Relationships between child life events stress, parenting stress, and child verbal cognitive functioning of low income p...

MISSING IMAGE

Material Information

Title:
Relationships between child life events stress, parenting stress, and child verbal cognitive functioning of low income preschoolers
Physical Description:
Unknown
Creator:
McManis, Lilla Dale
Publication Date:

Record Information

Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 23228911
oclc - 35069747
System ID:
AA00020425:00001

Full Text










RELATIONSHIPS BETWEEN CHILD LIFE EVENTS STRESS,
PARENTING STRESS, AND
CHILD VERBAL COGNITIVE FUNCTIONING
OF LOW INCOME PRESCHOOLERS










By

LILLA DALE McMANIS


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


1996













AKNOWLEDGMENTS

I would like to thank several people who were instrumental in seeing me through

my educational endeavor. There are many steps that are taken before even getting to the

point of working on a doctorate.

I would like to thank particularly my parents, Ella Mae and Ralph Rosenberger and

Gordon Wadsworth, and my sister, Brenda Reeves. They always encouraged me. More

recently with respect to my earning a Ph.D., I had the warm and loyal support of my

husband, Mark McManis and the unconditional love of our little boy, Perry McManis.

My committee was professional and patient. I would especially like to thank my

chairman, Dr. Gordon Greenwood for his availability and attention to detail. I am

indebted to Dr. David Miller for his teaching ability and statistical advice.

I would like to thank God for bringing all these people into my life and carrying me

through this journey. I have learned a great deal about education, psychology, and myself.







scientist and researcher. I owe much of my success to them for taking the time
to teach me how to be a professor. I hope that my career reflects the principles
they have taught me.
This dissertation is the product of many years of effort by many people.

The specific content of this dissertation, however, is mine, as are any errors the
reader may find. I hope that my efforts reflect well on those who have helped
me complete my dissertation. I have not been able to recognize all who have

contributed to my education here, but those who have helped can be assured I
have not forgotten them and I am grateful for their assistance.









Instrumentation...................................................................... 84
Child Verbal Cognitive Functioning......................................... 84
Child Life Events Stress....................................................... 89
Parenting Stress ................................................................ 92
Demographic Variables........................................................ 97
Procedure............................................................................ 98
Pilot Study.......................................................................... 100
Statistical Analysis................................................................. 102

4 RESULTS.......................................................................... 105

Purpose ............................................................................. 105
Tests of the Hypotheses........................................................... 108
Hypothesis 1-Overall Model................................................. 110
Hypothesis 2-Child Verbal Cognitive Functioning........................ 113
Hypothesis 3-Demographics and Child Verbal Cognitive Functioning. 116
Hypothesis 4-Parenting and Child Life Events Stress .................... 116
Hypothesis 5-Child Life Events Stress and Demographics.............. 133
Hypothesis 6-Parenting Stress and Demographics........................ 136
Supplemental Analyses............................................................ 145
Summary............................................................................ 146

5 SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS........... 153

Summary............................................................................ 153
Purpose......................................................................... 153
Subjects......................................................................... 153
Instrumentation................................................................ 153
Results.......................................................................... 154
Summary....................................................................... 155
Conclusions ........................................................................ 156
Child Verbal Cognitive Functioning ........................................ 156
Maternal age............................................................... 156
Child acceptability........................................................ 157
Maternal depression ...................................................... 157
Child life entrance events................................................ 159
Maternal employment..................................................... 160
Concluding Statements about Child Verbal Cognitive Functioning..... 161
Child Life Events and Parenting Stress..................................... 162
Vulnerability of mothers................................................. 163
Vulnerability of children................................................. 165
Resiliency of mothers.................................................... 166
Resiliency of children .................................................... 167
Concluding Statements about Vulnerability in Mothers and Children... 169
Mothers .................................................................... 169
Children.................................................................... 170
Concluding Statements about Resiliency in Mothers and Children...... 171
Mothers .................................................................... 171
Children.................................................................... 172
Demographic Variables and Stress.......................................... 173
Ethnicity.................................................................... 173
Maritial status.............................................................. 174
Education level............................................................ 174








Age ......................................................................... 175
Employment status........................................................ 176
Concluding Statements about Demographics and Stress................. 176
Closing Thoughts about Findings........................................... 178
Recommendations................................................................. 179

REFERENCES......................................................................... 182

APPENDICES.......................................................................... 200

A. DIAL-R......................................................................... 200
B. LER-P .......................................................................... 205
C. PSI.............................................................................. 207
D. ADMINISTRATION PROCEDURES ..................................... 215
E. INFORMED CONSENT..................................................... 219
F. ZERO ORDER CORRELATIONS.......................................... 221

BIOGRAPHICAL SKETCH......................................................... 229








LIST OF TABLES

Table pag

1 Means and Standard Deviations of All Variables................................. 107

2 Intercorrelations of Parent-Child Relationship Stress
(Parenting Stress Index) Factors........................................... 109

3 Relationships between DIAL-R and Child Life Events Record
Factors, Parenting Stress Index Factors, and Maternal
Demographic Variables...................................................... 111

4 Relationships between Child Life Events Loss Factor and Parenting
Stress Index Factors and Maternal Demographic Variables ............ 119
5 Relationships between Child Life Events Entrance Factor and
Parenting Stress Index Factors and Maternal Demographic
V ariables...................................................................... 121

6 Relationships between Child Life Events factor Family Troubles
and Parenting Stress Index Factors and Maternal Demographic
Variables ..................................................................... 123

7 Relationships between Child Life Events Factor Positive Events and
Parenting Stress Index Factors and Maternal Demographic
V ariables...................................................................... 125

8 Relationships between Child Life Events Factor Physical Harm and
Parenting Stress Index Factors and Maternal Demographic
Variables...................................................................... 125
9 Relationships between Child Life Events Factor Sibling Problems and
Parenting Stress Index Factors and Maternal Demographic
V ariables ..................................................................... 127

10 Relationships between Child Life Events Factor Primary Environment
Change and Parenting Stress Index Factors and Maternal
Demographic Variables ..................................................... 130

11 Relationships between Parenting Stress Index Factor Child Adaptability
and Child Life Events and Maternal Demographic Variables........... 137

12 Relationships between Parenting Stress Index Factor Child Acceptability
and Child Life Events Stress Factors and Maternal Demographic
Variables...................................................................... 137
13 Relationships between Parenting Stress Index Factor Child
Demandingness and Child Life Events Stress Factors and
Maternal Demographic Variables........................................... 138
14 Relationships between Parenting Stress Index Factor Child Mood and
Child Life Events and Maternal Demographic Variables................ 139








15 Relationships between Parenting Stress Index Factor Child
Distractibility and Child Life Events Factors and Maternal
Demographic Variables...................................................... 139

16 Relationships between Parenting Stress Index Factor Child Reinforces
Parent and Child Life Events Factors and Maternal Demographic
V ariables...................................................................... 140

17 Relationships between Parenting Stress Index Factor Parent Depression
and Child Life Events Factors and Maternal Demographic
V ariables...................................................................... 140

18 Relationships between Parenting Stress Index Factor Parent Restriction
of Role and Child Life Events and Maternal Demographic
V ariables...................................................................... 141
19 Relationships between Parenting Stress Index Factor Parental Sense
of Competence and Child Life Events Stress Factors and Maternal
Demographic Variables...................................................... 142

20 Relationships between Parenting Stress Index Factor Parent Social
Isolation and Child Life Events Factors and Maternal Demographic
V ariables...................................................................... 142

21 Relationship between Parenting Stress Index Factor Parent Relationship
with Spouse and Child Life Events Factors and Maternal
Demographic Variables...................................................... 143

22 Relationships between Parenting Stress Index Factor Parent Health
and Child Life Events Factors and Maternal Demographic
V ariables...................................................................... 144
23 Summary of Direction and Percent Variance Accounted for in
Relationships between the DIAL-R and the PSI Factors,
CLER Factors and Maternal Demographic Variables.................... 147

24 Summary of Direction and Percent Variance Accounted for in
Relationships between the CLER Factors and the PSI Factors
and Maternal Demographic Variables...................................... 148

25 Summary of Direction and Percent Variance Accounted for in
Relationships between the PSI Factors and Maternal
Demographic Variables...................................................... 149













Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

RELATIONSHIPS BETWEEN CHILD LIFE EVENTS STRESS,
PARENTING STRESS, AND
CHILD VERBAL COGNITIVE FUNCTIONING
OF LOW INCOME PRESCHOOLERS

By
LILLA DALE McMANIS

May 1996
Chairman: Gordon Greenwood
Major Department: Foundations of Education

This study examined child life events stress, parenting stress, child verbal cognitive

functioning, and the demographic variables maternal age, employment status, marital

status, education level, and ethnicity in 49 randomly selected Head Start preschooler-

mother dyads, extending research with older students. Child verbal cognitive functioning

(language and concepts) was measured by the Developmental Indicators for the Assessment

of Learning-Revised (DIAL-R), child life events by the Coddington Life Events Record-

Preschool Version (30 events in seven areas), and parenting stress by the Parenting Stress

Index (101 items in thirteen areas). Mothers completed the stress measures and

demographics under the researcher's supervision. Head Start provided DIAL-R scores.

Data were analyzed using multiple regression.

Results revealed that the variables more indicative of higher child verbal cognitive

functioning were those revolving around relationships between the preschooler and

important adults, namely mothers' acceptance of child, and entrance of stepfather and

relatives; along with low-income mothers' dissatisfaction with the circumstances under

which they raised their child. However, sibling births were related to lower








DIAL-R scores. Increasing maternal age was a strong predictor of higher DIAL-R scores

as was mother being employed. Both significant positive and negative correlations were

found between child life events stress and parenting stress. A positive relationship was

discovered between life events and parenting stress in the form of mothers experiencing

more stress in work and personal relationships with adults. On the other hand, negative

relationships were found between life events and parenting stress with particular focus on

lower stress in mother-child psychological closeness. White mothers reported more life

events stress while minority mothers reported more parenting stress.

The theory of viewing stress as change and the differing reaction to stress either

being one of resiliency or vulnerability was confirmed at least by this sample of low-

income preschoolers and their mothers. Overall, both higher and lower stress were related

to higher and lower child verbal cognitive functioning. Perhaps the most significant

conclusion was that low-income, Head Start preschoolers with higher verbal cognitive

functioning scores had close relationships with a greater number of important adults.














CHAPTER 1
INTRODUCTION


Statement of the Problem
One solution to the differential performance between children from lower- and

upper-income backgrounds has been early intervention in the form of structured preschool

experiences. This approach is supported by studies that link cognitive functioning such as

language and conceptual knowledge tasks to children's experiences early in life (e.g.,

Bernstein, 1961; Hess & Shipman, 1965; Jones, 1972). In particular, verbal cognitive

performance has been linked to characteristics and aspects of the home and the quality of

the relationship between the parent and the child. The home environment, including

physical surroundings and parent-child interactions, is thought to substantially contribute to

cognitive functioning and performance in children (c.f., Caldwell & Bradley, 1984; Jones,

1972; Laosa, 1982a, 1982b; Sameroff & Seifer, 1983; Wemrner & Smith, 1982).

Early childhood intervention programs have had success in improving social skills

and providing temporary gains in cognitive functioning, but long-term intellectual gains in

low-income children have not been maintained overall for program participants

(e.g., Evard, 1988; Brown & Rosenbaum, 1984). For example, Evard (1988) reviewed

almost 40 studies assessing early childhood education program elements in home or

preschool intervention approaches. In Evard's review article, home interventions were

programs with home visits by staff to teach mothers how to interact to be more effective

teachers of their young children. The preschool interventions were school or classroom-

based methods of guiding young children in group situations such as those found in

nursery schools.








The conclusion of Evard's review of these 38 program evaluations was that the

home intervention programs were generally beneficial in their ability to consistently

produce gains in the child's IQ and cognitive functioning (e.g., achievement tests).

Benefits of parental involvement elements present in some of the school-based early

education preschool programs were seen, reiterating the importance of the parent when

these programs were evaluated (Lazar & Darlington, 1981). However, findings to support

the main effects of the school-based preschool participation program elements on IQ and

cognitive functioning were more inconclusive than findings from the home-intervention

programs. The evaluators in Evard's study were consistent in finding that

low socioeconomic participating preschool children when compared to like
counterparts, always fared better on IQ gains and academic achievement
scores. However, when these disadvantaged program participants were
compared to middle-class average students, in most cases, the treated
children's performance was not comparable. These results confirmed that
low socioeconomic students, even after intervention, were still in jeopardy.
(p. 43)

Because the home-based programs were generally more effective in terms of

advancing cognitive development for preschool children, an examination of the critical

components of home-based programs is essential. In particular, verbal performance

continues to be a major concern for programs serving low-income preschoolers in the

1990's. For example, a Head Start Application Grant (Alachua County School Board,

1993) stated that, "children from economically disadvantaged families display deficits in the

area of language development when they enter kindergarten, and these deficits are difficult

to remediate" (p. 2).

The home environment and the child-rearing atmosphere have recently been

analyzed in terms of child stress and parenting stress on how they relate to the performance

of low-income children as compared to their middle-class counterparts on verbal cognitive

tasks (speech and language and conceptual knowledge). Evard (1988) found that an

underlying theme in the assessment of early childhood intervention programs for low-

income children was stress in the child as a contributing factor to his/her lower cognitive








functioning. Stress developed as a combination of factors from the home environment and

from the school setting. Brown and Rosenbaum (1984) also contended that stress is

related to lower cognitive and intellectual functioning but believe that the context of the

environment, in particular the child-rearing home environment, is more influential than the

school environment in being a source of stress. It is known that the breakdown under

stress of areas related to speech and language can begin early in a child's life (e.g., Brown

& Rosenbaum, 1984; Humphrey & Humphrey, 1985; Murphy & Moriarty, 1976). For

example, Murphy and Moriarty (1976) studied indices of vulnerability which they defined

as "susceptibility to stress" in middle-class preschool children. The investigators wrote

that "poor speech was both a reflection of stress and a source of vulnerability" (p. 229).

Children in low-income families grow up under more stressful life conditions both
from the environment (e.g., noise, crowding, violence) and from stressful life events (e.g.,

birth and death of siblings and parents, loss of jobs by parents) than children from upper

income classes. Researchers have found that persons who are in low socioeconomic (SES)

groups experience more negative life events and have more problems with general

psychological well-being than those in higher SES groups (e.g., Dohrenwend &

Dohrenwend, 1974; Gersten, Langner, Eisenberg, & Simcha-Fagan, 1977; Makosky,

1982). Further, environmental conditions such as poverty (correlated with low income,

lack of education, unemployment, and single parenthood) have been shown to be stressful

and to have a detrimental impact on parental attitudes and behavior (e.g., Belle, 1982;

Conger, McCarty, Yang, Lahey, & Kropp, 1984; Hastings-Storer, 1991; Longfellow,

Zelkowitz, & Saunders, 1982).

Researchers have found that demographic variables such as parental educational
level, employment status, marital status, age, and ethnicity are related to the stress found in

adults and children. Parental educational level has been related to stress in parents (Belsky,

1984; Conger et al., 1984; Hastings-Storer, 1991) and in children (Gersten, Langner,

Eisenberg, & Simcha-Fagan, 1977; Werner & Smith, 1982). The general finding is the








lower the parental educational level, the more stress is present in both the parent and child.

Employment status and marital status have been connected with stress in parents (Conger et

al., 1984; Hastings-Storer, 1991; McLoyd, 1990; Thompson & Ensminger, 1989;

Weinraub & Wolf, 1983) and in children (Gersten et al., 1977; Shinn, 1978; Wemer &

Smith, 1982). In this case unemployment or single parent status is related to higher levels

of stress present in the parent and the child. Age of the parent has been found to be related

to levels of stress in adults (Beard, 1982; Samuelsson, 1982; Swanda & Kahn, 1986) and

children (Abidin, 1990; Belsky, Lemrner, & Spanier, 1984). As age of the mother

increases, less stress is present in the adult and her child. Finally, ethnicity has been found

to be associated with stress in parents (Hastings-Storer, 1991; Norris, 1992) and children

(Coddington, 1972a, 1972b; Hastings-Storer, 1991). The general trend is that whites

report more life events while minorities report more stress about relationships.

Numerous researchers have also found these variables to be related to the child's

cognitive functioning. Parental educational level and employment status have been related

to cognitive functioning in children in many studies. The two variables are also linked

together as well. Low educational level of parents and/or unemployment in parents has

been connected with lower cognitive functioning in the child (e.g., Laosa, 1982a, 1982b;

Mclntire, 1991; Wemrner & Smith, 1982; ZiUll & Peterson, 1982). Marital status has also

been connected to child cognitive functioning (Allison & Furstenberg, 1989; Prince, Kiely,

Boros, & Engelsmann, 1972; Shinn, 1978; Wemrner & Smith, 1982). Being raised by a

single parent and/or the absence of the father or father figure are related to lower cognitive

functioning in children. Young maternal age has been related to lower cognitive

functioning in children (Belmont, Cohen, Dryfoos, Stein, & Zayac, 1981; Broman, 1981).

Ethnicity as well as been studied in relation to children's cognitive functioning, with

findings that minority children score below white children on measures of cognitive

functioning (Broman, Nichols, & Kennedy, 1975; Deutsch & Brown, 1967; Schooler &

Anderson, 1979).








Brown and Rosenbaum (1984), who conducted an extensive review of the

literature, explain the plight of these children and their families:

Disadvantaged children suffer from inadequate health care, stressful home
conditions including large family size, and disordered, undisciplined
schools. Disadvantaged families are at-risk from domestic breakup,
maternal depression, poor health (especially from stress-related disease),
recurrent financial crises and disordered and uncontrollable life conditions.
(pp. 131-132)
A considerable body of literature indicates that when children are aroused,

distracted, and confused by stressful life events and conditions (e.g., noise, crowding,

marital disruption, etc.), their intellectual/cognitive functioning in the form of IQ scores and

school functioning in academics tends to be below average (e.g., Belle, 1982; Blom,

Cheney, & Snoddy, 1986; Brown & Rosenbaum, 1984; Mclntire, 1991; Prince et al.,

1972; Werner & Smith, 1982). As Blom et al. (1986) have written

it is difficult to quantify the impact of stress on children's academic learning
and school adjustment, but it is safe to say that for some children it is a
major impediment to achievement and that most children occasionally are
diverted by the effects of stress. (p. xi)

A number of researchers believe that the family is the appropriate unit of study in

understanding the impact of stress on a child's cognitive functioning such as IQ or school

achievement (e.g., Blom et al., 1986; Brown & Rosenbaum, 1984; Upshur, 1988).

Brown and Rosenbaum (1984) believe "the key to the overall improvement of intellectual

competence lies in dealing with family functioning as part of the general reduction of

stressors in the child's environment" (p. 138). In low socioeconomic families, the parent's

ability to cope well with the stress of child-rearing may be the most effective buffer against

a child's life events stress impacting negatively on the child's cognitive coping (e.g.,

Rutter, 1981; Rutter, 1983; Sameroff & Seifer, 1983; Werner & Smith, 1982; Work,

Cowen, Parker, & Wyman, 1990; Wyman, Cowen, Work, & Parker 1991). For example,

according to Humphrey and Humphrey (1985)

there is an increasing amount of evidence that shows that children
supervised by adults who themselves do not cope well with stress, will pass
along to children this same inability to cope. In addition, it has been found








that when these same adults improve upon their ability to control stress, this
skill is also passed along to children. (p. vii)

By serving as a positive or negative influence or mediator, the level of parenting

stress partly determines the relationship between the stress the child is experiencing and the

child's cognitive functioning. Many researchers see the child's response to stress as falling

on a continuum ranging from vulnerability to resiliency (e.g., Brown & Rosenbaum, 1984;

Humphrey & Humphrey, 1985; Garmezy, 1983; Johnson, 1986; Murphy & Moriarty,

1976; Rutter, 1983; Sameroff & Seifer, 1983; Wemrner & Smith, 1982). The attitudes,

psychological characteristics, and behaviors of the parent in response to childrearing are the

channels through which the child is enabled (resiliency) or debilitated (vulnerability) when

dealing with life events stress which he/she encounters.

