11

11

MICROCOPY RESOLUTION TEST CHART

NAIIONAL BUREAU 01 SIANDARDS-1961 A

DOCOBENT BESUMB

PS 010 732

ED 174 330
AUTHOR
TITLE

SPONS AGENCY
PUB DATE
NOTE

El:in PRICE

DESCRIPTORS

Rescorla, Leslie A.; And Others

The Yale Child Welfare Research Program: Description
and Results.
Ford Foundation, New York, N.Y.
Mar 79

46p.; Paper presented at the Biennial Meeting of the

Society for Research in Child Development (San
Francisco, California, March 15-18, 1979)
MF01/PCO2 Plus Postage.
Day Care Services; *Developmental Programs; Early
Childhood Education; *Family Programs; Home Visits;
*Intervention; Language Development; Longitudinal
Studies; *Low Income Groups; Medical Services;
*Program Effectiveness; Program Evaluation; Quality
of Life

ABSTRACT

The Yale Child Welfare Research Program was a

comprehensive, service centered, longitudinal, intervention project
from
for low-income families and their children. Eighteen children
participated in
predominantly
black
families
inner city, low-income,
Each
the intervention program from before birth to 30 months of age.
family was assigned a team of project staff members who provided
services and recorded their work with and observations of the
families. Data were collected on four major components of the
(3)
program: (1) home visitor program, (2) Pediatric care, school. One
developmental evaluation, and (4) day care and toddler
of 18
year after the project ended, a matched comparison sample
selected.
Each
comparison
mother
was
children 30 mcnths of age was
comparison
child
was
seen for a single in-depth interview and each
the
seen once for a developmental evaluation. Five years after of
the
interviews
were
conducted
with
15
project ended, follow-up
and
each
of
the
research
group
original research group mothers
children was seen in a follow-up testing session. Findings are
presented in detail. Results were interpreted as suggesting that the
intervention had served to offset the detrimental effects of
socio-cultural deprivation on the research group children's verbal
abilities and had helped the project families improve their quality
cf life. (JMB)

************************************************************************
Reproductions supplied by EDPS are the best that can be made
*
*
from the original document.
*
***********************************************************************

WITHOUT AUTHORS' PERMISSION

'UNPUBLISHED MANUSCRIPT NOT TO BE QUOTED
DEPARTMENT
OF HEALTH,
EDUCATIONS. WELFARE
NATIONAL INSTITUTE

US

EDUCATION

The Yale Child Welfare Research Program:

OF

THIS DOCUMENT
DUCED EXACTLY HAS BEEN REPROA5 RECEIVED
THE PERSON
FROM
DR DRG4N/Z4TrON
ORIGINATING IT POINTS
OF VIEW OR OPINIONS
STATED DO NOT NECESSARILY
SENT OFFICIAL
REPRENATIONAL NSTi TUTE
EDUCA TioN POSITION
OF
OR POLICY

Description and Results

Audrey Naylor
Leslie A. Rescorla, Sally Provence and

comprehensive inThis report presents some of the results of a

families and their children: the
tervention project for low-income
The program was carried out at
Yale Child Welfare Research Program.
1967-72 supported by a grant from the
the Yale Child Study Center from
(later the Office of Child Development).
United States Children's Bureau
detail in the Challenge of Daycare
The project has been described in
Sally Provence, M.D. and
(Provence, Naylor, and Patterson, 1977).
director and assistant project
Audrey Naylor, MSW served as project
data wazi
condensing, analyzing, and reporting the
director. The job of
psychologist who was hired severcarried out by Leslie Rescorla, Ph.D. a

(IN

al years after the project ended.

The data reported here were collected

five years after the project

02)during the intervention program as well as
for follow-up.
619 ended when the families were seen

of a number of ambitious,
The Yale Child Welfare Program was one
1'1
in the late Sixties to
(2)comprehensive demonstration projects undertaken
Syracuse Family Develop00 help disadvantaged families and children: the
,
Frank
Porter
Graham Child De(Lally, 1971), The
gag merit Research Program
Peabody Early Training Project
velopment Center (Robinson, 1968), the
"PERMISSION TO REPRODUCE THIS
MATERIAL HAS BEEN GRANTED BV

ts-QSq...

R. Recco-\a..

TO THE EDUCATIONAL RESOURCES

INFORMATION CENTER (ERIC)."

-2-

Project
(Klaus & Gray, 1968), the Ypsilanti Carnegie Infant Education
Childhood Stimulation
(Lambie, Bond & Weikart, 1974), the Florida Early
Washington, D.C. Infant Education
program (Gordon & Guinagh, 1977), the

1972), as well as many other proResearch Project (Schaefer & Aaronson,

intervention programs has been reviewed
grams. The literature on early
Only the
amply elsewhere (Brcinfenbrenner, 1974; Day & Parker, 1977).

major issues and findings in the area will be outlined here.

has been
The most commonly used index of program effectiveness
few programs with massive "ecochild's IQ. With the exception of a
(25-30
logical" interventions where IQ gains reported are enormous
of
points in the Heber program (Bronfenbrenner, 1974)) the majority

but modest level

early intervention programs have achieved a significant
(Klaus & Gray, 1968; Levenof IQ superiority relative to control groups
Lambie, Bond & Weikart, 1974;
stein, 1977; Schaefer & Aaronson, 1972;
A few studies have reported particular gains on
Lally & Honig, 1977).
language ability (Clarke-Stewart in press).
studies has been a fadA prominent outcome in many intervention
While some programs
ing out of IQ gains after the program terminates.
achievement test scores
report continued superiority on IQ measures or
1968; Gordon & Guinach, 1977;
by third and fourth grade (Klaus re:. Gray,
effects
Levenstein, 1977) many other studies have found that program

(Schaefer & Aaronson, 1977;

were no longer significant at follow-up
Though some studies
Lambie, Bond & Weikart, 1974; Lally & Honig, 1977).
continued enhave demonstrated that participation in follow-through or
4p1 (Abelson, 1974), other
richment programs served to maintain early r
still enrolled in the
studies reported IQ declines while children were

original intervention program (tally & Honig, 1977; Klaus & Gray,

968).

which involve parBronfenbrenner (1974) has argued that projects

programs without parental
ents are more effective than center-based
(1977) which involved
involvement. Projects such as Levenstein's
(Schaefer &
structured work with the mother-child dyad, or Schaefer's
mother to observe enrichment sessions
Aaronson, 1972) which provided for a
seen by Bronfenbrenner to
between her young child and a home tutor, are
(in press) has reshow more durable gains. However, Clarke-Stewart
It seems valid
cently argued that evidence on his point is equivocal.
that a structured, cognitive childto conclude from the literature
and early school compecentered program may maximize test performance
impact on family attitence but that programs which do not make some
will not produce longterm sustained
tudes and patterns of interacting
before school entrance
As Gray has said, "Intervention programs
gains.
(Ryan, 1974).
cannot carry the entire burden of improving educability.."
magic period for starting
While reviewers agree that there is no
when they involve
interventions, programs seem to be most effective
Bronfenbrenner
younger children (Bronfenbrenner, 1974; Clarke-Stewart).
for intervention is in the first
has concluded that the optimal period
family based program to enhance
three years of life. He proposes a
followed by a preschool
mother-child interaction in the early years,
child for school but still stresses
enrichment program which prepares the
Clarke-Stewart concurs with
the crucial value of paren involvement.
begin and the longer they run
Bronfenbrenner that the earlier programs
effective'they tend to be.
(up to three years duration), the more

has been dissatisfaction with

A repeated theme in the literature
Zigler
of program effectiveness.
IQ gain as the ultimate criterion
that gains in a com(Zigler & Trickett 1978) has recently proposed

