See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/226384465

Psychometric Properties of the Family Support Scale with Head Start Families

Article  in  Journal of Child and Family Studies · March 1998

DOI: 10.1023/A:1022912130180

CITATIONS READS

29 1,519

4 authors, including:

Marc J. Tassé Michael G Aman

The Ohio State University The Ohio State University
108 PUBLICATIONS   2,545 CITATIONS    366 PUBLICATIONS   16,018 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Defining intellectual disability and its diagnostic process View project

National Core Indicators Adult Consumer Survey. Random selection and satisfaction survey and assessment of outcomes of adults receiving DD services in the state of
Ohio. View project

All content following this page was uploaded by Marc J. Tassé on 28 August 2014.

The user has requested enhancement of the downloaded file.

Journal of Child and Family Studies, Vol. 7, No. 1, 1998, pp. 69-77

Psychometric Properties of the Family Support

Scale with Head Start Families

Barbara Hanley, Ph.D.,1,5 Marc J. Tasse", Ph.D.,2 Michael G. Aman,

Ph.D.,3 and Pamela Pace, B.A.4

Social support systems can have either a beneficial or adverse effect when a
family faces stress. Hence, support perceived by family members is an important
construct. We studied the psychometric properties of the Family Support Scale
(FSS) with a sample of 244 low-income families of children in a Head Start
program, and we conducted an exploratory factor analysis with the scale. We
obtained a 5-factor solution for the FSS when assessing the family support of
low-income families. The internal consistency and test-retest reliability were
moderately high for the 5 proposed subscales. The factor structure obtained
may prove more valid for Head Start families than structures from previous
studies, although there is still a need for further psychometric study of the FSS.
KEY WORDS: social support systems; Family Support Scale; Head Start program; family
support; low-income families.

A family's social support network, whether formal or informal, has

been associated with a family's physical and emotional health maintenance.
1
Director of Social Work, University Affiliated Cincinnati Center for Developmental
Disorders, and Field Service Assistant Professor of Pediatrics, College of Medicine and School
of Social Work, University of Cincinnati, Cincinnati, OH.
2
Assistant Professor, Department of Psychology, Universite du Quebec a Montreal, Montreal,
QC, Canada.
3
Professor, Department of Psychology and Psychiatry, The Ohio State University Nisonger
Center UAP, Columbus, OH.
4
Program Manager, University/Head Start Partnerships, The Ohio State University Nisonger
Center UAP, Columbus, OH.
5
Correspondence should be directed to Barbara Hanley, UACCDD, University of Cincinnati,
3333 Burnet Avenue, Cincinnati, OH 45229-3039. Electronic mail may be sent to
[email protected].

69
1062-1024/9&!D300-0069$15.00/0 » 1998 Human Sciences Press, Inc.
70 Hanley, Tasse, Aman, and Pace

Cohen and Willis (1985) reviewed the literature and concluded that social
support systems can work in either beneficial or adverse ways when a family
faces stress. However, the type of effect the social supports have is not so
important as the type of system present. Chronic stressors families may live
with include poverty, single parenthood, being a minority and/or female,
inadequate standards of living, and role overload. Studies that have focused
on these types of chronic stress in families have also reported positive ef-
fects of a good social support network in respect to successful family
functioning.
The Head Start research database from 1980 to 1991 emphasized
the following: (a) the physical, emotional, and social development of
Head Start children (Dixon, 1980; Dubrow & Garbadino, 1989; Jensen,
1983; Keltner, 1990; Quay, 1989; Stone et al., 1982); (b) minority Head
Start families and upward mobility (Oyemade, 1985; Washington, 1985;
Zigler, 1985); and (c) Head Start parent participation and resources
(Oyemade, 1985; Payne, 1989; Sharni, 1980; Slaughter, Lindsey, Naka-
gawa, & Kuehne, 1989). Two studies emphasized family functioning,
strengths, and a family systems approach (Light & Martin, 1986; Rues-
chenberg & Buriel, 1989). In the study of Native American families, the
results indicated that Native American families have some family
strengths that far exceed those of families in the general American popu-
lation (Light & Martin, 1986). A study of Mexican-American families us-
ing a family systems perspective concluded that "acculturation is not an
all-or-none phenomenon and that adjustment to United States society
can take place with the basic integrity of the family remaining intact"
(Rueschenberg & Buriel, 1989, p. 242).
One aspect of successful family functioning is related to the social
support system of the adults in the family (Lendblad-Goldberg & Dukes,
1985; Gladow & Ray, 1986; D'Ecole, 1988; Stevens, 1988). Hence, it is of
utmost importance to assess family supports accurately to better
understand how the family functions, especially in low-income and/or
Head Start families.
We addressed the use of the Family Support Scale (FSS: Dunst,
Jenkins, & Trivette, 1984, 1988) with low income families of children in
Head Start programs. We present statistical analyses from data on the FSS
obtained from low-income Head Start families and provide a factor struc-
ture that appears to reflect the construct of family supports for these
families. We also examined the internal consistency and test-retest reliabil-
ity of the FSS's items when redistributed according to our proposed factor
structure.
Psychometric Properties 71

