Parent characteristics, economic stress and neighborhood context as predictors of parent

involvement in preschool children's education

By: Christine Waanders, Julia L. Mendez, and Jason T. Downer

Waanders, C., Mendez, J.L, & Downer, J. (2007). Parent characteristics, economic stress, and
neighborhood context as predictors of parent involvement in preschool children’s
education. Journal of School Psychology, 45, 619-636. doi:10.1016/j.jsp.2007.07.003

Made available courtesy of Elsevier: http://www.sciencedirect.com/science/journal/00224405

***Note: Figures may be missing from this format of the document

Abstract:
This study examines factors related to three dimensions of parent involvement in preschool:
school-based involvement, home-based involvement, and the parent–teacher relationship.
Participants were 154 predominantly African American parents recruited from two Head Start
programs. Results of bivariate and canonical correlation analyses support the validity of a multi-
dimensional, ecological conceptualization of parent involvement. Perceived context variables,
including economic stress and neighborhood social disorder, related negatively to parent
involvement. Parent characteristics, including sense of efficacy regarding education and level of
education, related positively to parent involvement. Regression analyses detected different
patterns of association between predictors and the three dimensions of parent involvement.
Parent characteristics were associated with home involvement, while perceived context variables
were predictive of the teacher–parent relationship. Implications of differential predictors for
different domains of parent involvement and directions for future research and intervention with
low-income families are discussed.
Keywords: Parent–school relationship; Project Head Start; Neighborhoods; Ecological factors;
Parental attitudes; Preschool teachers

Article:
Introduction
Children's development is influenced by factors at different ecological levels, including the
family, the school, the neighborhood, and society (Aber, Gephart, Brooks-Gunn, & Connell,
1997). For many years, researchers examined the separate impacts of family and school on
developmental trajectories, but focus has shifted to studying the link between these two settings
as a determinant of child outcomes ([Epstein, 1996] and [Grolnick and Slowiaczek, 1994]). This
home–school connection is represented in the early childhood and educational literature by the
construct of parent involvement, which refers to parents' participation in the education of their
children through behaviors that range from ideological support of education to active
communication with school personnel. For children from low-income families, parent
involvement in education can be a key protective factor that fosters cognitive and emotional
resilience in the face of multiple stressors ([Garmezy, 1991], [Myers and Taylor, 1998] and
[Shumow et al., 1999]). Thorough investigation into parent involvement and the determinants of
a high quality home–school connection is of considerable importance for understanding
preschool children's development ([Comer and Haynes, 1991] and [Epstein and Dauber, 1991]).
Head Start, an educational program initiated in the 1960s, has developed into the largest
federally funded program promoting school readiness for low-income preschool children. The
enriched preschool services offered by Head Start, including formalized involvement of parents,
help children enter kindergarten with better-developed cognitive and social skills (Takanishi &
DeLeon, 1994). Because Head Start programs have long recognized parent involvement as a key
component of school success for low-income children (U.S. Department of Health and Human
Services, 2000), they provide an ideal context in which to examine the complexities of the
home–school relationship during the preschool years. Using survey data obtained from over 150
minority families with children attending a Head Start program, the current paper extends the
literature by using a multi-method approach to study parent characteristics and contextual
variables, including neighborhood features, which may be predictive of parent involvement.

Involving parents in the educational process is particularly important for maximizing low-
income children's opportunities for academic success, as it has the potential to lessen the
discontinuity between the home and school environment ([Mendez and Fogle, 2002] and
[Slaughter-Defoe, 1995]). By involving parents, teachers' knowledge of their students' socio-
cultural context is enhanced, thereby helping them to deliver more culturally appropriate
educational services. Parents are also exposed to teachers who may model age-appropriate,
educational interactions with children (Haynes & Ben-Avie, 1996). Parent involvement can
promote positive adaptation to school and protect against negative outcomes for low-income
children, such as conduct problems or school failure (Alexander & Entwisle, 1996). In a study of
resilience among elementary-aged children, parent involvement was found to offset the negative
effects of living in a low-income, high-crime neighborhood on children's academic performance
(Shumow et al., 1999). Unfortunately, rigorous studies of the specific benefits associated with
parent involvement for low-income families during preschool are lacking (Mendez, submitted for
publication).

Multi-faceted nature of parent involvement

Parent involvement is frequently defined in school-centered terms, such as the frequency of
parents' visits to the school to volunteer or attend a conference with a teacher (Fantuzzo, Tighe,
& Childs, 2000). However, parent involvement in children's education can take a number of
forms, both within the home and at school ([Grolnick and Slowiaczek, 1994] and [Parker et al.,
1999]). As increasing numbers of low-income parents are experiencing significant time
constraints related to work, it is important for schools to offer ways for parents to be involved at
home (Marcon, 1999). Teacher and school characteristics may also be related to levels of parent
involvement (Eccles & Harold, 1996). Research has shown that teachers who hold more positive
attitudes toward parent involvement are more successful in involving ―hard-to-reach parents,‖
including working parents, single parents, and parents with low levels of education (Epstein &
Dauber, 1991). One study of Head Start staff practices found that when teachers received more
in-service trainings and offered more academically oriented activities for children at school,
parents of their students engaged in a greater variety of home-based learning activities (Barnes,
Guevara, Garcia, Levin, & Connell, 1997). Assessment of parent–teacher relationships may
therefore play an important role in enhancing our understanding of parent involvement.
Most recent models of parent involvement incorporate both home-based and school-based
activities ([Epstein, 1996] and [Fantuzzo et al., 2000]), yet few studies include objective
measures of involvement or consider the quality of the parent–teacher relationship. A major
contribution of the current paper is the use of a multidimensional, ecologically based, multi-
informant approach to the study of parent involvement. To this end, we obtained parent ratings
of their involvement in education at home and school, teacher ratings of their relationship with
each parent, and objective records of parent participation in Head Start center events and
meetings. These four dimensions represent overlapping yet distinct components of our
conceptual model of parent involvement during the preschool years.