Stress and Child Cognitive Functioning
Brown and Rosenbaum (1984) have developed a theory and model to test the
hypothesis that stress is related to cognitive and intellectual functioning in children. This

occurs through a process of brain regression transmitted to the child through parental

modeling of effective coping with stress. The part of the brain that controls emotion

(limbic) is more primary than that which controls cognition (neocortex). When people live

under stressful life conditions, they more often rely on emotionally-based coping than on

cognitively-based coping. This follows an initial assessment which uses cognitive

processes but which, under stress, moves increasingly to the control of the emotional

system (or more primary limbic brain).

Under stress, verbal cognitive problem solving skills (linked to speech, language,
and conceptual knowledge) may become depressed or may not develop fully because the

individual is emphasizing emotional coping. Emotional coping (e.g., aggression, denial,

withdrawal), while sometimes necessary and effective for daily living under stressful

conditions, is not as conducive to success in the formal educational system as is cognitive

coping based on the use of symbols (e.g., use of selective attention and ability to access

memory, knowledge base, metamemory, and skill in manipulating language). It should be








noted that verbal cognitive functioning may be distinguished from nonverbal cognitive

functioning. Verbal cognitive functioning includes speech and language use, and

understanding of concepts linked to language. These abilities consist of perception,

memory, and previous learning association. By contrast, nonverbal cognitive functioning

includes such tasks as those related to gross and fine motor skills and spatial skills.

Brown and Rosenbaum (1984) further explained how stressors affect intelligence in

children within the context of the family under stress:

Developmental insults affecting the growth and level of intelligence appear
to be the result of three epidemiological vectors that translate stress into
dysfunction. These are: (1) high stressor levels, especially those associated
with poverty and illness, which decrease ability to deal with daily stressors
and complex tasks; (2) inadequate family function, which increases the
incidence of stressful events and vulnerability to stress and; (3) poor coping
skills, which result in reduced self-management, social awareness, and
cortical control over emotion. (p. 145)

Brown and Rosenbaum believed that the negative influence of stress on cognitive

functioning develops early in life and is linked closely with the home environment. They

begin their emphasis on schooling with early intervention programs and progress to follow-

up studies of children in elementary and high school. Most of the studies in support of

their hypothesis involve older children. A few studies are available with younger children

examining relationships between stress, parenting, and child cognitive functioning but

these studies are not as clearly designed to test the Brown and Rosenbaum hypothesis as

those with older children.

Numerous researchers have shown children's cognitive functioning (both verbal

and nonverbal) to be impaired under stress. Brown and Rosenbaum (1984) examined the

relationship between specific stressful child life events and IQ in low-income elementary

students and obtained an inverted U distribution between the number of stressful life events

experienced by the students and their IQ performance. Higher peaks were found in

children from low socioeconomic groups, suggesting higher emotional arousal. They also

found that high school students who had been in early education preschool programs still








showed effects of stress on IQ in the form of lower cognitive functioning than students

who had been under less stress.

Other researchers have found the number of stressful life events are correlated with

maladjustment in school as demonstrated by acting out, being moody-withdrawn, having a

learning problem, and by a combination of all these elements during an entire year in a

sample of inner-city elementary children (Sandier & Block, 1979). In addition, in a

longitudinal study including preschoolers experiencing normal daily stress, Murphy and

Moriarty (1976) found that some of the children under stress reacted with disturbances in

perceptual and cognitive functioning. They found that "children whose IQs varied were

often sensitive and more aroused" (p. 344). The expressive speech of these preschool

children had become impaired. This is a clear example of the relationship between stress

and language.

Parenting Stress and Child Cognitive Functioning
The child's stress is not the only factor relevant when considering the relationship

between stress and cognitive functioning in the child. Brown and Rosenbaum and others

(e.g., Wemrner & Smith, 1982; Wyman, Cowen, Work, & Parker, 1990) have hypothesized

that the relationship between the stress the child experiences and the child's cognitive

functioning is mediated by the coping ability of his/her parents with regard to parenting

(childrearing) stress. The parent's coping ability, in turn, buffers the child from an amount

of and response to acute and chronic stress so great that the child's cognitive functioning

(necessary for coping) is impaired. The degree to which parents can effectively model and

help a child adjust and cope better with the child's life stress may be related to the parents'

ability to manage their own stress in parenting (Humphrey & Humphrey, 1985; Murphy &

Moriarty, 1976; Wemer & Smith, 1982).

Parenting stress is composed of cognitive, affective, and behavioral components
(Mash & Johnston, 1990). Parenting stress can be related to specific characteristics of the

child, such as the presence of disorder or the temperament of the child. Parenting stress is








also linked to certain characteristics of the parent (Abidin, 1985, 1990). For example,

Upshur (1988) stated that parents who suffer from health problems, poor self-concept,

feelings of being out-of-control of their lives, or who live under an array of stressors "are

not as likely to be able to provide optimal conditions for child development" (p. 135).

Reciprocity operates, and both parent and child factors may substantially contribute to the

stress in the parent-child relationship. In addition, there are substantial contextual sources

of stress operating in determining the total amount of parenting stress (Abidin, 1985, 1990;

Belsky, 1984).

Several studies have found parenting stress to be related to negative perceptions of

child-rearing and negative parent-child interactions (e.g., Bugental, Blue, & Cruzcosa

1989; Bugental & Shennum, 1984; Hastings-Storer, 1991; Patterson, 1983). Hastings-

Storer found that low-income mothers had more overall parenting stress than middle-class

mothers. Those areas related to attachment were especially stress-ridden. High levels of

parenting stress have been associated with higher levels of learning problems in elementary

school children (Belle, 1982), decreased language ability and achievement in children

(Murphy & Moriarty, 1976; Werner & Smith, 1982), and low child intellectual functioning

(Mclntire, 1991; Prince et al., 1972). For example, Mclntire (1991) in a study of parenting

stress and cognitive outcomes in early elementary school-age boys, found that parenting

stress was significantly and negatively related to the IQ of the child. He concluded

it should also be noted that just as lower intelligence in children may
influence stress levels in mothers, higher levels of chronic stress in mothers
might have some impact on the cognitive development of children, as
suggested by the results of the multiple regression on children's IQ scores.
... Chronic parental stress is likely to have greater effects on child
development, possibly even cognitive development. (p. 89)

Among low-income families, children with more attentive, nurturant, and
intellectually stimulating parents have had larger intellectual gains (e.g., Prince et al., 1972;

Wemrner & Smith, 1982; Wyman et al., 1991). Findings reported in a series of studies

which addressed resilience in families of low-income inner-city elementary school children

support these differences (Parker, Cowen, Work, & Wyman, 1990; Work, Cowen,








Parker, & Wyman, 1990; Wyman, Cowen, Work, & Parker, 1991). High quality parent-

child relationships when the child was an infant and a toddler characterized by attachment

and nurturance seemed to help children to be "stress resistant" to the negative life

conditions associated with poverty. These children's school functioning during the

elementary school years was less impaired and the children had developed personality

characteristics such as an internal locus of control which seemed to help them cope.

Although researchers contend that the effects of stressful events begin early in a

child's life and are related to the quality of the parent-child system, only a few studies are

available that have explored the relationship between these types of stress and cognitive and

intellectual functioning in preschool children. Prince, Kiely, Boros, and Engelsmann

(1972) examined preschool children who were under severe stress, measured aspects of the

parent-child relationship, and related both to child cognitive intellectual functioning.

Although the study suffered from unreliable instrumentation, the results support Brown

and Rosenbaum's theory that parental stress relates to child stress which in turn relates to

the child's verbal cognitive functioning. Prince et al. (1972), whose study was an

evaluation of Family Life Education Services in Canada, found

the mean WPPSI for the high income children to be 17 points above that of
the lower income children; the mothers' stress symptoms are twice as high
among the low-income groups and their attitudes are much more
authoritarian. (p. 13) ...In the low-income sample there is some evidence
that the highly distressed mother may be related to impaired cognitive
abilities of her child; this is evident in the verbal IQ. (p. 26) [Further, with
regard to authoritarianism] ...The relationships are stronger in the low-
income sample and are to be found both in the verbal and performance
aspects of IQ. (p. 31)

Murphy and Moriarty (1976) conducted further research on daily stress in middle-

class preschoolers. They found that an accumulation of stress in the child's life brought

changes in children's usual adaptation level and style of functioning. Depression,

exhaustion, or illness in mothers who had earlier been supportive in child-rearing were

"followed by drastic changes in the child's style and level of adaptation and even in IQ








level" (p. 187). They concluded that the stress these children were under was not severe,

yet the children were impacted enough for measurable change to take place.

By the same token, families can be resilient to severe stress. Werner and Smith

(1982), in a study assessing longitudinal child outcomes for Hawaiian children living in

poverty, compared children who were able to cope with those who were not so able. The

quality of the mother-child relationship when the child was a toddler was critical in relation

to the child's ability to cope throughout life and strongly related to IQ, cognitive

functioning, and academic performance in middle childhood and adolescence.

Not all low-income children will show low intellectual/cognitive functioning. Many

families will be resilient to a greater degree than others. Unfortunately, more low-income

children and their parents are susceptible to vulnerability as these families live under more

stressful environmental conditions and suffer more negative life events than members of

middle-class families do (Dohrenwend & Dohrenwend, 1974; Longfellow & Belle, 1984).

Therefore, it is critically important to identify factors that are related to resilience in those

children who cognitively cope well while under severe environmental and life events stress.

Because language is one of the most important cognitive skills a child needs to use

to manage his environment and feelings, the relationship of stress to language and verbal

cognitive functioning needs attention. One of the most powerful influences on language

development is the quality of the mother-child relationship. Several early researchers have

done extensive work on the relationship between verbal development and the home

environment.

Early work by Bernstein (1961) on social class (mainly lower and middle strata)

and linguistic development supports the theory that the SES status of the family influences

verbal functioning in such a way as to accentuate emotional coping and minimize cognitive

coping. This is along the lines of Brown and Rosenbaum's theory. Bernstein's theory

revolves around the concept that "two distinct forms of language use [formal and

public/restricted] arise because the organization of these two strata is such that different








emphases are placed on language potential" (p. 291). He suggested that middle class

speech is a language in which "the formal possibilities of sentence organization are used to

clarify meaning and make it explicit" (p. 291). The speech of lower class individuals is

termed a public language where the speaker "operates within a speech mode in which

individual selection and permutation are severely restricted" (p. 291). Bernstein proposed

that in middle class families as compared to lower class families, "the early linguistic

relationship between mother and child is essentially one which maximizes cognitive and

affective differentiation, rather than affective inclusiveness and identity" (p. 294).

Early Childhood Programs. Stress. and Child Functioning

Once the child begins school, the home situation Bernstein outlined involving

restricted language can be exacerbated. Bernstein (1961) posited that within a formal

learning situation, the public or restricted code of communication is often not recognized by

"authority figures." This leads to a situation in which

the attempt to substitute a different use of language and to change the order
of communication creates critical problems for the lower working-class
child, as it is an attempt to change his basic system of perception,
fundamentally the very means by which he has been socialized. (pp. 304-
305)

Stress in the home environment can thus be carried over to become stress in the

school setting. Even though much work has been done to accept dialects and languages in

an attempt to reduce stress associated with producing language unfamiliar to the student,

Evard (1988) found, after an extensive review, that stress was reported as a debilitating

side effect from early education programs. Further, the effect of life events on children has

not been explored to the extent it has with adults. This is especially true with preschool

children, many of whom participate in early childhood compensatory education programs.

These programs are considered particularly important for low-income children.

Blom, Cheney, and Snoddy (1986) reported that "children are subject to stress-

inducing environmental events and situations just as adults are. Parents and educators have

been slow in recognizing this, at least to the extent required to develop the knowledge and








skills necessary for effective intervention" (p. ix). Teachers often find it difficult to deal

with the stress which comes from the home into the school setting and may be making the

school situation more stressful as a result. In a survey study for example, elementary

school teachers were found to be aware that children in their classrooms experience acute

and chronic stress situations at home and in school (Zucker & Snoddy, 1980-cited in Blom

et al., 1986). These teachers reported that they listened to the students and felt the students

saw them as a potential source of support. However, the teachers stated that they are not

provided by the educational system with curriculum, materials, facts, or skills which could

be used to provide beneficial interventions. Yet teachers are in a excellent position to work

with children to reduce stress since they are often very significant adults in the child's life;

are with the child for extended periods of time; and can observe, interact with, and impact

on the child.

The following conclusions can be drawn:

1. A number of empirical studies support the conclusion that life events and

conditions relate to child stress and parenting stress, which in turn are related to the child's

verbal cognitive performance. The majority of these studies have been conducted with

school age students. Little is known about the relationships between child stress, parenting

stress, and the child's verbal cognitive functioning during the preschool years, especially in

low-income population children who attend early childhood compensatory education

programs.

2. Little research has been done on the relationships between child stress,

parenting stress, and the child's verbal cognitive functioning at the time of the child's entry

into early childhood compensatory education programs for low-income families. These

programs recognize that language acquisition is one of the most significant developmental

tasks during early childhood and that verbal cognitive functioning is valuable in problem

solving. Information about the stress present in the child's life, both from events which

have occurred with him/her and stress present in the parent-child relationship, could be








added to background knowledge about the family. If a significant relationship is found

between stress and the child's cognitive functioning it might raise questions about how

large a role stress management could play in early childhood compensatory education and

in the family setting.

3. Attention to demographic variables is a part of almost every study in the area

of stress and functioning. In almost all cases, middle class parents and children cope more

effectively with stress with regard to the child's verbal cognitive functioning than do lower

class parents and children. The demographic variables which along with income categorize

social class (e.g., educational level, employment status/occupation, and family

composition, e.g., number or parents in the home, maternal age, and ethnicity) are seen

both as relating to a higher or lower incidence of stressful life events (acute stress) and as

determining stress levels from the environment (chronic stress) for these families. Many of

these findings come from studies conducted with school age children. How these

demographics are related to resiliency or vulnerability in preschoolers' verbal cognitive

functioning (especially for low income children) is less clear.

4. The studies that have focused on preschoolers' stress, parenting stress, and

the child's verbal cognitive functioning suffer from such measurement problems as a lack

of attention to life events specific to preschool children, unreliability of instrumentation for

stress, no formal measurement of acute life events stress, parenting stress measures which

do not include characteristics of the child, and a lack of extreme poverty children as

subjects. These methodological problems necessarily limit generalizability, but not all

studies have all these problems. Overall, the results of the studies support the existence of

these relationships and thus continued research along these lines (with replication mindful

of the methodological problems noted) is warranted.

Purpose of the Study
Numerous K-12 studies have explored relationships between child stress, parenting

stress, and child verbal cognitive functioning once children have entered elementary school.








However, it is well established that a significant amount of a child's language development

occurs between the time the child begins combining words and age five (Eisenson, 1977).

Further, a child's language development is strongly influenced by the quality of the parent-

child relationship (Bernstein, 1961; Hess & Shipman, 1965; Wulbert, Inglis, Kriegsmann,

& Mills, 1975). The argument that low-income individuals experience more undesirable

life events (negative stressors) has a long history of confirmation in the literature

(Dohrenwend & Dohrenwend, 1974; Pearlin, Liebermann, Menaghen, & Mullin, 1981;

Pearlin & Schooler, 1978). It has also been noted that the preschool years are the most

stressful for parents (Abidin, 1985). Finally, preschool children's reactions to life events

are influenced by their particular stage of cognitive and emotional development (e.g.,

preoperational). They are especially susceptible to impaired functioning when these life

events are related to their primary caretakers. For most child, the primary caretaker is the

mother (Barton & Zeanah, 1990).

Of the studies found which did examine child stress in children before they entered

elementary school, only two studies focused on preschool children and their parents under

substantial stress which assessed the preschool years currently (Bee, Barnard, Eyres,

Gray, Hammond, Spietz, Snyder, & Clark, 1982 and Prince et al., 1972). Of these two

which studied preschoolers currently, the first study suffered most severely from

unreliability of instrumentation for stress, and the second study suffered from a weak

(global and general) measure of stress as well as from the exclusion of poverty level

families. Other studies lacked certain criteria with regard to the variables of child stress,

parenting stress, and verbal cognitive functioning in low-income preschool children.

Perhaps the most common of these limitations concerned collecting stress data

retrospectively to the preschool years. As a result, the child's cognitive functioning as a

preschooler is unknown. The link between stress in the preschool years and cognitive

functioning is extrapolated downward from the cognitive functioning of the child during the








school years when the study was being conducted to his/her cognitive functioning as a

preschooler.

Despite the limitations of individual studies, the literature as a whole suggests that

the relationship between child stress, parenting stress, and the child's verbal cognitive

functioning is active during the preschool years in low-income groups. Several stressful

life events measures for children are now available. In the last few years increased

attention has been paid to developing a more consensual definition of parenting stress. For

example, a 1990 issue of The Journal of Clinical Child Psychology ran a special issue on
"parenting stress". In addition, more reliable and valid measures of parenting stress have

been developed and normed. Children in such compensatory early education programs as

Head Start have had their verbal cognitive performance assessed as a pretest/posttest in the

programs, making this type of data more accessible.

The purpose of this study will be to examine the relationships between (a) child

stress (stressful child life events), (b) parenting stress (the main caretaker's ability to cope

with stress in the parent-child relationship), and (c) the child's verbal cognitive functioning

(speech and language and concept functioning) in late preschool age children in an early

childhood compensatory education Head Start program for low-income families.

This study extended the research done with low-income elementary and high school

students to preschool age students to test Brown and Rosenbaum's theory that

environmental stress (i.e., acute and chronic) and parenting stress are related to verbal

cognitive competence in the preschool age group and in the low socioeconomic group, with

the parent-child relationship mediating the impact of stress on the child's verbal cognitive

functioning. This study was designed to reduce limitations found in previous research. To

summarize, these limitations are

1. A lack of attention to life events specific to preschool children;

2. Unreliability of instrumentation for stress;

3. No formal assessment of acute life events stress;








4. Parenting stress measures which do not assess child characteristics;

5. Retrospective collection of data during the preschool years, also making it

impossible to have information on cognitive functioning at that developmental period; and

6. No inclusion of extreme low-income families.

Besides examining the relationships between child stress, parenting stress, and the

child's verbal cognitive functioning, the analysis included demographic variables which

have been shown in past studies to most strongly and consistently influence stress and

cognitive functioning in low-income families: namely, parental education, employment

status, marital status, age, and ethnicity. At times, these variables have also been seen as

sources of stress.

Significance of the Study

There is considerable support in the literature for conducting a study such as the one

proposed here. Johnson (1986) wrote that most attention had been paid to the life events of

adolescents. In addition, mediator variables (e.g., the parent-child relationship; the

temperament of the child) have been given inadequate attention. Johnson wrote:

Especially neglected in this regard are younger children, below the age of 10
or so. Not only has less attention been given to the assessment of life
changes in younger children, but there have been fewer studies that have
focused on the correlates of life events and the role of variables that may
moderate the impact of life changes in this age group. (p. 128)

Webster-Stratton (1990) described the situation in the following way:

The task for future research in this area is to continue to conceptualize the
complex and dynamic relationships between stressors and the family
interaction system, as well as to identify those factors that can serve to
increase or decrease a maladaptive outcome for the parents and the child.
For the ultimate challenge is to recognize those families most at risk, those
most vulnerable to disruption by life stressors, and to help them develop
resources and coping skills that will minimize the disruption.
(p. 310)
How children respond is likely to be tied to the relationship they have with their

family. Compas (1987) contended that "the relation between various social contexts or

ecologies and the coping behavior of children... needs to be examined in greater detail.

Foremost among these is the role of the family" (p. 401).








Measuring cognitive functioning in general as an indicator of coping with stress has

considerable support in the literature. For example, Werner and Smith (1982) with respect

to cognitively based coping argued that "we need to identify more systematically the

positive effect of these variables in contributing to 'resiliency' and 'invulnerability'"

(p. 160). Evard stated that the element of stress in relationship to cognitive functioning

"does warrant the need for further investigation" (p. 41). Murphy and Moriarty wrote that
"zones of disintegrative responses to stress (cognitive, as in IQ changes, motor, as in the

loss of control, including speech problems) call for study" (p. 350).

