least some of the time than children living at home (X2=8.8, p<.01). (back)

c M/Gs of children living out of the home are more likely to “do something else” in response to a child's temper tantrum than M/Gs of children living at home (X2=9.7, p<.01). (back)

d Foster families of children living out of the home are more likely to receive a daily newspaper than families of children living at home (X2=6.2, p≤.01). (back)

Table 6-23. Total HOME-SF Scores from NSCAW and NLSY
		0-12.9	13-15.9	16-18.9	19-22.9	23-28
		Range of Total HOME-SF Scores
0-2 years	NSCAW	26	46	28	--	--
0-2 years	NLSY	10	42	49	--	--
3-5 years	NSCAW	4	11	25	40	20
3-5 years	NLSY	2	4	10	34	51
6-10 years	NSCAW	8	16	31	38	7
6-10 years	NLSY	2	6	17	38	37

Mean HOME-SF scores are shown by child age in Table 6-24. To create the raw score (shown below), interviewer and observational items were added together. The number of items in the scales varies by age group, so comparisons between ages cannot be made using the raw scores (comparisons using standardized scores follow).

6.4.3 Race/Ethnicity and the HOME-SF

As we have in other sections of the report, we broke out the HOME-SF scores by child age, CWS setting, and race/ethnicity for multivariate analyses. But using this approach is somewhat more complex with regard to the HOME-SF because the characteristics of the home environment and parent-to-child interactions—as critical as they are to children's development across cultural groups are very difficult to measure consistently and accurately. Because the HOME-SF requires interpretation of features of the home and of the interaction between caregiver and child, the scores are more subjective than other standardized measures used in NSCAW. There is particular concern over the ability of the HOME-SF to validly measure the home environments of children of color (Bradley et al., 2001). Parents may value certain types of parenting behavior more or less depending on their cultural identity and familial history, although the total score seems to be quite robust across family structure, family status, and child outcome measures across cultures (Bradley, Corwyn, & Whiteside-Mansell, 1996). Socioeconomic status may also contribute substantially to the characteristics of the home environment. Because socioeconomic status and race/ethnicity are confounded (McLoyd, 1998), the interpretation of HOME-SF scores according to race/ethnicity must always be made with the understanding that any differences that occur cannot disentangle the fact that families of color are, generally, poorer than White families. In our study, this confounding element may be lessened because so many of the families, of all races/ethnicities, are poor.

Table 6-24. Mean HOME-SF Scores by Child Age
Age	Total	Cognitive Stimulation	Emotional Support	Physical Environment^
0-2 years	13.7 (0.2)	6.8 (0.1)	6.9 (0.1)	--
3-5 years	19.2 (0.2)	10.7 (0.2)	8.5 (0.1)	6.2 (0.1)
6-10 years	17.8 (0.2)	8.5 (0.1)	9.2 (0.2)	6.4 (0.1)

^ There is no physical environment score for children less than 3 years old. (back)

There are two ways that the HOME-SF could introduce systematic bias into the scores. The first would occur if the HOME-SF itself is a biased instrument. Sugland and colleagues (1995) examined differences in the internal consistency of the subscales, the underlying structure or the patterns of prediction by the HOME-SF, and the HOME Inventory for 3- to 5-year-old children in White, African American, and Hispanic subgroups. Factor analyses for the entire sample and each racial/ethnic group revealed strong similarities across groups and revealed that levels of internal consistency of the HOME-SF parenting subscales were comparable across racial/ethnic groups. HOME-SF subscale scores, however, predicted cognitive child outcomes (measured by the Peabody Individual Achievement Test) and socioemotional outcomes (measured by the Behavior Problems Index) significantly better for White children than for children of color and significantly better for cognitive child outcomes than for socioemotional outcomes.

Another way that the HOME-SF might provide systematically biased results that disadvantage the scores of children of color is if the interviewer is introducing bias. That is, the interviewer's understanding of the various ways that emotional support and cognitive stimulation may be provided could influence results. Specifically, the race/ethnicity of the interviewer may be correlated with the HOME-SF scores. To determine whether the interviewer's race/ethnicity might be correlated with HOME-SF scores, we did a simple regression of the interviewer's race/ethnicity on the total, cognitive stimulation, and emotional support scores. Total scores recorded by African American interviewers were significantly lower than total scores recorded by White interviewers (t=-2.7, p≤.01). Race/ethnicity of the interviewer did not significantly affect the cognitive stimulation or emotional support scores.

We also conducted a simple regression of the interviewer's race/ethnicity on the standardized total score separately for African American, White, and Hispanic children. These analyses showed that African American interviewers scored African American children lower than they scored White children (t=-2.9, p<.01). This may have a substantial effect on the HOME-SF scores, since African American interviewers interviewed 31% of African American children. To compensate for this potential bias, we have controlled for interviewer race/ethnicity in the multivariate analyses that follow.

The discussion above certainly does not exhaustively treat the issue of the HOME-SF and its ability to measure the home environment of American children of all races/ethnicities. It does provide a brief analysis of how the HOME-SF could provide biased results and addresses these concerns to a limited extent. The following information regarding differences in the HOME-SF scores by the race/ethnicity of the child should be interpreted in the context of the discussion above, and the reader should be aware of the concern that the HOME-SF could provide biased results across some racial/ethnic groups. Comparisons of HOME-SF scores that include households from diverse racial/ethnic groups should be made cautiously.

A final issue in interpreting the HOME-SF has to do with the scaling. In these analyses we interpret the values as simple linear functions; that is, a higher score is better. This is consistent with scores of studies that run simple correlations between HOME-SF scores and behavioral or cognitive outcomes and show a strong linear relationship. This does not mean, of course, that there is not some threshold for each child at which the provision of a lower level of cognitive stimulation or emotional responsive results in no significant difference in outcome. Knowing that two groups differ significantly in their mean score does not necessarily mean that the lower score will result in significantly worse development for that child. At the same time, two homes could have nonsignificantly different qualities (as measured by the HOME-SF) but have a significantly different impact on a child, depending on the child's needs (and, of course, unmeasured qualities). We have not tested for these threshold effects, but we believe an appropriate approach is to treat the scores that are significantly higher as likely to have an important meaning for children.

6.4.4 Child Welfare Services Setting and the HOME-SF

Cognitive stimulation, emotional support, physical environment, and total scores by CWS setting are shown in Table 6-25. Because there are a different number of items in the scales depending on the age group, standardized scores were created for the purpose of making comparisons across age groups. These scores were calculated by dividing the total score by the number of questions answered. (Note that although the mean scores displayed in Table 6-25 are not standardized, allowing comparison to other studies using the HOME-SF, the differences indicated in the footnotes to the tables are based on comparisons of the standardized scores.)

Out-of-home environments received higher total and emotional support scores than in-home environments for children from birth to age 10 (Table 6-25). In addition, the out-of-home environments had higher physical environment scores than the in-home environments for all children between 3 and 10 years old.

Table 6-25. Total HOME-SF and Component Scales by Age
		Total	CS	ES	PE^	Total	CS	ES	PE^
		In-Home				Out-of-Home
		Mean(SE)
Age	0-2	13.7 d e (0.2)	6.8 i (0.1)	6.7 m n (0.1)	--	13.7 g (0.3)	6.5 k (0.2)	7.2 p (0.1)	--
	3-5	19.2 f (0.2)	10.6 j (0.2)	8.5 o (0.2)	6.2 c s (0.1)	20.4 h (0.5)	11.4 l (0.5)	9.0 q (0.2)	6.9 (0.2)
	6-10	17.7 (0.2)	8.5 (0.1)	9.2 (0.2)	6.4 t (0.2)	18.6 (0.6)	9.1 (0.4)	9.4 (0.4)	7.0 (0.4)
Total*		17.1 a (0.2)	8.7 (0.1)	8.4 b (0.1)	6.3 (0.2)	16.7 (0.4)	8.3 (0.2)	8.4 (0.2)	7.0 r (0.3)

Note: CS is Cognitive Stimulation, ES = Emotional Support, and PE = Physical Environment. Children in group homes and other out-of-home settings were excluded.