Determinants of parent involvement

Eccles and Harold (1996) developed a model for examining determinants of parent involvement
that takes into account the multiple ecological systems influencing children. According to this
model, parent involvement is determined at the most proximal level by parents' beliefs and
values, as well as teachers' beliefs and practices specific to parent involvement. At more distal
levels, other child, parent, teacher, school, and neighborhood characteristics may have both
direct and indirect effects on parent involvement. For a fuller understanding of the factors that
lead some parents to be more involved than others, the current study considered multi-level
correlates of parent involvement, including a range of proximal and more distal factors. This
section reviews some prior work involving key parent characteristics and contextual variables
that are included as predictors of parent involvement in our study.

Family demographics are consistently related to levels of parent involvement in education. For
example, single parents tend to be less involved in educational activities with their children than
married parents (Zill, 1996). Studies of the role of socioeconomic status (SES) in parent
involvement suggest that lower SES parents are typically less involved in their children's schools
than middle or high SES parents (Dornbusch & Ritter, 1988). Parents with higher levels of
education have also been found to be more involved in their children's learning than parents with
lower levels of education (Fantuzzo et al., 2000). Hoover-Dempsey and Sandler (1995) argue,
however, that while demographic factors play a role, they are not the primary determinants of
whether and how parents become involved in their children's schooling. Instead, it is likely that
demographic variables serve as proxy variables for more complex dynamics within individuals
and communities (Coulton, Korbin, & Su, 1996), such as parenting efficacy, perceived economic
stress, and neighborhood context.

Parenting efficacy (i.e., a person's belief in his or her own competence to achieve a desired
parenting outcome) has been identified as a key determinant of parent involvement (Hoover-
Dempsey & Sandler, 1995). Downer and Mendez (2005) found significant relations between
African American fathers' self-reported efficacy regarding education and frequency of home-
based educational activities with their children enrolled in Head Start. Similarly, there is
evidence that parents with internal locus of control are more involved in educational activities at
home and at school than parents with external locus of control (Schaefer, 1991). It seems that
low-income parents are more likely than middle- and upper-income parents to view teachers as
the ―experts‖ in education, which may lead to a lower rate of involvement in educational
activities with their children (Crozier, 1999).
Only recently have models of parent involvement acknowledged the influence of more distal
factors, including neighborhood context, on parent involvement in education ([Eccles and
Harold, 1993] and [Smith et al., 1997]). Neighborhood structural factors such as residential
mobility, family disruption, housing and population density, and resource deprivation all
contribute to weakened community processes in low-income neighborhoods (Sampson, 1997).
Parents from higher-risk, lower resource neighborhoods may focus more on protecting children
from dangers than on fostering children's skill development ([Eccles and Harold, 1993],
[Furstenberg, 1993] and [O'Neil et al., 2001]). In one of the few identified empirical studies of
the issue, Smith et al. (1997) found that neighborhood climate was significantly associated with
parent involvement at school and at home for elementary school students. Given the growing
evidence for neighborhood effects on other family processes, further examination of the
relationship between perceived neighborhood context and parent involvement is warranted,
particularly among parents of younger children (Mendez, Stillman, LaForett, Wandersman, &
Flaspohler, 2004).

Research questions and hypotheses

The present study examines parent characteristics and perceived context in relation to the
multidimensional construct of parent involvement. It is hypothesized that parent involvement in
Head Start programs is best conceptualized as a set of interrelated dimensions involving home
activities, school-related contact, and the relationships between teachers and parents. We
intended to test whether parent involvement could be predicted by parents' perceptions of their
neighborhood context, parents' perceived economic stress, and parents' self-concept-specifically,
their sense of efficacy regarding their children's education. Thus, we investigated differential
patterns of prediction for the three specific dimensions of parent involvement (home-based
involvement, school-based involvement, and parent–teacher relationships). Analyses addressed
the following questions: (1) How are parent-report, teacher-report, and an objective record of
parent involvement activities associated? (2) How do parent characteristics (education, efficacy)
and perceptions of context (street crime, neighborhood disorder, local social networks, economic
stress) relate to parent involvement in Head Start? (3) What are the relative contributions of
parent characteristics and perceived context to different dimensions of parent involvement?

Method
Sample
Participants in the study were 154 caregivers or parents and 12 classroom teachers from two
Head Start centers in a medium-sized metropolitan area in the southeastern United States. The
children of these caregivers ranged in age from three to five and boys and girls were equally
represented. Ninety-five percent of the participating caregivers identified themselves as African
American, 5% reported they were European American, and 1% identified as Bi-racial. Sixty-two
percent of the participants were single, 22% percent were married, 6% were divorced, 7% were
separated, and 4% were widowed. A majority of the participants were mothers (87%), but the
sample also contained other primary caregivers, including fathers (9%), grandmothers (3%), and
aunts (1%). Most participants (67%) lived in rented homes, 20% percent were homeowners, and
another 9% were staying with friends. Twenty-eight percent of the participants had not moved at
all in the past 3 years, whereas 72% of the families had moved one or more times during the
same period. The majority of participants (56%) were employed full-time, while 16% percent
were employed part-time, 14% were unemployed and looking for work, 8% identified
themselves as homemakers, and 5% indicated that they were supported by disability benefits. All
participants had at least some high school education, and almost 80% had earned a high school
degree or reached a higher level of education. Center One had five classrooms, with a total of 96
children enrolled. Center Two had eight classrooms, with 158 children enrolled. Participation
rates were acceptable at both centers (64% and 63%, respectively). All participating teachers
were African American women.

Measures
Demographics
Caregivers completed a brief demographic survey regarding their relationship to the Head Start
student, their ethnicity, marital status, employment status, education level, living situation, ratio
of adults to children in the household, and number of residential moves in the past five years.

Economic stress
To assess parents' perceptions of economic stress, we administered two items developed and
validated by Conger and colleagues (1992). These items are rated on a 4-point Likert scale,
assessing the degree to which parents' income is sufficient to meet their expenses. Prior studies
found these items to correlate with each other at .65 (Conger et al., 1992), and to have a
Cronbach's alpha of .81 (Whitbeck et al., 1997). In the current study, the economic stress scale
had a Cronbach's alpha of .84.