Table 1. Participant Characteristics

Raters % Religion %
Mother 73.6 Baptist 46.6
Father 18.0 Other Protestant 14.4
Step-parent 2.4 Catholic 9.3
Grandparent 3.0 Muslim/Islam 1.9
Other 3.0 None 14.9
Other 12.9

Marital Status %
Single 45.6 Ethnicity %
Married 32.8 Asian 4.2
Separated/Divorced 20.7 Black 66.0
Widowed 1.0 White 26.8
Other 3.0
Income per year %
$0-5,999 46.4
$6,000-11,999 23.4
$12,000 or more 30.2

METHOD

Participants

The participants were recruited from the 15 Head Start Centers of

the Child Development Council (CDC) of Franklin County, Inc. in Colum-
bus, Ohio. Recruitment of parents was done through telephone solicitation,
presentations during parent orientations and meetings, and through the dis-
tribution of flyers. Participant recruitment and data collection were carried
out over three years (1992-1995). Out of a possible 1,392 families with at
least one child enrolled in Head Start in Central Ohio, 244 (18%) partici-
pated in the study. Table 1 presents demographic characteristics of the
participants, family income, sources of family income, marital status, eth-
nicity, parents' education level, and religion. Parents' mean age was 30.4
years (SD = 8.1).

Procedure

Parents participating in this study were asked to complete the FSS at

their CDC Head Start center or at the university center where the study
72 Hanley, Tasse, Aman, and Pace

was administered. In addition to the FSS, participants completed two

demographic questionnaires and three self-report scales. If a participant
needed assistance, a research assistant was available to answer questions
or read the survey forms to the participant.
Participants were also asked to complete a second FSS two weeks later.
This interval was chosen to minimize the possibility of responding from
rote memory, while the times were still close enough that the same family
circumstances should have been in effect.

Instrument

The Family Support Scale is a measure of how helpful different sources

of social support have been to the family rearing a young child who is at
risk and/or presenting developmental delays. The measure consists of 18
items that are rated on a 5-point Likert scale ranging from "Not At All
Helpful" (0) to "Extremely Helpful" (4). Respondents are instructed to cir-
cle the response that best describes how helpful the source was to the
family during the previous 3 to 6 months. If the source of support was not
available, they could circle "Not Available" (NA).
Dunst et al. (1984) reported a six-factor solution derived from a sample
of 139 parents of preschool children with mental retardation, with physical
handicaps, and children considered developmentally at-risk. They labeled
their six factors: informal kinship, social organizations, formal kinship, nu-
clear family, specialized professional services, and general professional
services. Their factor solution accounted for 62% of the total score vari-
ance. Dunst and Trivette (1986) reported deriving a second factor structure
of the FSS. This second factor structure consisted of a five-factor solution
(formal kinship, informal kinship, social groups, professionals, and profes-
sional groups). A third study reporting on the factor solution of the FSS
was published by Taylor, Crowley, and White (1993). They reported on the
psychometric properties of the FSS using an aggregated sample of 990 fami-
lies. Their exploratory factor analysis procedure resulted in a four-factor
solution. They labeled their four social support factors: familial, spousal,
social, and professional.
Reliability and validity measures were originally estimated from the
Dunst et al. (1984) sample of 139 parents. Cronbach's coefficient alpha
was reported to be .77 for the total score, split-half reliability was .75, and
test-retest reliability ranged from .41 to .75 across subscales. The internal
consistency estimated by Taylor et al. (1993) for their four-factor solution
ranged from .35 to .76 across subscales and was .80 for the total score.
Concurrent validity was determined by correlating the FSS with a number
Psychometric Properties 73