^ There is no PE scale for children aged 0 to 2 years. (back)

* Standardized scores for the Total, CS and ES scales were used for significance testing. (back)

a Total scores for out-of-home environments of children 0 to 10 were higher than total scores for in-home environments of children aged 0 to 10 years (t=-5.0, p ≤ .001). (back)

b ES scores for out-of-home environments of children aged 0 to 10 years were higher than emotional support scores for in-home environments of children aged 0 to 10 years (t=-6.1, p ≤ .001). (back)

c PE scores for out-of-home environments of children aged 3 to 5 years were higher than scores for in-home environments of children aged 3 to 5 years (t=3.6, p ≤ .01). (back)

d Total scores for in-home environments of children aged 0 to 2 years were higher than total scores for in-home environments of children aged 3 to 5 years (t=2.8, p ≤ .01). (back)

e Total scores for in-home environments of children aged 0 to 2 years were higher than total scores for in-home environments of children aged 6 to 10 years (t=24.9, p ≤ .001). (back)

f Total scores for in-home environments of children aged 3 to 5 years were higher than total scores of in-home environments of children aged 6 to 10 years (t=18.2, p ≤ .001). (back)

g Total scores for out-of-home environments of children aged 0 to 2 years were higher than total scores of out-of-home environments of children aged 6 to 10 years (t=9.0, p ≤ .001). (back)

h Total scores for out-of-home environments of children aged 3 to 5 years were higher than total scores of out-of-home environments of children aged 6 to 10 years (t=6.1, p ≤ .001). (back)

i CS scores for in-home environments of children aged 0 to 2 years were higher than scores of in-home environments of children aged 6 to 10 years (t=13.7, p ≤ .001). (back)

j CS scores for in-home environments of children aged 3 to 5 years were higher than scores of in-home environments of children aged 6 to 10 years (t =11.6, p ≤ .001). (back)

k CS scores for out-of-home environments of children aged 0 to 2 years were lower than scores for out-of-home environments of children aged 3 to 5 years (t=-2.9, p ≤ .01). (back)

l CS scores for out-of-home environments of children aged 3 to 5 years were higher than scores for out-of-home environments of children aged 6 to 10 years (t=3.7, p ≤ .001). (back)

m ES scores for in-home environments of children aged 0 to 2 years were higher than scores for in-home environments of children aged 3 to 5 years (t=6.5, p ≤ .001). (back)

n ES scores for in-home environments of children aged 0 to 2 years were higher than scores for in-home environments of children aged 6 to 10 years (t=24.8, p ≤ .001). (back)

o ES scores for in-home environments of children aged 3 to 5 were higher than scores for in-home environments of children aged 6 to 10 years (t=16.2, p ≤ .001). (back)

p ES scores for out-of-home environments of children aged 0 to 2 years were higher than scores for out-of-home environments of children aged 6 to 10 years (t=15.2, p ≤ .001). (back)

q ES scores for out-of-home environments of children aged 3 to 5 years were higher than scores for out-of-home environments of children aged 6 to 10 years (t=8.9, p ≤ .001). (back)

r PE scores for out-of-home environments of children aged 3 to 10 years were higher than scores for in-home environments of children aged 3 to 10 years (t=3.0, p<.01). (back)

s PE scores for in-home environments of children aged 3 to 5 years were higher than scores for in-home environments of children aged 6 to 10 years (t=2.7, p<.01). (back)

t PE scores for out-of-home environments of children aged 6 to 10 years were higher than scores for in-home environments of children aged 6 to 10 years (t=2.7, p<.01). (back)

6.4.5 Age of the Child and the HOME-SF

Comparisons across age groups showed that the home environments of younger children had higher HOME-SF scores than the home environments of older children, regardless of CWS setting (Table 6-25). The age-related findings are summarized in Table 6-26, where the age group with the statistically higher score is listed. For instance, children from birth to 2 years have significantly higher total scores than children between 6 and 10 years old.

Table 6-26. Significant Differences In-Home-SF Scores between Child Age Groups
Age	Total	CS	ES
0-2 vs. 3-5			0-2
0-2 vs. 6-10	0-2	0-2	0-2
3-5 vs. 6-10	3-5	3-5	3-5

Note: CS is Cognitive Stimulation, ES = Emotional Support.

Multivariate analysis of the total scores showed that child age, child race/ethnicity, CWS setting, household income, and caregiver education all significantly affect the score (Table 6-27). More specifically, the results indicate that the home environments of younger children have significantly higher scores than the home environments of 6- to 10-year-olds, and the home environments of African American and Hispanic children have significantly lower scores than the home environments of White children. In addition, in-home environments of children who received services had significantly lower total scores than in-home environments of children who did not receive services. Home environments where the household income was $50,000 or more received significantly higher home scores than home environments where the household income was below $25,000. Home environments where caregivers had some type of degree had significantly higher total scores than the home environments where caregivers had no degree. Finally, the home environments of children whose interviewers were of other races/ethnicities received lower scores than the home environments of children whose interviewers were White.

Multivariate analysis of the cognitive stimulation score shows that child age, child race/ethnicity, CWS setting, household income, and caregiver education affect the score (Table 6-28). As we found in the multivariate analysis of the total score, being young, having a higher household income, and more caregiver education are associated with higher cognitive stimulation scores. Unlike the analysis on the total scores, the interviewer race/ethnicity did not have a significant effect on cognitive stimulation scores.

A multivariate analysis of emotional support scores showed that child age, child race/ethnicity, CWS setting, and caregiver education all have a significant relationship to the score (Table 6-29). The race/ethnicity of the interviewer does not have a significant relationship to the emotional support score.

Finally, a multivariate analysis was conducted on the physical environment score, where child age, child race/ethnicity, income, caregiver education, and interviewer race/ethnicity were found to have a significant relationship (Table 6-30). Home environments of children between 3 and 5 years received higher scores than home environments of children between 6 and 10 years (there is no physical environment scale for children 0–2 years). Home environments of children of other races/ethnicities received higher physical environment scores than homes of White children. The physical environments of children whose caregivers' income was between $25,000 and $49,999 received higher scores than the physical environments of children whose caregivers' income was below $25,000. Home environments with caregivers who received a Bachelor's degree or more received higher scores than home environments with caregivers who received no degree. Children who were interviewed by persons of other races/ethnicities received lower scores than children whose interviewers were White.

Table 6-27. Regression Results for Explaining Total HOME-SF Score^
		Beta Coefficient (SE)
Child Age	0-2 years	0.41 (.02) a
	3-5 years	0.36 (.02) b
	6-10 years	(reference group)
Child Race/Ethnicity	African American	-0.07 (.02) c
	White	(reference group)
	Hispanic	-0.03 (.02) d
	Other	0.01 (.03)
Setting	No Services	(reference group)
	Services	-0.04 (.01) e
	Foster Care	-0.00 (.03)
	Kinship Foster Care	0.05 (.03)
Child Gender	Male	-0.00 (.01)
Child Gender	Female	(reference group)
Household Income	< $25,000	(reference group)
	$25,000 to $49,999	0.04 (.02)
	≥ $50,000	0.07 (.03) f
Caregiver Education	No Degree	(reference group)
	High School Diploma/GED	0.05 (.02) g
	Associate's or Vocational	0.10 (.02) h
	Bachelor's Degree or Higher	0.17 (.03) i
	Other	0.07 (.05)
Interviewer Race/ Ethnicity	African American	-0.03 (.02)
	White	(reference group)
	Hispanic	0.03 (.03)
	Other	-0.06 (.02) j