-4-

is a more appropriate
posite of areas summarized as "social competence"
Clarke-Stewart has also recently argued
index of program effectiveness.
assessing outcomes,
for more "complex and*multi-variate strategies" for
Some profactors.
including emotional, motivational, and achievement
important outcomes
grams have reported positive changes in mothers as
& Honig, 1977),
(Karnes & Badger, 1969; Gordon & Guinagh, 1977; Lally
gains, becoming more
with mothers making'educational and occupational
constructive childinvolved in community affairs, and developing more
A recent report on a collaborative long-term followrearing approaches.
projects (Consortium, 1978) argues that
up of twelve early intervention
in both a statistical
the most significant benefits of preschool programs
children have shown less
and a social sense are that intervention group
special education classes than
grade retention and less assignment to
higher vocational
control children and that intervention mothers express
aspirations for their children than control mothers.
literature indiAs this brief summary of the early intervention
While significant intercates, findings are complex and equivocal.
studies, much remains to be
vention effects have been found in many
most effective for diflearned: which strategies of intervention are
impact yield the most
ferent areas of outcome; which areas of project

longstanding gains; how do different kinds of families and children

variables are the best
utilize interventions; and which factors or
predictors of intervention outcome. While the intervention program
involved fewer families than
reported here was of much smaller scale and
more intensive and clinmany programs reported in the literature, .1..ts
to illuminate some of
ically oriented case approach may be helpful
with progam impact.
these important and subtle issues associated

_5_

PROJECT DESCRIPTION
Goals of the Project
First, through various kinds
The project had three major goals.
human needs and of child developof intervention based on knowledge of
of human potential often
ment, we wanted to try to diminish the erosion
inadequate care in the earliest
associated with conditions of poverty or
Finding effective ways of helping young children whose developyears.

likely to become so was related to that

ment was already at risk or was
helping young parents with child rearing and
goal. We believed that
would increase their ability to rear
with the stresses they experienced

would be most effective if it behealthy children and that intervention

We expected that interest in pargan early in the life of the child.
would enhance the quality of
ents and appropriate assistance to them
that providing services in a
their own lives. Secondly, it was hoped
development of more effective methods
research context would lead to the
intellectual impairment and personality
of preventing or alleviating the
could be usefully disseminated
damage referred to above: methods which
providers. While it will not
to clinicians and other human services
goal was related to child developbe discussed in this paper, a third
that involvement of an interment theory in general. It was expected
project would permit the examinadisciplinary research team in such a
aspects of developmental
tion, extension and elaboration of certain
and a servicecentered investigation were
theory. A clinical approach
of the project.'
chosen as methods most suited to the goals
Theoretical Approach
activities can best be charThe approach to research and service
In this approach, clinical assessacterized as clinical-developmental.

went provided the basis for individualizing services to a child and

and constructs were used in
family. Additionally,, clinical methods
collecting and organizing the research data. The approach was developmental growth
mental as well -- developmental in its view of physical and
being in a family
and in its view of the child's progress as a social
and community.

The biological dimensions of our view can best be briefly stated

such conin terms of embryological developmental processes encompassing
The
cepts as endowment, phase specificity, maturation and adaptability.
interaction of innate and experiential factors as co-determinants of

development was an important theoretical construct of the approach.

concepts
The theoretical assumptions that underlie many of the
child psychology,
of development utilized are those of psychoanalytic
influence of the
in particular those propositions that concern the
emotional deparent-child relationship as manifested in, the child's
characteristics of thought,
velopment, his attitudes toward learning, his
learning person, and his
his view of himself as a thinking, feeling and
In this approach the child's
ability to form relationships with others.
material world is eminteraction with the human environment and the
spheres
phasized, as is the interaction between cognitive and affective
of development.

share the view

In regard to the education of young children we
cognitive functions- of Shapiro and Bibz:r (1973) that "the growth of
information, judging, reasoning, problem solving,

acquiring and ordering

of personal
using systems of sybols--cannot be separated from the growth
of self-esteem and a sense
and interpersonal processes--the development

t.

-7-

control, capacity
of identity, internalization of impulse
Educational
mous response, relatedness to other people."
mainly
conceived in terms of developmental processes, not
We
achievements on the way to specified accomplishments.

for autonogoals are

as concrete
have relied,

of intelligence, finding it
too, on Piagetts view of the development
guidelines for action.
helpful both theoretically and in developing

combined with an
The clinical and developmental viewpoints are
influence children and their
awareness that the societal systems that
health and social
families are expressed through institutions such as
be the concern of those who work to
services and schools, and must
and adults.
enhance the well-being of children

Subject Selection
The first phase was a
The project was divided into two phases.
The pilot proto 4 years.
pilot group of 23 children ages 14 months
practice working togram was used to recruit and train personnel, to
and to gain the benefits of exgether Aoo refine research methodology
and modifying tradiperience in trying new ways of providing services
the pilot group program will not be retional ones. The children in
this report are eighteen children from
ported here. The subjects of
participated in the intervention
seventeen low income families who

program from before birth to 30 months of age.

of working with the
It was recognized that the effectiveness
began before environmental infamilies could best be evaluated if it
during pregnancy or at
fluences had affected development, that is,
least at the time of birth of the infant.

First born children were

-8early in their parchosen, in order that parents might be influenced

enthood.

Mothers were selected from those registered for obstetrical

Haven Medical Center. A

care in the Women's Clinic of the Yale-New
those that
project staff member reviewed clinic records and selected
defined
(1) lived in the inner city, (2) were in the poverty group as

serious complications of pregnancy

by the federal government, (3) had no
After that
and (4) were not markedly retarded or actively psychotic.
day of each
selection, a staff social worker went to the clinic on the
her inexpectant motherl.s next appointment and interviewed her about
terest in joining the study.

She was told about the project, the ser-

would impose on the family.

,Vices it would provide and the obligations it
The services they would reThe fathers, when available, were included.
ceive were described in concrete terms. They were represented as not
be asked to give throughfree but in return for the time parents would
children
out the study by talking regularly with staff and by bringing
daycare if they needed
for physical and developmental examinations, for
Each prospect
it, or for shorter periods similar to nursery school.
simple and direct terms the
was also given a brochure that stated in
and the obligations
purpose of the study, the services it would provide
The final agreement to join the project was, in effect,
of membership.
the project staff.
a verbal contract between the parents and
Twenty-five women were interviewed and twenty agreed to participate.

One child was stillborn, and one child is excluded from this

biological handicap
report though he remained in the project because .a
months of age; the third family was
was discovered when he was several
The fact that
study.
the only one that dropped out after joining the

-9_
reasons) was due, at
only one family dropped out later(for unknown
from first
least in part to the care taken in the admission process
was on thecontact to final agreement. The emphasis in this process
their child develop, on the
importance of the parents' role in helping
obligations and roles of parent
mutuality of the work to be done--the
be provided.
and staff--as well as'on the services to

Families were selected for the project in three waves between the
most favorable for
fall of 1968 and the spring of 1970 because this was
when the children were 30
the staffing pattern. Data collection ended
children from the seventeen families
months old. Of eighteen research
mothers and
twelve were black, two were white, two were mixed--white
Rican, the last family being the one
black fathers--and two were Puerto
There were eleven boys and
to have a second child during the study.
In
intact and nine one-parent families.
seven girls. There were eight
child's. father or another man was a more
seven of the latter either the
At the point of admission six
or less regular member of the household.
and two were on public
of the eight intact families were self-supporting
one was entirely self-supporting,
welfare. Of the single young woman
and seven by public welfare. Their age
one was supported by her parents
18, two were 19, two were 20, seven were 21
range was 18-24: three were
completed high school;
and three were 24 years old. Eleven of them had
family continued during the period besix had not. Contacts with each
greatly from birth on.
fore the infant was born and increased

Project Setting
old residence called
The project wao located in a remodeled
The
Children's House in one of L. :e inner city slums of New Haven.
1 square mile into which
residential section of this area is less than-

10-

-10The district
according to the 1970 census 21,628 persons were crowded.
suffered from the blight common to such areas all over urban America:

overcrowding in deteriorated buildings, lack of playspace, littered

the
streets and a generally dreary appearance. It is an old part of
settlement of migrants from
city and is the main place in New Haven for
Southern United States and Puerto Rico.