Neighborhood quality
The Neighborhood Characteristics Questionnaire (NCQ), consisting of four subscales, was used
to assess parents' perceptions of their neighborhoods on both structural and social dimensions
(Barnes McGuire, 1997). Items from three of the subscales were utilized in the current study.
The Street Crime and Neighborhood Quality subscale assesses parents' perceptions of the
frequency of violent crime in their neighborhoods and their general perceptions of the quality of
their neighborhoods as places to live and raise children. The Neighborhood Disorder subscale
assesses the presence of ―incivilities‖ in parents' neighborhoods, including litter, graffiti, public
drug and alcohol use, and abandoned buildings. The Local Social Networks subscale provides a
measure of social cohesion by assessing the number of people parents know in their
neighborhoods, the number of friends they have in the neighborhood and the quality of their
contact with neighbors. A previous study showed that the NCQ has good internal consistency:
the Street Crime and Neighborhood Quality Scale had a Cronbach's alpha of .85, the Local
Social Networks Scale had a Cronbach's alpha of .82, and the Disorder Scale had a Cronbach's
alpha of .77 (Barnes McGuire, 1997). This measure was designed for use with parents of young
children, and has been shown to be sensitive to neighborhood variability in communities with
many risk factors (Barnes McGuire, 1997). In the current study, the internal consistency of all
three subscales was adequate: Cronbach's alpha for street crime was .86, for neighborhood
disorder was .76, and for local social networks was .90.

Parent efficacy
The About Being a Parent Scale (ABPS; Wentzel, 1993) was adapted from a measure of teacher
efficacy developed by Hoy and Woolfolk in 1993 (see Seefeldt, Denton, Galper, & Younoszai,
1998). The ABPS assesses parents' beliefs about their ability to influence their children's
educational outcomes, and it includes such items as, ―Even a parent with good teaching abilities
cannot teach his or her child as well as a classroom teacher,‖ and ―Parents do not have a
powerful influence on children's achievement when all factors are considered.‖ Parents rate their
agreement with each item along a 6-point Likert scale. The scale has demonstrated good internal
consistency with a Head Start population in past research (Seefeldt et al., 1998), and in the
current study Cronbach's alpha for the parent efficacy items was .79.

Parent involvement
Parent involvement was assessed using data from three sources: parent self-report, teacher
ratings, and an objective count of parent attendance at center events and meetings. The parent
self-report measure was the Family Involvement Questionnaire (FIQ), a multidimensional
measure of parent involvement in early childhood education (Fantuzzo et al., 2000). The FIQ
was developed for and field-tested with low-income families of preschool children, ensuring its
validity for use with the current target population. In the measurement development study,
content validity was established by using focus groups of ethnic minority parents of preschool
children to generate items for types of involvement. Factor analytic techniques confirmed three
independent constructs of parent involvement for this measure: Home-based, School-based, and
Home–School Conferencing. The Home-based Involvement subscale assesses behaviors that
parents engage in at home to promote learning, including provision of learning materials and
initiation of learning activities for their children at home or in the community. The School-based
Involvement subscale assesses parents' participation in activities such as volunteering in the
classroom and going on class trips with the children. The Home–School Conferencing subscale
assesses communication between school personnel and parents regarding children's difficulties
and accomplishments in the classroom. Previous research has shown the internal consistency of
the FIQ subscales to be high, with alpha coefficients greater than .80 (Fantuzzo et al., 2000). In
the present study, the School-based Involvement and Home–School Conferencing subscales
were composited to create a single school-based involvement variable. For parsimony, this
composite (α = .90) and home-based involvement (α = .88) were utilized in regression analyses.

The Connection Sort, a new method for obtaining teacher ratings of parent involvement, was
utilized in this study to assess teachers' relationship with parents. Teachers were asked to place
parents into one of four categories: ―strongly connected,‖ ―moderately connected,‖ ―a little
connected,‖ or ―not connected‖ with the process of their children's education. First, the names of
every child in each class were written on individual cards. Next, classroom teachers sorted their
students into four piles based on their level of connection with each child's parent or caregiver.
Teachers were provided with short definitions for the four categories, including behavioral
examples generated by the authors, and were asked to consider parents' involvement and
interactions with them from the start of the school year until the time of the assessment. For
example, ―strongly connected‖ parents have contact with the teacher ―once a week or more,‖ and
the teacher knows them ―quite well.‖ Parents who are ―a little connected‖ have contact with the
teacher only ―once or twice‖ during the year and are ―usually hard to reach‖ (full category
definitions are available from the authors). Parents were then assigned a score of 1, 2, 3, or 4,
based on the teacher's rating. The advantage of the sorting technique is that it offers teachers a
visual representation of the categories of involvement to help them consider the quality of each
parent's involvement. Teachers are able to modify the placement of parents in a more interactive
fashion than is typically possible with a Likert-type rating scale.
Finally, parent attendance was recorded using sign-in sheets for each event or meeting at the
Head Start center. The number of events attended by each parent was totaled. These data were
only available at the individual level for parents from Center One, due to the centers' record
formatting system.

Procedures
This study was undertaken in a collaborative manner with Head Start center families and staff.
Several meetings were held with staff and parents to provide an overview of the research
questions regarding barriers and benefits of parent involvement. A letter was also sent home in
children's school bags to inform parents briefly about the study and invite them to participate. At
the suggestion of parent leaders, researchers were available so that parents could complete the
study measures while bringing their children to school. All parents in the center were given a
packet containing instructions, a consent form, and the measures to complete independently. Our
use of several strategies to accommodate parents' schedules yielded a 64% participation rate
across centers. Following the study, two books were given to each child whether or not the
parent had completed the measures, in support of the university–community research
partnership.

The Connection Sort was administered individually to teachers by three trained research
assistants. To reduce social desirability bias, teachers were assured of the confidentiality of their
responses and encouraged to respond honestly, with the rationale that their responses would help
identify the challenges inherent in involving parents in the learning process. Teachers were
compensated for their participation with a small honorarium.

Data analytic plan

Analyses were conducted in two steps. First, Pearson product moment correlations and canonical
correlations were computed to assess the associations among different measures of parent
involvement and predictor variables (see Table 1). Canonical correlation is useful for offering a
parsimonious summary of the overall association between the set of personal/ contextual
predictors (which included perceived economic stress, parent efficacy regarding education, and
neighborhood features), and the set of parent involvement measures. We expected that each
dimension of parent involvement would have a significant loading within the canonical structure,
indicating that these dimensions were related, but distinct, aspects of parent involvement during
preschool.

