You are on page 1of 16

Pergamon

Child Abuse & Neglect 6 (2001) 787 802

Antecedents and socioemotional consequences of physical


punishment on children in two-parent families
Mary Keegan Eamon*
School of Social Work, University of Illinois at UrbanaChampaign, 1207 West Oregon Street,
Urbana, IL 61801, USA
Received 7 March 2000; received in revised form 13 October 2000; accepted 21 October 2000

Abstract
Objective: The main objective of this study was to test a structural model of the antecedents and
socioemotional consequences of mothers use of physical punishment on children in two-parent
families.
Method: Mother-child data from the National Longitudinal Survey of Youth, based on a sample of
1397 4- to 9-year-old children, were used to test a structural model derived by the author from
previous research. The hypothesized model was revised; the revised model was cross validated on a
split-half sample, and estimated separately by age group, gender, and race/ethnicity.
Results: The revised model fit the data well and was supported by cross-validation. Poverty, maternal
birth age, parents education, maternal depression, and marital conflict were directly or indirectly
related to mothers frequent use of physical punishment. Frequent use of physical punishment was
directly related to childrens socioemotional problems, as were maternal depression and marital
conflict. Few subgroup differences were found.
Conclusions: Main findings indicate that the effect of poverty on mothers use of physical punishment
is indirect, and is mediated by maternal depression and marital conflict. Depressed mothers spank their
children more frequently and experience higher levels of marital conflict, which, in turn, is directly
related to their use of physical punishment. Younger, more educated mothers spank their children less
often. Children who are spanked more frequently exhibit more socioemotional problems. 2001
Elsevier Science Ltd. All rights reserved.
Keywords: Physical punishment; Socioemotional problems; Depression; Marital conflict; Poverty

* Corresponding author.
0145-2134/01/$ see front matter 2001 Elsevier Science Ltd. All rights reserved.
PII: S 0 1 4 5 - 2 1 3 4 ( 0 1 ) 0 0 2 3 9 - 3

788

M.K. Eamon / Child Abuse & Neglect 6 (2001) 787 802

Introduction
Children who are harshly or physically punished are more likely to experience socioemotional
problems than are children whose parents use other forms of discipline (DeaterDeckard, Dodge,
Bates, & Pettit, 1996; Larzelere, 1986; McLeod & Shanahan, 1993; Strassberg, Dodge, Pettit, &
Bates, 1994; Straus, Sugarman, & GilesSims, 1997; Turner & Finkelhor, 1996). These problems
include antisocial behaviors, low self-esteem, internalizing symptoms (such as anxiety,
withdrawal, and depression), and externalizing behaviors (such as aggression, disobedience,
and impulsiveness). Childhood effects of physical punishment may carry over into adulthood
in the form of depression, suicide, alcohol abuse, and family violence (Straus, 1991; Straus
& Kantor, 1994). Parents use of physical punishment also may increase the potential for
child abuse (Whipple & Richey, 1997; Whipple & WebsterStratton, 1991).
Previous research indicates that cultural, demographic, marital, psychological, and socioeconomic factors predict parents use of physical or harsh punishment (Conger et al., 1993;
DeaterDeckard et al., 1996; McLeod & Shanahan, 1993; Straus, 1991), which is linked to
childrens socioemotional problems. The main objective of the present study was to test a
structural model of the antecedents and socioemotional consequences of mothers use of
physical punishment on a national sample of 4- to 9-year-old children in two-parent families.
In this study, antecedents of physical punishment included poverty, maternal birth age and
education, fathers education, maternal depression, and marital conflict. The socioemotional
consequences of physical punishment, as well as of maternal depression and marital conflict,
were measured by childrens externalizing and internalizing problems. The hypothesized
relations among these variables were derived from the literature review that follows.
Influences of household and parental characteristics
A number of household and parental characteristics, including poverty, maternal birth age,
and mothers and fathers education, have been shown to influence maternal depression,
marital conflict, and mothers use of physical punishment. Income or absence of financial
strain, age, and years of education are all inversely related to adult depressive symptoms
(Aneshensel, Rutter, & Lachenbruch, 1991; Blazer, Kessler, McGonagle, & Swartz, 1994;
Bruce, Takeuchi, & Leaf, 1991; Kessler et al., 1994; Muntaner, Eaton, Diala, Kessler, &
Sorlie, 1998). Stress caused by higher exposure to discrete life events (such as crime
victimization, family illness, and job loss) and chronic strains (such as inability to fulfill
family role obligations and responsibilities for young children) is experienced more frequently by poor or young parents. High levels of stress seem to at least partially explain
relations between economic hardship and age and maternal depression (Belle, 1990; McLeod
& Kessler, 1990; Ross & Huber, 1985; Turner, Wheaton, & Lloyd, 1995). Education is a
coping resource (Kessler, 1982; Ross & Mirowsky, 1989); it facilitates effective problem
solving, thus reducing the risk of depression.
Some studies suggest that more educated couples have fewer marital conflicts and
negative interactions than do less educated couples (Harrell, 1990; Stets, 1997). Couples with
more education may be more likely to use nonconflictual methods to resolve differences.
Alternatively, negative reactions to low-status occupations or social positions held by less