As previously discussed, language is identified as critical for successful school

functioning by governmentally funded early childhood education programs. Projects such

as Head Start place a heavy focus on improving language and conceptual functioning in

preschoolers before they begin the formal educational process. Some researchers in Head

Start recognize the role stress plays, reflected in the writing of Slaughter (1980):

In short, the very context, attitudes, and behaviors attributed to Head Start
children and their poverty-stricken families fifteen years ago, have become
increasingly widespread in the nation as a whole, particularly its urban
areas. The social problems Head Start was created to partially resolve now,
in varying degrees, afflict the nation. Head Start is the nation's only
existing laboratory with a potential for dealing with these problems if, and
only if, it can reveal what it has learned, and is learning, about how
embattled children and families cope with environmental stress, and how the
families of these children, despite this stress, can often serve as incubators
for positive early cognitive and social development The future of Head
Start, I believe, lies in how well it can inform and serve the nation in this
regard. (p. 15)
This study may help preschool educators understand several critical issues.

1. The role of the parent in the child's language and conceptual development,

and the susceptibility of language and conceptual development to stress. Language is

known to be critical for learning. Eisenson (1977) wrote that the ability of parents to use

the parent-child relationship to convey to children how talking can help them structure their
world and gain information sets the foundation for a child's willingness to use language, or

the development of anxiety about communication. The parent also plays a crucial role in








establishing concepts and conceptual thinking in children by the number and quality of

experiences provided to the child to learn concepts.

When children have experienced numerous stressful life events, their verbal

cognitive functioning is often below average. It is well established that low-income groups

experience more undesirable life events. This study will extend findings from studies

examining child stress and cognitive functioning in elementary school children downward

to the preschool period of childhood, with a focus on low-income families.

2. The role of the parent-child relationship as a mediator of child stress on verbal

cognitive functioning. Most child stress researchers have found the parent-child

relationship to be a powerful mediator of life stress for the child (especially the young child

with attachment needs). Children who have parents less stressed by such variables as social

isolation and lack of marital support exhibit higher verbal cognitive functioning than

children of parents experiencing large amounts of stress. This follows the thinking that

when the parent is handling the different elements of being a person and a parent, the child

may benefit from a lessened number of stressful life events present and the positive

modeling provided by a coping parent. This study will hopefully shed light on these

mediated relationships of preschool age low-income children and their primary caretaker,

the mother.

3. The importance of early intervention with regard to both parenting and child

stress and the possible role of the school, especially the early education teacher, in this

intervention. Early childhood compensatory education programs such as Head Start have

always had a strong parental involvement component. This has most often been directed at

helping the parent teach the child in the hope of improving speech, language, and

conceptual knowledge in the child; these are major areas of concern for low-income

children. Such programs have not often attempted to relieve the stress of the parent and/or

child to increase verbal cognitive functioning in the areas of speech, language, and

conceptual knowledge in the child through the parent-child relationship. The early








childhood teacher can play a pivotal role in alleviating stress reactions in both the child and

the parent This study may be used to provide teachers with information about these issues

to help them be more aware of the stress the child brings to school. Equally, if not more

important, programs must work with the family to alleviate some of this stress from arising

in the first place (Blom et al., 1984) due to its high cost to the child, family, and society

(Abidin, 1990).

Limitations of the Study

According to a number of researchers (e.g., Blom et al., 1986; Brown &

Rosenbaum, 1984), the relationship between stress and functioning begins early in the

child's life. One limitation, however, of a study which focuses on preschool children, was

related to the difficulty of studying stress in young children. As Chandler (1984) pointed

out, there were two main problems. First, a very young child cannot say that an event has

been stressful. Therefore, the parent self-reported stress events for the child. Also, there

was less likelihood of showing a relationship between stress and an outcome variable

because life events are necessarily cumulative. Older children would therefore have a

higher probability of having experienced more life events than preschool children.

A second limitation was with the instrument chosen to measure child life events

stress, the Coddington Life Events Record. This instrument was normed in the early

1970s. The author commented that gathering data every few years on the extent of social

readjustment required by children would give a measure of changes occurring in our

society. This has not been done with this scale. However, this was the only instrument

available which was specifically designed to measure life events experienced by preschool

children. Reliability information was lacking on the preschool age group. To address this,

a reliability study was conducted and is discussed in Chapter III.

A third limitation of the study involved the measurement of language development
in general as an outcome measure. According to Seymour and Wyatt (1992), speech and

language in children this age group have not been completely mastered and are in a state of








flux. There was considerable variability in the developmental rate among individual

youngsters during the language acquisition period. However, developmental milestones

were much more stable. Using milestones allowed researchers a predictable alternative in

the form of classifying children by language level rather than rate.

A final limitation involved the use of a local sample (one county) for the study. For

instance, these mothers and children were likely to be on the very low end of the income

scale as they resided in the fourth poorest county in the United States. As poverty is a

strong criteria of risk, perhaps the results from this sample will be meaningful to this issue.

Definition of Terms
Verbal cognitive functioning. Verbal cognitive functioning referred to performance

on language and conceptual knowledge tasks. The Developmental Indicators for the

Assessment of Learning-Revised (DIAL-R) is a widely used screening measure for

preschool children, particularly those in government sponsored early childhood programs

(e.g., Head Start). This instrument assessed the abilities of perception, memory, previous

learning association, kinesthetic awareness and coordination, and language in young

children.

Two subscales, Concepts and Language, had 8 items each, a total of 16 items.

Each item within Concepts and Language further represented a subscale. All Language

subscales were related to memory and previous learning association. Seven of the eight

Concepts subscales were related to language, memory, and previous learning association.

The only subscale not related to language in the Concepts scale was sorting. This subscale

was related to perception, memory, and previous learning association. This instrument

was designed to measure competencies that nursery-school, kindergarten, and first-grade

teachers indicated were needed for success in the regular classroom. The operational

definition of verbal cognitive functioning for this study was the score obtained by a child

on the DIAL-R Language and Concepts developed by Mardell-Czudnowski and

Goldenberg (1990).








Child life events stress. Child life events stress was defined as an identifiable

external life event which occurred in relation to a single event, at times of crisis proportion

(Blom et al., 1986). Child life events stress measures have been developed to parallel the

work on life events stress of adults which found strong relationships between life events

and functioning. The difference in applying this procedure to children was that the child

life events measures were completed by parents when young children are involved.

Child life events instruments quantify the number of life events, the amount of

change required on the part of the child in response to the event (known as social

readjustment), or a group of events qualitatively different from one another. The

Coddington Life Event Record was the most widely known arid used life events measure

for children. For the purpose of this study, child life events stress was operationally

defined as the the number of life events factors representing the areas of Loss, Entrance,

Family Troubles, Positive, Physical Harm, and Primary Environment Change occurring at

any time in a child's life as measured by the Life Event Record--Preschool (LER-P)

developed by Coddington (1972a). The factor analysis from which the areas of life events

were taken was conducted by Sandier and Ramsey (1980).

Parenting stress. Parenting stress was comprised of affective, cognitive, and

behavioral components of the parent that related to the demands of child-rearing in an

interactive and reciprocal manner with the child (Mash & Johnston, 1990). The parent-

child relationship can be thought of as a mediator of stress for the child (Johnson, 1986;

Mash & Johnston, 1990; Webster-Stratton, 1990). When a great deal of stress was present

in the relationship, the level of dysfunction in the individuals) was greater.

As measured by the Parenting Stress Index (PSI) developed by Abidin (1985,
1990), parenting stress was the stress present in the parenting system defined as a function

of certain child characteristics, parent characteristics, and situational variables directly

related to the role of being a parent. High stress was consistently associated with

dysfunctional parenting (Abidin, 1985). The PSI has been used in more than one hundred








studies and is the most visible measure of parenting stress in the literature. Operationally

defined for this study, parenting stress was the thirteen factor scores obtained on the

Parenting Stress Index developed by Abidin (1985, 1990).

Demographic variables. Five current demographic variables were examined in this

study. Maternal educational level was determined by the number of years schooling the

mother had completed. Employment status was determined by whether the mother was

employed or unemployed. Marital status was measured as married or unmarried. Maternal

age was measured continuously. Ethnicity was measured as minority or white.

Research Ouestions
The following questions were posed in examining these relationships:

1. Is there a relationship between child verbal cognitive functioning and child life

events stress, parent-child relationship stress, maternal education level, maternal marital

status, maternal employment status, maternal ethnicity, and maternal age?

2. Is there a relationship between child verbal cognitive functioning, child life

events stress and parent-child relationship stress, after controlling for the maternal

demographic variables of education level, marital status, employment status, ethnicity, and

age?

3. Is there a relationship between child verbal cognitive functioning and the

maternal demographic variables of education level, marital status, employment status,

ethnicity, and age, after controlling for child life events stress and parent-child relationship

stress?

4. Is there a relationship between child life events stress and parent-child

relationship stress, after controlling for the maternal demographic variables of education

level, marital status, employment status, ethnicity, and age?

5. Is there a relationship between child life events stress and the maternal

demographic variables of education level, marital status, employment status, ethnicity, and

age, after controlling for parent-child relationship stress?








6. Is there a relationship between parent-child relationship stress and the

maternal demographic variables of education level, marital status, employment status,

ethnicity, and age, after controlling for child life events stress?

Hypotheses

The following hypotheses were tested to answer the research questions posed:

1. There will be a significant relationship between child verbal cognitive

functioning, child life events stress, parent-child relationship stress, maternal education

level, marital status, employment status, ethnicity, and age.

2. There will be a significant relationship between child verbal cognitive

functioning, child life events stress, and parent-child relationship stress, after controlling

for the maternal demographic variables of education level, marital status, employment

status, ethnicity, and age.

3. There will be a significant relationship between child verbal cognitive

functioning and the demographic maternal variables of education level, marital status,

employment status, ethnicity, and age, after controlling for child life events stress and

parent-child relationship stress.

4. There will be a significant relationship between child life events stress and

parent-child relationship stress, after controlling for the maternal demographic variables of

education level, marital status, employment status, ethnicity, and age.

5. There will be a significant relationship between child life events stress and the

maternal demographic variables of education level, marital status, employment status,

ethnicity, and age, after controlling for parent-child relationship stress.

6. There will be a significant relationship between parent-child relationship stress

and the maternal demographic variables of education level, marital status, employment

status, ethnicity, and age, after controlling for child life events stress.













CHAPTER 2
REVIEW OF THE LITERATURE

This literature review will be presented in six parts: (a) general background

literature on stress in childhood and adulthood; (b) direct literature on the relationships

between child life events stress, parenting stress, and child cognitive functioning as it is the

most directly related to the proposed study; (c) the literature on the relationship of child life

events stress and child cognitive functioning; (d) the literature on child life events stress and

parenting stress looking at the influence of these two variables on one another; (e) studies

on the relationship between parenting stress and child cognitive functioning; and (f)

relationships between five demographic variables related to stress in children and parents

and cognitive outcomes in children: namely, maternal educational level, marital status,

employment status, maternal age, and ethnicity.

Historical Background on Stress
Stress is a term originally used to describe breakdowns in human functioning.

Early stress research focused on physiological stress (e.g., Cannon, 1929; Selye, 1978).

Selye's definition of stress is "the nonspecific response of the body to any demand" (p. 1).

Cannon's work focused on the fight or flight response to stress. Recently, a shift has

taken place emphasizing psychological stress arising from the environment. For example,

according to Chandler (1985) "stress may be defined as a state of emotional tension arising

from the failure of the environment to meet the individual's needs, and/or from events and

situations that are seen as threatening" (p. 6). This view has made stress more generalizable

across outcomes, allowing it to take more than just a physical form.

Blom, Cheney, and Snoddy (1986) defined these events and situations in the

following way: as stressors take the form of an environmental geventL "in such cases the








term 'acute' is often associated with them" or for situations or conditions, "the term

'chronic' is often used to characterize this type of stress" (p. 26). This distinction between

events and conditions clarified the context in which the events took place. For example,

research has shown environmental socioeconomic factors are related to the number of

negative life events which occur in an individual's life and the ability to cope effectively

with them when they do arise (Dohrenwend & Dohrenwend, 1974; Pearlin & Schooler,

1978).

Of acute life events, children share some events with adults but experience others

unique to childhood. The area of life events has become the focus of much child stress

research. Additionally, there is the stress of the normal physical, cognitive, and emotional

developmental milestones of childhood (e.g., Chandler, 1984; Garmezy, 1983). Different

events heavily tax availability of the child's resources for dealing with both acute life events

stress and chronic environmental stress as a function of the child's current developmental

level. Barton and Zeanah (1990) describe the events most stressful to young children:

During the period between 1 and 4 years, stressors that threaten the primary
attachment relationship are likely to be experienced as particularly difficult.
... Control over the immediate environment is critical at many
developmental stages, but is particularly important to preschoolers, who
have struggled only recently to gain control over their behavior and who
continue to struggle toward a wider range of competencies. The threat
posed by many events experienced as stressful by the preschool child
centers around a loss of control, including the inability of trusted adults to
control environmental events. (p. 196)

In adulthood, the role of parent becomes a major part of life for most adults,
bringing with it its own unique set of stressors. Many researchers believed having young

children was a significant source of stress in itself (e.g., Cappell & Mays, 1973; Gil &

Noble, 1967; Longfellow, Zelkowitz, & Saunders, 1982; Mash & Johnston, 1990;

Webster-Stratton, 1990). For example, Gil and Noble (1967) found in a large statistical

compilation, that in 59% of the cases of child abuse, 'mounting stress on perpetrator due to

life's circumstances' was indicated.








Women may be at increased risk for stress in parenting. On this subject

Longfellow et al. (1982) suggested that

the stress factor is compounded by the fact that the mother of young
children is herself usually young, often poor, and in a growing percentage
of cases, shouldering responsibility alone. (p. 163)

Recent Census data support this claim. In 1986, 51.4% of female-headed-

households were below the poverty line. In 1985,61.3% of all mother-only families were

awarded child support compared to 40% of poor women with children (Congress of the

United States, 1988). Combining the need of the preschooler for an adult who can control

the environment for him/her and the environment that many mothers are being asked to

control, is likely to increase both the amount of stress present and adverse responses to this

stress in both of them.

Researchers in the field of stress have several suggestions and guidelines for

studying psychological stress. One guideline is to determine if an event or a situation can

be considered a stressor. To qualify as such, three conditions must be met, according to

Blom, Cheney, and Snoddy (1986). The event or situation must be external, it must be

clearly identifiable, and it must yield a "psychological disequilibrium ... sufficient to result

in a behavioral reaction (response)" (p. 17). Once the stressor(s) is/are identified, stress

theorists (e.g., Compas, 1987; Lazarus, 1977) stated that it was essential to find principles

for predicting the stress reaction from the person-environment relationship. The

relationships which may be present between the child, the mothering one, the childrearing

environment, and the stress response manifesting in the child's verbal cognitive functioning

is the subject of this literature review.

Child Verbal Cognitive Functioning. Child Life Events Stress. and Parenting Stress

While most research on child cognitive functioning, child stress, and parenting

stress has focused on elementary age children and up, a few studies have concentrated on

or included children who are preschool age. In this section of the literature review, seven

studies will be reviewed. These studies included young children, stress the child and








mother were under, and child verbal cognitive functioning and/or school functioning. Of

the seven, six assessed life events as stressors and one assessed stress symptoms. Three

of the studies were conducted with preschool age children and two others assessed the

preschool years retrospectively. Two studies in the section focused exclusively on the

school years, but are included because they illustrate the relationships between the variables

well. Four of the studies were longitudinal, ranging from four to twenty years.

Prince, Kiely, Boros, and Engelsmann (1972) studied child IQ and cognitive

functioning, stress in children, and maternal stress. Subjects were 240 dyads each of

which consisted of a mother and her 4-year-old child and which were randomly selected to

represent equal numbers of high-and low-income groups. As part of an evaluation of

preschool programs in Canada, these mother-child dyads were randomly assigned to

groups. Treatment children attended traditional nursery school programs. A nontreatment

group did not attend school and received no treatment. The middle-class children attended

programs already in place and accessible to them in the community. Low-income children

did not have adequate programs available. As part of the study these children attended

programs designed in the manner of Head Start in the United States. Mothers were also

assigned to treatment or control matching their children's participation or nonparticipation.

The instruments were: the Wechsler Preschool and Primary Scale of Intelligence (WPPSI),

Goodenough Draw-A-Person-Test (used to measure concept development, not as a

projective technique), Child Stress Symptom Checklist, Langner Psycho-physiological

Stress Scale for mothers, and Hereford Attitude Survey for authoritarian versus permissive

attitudes for mothers.

Pretest results showed that on both the WPPSI (including almost all its subtests)

and the Draw-a-person test, the scores were significantly poorer for low-income children.

The authors found that a trend was present "consistent with the idea that verbal abilities are

relatively more developed in higher income families" (p. 13). With regard to the measures

of stress, significant differences were found between the low- and high-income groups for








both the mothers and the children. The group differences in the mothers' reported stress

symptoms, however, were larger. The authors reported their finding as a "common"

relationship; that is, the lower the income (or class level) the higher the symptoms. They

wrote that "this relationship has been well-documented and is one of the most consistent

findings using self-report stress symptom checklist techniques (Phillips, 1966; Roberts et

al., 1966; Starr, 1950)" (p. 24). No significant relationships were found between the

reported stress symptoms in the high-income group of children and IQ; however, several of

the correlations were significant in the low-income children. The authors stated that "this

might suggest that symptoms so intense as to be worthy of note to the low-income mothers

are also severe enough to interfere with the low-income child's intellectual functioning"

(p. 28).
When looking at the relationship between parenting stress and child cognitive

functioning, the authors found that

in the low-income sample there is some evidence that the highly distressed
mother may be related to impaired cognitive abilities of her child; this is
evident in the verbal IQ and the Draw-a-person test. Such a relationship is
not apparent in the high-income sample. (p. 26)

Also included in the Prince et al. (1972) study were measures of authoritarian

versus permissive childrearing beliefs and attitudes shown to be connected with many child

outcomes. The relationship between the mother's symptoms of stress and her authoritarian

attitudes toward child-rearing was stronger in the low-income group. For the relationship

between authoritarian attitudes and child cognitive functioning, again the correlations were

higher in the low-income group than in the high-income group. In addition, both verbal

and performance aspects of IQ were related to authoritarianism in the low-income group;

whereas in the high-income sample only the verbal performance was related to

authoritarianism. In the Prince et al. study, the child's verbal cognitive functioning was

more highly related to his/her mothers' stress symptoms and authoritarianism than to

his/her nonverbal cognitive functioning.








In the course of the study, the question of reliability of instrumentation arose with

both of the parent measures. The authors looked more closely at the Langer instrument

which was designed to measure stress using a recent review of the measure. While

procedures during administration were not problematic or troublesome, 10 of the 22 items

on the instrument itself were about past history, with 3 other items phrased to reflect

character traits. These questions were stable and unlikely to register change. Upon further

investigation by the authors, the Hereford scale for measuring authoritarianism versus

permissiveness, was revealed to have low test-retest reliability.

Another weakness was with the stress measures. The score was heavily influenced

by social class and this was not taken into account. The authors found a stronger

relationship between the frequency of mothers' symptoms and frequency of children's

symptoms in the high-income group. They stated that based on knowledge of low-income

children, these children would be expected to suffer from more symptoms related to stress

than high-income children. As an explanation, the authors theorized that low-income

mothers had so many other problems that they were not as "aware of or concerned by these

kinds of symptoms in their children" (p. 28) and these symptoms would have to be much

more severe than for the high-income mothers in order to become "noteworthy". While

this may be so, it is clear that this was not an anticipated finding. The measures were not

independent with respect to social class as they were thought to be.

A final weakness of this study was in not measuring the mother's perceptions of the

child's characteristics. These characteristics, such as temperament and how acceptable the

child was to the parent were very important in getting an accurate assessment of how much

stress the mother experienced as she dealt with her child in the context of childrearing.

