Search Type Title Author/Investigator Search Advanced Search

Table of Contents | Previous | Next

2. FINDINGS

2.1 The Child Welfare System and Child Welfare Agencies

In every community, child welfare agencies try to ensure safety and promote permanency and well-being for abused and neglected children. Such services, however, are not uniform, and different types of child welfare organizations exist. A more precise organization and program context is critical to understanding the achievement of child-level outcomes (Glisson & Hemmelgarn, 1998). For example, public and private agencies historically have worked together to provide child welfare services. Such factors may affect child outcomes and thus are important to study.

2.1.1 Organization of Child Welfare Agencies

A relevant organizational factor is the amount of local control exercised by the child welfare agency. The weighted results indicate that roughly one-quarter (27%) of the directors of local child welfare administrative units were appointed by a state administrator, whereas about one-third (33%) were decided upon by an elected county board (see Table 2-1). Roughly another one-third (34%) were appointed by some entity other than a state administrator, county executive, elected county board, or appointed county board. There is some evidence that appointment by county-elected officials was more likely for county-administered agencies (69%) than for state-administered agencies (2%) and that appointment by some entity other than those mentioned was more likely for state-administered agencies (57%) than for county-administered agencies (4%) (x2 = 10.3, p = .04). No differences were noted with regard to county size, degree of urbanicity, or relative proportion of poor families.

Table 2-1. Responsibility for local child welfare agency (%)

Who is responsible for appointing the director of the local child welfare administrative unit?	Percent (CI Limits)
State administrator	27 (-15, +22)
County executive	1 (-1, +5)
Elected county board	33 (-22, +33)
Appointed county board	5 (-4, +17)
Other	34 (-18, +24)

 

The researchers estimate that in almost two-thirds (65%) of the nation’s counties, the county government had “substantial control” over decisions about how child welfare services funding was spent, but in a substantial minority (32%), county government had “very little or no control” (see Table 2-2). Differences do not appear to be significantly related to county size, degree of urbanicity, relative proportion of poor families, or whether the agency was county or state administered.

Table 2-2. Local control over CWS spending (%)

Local control over how funding is spent	Percent (CI Limits)
Almost complete	3 (-2, +12)
Substantial	65 (-23, +17)
Very little or none	32 (-16, +22)

The majority (65%) of child welfare agencies were units within larger agencies rather than freestanding entities (see Table 2-3). There is some evidence that child welfare agencies within state-administered systems (97%) were more likely than agencies within county-administered systems (30%) to be part of a larger agency (x2 = 4.79, p = .03). No differences were noted with regard to county size, degree of urbanicity, or relative proportion of poor families.

Table 2-3. Organizational position of local CWS agency (%)

Description of agency	Percent (CI Limits)
Freestanding	35 (-22, +31)
Unit within larger agency	65 (-31, +22)

In terms of other services provided by the larger agency, income maintenance/Temporary Assistance for Needy Families (TANF) appears to be the most commonly provided service (92%), with substance abuse treatment (46%), mental health (42%), juvenile justice (41%), and public health (20%) endorsed less frequently (see Table 2-4). Almost two-thirds (65%) of child welfare systems have been integrated into omnibus human services agencies. This is a noteworthy finding, because many have argued that greater integration of services would result in better outcomes for children and families. The trend seems to have accelerated recently, at least for the agencies in this sample, with nearly as many agencies beginning this practice in the 1990s as had been integrated into an omnibus agency in the previous four decades. No differences were noted for any of these variables with regard to county size, degree of urbanicity, relative proportion of poor families, or whether the agency was county or state administered.

Table 2-4. Organizational integration of
local CWS with other services (%)

Services contained within the larger agency	Percent yes (CI Limits)
Substance abuse treatment	46 (-26, +28)
Mental health	42 (-25, +31)
Public health	20 (-12, +22)
Income maintenance/TANF	92 (-14, +5)
Juvenile justice services	41 (-22, +27)
Integrated into omnibus human services agency	73 (-27, +17)

 

There are other indications of changes in organizational structure designed to improve service integration. For example, 40% of respondents indicated that their agencies had developed one of the following in the past 12 months: specialized units, blended teams, or community-based offices. There is some evidence that state-administered agencies (63%) were more likely to have developed a specialized unit than were county-administered agencies (16%) (x2 = 4.05, p < .05). There is some indication that such changes were also associated with county size, with large counties (67%) being more likely than others (35%) to have developed specialized units, blended teams, or community-based offices (x2 = 3.64, p = .06). No differences were noted with regard to degree of urbanicity or relative proportion of poor families.

2.1.2 Provision and Delivery of Child Welfare Services

Service integration is but one of many contemporary child welfare innovations or efforts to restructure child welfare services (see Table 2-5). Such innovations as neighborhood-based services or specialized service units may have a substantial influence on the character and effects of child welfare services.

Table 2-5. Recent innovations and reorganization

Types of innovations/restructuring efforts	Percent yes (CI Limits)
Regionalizing agency services	50 (-31, +32)
Changing casework practice	15 (-11, +26)
Offering neighborhood services/opening satellite offices	36 (-23, +31)
Creating specialized service units	59 (-39, +30)
Creating multidisciplinary or agency teams with TANF	35 (-21, +30)

The most commonly implemented innovation was the creation of specialized service units (59%). Other common efforts to restructure included regionalizing agency services (50%), offering neighborhood services or opening satellite offices (36%), and creating multidisciplinary or agency tea ms with TANF (35%). Only a small minority (15%) appears to have changed casework practice so that only one caseworker was assigned to a case from beginning to end. The mean number of innovations implemented was 2 (out of 5 listed in the questionnaire).

The creation of multidisciplinary or agency teams with TANF and the development of neighborhood services or opening satellite offices appear to be the most recent efforts to restructure services; regionalization of services and creation of specialized service units seem to be older strategies, with increasing numbers of agencies gradually implementing them over the past three to four decades.