^ Children in group care and other types of out-of-home care were excluded. (back)

a The home environments of children aged 2 and under received higher total scores than the home environments of children between 6 and 10 years old (t=25.9, p<.001). (back)

b The home environments of children between 3 and 5 years old received higher total scores than the home environments of children between 6 and 10 years old (t=20.3, p<.001). (back)

c The home environments of African American children received lower total scores than the home environments of White children (t=-3.9, p<.001). (back)

d The home environments of Hispanic children received lower total scores than the home environments of White children (t=-2.7, p<.01). (back)

e The home environments of children who lived at home and received services received lower total scores than the home environments of children who lived at home but did not receive services (t=-2.9, p<.01). (back)

f The home environments of children whose caregivers' income was $50,000 or above received higher total scores than the home environments of children whose caregivers' income was less than $25,000 (t=2.7, p<.01). (back)

g The home environments of children whose caregivers received a high school diploma or GED received higher total scores than the home environments of children who caregivers received no degree (t=3.5, p<.001) (back)

h The home environments of children whose caregivers received an Associate's or vocational degree received higher total scores than the home environments of children whose caregivers received no degree (t=4.5, p<.001). (back)

i The home environments of children whose caregivers received a Bachelor's degree or more received higher total scores than the home environments of children whose caregivers received no degree (t=5.1, p<.001). (back)

j The home environments of children whose interviewers were of other races/ethnicities received lower scores than the home environments of children whose interviewers were African American (t=-3.3, p<.01). (back)

Table 6-28. Regression Results for Explaining Cognitive Stimulation Score
		Beta Coefficient (SE)
Child Age	0-2 years	0.15 (.01) a
	3-5 years	0.16 (.01) b
	6-10 years	(reference group)
Child Race/ Ethnicity	African American	-0.03 (.01) c
	White	(reference group)
	Hispanic	-0.07 (.02) d
	Other	-0.02 (.02)
Setting	No Services	(reference group)
	Services	-0.02 (.02)
	Foster Care	-0.05 (.02) e
	Kinship Foster Care	0.01 (.02)
Child Gender	Male	0.00 (.01)
Child Gender	Female	(reference group)
Household Income	< $25,000	(reference group)
	$25,000 to $49,999	0.04 (.01) f
	≥ $50,000	0.06 (.02) g
Caregiver Education	No Degree	(reference group)
	High School Diploma/GED	0.05 (.01) h
	Associate's or Vocational	0.08 (.02) i
	Bachelor's Degree or Higher	0.12 (.02) j
	Other	0.04 (.03)
Interviewer Race/ Ethnicity	African American	-0.02 (.01)
	White	(reference group)
	Hispanic	0.01 (.02)
	Other	-0.04 (.02)

Note: CS is Cognitive Stimulation.

^ Children in group care and other types of out-of-home care were excluded. (back)

a The home environments of children aged 2 years and under received higher CS scores than children between 6 and 10 years old (t=12.9, p<.001). (back)

b The home environments of children ages 3 to 5 years received higher CS scores than children between 6 and 10 years old (t=12.6 p<.001). (back)

c The home environments of African American children received lower CS scores than the home environments of White children (t=-2.7, p<.01). (back)

d The home environments of Hispanic children received lower CS scores than the home environments of White children (t=-4.1, p<.001). (back)

e The home environments of children living in foster care received lower CS scores than the home environments of children living at home who did not receive services (t=-3.4, p<01). (back)

f The home environments of children whose caregivers' income was between $25,000 and $49,999 received higher CS scores than the home environments of children whose caregivers' income was less than $25,000 (t=3.3, p<.01). (back)

g The home environments of children whose caregivers' income was $50,000 or higher received higher CS scores than the home environments of children whose caregivers' income was less than $25,000 (t=2.6, p<.01). (back)

h The home environments of children whose caregivers earned a high school degree or GED received higher CS scores than children whose caregivers received no degree (t=3.3, p<.01). (back)

i The home environments of children whose caregivers earned an Associate's or vocational degree received higher cognitive stimulation scores than the home environments of children whose caregivers received no degree (t=4.4, p<.001). (back)

j The home environments of children whose caregivers earned a Bachelor's degree or higher received higher cognitive stimulation scores than the home environments of children whose caregivers received no degree (t=5.2, p<.001). (back)

Table 6-29. Regression Results for Explaining Emotional Support Score
		Beta Coefficient (SE)
Child Age	0-2 years	0.25 (.01) a
	3-5 years	0.18 (.01) b
	6-10 years	(reference group)
Child Race/Ethnicity	African American	-0.04 (.01) c
	White	(reference group)
	Hispanic	0.01 (.01)
	Other	0.01 (.01)
Setting	No Services	(reference group)
	Services	-0.02 (.01) d
	Foster Care	0.04 (.02)
	Kinship Foster Care	0.04 (.02)
Child Gender	Male	-0.00 (.01)
Child Gender	Female	(reference group)
Household Income	< $25,000	(reference group)
	$25,000 to $49,999	0.00 (.01)
	≥ $50,000	0.03 (.01)
Caregiver Education	No Degree	(reference group)
	High School Diploma/GED	0.01 (.01)
	Associate's or Vocational	0.02 (.01)
	Bachelor's Degree or Higher	0.04 (.02) e
	Other	0.02 (.02)
Interviewer Race/Ethnicity	African American	-0.01 (.01)
	White	(reference group)
	Hispanic	0.02 (.01)
	Other	-0.02 (.01)

Note: ES is Emotional Support.

^ Children in group care and other types of out-of-home care were excluded. (back)

a The home environments of children aged 2 years and under received higher ES scores than children between 6 and 10 years old (t=29.0, p<.001). (back)

b The home environments of children ages 3 to 5 years received higher ES scores than children between 6 and 10 years old (t=19.9, p<.001). (back)

c The home environments of African American children received lower ES scores than the home environments of White children (t=-3.5, p<.01). (back)

d The home environments of children who live at home and received services received lower ES scores than the home environments of children who live at home and did not receive services (t=-2.5, p ≤ .01). (back)

e The home environments of children whose caregivers earned a Bachelor's degree or higher received higher ES scores than the home environments of children whose caregivers received no degree (t=2.9, p<.001). (back)

Table 6-30. Regression Results for Explaining Physical Environment Score
		Beta Coefficient (SE)
Child Age	3-5 years	0.07 (.02) a
Child Age	6-10 years	(reference group)
Child Race/Ethnicity	African American	-0.02 (.03)
	White	(reference group)
	Hispanic	-0.06 (.03)
	Other	0.09 (.03) b
Setting	No Services	(reference group)
	Services	-0.03 (.02)
	Foster Care	0.04 (.03)
	Kinship Foster Care	0.06 (.03)
Child Gender	Male	<0.01 (.02)
Child Gender	Female	(reference group)
Household Income	< $25,000	(reference group)
	$25,000 to $49,999	0.07 (.02) c
	≥ $50,000	0.03 (.03)
Caregiver Education	No Degree	(reference group)
	High School Diploma/GED	0.04 (.02)
	Associate's or Vocational	0.06 (.03)
	Bachelor's Degree or Higher	0.13 (.03) d
	Other	0.04 (.05)
Interviewer Race/ Ethnicity	African American	-0.05 (.03)
	White	(reference group)
	Hispanic	0.03 (.04)
	Other	-0.18 (.05) e

Note: PE is Physical Environment.

^ Children in group care and other types of out-of-home care were excluded. (back)

a The home environments of children between 3 and 5 years received higher PE scores than the home environments of children between 6 and 10 years (t=3.9, p<.001). (back)

b The home environments of children of other races/ethnicities received higher PE scores than the home environments of White children (t=2.6, p<.01). (back)

c The home environments of children whose caregivers' income is between $25,000 and $49,999 received higher PE scores than the home environments of children whose caregivers' income is below $25,000 (t=2.8, p<.01). (back)

d The home environments of children whose caregivers received a Bachelor's degree or more received higher PE scores than the home environments of children whose caregivers received no degree (t=4.1, p<.001). (back)

e The home environments of children whose interviewers were of other races/ethnicities received lower PE scores than the home environments of children whose interviewers were White (t=3.6, p<.001). (back)

6.4.6 Caregiver Age and the HOME-SF

The ages of caregivers may be associated with the type and quality of care they receive (Thomas, Sperry, & Yarbrough, 2000). As noted in Section 6.1, the ages of caregivers of children involved with CWS vary markedly. Caregivers of children living at home are much younger than those living in nonkinship or kinship foster care. Because the ages of out-of-home caregivers differ from their counterparts in the general population, their in-home counterparts, we examined differences in in-home-SF scores according to the age of the caregiver. The results are presented below by the age of the child.