One of the clearest patterns that emerged in the Prince et al. study concerned the

outcome measure. Overall, in both social classes children clearly exhibited higher

susceptibility to impaired verbal cognitive functioning than nonverbal cognitive functioning








under stress. This was consistent with the findings of other researchers (e.g., Murphy &

Moriarty, 1976; Wemrner & Smith, 1982).

Resiliency and vulnerability were investigated in a 20 year longitudinal study

conducted by Werner and Smith (1982) on children in Hawaii. From previous research, it

became clear that many of the variables on which well-functioning children in general

differed from poorly functioning children were difficult to change (e.g., poverty, low

educational level of parents, low self-efficacy/external locus of control). The researchers

turned their attention to the "resilient" child to provide understanding of coping in high risk

groups. The resilient child was the child who performed well under adverse environmental

conditions.
All live births (N=698) during the year 1955 in Kauai, Hawaii served as subjects.

The data collection points occurred during the children's toddler years and at age 10 in

reference to the years ages 2 through 10. In addition, data was collected in late

adolescence, after which the study ended. Since the study was longitudinal and because

there was no data collection during the preschool years, information about other periods of

childhood has been included here to illustrate trends. Measures included: (age 2) an

assessment of pre/perinatal complications, home visits, temperament, mother's interaction

with infant and toddler, stressful life events, Cattell Infant Intelligence Scale and the

Vineland Social Maturity Scale; (age 10) assessment of the early family environment,

grades, previous intelligence and achievement tests, information from current teacher on

achievement and intellectual problems observed in the classroom, and the results of two

group tests: the Bender Gestalt and the Primary Mental Abilities Test (sampling reasoning,

verbal, numerical, spatial, and perceptual-motor skills).

Before presenting results, Werner and Smith (1982) began by saying that:

Although most of the children and youth with serious and persistent
learning and behavior problems in this community were poor, it needs to be
kept in perspective that poverty alone was not a sufficient condition for the
development of maladjustment. (p. 32)








The specific stressful life events found at age two indicated that among the females,

those events that differentiated the resilient children from those who subsequently exhibited

difficulties were: parental mental health problems, parental conflict, illnesses, and

accidents. With regard to the mother-child interaction, mothers of high-risk resilient girls

were observed to interact in a more positive manner with them than did mothers of girls

considered high-risk who later developed problems. Resilient toddlers scored significantly

higher on both the Cattell and the Vineland.

During the 2- to 10-year period, the most often cited stressful life events among

these families included: mother's employment outside of the household on a long-term

basis, childhood illnesses or accidents requiring medical treatment, chronic illnesses,

handicaps, or death of a sibling, parental illnesses, the departure of a sibling from the

household, the absence of the father, and chronic family discord during the early and

middle childhood years. Many of these events represented threats to relationships with

close attachment figures. With regard to the quality of family life, the authors reported

most of the children who developed serious ...problems received little
emotional support in early and middle childhood. They had poorer
relationships with their parents, fewer opportunities for satisfactory
identification; and were exposed to more inconsistent methods of discipline
than the resilient children. (p. 69)

Significantly higher scores on the Primary Mental Abilities Test (PMA) were found

for both male and female resilient children than for those children who were manifesting

serious problems. These differences were greatest on the Verbal Comprehension (V) factor

and on the Reasoning (R) factor, a measure of problem-solving ability, of the PMA.

This study demonstrated many of the aspects of the parent-child relationship that

relate to life events and cognitive functioning in low-income children. As it is longitudinal,

it has advantages for generalizability that the cross-sectional design of the Prince et al.

(1972) study did not. The problem of reliability of instrumentation was not as pronounced.

However, like the Prince et al. study there were certain weaknesses with the operational

definition of parenting stress. Stress for the parents was a measure of family life events








more than the stress present in the parent-child relationship. During physical and

psychological exams, the parent-child relationship was observed and assessed. The

researchers observed the children in their homes on a limited basis. In addition, the

variables of parenting stress and child stress were not clearly separated, instead the same

events applied to both the adults and the children. Most researchers in child stress felt that

the life events that occurred with a child were qualitatively different than those for an adult

However, there is some overlap which helps offset this criticism.

The major weakness of the Wemrner and Smith study was that while the research did

encompass children from birth through adolescence, there is the lack of a separate data

collection point for the preschool years. The years between ages 3 and 5 were considered a

separate and distinct period of childhood from infancy/toddlerhood, childhood, and

adolescence, particularly with respect to cognitive functioning (e.g., Piaget). There was no

way to know the age of the child at the time of occurrence of events and no cognitive

outcomes were available for the children during the preschool years using this design.

Also, parents had to go back through seven years to answer questions in this study.

Therefore, the information gathered, particularly about the home environment (more than

life events), could be strongly susceptible to response bias in favor of the present

circumstances.

Bee, Barnard, Eyres, Gray, Hammond, Spietz, Snyder, and Clark (1982) studied

prediction of IQ and language skill with attention to parental input in a four year

longitudinal study of 193 mother-infant dyads. Socioeconomically, the group as a whole

had an education above high school, mothers were in their twenties, were mostly married,

mostly white, and with average family incomes. Instruments used were the Brazelton

Neonatal Behavioral Assessment Scale, Bayley Scales of Infant Development, Denver

Developmental Screening Test, McCarthy Scales of Children's Abilities, Sequenced

Inventory of Communication Development, Fluharty Speech and Language Screening Test,

Preschool Behavior Questionnaire, Stanford-Binet, maternal education, Schedule of Recent








Events, Neonatal Perception Inventory, and the Home Observation For Measurement of the

Environment (HOME) Inventory.

The results indicated that from the "family ecology cluster", (i.e., stress, social

support, developmental expectations, and mother's education) measured in infancy,

additional information about stress and social support improved the prediction of child IQ

and language functioning over that which could be made from just knowing the number of

years of education the mother had. The authors concluded, "thus, it is not 'social class'

alone that is of importance, but rather the perceived supportive or stressful quality of the

family interaction" (p. 1148).

However, developmental expectations, social support, and life change revealed

correlations which were consistently stronger in the low-education sample than the high-

education group. The explanation given was that a different interactional dynamic may

have been operating which was dependent on the education level of the family. The

authors suggested that mothers with less education may respond differently to high life

change levels or levels of social support which are low. "These mothers may be less able

to 'buffer' the child against vicissitudes in their own personal relationships" (p. 1152).

Bee et al. noted that several researchers have recently proposed that adults vary in a

predictable fashion in their ability to cope with high personal stress levels. In particular,

better personal support systems and more education in parents would help them "weather

the inevitable episodes of high life change (Dohrenwend & Dohrenwend, 1978;

McFarland, Norman, Streiner, Roy, & Scott, 1980)" (p. 1152). With respect to their own

study, Bee et al. (1982) reported that

our results are at least consistent with the hypothesis that one of the specific
features of low-education families that increases the risk of later poor
intellectual or linguistic performance for the infant is a lesser ability of the
parents to adapt to the changing demands for the child, and to create or use
helpful support systems for themselves. (p. 1152)

One limitation of the study was that the measure of parenting stress was somewhat

weak, because it was actually a measure of the quality of the home environment (HOME)

on the one hand, and recent stressful life events were not separate for parent and child on








the other. However, the parent-child interaction was assessed with this method. The

HOME was developed for use with Head Start, so the results may generalize more

effectively to populations of low-income preschoolers. In this study, however, the sample

was most representative of white middle-class families.

A retrospective data collection design was used by Work, Cowen, Parker, and
Wyman (1990) on inner-city low-income elementary school children aged 10 tol2 and their

families in the Rochester Child Resilience Project (RCRP). Parents were interviewed about

the child's stressful life events in infancy and the preschool years and the parent-child

relationship at those times as well as in the current school years. Parents completed the

Stressful Life Events Checklist for children (SLE-C) and several ratings of the child's

adjustment. Teachers both from the current year and the year before independently rated

the child's adjustment, comparing them with same-sex classmates on a global rating of their

standing in class (separately for overall class adjustment and for academic performance)

using the Teacher-Child Rating Scale with the following factors: Acting-Out, Shy-Anxious,

and Learning Problems (e.g., poor concentration, limited attention).

The criterion used to define Stress Affected (SA) and Stress Resistant (SR) groups
were that, for both groups, the parent must have checked at least four stressful life events

on the SLE-C. SRs must have placed in the top third on at least two of the three adjustment

ratings, i.e., by parent, former teacher, and current teacher; and no lower than the middle

third on the other adjustment rating. To be classified as SA, children must have placed in

the bottom third on at least two of the three adjustment ratings, and no higher than the

middle third on the other.

Identified were 40 SRs and 37 SAs. A low stress (LS) comparison group
consisting of 45 children who had experienced 0 or 1 Severe Life Events in childhood

(SLE-C) was matched on sex and grade and included in the study. Analyses revealed that

on the Teacher-Child Rating Scale; the SRs scored as best adjusted, SAs as least well

adjusted, with LSs in between. This supports a popular theory that when a child








experiences no stress or very little she/he does not get a chance to learn and practice coping

skills (e. g., Brown & Rosenbaum, 1984).

Identified SR and SA children completed 11 measures consisting of child self-rated

adjustment, empathy, self-esteem, perceived competence, locus of control, and perceived

support. Interviews with both the children and parents were conducted (Wyman, Cowen,

Work, & Parker, 1990). Parents were interviewed on core clusters of variables examined

within a developmental framework to reflect possible developmental shifts in protective

factors occurring over time. These were: temperament, caregiver efficacy, father

involvement, separation, childcare support, and parent-child relationship. Parents who

were effective in handling these areas were seen as coping better with the stresses of child-

rearing than parents who were not coping well.

SR children had significantly higher self-reported current perceived competence

scores than SAs on all subscales except Athletic Competence. Cognitive strategies showed

that SAs endorsed more ineffective problem-solving solutions than SRs. SRs had a

stronger endorsement of more effective solutions overall. The trend was for SRs to have

more positive and fewer negative coping strategies than SAs. SRs scored as significantly

better adjusted than SAs on CRS Rules, Social Skills, and School Interest, and

directionally higher on locus of control. SRs had significantly higher realistic control

scores than SAs. On affective factors, the SRs had higher empathy scores than SAs, but

did not differ on anxiety or depression. Perceived support failed to reach significance.

The family milieu variables showed that in infancy, SRs had easier temperaments,

more father involvement, fewer separations, and more childcare support. For the

preschool period, SRs again were rated as having had easier temperaments than SAs. In

addition, there was a more positive parent-child relationship between SRs and their parents

than in families with SA children. For the school-age period, the primary caregivers of SRs

had greater perceived parenting efficacy and more positive parent-child relationships than

the primary caregivers of SAs.








The parents of SRs had aspirations for more fulfilling educational, employment,
and interpersonal futures for their children. With regard to caregiver attributes, parents of

SR compared to SA children had greater perceived resources in the form of more available

support, greater life satisfaction, and a positive self-image. Interviewers who worked with

the families rated SR compared to SA children as also having parents with more positive

attributes, positive parent-child relationships, and support available for the family.

Socioeconomic status was also analyzed by looking at family income level (FIL)

and parent educational level (PEL). The parent occupation level was not included because

40% of the families had mothers as head-of-household and of these, 50% were not

regularly employed. While SAs and SRs were similar in FIL (average monthly income

$600-$900), these groups were significantly different on Parent Educational Level. The

parents of Stress Resistant children had completed more years of education than the parents

of Stress Affected children.

On the topic of resilience, Work et al. (1990) summarized:

many children in this school come from families that have experienced
multiple, chronic life-stressors including poverty, drug and alcohol
problems, disrupted marriages, serious emotional problems, and histories
of abuse or neglect. The undercurrent of stress is pervasive. Yet within it
are resilient children who adapt, indeed adapt well, to difficult life
circumstances. (p. 3)
The authors went on to say that evidence was increasingly supportive of the view

that a high quality early caregiver-child relationship provided a sound foundation for a

child's later adjustment, citing such researchers as Ainsworth (1989), Bowlby (1982), and

Sroufe (1983). The findings of Work et al. (1990) suggested that this relationship was

facilitated by the combination of certain child qualities and an effective
caregiver with adequate support. This configuration may provide a set of
conditions that help the child to learn, and internalize, attributes, e.g.,
empathy, perceived self-worth, a sense of efficacy, that favor effective
coping with adversity.
The authors stated that while the results were "informative and heuristic" there were

also important limitations to the study. The main one was that the interview areas assessed

information retrospectively for the preschool years e.g., temperament and family milieu,








and could be affected by response bias influenced by current functioning. In this respect the

study had the same flaw as the Werner and Smith study.

A second lesser weakness of the study was its illustration of the quality of the

parent-child relationship and the child's school functioning rather than only cognitive

functioning. School functioning can include many social skills. Not separating this out

makes it difficult to assess a child's cognitive functioning.

Following are two studies which used elementary age children as subjects.

However, they have good comparability with the other studies in this chapter as they both

illustrate social class concerns, child stressful life events, the mediating role of the parent-

child system, and competence in children's learning.

Masten, Garmezy, Tellegan, Pellegrini, Larkin, and Larsen (1988) studied children

at risk for psychopathology, the effects of stressful life events on the functional competence

of children, and protective/competence factors, influencing stress resistance in children. In

this study "stress resistance" was defined as competent manifestations in children despite

experiencing stressful events. Dispositional attributes, characteristics of the family milieu,

individual developmental characteristics, and parental attributes were hypothesized to relate

to resiliency and coping in the children.

Subjects were 205 third- through sixth-grade children and their families drawn from

two urban community samples involved in Project Competence. The measures were the

Life Events Questionnaire (Coddington), interviews for a global subjective rating of stress

level of the family, socioeconomic status, grades, standardized test scores, Peabody

Individual Achievement Test (PIAT), vocabulary and block design of the Wechsler

Intelligence Scale for Children (WISC), and interviews assessing child competence and

milieu and child attributes.

In one study with this group, Garmezy, Masten, and Tellegen (1984) found that in
PIAT scores, IQ played a large role which the authors acknowledged was "predictable". In

addition, SES and stress accounted for small but significant additional amounts of variance








in the PIAT. An interaction effect was found for IQ and stress to PIAT performance. The

relationship appears to be conditional and is discussed in terms of models:

... The interaction provides an example of the immunity-versus-
vulnerability model. It also indicates a compensatory effect As for the
latter, PIAT performance, regardless of stress level, covaries with IQ. As
for the interaction, IQ also appears to function to some degree as a
protective factor. High IQ children maintain good PIAT performance at
both low and high levels of stress, whereas for low-IQ children, PIAT
performance drops off as a function of higher stress. This interaction is
also significant for the other achievement measure based on school records.
(p. 107)

In a second study with the same group that included family milieu and stress

(Masten, Garmezy, Tellegan, Pellegrini, Larkin, & Larsen, 1988), the PIAT was dropped.

Child competence was defined as scores on an instrument developed by Garmezy to

measure engagement or disruptiveness in school functioning. Multiple regression was used

to analyze the data. Gender was entered first since it was fixed; IQ was next because of its

well established link with child competence and because it was considered less changeable

than environment; then family characteristics, such as quality of parenting, were entered

over SES because they were psychologically more informative; and finally the life events

were entered after all these personal and environmental variables so as to incorporate their

longer-term role in competence before considering the effects of recent life events.

The results indicated that parenting quality covaried highly with IQ, but even after

IQ was partialled out, parenting quality and child competence were significantly related.

Likewise, controlling for SES, parenting quality still showed a significant relationship with

the engaged criterion. This condition was true for the disruptive criterion among the girls.

After removing IQ and parenting quality first, SES differences for all criterion measures

were nonsignificant.
When the first four variables (gender, IQ, parenting quality, and SES) were

partialled out, life events was a significant predictor for the disruptive criterion. An

interaction for parenting quality by life events was observed for girls only. The increase of

disruptiveness in girls when low quality of parenting was present along with high life








events stress indicated that the increase was an accelerating function. Masten et al. (1988)

found that:
For the engaged-disengaged criterion of school competence, the level of life
events was inversely related to competence for children high on parenting
quality, in keeping with the vulnerability model. For the criterion
disruptive, life events related substantially to the criterion only for children
low on the attributes of IQ and SES, in keeping with a bipolar protective-vs-
vulnerability model. (757-758)
IQ was the most "salient" predictor of academic achievement, with no interactions

present. The authors did not treat IQ as an outcome measure at any time and it is

unfortunate that the PIAT as an outcome measure was dropped. However, the authors did

replace this with the engaged/disruptive measure of competence as related to school

functioning, making it more similar to the Work et al. study with respect to type of outcome

measure. This study was helpful in demonstrating that a relationship existed between child

stress, the quality of parenting, and child functioning. As mentioned at the outset,

however, the children in this study were elementary school age as opposed to preschool

age and this limited the generalizability of the results to a younger developmental period.

Longfellow and Belle (1984) studied stressful environments and their impact on

children in 160 mother-elementary school age child dyads from low-income families. An

interview was conducted with mothers on: demographics; sources of stress across different

areas of life; availability of support; feelings of stress and strain; current mental health

status; and any physical, emotional, and learning problems their children might have

(termed child adjustment problems). Children also completed an interview which asked

about their relationship with their parents, among other things.

The frequency of adjustment problems had a significant relationship with several

stressors; health problems, emotional problems, and social network problems. Those

mothers whose children had more adjustment problems had less education and were not as

likely to be employed. Little variation was present from one stressor to the next in

magnitude. Because so many of the stressors (disadvantaged backgrounds, lower

educational attainment, and less likely to be employed) associated with child adjustment








were also related to maternal depression/mental health, the authors investigated "the

possibility that maternal mental health mediated the impact of the other life stressors on

children" (p. 68).

Path analysis showed that 14% of the variance was explained by the two sources of

stress (situational environmental stress and maternal stress related to physical and emotional

health) and each made a unique contribution to the variance in child outcome adjustment.

The authors concluded

thus, the data support a model in which both situational stressors and
maternal well-being have a direct effect on children's adjustment. In
addition, the two sources of stress were highly correlated with each other;
the path analysis further suggests that some of the impact of situational
stressors on children is mediated by the mother's well-being. (p. 68-69)

A significant interaction effect was found for gender, with stressors adversely affecting

boys more than girls, regardless of number. Both boys and girls experienced the same

stress from the situation and their mother's well being but were differentially impacted by

this stress. In a study by Pianta and Egeland (1990), however, it was girls who were more

adversely impacted than boys. Gender of the child was a variable which was inconsistent

in showing differences. Children in general appeared to be influenced by stress.

A limitation of this research was that the adjustment problems were not reported

separately to investigate the relationships between stress and parenting as opposed to just

learning problems. In addition, life events and conditions explained less of the variance in

outcomes for children than for adults. The authors explained

in part this may reflect the fact that the study was originally designed to
assess the impact of stressors on maternal mental health and therefore the
interview was geared towards measuring those events and conditions
believed to be most stressful to mothers. Most likely, though, is the fact that
school age children are buffered from the impact of life stressors in a variety
of ways. Most important is the physical and emotional status of their
mothers, but other factors, not included in our analysis are also significant
such as the supportiveness of the relationship between the parent and child.
(p. 76)

The last study examining relationships among the three variables was a longitudinal

study conducted by Murphy and Moriarty (1976). These researchers studied 32 randomly








selected normal preschoolers under daily stress and its relationship to coping and

vulnerability in functioning through variables explaining this relationship (e.g.,

temperament, parent-child relationship). The sample, small to begin with, included no

minority families, few families on the lower or upper ends of the SES range, and the

children had IQs just slightly above average (116). This study however, had much to offer

as research on "normal" children (e.g. Anna Freud) exhibiting disruption of verbal

cognitive functioning under stressful life events. The research methodology was

observational and done as a type of case study on each of the 32 children.

Observation, psychological, psychiatric, and pediatric exams were collected during

preschool, latency (prepuberty), puberty, and late adolescence. Full scale intellectual

testing (WISC) was done on the children and their home situations were assessed. The

researchers kept track of life events by word of mouth (as they happened) instead of using

a formal instrument. While the main focus was on the children's lives, the relationship

between the child and mother was a large part of the study. With respect to parenting

stress, a handful of parents engaged in interviews when the children were in latency.