There is some evidence that state-administered agencies (65%) were more likely to have begun offering neighborhood services or opened satellite offices than were county-administered agencies (6%) (x2 = 4.06, p < .05). There is also some evidence that urban counties (73%) were more likely than rural ones (28%) to offer such services (x2 = 3.93, p = .05). Finally, it appears that urban counties (73%) were more likely than rural counties (27%) to have created multidisciplinary teams with TANF (x2 = 4.23, p = .04). No differences were noted with regard to county size or relative proportion of poor families.

The majority (87%) of agencies has already implemented concurrent planning (see Table 2-6)—which involves making plans for reunification and adoption at the same time and is expressly allowed by the Adoption and Safe Families Act (ASFA). There is some evidence that large counties (70%) were less likely to have completed implementation of concurrent planning than were other counties (90%) (x2 = 6.49, p = .04) and that urban counties (65%) were less likely to have completed implementation of concurrent planning than were rural counties (92%) (x2 = 6.67, p = .04). No differences were noted with regard to relative proportion of poor families or whether the agency was county or state administered.

Table 2-6. Status of concurrent planning (%)

Concurrent planning in agency	Percent (CI Limits)
Already implemented	87 (-14, +8)
Begun but not completed	9 (-5, +11)
No	5 (-4, +13)

 

2.2 Staffing and Training

2.2.1 Required Training and Academic Background

The LAS collected detailed data on agency staffing and training resources, including information on types of workers employed in the agency and their educational levels. No significant differences were found among agencies on requirements for workers in relation to college degrees. Agencies almost universally required a college education for investigators, those workers who make the sensitive decisions about whether a case will be opened and which services will be provided. The majority of agencies required workers of other types to have a college degree as well (Table 2-7). Though information on years of experience was not collected, it is likely that years of experience substituted for degree requirements for some portion of the sample (e.g., an MSW or a bachelor’s degree and a set number of years of experience).

Agencies were asked what degrees were required for various types of workers. Because the question posed was “What degrees are required?”, respondents could indicate more than one type of degree. Consequently, only the degrees that are possible were known, not the distribution of degrees held by the workers in the agency sample.

<

Table 2-7. Lowest degree accepted for workers (%)

	Investigators	In-home service workers	Foster care workers	Adoption workers
	% (CI)	% (CI)	% (CI)	% (CI)
No degree requirement	3 (-2, +8)	13 (-8, +16)	11 (-7, +16)	14 (-8, +16)
Other bachelor's degree	80 (-22, +12)	81 (-19, +11)	83 (-18, +10)	74 (-22, +14)
Bachelor's in social work (BSW)	16 (-10, +23)	6 (-5, +16)	6 (-5, +15)	12 (-9, +23)
Other master's degree	0	0	0	0.17
Master's in social work (MSW)	1 (-1, +5)	0.2 (-0, +1)	0.09 (-0, +1)	0

 

A social work degree was generally not required. A master’s in social work was rarely needed to begin work as a child welfare worker. The lowest degree required for all types of workers was most frequently a bachelor’s unrelated to social work (Table 2-7). This lower degree requirement may be due to the previously mentioned practice of accepting years of relevant experience in lieu of a higher degree.

On-the-Job Training Requirements

Training of child welfare personnel also occurs on the job. The majority of agencies (91%) required new workers to have four or more days of pre-service training. There was some evidence of a difference between agencies serving nonpoor counties and poorer counties (x2 = 6.44, p < .05). Agencies serving poorer counties (i.e., those with greater than 5% of the county’s children living below the 50% poverty level) tended to require more pre-service training, with 81% requiring more than two weeks pre-service training for new workers, whereas only 32% of agencies serving nonpoor counties required that much training. No differences were noted with regard to agency administration, county size, or degree of urbanicity. For required in-service training, no differences were found between agencies that were associated with county characteristics. Less in-service than pre-service training was required (Table 2-8).

Table 2-8. Amount of required training for new workers

	Pre-service	In-service
	% (CI)	% (CI)
None	2 (-2, +8)	20 (-12, +21)
Less than 1 day	0.0007 (-0, +0)	51 (-24, +23)
1-3 days	7 (-5, +12)	24 (-13, +20)
4-10 days	38 (-23, 30)	3 (-3, +13)
11-14 days	53 (-25, +23)	2 (-1, +4)
15 days or more	0	0
TOTAL	100	100

Caregiver Training Requirements

Counties of different characteristics also showed differences in caregiver pre-service training requirements. Caregivers for children under the agency’s supervision, even if not employees of the agency, may have also been required to obtain training. There was some indication of a difference between state- and county-administered agencies on required pre-service training for licensed or approved kinship caregivers (x2 = 4.7, p = .10). Most noticeably, 98% of county-administered agencies required little (less than one day) or no training for licensed or approved kinship caregivers, whereas state-administered agencies tended to require more training. Only 58% of state-administered agencies required less than one day of pre-service training for licensed or approved kinship caregivers. No differences were noted with regard to county size, degree of urbanicity, or relative proportion of poor families.

For required pre-service training for caregivers, no differences were found between agencies on any of the four comparison variables. Caregivers include licensed and nonlicensed kinship caregivers, foster care, and adoptive families. A total of 61% of all agencies tended to require less than one day of pre-service training for caregivers (Table 2-9).

Table 2-9. Average amount of required pre-service training days for caregivers (%)

Training amount	Percent (CI Limits)
Less than 1 day	61 (-23, +9)
1-3 days	22 (-13, +22)
4 or more days	17 (-10, +20)
TOTAL	100

Tables 2-10 through 2-12 provide information from all agencies about pre-service training requirements broken out by the various types of caregivers. Agencies required the majority of caregivers to receive at least a minimum level of pre-service training. The major exception was for nonlicensed caregivers, for whom only about 1% of agencies reported a pre-service training requirement (Table 2-10).

The majority of agencies (78%) required kinship caregivers to receive less than one day of pre-service training. (All of the agencies reported less than one day of pre-service training for nonlicensed kinship caregivers; consequently, this group was not included in Table 2-11.) Agencies were more likely to require additional pre-service training for foster and adoptive parents.