Children Birth to 2 Years Old

Caregiver age had little relationship to the total, cognitive stimulation, or physical environment scores among children less than 3 years old (Table 6-31). However, emotional support scores for the home environments with caregivers less than 30 years old were lower than emotional support scores for home environments with caregivers between 30 and 45 years old.

Children 3 to 5 Years Old

Among children aged 3 to 5 years, home environments differed by caregiver age in only one instance (Table 6-32). This difference showed that emotional support scores for out-of-home environments with caregivers from 30 to 45 years were higher than for out-of-home environments with caregivers over 45 years.

Children 6 to 10 Years Old

Among children aged 6 to 10 years old, no differences exist by caregiver age in the total, cognitive stimulation, or emotional support scales (Table 6-33). However, the physical environment scores for out-of-home environments with caregivers less than 30 years old were higher than scores for in-home environments with caregivers less than 30 years old.

6.4.7 Relationships between HOME-SF Scores and Other Measures

To better understand the meaning of the HOME-SF scores for this population, we tested whether the total scores were related to case substantiation, the total risk assessment score, or the proportion of clinical developmental scores. Among children less than 3 years old, we found that the total score tended to be inversely related, the predicted direction, to case substantiation (p<.05) and to the proportion of clinical scores (p<.001). Among children aged 3 to 5 years, the total score tended to be inversely related to the proportion of clinical scores (p<.05). Finally, we found that the total score for 6- to 10-year-olds was inversely related to both the total risk assessment score (p<.001) and the proportion of clinical scores (p<.001). Thus, the home environments that scored lower were also the home environments with children with the greatest number of problems and most likely to have a substantiated allegation of maltreatment.

6.4.8 Punitiveness

High levels of parental punitiveness are associated with attitudes and behaviors related to maltreatment (Milner & Chilamkurti, 1991) and in many studies (e.g., Smith & Brooks-Gunn, 1997) are also found to have an association with poor developmental outcomes. Although this relationship is not inviolate, evidence from a variety of sources shows that parental punitiveness can become child maltreatment (e.g., Straus et al., 1998). Prior research suggested the value of using the HOME-SF to estimate the level of punitiveness in the homes in which the children reside. Using a punitiveness/hostility subscale based on the observational items of the HOME-SF (Linver, Filigni, & Brooks-Gunn, 2001), we assessed punitive caregiving environments for children younger than 6 years. The items in the scale included observations of whether the mother/guardian shouted, expressed annoyance or hostility, criticized, slapped or spanked, and restricted the child multiple times during the interviewer's home visit. About two-thirds of the caregivers of children less than 3 years old displayed some punitiveness during the observation by the interviewers. Among children less than 3 years old, caregivers of children living at home and living out of the home showed no significant differences in punitiveness. Among children less than 3 years old who were living out of home, there were no differences between kinship and nonkinship caregivers. Among children 3 to 5 years old, 49% of caregivers displayed some punitiveness during the observation by the interviewers. The punitive behaviors that occurred most commonly during the interview most often included criticism and annoyance, whereas slapping and spanking were fairly rare. There were no differences in punitiveness by CWS setting or by race/ethnicity among caregivers of 3- to 5-year-olds.

Table 6-31. HOME-SF Scores for Children Less Than 3 Years, by Caregiver Age
			In-home			Out-of-home
		TOTAL	Setting
			No Services	Services	TOTAL In-Home	Foster Care	Kinship Foster Care	TOTAL Out-of-Home
		Mean / (SE)
Caregiver Age <30 years	Total	13.5 (0.2)	13.7 (0.3)	13.1 (0.2)	13.5 (0.2)	15.2 (0.4)	11.4 (1.0)	13.9 (0.7)
	CS	6.8 (0.1)	6.9 (0.1)	6.6 (0.1)	6.8 (0.1)	7.6 (0.3)	5.5 (0.5)	6.8 (0.5)
	ES	6.8a (0.1)	6.8 (0.2)	6.6 (0.1)	6.8 (0.1)	7.6 (0.3)	5.9 (0.6)	7.0 (0.3)
Caregiver Age 30-45 years	Total	14.1 (0.3)	14.6 (0.5)	13.5 (0.3)	14.3 (0.3)	14.2 (0.3)	13.4 (0.7)	13.9 (0.3)
	CS	6.8 (0.1)	7.2 (0.2)	6.6 (0.2)	7.1 (0.2)	6.6 (0.2)	6.3 (0.2)	6.5 (0.2)
	ES	7.3 (0.2)	7.4 (0.3)	7.0 (0.2)	7.2 (0.2)	7.7 (0.1)	7.1 (0.5)	7.4 (0.2)
Caregiver Age >45 years	Total	13.7 (0.3)	13.7 (1.1)	14.9 (0.5)	14.3 (0.6)	13.4 (0.4)	13.6 (0.6)	13.6 (0.4)
	CS	6.7 (0.3)	7.1 (0.3)	7.3 (0.3)	7.2 (0.2)	6.4 (0.2)	6.7 (0.6)	6.5 (0.3)
	ES	7.0 (0.1)	6.6 (1.0)	7.6 (0.3)	7.1 (0.5)	7.1 (0.2)	7.0 (0.1)	7.0 (0.1)

Note: CS = Cognitive Stimulation, ES = Emotional Support, and PE = Physical Environment. Children in group homes and other out-of-home placements were excluded.

a ES scores for the home environments of all children living with caregivers between 30 and 45 years had higher ES scores than the home environments of caregivers less than 30 years (t=3.0, p<.01). (back)

Table 6-32. HOME-SF Scores for Children Aged 3 to 5 Years, by Caregiver Age
			In-home			Out-of-home
		TOTAL	Setting
			No Services	Services	TOTAL In-Home	Foster Care	Kinship Foster Care	TOTAL Out-of-Home
		Mean / (SE)
Caregiver Age <30 year	Total	18.8 (0.3)	19.1 (0.3)	18.0 (0.4)	18.8 (0.3)	---	---	20.4 (1.0)
	CS	10.4 (0.2)	10.6 (0.2)	10.1 (0.3)	10.4 (0.2)	---	---	10.9 (0.6)
	ES	8.4 (0.2)	8.5 (0.2)	7.9 (0.2)	8.4 (0.2)	---	---	9.6 (0.8)
	PE	6.1 (0.2)	6.2 (0.2)	5.8 (0.3)	6.1 (0.2)	6.2 (0.2)	7.1 (0.1)	6.6 (0.3)
Caregiver Age 30-45 years	Total	19.8 (0.4)	19.9 (0.5)	19.5 (0.6)	19.8 (0.4)	21.0 (0.8)	20.2 (0.5)	20.6 (0.5)
	CS	11.0 (0.3)	11.1 (0.4)	10.8 (0.5)	11.0 (0.3)	11.2 (0.6)	10.5 (0.4)	10.9 (0.4)
	ES	8.9 (0.3)	8.9 (0.4)	8.7 (0.3)	8.8 (0.3)	9.8 (0.4)	9.7 (0.6)	9.8a (0.3)
	PE	6.3 (0.2)	6.1 (0.3)	6.5 (0.3)	6.3 (0.2)	7.4 (0.2)	6.2 (0.6)	6.9 (0.4)
Caregiver Age >45 years	Total	19.5 (0.5)	18.7 (0.9)	18.8 (1.4)	18.7 (0.8)	19.5 (0.7)	20.6 (0.9)	20.3 (0.7)
	CS	11.2 (0.6)	10.9 (0.7)	10.4 (0.7)	10.8 (0.6)	10.0 (0.4)	12.1 (0.7)	11.7 (0.7)
	ES	8.3 (0.3)	7.8 (0.6)	8.4 (0.8)	7.9 (0.5)	9.5 (0.4)	8.4 (0.4)	8.7 (0.3)
	PE	6.5 (0.4)	6.2 (0.8)	5.3 (1.1)	6.1 (0.6)	6.9 (0.3)	7.0 (0.3)	7.0 (0.3)