Table 1.
Means, standard deviations, and bivariate correlations among predictors and parent involvement
measures
1 2 3 4 5 6 7 8 9 10 11 12

1-Centera –

2-Parent .04 –
education
level
 1 2 3 4 5 6 7 8 9 10 11 12

3-Parenting .01 .26 –

efficacy

4-Difficulty − .0 − .2 − .1 –
making ends 5 7 3
meet

5-Street crime .09 − .1 − .2 .23 –

5 3

6- − .0 − .1 − .1 .16 .69 –
Neighborhoo 1 3 7
d disorder

7-Local social − .1 .01 .07 − .0 − .0 − .1 –

networks 6 3 8 3

8-Parent– − .1 .15 .07 − .1 − .0 − .2 .02 –

teacher 5 9 8 0
relationship
(Q-sort)

9-Home- − .0 .20 .21 −.11 − .1 − .1 .20 .15 –

based parent 4 0 4
involvement
(FIQ)

10-School- − .1 .05 .03 − .0 − .1 − .1 .27 .20 .44 –

based parent 3 9 0 2
involvement
(FIQ)

11-Home– − .0 .14 .08 − .1 − .0 − .0 .20 .15 .60 .74 –

school 6 2 3 9
conferencing
(FIQ)

12-Parent – .31 .18 − .1 − .1 − .0 .14 .30 .23 .41 .34 –

attendance at 8 3 8
Head Start
center

Mean – 6.27 4.78 2.60 4.28 8.49 1.5 3.3 3.1 1.7 2.0 4.3
6 8 3 7 6 8
 1 2 3 4 5 6 7 8 9 10 11 12

Standard – 1.48 1.04 .82 4.03 5.42 1.7 .76 .58 .55 .59 3.7
deviation 7 6
FIQ = Family Involvement Questionnaire.
p < .05, p < .01, p < .001.
a
Mean and standard deviation not reported because this is a dichotomous variable.

Following this step, individual hierarchical linear regression analyses were used to more
precisely examine the relative predictive power of each variable in explaining variance
associated with the three dimensions of parent involvement: home-based involvement, school-
based involvement, and parent–teacher relationship. Predictor variables were entered into each
model in the following order: Center, Parent Characteristics (education and efficacy) and
Perceived Context (street crime, neighborhood disorder, local social networks, and economic
stress).

Results
Associations among measures of parent involvement
In examining the pattern of intercorrelations across measures, the three subscales of the Family
Involvement Questionnaire (FIQ) were highly correlated with one another, though more so for
School-based Involvement and Home–School Conferencing (r = .74, p < .001). Teachers' ratings
of their relationship with parents were significantly correlated with parent reports of School-
based Involvement (r = .20, p < .05), but were not significantly correlated with parent reports of
Home-based Involvement or Home–School Conferencing. Additionally, parents' attendance at
center events (for Center One only) significantly correlated with teacher ratings of their
relationship with parents (r = .30, p < .05), School-based Involvement (r = .41, p < .01) and
Home–School Conferencing (r = .34, p < .01). Attendance was not significantly correlated with
Home-based Involvement (r = .23). Because the objective count of parent attendance was only
available from parents at Center One, we did not further analyze these data. We created a
composite variable due to the high degree of overlap and conceptual similarity between Home–
School Conferencing and School-based Involvement.

Relations among parent characteristics, perceived context, and parent involvement

Canonical correlation analysis confirmed the association between the set of predictors and our
multidimensional assessment of parent involvement. Specifically, this analysis yielded four
variate pairs within the canonical structure; only the first canonical correlation (.35) accounted
for a significant amount of the overlapping variance (12%) between the sets of variables. With
the four canonical correlations included, the model was statistically significant (F = 1.81,
p < .01). Without the first variate, the model was no longer significant (F = 1.64, p < .06). This
sole significant variate, named ―global parent involvement,‖ reveals a significant association
between several parent characteristics/perceived context variables and the set of parent
involvement dimensions. Variables with a salient loading of .40 or greater (Weiss, 1972) were:
(1) parent efficacy; (2) economic stress; and (3) neighborhood disorder. For the variables
comprising the parent involvement set, all four dimensions contributed to the solution, with
salient loadings exceeding .40 (see Fig. 1). This analysis confirms that each measure of parent
involvement makes a contribution to the overall conceptualization of a multidimensional
construct. Also, global parent involvement is positively associated with parents' efficacy
regarding education, and negatively associated with perceived economic stress and neighborhood
disorder.

Fig. 1.

Relative contribution of parent characteristics and perceived context to variation in three

specific dimensions of parent involvement
To provide greater specificity to our understanding of each dimension of parent involvement, we
examined three hierarchical regression models. Parent–teacher relationship, home-based
involvement, and school-based involvement (a composite of School-based Involvement and
Home–School Conferencing of the FIQ) served as the dependent variables. Center was entered
first in each model as a covariate, followed by parent characteristics (education, efficacy) and
perceived context (street crime, neighborhood disorder, local social networks, economic stress).
Table 2 displays the standardized regression coefficients from final models and significance tests
for each model, as well as total R2's and changes in R2 for each block of predictors.

Table 2.
Hierarchical linear regression results from predicting three dimensions of parent involvementa
Parent–teacher Home-based School-based
relationship involvement involvement

Step 1 (R2 change) .02 .01 .01

Center − .19 − .02 − .07

Step 2 (R2 change) .03 .07 .01

Parent education .11 .15 .08

Parenting efficacy .01 .15 .01

Step 3 (R2 change) .08 .04 .07

 Parent–teacher Home-based School-based
relationship involvement involvement

Street crime .18 .05 .04

Neighborhood − .32 − .11 − .11

disorder

Local social − .05 .18 .22

networks

Economic stress − .16 − .03 − .08

R2 .13 .12 .09

Model F 2.77 2.62 2.00

N = 141 for parent–teacher relationship model and 148 for home- and school-based involvement
models.
p < .10, p < .05, p < .01.
a
Standardized betas from final model.

The regression model for teachers' relationship with parents was significant (F[6,140] = 2.72,
p < .01), accounting for 12% of the total variance. Only the perceived context block contributed
to a significant R2 change (8%); specifically, parents reporting higher levels of economic stress
and disorder in their neighborhoods were rated lower by teachers in regard to the quality of the
parent–teacher relationship. There also was a slight trend toward teachers in one center providing
higher relationship ratings as compared with teachers in the other center.