Adoptive and licensed foster parents had greater training requirements prior to having a child placed in their care, whereas in-service training requirements for caregivers tended to be for less than one day. A total of 100% of agencies reported less than one day of in-service training for nonlicensed kinship caregivers. This group was not included in Table 2-12 (p. 22).

Table 2-10. Pre-service training requirement for caregivers (%)

	Approved kinshipcaregivers
	Licensed	Nonlicensed	Foster parents	Adoptive parents
	%	%	%	%
	(CI)	(CI)	(CI)	(CI)
Yes	76 (-19, +12)	1 (-1, +4)	98 (-6, +1)	87 (-14, +7)
4 or more days	24 (-12, +19)	99 (-4, +1)	2 (-1, +6)	13 (-7, +140)
TOTAL	100	100	100	100

Table 2-11. Amount of pre-service training required for caregivers (%)

	Licensed/Approved Kinship caregivers	Foster parents	Adoptive parents
	%	%	%
	(CI)	(CI)	(CI)
Less than 1 day	78 (-24, +14)	60 (-24, +20)	41 (-20, +23)
1-3 days	14 (-10, +23)	24 (-14, +22)	42 (-23, +28)
4 or more days	8 (-6, +24)	16 (-10, +20)	17 (-10, +20)
TOTAL	100	100	100

Table 2-12. Amount of required in-service training for caregivers (%)

	Licensed/Approved Kinship caregivers	Foster parents
	%	%
	(CI)	(CI)
Less than 1 day	97 (-5, +2)	95 (-12, +4)
1-3 days	2 (-2, +5)	4 (-3, +12)
4 or more days	1 (-0, +1)	1 (-1, +4)
TOTAL	100	100

2.2.2 Use of Structured Decision Making or Risk Assessment

One way to complement training and bring consistency to child welfare agencies is to provide tools for decision making so that similarly situated children receive similar services that best protect their interests. Agencies often call these “risk assessment instruments” or “structured decision making” procedures because they help provide a framework for child welfare practice (Shlonsky & Gambrill, 2001). Agencies did not differ significantly in terms of implementation of risk protocols and approaches. Although almost every agency indicated that it used “structured decision making,” as Table 2-13 (p. 22) indicates, only 7% of agencies used an approach that resulted in point totals to help determine decision making. Another 36% used some sort of guidelines that indicated which issues should be addressed in the investigation/assessment. A total of 27% used a combination of point totals and guidelines approach. Finally, 30% of agencies used some other risk assessment approach.

There was some indication of a difference between state- and county-administered agency use of a risk assessment approach when deciding whether a case was substantiated (x2 = 3.37, p = .07). State-administered agencies were more likely to use a risk assessment approach: 93%, versus 36% of county-administered agencies. No differences were noted with regard to county size, degree of urbanicity, or relative proportion of poor families.

There was also some indication of a difference between state- and county-administered agency use of a risk assessment approach when deciding whether to reunify a child with the caregiver/family from which the child was removed (x2 = 3.44, p = .07). State-administered agencies (96%) were more likely to use a risk assessment approach than county-administered agencies (37%). No differences were noted with regard to county size, degree of urbanicity, or relative proportion of poor families.

As indicated in Table 2-14, agencies used a risk assessment approach at a variety of times, with most (96%) using one when deciding what services to provide. Far fewer agencies also used a risk assessment when making decisions about whether to close in-home services cases, reunify a child, or substantiate a case. No significant differences were found among agencies in their reasons for risk assessment use.

Table 2-13. Risk assessment approaches used (%)

	Percent (CI Limits)
Structured decision making with point totals	7 (-5, +16)
Guidelines	36 (-18, +24)
Combination of points and guidelines	27 (-15, +22)
Other	30 (-21, +35)
TOTAL	100

Table 2-14. Point in child welfare case at which risk assessment approach is used (%)

	Percent (CI Limits)
When deciding to investigate	59 (-20, +17)
When deciding whether the case is substantiated	88 (-14, +7)
When deciding what services to provide	96 (-6, +2)
When deciding whether to close an in-home services case	94 (-9, +4)
When deciding whether to reunify a child	92 (-9, +4)
At other times	54 (-29, +25)

The level of risk assessment use was calculated by totaling the reasons (listed in Table 2-14) agencies reported using a risk assessment. No significant differences were found among agencies on the level of risk assessment use (see Table 2-15). A little more than half of agencies (54%) reported five to six different reasons for the use of a risk assessment approach.

Table 2-15. Level of risk assessment use (%)

	Percent (CI Limits)
Low (0 - 2 uses)	31 (-21, +34)
Medium (3 - 4 uses)	15 (-8, +15)
High (5 - 6 uses)	54 (-25, +23)

Turnover for Child Welfare Workers

The extent of worker turnover and the effect this might have on the quality of service is a source of concern in the child welfare field. Although there have been numerous news reports of turnover rates of 50% per year and higher, neither the source of these estimates nor the procedures for calculating them is generally described; neither appear to be standardized.

Turnover rate is defined as the rate of separation from an employing organization. “The standard method the Bureau of Labor Statistics (BLS) uses to compute net monthly turnover rates is: the number of separations during the month, divided by the average number of employees on the payroll during that month, times 100” (International Personnel Management Association: http://www.ipma-hr.org/pubs/turnover.html). Because monthly figures were unavailable, the researchers used annual figures to calculate a turnover rate defined as follows: the annual number of separations, divided by the average annual number of employees, multiplied by 100.

The weighted estimates indicate that the average number of workers who separated during the most recent fiscal year was eight. Using the above formula, this translates into an average turnover rate of 13% (range = 0% to 86%, n=38)—considerably lower than what had been expected, given the journalistic accounts. The turnover rate was not significantly associated with county size, type of agency administration, urbanicity, county wealth, degree of local autonomy, or number of agency innovations.

2.2.3 Staff Resources and How They Have Changed

State and local funding for child welfare services and child welfare services’ caseloads have grown considerably in recent decades. As child abuse reports and foster care caseloads have expanded, many agencies have received funding for additional positions, although the extent of the growth is not really known. Nor is much known about the ability of agencies to fill positions and whether their capacity has declined or increased.