There are large numbers of

Italians, Blacks, Puerto Ricans and Irish; a smaller number of

individual exAnglo-Saxons, Germans, Jews, and Slays. While there are
of low
ceptions, most residents of this area suffer the consequences
educational level, low incomes and underemployment.

Staff and Services

clinicians in social
The project staff was composed of experienced
psychoanalysis augmented by
work, psychology, nursing, pediatrics and
For each family
early childhood educators and research psychologists.
services and to
there was a "family team" whose function was to provide
The family
'record their work with and-observations of the families.
team members were a home visitor, pediatrician and developmental exhis
aminer, and as soon as the child was in daycare or toddler school,
staff members who observed the reguteacher. There were other research
and the daycare and toddler
lar pediatric and developmental examinations
Data colschool sessions but had no direct contact with the families.
observers were pooled
lected by both pirticipant and non-participant
components of
for the analyses reported here. There were four major
collected: the (1) home
the service around which research data were
and
visitor program (2) pediatric care (3) developmental evaluation
components has been fully
(4) daycare and toddler school. Each of these
described in The Challenge of Daycare and will be briefly summarized here.

11

Home Visitor Program:

The primary goal of the study, to promote

children could not be achieved,

the development of disadvantaged young
partnership with parents that
it was believed, without entering into a
in their child's development, but
recognized the parents' crucial role
prepared to help. Thus, each family
a role with which the staff was
than any other staff member, was
was assigned a home'visitor who, more
with parental needs.
"the parents' person," identified more clearly
the means of getting
Frequent regularly scheduled interviews* were
relationship in order to try to be of
to know parents, of developing a
In order to do so home
faced.
help to them with whatever problems each
problems due to poverty and
visitors needed to assess which were reality
which were the
consequent lack of the resources available to others,
self-confidence, feelconsequence of second-class citizenship--lack of
related to more deep-seated personings of hopelessness--and which were
be accessible to psychological inality factors that might or might not
in whatever areas of need parents
tervention. Interviews were focussed
given with a
In some situations much tangible help was
identified.
affected parenting. Building
view to reducing stress that adversely
aspirations for a better life were among
of self-esteem and supporting
rarely given gratuitously
the goals of the work. Child care advice was
solve a problem, but some parents sought
unless there was no other way to
help to deal with psychological probsuch help, and a few also sought
affected child care.
lems whether or not they directedly
the first year;.thereafter the schedule
*At least every other week during
contacts were in most cases
called for monthly contacts. In fact the
to feel free to call whenever they
more frequent since parents came
needed or simply wished to do so.

12

-12Finally, through knowledge of the parents and the day-to-day life

of the family, the home visitor provided liaison between the home and
the center.

One important aspect of such liaison was that when a family

visitor could often

team member became concerned about a child, the home

shed light on probable causes as a result of knowing the parents and

what was currently happeRing in the child's family life.

Pediatric Care:

The pediatricians were responsible for periodic

well child examinations and the care of the children when they were sick,
from birth on.

Routine visits were scheduled monthly for the first 12

months and thereafter at age 15, 18, 21, 24, 27 and 30 months.

There

pediatrician who also

was one main pediatrician for each and a backup
became well-known to the family. Protection of the health of the child
in a comprehensive sense was as much a part of the pediatrician's role
as was the treatment of illness.

Because the project pediatricians

records were rich in dewere also child development specialists their

velopmental as well as health data. In their relationship with parents
their
there was emphasis from the beginning on helping parents bring

questions and observations about their children and increasingly to

feel confident of their own ability to decide when they needed to seek
or talk with the doctor.

Respectful listening, helping to clarify ques-

of the ways
tions or concerns and inviting parents' opinions were some
in the health
in which parents were encouraged to participate actively
care of their children.

The time allotted for the periodic examinations

(up to an hour) facilitated this process.

loth the information provided

important.
and the psychological support were seen as

Pediatricians

-13well as about illness, seeking to

gave advice about child rearing as
relationship to inprovide this not in recipe form but in specific

dividual children and parents.

Developmental Examination:

Developmental examinations, using

when the children

the Yale Developmental Schedules were administered
months of age by examiners (pediawere 4, 3, 6, 9, 12, 18, 24 and 30
providing other services for
trician and psychologist) who were not
examinations took place
the children they tested. .The developmental
regular pediatric evaluations in the
in conjunction with but preceding
The results of the test were passed
presence of one or both parents.
visitor, pediatrician or the tester.
on to parents either by the home
parents became interested in
One result of these sessions was that
their way of solving
specific characteristics of their children such as
It provided
their sensitivities.
problems, their interests and often
child and the fact that the examiner
another way for them to know their
observations and inwas a familiar person facilitated the sharing of
by
Pediatric and developmental examinations were observed
formation.
mirror. All of this innon-participant observers through the one-way
each family and contributed
formation was funneled into the record on
to understanding.

Daycare and Toddler School:

The program of child care and educa-

children develop and

tion for study children was based on a view of how
with developmental needs- learn and a commitnent to plan in accordance
Developmental proposiattitudes that we, of course, share with others.
to the relevance of
tions translated into practice included attention
specific tasks, competencies,
(1) the phase concept with its emphasis on

i 4

-14-

styles of interaction, needs and vulnerabilities

(2) the central role

better or
of human relationships. and how these influence learning for

worse

(3) the interdependence of cognitive, emotional and social de-

velopment.

Derivatives and corollaries of these constructs included

elements at
the rationale behind the introduction of certain program
particular times, the handling of the separation experience for parents

and child, the staffing pattern planned in relation to the child's need
for continuity of care from a principal staff member and the importance
of recognizing the individuality of the child.

In full daycare, more

intimately conthan in contacts of shorter duration, teaching must be

knowing the child's
cerned with the child's bodily needs in addition to
nurturing and teachtempo, feelings, style of learning--thus combining
ing roles.

if much
In summary the educational approach was--to repeat an apt
strong efforts were made
used phrase--addressed to the whole child, and
intellectual and
to arrange experiences that would enhance physical,
emotional development.
Twelve study children. were in the daycare program for varying

lengths of time over the 2i years of the study.

Five had 20 or more

two spent 5 months or

months in daycare; five had from 10-19 months and
Some began as
One was in neither daycare nor toddler school.
less.

early as 6 weeks of age, others not until well into their second year.
daycare, toddler school, the equivaFor five children who did not need
characteristics, was
lent of nursery school adapted to their age and
-held twice a week for an hour and a half with their mothers always
present.

Children oegan toddler school between 15 and 18 months of age

and continued until they were 30 months of age.