The regression model predicting parents' reported home-based involvement was also significant
(F[6,147] = 2.62, p < .05). The set of predictors accounted for 12% of the variance in home-
based involvement, but only the parent characteristic block contributed to a significant R2 change
(7%). In particular, there was a trend toward parents reporting greater home involvement when
they were more educated and reported greater feelings of efficacy regarding their children's
education. Also, though the perceived context block was not significant, local social networks
were positively linked to home-based involvement.

Finally, the overall school-based involvement model showed marginal significance

(F[6,147] = 2.00, p < .06). As a set, the predictors accounted for 9% of the variance in school-
based involvement, but only the perceived context block resulted in a significant R2 change (7%).
Specifically, parents who reported being more involved at school also perceived that they lived
in more socially cohesive, supportive neighborhoods.

Discussion
The purpose of this study was to gain a better understanding of the parent characteristics and
perceptions of context associated with parents' involvement in education within Head Start
programs. Several significant predictors of parent involvement were identified for this sample,
and results confirmed the importance of considering multiple dimensions of involvement during
preschool. Moreover, multiple regression analyses detected combinations of factors that
predicted different dimensions of parent involvement, which may be relevant for policy and
practice in early childhood education programs serving low-income populations.

Multidimensionality of parent involvement

The canonical correlation analysis was useful in providing empirical support for our
multidimensional model of parent involvement during preschool. The dimensions of home
involvement, school involvement, home–school conferencing, and teachers' connection with
parents were loaded significantly on a single canonical variate. This suggests that, despite
representing conceptually distinct dimensions of parent involvement, all four variables are part
of the same general construct of parent involvement. The inter-relatedness of these dimensions
argues for the multidimensional conception of parent involvement employed in this study, and
speaks to the importance of considering aspects of the surrounding context that influence
different types of parent involvement. However, some dimensions (e.g., school involvement and
conferencing) may have more overlap than other more distinct elements, especially activities in
the home and the parent–teacher relationship.

Considering the associations among the dimensions of parent involvement is also interesting,
particularly given the use of a multi-modal assessment strategy. First, the objective attendance
records and teacher reports of connection validate the parent reports of school involvement and
home–school conferencing. The converging pattern of associations reinforces the idea that these
dimensions tap into activities that are dependent upon parent interaction within the school
setting. Based on these data, it is likely that a stronger relationship between teachers and parents
is fostered by interactions within the school setting, which may run counter to a standard
preschool practice of conducting a home visit twice per year. Therefore, early childhood practice
that increases opportunities for parent visitation with teachers in school settings may be an
underutilized approach to fostering healthy home–school connection during preschool.

A second noteworthy pattern is that parents' reports of their involvement at home are unrelated to
objective indicators of attendance at school events or to the teacher–parent relationship. We
believe these results show the relative independence of educational involvement in the home and
school contexts. For example, teachers may be relatively unaware of parents' home involvement
in educational activities, particularly if work schedules or other barriers prevent high quality
teacher–parent communication from developing. Alternatively, home involvement may have
more to do with individual (e.g., efficacy) and family factors (adult–child ratio, parenting style)
than factors associated with the preschool setting. In summary, the results are consistent with our
emphasis on an ecological perspective that guides the study of parent involvement across the key
settings of home and preschool.

In recent years, there has been an increase in research examining multiple dimensions of parent
involvement (see [Fantuzzo et al., 2000], [Grolnick et al., 1997] and [Smith et al., 1997]). The
challenge remains, however, to understand more fully the proximal and distal ecological factors
that influence different dimensions of parent involvement, in order to inform policies and
practices aimed at increasing parent involvement. Prior research on barriers to parent
involvement has shown that several factors have a limiting effect: time constraints, including
schedule conflicts related to work or school, having a baby or toddler at home, and a lack of
energy or interest ([Gettinger and Guetschow, 1998] and [Parker et al., 2001]). Some parents
have more time and interest in center involvement, while others are more comfortable with or
available for home involvement. Current findings point to the importance of creating a variety of
parent involvement opportunities within Head Start. Historically, parent involvement
opportunities have been designed to meet school needs (Gettinger & Guetschow, 1998). Instead,
staff should make a concerted effort to tap parents' diverse abilities and interests and to consider
the limiting circumstances for families, in order to engage as many parents as possible.

Prediction of specific dimensions of parent involvement during preschool

Using a set of parent characteristics and contextual predictors, separate regression analyses were
conducted for three dimensions of parent involvement in order to determine whether prediction
models are distinctive, as reported in past research with an elementary school sample (Grolnick
et al., 1997). Results showed different patterns across the predictors, which may have
implications for the development and implementation of school- or home-focused parent
involvement interventions.

The parent–teacher relationship

The model predicting teachers' sense of connection with parents accounted for the greatest
amount of variance. Significant predictors within this model were perceived context variables,
namely parents' reports of neighborhood social disorder and economic stress. There were also
significant differences associated with center, with teachers from Center One rating parents as
more connected on average than those at Center Two. Size of program may have been an
unmeasured variable that could account for these differences. Center Two had almost twice as
many classrooms as Center One, which may have contributed to a different school climate or
culture—one less conducive to parent involvement. Barker's theory of behavior settings suggests
that people in smaller schools are more likely to get involved, and to get involved in multiple
ways (Barker, 1978). These data appear to support Barker's theory, though school climate within
Head Start centers is an area that merits more careful empirical investigation.

The additional predictors in this model, particularly neighborhood characteristics, are often
overlooked in studies of family functioning. Living in a community with greater social disorder
has been found to exacerbate parents' psychological distress, very likely leaving them with less
energy for activities like developing relationships with their children's teachers (Wandersman &
Nation, 1998). In addition, neighborhood social disorder may cause people to stay inside their
homes more, thereby depriving their children of potentially enriching educational experiences
outside the home (Furstenberg, 1993). Prior research also shows that, as a result of the emotional
distress that often accompanies economic hardship, parents are equipped with less energy for
involvement in educational activities (McLoyd, 1998). Additionally, the time burden faced by
low-income families trying to make ends meet and the greater inflexibility of many low-wage
jobs likely interfere with the quality of parent–teacher interactions. Parents working in the
service industry, for example, tend to have less control over their work schedules and are less
able to take time off for activities with their children (Wright & Smith, 1998). Overcoming such
barriers requires schools to engage in alternative, non-traditional practices to promote greater
involvement among these parents.