The researchers projected that child welfare agencies had an average of 55 authorized full-time equivalent (FTE) positions. Approximately 32% of these positions were noninvestigative direct service workers (e.g., in-home services or foster care workers), 24% were CPS investigative workers, and 13% were casework supervisors—the remainder being administrative and other workers (see Table 2-16). On average, 97% of authorized positions were filled: almost 100% (99%) of supervisor positions, 97% of direct service positions, and 95% of investigative positions. Between the beginning and the end of the fiscal year in question, the proportion of filled FTEs did not change substantially for any category of worker.

Table 2-16. Hiring of staff

Staff and staff vacancies	Mean (SE Mean)
Proportion of authorized positions: Direct service workers	0.32 ( 0.09)
Proportion of authorized positions: CPS investigative workers	0.24 ( 0.04)
Proportion of authorized positions: Casework supervisors	0.13 ( 0.01)
Proportion of authorized positions filled: Total	0.97 ( 0.02)
Proportion of authorized positions filled: Direct service workers	0.97 ( 0.02)
Proportion of authorized positions filled: CPS investigative workers	0.95 (0.04)
Proportion of authorized positions filled: Casework supervisors	0.99 (0.02)
Change in vacancy rate: Total	-0.01 (0.01)
Change in vacancy rate: Direct service workers	-0.01 (0.01)
Change in vacancy rate: CPS investigative workers	0.01 (0.01)
Change in vacancy rate: Casework supervisors	-0.00 (0.00)

 

There is strong evidence that large counties had a higher proportion of authorized direct service positions (.53) than did other counties (.28) (t = -2.55, p = .01). There is also some evidence that large counties had a lower proportion (.16) of authorized CPS positions than did other counties (.26) (t = 2.11, p = .04) and that urban counties had a lower proportion (.16) of authorized CPS positions than did rural counties (.27) (t = 2.07, p = .04). No differences were noted with regard to relative proportion of poor families or agency administration.

Table 2-17. PSU characteristics and CWS staffing patterns

	Higher CPS to direct services	Higher direct services to CPS	Higher supervisor to direct services
	Ratios	Ratios	Ratio
State/county administered			County
Larger/smaller size county	Smaller		Smaller
Urban/rural	Rural		Rural
Poor/nonpoor

To understand the configuration of child welfare services, the researchers examined proportions of CPS workers to direct service workers, CPS workers to supervisors, and supervisors to direct service workers (these are summarized in Table 2-17).

There is strong evidence that the ratio of CPS workers to direct service workers (t = 2.85, p = .006), as well as the ratio of child welfare supervisors to direct service workers (t = 2.59, p = .01), was higher in rural counties than in urban counties. There is also strong evidence that the ratio of CPS workers to direct service workers was significantly lower in large counties than in other counties (t = 2.62, p = .01) and some evidence that the ratio of supervisors to direct service workers was lower in large counties (t = 2.28, p = .03). There is only some evidence that the ratio of supervisors to direct service workers was higher in county-administered agencies than in state-administered agencies (t = -2.08, p = .04).

Child welfare staff were not increasing during this period (see Table 2-18). In fact, the vast majority of agencies reported experiencing no change in the number of authorized positions reported (regardless of type) during the course of the most recent fiscal year (mode=0). Overall, however, there was an average increase of just one authorized position per agency. The just/per agency number of noninvestigative direct service worker positions showed the greatest average increase (.44). No differences were noted with regard to county size, degree of urbanicity, relative proportion of poor families, or agency administration.

Table 2-18. Changes in child welfare staff size

Change in number of authorized positions	Mean (SE Mean)
Total	0.9 (0.47)
Direct service workers	0.44 (0.27)
CPS investigative workers	0.19 (0.13)
Casework supervisors	0.04 (0.03)
Administrative workers	0.14 (0.14)
Other workers	0.04 (0.08)

2.2.4 Subcontracting of Services

“Privatization” of public services has been an important trend in some sectors and has certainly received much attention in the child welfare field (e.g., Smith, 1989; Petr & Johnson, 1999). Although there has been substantial private provision of residential care throughout the history of child welfare services in the United States, privatization has also grown to include significant amounts of family preservation, treatment foster care, and adoption home-finding services in recent years. The full extent of privatization in child welfare services, however, has not been known.

In child welfare agencies nationwide, family preservation/in-home services was the most commonly subcontracted service, done by 58% of agencies (see Table 2-19). Residential treatment was the second most commonly subcontracted service, with 42% of agencies using such an arrangement. Subcontracting for other types of services, however, was relatively infrequent; investigative services, in particular, were contracted out by only .1% of agencies. In general, agencies were more likely to subcontract for service delivery than for case management services—with the notable exception of adoptive placements, for which 52% of agencies reported subcontracting case management.

In general, it appears that subcontracting was most often associated with agencies in large, urban counties. For example, there is strong evidence that subcontracting for family reunification services was related to county size, with large counties (54%) significantly more likely than others (8%) to subcontract for such services (x2 = 6.80, p = .01). Subcontracting for family reunification services was also associated with degree of urbanicity, with urban counties (39%) more likely than rural counties (8%) to contract out this service (x2 = 4.51, p = .04).

There is some evidence that subcontracting for foster care services was likewise associated with urbanicity and county size: urban counties (41%) were more likely than rural ones (8%) (x2= 4.65, p = .03), and large counties (44%) more likely than other counties (12%), to subcontract for foster care services (x2= 4.95, p = .03). Large agencies (81%) were more likely than other counties (35%) to subcontract residential treatment services (x2 = 4.75, p = .03). Subcontracting for adoptive placement services was also linked to county size and urbanicity: large counties (51%) were more likely than other counties (7%) (x2 = 5.30, p = .02), and urban agencies (37%) more likely than rural agencies (7%), to use such an arrangement (x2 = 3.99, p <.05). Finally, there is some evidence that subcontracting of recruitment of foster/adoptive parents was related to degree of urbanicity, with agencies in urban counties (40%) more likely than rural counties (4%) to contract out such activities (x2= 4.82, p = .03).