of parent and family outcomes and obtained values that ranged from r =
.18 (p < .01) to r = .40 (p < .001) (Dunst et al., 1984).

RESULTS

Factor Analysis

All analyses were done using SPSS, version 4.1, on a mainframe com-
puter. Exploratory factor analysis was done using a principal components
extraction procedure with a varimax rotation. All factors with an eigenvalue
equal to or greater than 1.0 were kept as part of the factor structure. An
item was assigned to a factor if its loading was > .40. This procedure was
identical to that used by Dunst et al, (1984) when they conducted the origi-
nal factor analysis of the FSS. Our factor analysis resulted in a five-factor
solution that accounted for 61% of the variance as follows: Factor 1, 30.6%;
Factor 2, 11.5%; Factor 3, 7.1%; Factor 4, 5.7%; and Factor 5, 5.7%. In
general, there was very little tendency for items to load heavily on more
than one factor. The five factors were given the following labels: (1) com-
munity; (2) spouse and in-laws; (3) friends; (4) specialized/professional; and
(5) own parents and extended family. These factors will henceforth be re-
ferred to as subscales. The factor loadings for the 18 items are presented
in Table 2. The mean factor loading across all 18 items was .57, which is
quite high.

Reliability

Internal Consistency

Cronbach's coefficient alpha was used to estimate the internal consis-

tency for each of the subscales and the total score. The split-half procedure
was also used to estimate the total scale internal consistency. The split-half
method used alternating items and corrected for length using the Spear-
man-Brown formula (Anastasi, 1988). This is consistent with the procedure
used by the original authors of the FSS. Cronbach's alphas for the total
score was .85 and the total score split-half was .72. The subscale alphas
were: community a = .74; spouse and in-laws a = .78; friends a = .73;
specialized/professional a = .60; and own parents and extended family a
= .65.
74 Hanley, Tasse, Aman, and Pace

Table 2. Item Loadings for the Family Support Scale's 18 Items

Item Factor I Factor II Factor III Factor IV Factor V

1-parents 0.09 0.11 0.04 0.13 0.86

2-spouse's parents 0.08 0.82 0.05 0.03 0.24
3-relatives/kin 0.05 0.2 0.42 0.16 0.68
4-spouse's relatives 0.15 0.79 0.14 0.05 0.21
5-spouse -0.08 0.81 0.03 0.22 -0.06
6-friends 0.11 0.04 0.71 0.26 0.18
7-spouse's friends 0.16 0.52 0.54 0.22 -0.2
8-children 0.09 0.16 0.31 0.41 0.07
9-other parents 0.23 0.06 0.72 0.17 0.11
10-co-workers 0.47 0.2 0.55 -0.15 0.19
11-parent groups 0.69 -0.05 0.34 0.08 -0.03
12-social groups 0.84 0.14 0.12 0.06 0.02
13-church 0.55 -0.02 0.35 0.07 0.1
14-family/child's physician 0.31 0.23 0.36 0.46 0.22
15-early interv. program 0.5 0.08 0.2 0.38 -0.03
16-school/daycare 0.03 0.07 0.1 0.7 0.16
17-professional help -0.44 0.09 0.07 0.62 -0.01
18-professional agency 0.68 0.04 -0.08 0.22 0.14

Test-Retest

The Pearson product-moment correlation was used to estimate the de-

gree of reliability between the test and retest FSS ratings. In all, 54 parents
completed two FSS forms. Stability was estimated on the total score and
the subscale scores. The test-retest reliability for the total score was .73.
The test-retest reliability for each subscale was: community r = .60; spouse
and in-laws r = .65; friends r = .70; specialized/professional r = .78; and
own parents and extended family r = .70. All reliability coefficients were
significant at the p < .001.