Parent involvement in education at home

The model for home involvement revealed that the block of parent characteristics was most
useful in explaining variance, along with the contribution of local social networks. Specifically,
parents who were more involved in home educational activities were those parents with more
years of education, a greater sense of efficacy regarding their children's education, and a strong
social network. It is noteworthy that home involvement is the only dimension of parent
involvement for which parent efficacy regarding education was a significant predictor. This may
be due to the fact that home involvement is the one dimension that is comprised of decisions
driven exclusively by parents. Perhaps more initiative is required on the part of the parent to
engage in independent educational activities at home, as compared to attending programs held at
the center. In considering relevant intervention strategies, it may be that a parent's strong belief
in the importance of his or her role in a child's education is a necessary precursor to extending
educational involvement beyond the confines of Head Start and into the home. The present data
show that parents who perceived themselves as important agents in the education of their
children were more likely to be involved in educational activities as a whole. Parents who
endorsed the belief that education of their children was solely the job of the teacher and the
school tended to be less involved.

These results are generally consistent with prior work examining ethnic minority families and
parental efficacy. For example, Smith and colleagues (1997) found that parents' attitudes toward
involvement were related to home involvement but not school involvement. Similarly, Downer
and Mendez (2005) showed that African American fathers' ratings of efficacy using the same
measure as the present study were associated with greater involvement in education at home, but
not at school. In contrast, Grolnick et al. (1997) found that school involvement was affected by
parents' efficacy regarding education, while ―personal‖ involvement (analogous to Home
Involvement) was not. Perhaps the converging results for ethnic minority families in particular
suggest that promoting efficacy or control over one's own environment is a salient dimension
upon which to build intervention programming.

Prior work has also confirmed the role of higher educational status in parent involvement, even
within a low-income sample. Kohl, Lengua and McMahon (2000) found that parents' education
level was positively associated with home involvement, school involvement, and parent–teacher
contact, but not with the quality of the parent–teacher relationship. Dauber and Epstein (1993)
found both home and school involvement to be significantly related to parents' education level.
Fantuzzo and his colleagues (2000) found that home–school conferencing and school
involvement were impacted by parental education level, while home involvement was not. The
results of the current study confirm some of these findings, but contradict others.

Home involvement was associated with parents' education, while parent ratings of school
involvement and home–school conferencing were not. It has been argued that parents who are
more highly educated place a greater value on education, and therefore, get more involved in
their children's educational activities (Kohl et al., 2000). This would explain the greater degree of
home involvement by more educated parents in the present sample, but it does not explain the
lack of effect of education on parent-reported school involvement. Again, it may be that school
involvement, which is mandated in Head Start programs, is less variable and less sensitive to
differences in parents' personal characteristics. The inconsistencies between present findings and
existing literature indicate the need for further research into the impact of education level on
different types of parent involvement.

Parent involvement in education at school

The proposed ecological model explained much smaller amounts of variance in parent-reported
school involvement/conferencing than in the other dimensions. A strong local social network,
defined as perceived social cohesion within parents' neighborhoods, was the only predictor that
accounted for a significant amount of variance in school involvement. When parents knew many
of the people in their neighborhoods and reported having positive interactions with neighbors,
they were more involved at school and participated in home–school conferences. Thus, it may be
that having neighbors who can help with child monitoring, for example, makes it easier to get
involved at the Head Start center. Alternatively, social networks may be fostered within a
smaller, more intimate community, which paves the way for parents to expect social interactions
within the school context as well. Prior research has confirmed that neighborhood social support
is vital to single mothers' well-being (Goldberg, Greenberger, Hamill, & O'Neil, 1992), a group
with a significant representation in the sample for this study. Lastly, the limited variability in
reported school involvement (as measured closer to the end of school year) may account for the
lack of other significant relationships involving perceived context variables or parent
characteristics.

Unique contributions and limitations of the present study

This study extends our understanding of the multidimensionality of parent involvement and
ecological factors that may serve as barriers or supports to different types of educational
involvement in low-income families during preschool years. A limitation of the study is that the
sample of low-income parents may be more involved than usual, due to their participation in
Head Start and their involvement in this study. We acknowledge that these results may not
generalize to other populations, and the restricted range of involvement can limit our
understanding of the phenomenon. We are also unable to conclude that the parent characteristics
and ecological variables included in this study are directly responsible for parental involvement
levels, which speaks to the need for intervention designs that can test the relative importance of
these variables in producing changes in parent involvement (Mendez, submitted for publication).
Many of the measures employed in the study were developed and validated for use with a
population of African American mothers, which is a strength of this research. Unfortunately,
there are likely method effects involving the self-report measures. These could be remedied with
additional observational or objective measures (e.g., census data for neighborhood context).

A unique contribution of this study is the support it lends for assessing the personal connection
between teachers and parents when considering parent involvement. This dimension of parent
involvement deserves further attention, as it appears to be distinct from home involvement and
home–school conferencing (Kohl et al., 2000). The Teacher–Parent Connection Sort technique
developed for this study emerged as a valuable measurement tool for assessing teachers'
perceptions of their relationships with parents. This measure was validated through correlations
with the FIQ School Involvement dimension and objective attendance data. This rating technique
assesses the quality of parent involvement rather than measuring quantity, and previous research
has found quality of parent involvement to be more strongly related to child outcomes than
quantity (Kohl et al., 2000). Beyond its usefulness for parent involvement research, this measure
may be used to reflect upon and improve teacher practices. Teachers could use the Connection
Sort periodically throughout the school year, in order to assess how well they are reaching the
parents in their classroom. Future research could track changes in this relationship over time, as a
measure of the outcome of parent involvement interventions. For both practical and research
purposes, it would also be interesting to create a similar Connection Sort system by which
parents could rate their relationships with teachers.