Table 2-19. Subcontracting of child welfare service functions (%)

Subcontracting by type of service	Percent yes (CI Limits)
Investigation/assessment	0.001 (-0, +1)
Family preservation/in-home services	58 (-28, +24)
Family reunification services	16 (-9, +17)
Foster care	16 (-10, +20)
Residential treatment	42 (-23, +27)
Adoptive placements	13 (-9, +19)
Recruitment	12 (-7, +16)

 

In addition to county size and urbanicity, type of agency administration may also be associated with subcontracting of services. There is some evidence that state-administered agencies (26%) were more likely than county-administered (2%) to subcontract recruitment services (x2 = 4.40, p = .04) as well as case management of adoption placements (state-administered: 69%; county-administered: 0%; x2 = 5.07, p = .03). There was no indication that subcontracting of services was associated with relative proportion of poor families.

2.3 Services and Service Dynamics

This series of questions and analyses sought to describe the variation in child welfare service patterns and transitions and how those are related to county characteristics. To better understand child outcomes, service patterns within child welfare must be documented, creating a baseline regarding the types of care provided to children. From an equity perspective, it is also important to document variations in the amount and types of intervention that may be associated with agency and/or community characteristics.

2.3.1 Referral for Investigation

On average, 65% of all child abuse reports were referred for investigation. There is strong evidence that the proportion of reports not investigated relative to the number of poor families with children in the county differed by county poverty levels: nonpoor counties investigated a lower proportion of reports relative to the number of poor families in the county (.13), than did poor counties (.05) (t = 2.60, p = .01), possibly indicating that poor families in nonpoor counties are more likely to be investigated than are poor families in poorer counties. There was some evidence that the ratio of reports not investigated to the total number of reports also differed by county poverty levels: nonpoor counties had an overall higher rate of reports not investigated (.44) than did poor counties (.30) (t = 2.11, p = .04). There was also some evidence that the ratio of total number of reports to the number of poor families with children in the county differed by county poverty levels: nonpoor counties had an overall higher rate of reports (.30) than did poor counties (.16) (t = 2.46, p = .02). No differences were noted with regard to county size, degree of urbanicity, or agency administration. When allegations were unsubstantiated, 94% of agencies referred families to voluntary services.

2.3.2 Family Preservation and Family Support

Family preservation services are typically designed to help families at risk or in crisis. Services may be designed to prevent foster care placement, reunify families, or support adoptive families. This report defines family support services as those services that are primarily community-based preventive activities designed to promote the well-being of children and families. Services are designed to increase the strength and stability of families, increase parents’ confidence and competence in their parenting abilities, afford children a stable and supportive family environment, and otherwise enhance child development.

Table 2-20. Family preservation and family support use

Receipt of family preservation and family support services	Mean (SE Mean)
Number in family preservation/child population	0.01 (0.00)
Number in family support/child population	0.01 (0.00)
Number in family preservation/number poor families with children	0.05 (0.01)
Number in family support/number poor families	0.07 (0.02)

 

The researchers’ weighted estimates indicate that child welfare agencies provided family preservation services to an average of 248 families per year and family support services to an average of 325 families per year. This means that about 1 of every 100 children received family preservation services and that about 1 of every 100 children received family support services (this assumes that only 1 child per family received the service). This represents an average of 5 children of every 100 children in families below poverty level receiving family preservation services and 7 children of every 100 children in families below poverty level receiving family support services (see Table 2-20).

There is strong evidence that nonpoor counties had a higher rate of family preservation services relative to the number of families with poor children (.07) than did poor counties (.01) (t = 3.38, p = .001). There is some evidence that nonpoor counties had a higher rate of family preservation services relative to the child population (.01) than did poor counties (.002) (t = 2.34, p = .02). There is also some evidence that nonpoor counties had a higher proportion of family preservation services relative to total number investigated (.60) than did poor counties (.06) (t = 2.06, p = .04). In addition, there is some evidence that urban counties had a higher overall rate of family preservation services (.02) than did rural counties (.004) (t = -2.31, p = .02). No differences were noted with regard to county size or agency administration.

2.3.3 Voluntary Placement

Voluntary placement of children into foster care was quite uncommon, representing only 0.3% of the total investigated. An average of only four voluntary custody placements occurred per year per agency, with a median of zero. There is strong evidence that the rate of voluntary placement relative to the number of poor families with children in the county was associated with urbanicity, with urban counties reporting a higher average rate (.01) than rural counties (.0003) (t = -2.61, p = .01). No differences were noted with regard to county size, relative proportion of poor families, or agency administration.

Table 2-21. Out-of-home placement rates

Out-of-home placement rates	Mean (SE Mean)
Total children in care/child population	0.01 (0.00)
Total children in care at year's end/poor families with children	0.03 (0.01)

 

2.3.4 Children in Out-of-Home Care

At the end of the most recent fiscal year, the researchers project that the mean number of children in out-of-home care was 91 per child welfare agency. This represents an average out-of-home placement rate of about 1 child per 100 children and about 3 children per 100 children living in families below poverty level (see Table 2-21).

2.3.5 Specialized Care

There is evidence from several states that out-of-home care is changing in character and that a growing proportion of children are no longer residing in traditional nonrelative foster care at the traditional foster care board rate. Instead, growing proportions of children are being served in kinship foster care, in foster care with specialized board rates, and in treatment foster care.

The researchers estimate that foster care expenditures accounted for almost half of all child welfare expenditures (45%). The average annual per-child board rate for out-of-home placements was $7,283 (not including administrative costs).

All agencies gave preference to relatives when considering placement options. In terms of adoption, in about three-quarters (76%) of the PSUs, relatives were recruited as foster and adoptive parents. In the remaining PSUs, relatives were apparently not considered a priority for foster care and adoption. About two out of three (63%) agencies expected relative caregivers who plan to care for a child for a long time to become the legal guardian or adoptive parent. Only about one-third indicated that they accepted a plan of long-term foster care by relatives without first providing encouragement and expectations to provide legal permanence.