These interviews were about the "parent's ways of coping with stress".

A major construct studied in the children was "the capacity to maintain internal

integration". During the preschool stage, deterioration or inhibition of functioning in

integrative, cognitive, motor, or affective areas signalled vulnerable areas. Vulnerability

was defined in the study as "susceptibility to stress". In summarizing the influence of

stress throughout developmental periods, the authors wrote that:

... infancy vulnerability is negatively related to a wide range of coping
patterns at prepuberty, and preschool vulnerability is negatively related to
still more prepuberty items; evidently, sequelae of traumata of the first six
years greatly influences the later capacity to deal with stress, while the
earliest patterning foreshadows the level of perceptual clarity and cognitive
functions. (p. 149)
A system to measure coping using hundreds of observational categories was factor

analyzed and yielded two factors. The second factor, named Coping II, was of particular

interest because of the many cognitive processes involved. It was defined as "maintenance








of internal integration--that is, capacity to manage one's relation to the environment so as to

maintain integrated functioning (free from marked tenseness, unmanageable anxiety, loss

of motor coordination, deterioration of speech, disorganization of thought processes, and

so on)" (p. 116). Coping II was related to child functioning items such as "achieving clear

structure has priority over expressing affect," competence, task involvement, low

impulsiveness, and sense of self-worth, among others.

Those children who were judged as more vulnerable from infancy to preschool age

had experienced illness which "diminished the child's adequacy" but more important,
"whose mothers' difficulties (depression or other emotional disturbance, extreme fatigue,

or anxious over-restrictiveness) contributed to the child's vulnerability. Less resourceful

coping capacities were seen in the children who showed greater vulnerability" (p. 131). For

example, in studying the mother-infant interaction and coping at puberty, the authors found

the amount of attention given by mothers to young children had a curvilinear relation with

prepuberty phase functioning. A "balanced amount of attention is related to orientation,

perceptual clarity, internalization of standards, trust, and internal integration" (p. 144).

With respect to cognitive outcomes, the children's IQ variability was related to

coping, but an even more sensitive measure of vulnerability to stress was found in

language, especially in speech. The authors reported that they unexpectedly found 75

percent of these normal children exhibited deviations severe enough to interfere with

communication with others in their prepuberty years. Further, they found that in many

cases "vulnerability" in speech and vocalization went back in time to the infancy years and

continued to manifest all the way through to prepuberty.

The authors noted that vulnerability in the speech-language areas might persist in

relation to temperament (biological) bases as the children rated most vulnerable in speech
and language were this way during infancy and preschool. "On the other hand, it was

equally clear that changing developmental pressures and conflicts, as well as family

stresses, particularly from relationships with mother, contributed to continuing








vulnerability in speech" (p. 238). Some of these stresses were: long hospitalization in

infancy, repeated illnesses, marital discord or instability, inconsistent maternal handling,

temporary loss of mother, or maternal care style mismatched with babies needs.

Finally, for coping in children, the capacity to seek help was a characteristic of

resilience found to be related to life events stress. Murphy and Moriarty report that those

children who did not exhibit a strong drive for recovery and mastery had experienced
"prolonged or multiple stress in addition to the mother's early tension; family stress and

long drawn-out periods of suffering left emotional scars and decreased resilience"

(p. 291). These results were especially powerful in light of the small sample size, the fact

that the stimulation in the homes was reported overall as very favorable with moderate to

high family interaction, and that these families had many socioeconomic resources.

Summary

The studies reviewed in this section illustrate the role played by parent-child system

stress in the relationship between the stressful life events of a child and his/her cognitive

functioning. The following can be said in summarizing this section of the literature review:

(a) There is a relationship between the child's stressful life events and his/her

cognitive functioning. These correlations are in the range of -.02 to -.33.

(b) Parenting stress appears to be a mediating variable which influences the

relationship between child life events stress and child verbal cognitive functioning. When

there is high parenting stress, there is often high child stress and lowered child verbal

cognitive functioning. The reverse has also been found. When parenting stress is low,

there is often lower child stress and higher child verbal cognitive functioning. The

correlation between parenting stress and child stress is .11 to .56 and that between

parenting stress and child cognitive functioning is -.04 to -.3.

(c) Analyses suggest these relationships are present in the preschool age child.

However, these relationships are not as clearly confirmed as in those studies with school

age children. Young children for the most part have not been well researched with respect








to their life events stress, the stress present in the parent-child relationship, and their verbal

cognitive functioning. The main reason given for the lack of study on preschool children

was that children of this age were more difficult to study than school age children and

adolescents. This was because fewer life events had happened to younger children and

young children themselves could not complete the measures, making them more difficult to

sample. However, many researchers felt the preschool years were critical in understanding

stress and functioning as most issues of childhood were ruled to some extent by the child's

developmental period. Therefore many stress researchers assessed stress retrospectively

from the school years to the preschool years. Unfortunately, data collected in this way was

strongly susceptible to response bias. In the studies conducted with both preschoolers and

school age children, there were many types of stress measures used. In many cases there

was a lack of distinction between the parenting stress and the child's life events stress

measures.

(d) The studies with preschoolers used well-established indicators of cognitive
functioning (e.g., verbal IQ tests and speech and language performance). Several of the

studies were conducted with both low and high or exclusively low socioeconomic status

groups of school age children following the literature which indicates that low income

individuals experience more negative life events and lower cognitive functioning than

individuals in higher socioeconomic groups, but that resilience was present in low income

families.

A model described by Cowen and Work (1988) seems useful for summarizing the

major points of the research on resilience:

This largely developmental/transactional model, similar to one proposed by
Lorion, Tolan, & Wahler (1987) suggests that resilient outcomes are
favored by interplays between qualities of the child and the family context in
which s/he is reared. It stresses that the availability of adaptive models who
provide care, warmth, support and understanding, helps the young child to
acquire resources and strategies needed to cope effectively with major life
stress. Whereas, initially, the model's support is largely protective, later, as
the child's awareness grows, models) provide concrete examples of
effective coping strategies in difficult circumstances, that the child comes to
adopt in dealing with future stressors. (p. 5)








The next sections will be focused on the variables of interest in pairs. While in each
instance a piece of the whole model as outlined above will necessarily be missing, these

studies added to the understanding of the proposed study. In particular the lack of attention

to preschool children was prevalent.

Child Life Events Stress and Child Cognitive Functioning

This section of the literature review was concerned with studies which measured the

relationship between child stress and child cognitive functioning. Three studies will be

reviewed. One focused on intellectual functioning and two on school functioning. All three

were conducted with young elementary age children since no studies on these two variables

were found which used preschool children.

Using data collected in the 1960's, Brown and Rosenbaum (1984) studied the

effects of stress on IQ in 4,154 middle and lower class seven-year-old children drawn from

the NINCDS Collaborative Perinatal Project. The measure of stress was an index score

composed of the number of medical/psychological problems found in a child's life. These
included: mother's marital status, employment, family configuration, history of illness,

death, and divorce in the family, measures of autonomic function, achievement measures,

and physician-identified disorders. IQ was measured by the WISC. Scores on the WISC

versus stress level of the child produced an inverted U relationship. For low

socioeconomic children, the curves were the same shape, but "peaked at lower stress levels

suggesting higher arousal" (p. 136). In this study, SES interacted significantly with

stressful life events.

Sander and Block (1979) studied kindergarten through third grade children who

had been identified by their teachers as having manifested adjustment problems in the

classroom. As subjects, 99 children in four inner-city elementary schools were chosen. A

control group of 44 children was randomly selected within each school and matched the

maladaptive group for grade, gender, and ethnic membership. The sample was mostly

ethnic minority, poor, and from single-mother homes.








The children were rated by their teachers on dimensions of maladaptive behavior

using the AML, with scores for: A (acting out), M (moody-withdrawn), L (learning

problem), and a sum AML. Parents completed the Louisville Behavior Check List

composed of factors such as sensitivity, anti- or prosocial, and academic disability. The

stressful life events of the children were measured by a life event schedule with items

combined from the Coddington (1972) Life Event Record and a scale created by Gersten,

Langer, Eisenberg, & Simcha-Fagan (1977).

The results indicated that those children identified by their teachers as having

adjustment problems in the classroom had more stressful life events occur in the previous

year than controls. Further, the group considered to be maladaptive was significantly

higher on stress measures of Ambivalent, Undesirable, Total Events, and Weighted Sum

Scores than the controls.

In a subsample of the school maladjusted group, life stress events were

significantly related to parent ratings of child maladjustment and Total Stress Events,

Weighted Sum Scores, Undesirable Events, and the Undesirable-Desirable score for the

maladapted children. The Total Events score tended to have the highest correlations with

the dimensions of maladjustment. The Undesirable Events score was the most consistent,

being significantly correlated with all nine dimensions of maladjustment. The authors drew

the conclusion that "the results of the present study can be interpreted as supporting the

view that it is the desirability rather than the change per se inherent in an event that mainly

determines its stress value" (p. 437).

A limitation of this study is that the measure of cognitive functioning is mixed with

behavior and mood of the child. These influence the social/emotional aspect of school

functioning. It is difficult to see the relationship of stress and cognitive functioning with

these variables included. A second limitation is having input about the child from the

parents which at first glance might be considered a strength of this study. However, the

implicit limitation of all the studies in this section (no assessment of parenting stress) may








have more impact in this study. Not assessing characteristics of the parent failed to give

insight into parental perceptions which might have influenced responses about children, as

was seen in the Prince et al. (1972) study.

In 1985, a study was conducted by Sterling, Cowen, Weissberg, Lotyczewskil,

and Boike to examine the recent stressful life events of first through fourth grade children

and their school adjustment. Out of 974 children, 211 were identified by their teachers as

having had one or more of 11 Stressful Life Events (SLEs) (cf. below) occur during the

present school year. A group of 211 children designated by teachers as having experienced

no stressful life events during the current year served as a comparison group. This group

was closely matched to the SLE group for location (urban, suburban), school, grade,

repeat in grade, age, gender, and ethnic background.

Teachers completed all measures. The 11 stressful life events were taken from two

child life-events inventories, the Coddington Life Events Record and the Gersten scale.

The events studied were death of a parent, sibling, or close relative; serious illness of a

parent, sibling, or close relative; lengthy illness and/or hospitalization of child; school

transfer; parents separated or divorced; parent remarried; parent lost job; family

experiencing severe economic difficulties; change in home residence; new child born into

family; new adult or child moved into the home. The Classroom Adjustment Rating Scale

(CARS) was used with factors consisting of acting-out, shy-anxious, and learning

difficulty; and the Health Resources Inventory (HRI) which measured the following

elements: good student, copes well with failure and other school pressures.

There were two main findings of the study. First was that those children who were

reported by their teachers as being more maladjusted and less competent than peers matched

on demographics, experienced one or more recent SLEs. On 8 of the 10 measures (with

the exception of the CARS Acting-out and the HRI Rules), significant differences were

found, with more severe problems and more incompetencies rated for the group with SLE.








Second, those children who experienced multiple recent stressful events were rated

as more maladjusted and less competent than their peers who had experienced fewer of
these events. Three subgroups were formed based on number of SLEs experienced by the

child in the current school year, based on one such event, two such events, and three or

more such events. Significantly fewer competencies and more serious problems were

reported for the three or more SLEs group than for the other two groups. The authors

further commented that the problem of stress in children was of concern. Within their

large, representative sample, 22% of these young children experienced in the preceding 8

months, one or more of the SLEs rated.

A major limitation of this study was that it did not address income level. Brown

and Rosenbaum (1984) and Chandler, Million, and Shermis (1985) found the presence of

stressful life events to be higher the lower the SES status. In all fairness, many of the

studies showing these relationships were published at the same time as this article being

discussed. A second limitation was that while it may be assumed that the teacher would

know whether the 11 life events measured had occurred, there was a possibility that events

occurred in the child's life which the teacher was unaware of or misinformed of. Finally,

similar to the study by Sandier and Block, the measure of competence was school

competence (including social skills) rather than cognitive functioning.

Summary. These three studies together show that intellectual and school

performance are related to the number of stressful life events experienced by early school

age children. In summarizing this section, the following conclusions have been developed:

(a) Cognitive functioning often decreases when children experience severe and frequent
stressors. This was reflected in IQ scores and school adjustment in the form of academic

motivation and performance problems, ineffective learning skills, and academic disability;

(b) A curvilinear relationship was present in the case of IQ and stress. The relationship
was linear with school achievement and stress. The difference may be that IQ was more

clearly or closely related to cognitive functioning than school functioning which could








include social skills, thus offsetting or masking the downward trend in cognitive

functioning; (c) Little research has been done on the relationship between stressful life

events unique to preschoolers and child cognitive functioning during this developmental

period.

Child Life Events Stress and Parenting Stress
In this section of the literature review, two studies examining the relationship
between child life events and parenting stress will be presented. Both these studies were

conducted with young elementary school age children. A third study examined parenting

stress in detail in mothers of Head Start children as it fit in with the findings of the first two

studies and more fully developed some of their points. Webster-Stratton (1990) wrote in a

review article that one of the factors "shown to influence parental perceptions and

behaviors" (p. 302) was negative life events. Mash & Johnston (1990) indicated that "in

particular,... adverse life circumstances (e.g., life stressors) have been shown to have a

detrimental impact on parenting and to be associated with increased parent-child interactive

stress" (p. 318).

A central question was the relationship between the child's stressful life events and
his/her parent's functioning in the parent-child relationship. This was investigated by

Jensen, Bloedau, Degroot, Ussery, and Davis (1990) in studying the referral of children to

a psychiatric clinic. Subjects were 134 six through twelve-year-old children referred to a

military psychiatry clinic and 134 control children. The Coddington Life Events Record

was used to measure life stress in the children and parents completed the Hopkins'

Symptom Checklist for psychopathology in themselves.

Regression analysis showed that the life stress contributed from 2.9% to 7.1% of
the variance in children's symptoms when parent symptomology was in the model. "When

parental symptoms were forced out of the model in further multiple regression, stress

explained as much as 11% of the variance in children's symptoms. This suggests that

parents react to the life events stress purported to affect their children" (p. 57).








In a second study by Jensen, Richters, Ussery, Bloedau, and Davis (1991) on this

same sample, findings were that in children referred to the clinic, higher levels of events

confounded with their own maladjustment were found than for the control children. The

authors wrote that "more interesting, however, is the fact that, compared with the control

children, clinic children had also experienced higher levels of events confounded with their

parents) functioning (termed parent-marker events) as well as higher levels of normative

child life events" (p. 305).
The level of parent-marker events was significantly associated with the level of

normative events in both the clinic and control children. The authors tested for the

possibility that differences in normative events between the clinic and control groups might

"be attributable to parent functioning as indexed by the parent-marker events" (p. 305).

Results supported this by showing that the normative event differences were no longer

significant when the parent-marker events were controlled.

Parental psychopathology therefore was a key parenting stress indicator (Abidin,

1985, 1990). Hastings-Storer's (1991) work on 85 low-income Head Start mothers gave

insight into parent psychopathology, along with the influenced perceptions of the child's

temperament found to influence the parent-child relationship stress level (Abidin; 1985,

1990) and coping for the child himself (Rutter, 1981; Compas, 1987). The Parenting

Stress Index was used as the measure of parenting stress. Three groups with different

stress levels (high, normal, and low) were formed, using cutoff scores recommended by

the author of the instrument.

Overall, with respect to the Child Domain Stress in the Adaptability subscale, focus

(or study) group data showed that only a minimally positive relationship was present

between the parent and child for a significant group of the mothers in the study. An

explanation was that mothers might have a hard time relating to the child's point of view

and world experience (e.g, in needing routine), decreasing the sensitivity of the mother to

the needs of the child. The author quoted Abidin on parent-child attachment: "When the








Adaptability/Plasticity scale is elevated along with the Child Reinforces Parent and Parent

Attachment, there exists a strong indication that a minimally positive relationship exists

between parent and child" (p. 59). The Child Reinforces Parent subscale scores showed

high stress levels with over half of the experimental group scoring in the High stress

category. The absence of reinforcement from the child could threaten the parent-child

relationship. Possible reasons for a high score here were: the child might have been

depressed, the parent might have been misinterpreting the child, the parent was depressed

or not supported, or the parent expected too much from the child. Abidin stated that when

the parent did not receive reinforcement from the child, the parent-child bond was

threatened. The Mood score was also high in these children. High scores were commonly

found with children who had dysfunctional affective expression and extreme scores

indicated impairment in maternal attachment to the child.

When studying the Parent Domain Stress scores, a large percentage of mothers fell

in the high stress group on Attachment as well. Hastings-Storer wrote that this was

disturbing due to the many connections parental attachment had with child functioning of a

cognitive, social, and emotional nature.

Attachment problems were a strong theme in this sample. A significant number of
mothers also had elevated scores in Acceptability. Abidin writes this means, "that the child

is not as attractive, intelligent, or pleasant as the parent expected or hoped. Poor attachment

and/or rejection may consciously or unconsciously be issues in the relationship between the

parent and the child" (p. 60). Hastings-Storer commented that none of the children scored

in the potential problem category range on the DIAL-R and that the child not being as

intelligent as the parent had hoped was not a valid explanation for high scores in

Acceptability.

The DIAL-R was designed to tap severe problems such as those which might
warrant special education. Within the range operating in this sample, analysis for

differences on the DIAL-R among children whose mothers varied on stress levels was not








carried out. To be a point or two above the cutoff for "Potential Problems" on the DIAL-R

or at the top of the range for "OK" represented a wide range of cognitive functioning. In

addition, the DIAL-R had a domain for motor. If the total score were used, a high Motor

score could mask a low Language or Concepts score (the other two domains).

Social Isolation was another Parent Domain subscale where the focus group had

high scores. Hastings-Storer quotes McLoyd (1990) as saying psychological distress in

parents in relation to isolation "increases the tendency of parents to be punitive, erratic,

unilateral, and generally nonsupportive of their children" (p. 330). The third subdomain in

which the subjects were significantly different from the norm group was on Relationship

with Spouse. Here dissatisfaction with the support given by the children's father was

reported in the study groups. While many fathers sent material help, the mothers reported

wanting active support in rearing the child. Hastings-Storer felt that program services

should encourage father-child involvement rather than zeroing in on marriage counseling.

One drawback of using the Hastings-Storer study to explore further differences in

parental psychopathology found in relation to child stress studied by Jensen et al. was the

difference in the two groups of subjects. All of the mothers in Hastings-Storer's research

project were low-income African American, whereas the subjects from Jensen et al.'s study

were more evenly distributed across socioeconomic and ethnic lines. On the other hand,

since psychopathology is more prevalent among low-income individuals, the fact that the

Hastings-Storer sample was low-income as opposed to middle or high helped offset this

limitation.

Summay. The following remarks can be made with respect to this section on child
life events stress and parenting stress: (a) A clear relationship has been found between the

child's stressful life events and aspects of parenting stress such as parental

psychopathology in elementary school age children; (b) Parents were both impacted by the

life events their children experienced and the parents influenced the occurrence of these

events in the first place in elementary school age children; (c) Research on parenting stress








with respect to child and parent characteristics revealed a great deal of stress present in the

parent-child relationship between low-income mothers and their preschool children.

Attachment between the mother and child was a major area of stress for mothers and their

young children. In the literature on child stress, attachment issues were considered the

most stressful for preschool children (Barton & Zeanah, 1990); (d) The work on parental

psychopathology as a contributing factor to child life events stress has focused on

elementary school age children. Little research has been done on preschool populations

with respect to the interplay between parenting stress and child life events stress.

Parenting Stress and Child Cognitive Functioning

There are a few recent studies which have focused on the mediating variable of

parenting stress with regard to the child's cognitive functioning. In this section are two

studies. The first measured parenting stress in the parent-child system and its relationship

with child cognitive functioning (particularly verbal) in early elementary school age

children. The other study in this section used large data bases of life events in low-income

mothers when their children were preschoolers and school age.