M.K. Eamon / Child Abuse & Neglect 6 (2001) 787 802

789

educated spouses may spill over into their marital relationship. Older, more educated
mothers are more likely to be emotionally mature and knowledgeable of positive parenting
practices, and thus use physical punishment less frequently than younger, less educated
mothers; however, research findings are inconsistent (cf., DeaterDeckard et al., 1996;
McLeod & Shanahan, 1993; Straus, 1991).
Influences of maternal depression and marital conflict
Maternal depression is related to marital conflict, negative parenting practices, and
childrens socioemotional problems (Conger et al., 1993; Davies, Dumenci, & Windle, 1999;
Gelfand & Teti, 1990; Harnish, Dodge, & Valente, 1995; Miller, Cowan, Cowan, Hetherington, & Clingempeel, 1993). Maternal depression appears to be directly related to childrens socioemotional problems through a genetic link, or children might imitate their
mothers depressive symptoms, such as facial expressions and social withdrawal. Maternal
depression also is related indirectly to childrens socioemotional problems, by escalating
marital conflict and increasing parenting behaviors that are hostile, emotionally unresponsive, inattentive, and physically abusive (Cummings & Davies, 1994; Dodge, 1990; Downey
& Coyne, 1990; Gelfand & Teti, 1990; Turner & Finkelhor, 1996). Children can respond to
hostile and aggressive parenting behaviors with fear, withdrawal, or by developing insecure
parent-child attachments. Maternal use of physical punishment might also increase childrens
externalizing behaviors, by legitimizing violence or modeling aggressive behaviors.
Marital conflict is also directly and indirectly linked to childrens socioemotional problems (Buehler et al., 1997; Emery, Fincham, & Cummings, 1992; Engfer & Schneewind,
1982; Harold, Fincham, Osborne, & Conger, 1997; Katz & Gottman, 1993; Rogers, 1996;
WebsterStratton & Hammond, 1999). Marital conflict might have a direct influence on
childrens internalizing symptoms, because children react with fear or anxiety; they might
worry about the stability of their parents relationship or blame themselves for the conflict.
Children can externalize behaviors by imitating parents conflict resolution styles (Davies &
Cummings, 1994; Fincham, 1998). Marital conflict indirectly affects childrens socioemotional functioning by disrupting effective parenting practices. Parents might withdraw from
or reject the child, or allow aggression to spill over into the parent-child relationship, such
as in the form of physical punishment (Engfer & Schneewind, 1982; Erel & Burman, 1995;
Fauber, Forehand, Thomas, & Wierson, 1990; Harrist & Ainslie, 1998).
Influences of childs age, gender, and race or ethnicity
Although the research literature offers no definitive findings, the antecedents and socioemotional consequences of physical punishment, maternal depression, and marital conflict
might vary by the childs age, gender, or race/ethnicity (Cummings & Davies, 1994; Emery,
1982). Younger childrens adjustment might have a stronger relation with parental behaviors,
because younger children are dependent upon their parents and have few coping skills. On
the other hand, older children are more likely to be aware of marital conflict and maternal
depression, and might feel pressured to become involved in the conflict or to assist the