A variety of kinship care arrangements was reported. The most common ones provided TANF for the child in care but no foster care payment (24% of all children in kinship care were identified as having this type of arrangement) or provided the relative with a foster care payment while requiring the relative to meet standard licensing requirements (23%). About 12% received TANF for both the caregiver and child. Other varieties of arrangements were rare: only 1% of placements involved no assistance, and virtually none were with nonrelative or “fictive kin” caregivers that involved no payment (0.5%) or provided a foster care payment without the caregiver having to meet normal licensing requirements (0.1%). No placements were assisted kinship guardianships. About half (52%) of kin placements received specialized/difficulty-of-care payments for children with behavioral, emotional, developmental, or medical special needs. There is some evidence that a higher proportion of foster care homes from county-administered agencies received specialized payments (64%) than foster care homes from state-administered agencies (25%) (t = -1.99, p < .05). No differences were noted with regard to county size, degree of urbanicity, or relative proportion of poor families.

There is some evidence that the use of fictive kin placements was related to county size and urbanicity: large counties (66%) were more likely than others (8%) (x2 = 4.15, p < .05) and urban counties (57%) more likely than rural (6%) to use such a placement (x2 = 4.82 p = .03). There is some evidence that receipt of TANF only for the child in care was related to urbanicity: kin caregivers in urban counties (90%) were more likely than those in rural ones (20%) to receive TANF for the child but no foster care payment (x2 = 5.24, p = .03). There is some evidence that provision of a foster care payment plus normal licensing requirements was also related to urbanicity: urban counties (81%) were more likely than rural ones (19%) to require licensing and provide a foster care payment (x2 = 5.49, p = .02). There is strong evidence that state-administered agencies (84%) were more likely than county-administered agencies (7%) to require licensing and provide a foster care payment (x2 = 7.66, p < .01). No differences were noted with regard to relative proportion of poor families.

2.4 Client Characteristics and Caseload Dynamics

2.4.1 Response to Policy Changes

The last decade had an unprecedented set of major child welfare policy initiatives, including the Adoption and Safe Families Act and the Multiethnic Placement Act (MEPA). Agencies are required to implement procedures consistent with these policies.

Adoption

Adoption-related activity has increased sharply in recent years (AFCARS, 2000). These agencies reported that on the first day of the most recent fiscal year, the mean number of children in care with a goal of adoption was 46 (68% of whom were legally free for adoption). By the end of that fiscal year, an average of 33% of those with an adoption goal were placed for adoption, and an average of 22% were legally adopted (see Table 2-22).

Table 2-22. Adoption dynamics

Adoption services	Mean (SE Mean)
Children legally adopted/number with goal of adoption	0.22 (0.07)
Children placed for adoption/number with goal of adoption	0.33 (0.11)
Children legally free for adoption/number with goal of adoption	0.68 (0.09 )

 

There is strong evidence that state-administered agencies had higher rates of adoption placements (as compared to the number of children legally free for adoption) (.54) than county-administered agencies (.17) (t = 2.53, p = .01). There is strong evidence that the rate of legal adoption (as compared to the number of children legally free for adoption) differed by type of agency administration, urbanicity, and county poverty: state-administered agencies had a higher rate (.37) than did county-administered agencies (.08) (t = 3.13, p = .002); urban counties had a higher rate (.41) than rural counties (.14) (t = -2.51, p = .01); and nonpoor counties had a higher rate (.36) than poor counties (.10) (t = 2.76, p = .007). No differences were noted with regard to county size.

Disproportionality of African American Children in Care

There is considerable federal, state, and local concern that the proportion of African American children in foster care is substantially greater than the proportion of African American children in the general public. A number of state and local agencies have undertaken initiatives to address this perceived racial imbalance. These analyses are intended to provide information about what is being done.

Weighted estimates indicate that only a small minority (14%) of agencies had identified as a concern the over- or under-representation of African American children in certain services, such as foster care or community-based services, while 86% had not identified it as a concern (see Table 2-23).

Table 2-23. Addressing representation
of African American children

Concerns about and initiative to address over-representation	Percent yes (CI Limits)
Concerns about over- or under-representation of African American children	14 (-7,+14)
Initiatives to address concerns: Training	15 (-11,+26)
Initiatives to address concerns: Efforts to racially match child welfare workers and families	10 (-8,+25)
Initiatives to address concerns: Performance measures	2 (-1,+8)

 

Likewise, only 15% of agencies were involved in training initiatives designed to address over- or under-representation of African American children in services, while 10% made efforts to racially match child welfare workers and families. Two percent used performance measures to reduce racial imbalance in placement.

There is strong evidence that concerns about over- or under-representation of African American children were associated with county size and urbanicity: large counties (54%) were significantly more likely than others (6%) (x2 = 9.12, p = .003), and urban counties (45%) more likely than rural (5%), to have such concerns (x2 = 9.32, p = .003). There is some evidence that state-administered agencies (25%) were more likely than county-administered ones (4%) to have concerns about representation of African American children (x2 = 4.36, p = .04). No differences were noted with regard to relative proportion of poor families.

There is also strong evidence that participation in special training initiatives to address over- or under- representation of African American children was associated with county size: large counties (58%) were significantly more likely than others (6%) (x2 = 7.79, p = .007) to participate in such trainings. State-administered agencies (52%) were also significantly more likely than county-administered ones (2%) to participate (x2 = 7.48, p = .008). There is some evidence that poor counties (49%) were more likely than nonpoor counties (4%) to participate in these initiatives (x2 = 5.50, p = .02). No differences were noted with regard to degree of urbanicity.

2.5 Budget and Expenditures of Child Welfare Agencies

Child welfare expenditures differ in amount, program focus, and flexibility. Each could be a significant contributor to the types of services delivered and the outcome achieved.

(The study researchers’ experience with the data collection was that these estimates of expenditures varied in accuracy—as child welfare expenditures come from many sources, and there is no standard way of accounting for child welfare expenditures.)