Much of the stress the mother felt came from her relationships with those people in

her life who were important to her, such as her child and significant others. This in turn

influenced the child. Trad and Greenblatt (1990) stated, "the status of a child's attachment

to parents and the milieu or origin are crucial factors in stress resistance and vulnerability"

(p. 35). Similarly, Barton and Zeanah wrote, "important in mediating the effects of stress

on the young child is the nature of the child's tie to primary caregivers" (p. 211).

One approach to studying parenting stress (and the one taken in the proposed study)

was to look at the stress present in the parent-child relationship. Relationships were a part

of parenting stress; so too were mothers' feelings about parenting, as well as the

contribution of the child with respect to temperament and attachment. Mclntire (1991)

conducted a study of parenting stress and child cognitive outcomes with middle and lower

class subjects. This study utilized Jay Belsky's (1984) model of parent, child, and








situational determinants and the influence they have on parenting. Belsky's work

suggested that the parent's personal psychological resources were the most critical in the

protection of the relationship between the parent and child from stress. Psychopathology in

the parent placed children at risk for poor developmental outcomes while parents who were

healthy were able to withstand enormous adversity and could assist children in overcoming

behavioral or developmental handicaps.

The mean age for this sample of boys in Mclntire's study was 8 years, 9 months,

and ranged from 7 to 13 years. One parent group consisted of 42 single mothers, a second

group of 52 married mothers whose husbands did not respond or refused to participate in

the study, and a third group of 69 married mothers and their husbands who agreed to be in

the study. Measures used were Conner's Teachers Rating Scale (CTRS), Peabody Picture

Vocabulary Test-Revised (PPVT-R), Wide Range Achievement Test-Revised (WRAT),

Parenting Stress Index (PSI), Parental Locus of Control Scale (PLOC), Knowledge of

Behavioral Principles as Applied to Children (KBPC), Parenting Alliance Scale (PAS), and

the demographic variables socioeconomic status, education, and family income.
The results may be presented in terms of stepwise multiple regression. The child's

IQ was predicted by Maternal Knowledge of Behavioral Principles and Child Domain

Stress. The child's Reading level was predicted by Child Domain Stress and Maternal

Knowledge of Behavioral Principles. The Spelling level was predicted by Child Domain

Stress and Maternal Knowledge of Behavioral Principles. Finally, the Teacher's

Perception of the Child was predicted by Maternal Perception of the Child and Parent's

Affective Response.

In addressing the question of predictors of maternal stress, McIntire followed
Belsky's (1984) hypothesis that the order of predicting psychological factors would reveal

that parent psychological factors would have the most power, contextual factors would be

secondary, and child characteristics would have the least amount of power for prediction of

parenting stress. The results of the regression analyses generally supported this order, but








in several cases the child characteristics were stronger predictors than the contextual

factors. Parenting Stress Index (PSI) Total Score was predicted primarily by Maternal

Locus of Control, followed by Maternal Perceptions of the Child, and Income; Child

Domain Stress on the PSI was predicted by Maternal Perception of the Child, Maternal

Locus of Control, and Child's IQ; and Parent Domain Stress on the PSI was predicted by

Maternal Locus of Control, Maternal Perception of the Child, and Family Income.

Separate analyses on the single and couples groups found parental support added

significantly to prediction equations for total and contextual sources of stress. This was in

keeping with the overall model that marital status was a critical factor in the amount of

stress present in the parent-child relationship and in child outcomes. To summarize the

relationship between parental stress and child outcomes, Mclntire (1991)writes:

All measures of stress are negatively, and significantly, related to the child's
IQ, suggesting that parents of intelligent children experience less stress in
parenting. (p. 89) The Teacher's Perception of the Child's adjustment was
positively related to each stress measure...,indicating that children who are
seen by their teachers as being deviant or poorly adjusted in the classroom
in the direction of exhibiting hyperactivity, aggression, or defiance, have
parents who are experiencing elevated levels of stress in coping with the
child. (p. 90)
As mentioned at the onset, if demographic variables are a part of the studies with

any of the variables (child verbal cognitive functioning, child stress, and or parenting

stress) they will be reported when that study is reviewed. McIntire found with regard to

demographics that the majority of the mothers in the study were high school graduates.

Couples group mothers had obtained more education than mothers in the single mothers

group. Percentages showed that 17% of single mothers were not high-school graduates

compared with 8% of the mothers in the mothers-only group, and 6% of the mothers in the

couples group. Fathers had significantly more education than mothers. Significant group

effects for income were found, with the single mothers making less money than either of

the other groups. In addition, couples group mothers held significantly more prestigious

occupations than did single mothers. There were no group effects for number of children

in the home nor age of the child. Education of the mother did not distinguish between








groups of mothers. One reason may have been a restriction of range operating with this
variable as the majority of mothers were high school graduates. Other researchers (e.g.,

Hastings-Storer, 1990; Conger et al., 1984) found that mother's education was very strong

in predicting parenting stress.

A second approach to studying parenting stress was to look at the number of life

events the parent was experiencing. This gave the added benefit of knowing some of the

events which happened with the children, but a drawback was that little was seen with

respect to the processes that regulate stress in the parent-child relationship. In defense,
however, the quality of the home environment was assessed in the studies to be reviewed

and found to be important. Pianta and colleagues (Pianta & Egeland, 1990; Pianta,

Egeland, & Sroufe, 1990) studied cumulative maternal life events stress on young

children's functioning (birth through second grade) in the Mother-Child Interaction Project.
Due to their disadvantaged economic status, these mothers were seen to be "susceptible to

caretaking problems". Further risk factors included: low educational level, lack of marital

support, generally unstable living conditions, and exposure to a variety of environmental
stressors. The majority of the sample was white and most of the pregnancies unplanned.

The measure of maternal stress was an interview asking if a particular event or

condition had occurred since the previous assessment and if it was still going on, along

with assessment of the extent to which the family functioning was impacted. These items

mostly tapped stressful experiences in the mothers' interpersonal relationships, as

discovered by factor analysis, so the subscale formed from summing these items was

named Personal Stress. Items with loadings of greater than range .30 were included in the

factor. The factor accounted for the greatest percentage of variance in each of the solutions
(between 35% and 50%). These scores were gathered approximately every year and a half.

Pianta and Egeland's conclusion was that stress in economically disadvantaged mothers

having to do with interpersonal relationships was "a particularly salient type of stress....








This factor was a significantly better predictor of children's adjustment in first grade than

was the life event total score obtained from the same scales" (p. 330).

Pianta, Egeland, & Sroufe (1990) studied competence in the children of these same

mothers by looking at intelligence (Weschler Preschool and Primary Scale of Intelligence),

language ability (Zimmermann Preschool Language Scale), and problem-solving (barrier-

box situation) in children. Competence in children was determined by the Child Behavior

Checklist for Teachers and Teacher's Ratings of Cognitive Competence. When looking at

mothers who were highly stressed, competent boys were distinguished from their peers

who were incompetent by several factors. Competent boys compared to less competent

boys had higher intelligence, language ability, showed more positive affect and more

creativity in problem solving. In the problem solving situation for the less competent boys,

Pianta, Egeland, & Sroufe observed that the situation was "basically a negative one for the

less competent boys" (p. 226). This group avoided their mothers, was less persistent and

enthusiastic, showed less affection for their mothers, and had more negative affect. The

home environments of competent boys differed from less competent boys in the "extent to

which the mothers provided a structured, responsive home environment despite high

stress"(p. 226). Differences on ratings reflecting positive mother-child interaction were

noted with respect to a lack of respect for the child's "autonomy, poorer quality of

instruction, lack of structure and limit setting" (p. 226) as well as emotional warmth of the

home. Further, "These data suggested that competence in boys of highly stressed mothers

was in part due to the mothers' ability to buffer their sons from the effects of stress" (p.

331).

Pianta et al. (1990) wrote about the two genders, "Like the boys, the competent
girls were more intelligent, had better language skills, and had more positive home

environments." (p. 331). However, in the girls, competence was most highly related to

positive maternal personality characteristics as assessed by the Sixteen PF Personality

Assessment gathered when the children were 64 months old. In summary, "These data








were interpreted to suggest that competence in girls depended on their mothers' personal

adjustment, which may have had the double benefit of buffering their daughters from the

negative effects of stress and also providing their daughters with a role model for positive

coping" (p. 331).

The findings suggested that "maternal stress was related to markedly maladaptive

mother-daughter interaction" (p. 332). The interaction between mother and child showed

mothers' lack of support for their daughters, use of intrusive and hostile attempts to

interact, and a lack of persistence by the girls, low enthusiasm toward the task,

noncompliance, and avoidance of their mothers. Personal Stress, which was experienced

by the mothers between the birth of their daughter and 42 months, accounted for almost

one-quarter of the variance of avoidance of their mothers by girls.

Summary

Summarizing this section on parenting stress and child cognitive functioning, the

following statements can be made:

(a) When the parent-child system was stressed, whether by threats to the

interpersonal relationship of the mothers with their children and/or significant others, the

child was often found to have lower cognitive verbal and school functioning.

(b) Parenting stress in relation to child cognitive outcomes revealed an inverse

pattern. When the parenting stress was low, the child's verbal cognitive functioning was

higher than when parenting stress is high.

(c) The studies in this section on parenting stress and child cognitive functioning

examined the parent's rather than the child's unique life events stress and may have missed

the more direct influence this stress could have on the child's verbal cognitive functioning.

However, it should be kept in mind that in many cases if the parent experiences a life event,

the child does too (e.g., divorce, birth of a sibling, moving). As mentioned in the section

on child stress and parenting stress, it is often the actions or the pathology of the parent that

are influential in the type and amount of the child's life eventss.








(d) The relationship between parenting stress and child cognitive functioning is

unclear in preschoolers. This is due to a measure of parenting stress made up exclusively

of stressful life events for the mother in the study with preschoolers (Pianta et al.). Issues

such as the child's characteristics were not assessed. Mclntire's study on school age

children used a more well-rounded measure of parenting stress, but was limited in

generalizability to preschoolers as it measured children in a different developmental level.

The preschool period is considered to be the most stressful for parenting (Abidin).

Child Cognitive Functioning and Demographics
Many studies have found children differ in cognitive functioning based on social

class demographic indicators. The five variables of interest in this study were parental

educational level, parental employment status, parental marital status, parental age, and

ethnicity. In this section, fourteen studies will be reviewed. They are reported in the order

of those studies which included the most demographic variables of interest (parental

educational level, employment status, marital status, age, and ethnicity) to those that looked

at the demographics individually. About half of the studies in this section were conducted

with preschool children and the other half with school age children.

Broman, Nichols, and Kennedy (1975) studied 26,760 four-year-old children in

the Collaborative Perinatal Project of the National Institute of Neurological Diseases and

Stroke. Eighty-two variables were found to be statistically related with Stanford-Binet

scores. The results of the multiple regression analyses showed that maternal education and

the socioeconomic index (a composite of head of household education and occupation, and

family income) contributed the largest proportion of explained variance when prenatal,

neonatal, and childhood variables were added. This effect held across sex and race of

child. In addition, for some subsets of children, presence of father in the home was also

related to cognitive functioning. Broman et al. also compared the Stanford-Binet scores of

the children and found that white children had a higher mean IQ score than Black children,

most noticeable in the middle and upper class children.








Preschool girls and their parents above average in socioeconomic status (N=33)

were studied by Poresky and Whitsitt. The Peabody Picture Vocabulary Test-Revised was

found to be significantly associated with the level of formal education of their mothers.

Laosa (1982a) used a sample of 50 white intact families of children approximately 3

years of age to measure predictors of intellectual competence in preschoolers. The sample

was chosen to be representative of this family composition with respect to socioeconomic

characteristics. The Preschool Inventory was used which tests among other things, verbal

skills "defined by teachers as expected of children in kindergarten" (p. 19). Predictors of

cognitive functioning were mother's "Socioeducational Value" defined as mother's

education and occupation, how much she reads to the child, maternal modeling, how much

others read to child, child sex and age. Using path analysis, Laosa found that "the

influence labeled for easy reference 'mother's socioeducational values' is the most

important determinant of the child's performance on the Preschool Inventory, at least 0.6

times more important than any other source" (p. 32).

Whiteman and Deutsch (1968) studied social disadvantage and intellectual and

language development in a sample of 165 elementary school children. SES was measured

with the Empey Scale of Occupational Prestige, which assessed education and occupation

for head of household. Language and intellectual development of the children was

measured by the Gates Reading Score, the Lorge-Thorndike IQ, the WISC Vocabulary

subscale, and the Orientation Scale (a verbal test tapping the child's fund of general

information and conceptual understanding). The SES index of education and occupation

was significantly correlated with all the cognitive measures used in the study.

Zill and Peterson (1982) reported on data collected in the Foundation for Child

Development's National Survey of Children (Zill, 1983) on 2,279 7-1 1-year-old children.

Subjects were chosen based on a national probability sample, but minorities were

oversampled for this study. Measures included the Peabody Picture Vocabulary Test and a

test of Practical Skills, community size and type, family income, parent educational








attainment, family size, and ethnic group. Education was the best predictor of the child's

picture-vocabulary test of the family characteristic examined. The authors reported the

following results:

The relationship between parent educational attainment and the child's
vocabulary score was essentially linear, with more than a full standard
deviation separating the mean vocabulary score of children in the highest
parent education category (college graduates) from those in the lowest
education category (grade school dropouts). When parent education was
entered into a multiple classification analysis along with the other family
characteristics the strength of the relationship between education and
vocabulary score was reduced somewhat. However, there was still a
difference of nearly three-quarters of a standard deviation in vocabulary
scores between top and bottom education categories. (p. 354)

Maternal Employment with respect to preschoolers' cognitive outcomes were

examined by Baydar and Brooks-Gunn (1991) using data from the Children of the National

Longitudinal Survey of Youth. This topic was of interest to the authors because, "the

number of women who were employed and who had children under age 6 increased from

2.3 million in 1960 to 7.1 million in 1988 (U.S. Bureau of the Census, 1989)" (p. 932).

Subjects were 572 three to four-year-old white children of young mothers, most of whom

had not completed high school, and most of whom were married at the birth of the child.

The majority of the group lived in poverty. Children were given the Peabody Picture

Vocabulary Test-Revised. In this study, maternal education and ability were used as

covariates. The coefficient which indicated that the mother was continuously employed on

the PPVT-R was significantly positive. "Hence, these data suggest that given the mother

was employed in infancy, continuous maternal employment during early childhood was not

detrimental and might be slightly more beneficial than intermittent employment" (p. 938).

In looking at continuity another way, children showed more negative effects when their

mothers worked an average of 10-19 hours than children whose mothers worked an

average of 20 or more hours. A job with fewer hours might be more variable from day to

day or week to week.
Shinn (1978) did a meta-analysis to review father absence and children's cognitive

performance as measured by tests of IQ and school achievement Twenty-eight (the great








majority on elementary and high school age students) of 54 studies reviewed met the

methodological adequacy criteria. Of these, detrimental effects of father absence were

found in 16, no significant effects were found in 9, and positive or mixed positive and

negative effects were revealed in 3. The authors report, "the cognitive differences between

children from intact and fatherless families were of some consequence. They ranged up to

1.6 years in achievement, .9 standard deviation units in IQ and aptitude, and .8 of the

difference between a B and a C in grade point average" (p. 312).

A study with preschoolers was conducted by Eiduson (1983) in which traditional

versus non-traditional families were studied with regard to the child's cognitive and

intellectual development Subjects were 200 families "generally successful and affluent"

(p. 426) with an equal number of nuclear, single mother, social contract (unmarried), and

communal families. Overall, the author reports that the child who experiences "elevated

levels of stress in the family is likely to have lower scores on intelligence tests, such as the

Stanford-Binet, by age three" (p. 431). The events determined to affect cognitive

functioning at this age were those related to the personal sense of well-being of the parent,

as well as those that pointed to a troubled parent-child relationship. The author concluded

that, "while this is true for the sample as a whole, in the first three years, the child most

susceptible to low cognitive scores if the mother herself has psychological problems, is the

child of a single parent" (p. 432).

Allison and Furstenberg (1989) studied how marital dissolution affected children's

functioning. Subjects were 1,197 seven through 11-year-old children in the National

Survey of Children, which chose subjects based on census proportions nationally. Of

these, 328 had experienced a marital dissolution. Parents, teachers, and children

contributed to assessment of well-being of the child, and academic difficulty among others.

Findings indicated that dissolution of the marital relationship had a significant relationship

to reduced well being, particularly in teachers' reports. Also, "not surprisingly," the








impact of the mother's education was stronger than the effects of marital dissolution for

academic performance, while weaker for the other outcome measures (e.g., behavior).

The authors came to the following conclusion:

A decline in effects with age at separation suggests that young children are
especially vulnerable, either (a) because they are more dependent on their
parents and hence, are less protected by extrafamilial supports such as
teachers or peers or (b) because they are in a more formative stage of
development and are therefore less resilient when faced with a traumatic
event. On the other hand, an increase in effects with time since separation
suggests that marital dissolution is not an isolated event but only the
beginning of a continuous exposure to a long-lasting adverse situation that
produces cumulative effects on the child. (p. 545)

Several studies have found that the age of the child's mother was related to the
child's cognitive functioning. For example, Belmont, Cohen, Dryfoos, Stein, and Zayac

(1981) found in a sample of 67,000 that as maternal age increased, 6-11-year-old's WISC-

R scores increased. Broman (1981) found this same relationship between maternal age and

4-year-old's Stanford -Binet scores in a sample of approximately 36,000 from lower to

upper class groups. Maracek's (1980) study (cited in Baldwin & Cain, 1981) looked at

just the low-income group of the sample Broman (1981) studied (approximately 19,000)

children and found that Stanford-Binet scores at age 4 increased as mother's age increased.

Deutsch and Brown (1967) studied 543 first- and fifth-grade children who had

scores from the Lorge-Thorndike Primary Battery. They found that mean IQ scores were

higher for white children when compared to black children, with these differences

increasing as socioeconomic status rose. Schooler and Anderson (1979) reported that in

the preschoolers they studied, white children scored higher on the Slosson Intelligence

Test.

S mmat
In concluding this section on child cognitive functioning and demographics the

following generalizations seem warranted:

(a) The demographic variable of parental educational level was one of the most
powerful and consistent varaibles related to child cognitive functioning.








(b) Parental age followed education level closely with being related in several

large and well-controlled studies to child cognitive functioning.

(c) Occupation and income were related to child cognitive functioning, but they

lost their potency when "family factors" were removed. As Zill and Peterson (1982)

commented, "this indicates that poverty or affluence is less important as a determinant of a

child's intellectual development than are other family factors, such as parent education and

family size, with which family income is associated" (p. 355).

(d) The relationship between maternal employment and child cognitive

functioning may change as a function of the child's dependency on the adult. In addition, it

appeared a stable job or continuous employment was better than erratic employment for the

child's cognitive functioning.

(e) Father absence was clearly related to both IQ and school achievement

Further, marital dissolution was often only the beginning of cumulative stress on the child.

(f) Ethnicity was related to child cognitive functioning with minority children

scoring below white children. However, this difference became greater when comparing

children of different ethnic backgrounds who were in the middle to upper class groups.

For children in the low socioecomic status group, the differences were not as great.

(g) Eight studies in this section were conducted with preschool age children and

six with young elementary school age children, making the generalizability of a relationship

between sociodemographic variables and child cognitive functioning in preschoolers at least

moderate.

Child Life Events Stress and Demographics

Studies have also shown the incidence of child stressful life events to be related to

social class demographic variables. This section presents eight studies, two with

elementary school age children, two with preschool age children, and four with adults.

The demographic variables were parental educational level, parental employment status,








parental marital status, parental age, ethnicity, and occupation and place of residence (these

last two will not be measured in the proposed study).

Chandler, Million, and Shermis (1985) measured the incidence of stressful life

events in a sample of 277 children aged 5 through 14, randomly selected from a school

district representative on socioeconomic status. The study's purposes were to collect

baseline data on the incidence of stressful life events of children, and to study the

relationship of age and SES to the number of events reported for the children.