790

M.K. Eamon / Child Abuse & Neglect 6 (2001) 787 802

Fig. 1. Conceptual model of the antecedents and socioemotional consequences of physical punishment on children
in two-parent families.

depressed parent. Although the use of physical punishment declines after age 4 (Straus, 1991),
whether physical punishment differentially affects children at various ages is unknown.
Reviews and meta-analyses of studies that have examined gender differences in the
developmental influences of parental depression, marital conflict, and parenting practices
find contradictory evidence regarding whether boys or girls are more vulnerable (Erel &
Burman, 1995; Reid & Crisafulli, 1990; Rothbaum & Weisz, 1994). Gender-linked vulnerabilities appear to depend on the childs age. Boys are more susceptible to family risk factors
during childhood, and girls are more susceptible during adolescence (Davies & Windle,
1997; Luthar, 1999). Nonetheless, no clear relation between gender and marital conflict,
parenting behaviors, and child adjustment has been established (Snyder, 1998).
Whether developmental processes vary by race/ethnicity is also controversial (cf., Ogbu,
1981; Rowe, Vazsonyi, & Flannery, 1994), but influences of specific parenting practices on
children might be relative to particular cultures. DeaterDeckard et al. (1996), for example,
found that mothers use of physical punishment predicted the externalizing behaviors of
White, but not Black, children. The researchers speculated that the effects of a particular
parenting practice can be contingent upon its prevalence and acceptance, the manner in
which it is used, or the childs perception of its appropriateness. Other studies, however,
report relations between physical punishment and childrens socioemotional problems for
both White and Black children (McCabe, Clark, & Barnett, 1999; Nix et al., 1999). Despite
such inconsistencies, research suggests that study findings might not generalize to children
of different age, gender, and race/ethnicity.
Consistent with current research, a conceptual model of the antecedents and socioemotional consequences of physical punishment on children in two-parent families appears in
Fig. 1. Poverty, maternal birth age, and parents education are shown to exert direct and/or
indirect influences on mothers use of physical punishment, and indirect influences on
childrens socioemotional problems. As indicated in the path diagram, the effect of poverty
on physical punishment is mediated by maternal depression and marital conflict. Maternal
depression and marital conflict mediate the effects of maternal birth age and mothers

M.K. Eamon / Child Abuse & Neglect 6 (2001) 787 802

791

education on physical punishment, and are also directly related to physical punishment.
Fathers education influences maternal use of physical punishment indirectly, by affecting
marital conflict. Physical punishment, marital conflict, and maternal depression were hypothesized to directly affect childrens socioemotional problems.
As shown in Fig. 1, maternal depression, marital conflict, and childrens socioemotional
problems were conceptualized as latent variables; they are represented by ovals in the path
diagram and measured by multiple indicators. A latent variable is defined by the nonerror
variance of its indicators. The remainder of the variables in the model, represented by
rectangles, were measured by single indicators and assumed to be measured without error.