On average, child welfare agencies spent a total of $5,986,411 each on child welfare services in the most recent fiscal year. This represents an average per-child expenditure rate of $8,234 (the standard error is $1,964). On average, foster care services and group care accounted for the largest percentage of total spending (45%), followed by child protective services (28%) and in-home services—commonly known as “family preservation” (14%). Family support services (primary child-abuse prevention) accounted for 5% of total expenditures, and adoption services accounted for 7%. The smallest category, at 1%, was independent living services.

There is strong evidence that nonpoor counties had a significantly higher average per-child child welfare expenditure ($10,739) than did poor counties ($2,689) (t = 2.73, p = .008). In addition, there is strong evidence that the ratio of CPS dollars spent to the total number of children investigated was significantly higher for county-administered agencies ($1,192) than for state-administered agencies ($487) (t = -2.70, p = .008); it was also significantly higher for nonpoor counties ($1,046) than for poor counties ($307) (t=3.64, p=0.001). No differences were noted with regard to county size or degree of urbanicity.

2.5.1 Changes in State Funding

Local agency directors indicated the extent to which their states had increased funding for a variety of programs. The majority reported no changes in state funding, and a minority had decreases. The most common area experiencing a decrease (16%) was child abuse prevention services, which had also been reported to increase in many PSUs, indicating that this area had the greatest variability (see Table 2-24). Flexibility in using state funds for services stayed the same for the majority of agencies, and an estimated 27% noted an increase in flexibility. The largest proportion of PSUs indicated that state funding had increased for child abuse prevention services (45%) and adoption services (42%); child protective services (35%), in-home services (24%), and placement services (20%) had also increased but in somewhat fewer PSUs.

There is strong evidence that county-administered agencies (92%) were more likely than state-administered agencies (24%) to have no changes in state funding for adoption (x2 = 13.43, p = .002), while state-administered agencies (76%) were more likely than county-administered agencies (8%) to have an increase. There is also some evidence that county-administered agencies (93%) were more likely than state-administered ones (36%) to have no changes in state funding for in-home services (x2 = 6.56, p = .04); state-administered agencies were more likely to experience both increases and decreases. No differences were noted with regard to degree of county size, urbanicity, or relative proportion of poor families.

Table 2-24. Funding of child welfare services in prior year (%)

Changes in state funding	Percent (CI Limits)
Funding changed past 12 months (child abuse prevention services)
Decreased	16 (-11, +23)
Remained unchanged	39 (-19, +23)
Increased	45 (-24, +26)
Funding changed past 12 months (child protective services)
Decreased	12 (-9, +25)
Remained unchanged	53 (-24, +23)
Increased	35 (-18, +24)
Funding changed past 12 months (in-home services)
Decreased	12 (-9, +25)
Remained unchanged	64 (-24, +19)
Increased	24 (-13, +21)
Funding changed past 12 months (placement services)
Decreased	9 (-7, +29)
Remained unchanged	71 (-24, +16)
Increased	20 (-11, +20)
Funding changed past 12 months (adoption services)
Decreased	0.003 (-0, +1)
Remained unchanged	58 (-25, +22)
Increased	42 (-22, +25)
Past 12 months agency's flexibility using state funds for services:
Decreased	10 (-8, +27)
Remained unchanged	64 (-24, +18)
Increased	27 (-14, +21)

2.6 Changes in Child Welfare Services

It is important to examine the effects of changing federal policy on the experience and functioning of local child welfare agencies. In addition to giving a national view of the effects of changing federal policy on agencies, this analysis provides background and insights for the later analyses of data on agency practices and services for children and families in the child welfare system.

2.6.1 Effects of Temporary Assistance for Needy Families (TANF)

The study researchers project that most child welfare agencies had few changes in service delivery programs after the welfare reform implementation of TANF that began in late 1996. For example, employment services became routinely linked to child welfare services for only an estimated 24%, and for only 22% were referrals from TANF to child welfare services routinely made when clients were sanctioned (see Table 2-25). The most common change was the creation of multiprogram teams that include both TANF and CPS caseworkers (38%).

There is some evidence that type of agency administration was related both to the creation of multiprogram teams with both TANF and CPS caseworkers: state-administered agencies (66%) were more likely than county-administered ones (14%) to create such teams (x2 = 4.33, p < .05). Likewise, state-administered agencies (46%) were more likely than county-administered agencies (2%) to routinely make referrals from TANF to child welfare services when clients were sanctioned (x2 = 4.69, p = .03). No differences were noted with regard to county size, degree of urbanicity, or relative proportion of poor families. For a little more than half of the agencies, after the implementation of TANF, the number of child welfare cases handled by the agency increased (52%); and for a little less than half (45%), the number of hours spent on a child welfare case increased.

Table 2-25. Effects of TANF on child welfare services (%)

Effects of TANF	Percent yes (CI Limits)
Multiprogram teams created that include both TANF and CPS caseworkers	38 (-20, +26)
Employment services linked to child welfare services	24 (-15, +27)
Referrals from TANF to child welfare services made when clients sanctioned	22 (-14, +25)
Other changes	17 (-11, +23)
No changes	44 (-25, +28)
Number of child welfare cases handled by the agency:
Decreased	5 (-4, +20)
Remained unchanged	43 (-23, +26)
Increased	52 (-26, +25)
Number of hours spent on a child welfare case:
Decreased	0 (N/A)
Remained unchanged	55 (-29, +26)
Increased	45 (-26, +29)

 

2.6.2 Effects of the Adoption and Safe Families Act (ASFA)

The study researchers’ analyses indicate that child welfare agencies saw a greater impact from ASFA, implemented in 1997, than they did from TANF. For example, for about 60% of the agencies, ASFA led to a greater emphasis on ensuring the child’s safety (vs. family preservation approach), and for 93%, it shortened time frames for decision making to less than 12 months (see Table 2-26, p.39). After ASFA, 54% of agencies noted an increased emphasis on adoption for older children, whereas for almost three-quarters (74%), there was increased emphasis on adoption for children living in kinship foster care. Smaller changes resulted regarding expedited access to drug treatment for clients, reported in just 33% of agencies. A little more than one-quarter (28%) of child welfare agencies saw an increase in the number of families that were precluded from receiving reunification services after ASFA. This may be attributable to the implementation of policies that do allow no reunification orders for some children or because meaningful reunification was not considered feasible given the ASFA time frames—the data do not allow us to tell the difference between these two possibilities.