-15-

partnership with parents in

In all of the services, developing a
them rather than competitive
behalf of the child, being supportive of
staff could work harmoniously with
and creating an atmosphere in which
parents were seen as vitally important.
each other and with children and
program were goals not always
Flexibility and adaptability of staff and
placed upon ourselves. This
achieved but always a part of the demand we
individual characteristics must be
meant first of all that each child's
developed must be attuned to his/her
known and respected and that the program
changed. It also meant that the
developmental needs, changing as the child
unexpected events in the lives of
staff had to be ready to respond to

families, especially crisis situations.

components of the service in
Having described separately the four
data were collected, it_i necessary
the course of which the major research
The unique ck..tribuand unity.
to again emphasize their interdependence
less value to the families if
tions of each component would have been of
integrated parts of a whole. The
they had not been brought together as
continuous contact with one another,
members of each family team were in
Of immediate value was the ongoing
sharing observations and information.
in working out ways to help
synthesis of data from all sources as an aid
This integrative
children and parents about whom concern developed.
formal and informal discussions, had a
process, carried out in frequent
the goals of the project and the
clarifying function with respect to
parents.
philosophy of how to work effectively with both children and

formulations,
The accumulation of information as well as the
it was the core of the research
speculations and actions in respect to
is on the importance of the
effort, of course, but here the emphasis

co-ordination and synthesis of data, and of a shared philosophy as

they influenced services.

Whatever good came out of the work with

children and parents was due, it is believed, not only to having a

to the cocompetent and committed interdisciplinary group but also
ordination of effort.

Comparison Group Study

recognized that the
After the intervention project ended, it was
project data would be more meaningful if some data on a comparable
group of children not involved in the project could be obtained.
Accordingly, one year after the project ended, a Comparison sample of
18 children 30 'months of age was selected.

They were chosen from records

Research mothers. Famiin the same hospital clinic used to select the
mother and race of parlies were matched on income, marital status of
All Comparison
ents; children were matched on sex and ordinal position.
children were full term and free from congenital defect or illness.at

birthpas were the Research children.

Each Comparison mother was seen for a single in-depth interview
project staff. The interby a psychiatric nurse who was part of the
characteristics of the family,
view covered family history, demographic
The
child's birth and health history, and child's general development.

interview was summarized in a narrative report.

Each child in the Com-

by a psycholparison group was seen once for developmental evaluation

The Yale
ogist who had not been involved in the intervention project.
report of the
Developmental Schedules were administered and a narrative
the child was written.
test findings and clinical impressions of

J.

In summary, the design employed was a static-group comparison

basis for
(Campbell and Stanely,, 1965). While the design provides some
evaluating the impact of the intervention program, it is- -acknowledged
in an exto have methodological weaknesses which would not be present
perimental paradigm.

Follow-Up Study
In order to assess the long-term effects of the Yale Child Welfare
after the
Research Program, a follow-up study was carried out five years
in collaboration with
program terminated. This follow-up was conducted
colleagues in the
Edward Zigler, Ph.D. and Penelope Trickett, Ph.D.,
daycare and its
Yale Psychology Department with an interest in early
described here, but is reeffects. The Trickett follow-up will not be
ported. elsewhere (Trickett, 1979).

In the Child Development Unit follow-up, the 16 Research families

still resident in New England were contacted, and all but one family
Each Research mother was seen for an interview
agreed to participate.
from the project or by
of about one hour's duration by her home visitor
but
another familiar project staff member. The interview was flexible
specified topics: changes in family unit,
was organized to cover a set of
the project's close;
residence, education, occupation, and health since
to the project;
the child's daycare and school experience subsequent
retrospective evaluathe child's general develcpment; and the mother's
recorded.in
tion of the intervention program. The interview data were
testing session by a
a narrative report. Each child was seen for one
Staff who had not been
child psychologist on the Child Development Unit
Intelligence Scale
part of the original project. The Revised. Wechsler

18

-18-

and the Drawfor Children, the Beery Test of Visual-Motor Integration,

each child was written.
a-person were administered and a test report for

RESULTS
The data analysis for the Yale Child Welfare Research Program consisted of three major components.

The first component was the analysis

18 children and the

of differences found between the Research group of
matched Comparison group at 30 months of age. The second group of
charanalyses concerned the interrelationships found between family
acteristics and aspects of the children's development, as well as util-

ization of the intervention project by the Research families.

Rather

these anthan being concerned with the overall impact of the project,
and
alyses probe the interdependency of family and child characteristics

relate these factors to project utilization.

Finally, the third cluster

families and
of analyses related to the characteristics of the Research
children at the five-year follow-up.

While these data were relevant to

central to exthe intervention program's lasting impact, they are also

characteristics
amining the relationships between early Family and child
and later outcomes.

select which data

The preliminary problem in data analysis was to
analyses contrasting the Research
should be analyzed. With regard to the

in volume and character

ynd Comparison groups, there was a great disparity
subjects consisted of
of data for the two groups. Data for the Research
approximately five volumes of narrative material for each family containreports, 7-8 developing monthly home visitor reports, 15-20 pediatric
visitor summery of 25-50
mental assessment reports, and a concluding home
consisted of a mother
In contrast, data for the Comparison group
pages.

1.9

-19--

test reinterview summary of about 5 pages and a 30 month developmental

for the two groups, a
To reduce the discrepancy in data sources
port.
analysis to the 30 month test
decision was made to restrict the two-group
mother interviews, and the
report for each child, the Comparison group
concluding home visitor summaries for the Research mothers.

how to convert the

The second major problem in data analysis was
To accomplish this
wealth of narrative material into analyzable form.
The Child Assessment
aim, four major rating instruments were developed.
Scale (MIS) were used for both
Scale (CAS) and the Mother Interview
developmental test report and mother
groups and were ratings based on the
Scale (pus) was
interview summary for each child. The Project Utilization
and was a distillation from the comused for the Research subjects only
Interview Scale (PIS) was
plete data file for each family. The Follow-up
narrative summaries of the follow-up
used to distill information from the
will be
interviews held with 15 Research mothers. These four instruments
described briefly below.

rating scale, with

The Child Assessment Scale (CAS) was a 17-item
qualitative characteristics of the child's
4 levels per item, covering
and
behavior such as attention span, anxiety, zest, coping skills,
presence of emotional problems.

Ratings were done independently by

developmental assessment after a

Rescorla and another rater skilled in
settled
period of training and pilot ratings. All disagreements were
from 55% to 82 %; reliability was
by consensus. Range of agreement was
significance according
better than chance agreement at a high level of
to the weighted Kappa (Cicchetti, 1975) statistic.

20

-20(MIS) was a 38 item rating inventory

The Mother Interview Scale
parental history, assesscovering demographic information, aspects of
child's health history, and child
ments of current parental functioning,
and a
Ratings were done independently by Rcscorla
rearing practices.
settled by consensus.
clinical social worker, with all disagreements
and 60-75% on qualitative judgeAgreement was 90-100% on factual items
by weighted Kappa.
ments, with good reliability as assessed
(PUS) was a 12 item
The Project Utilization Scale
four project
marizing each Research family's use of the
home visitor,
pediatric care, developmental assessment,
collaboration with the
Ratings were done by Rescorla in

inventory sumcomponents:

and daycare.
staff members

who best knew each of the families.

(FIS) was a 21 item inventory sumThe Follow-up Interview Scale
at follow-up. Topics included
marizing family and child characteristics
mother's assessment of her child's
family history since follow-up and the
by kescorla in collaboration with the
Ratings
were
done
development.

author of the interview summary.