The measure of life stress was the Children's Life Events Inventory (Chandler) and

socioeconomic status (SES) was determined by using Hollingsheads' Index of Social

Status based on education and occupation. Three SES groups were formed; low, middle,

and high. ANCOVA revealed significant differences among the number of parental

reported events in these three SES groups. Post hoc analysis showed "a significant

difference in the mean number of events reported between the low and high SES groups"

(p. 744) with the low SES group reporting more events.

Toomey and Christie (1990) reported that, "there are few studies that directly ask

children if they are concerned about unemployment, poverty, or financial losses in their

families" (p. 427). Brown, Cowen, Hightower, and Lofyczewski (1986) investigated

sociodemographic differences in the perception and experience of stressful events in a

sample of 503 fourth through sixth grade children. The measure of stress was taken from

existing life events measures with a few new events included, and asked the children to rate

the occurrence of the events. Of these, parental unemployment was ranked fifth in

frequency. This event occurred more often in urban children and was rated more upsetting

by urban children. Brown and Cowen (1988) found that these two groups were
significantly different on common sociodemographic indicators including family income,

median house value, percentage minority enrollment, and percentage at poverty level.

Further analysis revealed that the children living in the urban areas experienced more

stressful events than those children living in the suburbs.








Eiduson (1984) studied child stress in traditional versus non-traditional families

who were "generally successful and affluent" (p. 426). Data were from the Family Styles

Project involving 50 nuclear families, 50 single mother families, 50 social contract families

(unmarried parents), and 50 communal families. The traditional nuclear family was found

to experience a lesser number of stressful child life events than the other family types.

Single-mother families were the most "susceptible" to stress.

As many of the events on life events measures are similar to the ones on child life

events measures, children of mothers will probably also experience these events. Swanda

and Kahn (1986) studied rural persons between the ages of sixteen and sixty-six and found

significantly higher levels of stress in the younger subjects. Samuelsson (1982) initially

studied 800 women, followed up six years later with 677 of the subjects, and found that

recent undesirable life events decreased with age. Beard (1982) found in university

students, faculty, and employees that there were fewer recent stressful life events reported

with increased age.

Ethnicity has also been related to the number of life events experienced.

Coddington (1972b) found in a group of 806 preschoolers, that white children had both a
higher number of life events as well as a higher number of social readjustment units,

although not significantly so. Norris (1992) studied 1,000 adults and found that exposure

to life events over the life span was higher among whites than among blacks.


In summary of this section on child life events stress and demographics the

following conclusions offered:

(a) Education, unemployment, single motherhood, parental age, and ethnicity

were all related to the number of stressful life events the child experiences. When

education levels were high, and/or mothers were employed, and/or when the father was

present, the parental figure was older, and the family ethnic group was minority, the child

tended to experience fewer stressful life events than when educational levels were low,








mothers were unemployed, the father was absent, parent was younger, and family ethnic

group was white.

(b) Parental occupation and place of residence (urban v rural) were also related to

the child's stressful life events. The more prestigious the occupation of the parent the

lower the number of stressful life events for the child. Children who lived in urban areas

experienced more stressful life events than children who lived in a rural environment.

(c) The studies in this section were conducted with early elementary school age

children and to a lesser extent with preschoolers, so it is not known if the relationship

between child stress and these demographics is present to the same degree in preschoolers.

Parenting Stress and Demographic Variables
Studies of parenting stress have focused on the quality of the mothers' engagement

with the child under different environmental conditions. Seven studies, one with parents of

school age children, three with preschoolers and three with infants will be reviewed in this

section. As has been mentioned previously, there are several slightly different ways used

in the field to measure parenting stress.

A longitudinal study on poor urban black mothers was reported on by Thompson

and Ensminger (1989) examining psychological well-being. At the first interview, 1,241

mothers of first graders participated. At the ten-year follow-up, 76% were reinterviewed.

Some differences were found between the mothers who were reinterviewed and those who

were not on mothers' age (young mothers refused more often), geographic mobility, and

parochial school attendance. There were no differences on original well-being, early family

income, welfare status, or composition of the household. The measures included: a global

sadness and tension scale, items specifically related to recent psychiatric symptoms, five

stressful life events, social support, and social and background characteristics including age

of mother, number of children in home, adult household composition, education level,

employment, and poverty status.








At time 1, psychological well-being was not significantly related to the number of

adults in the home. However education, age, and employment had significant relationships

with mothers' well-being. At time 2 however, the presence of another adult in the home

was significantly related to well being. Education and employment continued to be related

to well-being. Stressful life events had a significant relationship with mothers' well being

at time 2. Social support (e.g., close friend, church attendance) appeared to mediate this

stress on mothers' psychological well-being.

Parenting stress was explored in a sample of 85 African American low-income

Head Start mothers by Hastings-Storer (1991). One purpose of this study was to compare

the scores of these mothers to the original norm group of the Parenting Stress Index

(Abidin, 1983), which was predominantly white middle class. In comparing the

experimental group to the norm group, there was a significant difference on stress levels

between groups on the demographic variables: race, marital status, educational level,

employment status, income level of household, educational level of father, and employment

status of father. Between the experimental and norm groups, there was a significant

difference on PSI total stress and for each of the PSI Child Domain Stress scores (overall

and for all Child Domain subscales). This was also true between groups on the Parent

Domain Stress overall, and the subscales: Attachment, Sense of Competence, Social

Isolation, and Relationship with Spouse. No differences were found for Depression,

Restriction of Role, or Parent Health.

Three study or focus groups were formed into the categories of Normal,

Borderline, and High Stress from among the experimental mothers, using cutoff scores

defined by Abidin (1990). Significant differences among these groups were found on the

demographic variables: income level of the household, educational level of the mother,

employment status of the mother, and marital status. There was a significantly higher

percentage of mothers in the high stress group with household incomes below $10,000.

The relationship between educational level of mother and stress group shows that mothers








with lower levels of education had higher amounts of parenting stress. A significantly

greater percentage of mothers who were unemployed were in the high stress group. A

significantly greater percentage of the never married and separated/widowed/divorced

mothers were in the borderline or high stress groups compared to the married mothers.

There were no significant differences for age of mother, age of father, age of index child,

or number of children in home.

Focusing on "chronic" stressors involves studying demographic variables to a large

degree. Conger, McCarty, Yang, Lahey, and Kropp (1984) studied chronic stressors and

the influence of psychological characteristics as mediators of mothers' parental teaching

actions with their preschool child in a sample of mothers. Subjects were 74 families who

were selected to be heterogeneous on income, race, and number of children per families.

While attempting to draw a heterogeneous sample, the authors described their final sample

as having below-average incomes, above-average number of children, and as being

disproportionately nonwhite in terms of the national standard; but this was not unusual in

the rural South where the research took place. Two subsets of participants were those

studied in a preschool school-based setting and those studied in their homes. Within each

setting some families had been identified by a county agency as being physically abusive to

one or more children in the home. The demographic conditions as measures of chronic

stress were: (a) financial stress (income and dependence on public support); (b) family

structure (number of children and single-parent head of household); and (c) past events that

may be associated with continuing stressful life conditions (educational achievement and

mother's age at first birth).

Results indicated that the subgroups of blacks and whites, abusive and nonabusive,

mothers of children in the school-based preschool and mothers of children at home, older

and younger children, and male and female children were not significantly different from

one another. Positive relationships between psychological risk for elevated levels of

aversive maternal behavior, depressed rates of supportive interactions, and increasing








environmental stress (public assistance and number of children) were found in all cases.

Further, with only one exception (negative perceptions and number of parents), elevated

levels of aversive maternal behavior and depressed rates of supportive interactions were

inversely related to decreasing stress (i.e., income, education, age at first birth, number of

parents in the home). Almost 53% of the variation in the psychological measures was

accounted for by these demographic circumstances. A negative relationship between

increasing environmental stress and positive maternal behaviors was observed. Decreasing

environmental stress was positively related to supportive maternal interactions. The

demographic predictors which were the most consistent for maternal behavior were age of

mother at first child's birth and mother's educational level.

Using hierarchical regression, the psychological variables accounted for about

14.5% of maternal behavior variance when entered first into the prediction equation. From

25% to almost 37% of the variance in maternal behavior was accounted for when the set of

environmental stressors (demographics) was entered first. However even after the

environmental stressors were controlled, a significant relationship was present between the

mothers' psychological characteristics and the mothers' behaviors. This taps the multi-

dimensional definition of parenting stress in which both psychopathological and situational

variables contribute to parenting stress.

In a sample of 312 women who had recently become mothers, Turner and Noh

(1983) studied the link between social class and psychological vulnerability among women.

Social class was determined by husband's occupation. Psychological distress and stressful

life events for mothers were measured. The authors found social class to be significantly

related to life stress. Further, regression showed that "within the lower class, stress

appears to be translated into distress at a rate approximately 1.6 times that observed in the

remainder of the population" (p. 6). The results of this section support the findings of

Roghmann, Hecht, and Haggerty (1975; cited in Webster-Stratton, 1990) who reported








that the incidence of major stressors was two to four times greater for poor or lower-class

families than for middle-class families.

Abidin (1990) who constructed the Parenting Stress Index, states that mothers who

are young tend to have higher Parenting Stress Index scores, in the Parent Domain in

particular. Belsky, Lerner, and Spanier (1984) cited statistics from the Alan Gutthmacher

Institute (1981), which found that the majority of pregnancies of teenage mothers were

unplanned. Younger (1984; cited in Abidin, 1990) discovered when studying new mothers

that the scores on the Parenting Stress Index were significantly higher with both unplanned

and unwanted pregnancies.

As mentioned above, Hastings-Storer (1991) compared the scores on the Parenting

Stress Index of black Head Start mothers with the scores of the norm group of the PSI who

were mainly white middle-class females. She found the black mothers to be significantly

higher on parenting stress on the Total score, the Child Domain score, and the Parent

Domain score. Of the thirteen factors which make up the PSI, there were only three on

which there were not significant differences between the two ethnic groups. Millar (1986;

cited in Abidin, 1990) found significant differences for ethnicity on the PSI also on the

Total, Child Domain, and Parent Domain scores when studying black Bermudian mothers

and using a cutoff score of stress at or above the 90th percentile.

Summary

The following remarks can be made in summary of this section on parenting stress

and demographics:

(a) All five of the demographic variables, parental educational level, parental

employment status, parental marital status, parental age, and ethnicity were related to

parenting stress.

(b) Of these demographic variables, educational level was the most consistent in

its relationship to parenting stress. The higher the mother's education, the lower her

parenting stress. Having more education may mean less stress, while there are sometimes








situations when mothers work or are in a relationship which brings increased stress into the

parent-child relationship. Age may function in a similar way with respect to parenting

stress than it does to life events stress, that is being older was related to experiencing less

stress, while being of a minority ethnic background was associated with more parenting

stress.

(c) The demographic variables of income and occupation were also related to

parenting stress. The higher the income and prestige of the job, the lower the level of

parenting stress.

(d) All but one of the studies in this section were conducted with children below

school age and thus have good generalizability to preschoolers and point to a relationship

between the sociodemographic indicators and the amount of stress present in the parent-

child relationship. However, the problem of agreeing on an operational definition of

parenting stress somewhat cloud these findings.

Concluding Statements
A strong relationship between child stressful life events (child stress), parent-child

relationship stress (parenting stress), and child verbal cognitive functioning has been

reported in numerous studies with middle and low income children who have already

entered school. Far less is known about these relationships in preschool children,

however, especially children who are socioeconomically low income. It is within this

group, most at risk for experiencing stress, that research needs to be done. Wyman et al.'s

(1991) comment is illustrative:

These findings suggest that a rich understanding of resilient outcomes under
stressful life conditions must look beyond child qualities, i.e., it must also
reflect the family context in which the child develops and the matrix of
his/her relationships with primary caregivers. One factor that may potentiate
negative developmental outcomes for children who grow up under stress
may be the disruption of supports and structures that facilitate adaptive
caregiver-child interactions. Conversely, an environment that facilitates
consistent, positive attachment to a caregiver may provide resources that
enable the child to adapt well in spite of such stress. Although it remains
for future studies to clarify precisely how such relationships enhance a
child's ability to cope, our findings suggest that this may happen in different
ways at different developmental stages. (p. 423)








With this research history in mind, a number of overall conclusions may be derived

from the literature reviewed in this chapter. The review of the literature examining the

relationships between child stress, parenting stress, child cognitive functioning (the major

variables in this study) and the demographic variables revealed a total of forty-four studies.

The breakdown of these studies is as follows: seven studies with all three major variables;

three studies of child stress and child cognitive functioning only; three studies of child life

events stress and parenting stress only; two studies of parenting stress and child cognitive

functioning only. This yields a total of fifteen studies of the three major variables examined

both together and in pairs. In addition, there are twenty-nine studies of the demographic

variables of maternal education level, maternal employment status, maternal marital status,

maternal age, and ethnicity as related to the three major variables.

1. From this literature, it can be concluded that there is at least a moderate

inverse relationship in the range of -.02 to -.33 between the stressful life events a child

experiences and his/her cognitive performance. This relationship is mediated by the

interaction between parent and child in two ways. First, high stress in the parent-child

relationship is related to a higher number of stressful child life events. Correlations range

from .11 to .56. Secondly, after high stress events occur, high stress in the parent-child

relationship is related inversely to lower cognitive functioning in the child. These

correlations are in the range of -.04 to -.36.

2. In the fifteen studies in this literature review which examined child stress,

parenting stress, and child cognitive functioning, more than twice as many studies have

been conducted with school age children than with preschool age children. Only four

studies were found which used preschool children exclusively as subjects. These four

studies indicated that preschool children also experienced stress and that such stress was

inversely related to their verbal cognitive performance. Correlations are a bit lower in the

range of -.02 to -.26. These correlations are influenced partially by the fact that the number

of life events was partly a function of age, with younger children having experienced fewer








events than older children. The direction of the relationship is the same as that seen with

older children.
When the mothers of these children are experiencing high stress, the quality of the

parent-child relationship suffers. This in turn is related both (a) positively to the amount of
stress the child experiences (correlations in the range of .11 to .32) and (b) negatively to the

childs' verbal cognitive functioning (correlations between -.04 to -.28). These correlations

are again slightly lower than when older children are included. The directions of these
relationships are similar to those found with school age children and their parents. That is,

when child life events stress and parenting stress are high, the child's verbal cognitive

functioning is lower than when child stress and parenting stress are low.
However, all four of these studies with preschoolers (Bee et al., 1982; Murphy &

Moriarty, 1976; Pianta et al., 1990; and Prince et al., 1972) have at least one of the

following methodological problems: (a) using a symptoms measure of stress (more

subjective) rather than a life events measure of stress (more objective) for both the mother

and the child; (b) using unreliable instrumentation to measure stress; (c) collecting data
about the child's stressful life events by word of mouth rather than using a formal measure;

(d) assessing the stressful life events of the "family" rather than events specific to the

developmental period of the preschool child and to the context of parenting. This does not

yield the most sensitive measure of child stress and parenting stress; (e) not including

characteristics of the child (e.g., temperament) when measuring parenting stress. The

contribution of the child is important in understanding the total amount of stress present in

the parent-child relationship; and (f) not including poverty level families as subjects,
limiting generalizability to this social group. Stress research in general has found higher
levels of negative life events stress and parenting stress in low-income families than in
higher income families. There is substantial research documenting lower cognitive

functioning in lower income children compared to higher income children.








3. One group of preschoolers who may be particularly susceptible to stress

influencing verbal cognitive functioning is low-income children in compensatory education

programs. Only the Prince et al. (1972) study which had both low- and high-income

subjects included low-income children in compensatory education programs. The children

were exposed to either DISTAR (a Head Start model) or Sesame Street. While significant

relationships were found between child stress, parent stress, and child verbal cognitive

functioning in the low income families, unreliable instrumentation was used to measure

stress. The measures of stress (both completed by the mother) were not independent as

thought to be. The stress symptoms reported by the mother for her child were heavily

influenced by her own reported stress symptoms. The stress measures were of symptoms

of stress which is more subjective than reporting life events. Further, the parent stress
measure did not include the child's characteristics along with the parent's characteristics,

both found to be important in accurately assessing the amount of stress present in the

parent-child relationship.

4. Certain demographic variables have been found to be related to stress and

the child's verbal cognitive functioning. This literature review indicates that across both
low and high income groups, the five sociodemographic variables of maternal educational

level, maternal employment status, maternal marital status, maternal age, and maternal

ethnicity facilitate coping or promote vulnerability in both children and their parents. They

relate to (a) the number and type of life events occurring in the family, and (b) the

availability of personal and social resources for coping with stress. Findings indicate that

maternal educational level has an inverse relationship with child stress and parenting stress

and a positive relationship with child verbal cognitive functioning. Studies on employment

reveal that when mothers are employed, child stress and parenting stress are lower and

child verbal cognitive functioning is higher. When mothers have a spouse or significant

other in the home, child stress and parenting stress are lower and child verbal cognitive

functioning is higher. When mothers are older, child stress and parenting stress are lower








and child verbal cognitive functioning is higher. Finally, when mothers are of a minority

ethnic background, parenting stress appears to be higher; when they are white, child stress

is higher, but so is child verbal cognitive functioning.

Less is known about how these demographic variables relate to stress resiliency and

vulnerability in children and their families within the exclusively low socioeconomic group

than is known about differences between socioeconomic groups. This is especially true

with preschoolers and their families. Twenty-nine studies were done which examined

demographics in relation to the three major variables of child stress, parenting stress, and

child cognitive functioning separately. Of these, thirteen were conducted with preschoolers.

Of the thirteen, ten used both low and middle/high or just middle/high income families. Just

three of the thirteen used an exclusively lower class group.

Two studies (Conger et al., 1984; Hastings-Storer, 1991) which focused on

exclusively low-income families were studies of parenting stress. There was no study

found which examined child life events stress in relation to the demographic variables.

Only one study was found using a low-income sample which looked at child cognitive

functioning and a demographic variable, this was age (Maracek, 1980; cited in Fein & Fox,

1990).

5. Methodological flaws were also found in the literature on school age children,

though they were less numerous than those with preschoolers. These were: (a) assessing

child stress and parenting stress retrospectively (5 to 7 years) to the preschool years; and

(b) measuring school success (e.g., classroom problem behaviors and competencies,

learning problem, good student, peer sociability) rather than cognitive functioning. School

success can be influenced by social skills, making it difficult to clearly identify the child's

level of cognitive functioning instead of his/her social/emotional functioning.

6. No studies were found which measured child stressful life events, parent-
child relationship stress, and child verbal cognitive functioning in low-income preschoolers
involved in early education compensatory programs and their mothers (or fathers). Only








two studies (Pianta et al., 1990 and Hastings-Storer, 1991) were found which used an

exclusively low-income group of preschoolers and their mothers. The focus of stress in

both of these studies was that being experienced by the mother. In the case of Pianta et al.,

mother's stress was correlated with child verbal cognitive functioning.

This lack of attention is surprising since children experience events just as adults do

and low-income children are the group considered most at-risk for problems in verbal

cognitive functioning in school. Only one study (Prince et al., 1972) measured child

stress symptoms, parent stress symptoms, and child verbal cognitive functioning in low

income children who were also in compensatory early education programs. As previously

mentioned, this study compared low-income children in compensatory programs with

middle-income children in preschool programs and had several major methodological

flaws.














CHAPTER 3
METHODOLOGY

Purpose

The purpose of this study was to examine the relationships between child verbal

cognitive functioning, child life events stress, and parent-child relationship stress in a

sample of low-income preschoolers and their mothers (or mothering one) in early

childhood compensatory education programs. In addition, the influence of the following

demographic variables was examined: mother's education level (number of years of school

completed), mother's employment status (employed or unemployed), mother's marital

status (married or unmarried), mother's ethnicity (minority or white), and mother's age

(continuous).

Research Questions

The following questions were posed in examining these relationships.

1. Is there a relationship between child verbal cognitive functioning and child life

events stress, parent-child relationship stress, maternal education level, maternal marital

status, maternal employment status, maternal ethnicity, and maternal age?

2. Is there a relationship between child verbal cognitive functioning, child life

events stress and parent-child relationship stress, after controlling for the maternal

demographic variables of education level, marital status, employment status, ethnicity, and

age?