Note: CS = Cognitive Stimulation, ES = Emotional Support, and PE = Physical Environment. Children in group homes and other out-of-home placements were excluded.

a ES scores for out-of-home environments with caregivers between 30 and 45 years were higher than ES scores for out-of- home environments with caregivers over 45 years (t=2.7, p ≤ .01). (back)

6.4.9 Neighborhood Context

To estimate the development-supporting characteristics of the community in which the child resided, the caregiver was asked about nine aspects of his or her community. Overall, the community environment of children living at home was worse than the community environment of children living out of the home (p ≤ .001). (See Table 6-34.) In general, children living at home reside in communities that experienced significantly more specific problems, including open drug use or dealing, unsupervised children (p ≤ .001), teenagers hanging out and being a nuisance (p≤.001), neighbors helping each other out less (p ≤ .001), and parents involved less (p≤.001); in addition, they were less safe than most neighborhoods (p ≤ .001), and fewer parents thought their neighborhood was a better place to live than most (p ≤ .001) compared with communities of children who live out of the home. There were no differences in the community environment between foster, kinship, or group homes.

Table 6-33. HOME-SF Scores for Children Aged 6 to 10 Years, by Caregiver Age
		TOTAL	In-home			Out-of-home
		Setting
		TOTAL	No Services	Services	TOTAL In-Home	Foster Care	Kinship Foster Care	TOTAL Out-of-Home
		Mean / (SE)
Caregiver Age <30 years	Total	17.3 (0.3)	17.4 (0.3)	17.0 (0.5)	17.3 (0.3)	18.2 (0.9)	---	18.6 (0.6)
	CS	8.2 (0.2)	8.1 (0.3)	8.6 (0.3)	8.2 (0.2)	8.3 (0.4)	---	8.5 (0.4)
	ES	9.1 (0.2)	9.3 (0.2)	8.4 (0.3)	9.1 (0.2)	10.0 (0.6)	---	10.0 (0.4)
	PE	6.3 (0.2)	6.3 (0.3)	6.1 (0.4)	6.3a (0.2)	7.1 (0.4)	7.5 (0.6)	7.3 (0.4)
Caregiver Age 30-45 years	Total	17.9 (0.3)	18.1 (0.4)	17.3 (0.5)	17.9 (0.3)	18.5 (0.5)	18.0 (1.0)	18.4 (0.5)
	CS	8.7 (0.2)	8.8 (0.2)	8.1 (0.2)	8.7 (0.2)	9.4 (0.2)	9.1 (0.5)	9.3 (0.2)
	ES	9.2 (0.2)	9.3 (0.3)	9.1 (0.3)	9.2 (0.2)	9.1 (0.5)	8.9 (0.6)	9.1 (0.4)
	PE	6.5 (0.2)	6.4 (0.3)	6.4 (0.2)	6.4 (0.3)	7.5 (0.6)	7.6 (0.5)	7.5 (0.5)
Caregiver Age>45 years	Total	18.6 (0.7)	18.8 (0.7)	17.9 (2.0)	18.5 (0.8)	16.7 (1.4)	19.6 (1.1)	18.7 (1.0)
	CS	9.0 (0.4)	8.7 (0.4)	9.1 (1.3)	8.8 (0.6)	8.4 (0.6)	9.4 (0.7)	9.1 (0.6)
	ES	9.6 (0.3)	10.1 (0.4)	8.9 (0.8)	9.7 (0.4)	8.3 (0.9)	10.2 (0.5)	9.6 (0.5)
	PE	6.6 (0.3)	6.6 (0.5)	5.8 (0.5)	6.4 (0.4)	7.1 (0.4)	6.6 (0.8)	6.8 (0.5)

Note: CS = Cognitive Stimulation, ES = Emotional Support, and PE = Physical Environment. Children in group homes and other out-of-home placements were excluded.

a PE scores for out-of-home environments with caregivers <30 were higher than PE scores for in-home environments with caregivers less than 30 years (t=3.0, p<.01). (back)

6.4.10 Discussion of Children's Living Environment

Bivariate analyses of the HOME-SF indicated that there are differences in in-home environment scores by CWS setting and child age. The bivariate analyses indicated that out-of-home environments received higher scores than in-home environments. When a multivariate analysis controlled for child age, child race/ethnicity, child gender, household income, caregiver education, and interviewer race/ethnicity, the difference in scores between out-of-home and in-home settings largely disappeared. Instead, the multivariate analyses indicate that the home environments of children who received CWS received significantly lower total and emotional support scores than the home environments of children who did not receive CWS. In addition, home environments of children living in foster care received lower cognitive stimulation scores than the home environments of children living at home who did not receive CWS.

Table 6-34. Caregiver Reports on Level of Problems and Strengths in Neighborhood Environment
	TOTAL	Total In-Home	Total Out-of-Home
	Setting
	Mean / (SE)
Assaults/Muggings	1.2 (0.02)	1.2 (0.02)	1.2 (0.03)
Gangs	1.3 (0.03)	1.3 (0.03)	1.2 (0.04)
Open drug use or dealing	1.4 (0.02)	1.4 b (0.03)	1.2 (0.04)
Unsupervised children	1.5 (0.02)	1.5 c (0.03)	1.3 (0.03)
Teenagers hanging out	1.4 (0.02)	1.4 d (0.02)	1.2 (0.03)
Safe as most other neighborhoods	1.6 (0.02)	1.6 e (0.03)	1.4 (0.03)
Neighbors help each other out	2.0 (0.02)	2.0 f (0.03)	1.7 (0.04)
Parents are involved	2.0 (0.02)	2.1 g (0.03)	1.8 (0.05)
A better place to live than most	1.6 (0.02)	1.6 h (0.03)	1.4 (0.03)
TOTAL	13.7 (0.2)	13.8 a (0.2)	12.3 (0.2)

Note: The total out-of-home category also includes children living in “other” out-of-home settings.

a Total community environment scores were higher among children living at home than among children living out of the home (t=5.4, p≤.001). (back)

b The likelihood of open drug use or dealing in the environment of children living at home is higher than for children living out of the home (t=3.2, p≤.01). (back)

c The likelihood of unsupervised children in the environment of children living at home is higher than for children living out of the home (t=4.2, p≤.001). (back)

d The likelihood of teenagers hanging out in the environment of children living at home is higher than for children living out of the home (t=4.9, p≤.001). (back)

e The likelihood of the neighborhood being as safe as most other neighborhoods among children living at home is higher than for children living out of the home (t=5.6, p≤.001). (back)

f The likelihood of neighbors helping each other out is higher in the environment of children living at home than for children living out of the home (t=6.0, p≤.001). (back)

g The likelihood of involved parents is higher among children living at home than for children living out of the home (t=4.4, p≤.001). (back)

h The likelihood of the neighborhood being a better place to live than most is higher among children living at home than for children living out of the home (t=4.9, p≤.001). (back)

These differences suggest that children in different CWS settings are receiving varied levels of exposure to environmental characteristics associated with developmental success. Although Cronbach's alpha ranges from .44 to .59 for the total score, the total score was significantly related to the family's total risk score and the child's proportion of clinical developmental scores. This association lends some credibility to the HOME-SF scores and may indicate that there is a set of children with many risks, poor developmental scores, and poor home environments that should be of particular concern to child welfare workers.

Finally, we found that the community environment scores of children living at home were lower than the community environment scores of children living out of the home. This may help to alleviate concern that most children in out-of-home care are sent to homes with community environments that are worse than the children's original environment. Conversely, these findings will not reassure those who hope that foster care will consistently provide an enriching environment for children.