Two-Group Comparison at 30 Months

Developmental test scores:

The Research and Comparison groups were

Yale Revised Developmental Schedules

compared on their performance on the
items from
This is a composite test scale including
at 30 months of age.
Binet tests organized into a
the Gesell, Viennese, Merrill-Palmer and
motor, fine-motor, adaptive, language
protocol of five categories: gross
Global developmental age and developand personal-social developmental.
well as DA and DQ in each of the five
mental quotient can be computed as

21

-21and
of variance was used to examine group
Two-way
analysis
categories.
developmental quotient (TDQ), adaptive developsex differences on total
developmental quotient (LDS), each
(ADQ),
and
language
mental quotient
were matched, Group was treated
with a norm of 100. Because the groups
repeated measures analysis of variance.
subjects
factor
using
a
as a within

there was a small but non-significant

As can be seen from Table 1,
the Comparison group and of girls
superiority of the Research group over
However, there was a highly significant
over boys for both TDQ and ADQ.
the Research group, (F=14.05, df=1,
difference on language DQ favoring
ADQ (99 vs
group LDQ was almost as high as
For
the
Research
P e...001).
were already markedly delayed in
106), whereas Comparison group children
adaptive performance (85 vs 101).
relative
to
their
language development
the two groups of children were comparable
would
suggest
that
This finding
capacities, but that the interin terms of basic congitive-perceptual
offset the detrimental effect socio-cultural
vention program had served to
in
language function which was evident
has
on
emerging
deprivation often
the Comparison group.

developmental test performance

An item analysis of the children's
superiority in language function was
revealed that the Research group
syntactic development: identification
manifest in both vocabulary and
of plurals in speech, labelling of
and labelling of pictures, presence
in speech,
the absence of "baby talk" jargon
an action in a book, and
Fisher's exact test (p 4.05).
all significant differences by
there were
Using a Group X Sex analysis of variance,
total Child Assessment Scale (CAS) score
no significant differences for
given the
item scores. In other words,
or for any of the component
CAS scores.

Gi

-22-

indistinguishable
procedures used to compare the two groups, they were
drive for mastery,
in terms of such characteristics as attention span,
relatedness to examiner, and presence of emotional problems.

MIS data.

scale items
Group differences on the 38 mother interview

were explored usingW analyses.

There were only 4 items which signifi-

related to demographic factors

cantly differentiated the two groups--one
in character.
and three, more experiential or psychological
Research father was
In the area of demographic factors, only one
to five Comparison fathers
the sole supporter of his family, co.,Ipared
This finding appears to reflect more inter( t2=5.89, df=2, p

4.052).

in the Research
mittent unemployment or a higher rate of underemployment
fathers made some economic contribution
group. However, more Research
It is worth noting
fathers (10 vs 8).
to the family than did Comparison
daycare in the intervention project,
that despite the availability of free
mothers worked (8 mothers).
the same number of Research and Comparison
as opposed to 8 Research
13 Comparison families were self-suporting,
supplementary benefits, although
-families supported without welfar,Ts. or
Finally, more Research children
this difference was not significant.
children (9 vs 5))
lived alone with their mothers than did Comparison
significant male figure (not the
and more Comparison children had a
(8 vs 4).
father) in contact with the family than did Research children

While most of these demographic findings were not statistically

tht the Comparison group
significant, the pattern of findings suggests
functioning group of families at
may have been a more intact and better
While this outcome reinforces
the outset than the Research families.
comparison group
Campbell's and Stanley's (1963) admonition against

2P

language superiority
matching as a procedure, it makes the finding of
be.
in the Research group more compelling than it might otherwise

which signiOne of the three psychological-experiential variables

points to some superiority
ficantly differentiated the two groups also
favoring the Comparison group. Eight Comparison group mothers were

and coping, as opposed to

rated as having good psychological adjustment
2
Corollary nononly three Research mothers 0(y =4.33, df=l, p 4.05).
significant trends were that Comparison homes were less frequently rated
either overstimulating or underas hostile or depressed in tone and
stimulating in quality than Research group homes. While these three
findings may reflect genuine group differences, it is also possible,

superficial acquaintance the

even probable, that they reflect the
It would not be suprising
project staff had with the Comparison group.
pathology were more evident in
if problems in functioning or areas of
much more inResearch mothers, because staff members knew them so
timately and had so much more information about them.
data analysis used
It is important to note that by the method of
found between the
for these analyses, no significant differences were
amount of cognitive
groups on important child-rearing variables such as
discipline.
stimulation, provision of play materials, nurturance, or
However, there was a highly significant difference between the two
groups of mothers on their expectations for their children.

Nine Re-

search mothers were rated as expecting their child to be more mature

No Comparison
than his age in beaavior, habit training, Jr development.
high standards ('W2 =12.86,
mothers were seen as setting this type of overly
Furthermore, ten Research mothers felt that their child
df=3, P 4-.005).

-24-

mothers
had some or many problems in development, but only two Comparison
2

expressed such concerns (Ny =6.12, df=1,

higher standThese two findings suggest that Research mothers set

and/or observant about
ards for their children and were more critical
What is not clear is
their child's problems than Comparison mothers.
whether the Research mothers were predisposed to having high expectations for their children and thus attracted to participation in the
in some
project, or alternatively,_ that participation in the project
way fostered these attitudes.

It is important to stress here that while

aspirations for school achievea main project goal was to foster parental
help parment and social advancement, an equally important goal was to
ents be realistic with regard to habit training, behavior, and development
accomplishments and to not demand overly mature behavior in these areas.

Interrelationshi s between fatall

characteristics

child variables and

project utilization:

Correlational analysis was used to explore relationships between

project utilizachild DQ, child characteristics, family variables, and
Factor analyses were done on the CAS and MIS, to look at the comtion.

position of the two inventories.

For the CAS, three major factors were derived from a principal com-

ponents analysis, accounting for the 73% of the variance.

Using a vari-

max rotation, the major factor (45% of variance) reflected compliant,

co-operative behavior during the testing session.

The second factor

objects
(16% of variance) reflected zest and animation in response to
and people.

The third factor (12% of variance) appeared to reflect

physical and cognitiv.- development.

4 0

-25-

components analysis'
Six major factors were derived from a principal
accounting for 68% of the total variand varimax rotation Of the MIS,
variance) related to intactness of family group;
ance: Factor 1 (18% of
mother's level of attrainment in edFactor 2 (14% of variance) involVed
(13% of variance) pertained to physical
ucation and employment; Factor 3
environment; Factor 4 (9% of variance)
and emotional quality of the home
of the mother and mother's exloaded on both psychological adjustment
5 (8% of variance) perpectations and handling of her child; Factor
lastly, Factor 6 (6% of variance)
tained to neonatal and child health;
such as cognitive stimulation,
related to child-rearing characteristics
appeared to valiplay materials, and nurturance. This factor solution
of the scale, in that the major
date the rationale used in construction
the topic headings of the
factors described corresponded closely to
scale items.

variables was a main focus

The relation between child DQ and other
When the data for the combined Research
of the correlational analyses.
variables were significantly
and Comparison groups were examined, only two
(r=66,
related to TDQ at 30 months: amount of cognitive stimulation
.01).
(r=.58, df=34, p
df=34, p Z.01) and provision of play materials
correlated with each other
These two variables were also significantly
correlations were found be(r=63, df=34, p.0 e01). Similar signifiCant
and both ADQ and LDQ, with cognitive
tween these childrearing variables
play materials more
stimulation more highly correlated with LDQ and

highly related to /JDQ.

also obtained when the ReThe findings on DQ described above were

Additionally, several intersearch group data were analyzed separately.
,

-26esting relationships

emerged between child DQ and family project utiliza-

by LDQ was posiIt appeared that language development as measured

months spent in the daycare
tively related in some measure to number of
(The correlation was only significant when one
program of the project.
in another daycare center when the
child was credited for months spent
to infamily moved out of town temporarily). Thus, it seems plausible
in language development may have
fer that the Research group superiority
enrichment provided by the daycare
been partially attributable to the
tion.

program.

indices of project
Child TDQ was significantly related to three
developmental
utilization: parental invo lvement and interest in child
df=16, p 4.01)1 parental involvement and inexam performance (r=.62,
dfg=10, p ef_.01), and parental
terest in the daycare program (r..7O,
Thus,
.Z...05).
positive relation to the daycare staff (r=.67, df=10, p
child was closely linked with pargeneral cognitive competence in the
effective utilization of aspects of the inental involvement in and

tervention program.

the two variables perBecause of their high correlation with TDQ,

examined in some
taining to parental relation to the daycare program were
correlated with
While the two variables were not significantly
detail.
to be some pattern evident in their relation
one another, there appeared
families who were most involved and interested
to other var Ibles. The
tended to be those in which fathers made an econin the daycare program
mothers.nad been married, and
omic contribution, mothers were employed,
ting (correlations of .79, .71, .77, .62,
the family was self-suppor
with a strong positive relation to
df=10, p eL,05). similarly, families

:e7

-22-

indistinguishable
procedures used to compare the two groups, they were
drive for mastery,
in terms of such characteristics as attention span,
relatedness to examiner, and presence of emotional problems.