In summary, the results of the present study provide support for the proposed ecological model
of parent involvement, and have important implications for Head Start practitioners interested in
increasing parent involvement. It is clear that parent involvement is a function of the interaction
between family, school, and community factors and certainly not the responsibility of parents
alone. Previous research has shown that parents do not tend to identify economic and
neighborhood factors as barriers to parent involvement (Parker et al., 2001). However, current
results demonstrate that these factors do, in fact, significantly relate to involvement for many
parents. Teachers, too, may be unaware of the significant impact of these contextual variables on
parent involvement. Increasing teachers' awareness of the economic and community issues that
impact families could foster more positive attitudes toward parents' involvement in their
children's education. Understanding the effects of contextual factors, such as perceived economic
stress and neighborhood social cohesion, could allow educators to better target their efforts to
promote parent involvement.

References
Aber et al., 1997 J.L. Aber, M.A. Gephart, J. Brooks-Gunn and J.P. Connell, Development in
context: Implications for studying neighborhood effects. In: J. Brooks-Gunn, G.J. Duncan and
J.L. Aber, Editors, Neighborhood poverty, Context and consequences for children Vol. 1, Russell
Sage Foundation, New York (1997), pp. 44–61.
Alexander and Entwisle, 1996 K.L. Alexander and D.R. Entwisle, Schools and children at risk.
In: A. Booth and J.F. Dunn, Editors, Family–school links: How do they affect educational
outcomes?, Lawrence Erlbaum Associates, Publishers, Mahwah, NJ (1996), pp. 67–88.
Barker, 1978 R.G. Barker, Theory of behavior settings. In: R.G. Barker, Editor, Habitats,
environments, and human behavior: Studies in eco-behavioral science from the midwest
psychological field station, Stanford University Press, Stanford, CA (1978).
Barnes et al., 1997 H.V. Barnes, M.D. Guevara, G. Garcia, M. Levin and D.B. Connell, How do
Head Start staff characteristics relate to parent involvement and satisfaction?, Family and child
experiences survey, U.S. Department of Health and Human Services: Administration on
Children, Youth, and Families (1997).
Barnes McGuire, 1997 J. Barnes McGuire, The reliability and validity of a questionnaire
describing neighborhood characteristics relevant to families and young children living in urban
areas, Journal of Community Psychology 25 (6) (1997), pp. 551–566.
Comer and Haynes, 1991 J.P. Comer and N.M. Haynes, Parent involvement in the schools: An
ecological approach, The Elementary School Journal 91 (3) (1991), pp. 271–278.
Conger et al., 1992 R.D. Conger, K.J. Conger, G.H. Elder Jr., F.O. Lorenz, R.L. Simons and L.B.
Whitbeck, A family process model of economic hardship and adjustment of early adolescent
boys, Child Development 63 (1992), pp. 526–541.
Coulton et al., 1996 C.J. Coulton, J.E. Korbin and M. Su, Measuring neighborhood context for
young children in an urban area, American Journal of Community Psychology 24 (1) (1996), pp.
5–32.
Crozier, 1999 G. Crozier, Is it a case of ―We know when we're not wanted‖? The parents'
perspective on parent–teacher roles and relationships, Educational Research 41 (3) (1999), pp.
315–328.
Dauber and Epstein, 1993 S.L. Dauber and J.L. Epstein, Parents' attitudes and practices of
involvement in inner-city elementary and middle schools. In: N. Chafkin, Editor, Families and
schools in a pluralistic society, State University of New York Press, Albany, NY (1993), pp. 53–
72.
Dornbusch and Ritter, 1988 S.M. Dornbusch and P.L. Ritter, Parents of high school students: A
neglected resource, Educational Horizons 66 (1988), pp. 75–77.
Downer and Mendez, 2005 J.T. Downer and J.L. Mendez, African American father involvement
and preschool children's school readiness, Early Education and Development 16 (2005), pp.
317–340.
Eccles and Harold, 1993 J.S. Eccles and R.D. Harold, Parent–school involvement during the
early adolescent years, Teachers College Record 94 (1993), pp. 560–587.
Eccles and Harold, 1996 J.S. Eccles and R.D. Harold, Family involvement in schooling. In: A.
Booth and J.F. Dunn, Editors, Family–school links: How do they affect educational outcomes?,
Lawrence Erlbaum Associates, Publishers, Mahwah, NJ (1996), pp. 3–34.
Epstein, 1996 J.L. Epstein, Perspectives and previews on research and policy for school, family,
and community partnerships. In: A. Booth and J.F. Dunn, Editors, Family–school links: How do
they affect educational outcomes? (1996), pp. 209–246.
Epstein and Dauber, 1991 J.L. Epstein and S.L. Dauber, School programs and teacher practices
of parent involvement in inner-city elementary and middle schools, Elementary School Journal
91 (3) (1991), pp. 289–305.
Fantuzzo et al., 2000 J. Fantuzzo, E. Tighe and S. Childs, Family Involvement Questionnaire: A
multivariate assessment of family participation in early childhood education, Journal of
Educational Psychology 92 (2) (2000), pp. 367–376.
Furstenberg, 1993 F.F. Furstenberg, How families manage risk and opportunity in dangerous
neighborhoods. In: W.J. Wilson, Editor, Sociology and the public agenda, Sage, Newbury Park,
CA (1993), pp. 231–258.
Garmezy, 1991 N. Garmezy, Resiliency and vulnerability to adverse developmental outcomes
associated with poverty, American Behavioral Scientist 34 (1991), pp. 416–430.
Gettinger and Guetschow, 1998 M. Gettinger and K.W. Guetschow, Parental involvement in
schools: Parent and teacher perceptions of roles, efficacy, and opportunities, Journal of Research
and Development in Education 32 (1) (1998), pp. 38–52.
Goldberg et al., 1992 W.A. Goldberg, E. Greenberger, S. Hamill and R. O'Neil, Role demands in
the lives of employed single mothers with preschoolers, Journal of Family Issues 13 (3) (1992),
pp. 312–333.
Grolnick et al., 1997 W.S. Grolnick, C. Benjet, C.O. Kurowski and N.H. Apostoleris, Predictors