DISCUSSION

Dunst et al. (1984) derived their six factor solution of the FSS on a
sample of 139 parents of preschool children with developmental disabilities.
A majority (60%) of the families from the Dunst et al. sample were of low
socio-economic status and all participated in an early intervention program.
Although our sample was comparable on some levels (children participated
in early intervention programs and families were from lower SES), we ob-
tained a slightly different factor solution. In a second factor solution (which
also resulted in five factors) with their initial sample of 139 parents, Dunst
and Trivette's (1986) item assignment shared 12 of 18 items (67%) with
Psychometric Properties 75

our factor structure. Whereas our factor structure contains the same num-
ber of factors, we obtained a slightly different item assignment, Taylor et
al. (1993) reported obtaining a four-factor solution derived from an aggre-
gated sample of 990 families recruited nationwide from several early
intervention studies. When we compared the Taylor et al. item assignment
with our own, we had similar item assignments on only 10 of 18 items
(56%). Taylor et al. also reported dropping one FSS item (i.e. school/day-
care center) due to its very low communality (<.20).
Several variables distinguish the Taylor et al. (1993) sample from our
own. Taylor et al.'s sample was predominantly white (84%) compared to
only 27% of our sample, and their mothers were mostly married (80%)
compared to 33% of married respondents in our sample. Overall, their sam-
ple seemed to be of higher socio-economic level, with an average income
of $23,000, whereas almost half of our sample (46%) had less than $5,000
annual income.
We suspect that our sample is more representative of Head Start fami-
lies than that of Taylor et al. (1943) and, thus, we feel that our five-factor
solution will be more stable and useful when using the FSS with Head
Start families or other low-income families. Clinically, the subscales make
good sense. Community agents loaded together under community supports.
All spousal items also loaded together, with the exception of "spouse's
friends" that loaded with "friends," "co-workers," and "other parents."
Items referring to family physician, daycare, and professional help loaded
under specialized and professional supports. Interestingly, the item refer-
ring to children loaded on this factor as well. As children are often not
considered a direct source of support for a parent, the children item may
reflect support groups directly associated with the child. The last factor
identifies parents and extended family as a fifth source of social support
for the family. Although this factor structure appears to be quite rational,
additional empirical research with low-income population is needed to con-
firm the stability of our proposed structure.
Overall, the reliability of the total and subscale scores was good. The
five subscales obtained in our study possessed moderate to high internal
consistency, expressed in the Cronbach's alpha (ranging from .60 to .78).
Internal consistency for the total scale (a = .85) compared favorably with
the total scale alpha of .77 reported by Dunst et al. (1984), and .80 by
Taylor et al. (1993).
The FSS is a reliable instrument for measuring family support. Family
support is a central construct related to a number of important research
and clinical issues, such as family functioning, well-being, parental percep-
tions of child functioning, and family integrity (Dunst, 1985). When
studying the support of low-income families, the five-factor structure re-
76 Hanley, Tasse, Aman, and Pace

ported in our study appears to be more appropriate than Dunst et al. (1984)
and Taylor et al. (1993) factor structures for assessing family functioning.

ACKNOWLEDGMENT

This research was supported by grant No. 90CD080901 from the De-
partment of Health and Human Services "University/Head Start
Partnerships" to Barbara Hanley.

REFERENCES

Anastasi, A. (1988). Psychological testing. New York: MacMillan.