MIS data.

scale items
Group differences on the 38 mother interview

were explored using'

analyses.

There were only 4 items which signifi-

related to demographic factors

cantly differentiated the two groups--one
in character.
and three, more experiential or psychological
Research father was
In the area of demographic factors, only one
to five Comparison fathers
the sole supporter of his family, co,ipared
This finding appears to reflect more inter( X2=5.89, df=2, p

G.052).

underemployment in the Research

mittent unemployment or a higher rate of
contribution
However, more Research fathers made some economic
group.
fathers (10 vs 8). It is worth noting
to the family than did Comparison
daycare in the intervention project,
that despite the availability of free
mothers worked (8 mothers).
the same number of Research and Comparison
as opposed to 8 Research
13 Comparison families were self-suporting,
supplementary benefits, although
-families supported without welfarTs. or
Finally, more Research children
this difference was not significant.
children (9 vs 5))
lived alone with their mothers than did Comparison
significant male figure (not the
and more Comparison children had a
children (8 vs 4).
father) in contact with the family than did Research
not statistically
While most of these demographic findings were
tht the Comparison group
significant, the pattern of findings suggests
functioning group of families at
may have been a more intact and better
While this outcome reinforces
the outset than the Research families.
comparison group
Campbell's and Stanley's (1963) admonition against

2P

seeking advice from the home visitor were all significantly inter.05).
correlated with each other (correlations of .57 to .79, df=10, pGof home visitors recommendaIt is interesting to note that mother's use
visitor variables.
tions was not closely related to these other home
significantly corHowever, use of home visitor recommendations was
psychological adjustrelated with rater's assessments of the mother's
The findings supports a widely
ment and coping (r=.72, df=lo, p G .01).
poorly adjusted clients have
held clinical impression that disturbed or
Alternatively, the finding might be interdifficulty utilizing advice.
professionals tend to regard clients
preted as showing that mental health
adjusted and present them as
who have difficulty using advice as poorly

such in their records.

Follow-Up Findings

follow-up study fall into two

The findings from the five year
Firstly, there are data on the children and families
broad classes.
to the question of the longat time of follow-up which are relevant
Secondly, there are the interm impact of the intervention project.
characteristics at 30
terrelationships found between child and family
months and follow-up outcome.
Long term impact of the intervention.

Data documenting some .tus-

obtained in
tained, long-term impact of the intervention program were
from the Trickett followboth the follow-up studies conducted. Findings
appear elsewhere in a separate
up will not be de.s.:ribed here, as they
however, that they show a
report (Trickett, 1979). It can be said,
continued effect of the intervention program on both child Peabody

achievement as measured by the

Picture Vocabulary Test IQ and school

-29-

Peabody Individual Achievement Test.

Intelligence test data from the

children
Child Development Unit follow-up suggested that the Research
continued to function somewhat above the norm for inner-city, disad-

vantaged children. The average score for the 15 Research children tested
on the WISC-R was 91.8, with only a negligible 5 point superiority of

Performance IQ over Verbal IQ.

This VIQ-PIQ pattern suggests that some

strength in language function continued to characterize the Research

children, relative to what is often found in disadvantaged samples.
The strongest findings obtained from the follow-up pertain to
general upward mobility of the Research families.

As can be seen in

indices of upTable 2, the families showed striking gains by several

beginning of the project.
ward mobility, relative to their position at the
sample, it cannot
As there are no comparable data available for a control
for these gains
be proven that the intervention project was responsible
rather than some other factor such as self-selection of the sample.
However, the data to be outlined below certainly are not typical of

the cycle of poverty associated with disadvantaged families.

In the area of educational advancement, 10 out of 17 research
mothers obtained some further education during tile project: 2 graduating
training course, and 2 taking
from high school, ti taking some form of
At the time of follow-up, eight mothers had continued
college courses.
training courses and four
to advance educationally, with four taking
working toward BA degrees.

toward econA parallel pattern was evident in terms of .progress

omic self-sufficiency.

At the end of the project, the number of families

with eight entirely selfon welfare had declined from nine to six,
At the time
supporting and three partially self-suppor.z.ing families.

-30-

still on welfare, eleven

of follow-up, there were only three families
self-supporting, and two
families were self-supporting, one was partly
grants.
were supported by government college education

in life
A qualitative assessment was made of general improvement

style and quality of life for the Research families.

The criteria for

improvement were one or more of the following: improvement in housing,

social
medical care, socioeconomic statusseducational or training status,
life, and engagement in community life.

By the end of the project,

using these
twelve families had improved moderately in quality of life
eleven families showed clear evicriteria. At the time of follow-up,
life and expressed attidence of tangible improvements in quality of
Three others did not
tudes of a belief in a progressively better life.
happier in their personal life and
seem materially better off but seemed
have deteriorated in
more positive in outlook. Only one mother seemed to

quality of life and general functioning.

birth
Finally, the project appears to have had a striking effect on

rate in the Research families.

At the end of the project 14 families

children and one

still had only one child, with two families having two
At follow-up, there were ten families still with
family having three.
three families had
only one child, four families had two children, and
twins).
three or more children (which in one family was a set of

Interrelationships between 30 month

and follow-up variables.

The

relationships between follow-up intelligen'e test scores (PPVT IQ, 441SC-R

PIQ, WISC-R VIQ) and 30 month

variables were examined by correlational

analysis and stepwise multiple regression.

Predictor variables used in

30 months, and the

the multiple regression analyses were ADQ and LDQ at

-31-

factor scores for each child on the three CAS factors and six MIS
factors described above.

The best predictor of PPVT IQ was the MIS factor related to cog72% of the varnitive stimulation and play materials, accounting for
iance with a highly significant F ratio (F=39.511 df=1,

15, pL_.001).

LDQ at 30 months was the only other predictor variable which met the

G .01)
_criterion for entering the regression equation (F=7.081 df=2, 14, p
raising the PPVT variance accounted for to 82%.

Correlational analysis

supplemented the multiple regression findings and demonstrated several

other significant relationships.

PPVT IQ was significantly correlated

(correlation of
with 30 month total DQ, adaptive DQ, and language DQ
PPVT IQ at follow-up was also positively
.76, .75, .77, df-15, p 4:.01).
behavior at the 30
correlated with mature, cooperative, and persistent

month testing session, as measured by CAS total score (r=.61m, df=15,

p

.01).

There were also significant correlations between PPVT IQ and

some 30 month demographic and project utilization variables: children

had higher PPVT IQs in families where mothers had been married, when
of the
fathers had made some economic contribution, and.when utilization

pediatric care and the developmental assessment components of the project

had been high.

These findings replicate some of the 30 month. findings,

notably that higher IQ's were found in children from homes with more
cognitive and play enrichment.

Relationships between the follow-up WISC-R and the 30 month variables were complex.

The correlation of WISC FIQ and TDQ at 30 months

was .58 (df=13, p 4=.05).

It is interesting that the correlation between

Adaptive DQ at 30 n.onths and Performance IQ at follow-up was .73 (df.13,

30 months and
p. 2:-.01), whereas the correlation between Language DQ at

32

Verbal IQ at follow-up was only .44 (n.s.).