6.5 In-Home Caregiver Self-Reports of Functioning

The health and well-being of caregivers is likely to contribute to the quality of parenting they provide (Pedro-Carroll, 2001; Wyman et al., 1999). In turn, children's safety, the stability of their home environment, and their overall well-being are directly related to the characteristics and behavior of their caregivers (Kitzman et al., 1997). CWS are designed to assist families with myriad difficulties related to child maltreatment. Child welfare workers intervene both remedially and preventively with families, ideally providing assistance that will strengthen families and prevent future child maltreatment. Caregivers may struggle with mental health issues, poverty, and substance abuse, which are frequently the focus of child welfare interventions (Polansky, DeSaix, & Sharlin, 1972). The degree to which caregivers involved with CWS experience these difficulties, however, is not known. This section presents information about current caregiver functioning across several domains: mental and physical health; social support; substance use; domestic violence; arrest history; and other characteristics of the living environment, including family poverty and household composition. This information will aid our understanding of family service needs.

Each of the caregiver and family characteristics described in this section have been associated with child maltreatment. Parental depression has repeatedly been shown to be associated with problems in parenting (Lovejoy et al., 2000). Alcohol problems have long been a known contributor to involvement with CWS (Gordon, 1988; Polansky, Hall, & Polansky, 1975). More recently, the proportion of caregivers involved with CWS who have a substance abuse problem is estimated to be in the 20% to 80% range (Semidei, Feig-Radel, & Nolan, 2001). These estimates, however, are almost all derived from child welfare worker case records or reports rather than from the self-report of child welfare clients. The NSCAW study is the first large-scale study to obtain this information from the caregivers themselves.

Other problems that caregivers involved with CWS may experience include domestic violence and trouble with the law. Domestic violence is a well-documented contributor to the risk of child maltreatment (Mills et al., 2000), and domestic violence reports are increasingly a source of child maltreatment reports (Fleck-Henderson, 2000). This section provides an estimate of the proportion of caregivers experiencing domestic violence, and Chapter 8 links this information to caregiver services for domestic violence.

Caregivers in CWS may also be involved with the criminal justice system. Whereas caregiver arrest for child maltreatment is a relatively infrequent occurrence across all child welfare cases (Loman & Siegel, 2000), caregiver arrest for other criminal behavior may be more likely, given the range of risk factors that may exist in a family involved with CWS (Shireman, Miller, & Brown, 1981). Caregiver involvement with the criminal justice system complicates the children's home environment, not only because of the criminal behavior, but because criminal justice sanctions often involve financial and community restitution. When arrest does occur, caregivers may spend time in jail or prison, leaving their children's care to family members or CWS.

Caregivers who have adequate support from family members, friends, and their community may be better able to resolve family problems related to child maltreatment. A supportive adult partner, extended family members, and other sources of social support can help caregivers with concrete needs (e.g., child care) and also with caregivers' psychosocial needs for connection and encouragement. Social support is associated with successful functioning and coping in a variety of life domains (Cochran & Starr, 2002).

Several measures were used to capture caregiver self-report of functioning. Results from the Short Form Health Survey (SF-12) provide estimates of the physical and mental health status of caregivers (Ware, Kosinski, & Keller, 1998). This measure was used with out-of-home caregivers as well as in-home caregivers, and these findings are included in this section.

In-home caregivers provided information, via audio computer-assisted self-interviewing (ACASI), about depression, substance use, and substance dependence. The Composite International Diagnostic Interview–Short Form (CIDI-SF) provides additional information about caregiver self-reported depression and substance use and dependence (both alcohol and other drugs). Caregiver experiences of domestic violence and arrest were also captured using the Conflict Tactic Scale (CTS Version I). And finally, in-home caregivers report on the types and degree of social support present in their lives.

All information in this section, with the exception of mental and physical health status, is for in-home caregivers only. Mental and physical health status was obtained on all caregivers and is included in this section because it is an indicator of functioning rather than a demographic characteristic.

6.5.1 Mental and Physical Health Status of Caregivers

In-home and out-of-home caregivers responded to items regarding their mental and physical health on the SF-12. Caregivers who were 55 years or older had significantly better mental health than caregivers less than 55 years old, regardless of the setting of the children (a higher score is better) (Table 6-35). Caregivers of children living in foster care had significantly better mental health than caregivers of children in kinship care. In addition, caregivers of children living out of the home had significantly better mental health than caregivers of children living at home. There were no significant differences in the caregiver's mental health by race/ ethnicity.

When a multivariate analysis is conducted, type of service setting and caregiver age each have a significant relationship to the mental health score (Table 6-36). Caregivers of children in foster homes or in kinship care had significantly higher scores than caregivers of children living at home who did not receive services. Caregivers age 55 and over had significantly higher mental health scores than caregivers between 35 and 44 years old. There were no differences in mental health by caregiver race/ethnicity.

6.5.2 Physical Health of Caregivers

The physical health of caregivers varied by a number of factors (Table 6-37). In general, young caregivers were in significantly better physical health than other caregivers. Caregivers of children in group care were in significantly better physical health than caregivers of children in foster or kinship foster care. Hispanic caregivers were in significantly better physical health than African American caregivers in all settings combined. In addition, Hispanic caregivers of children at home were in significantly better physical health than African American caregivers of children at home.

Table 6-35. Mean Mental Health Score of Caregivers on SF-12
			In-Home			Out-of-Home
		TOTAL	Setting
			No Services	Services	TOTAL In-Home	Foster Care	Kinship Foster Care	Group Care	TOTAL Out-of-Home^
		Percent/ (SE)
Caregiver Age	<25	47.5 c (0.9)	48.3 (1.0)	45.5 (1.8)	47.5 g (0.9)	---	50.6 (1.2)	49.1 (3.9)	47.1 (2.0)
	25-34	48.1 d (0.6)	48.1 (0.7)	47.4 (0.8)	47.9 h (0.6)	55.3 (1.0)	52.3 (1.2)	53.0 (1.7)	53.0 (1.0)
	35-44	48.2 e (0.6)	47.8 (0.9)	47.3 (0.9)	47.7 i (0.7)	56.2 (0.5)	49.7 (1.3)	54.0 (1.0)	52.9 j (0.8)
	45-54	49.8 f (0.9)	49.8 (1.2)	45.4 (3.1)	48.6 (1.2)	54.3 (0.9)	51.0 (1.0)	---	52.4 k (0.8)
	>55	54.0 (0.8)	54.1 (2.4)	51.4 (2.3)	53.2 (1.8)	53.4 (1.8)	56.1 (1.0)	---	54.4 l (1.0)
Caregiver Race/ Ethnicity	African American	48.2 (0.7)	46.9 (0.8)	49.2 (0.8)	47.7 (0.7)	55.2 (1.3)	52.0 (1.0)	47.6 (3.4)	52.5 (1.0)
	White	48.4 (0.4)	48.8 (0.6)	45.0 (0.8)	47.8 (0.5)	55.3 (0.5)	53.2 (1.1)	53.4 (0.8)	53.5 (0.7)
	Hispanic	48.3 (1.0)	47.5 (1.2)	49.9 (1.3)	48.1 (1.1)	53.3 (1.2)	48.9 (1.9)	55.9 (1.5)	50.9 (1.0)
	Other	49.7 (1.3)	50.3 (1.6)	44.8 (1.6)	49.2 (1.3)	54.4 (0.9)	56.7 (1.3)	---	55.3 (0.8)
TOTAL		48.4 (0.4)	48.3 (0.5)	46.9 (0.6)	47.9 b (0.4)	54.9 a (0.5)	52.6 (0.6)	50.9 (2.2)	53.1 (0.5)