Method
Data and sample
Data were extracted from the mother-child data set of the National Longitudinal Survey of
Youth (NLSY). The initial survey of 12,686 individuals was conducted in 1979. Data from the
1992 and 1994 survey years were used for this study. Only during these years were respondents
asked questions related to their depressive symptoms and the quality of their marital relationships.
Assessments of the female respondents children began in 1986. By 1994, 10,042 children were
born to the original 6,283 NLSY females. Approximately 91% of eligible women were interviewed in 1992 and 1994, and valid child assessment scores were obtained for approximately
90% of available children (Center for Human Resource Research, 1997).
The sample for this study included 4- to 9-year-old children who lived in intact families
in 1992 or 1994. Children younger than age 4 and older than age 9 were not included in the
study, because younger children were not assessed with the Behavior Problem Index (BPI,
measures of the dependent variable), and mothers report spanking older children with low
frequencies. Children remained in the sample if, in the most recent of the years, they were
assessed with the Home Observation for Measurement of the Environment-Short Form
(HOME-SF, measure of physical punishment) and the BPI, and their mothers answered at
least one question related to the quality of their marital relationship and one question on the
Center for Epidemiology Studies Depression (CES-D) scale. One child was randomly
selected from multiple-children families, and children who had more than two missing BPI
items were removed from the sample. The resulting sample of 1397 4- to 9-year-old children
included 246 Black, 317 Hispanic, and 834 non-Hispanic, White children.
Measures
Poverty. Family income the year before the survey was inflated to 1994 dollars and compared
to the official poverty threshold for the family size measured at the interview date. If total
family income for a given family size fell below the official threshold, the child was
categorized as poor (poverty 1), and 0 otherwise. Following the method of McLeod and
Shanahan (1993), if family income was missing (16% of respondents had missing income data)
the child was coded as poor if the mother reported receiving public assistance or Food Stamps

792

M.K. Eamon / Child Abuse & Neglect 6 (2001) 787 802

during the year. If the mother reported receiving neither of these sources of income, the child was
coded as nonpoor. After these imputations, only four respondents had missing income data.
Maternal birth age and parents education. Maternal birth age and the number of years of
education parents attained by the assessment year were measured as continuous variables.
Maternal depression. Mothers depressed mood was assessed using a shortened version of
the 20-item CES-D scale. Although the CES-D was not developed to diagnose clinical
depression, it can distinguish whether individuals are clinically depressed (Radloff & Locke,
1986). In 1992, the CES-D was administered to NLSY respondents; in 1994, a shortened,
7-item version of the scale was developed and administered. The shortened scale had an
alpha reliability of .76 and correlated .92 with the full CES-D (Ross & Mirowsky, 1989). The
seven common items administered in the two survey years were used in this analysis.
Respondents were asked to rate on a scale from 0 to 3 how many days during the past
week they had problems with seven depressive symptoms (0 rarely/none of the time or 1
day; 1 some/a little of the time or 12 days; 2 occasionally/moderate amount of the time
or 3 4 days; 3 most/all of the time or 57 days). The four indicators of maternal
depression used in this analysis were determined by a Linear Structural Relations 8 (LISREL
8, Joreskog & Sorbom, 1996a) confirmatory factor analysis. Items included I had trouble
keeping my mind on what I was doing (Keep Mind); I felt that everything I did was an
effort (Effort); I felt sad (Sad); and I could not get going (Get Going). Cronbachs alpha
for maternal depression .72.
Marital conflict. Measures of marital conflict were determined by a LISREL confirmatory
factor analysis, using 14 questions on the quality of the respondents marital relationship.
Indicators were how frequently the couple argued about chores, showing affection, children,
and money (1 never; 2 hardly ever; 3 sometimes; 4 often). Alpha .75.
Physical punishment. One item from the HOME-SF, a modification of Caldwell and Bradleys (1979) HOME inventory, measured physical punishment: the number of times the mother
reported spanking her child in the past week (0 none; 1 1; 2 2; 3 more than 2).
Socioemotional problems. The construct childrens socioemotional problems was measured
by two indicators, externalizing and internalizing behaviors. Items from the BPI measured
these two types of behaviors. The BPI was developed by Zill and Peterson of Child Trends,
Inc., who derived many of the items from Achenbach and Edelbrocks Behavior Problems
Checklist (Center for Human Resource Research, 1994). Mothers rated as often true,
sometimes true, or not true the occurrence in the previous 3 months of 30 common child
behaviors. BPI items were re-coded so that higher scores represent more behavior problems.
Factor analysis reveals that BPI items can be represented as two subscales, measuring a
childs tendency to internalize or externalize behaviors (Center for Human Resource Research, 1994). In this study, a principal components analysis determined 16 indicators of
externalizing behaviors (such as disobedience, impulsiveness, cruelty, bullying, stubbornness, and destructiveness), and 14 indicators of internalizing problems (such as worrying,