In other words, perform-

quite consistent
ance on perceptual-motor and form perception items was
measured
for these children from 21 to age 8 but'language proficiency as
did not preby the Yale Revised Developmental Schedules at 30 months
dict VIQ performance on the WISC.

There was a marginally significant

correlation of .51 (df=13, p .G.05) between LDQ at 30 months and the

Vocabulary subtest of the WISC-R, which is the Verbal subtest most
similar to the language items tested at 30 months.
There was no clear Pattern of relationship between WISC-R PIQ
variables
and VIQ and demographic, child-rearing, or project utilization
at 30 months.

PIQ and VIQ were each significantly correlated with a few

-scattered 30 month variables but there was no overlap between sets.

stepwise multiple regression confirmed that PIQ was significantly predicted by 30 month ADQ (53% the variance in PIQ), but there were no
significant predictors of VIQ.

However, there was a significant cor-

relation between WISC-R FIQ and CAS total at 30 months (r=.75, df-13,
p .4:.01).

As with the PPVT IQ score, children who were more cooperative

and mature and animated at 30 months scored higher on the WISC-R at

follow-up.

Follow-up WISC FIQ was correlated .74 with the PPVT given in

approximately the same month by a different examiner (df=13, pG .01),

with WISC scores being consistently lower than Peabody Scores.

There

was a very high correlation of .81 (df=13, pG .01) between Verbal IQ

on the WISC and performance on the Peabody Individual Achievement Test
ImplicaCorrelation of PIAT and Performance IQ was minimal.
(PIAT).

tions of these data are that the skills measured by the WISC Verbal
scales are those most central to school achievement in reading, spelling, math, and general information; furthermore, attainment in these
skills was not well predicted by the variables measured at 30 months.

38

-33The last relationship of interest to report is that mother's

correlated with
view of the child's development at follow-up was highly
It appears
her view of the ohild'at 30 months (r=.80, df=13, pd.! .01).
development
the mothers were basing their appraisal of their children's

in large measure on the child's academic and verbal achievement, as

shown by correlations of .60 and

.55 respectively between mother's view

and WISC-R
of development at follow-up and PIAT total achievement score
.05).

VIQ (df=13, p

The mother's perception of her mother's behavior

the child's
in school also seemed to contribute to her overall view of

development.(r=.531 df=13, p

Discussion

The Yale Child Welfare Research Program was designed as a servicestrongly influenced
centered longitudinal study, a choice of design that
its methodology.

It is recognized that each research plan offers access

to a limited number of issues and has advantages and disadvantages.

research, a term
The service-centered investigation is a form of action
that applies to situations in which the investigator not only observes
the process
but acts within the field of observation and is a part of
he studies.

Thus the study did not employ the paradigm of experimental

This
or laboratory research but derived from a clinical orientation.
approach must be taken into account in considering the study's methods

and findings.
study
To amplify, the position taken in this research was that a

clinicians and
in the context of provision of services by experienced
to the goals of
educators would Provide very rich data, highly relevant
the children and parents over a
the project. We chose, also, to study

-34development as continutime span in order to document the children's

adults as
ously as possible, as-well as the development of the young
investigaparents. There were also advantages in the continuity of the
increasingly able to
tors: participating families and family team became
communicate as time passed. The expectation that the contact over time
would result in progressive enrichment of the data was confirmed.

For

to
example, important material about the parents' own childhood came
Informalight especially around the relationship to the growing child.
tion-tended to become more personal and relevant as the contact continued

and trust in the,study staff increased.

It was also apparent that the

population ingrowing f,imiliarity of the staff members with the study

Thus we becreased their sensitivity and made them better observers.
data that reasonlieve that the clinical- developmental approach provided
of human development and discourages simably reflects ';.he complexity

plistic conclusions.

However, the very richness and complexity of the

As must be obvious,
data make its organization and analysis difficult.,
part of the data collected
the data preser.ttd here represent only a small
those data which were amenable to conventional
during the project -- namely

tabulation and statistical analysis.

The other data collected will serve

reference.
as the basis for forthcoming reports with a different frame of

The findings from the Yale Child Welfare Research Program constitute a modest but important addition to the literature documenting
disadvantaged families.
the effectiveness of early interventions for
broad scope, clinicallyThe research described here documents that a
child developoriented intervention program can be effective in fostering

ment for disadvantaged infants and their families.

A finding of particula]

-35-

interest is that without the intervention program's having a focussed

Levenstein (1977) or Schaefer
or structured curriculum such as that of
effect on the chiland Aaronson (1972), there was a hfi.ghly significant
dren's language development at 30 months of age relative to Comparison
subjects.

While the presence of a program effect at the termination of the

evidence for the
project is a gratifying finding, the more impressive
The Research
program's ultimate impact derives from the follow-up data.

children's performance on the WISC-R at 5 year follow-up was somewhat

furthermore, the children's
above the norm for disadvantaged samples;
performance skills, which is
verbal abilities had kept pace with their
inner city samples. Evidence from
a pattern not universally found in
indicated a long term
the Trickett follow-up (Trickett, 1979)
effect on PPVT IQ and PIAT school achievement. Thus, the program impact
after the
in child IQ and achievement did not appear to "fade out"
project ended.

to be the
The long term project impact on family patterns appear

A primary thrust of the promost striking findings of the research.

achieve their aspirations for an
gram was to help the project families
The follow-up data revealed impressive upimproved quality of life.
improvements in residence,
ward mobility of the families in terms of
The .low birth
educational advancement, and economic self-sufficiency.
further suggestion of a
rate in the Research families at follow-up is a

decisions E.ad
change toward more autonomous control of Laportant life
As the ratings on quality
a striving for improved social circumstances.
families appeared to
of life indicated, the majority of the Research
improvements in their lives at the
have made significant and substantial

-36-

time of follow-up.

While these changes cannot .be incontrovertibly

of many of the
attributed to the program, it was certainly the belief
family members, as well as of the project staff, that the program had

played a significant role in these improvements.

Child Welfare
In addition to the specific outcomes of the Yale
project has implications
Research Program decribed in this paper, the
early intervention. The
of a more general nature for the field of
effect was its improject data suggest that the intervention's primary

choices, and patterns of

pact on families--on their aspirations, life
It is certainly the impression of the clinicians who profunctioning.
ingredient of the provided the intervention services that the crucial
belief that both the families and the
gram was a committed, concerned
that the role of the
children were important in their own right, and
in the program realize
service providers was to help each individual
his or her potential as fully as possible.
The Yale Child Welfare Research Program served an educational
to use pediatric
role, by such activities as helping new mothers learns
negotiating for public housing,
care effectively, or aiding families in
appropriate stimulation for infants.
or encouraging mothers to provide
designed to teach parents
However, the program was quite definitely not
the manner of such
how to educate or interact with their children, in
Schaefer's
projects as Levenstein's (1977) toy demonstration model or
(Schaefer and Aaronson, 1972) home tutoring model. The fact that signi-

variety of programs
ficant project effects have been achieved .:.11 such a
crucial mediating
.-ith so many diverse approaches suggests that the
actors of many interventions may be interpersonal and motivational

ti7

-37-

intervention comes to believe that the

ones--that the recipient of the
and considers his or her deservice provider values him as a person
goal worth striving for.
velopment and achievement as an important
Research Program are conThe results of the Yale Child Welfare
that a project must signifisistent with Bronfenbrenner's (1974) view
attitudes in order to have long-standcantly impact family patterns and

The findings from the Trickett (Trickett, 1979)

families continued
follow-up-provide further evidence that the Research
advancement, in such ways as
to actively strive for their childrenls
parochial school or in assuring that
choosing to send their children to
Furthermore, the
attendance.
their children maintained good school
by showing that
project underscores the importance of early; intervention
ing effect.

comparison children in language

a difference between intervention and
development can be detected as young as 30 months of age.
point, it must be
While the preceding discussion has implied this
Program
reiterated that the results of the Yale Child Welfare Research
of outcome variables as
highlight the importance of using a multiplicity
such as advocated in Zigler and
an index of project effectiveness,
child DQ did not show a signifi--Trickett-(1978). For instance, overall
strong effect was obtained
cant program effect at 30 months, whereas a
We would argue that at the age of 30
in the area of language function.
tested reflect in large part
months, the cognitive-perceptual tasks
sensori-motor systems and would
maturation of the cognitive-perceptual and
cultural enrichment and envirGnmental
be less likely to show the impact of
within the range of relatively
stimulation than would language function,
in intervention studies of this
normal ezidwaent and environment sampled

-3843,

type.