^ Includes children in “other” out-of-home placement settings. (back)

a Caregivers of children in foster care had higher mental health scores than caregivers of children in kinship foster care (t=2.9, p<.01). (back)

b Out-of-home caregivers had higher mental health scores than in-home caregivers (t=8.8, p≤.001). (back)

c Caregivers 55 years old and over had higher mental health scores than caregivers less than 25 years old (t =6.0, p≤.001). (back)

d Caregivers 55 years old and over had higher mental health scores than caregivers between 25 and 34 years old (t=5.7, p≤.001). (back)

e Caregivers 55 years old and over had higher mental health scores than caregivers between 35 and 44 years old (t=5.7, p≤.001). (back)

f Caregivers 55 years old and over had higher mental health scores than caregivers between 45 and 54 years (t=3.5, p≤.001). (back)

g Caregivers 55 years old and over of children living at home had higher mental health scores than caregivers less than 25 years of children living at home (t =3.0, p≤.01). (back)

h Caregivers 55 years old and over of children living at home had higher mental health scores than caregivers between 25 and 34 years old of children living at home (t =2.8, p≤.01). (back)

i Caregivers 55 years old and over of children living at home had higher mental health scores than caregivers between 35 and 44 years old of children living (t =3.0, p≤.01). (back)

j Caregivers 35 to 44 years old of children living out of home had higher mental health scores than caregivers less than 25 years old of children living out of home (t =2.6, p≤.01). (back)

k Caregivers 45 to 54 years old of children living out of home had higher mental health scores than caregivers less than 25 years old of children living out of home (t =2.7, p≤.01). (back)

l Caregivers 55 years old and over of children living out of home had higher mental health scores than caregivers less than 25 years old of children living out of home (t =3.3, p≤.01). (back)

Table 6-36. Regression Modeling Mental Health of Caregivers
		Beta Coefficient (SE)
Child Setting/Services	No child welfare services	(reference group)
	Child welfare services	-1.26 (.65)
	Foster home	5.92 (.68)**_i
	Kinship care	3.13 (.90)**_ii
	Group home	2.18 (2.53)
Caregiver Race/Ethnicity	White	(reference group)
	African American	-.19 (.73)
	Hispanic	.07 (1.14)
	Other	1.36 (1.33)
Caregiver Age	35-44 years	(reference group)
	<25 years	-.31 (1.17)
	25-34 years	.13 (.83)
	45-54 years	.63 (1.14)
	55+ years	3.35 (1.18)*

Multiple R2 is .03.

** p<.001; (back: **_i, **_ii)

* p<.01 (back)

A multivariate analysis showed that service setting and caregiver age had a significant effect on the physical health scores of caregivers. This is consistent with the findings for mental health. The analysis showed that caregivers of children in group care had significantly better physical health scores than caregivers of children at home who did not receive services. There were differences in physical health among caregivers that had open in-home CWS cases and those with closed cases; those with closed cases had better physical health. Also, caregivers who were 24 years old or under were in significantly better physical health than caregivers between 35 and 44 years old (Table 6-38). Compared with the expected mean of 50 and a standard deviation of 10, with higher scores reflecting better health, the self-reported mental and physical health of caregivers, as a group, is decidedly average.

6.5.3 In-Home Caregiver Depression on the CIDI-SF

Almost one in four of the in-home caregivers (23%) reported experiencing major depression in the past 12 months (Table 6-39), based on results from the Composite International Diagnostic Interview – Short Form (CIDI-SF), a highly standardized interview that screens for mental health disorders using the criteria established in the Diagnostic and Statistical Manual of Mental Disorders (American Psychiatric Association, 1994). This compares with 13% of women experiencing a major depressive episode using the CIDI as reported in the National Comorbidity study (Kessler et al., 1994). There were no significant differences in depression among caregivers of different racial/ethnic groups, or on the basis of caregiver or child age, or whether services had been received. There was a trend (p<.03), however, toward decreased depression among older caregivers, with caregivers 55 and older being less likely to report depressive episodes. These findings of more robust mental health for older caregivers match the results from the self-reported mental health scores on the SF-12.

Table 6-37. Mean Physical Health Score of Caregivers
			In-Home			Out-of-Home
		TOTAL	Setting
			No Services	Services	TOTAL In-Home	Foster Care	Kinship Foster Care	Group Care	TOTAL Out-of-Home
		Percent/ (SE)
Caregiver Age	<25	50.0 c d (0.5)	50.3 (0.7)	49.5 (0.8)	50.1 f g (0.5)	---	37.8 (6.1)	---	47.6 (4.1)
	25-34	49.8 e (0.5)	49.8 (0.5)	50.1 (0.6)	49.7 h (0.5)	52.9 (0.8)	48.4 (2.8)	54.7 (0.6)	52.5 i j (0.8)
	35-44	47.9 (0.6)	47.7 (0.9)	47.8 (0.9)	47.8 (0.6)	47.2 (3.2)	51.8 (0.9)	50.3 (2.9)	49.3 (1.9)
	45-54	45.3 (1.0)	45.0 (1.6)	42.2 (2.0)	44.2 (1.3)	47.8 (2.0)	47.2 (1.8)	---	47.6 (1.2)
	>55	46.0 (1.6)	45.4 (3.5)	41.1 (2.7)	44.0 (2.8)	49.2 (1.8)	44.9 (2.4)	---	46.8 (1.9)
Caregiver Race/ Ethnicity	African American	47.8 (0.7)	47.9 (0.9)	47.3 (0.9)	47.7 (0.7)	46.2 (2.1)	48.0 (1.2)	54.0 (0.6)	48.5 (1.1)
	White	48.8 (0.4)	49.1 (0.5)	48.4 (0.7)	48.9 (0.4)	48.8 (2.2)	46.2 (2.1)	54.3 (1.1)	48.1 (1.3)
	Hispanic	50.3 k (0.6)	51.0 (1.1)	47.8 (1.6)	50.1 l (0.7)	52.9 (1.6)	52.2 (1.0)	---	52.6 (1.0)
	Other	47.4 (1.1)	47.9 (1.4)	45.5 (1.8)	47.5 (1.2)	47.9 (2.1)	45.6 (2.6)	---	47.0 (1.5)
TOTAL		48.7 (0.3)	49.1 (0.4)	47.9 (0.5)	48.7 (0.3)	48.6 (1.4)	47.2 (1.3)	54.0 a b (0.5)	48.6 (0.9)

a Caregivers of children in group care had higher physical health scores than caregivers of children in foster care (t=3.4, p≤.001). (back)

b Caregivers of children in group care had higher physical health scores than caregivers of children in kinship foster care (t=5.0, p≤.001). (back)

c Caregivers less than 25 years old had higher physical health scores than caregivers 35 to 44 years old (t=2.9, p≤.01). (back)

d Caregivers less than 25 years old had higher physical health scores than caregivers 45 to 54 years old (t=4.0, p≤.001). (back)

e Caregivers less than 25 to 34 years old had higher physical health scores than caregivers 45 to 54 years old (t=4.6, p≤.001). (back)

f Caregivers who are less than 25 years old and whose children live at home had higher physical health scores than caregivers who are 35 to 44 years old and whose children live at home (t=3.2, p≤.01). (back)

g Caregivers who are less than 25 years old and whose children live at home had higher physical health scores than caregivers who are 45 to 54 years old and whose children live at home (t=3.7, p≤.01). (back)

h Caregivers who are 25 to 34 years old and whose children live at home had higher physical health scores than caregivers who are 45 to 54 years old and whose children live at home (t=4.1, p≤.001). (back)

i Caregivers who are 25 to 34 years old and whose children live out of the home had higher physical health scores than caregivers who are 45 to 54 years old and whose children live out of the home (t=3.8, p≤.001). (back)

j Caregivers who are 25 to 34 years old and whose children live out of the home had higher physical health scores than caregivers who are 55 years old and over and whose children live out of the home (t=2.8, p≤.01). (back)

k Hispanic caregivers had higher physical health scores than African American caregivers (t=2.8, p≤.01). (back)

l Hispanic caregivers of children who live at home had higher physical health scores than African American caregivers of children who live at home (t=2.87, p≤.01). (back)

Table 6-38. Regression Modeling Physical Health of Caregivers
		Beta Coefficient (SE)
Child Setting/ Services	No child welfare services	(reference group)
	Child welfare services	-1.09 (.64)
	Foster home	1.76 (1.58)
	Kinship care	1.12 (1.49)
	Group home	5.91 (.96)**
Caregiver Race/ Ethnicity	White	(reference group)
	African American	-.92 (.82)
	Hispanic	1.36 (.74)
	Other	-1.42 (1.07)
Caregiver Age	35-44 years	(reference group)
	<25 years	2.37 (.75)*
	25-34 years	1.74 (.77)
	45-54 years	-2.98 (1.26)
	55+ years	-2.95 (1.73)

Multiple R2 is .03.