M.K. Eamon / Child Abuse & Neglect 6 (2001) 787 802

793

withdrawing, crying, anxiousness, sadness, and dependency). The 29 children who had fewer
than three missing BPI items were assigned the mean of their own values for available items
representing internalizing and externalizing behaviors, Cronbachs alpha .86 and .81, for
externalizing and internalizing behaviors, respectively. Externalizing and internalizing behaviors were measured as the sum of the items representing each type of behavior. Because
boys tend to have more externalizing behaviors than girls, and older children exhibit more
behavior problems than younger children (Center for Human Resource Research, 1994),
externalizing and internalizing scores were standardized by sex within each age group. A
preliminary analysis demonstrated that mediating variables were related to both types of
behaviors identically; therefore, socioemotional problems were measured as two indicators,
externalizing and internalizing behaviors.
As discussed above, missing information for poverty and externalizing and internalizing
behaviors were imputed. Although a number of other variables had missing data (ranging
from 2.2% to .1% of the sample), data were not imputed; the structural models were
estimated with pair-wise present cases. Models were also estimated using cases with no
imputed income and BPI data. Signs and statistical significance of path coefficients were
identical to estimates reported in the Results section.
Data analysis
Structural equation models consistent with Fig. 1 were estimated by LISREL 8. A one-half
split sample (n 698) was used to test and modify the hypothesized structural model, and
a one-half split sample (n 699) to test the stability of the revised model. Because the
majority of the variables in this analysis violate the linear structural equation modeling
assumptions that variables are continuous and measured on at least an interval scale, the
preprocessor PRELIS 2 (Joreskog & Sorbom, 1996b) was used to compute polychoric/
polyserial correlation matrices on pair-wise present cases in the split samples. PRELIS 2
adjusts the categories of an ordinal variable to thresholds of an underlying, normally
distributed, continuous variable. Four fit indexes were used to evaluate model fit. They
include the 2 goodness-of-fit statistic, root mean square residual (RMR), adjusted goodness
of fit index (AGFI), and comparative fit index (CFI). Good fitting models have low or
nonsignificant 2 values. Because the 2 statistic is biased upwards as samples get larger, a
suggested rough rule of thumb for good fitting models is a ratio of the 2 to the degrees of
freedom of less than 2. Values of .05 or less for the RMR, and values greater than .90 for
the AGFI and CFI indicate good fitting models. Relative to the other indexes, the CFI is
insensitive to sample size (Ullman, 1996).

Results
Revised model estimates
The results of the hypothesized structural model depicted in Fig. 1, with the exception of
the high 2 ratio, indicated an acceptable fit with the data (2[76, maximum N 698]

794

M.K. Eamon / Child Abuse & Neglect 6 (2001) 787 802

Fig. 2. Standardized path coefficients for the structural model of the antecedents and socioemotional consequences of physical punishment on children in two-parent families.