The majority of agencies (78%) experienced no change in the number of cases handled after the implementation of ASFA; similarly, 79% of agencies reported that AFSA had not affected the agency’s client base. On the other hand, for 60%, there was an increase in the average number of hours spent on a child welfare case (which they attributed to ASFA). Almost all agencies reported an increase in regulations and paperwork after ASFA.

There is some evidence that state-administered agencies (53%) were more likely than county-administered ones (7%) to have an increased number of families who would not get reunification services after ASFA (x2 = 4.61, p = .04). There is strong evidence that rural counties (84%) were more likely than urban counties (37%) to have increased the emphasis on adoption for children living in kinship foster care after ASFA (x2 = 7.12, p = .009). No differences were noted with regard to county size or relative proportion of poor families.

Table 2-26. Effects of ASFA on child welfare services (%)

Effects of ASFA	Percent yes (CI Limits)
Greater emphasis on child safety (vs. family preservation approach)	60 (-31, +24)
Shortened time frames for decision making to less than 12 months for some children	93 (-12, +4)
Increased number of families who will not get reunification services	28 (-16, +25)
Expedited access to drug treatment for clients	33 (-23, +35)
Increased emphasis on adoption for older children	54 (-27, +25)
Increased emphasis on adoption for children living in kinshipfoster care	74 (-21, +14)
Other changes	10 (-6, +12)
No changes	2 (-2, +10)
Number of cases handled:
Decreased	6 (-5, +17)
Remained unchanged	78 (-24, +14)
Increased	16 (-11, +25)
Number of hours spent on a child welfare case:
Decreased	0.002 (-0, +1)
Remained unchanged	40 (-21, +26)
Increased	60 (-26, +21)
Agency regulations and paperwork:
Decreased	0 (NA)
Remained unchanged	0.001 (-0,, +0)
Increased	100 (-0, +0)
Affected agency's client base	21 (-14, +26)
Other effects on services delivery	41 (-22, +26)

2.6.3 Effects of Multiethnic Placement Act (MEPA)

The Multiethnic Placement Act appears to have had the least impact on service delivery. Although 29% of agencies saw increased training, after MEPA, about the ways that race can be used in making foster care and adoption placements, 77% had no increase in the proportion of transracial foster care placements and no increase in the proportion of transracial adoption placements. Only 8% saw the creation of new recruitment resources (see Table 2-27).

Likewise, 100% of agencies experienced no change in the number of cases handled by the agency, and 98% reported no change in the agency’s client base. The great majority (97%) experienced no change in the average number of hours spent on a child welfare case, though 17% witnessed an increase in agency regulations and paperwork.

There is strong evidence that creation of new recruitment resources was associated with urbanicity, with urban counties (39%) significantly more likely than rural ones (0%) to experience this change (x2 = 8.58, p = .004). There is some evidence that creation of new recruitment resources was also related to county size, with large counties (48%) more likely than others (44%) to have such a change (x2 = 5.47, p = .02). There is also some evidence that county-administered agencies (94%) were more likely than state-administered agencies (38%) to see no changes in agency services (x2 = 5.95, p = .02). There is some indication that rural counties (75%) were more likely than urban counties (41%) to have no changes in agency services (x2 = 2.97, p = .09). No differences were noted with regard to relative proportion of poor families.

Table 2-27. Effects of MEPA and related
provisions on child welfare services (%)

Effects of MEPA	Percent yes (CI Limits)
Increased training about ways race can be used in making foster care and adoption placement	29 (-17, +26)
Creation of new recruitment resources	8 (-5 +11)
Increased proportion of foster care placements that are transracial	23 (-15 +27)
Increased proportion of adoption placements that are transracial	23 (-15, +27)
Other changes	1 (-1, +5)
No changes	67 (-26, +18)
Affected agency's client base	2 (-1, +9)
Number of cases handled by the agency:
Decreased	0 (N/A)
Remained unchanged	100 (-1, +0)
Increased	0.002 (-0, +1)
Number of hours spent on a child welfare case:
Decreased	0.002 (-0, +1)
Remained unchanged	97 (-8, +2)
Increased	3 (-2, +8)
Agency regulations and paperwork:
Decreased	1 (-1, +5)
Remained unchanged	82 (-22, +11)
Increased	17 (-11, +22)

 

2.6.4 Performance-Based Measures

In recent years, child welfare agencies increasingly have developed and implemented the use of performance measures to assess agency outcomes. (Federal outcome measures have also been instituted, although these data were collected before the final regulations were in place.) The study researchers estimated that a little more than half of child welfare agencies had begun using performance-based measures tied to financial incentives or sanctions (see Table 2-28, p. 42). The most common area in which performance-based measures were used was family preservation (38%), followed by investigation (32%). Roughly one-quarter used performance-based measures to assess family reunification services, length of time in foster care/out-of-home care, adoption services, reoccurrence of abuse or neglect, independent living services, use of least-restrictive placement, and permanency. Only 20% of agencies, however, used performance-based measures in the area of child fatalities. Implementation of performance measures occurred for the most part in the early to mid-1990s.

Table 2-28.Performance measures
introduced into childwelfare services (%)

Performance measures by area	Percent yes (CI Limits)
Investigation	32 (-22, +36)
Family preservation	38 (-26, +35)
Family reunification	26 (-19, +37)
Foster care	28 (-20, +37)
Adoption	24 (-18, +38)
Independent living	24 (-18, +38)
Abuse/neglect	26 (-19, +37)
Child fatalities	20 (-16, +39)
Permanency	27 (-20, +37)
Placements	24 (-18, +38)

 

There is some evidence that rural counties (63%) were more likely than urban counties (17%) to use performance-based measures that are tied to financial incentives (x2 = 4.48, p = .04). No differences were noted with regard to county size, relative proportion of poor families, or agency administration.

 

 

Table of Contents | Previous | Next