** p<.001; (back)

* p<.01 (back)

When logistic regression analysis was used to predict caregiver major depression, the finding of a decreased rate among the oldest caregivers was significant (Table 6-40). No other predictor was significant when examined within the multivariate model.

6.5.4 In-Home Caregiver Alcohol Use and Dependence

Alcohol use was examined by self-report of the maximum number of drinks consumed per day in the past year. Alcohol dependence was measured using the CIDI-SF, in which a stem question (consumption of four or more drinks in a single day in the past year) is used to screen out persons who are least likely to be alcohol dependent before they are asked further symptom questions (Nelson, Kessler, & Mroczek, 1998). About two out of five (41%) in-home caregivers reported some use of alcohol in the past 12 months, with slightly less than 1 in 10 reporting that they had used four or more drinks in 1 day (Table 6-41). Race/ethnicity and caregiver age were significantly related to report of any use of alcohol. Hispanic in-home caregivers were less likely to report use as compared with White caregivers (p<.001), and caregivers in the oldest age group (55 or more years) were less likely to report four or more drinks in one day as compared with caregivers in the 25-34 year age group.

Using the CIDI-SF, 2% of this population scored in the alcohol dependence range (Table 6-42), meaning that they had (1) consumed four or more drinks in a single day at least once during the past year and (2) reported at least three of the seven symptoms of alcohol dependence from the Diagnostic and Statistical Manual of Mental Disorders, Third Edition Revised (DSM-III-R) (American Psychiatric Association, 1987) and were likely to meet DSM-III-R criteria for alcohol dependence. This is lower than the national prevalence rate of 4% for women as measured using the CIDI in the National Comorbidity Study (Kessler et al., 1994). There was a significant difference by caregiver race/ethnicity, with African American caregivers more likely to report alcohol dependence than Hispanic caregivers (p<0.01); caregivers in other racial/ethnic groups also had a trend toward increased alcohol dependence as compared with Hispanic caregivers (p<0.04). There were no significant differences by caregiver age or child age. There were no significant differences in any categories or total percentages when the caregivers were compared on the basis of whether or not they were receiving in-home CWS.

Logistic regression analysis uncovered one significant predictor of alcohol dependence. Hispanic caregivers were significantly less likely to report alcohol dependence than those in the White reference group (Table 6-43).

Table 6-39. In-home Caregiver Report of Major Depression in the Past 12 Months
		Percent / (SE)
		No Services	Services	TOTAL
Caregiver Age	≤24 yrs	24.0 (4.2)	25.3 (2.4)	24.4 (3.0)
	25-34 yrs	21.6 (3.0)	25.5 (2.2)	22.6 (2.3)
	35-44 yrs	22.9 (3.3)	27.2 (3.7)	24.2 (2.8)
	45-54 yrs	22.7 (6.7)	32.0 (5.3)	25.2 (5.3)
	55+ yrs	0.0 (0.0)	11.1 (5.7)	4.0 (2.4)
Caregiver Race/ Ethnicity	African American	23.4 (3.6)	17.8 (2.1)	21.5 (2.6)
	White	24.1 (2.4)	32.3 (2.0)	26.2 (1.9)
	Hispanic	15.8 (5.7)	22.8 (4.6)	17.3 (4.6)
	Other	20.9 (5.4)	26.4 (5.0)	22.2 (4.3)
Child Age	0-2 yrs	21.5 (3.4)	23.0 (2.5)	21.9 (2.6)
	3-5 yrs	27.1 (3.8)	25.5 (3.2)	26.6 (2.8)
	6-10 yrs	17.6 (3.3)	27.4 (2.1)	19.8 (2.7)
	11+ yrs	25.9 (2.9)	28.0 (3.4)	26.5 (2.2)
TOTAL		22.2 (2.0)	26.2 (1.8)	23.3 (1.6)

6.5.5 In-Home Caregiver Drug Use and Dependence

Fifteen percent of in-home caregivers in the NSCAW population reported some abuse of “legal” drugs (including reported use of sedatives [6%], tranquilizers [5%], amphetamines [3%], and analgesics [8%]) within the past 12 months (Table 6-44). Approximately 1 in 10 (10%) reported use of “illicit” drugs (including reported use of inhalants [1%], marijuana [9%], cocaine [3%], LSD [1%], and heroin [1%]) within the past 12 months. Overall, 21% of in-home caregivers reported either abuse of legal drugs or use of illegal drugs in the past year.

Table 6-40. Logistic Regression Modeling Major Depression
		OR	95% CI
Caregiver Age	≤24 yrs	0.98	.60, 1.59
	25-34 yrs	0.93	.63, 1.37
	35-44 yrs	(reference group)
	45-54 yrs	1.10	.59, 2.07
	55+ yrs	0.13*	.04, .44
Caregiver Race/ Ethnicity	White	(reference group)
	African American	0.75	.50, 1.12
	Hispanic	0.59	.31, 1.10
	Other	0.79	.49, 1.27
Caregiver Gender	Male	0.73	.41, 1.31
Caregiver Gender	Female	(reference group)
Child Setting/ Services	In-home, no services	(reference group)
Child Setting/ Services	In-home, services	1.25	.95, 1.64

Cox and Snell pseudo-R2 is .02

* p<.001 (back)

There were no significant differences by caregiver age, race/ethnicity, or child age in the total reported abuse or use of any type of drug or in the abuse of legal drugs. The oldest age group, 55 years or older, was significantly less likely to report use of “illicit” drugs as compared with caregivers aged 25 to 34 years (Table 6-45).

Drug dependency as measured by the CIDI-SF requires very high thresholds of use, which may account for the overall relatively low rate of in-home caregivers (3%) meeting the criteria (Table 6-46). Nevertheless, this is higher than the national prevalence rate for women (2%), as measured using the CIDI in the National Comorbidity Study (Kessler et al., 1994). Caregiver race/ethnicity and child's age were not significantly associated with self-reported caregiver drug dependence. Significant differences in reports of drug dependence were present on the basis of caregiver age, with caregivers aged 45 years and older less likely to report drug dependence. Logistic regression analysis confirms the significance of older age in predicting lower rates of drug dependence but uncovered no other significant predictors.

6.5.6 In-Home Caregiver Domestic Violence

The physical violence section of the Conflict Tactic Scale (CTS Version I) was used to assess the frequency and extent of domestic violence (DV) as committed against and reported by the primary caregiver. Nine types of physical violence are grouped into minor and severe categories. The respondent was asked to indicate if a specific type of abuse occurred and then how many times it occurred. Table 6-47 presents the lifetime prevalence of caregiver-reported domestic violence. Almost half of the in-home caregivers (45%) reported experiencing either minor or severe domestic violence during their lifetime, with 43% reporting at least one incident of minor violence and 32% reporting having experienced at least one incidence of severe violence. These rates are much higher than those reported by the National Violence Against Women study (Tjaden & Thoennes, 2000), which also used the physical violence scale of the CTS and found lifetime prevalence rates of 22%.

Table 6-41. In-Home Caregiver Report of Alcohol Use in the Past 12 Months
		No Services	Services	TOTAL	No Services	Services	TOTAL
		Any			4 or More Drinks in One Day
		Percent(SE)
Caregiver Age	≤24 yrs	39.0 (4.6)	37.1 (2.3)	38.5 (3.4)	9.7 (2.4)	8.9 (2.1)	9.5 (1.8)
	25-34 yrs	42.3 (3.1)	42.2