3. Is there a relationship between child verbal cognitive functioning and the

maternal demographic variables of education level, marital status, employment status,

ethnicity, and age, after controlling for child life events stress and parent-child relationship

stress?








4. Is there a relationship between child life events stress and parent-child

relationship stress, after controlling for the maternal demographic variables of education

level, marital status, employment status, ethnicity, and age?

5. Is there a relationship between child life events stress and the maternal

demographic variables of education level, marital status, employment status, ethnicity, and

age, after controlling for parent-child relationship stress?

6. Is there a relationship between parent-child relationship stress and the

maternal demographic variables of education level, marital status, employment status,

ethnicity, and age, after controlling for child life events stress?

Hypotheses

The following hypotheses were tested to answer the research questions posed:

1. There will be a significant relationship between child verbal cognitive

functioning, child life events stress, parent-child relationship stress, maternal education

level, marital status, employment status, ethnicity, and age.

2. There will be a significant relationship between child verbal cognitive

functioning, child life events stress, and parent-child relationship stress, after controlling

for the maternal demographic variables of education level, marital status, employment

status, ethnicity, and age.

3. There will be a significant relationship between child verbal cognitive

functioning and the demographic maternal variables of education level, marital status,

employment status, ethnicity, and age, after controlling for child life events stress and

parent-child relationship stress.

4. There will be a significant relationship between child life events stress and

parent-child relationship stress, after controlling for the maternal demographic variables of

education level, marital status, employment status, ethnicity, and age.








5. There will be a significant relationship between child life events stress and the

maternal demographic variables of education level, marital status, employment status,

ethnicity, and age, after controlling for parent-child relationship stress.

6. There will be a significant relationship between parent-child relationship stress

and the maternal demographic variables of education level, marital status, employment

status, ethnicity, and age, after controlling for child life events stress.

Population
The population of parents studied were all mother-child dyads enrolled in the

Alachua County Head Start program during the 1993-1994 school year. For purposes of

examining the generalizability of the study it should be noted that Alachua County, Florida

has a large group of children and families living in poverty. The Gainesville Metropolitan

Statistical Area (GMSA) is ranked as the fourth lowest in family income in the United

States, with 23.6% of all families who live in Alachua County living below the poverty

level. In the county, 47 out of every 1,000 families with children are headed by females

mid live in poverty. There were 1,022 children in Alachua County on child care waiting

lists as of September 1991. Less than a third of the parents were employed full-time

(Alachua County School Board, 1993).

In the United States, Alachua County is ranked 97th out of 150 counties cited as

hunger and poverty counties by the Physician's Task Force on Hunger in America. In

Alachua County, approximately 42% of all students in the public schools receive free or

reduced breakfast and lunch. In several elementary schools, the percentages are near 80%.

Alachua county makes up 1.4% of the total population in Florida, but has 2.1% of the

state's monthly average number of families on public assistance (Alachua County School

Board, 1993).

Educationally in Alachua County, two-thirds of those who receive the public

assistance of Aid to Families with Dependent Children (AFDC) do not have high school








diplomas. An estimated 13% of the adults in the county have less than an eighth-grade

education (Alachua County School Board, 1993).

Looking specifically at the 1992-1993 Head Start population in Alachua County,

the population of eligible Head Start parents consists of 873 families. Of these parents,

653 were female and 220 were male. Sixty-eight had a 9th grade or less education, 93 had

a 10th grade education, none had just an 11th grade education, 465 had a 12th grade

education, 14 had a GED, 134 had some college or training, and 81 were college or

training graduates. For employment, 395 were full-time, 15 were part-time, four were

seasonal, 446 were unemployed, and 10 were retired or disabled (Alachua County Head

Start Summary Demographic Report 1992-1993, 1993).

Of these 873 who were eligible, 732 families actually participated in the Head Start
program. Of these, 456 were eligible for Medicaid, and 380 were Aid For Families With

Dependent Children (AFDC) recipients. With regard to parent status, 522 of the families

were headed by one parent, 158 by two parents, 4 by a foster parent, and 48 by a non-

parent. Of the 732 children, 522 were black, 10 were Hispanic, 1 was Native American,

14 were Pacific Islanders, and 185 were white. There were 372 female children and 360

male children (Alachua County Head Start Summary Demographic Report 1992-1993,

1993).

Subjects
Subjects were randomly drawn from a pool of approximately 653 female parents
participating in Head Start for the 1993-1994 school year. A sample of 51 mother-child

dyads were drawn from 6 schools randomly selected from Gainesville City schools. The

subjects were drawn from the five schools with the most students enrolled in Head Start.

The school or one of the schools with the fewest number of students would have served as

an additional pool should the need have arisen, but was not needed. Dyads were excluded

if the child had been diagnosed with or suspected of having a learning disability, mental

retardation, or any other neurological disorder (such as included in Public Law 94-142).








The 1992-1993 school year demographics show that 110 of the 732 Head Start children

had been diagnosed as handicapped. Additionally, one was suspected of being

handicapped, 6 had been referred but no diagnosis had been made, and 615 were non-

handicapped (Alachua County Head Start Summary Demographic Report 1992-1993,

1993).

Mothers (or mothering females) only rather than fathers (or fathering ones) were
asked to participate. According to Abidin (1985, 1990) females are most often the main

caretakers of children. The Parenting Stress Index (PSI; Abidin, 1990) questions are

written slightly more with the mother in mind. In addition this controlled the gender of the

parent as a confounding variable.

Design
This study used a between-subjects correlational design. The independent variables

in this study were: (a) child life events stress; (b) parenting stress-indicating the amount of

stress present in the parent-child relationship; and (c) five demographic variables: maternal

educational level, maternal employment status, maternal marital status, maternal age, and

maternal ethnicity. The dependent variable was the preschool child's verbal cognitive

functioning.

The operational definition of child life events stress was the seven factor scores

from the items on the Coddington Life Events Record-Preschool Version. Parenting stress

was operationally defined as the thirteen factor scores on the Parenting Stress Index.

Maternal educational level was defined as whether the mother had a high school education

or above or below a high school education. Maternal employment status was defined as

employed (full- or part-time) or unemployed at the time of the data collection. Head Start

asks if the home has one or two parents present. Marital status was defined as married or

unmarried. Ethnicity was defined as minority or white. Maternal age was measured

continuously.








In studying the demographic breakdown, it can be seen that in the total Head Start

population, the one-parent and two-parent home groups were disproportionate in size.

There are far more one-parent than two-parent homes. To address this, mothers were

drawn until a ratio of at least 20 two-parent to 30 one-parent homes was reached, so that

these two groups were more evenly sampled. This was important in the study of the

relationship this environmental condition had on stress in parents and children and

children's verbal cognitive functioning. In addition, the design included a reliability test

which used the Cronbach Alpha procedure on the Coddington Life Events Record-

Preschool Version. The results are presented later when the pilot study is discussed.

The dependent variable was the preschool child's verbal cognitive functioning. The

combined score from the Language and Concepts subscales on the DIAL-R served as the

operational definition of child verbal cognitive functioning.

Setting
The mothers' homes were suggested first for the location of data collection. If

mothers felt uncomfortable doing the interview in their homes, the Family Services Center

was available as a meeting place. For those parents who felt uncomfortable meeting either

at their homes or at the Family Services Center, the interview took place at the Head Start

center in which the child attended or the mother's place of employment. If none of these

locations was suitable, a site mutually determined by the researcher and the mother (e.g.,

libraries, churches, shopping malls) would have been chosen, but this situation did not

arise. The investigator encouraged the mother to choose a time of day for the data

collection when activity was likely to be low (e.g., children were in school).

Instrumentation
Child Verbal Cognitive Functioning

In measuring preschool child verbal cognitive functioning (e.g,. speech and

language and conceptual functioning), instruments which tap developmental milestones are

preferred over those which measure rate (Seymour and Wyatt, 1992). Milestones measure








what the child can do, not how quickly or slowly he can do it. This is important because

the rate of language development is highly variable in preschool children, while milestones

are more stable as they are sequential in nature. The Developmental Indicators for the

Assessment of Learning-Revised (DIAL-R) was constructed using the milestones approach

by Mardell-Czudnowski and Goldenberg (1990). The DIAL-R measures typical

developmental behaviors that children ages 2-0 through 5-11 can demonstrate. The

instrument is an individually administered comprehensive screening test which follows an

empirically based developmental sequence of language, conceptual, and motor behaviors.

The test consists of three areas--Motor, Language, and Concepts. Each area has

eight items and the items are subdivided into tasks. For this study, only the Language and

Concepts area scores were used. The DIAL-R is administered each year by the child's

Head Start teacher and/or Head Start Teacher-Student Liaisons. Alachua County Head Start

Administration scores the test and provides feedback to the classroom. Parents are

informed of their child's test results and their meaning from the Head Start Teacher. Test

results are kept in the Head Start classroom and become a part of that child's curriculum for

the year. Administration and scoring of the test takes place in the fall and again in the spring

of the school year. After pre- and post-testing, the scores are placed into the database of

the Alachua County School Board. The fall administration was of interest in this study.

The fall testing period for the 1993-1994 school year was set for late August through

October. Two weeks are allowed for testing to be completed once a classroom receives a

copy of the DIAL-R to allow rotation of the test kits.

Head Start teachers are trained on the use of the DIAL-R with extensive materials

developed by the test authors. The training is conducted by the Early Childhood Program

in Alachua County through a training manual and workshops lasting a full day before

school starts. Trainees learn from the manual the type of test the DIAL-R is, the purpose of

the DIAL-R, time to administer, materials, parents, what happens in each screening area,

motivating a child, participation, and closure. Individuals being trained to give the DIAL-R








must do modeling, role playing, testing over knowledge of how to administer the test, and

are observed while testing children before being allowed to test independently. Once the

individual begins testing, the Teacher-Student Liaison provides on-going training.

Establishing rapport is critical. The authors of the DIAL-R write that the operator

must be sure to have the child's attention before presenting the items. Further, the test

administer's attitude, verbal, and nonverbal communication are influential in getting scores

that are reliable. Operators are instructed to keep a positive and calm attitude regardless of

the child's responses. Finally, all attempts and performance, irrespective of what the child

does or doesn't do, are to be acknowledged with praise, positive body language, and

comments like "good" and "thank you". An example of a DIAL-R item is given here from

the Language section for problem solving:

Item 7. Problem Solving.
a) Say: OK, child's name, what do you do when you're hungry? If the
child does not respond or gives a response that implies there is not food in
the house to eat, prompt the child by asking: What do you WANT to do
when you're hungry? [directions say stop here if child still doesn't answer]
b) If the child did respond, continue by asking the following questions:
What do you do when you want to go into a room that is dark? What do
you do when you want to go outside but it's raining? What do you do when
you break something that belongs to someone else? (p. 13).

The DIAL-R represents one revision done in 1981 to standardize the test on a

national sample, extend the age range downward to include younger two-year-olds, and to

combine gross and fine motor skills into a single motor area. The instrument was

standardized on a national sample of 2,447 children on the variables of sex, geographic

region, size of community, and race. In addition, the norms were reanalyzed in 1990 to

adjust inappropriate scores. Three norm groups are available with which to compare

children. These are a Caucasian sample, a Minority sample, and a Census sample weighted

to resemble 1990 U.S. estimates. Alachua County School System uses the Census norms.

Reliability. Test-retest reliability on the DIAL-R for Concepts is (.90) and for
Language is (.77). Eight sites retested a total of 65 children within five weeks of the

DIAL-R standardization. Internal consistency reliability is reported as (.78) for Concepts








and (.79) for Language using the Minority Norm Sample; (.74) for Concepts and (.73) for

Language using the Caucasian Norm Sample; and (.76) for Concepts and (.72) for

Language using the Census Norm Sample. Interrater reliability is reported for the original

DIAL as between 81 and 99%. The authors report that "because 21 of the 24 DIAL-R

items are identical to, or revisions of, DIAL items, it is likely that a similar study with the
DIAL-R would produce similar results" (Mardell-Czudnowski & Goldenberg, 1990, p.

60).

Validity. Research done at Northwestern University during the DIAL development
was used to predetermine criteria for the selection of tasks to be included and the scoring

criteria. Chosen from an extensive list of possible consultants were eight well respected
professionals. This group conducted the review for the content and for item bias of the

DIAL. Reviewed were the design of the test development, the construction for each item,

and scoring criteria. There was total agreement with respect to the general test development

among all consultants. Only minor revisions based on individual suggestions were made.

The DIAL conceptual model underwent a revision. The original 315 test items were
reviewed with respect to research published since the DIAL was developed as well as

feedback received from DIAL coordinators and operators throughout the country. This

resulted in a reduction of test items from 315 items to 155 items. Field studies yielded data

which clustered most of these items into 31 items in a usable format. These 31 items

formed the normative battery used in standardization testing.

DIAL-R items were analyzed in terms of the abilities they were intended to assess:

perceptual, memory, previous learning association, kinesthetic awareness, coordination,

and language. Most items were multidimensional. The DIAL-R items were also analyzed

to determine their fit with an information-processing model based on input and output of
the visual and auditory systems. To aid in the identification of strong and weak modalities

across the DIAL-R, this model analyzed the items in these terms. In summary, for most
items, even though a child might not understand what was said, he or she sees a








demonstration of what is expected. The output information clearly showed that all

Language items required verbal responses.

The DIAL-R was compared with the Learning Accomplishment Profile-Diagnostic
(LAP-D) on Head Start children (Barnett, Faust, & Samir, 1988) to determine construct

validity. Significant validity coefficient correlations between similar scales on the two

instruments (i.e., motor-motor, concepts-cognitive, and language-language) were obtained.

Concurrent validity (a type of criterion validity) has been demonstrated for the DIAL-R.

This type of validity is concerned with the relationship of the test being validated with a test
with similar criteria given at the same time. Lichenstein (1981) found the DIAL-R to be

significantly correlated with the Denver Developmental Screening Test (DDST), the

Stanford-Binet Intelligence Scale (SBIS), the Peabody Picture Vocabulary Test (PPVT),

and the Woodcock Johnson Psychoeducational Battery (WJPEB).

Mardell-Czudnowski and Goldenberg (1990) also compared the DIAL-R with the

Stanford Binet because, "although the DIAL-R is not intended to be a test of intelligence, it
provides a gross estimate of the level of development of intellectual skills needed to succeed

in kindergarten and first grade" (p. 67). Using cutoff scores based on + and 1.5 standard

deviations, the DIAL-R had an agreement index of 89% with the Stanford Binet. In

another study using an ability and achievement test, the Kaufman Assessment Battery for

Children (K-ABC) correlated with the Total DIAL-R on the K-ABC scales of Mental

Processing Composite, Sequential Scale, and Achievement Scale (Parks-Trace, 1984; cited

in Mardell-Czudnowski & Goldenberg, 1990).

The DIAL-R has been examined for predictive validity (a second type of criterion
validity). Several studies have demonstrated that the DIAL-R is an excellent predictor of

kindergarten performance and achievement (Vilmure, Achenback, Woodard, & Sheehan,

1984; cited in Mardell-Czudnowski & Goldenberg, 1990; Smith, 1986; cited in Mardell-

Czudnowski & Goldenberg, 1990; Jacob, Snider, & Wilson, 1988). (see Appendix A for

the DIAL-R)








Child Life Events Stress

The Coddington Child Life Events Record (1972) has a preschool, child, and

adolescent scale. In a review of child life stress measures, Johnson (1986) stated the Life

Events Record was "the best-known and widely used life stress measure for younger age

groups" (p. 32). The Coddington Life Events Record-Preschool version is unique when

compared to other measures of child life events stress in that the events chosen to be

included on the instrument are events which happen specifically to preschool children as

opposed to children of any developmental period. The Life Events Record yields the

number of life events which have occurred and assigns life change unit scores to reflect

readjustment for each of the 30 events on the list. Alternatively to using a total number of

life events, a factor analysis has been done on the Coddington Life Events Record and

factors which represent qualitatively different types of events are available.

Parents are read the list of events and asked to indicate if this event has occurred in

the life of the child. If the answer is yes, the parent is asked how many times that event has

occurred and the age(s) of the child at occurrence. Examples of items on the Life Events

Record-Preschool Version are: birth of a brother or sister, mother beginning work, loss of

job by a parent, and divorce of parent. This can be done orally with the researcher marking

the responses when the interview is conducted one-on-one. Responses can be written

when interviewing in a group.

The amount of readjustment required for a preschooler on these events was

surveyed from a sample of professional workers which included 131 teachers, 25

pediatricians, and 87 mental health workers in academic divisions of child psychiatry.

There was high agreement among the sample regarding the average degree of readjustment

necessary for each event. The instrument was normed on a sample of 806 preschoolers.

The sample was chosen to be a valid cross-section of the population in and around

Columbus, Ohio with regard to race and socioeconomic class.








The LER yielded seven factors from an analyses by Sandier and Ramsey (1980)

using a group of clinical child psychologists as raters. These factors were Loss, Entrance,

Family Troubles, Positive, Physical Harm, Sibling Problems, and Primary Environment

Change.

Reliability. The test-retest reliability of the Coddington Life Events Record was

conducted with 120 high school football players (Coddington, 1984). The instrument was

completed at three months, seven months, and eleven months. While the reliability was

inversely related to the length of time between administrations; the correlation coefficients

were significant for family event scores, extrafamilial event scores both desirable and

undesirable events, and on the total scores for all three occasions and ranged from (.37) to

(.69). Evidence of reliability was lacking on younger ages. An internal reliability study for

the Coddington Life Events Record-Preschool version was incorporated into a pilot study

for this dissertation study and is described later.

The reliability for the factor analysis described above was .69 for the mean ratings

over 496 scores on the similarity of events. A matrix was formed and then factor analyzed

using principal components. An eigenvalue criterion of 1.00 yielded seven factors which

were rotated to a varimax solution and accounted for 63.6% of the total variance. A factor

loading criterion of .50 or greater was chosen to select items describing each factor.

Validity. Content validity of the LER has been investigated. The events were

drawn from the literature and from work with normal and abnormal children by the author

of the test, who is a professor of Psychiatry and Pediatrics at Ohio State University College

of Medicine and is Director of the Division of Child Psychiatry. In relation to all possible

events which could occur, Coddington (1984) studied the representativeness of a life

events list. The LER was given 724 times to 84 fourth graders and their parents (345 times

to the children, 379 times to the parents). The children and their parents were asked about

other events which may have occurred but were not found on the LER. Of the events

reported by the children, 97% were on the LER. Of the events reported by the parents,








eighty-nine percent were on the LER. No other or additional events were recorded in 84%

of the administrations to parents with the LER.

The degree to which a test measures theoretical constructs on which it is based is

construct validity. Construct validity is determined by test scores being positively

correlated with alternative measures of the construct The LER was compared and

discussed by Coddington (1984) to another measure of life events, the Life Event

Inventory by Van Houten & Golembiewski (1978) with respect to weighting of events. A

high correlation between weights derived by three independent methods supported the

validation of the Coddington.

Criterion validity consists of concurrent and predictive validity. The concurrent

validity of the LER was examined by the agreement present between parent and child on the

fourth grade sample introduced earlier. Coddington remarks that individuals within the

same family are expected to report family events in a similar fashion, keeping in mind

individual perceptions as a source of error. Coddington's research staff (1984) found the

correlations between the LER for the children and their parents to be significant. This held

for total scores, familial events, and extrafamilial events. Coddington believes that the

younger the child the more agreement will be present between the parent and the child.

"Concurrent validity clearly decreases with increasing disparity between ages, or more

correctly, with an increasing degree of individuality" (p. 111).

Concurrent validity is also determined by comparing the scores of the test being

assessed with some other measure administered at the same time. Sander and Block

(1979) studied inner-city school children (K-3) using the LER scale and found that:

children identified by their teachers as experiencing adjustment problems
experienced more stressful events in the preceding year than a matched
sample of normal controls. Within a subsample of the school maladjusted
group, life stress was also significantly correlated with parent ratings of
child adjustment problems. (p. 436)

Using this same sample and same measure of adjustment, Sandier and Ramsey
(1980) found the maladaptive group to be significantly different from the controls on the