Another example of the importance of multiple dependence variables

characteristics such as educational attainment,

are the data on family
mobility, and birth rate, which were among the more striking outupward mobility,
comes of this program.

is one derived from

Among the general implications of the program
the project services varied conthe fact that effective utilization of
As reported, the data suggest that the better-functioning
siderably.
this
families made best use of the services. However, looking beyond
reveals striking
correlation, examination of the individual families
illuminate the issue of why
exceptions to this general pattern which
to utilize the program for helping
some were better able than others
Despite some similarities in members of this disadvantaged
themselves.
them, as in other groups, with
group, there were large variations in
and
respect to general adaptive abilities, personality characteristics,
and developing as parents.
capacity for relating and trusting others
experiences, including the
They also varied widely in their childhood
the quality of their nurturance,
strengths in their families or origin and
as adults; This study sugconditions which influenced their capacities
available by qualified personnel
gests that once good services are made
human complexity, what will be
who respect their clients and understand
what they bring to the situation,
utilized depends upon the participants,
this study for social
and hence what each can use. One implication of
disadvantaged families should
policy is that intervention programs for
options which are responprovide a spectrum of quality services, offering
sive to the needs of individual participants.

39

References
Studies in New Haven,
Headstart graduates in school.
A Report on Lon0.tudinal EvaluaEd.
In Ryan, S.
Connecticut.

Abelson, W.

tions of Preschool Programs, Vol. I.

Washington, D.C.:

U.S.

Government Printing Office, 1974.

Is early intervention effective? In Bronfenbrenner,

A Report. on Longitudinal Evaluations of Preschool Programs,
U.
Vol. II. Washington, D.C.: U.S. Government Printing Office,

Bronfenbrenner, U..

1974.

Evaluating parental effects on child development.

In Shuman, L. Ed. Review of Research in Education (in press).

Clarke-Stewart, A.
1978.

Lasting Effects After Preschool.

Consortium for Longitudinal Studies:
Final Report, HEW Grant, 90C-1311 to the Education Commission
of the States,
Day, M. & Parker, R.
Bacon, 1977.
Gordon, I. & Guinagh,

1978.

The Preschool in Action. Boston:

Allyn and

A Home Learnin Center A roach to Earl

Final Report. Gainesville, Fl.: Institute for
B.

Stimulation:
Development of Human Resources, Dec. 1977.

Training mothers to instruct their infants at

Program on PreIn Karnes, M. Research and Development
home.
Final Re cwt. Washington, D.C.:
school Disadvantaged Children:
U.S. Office of Education, 1969:249-263.

Karnes, M. & Badger, E.

The Early Training Project for Disadvantaged

Klaus, R. & Gray, S.
DevelopA Report After 5 Years. Soc. Res. in Child
Children:

ment Ser. 020, 1968, 33:4.

Research Program.
Lally, J. The Famlly Development
Children's Center, Syracuse Univ., 1971.

Syracuse, N.Y.:

The Family Development Research Program: A

Lally, J. & Honig, A.
Enrichment.
Program for Prenatal, Infant & Early Childhood
Program, Syracuse
Syracuse, N.Y.: Family Development kesearch
Univ., 1977.

40

-400 .

Home Teaching of Mothers and Infants.

Lambie, D., Bond, J. & Weikart, D.
Ypsilanti, Mich.: High/Scope, 1974.
In Day, M. & Parker, R.
The Mother-Child Home Program.
Levenstein, P.
Boston: Allyn & Bacon,
The Preschool in Action (2nd ed.).
Eds.
1977.

The Challenge of Daycare.

Provence, S., Naylor, A. & Patterson, J.
New Haven, Ct.: Yale Univ. Press, 1977.
In
Center.
The Frank Porter Graham Child Development
Robinson, H.
New York: Atherton Press,
Dittman, L. Ed. Early Child Care.

1968:302-312.

A Report on Longitudinal Evaluations of Preschool Programs,

Washington, D.C.: U.S. Government Printing Office, 1974.
Vol. I.

Ryan, S., Ed.

project: ImpleSchaefer, E. & Aaronson, Mi. Infant education research

In
program.
mentation and implications of the home-tutoring
Allyn &
The Preschool in Action. Boston:
Parker, R. Ed.
Bacon, 1972:410-436.

An Independent
The Yale Child Welfare kesearch Program.
Trickett, P.
Paper presented at Society for
Follow-Up Five Years Later.
Francisco,
Research in Child Development Biennial Meeting, San
March, 1979.

evaluations of
Zigler, E. & Trickett, P. I.Q. social competence, and
early childhood intervention programs. Am. Psychologist, Sept.
.

1978, 33:9, 789-798.

41

-414

Footnotes

Gerber, Ph.D. for her service

The authors wish to thank Barbara Wolf
M.D. and Joyce Ruhloff,
as psychological examiner and Martha Leonard,
analysis. The many dedicated people
MSW who served as raters for the data
Child Welfare Research Program
who served as staff members for the Yale
without their enthusiasm and commitare too numerous to list here, but
Finally, we would like
ment the project would not have been possible.
shared their lives with us and
to thank the children and parents who
Project #PR900 of
cooperated in our efforts. This paper is based on
Child Development)
The United States Children's Bureau, Office of
A grant from The Ford
Department of Health, Education and Welfare.
organization and analysis.
Foundation, New York supported the work on data
form at The Society for Research
This report was presented in abbreviated
Francisco, March, 1979. Rein Child Development Biennial Meeting, San
A. Rescorla, Ph.D., Yale
quests for reprints should be sent to Leslie
New Haven, Ct., 06510.
University, Child Study Center, 333 Cedar Street,

-42-

Table

Developmental Quotients at ,10 Months

Language DQ

Ada five

Total DI

Boys

Girls

Research

Compar.

Group

Group

102.0

110.4

93.6

97.8

Research

Compar.

Research

Compar.

Group

Group

Group

Group

Boys

Girls

105.1

100,9

Boys

97.5

79.6

88,6

108.4

Girls

1023

94,7

98.5

* 99.4

85.5

107,8

105.3

98,1

106.2

101.5

*F

11 16 to 14.05, p

--

-43Table II

Project Start

Mothers'

Educational

'

11 high school graduates

6 some high school

Project End

10 with more education:

2 finished high school

6 took training course(s)

Five Year Follow-Up

8 with more education:

4 took training course(s)

4 working toward B.A. deg,

Level

2 took college course(s)

Source of
Support

7 self-supporting
1 supported by parents
9 on welfare

8 self-supporting
4 partially self-supporting
S on welfare

11 self-supporting
1 partially self-supportin,

2 on college grants
3 on welfare

Number of

17 had 1 child

14 had 1 child

2 had 2 children

children
1 had 3 children

10 had 1 child
4 had 2 children

3 had 3 or more children

in family

113