Cohen, S., & Willis, T. A. (1985). Stress, social support, and the buffering hypothesis.
Psychological Bulletin, 98, 310-357.
D'Ecole, A. (1988) Single mothers: Stress, coping and social support. Journal of Community
Psychology, 16, 41-54.
Dixon, G. H. (1980). Child development in the family health care center. Young Children, 35,
49-56.
Dubrow, N. F., & Garbadino, J. (1989). Living in the war zone: Mothers and young children
in a public housing development. Child Welfare, 68, 3-20.
Dunst, C. J. (1985). Rethinking early intervention. Analysis and Intervention in Developmental
Disabilities, 5, 165-201.
Dunst, C. J., Jenkins, V., & Trivette, C. M. (1984). The Family Support Scale: Reliability and
validity. Journal of Individual, Family and Community Wellness, 1, 45-52.
Dunst, C. J., Jenkins, V., & Trivette, C. M. (1988). Family Support Scale. In C. J. Dunst, C.
M. Trivette, and A. G. Deal (Eds.), Enabling and empowering families: Principles and
guidelines for practice (pp. 153-174). Cambridge, MA: Brookline Books.
Dunst, C. J., & Trivette, C. M. (1986). Family Support Scale: Supplemental Scoring Instructions.
Morganton, NC: Western Carolina Center.
Gladow, N. W., & Ray, M. P. (1986). The impact of informal support systems on the well
being of low income single parents. Family Relations: Journal of Applied Family Child
Studies, 35, 113-123.
Jensen, M. A. (1983). Self concept and its relation to age, family structure and gender in
Head Start children. The Journal of Psychology, 1, 89-94.
Keltner, B. (1990). Family characteristics of preschool social competence among black children
in a Head Start program. Child Psychiatry and Human Development, 21, 95-108.
Lendblad-Goldberg, M., & Dukes, J. L. (1985). Social support in black, low income, single
parent families: Normative dysfunctional patterns. American Journal of Orthopsychiatry,
55, 42-58.
Light, J., & Martin, R. (1986). American Indian families. Journal of American Indian
Education, 26, 1-5.
Oyemade, U. J. (1985). The rationale for Head Start as a vehicle for the upward mobility of
minority families: A false minority perspective. American Journal of Orthopsychiatry, 55,
591-602.
Payne, C. (1989). Programs to strengthen families: A resource guide. Monograph. Chicago, IL.
Quay, L. (1989). Social interaction styles of socially normal, low interacting and
socially-emotionally handicapped school children. Child Study Journal, 18, 285-294.
Rueschenberg, E., & Buriel, R. (1989). Mexican-American family functioning and
acculturation: A family systems perspective. Hispanic Journal of Behavioral Sciences, 11,
232-244.
 Psychometric Properties 77

Sharni, S. (1980). Groups of culturally deprived parents: A multidimensional intervention

model. Small Group Behavior, 11, 345-356.
Slaughter, D., Lindsey, W. W., Nakagawa, K., & Kuehne, V. S. (1989). Who gets involved?
Head Start mothers as persons. Journal of Negro Education, 58, 16-29.
Stevens, J. H. (1988). Social support, locus of control and parenting in three low-income
groups of mothers: Black teenagers, black adults, and white adults. Child Development,
59, 635-642.
Stone, N. W., Pendleton, V. M., Vaill, M. B., Slatin, M., Mitcham, C., & Georgette, F. (1982).
Primary prevention in mental health: A Heat Start demonstration model. American
Journal of Orthopsychiatry, 52, 360-363.
Taylor, M. J., Crowley, S. L., & White, K. R. (1993, April). Measuring family support and
resources: Psychometric investigation of the FSS and FRS. Paper presented at the Annual
Meeting of the National Council on Measurement in Education. Atlanta, GA.
Washington, V. (1985). Head Start: How appropriate for minority families in the 1980's.
American Journal of Orthopsychiatry, 55, 577-590.
Zigler, E. (1985). Assisting Head Start at 20: An invited commentary. American Journal of
Orthopsychiatry, 55, 603-609.

View publication stats