of parent involvement in children's schooling, Journal of Educational Psychology 89 (3) (1997),
pp. 538–548.
Grolnick and Slowiaczek, 1994 W.S. Grolnick and M.L. Slowiaczek, Parents' involvement in
children's schooling: A multidimensional conceptualization and motivational model, Child
Development 65 (1994), pp. 237–252.
Haynes and Ben-Avie, 1996 N.M. Haynes and M. Ben-Avie, Parents as full partners in
education. In: A. Booth and J.F. Dunn, Editors, Family–school links: How do they affect
educational outcomes?, Lawrence Erlbaum Associates, Publishers, Mahwah, NJ (1996), pp. 45–
55.
Hoover-Dempsey and Sandler, 1995 K.V. Hoover-Dempsey and H.M. Sandler, Parental
involvement in children's education: Why does it make a difference?, Teachers College Record
97 (2) (1995), pp. 310–331.
Kohl et al., 2000 G.O. Kohl, L.J. Lengua and R.J. McMahon, Parent involvement in school:
Conceptualizing multiple dimensions and their relations with family and demographic risk
factors, Journal of School Psychology 38 (6) (2000), pp. 501–523.
Marcon, 1999 R. Marcon, Positive relationships between parent school involvement and public
school inner-city preschoolers' development and academic performance, School Psychology
Review 28 (3) (1999), pp. 395–412.
McLoyd, 1998 V.C. McLoyd, Socioeconomic disadvantage and child development, American
Psychologist 53 (2) (1998), pp. 185–204.
Mendez, submitted for publication Mendez, J. L. (submitted for publication). The Companion
Curriculum: A community-based preventive intervention for enhancing parent involvement in
Head Start preschool programs. Manuscript.
Mendez and Fogle, 2002 J.L. Mendez and L.M. Fogle, Parental reports of preschool children's
social behavior: Relations among peer play, language competence and problem behavior,
Journal of Psychoeducational Assessment 20 (2002), pp. 374–389.
Mendez et al., 2004 J.L. Mendez, L. Stillman, D. LaForett, A. Wandersman and P. Flaspohler,
Neighborhood and community influences favoring the growth and development of young
children. In: T. Gullota and M. Bloom, Editors, A blueprint for the promotion of prosocial
behavior among young children, CWA Press, Stanford, CT (2004).
Myers and Taylor, 1998 H. Myers and S. Taylor, Family contributions to risk and resilience in
African American children, Journal of Comparative Family Studies 29 (1) (1998), pp. 215–230.
O'Neil et al., 2001 R. O'Neil, R.D. Parke and D. McDowell, Objective and subjective features of
children's neighborhoods: Relations to parental regulatory strategies and children's social
competence, Applied Developmental Psychology 22 (2001), pp. 135–155.
Parker et al., 1999 F. Parker, A.Y. Boak, K.W. Griffin, C. Ripple and L. Peay, Parent–child
relationship, home learning environment, and school readiness, School Psychology Review 28 (3)
(1999), pp. 413–425.
Parker et al., 2001 F.L. Parker, C.S. Piotrowski, A.J.L. Baker, S. Kessler-Sklar, B. Clark and L.
Peay, Understanding barriers to parent involvement in Head Start: A research–community
partnership, Early Childhood Research Quarterly 16 (2001), pp. 35–51.
Sampson, 1997 R.J. Sampson, Collective regulation of adolescent misbehavior: Validation
results from eighty Chicago neighborhoods, Journal of Adolescent Research 12 (2) (1997), pp.
227–244.
Schaefer, 1991 E.S. Schaefer, Goals for parent and future–parent education: Research on
parental beliefs and behaviors, The Elementary School Journal 91 (3) (1991), pp. 239–248.
Seefeldt et al., 1998 C. Seefeldt, K. Denton, A. Galper and T. Younoszai, Former Head Start
parents' characteristics, perceptions of school climate, and involvement in their children's
education, Elementary School Journal 98 (4) (1998), pp. 339–349.
Shumow et al., 1999 L. Shumow, D.L. Vandell and J. Posner, Risk and resilience in the urban
neighborhood: Predictors of academic performance among low-income elementary school
children, Merrill-Palmer Quarterly 45 (2) (1999), pp. 309–331.
Slaughter-Defoe, 1995 D.T. Slaughter-Defoe, Revisiting the concept of socialization: Caregiving
and teaching in the 90s: A personal perspective, American Psychologist 50 (1995), pp. 276–286.
Smith et al., 1997 E.P. Smith, C. Connell, G. Wright, M. Sizer, J. Norman and A. Hurley et al.,
An ecological model of home, school, and community partnerships: Implications for research
and practice, Journal of Educational and Psychological Consultation 8 (4) (1997), pp. 339–360.
Takanishi and DeLeon, 1994 R. Takanishi and P.H. DeLeon, A Head Start for the 21st century,
American Psychologist 49 (2) (1994), pp. 120–122.
U.S. Department of Health and Human Services, 2000 U.S. Department of Health and Human
Services. (2000). Head Start performance standards (1304.40). Washington, DC.
Wandersman and Nation, 1998 A. Wandersman and M. Nation, Urban neighborhoods and
mental health: Psychological contributions to understanding toxicity, resilience, and
interventions, American Psychologist 53 (6) (1998), pp. 647–656.
Weiss, 1972 D.J. Weiss, Canonical correlation analysis in counseling psychology research,
Journal of Counseling Psychology 19 (3) (1972), pp. 241–252.
Wentzel, 1993 K. Wentzel, About being a parent, University of Maryland, Department of Human
Development, College Park (1993).
Whitbeck et al., 1997 L.B. Whitbeck, R.L. Simons, R.D. Conger, K.A.S. Wickrama, K.A.
Ackley and G.H. Elder Jr., The effects of parents' working conditions and family economic
hardship on parenting behaviors and children's self-efficacy, Social Psychology Quarterly 60 (4)
(1997), pp. 291–303.
Wright and Smith, 1998 G. Wright and E.P. Smith, Home, school, and community partnerships:
Integrating issues of race, culture, and social class, Clinical Child and Family Psychology
Review 1 (3) (1998), pp. 145–162.
Zill, 1996 N. Zill, Family change and student achievement: What we have learned, what it means
for schools. In: A. Booth and J.F. Dunn, Editors, Family–school links: How do they affect
educational outcomes?, Lawrence Erlbaum, Mahwah, NJ (1996), pp. 139–174.