262.10, p .001, 2 ratio 3.45, RMR .036, AGFI .931, and CFI .943). The results
indicated a negligible path coefficient (B .013) for the relation between maternal education
and marital conflict, and this path was eliminated. Although fit statistics for the revised model
were almost identical (2[77, maximum N 698] 262.16, p .001, 2 ratio 3.40,
RMR .036, AGFI .932, and CFI .944), it is more parsimonious. Standardized path
coefficients (maximum likelihood estimates) for the revised model are presented in Fig. 2. As
the figure indicates, path coefficients for relations between maternal birth age and education
and depression also are not significantly different from zero. These coefficients were
estimated in the revised model because coefficients are in the expected direction, and the
relations are supported by past research. As shown in Fig. 2, maternal birth age (B .15),
education (B .19), and depression (B .18), and marital conflict (B .16) were directly
related to physical punishment. These coefficients indicate that mothers who were older, less
educated, and experienced more depressive symptoms and marital conflict used physical
punishment more frequently. Physical punishment, in turn, was directly related to childrens
socioemotional problems (B .20), as were marital conflict (B .21) and maternal
depression (B .29).
The effect of poverty on mothers frequent use of physical punishment was mediated by
its link with maternal depression (B .32), which was directly and indirectly related to
physical discipline (by increasing marital conflict, B .42). Fathers who had more education
had fewer conflicts with their spouses (B .10), which was related to mothers spanking
their children less frequently. With the exception of the positive relation between maternal
birth age and physical punishment, coefficients were in the expected direction.
Standardized direct, indirect, and total effects on physical punishment and childrens
socioemotional problems are summarized in Table 1. A direct effect is a relation between two
variables, unmediated by a third variable (e.g., the relation between physical punishment and
childrens socioemotional problems); an indirect effect is the relation between two variables
mediated by one or more other variables (e.g., the effect of poverty on physical punishment,
mediated by maternal depression and marital conflict). A total effect is the sum of the indirect
and direct effects. As indicated in Table 1, total effects of all antecedent variables on physical

M.K. Eamon / Child Abuse & Neglect 6 (2001) 787 802

795

Table 1
Standardized direct, indirect, and total effects on physical punishment and childrens socioemotional problems
Variable

Poverty
Maternal birth age
Mothers education
Fathers education
Maternal depression
Marital conflict
Physical punishment
R2

Physical punishment

Socioemotional problems

Direct
effect

Indirect
effect

Total
effect

Direct
effect

Indirect
effect

Total
effect

.145**
.190**

.175**
.163**

.078**
.003
.014
.016*
.069**

.144

.078**
.142**
.204**
.016*
.244**
.163**

.291**
.210**
.203**

.137**
.024
.062*
.024*
.138**
.033*

.279

.137**
.024
.062*
.024*
.429**
.243**
.203**

Note: Dashes indicate coefficient was not estimated.


* p .05; ** p .01 for two-tailed tests.

punishment were statistically significant, but the effect of fathers education (B .016)
was trivial. The effect of poverty was relatively small (B .078), compared to maternal
depression, education, and birth age (B .244, B .204, and B .142, respectively). In
addition to the direct effect of physical punishment (B .203) on childrens socioemotional
problems, total effects of poverty (B .137), maternal depression (B .429), and marital
conflict (B .243) were statistically and practically significant. The relatively large total
effects of depression on physical punishment and childrens socioemotional problems indicate the importance of maternal depression on the use of physical punishment and childrens
socioemotional problems.
Cross-validation and group comparisons
The results of the model on the cross-validation sample, with the exception of the lower
but still significant 2, indicated an acceptable fit with the data (2[77, maximum N 699]
280.95, p .001, 2 ratio 3.65, RMR .044, AGFI .923, and CFI .929). All but
one of the statistically significant path coefficients presented in Fig. 2 cross-validated (the
coefficient for the relation between marital conflict and socioemotional problems).
To determine if the revised model varied by age group, gender, and race/ethnicity, the
model was estimated separately for children 4 to 5, 6 to 7, and 8 to 9 years of age, and for
male, female, Black, Hispanic, and White children. Fit indexes for the revised model,
estimated on the split samples and multiple groups, are presented in Table 2. Because of the
small sample sizes for the multiple groups, cross-validation models were not estimated.
Although multisample procedures available in LISREL allow testing for statistical differences between groups, they require the use of covariance, not correlation, matrices. These
procedures, therefore, were not used, and any differences between group models should be
considered as only suggestive. As indicated in Table 2, the CFI, which is relatively unaffected by sample size, indicated at least a marginally acceptable fit for all groups. Fit indexes,
however, generally indicated that the model fit the younger age group samples slightly better

796

M.K. Eamon / Child Abuse & Neglect 6 (2001) 787 802

Table 2
Fit Statistics for models of the antecedents and socioemotional consequences of physical punishment by
group
Sample

2(77)

2/df

RMR

AGFI

CFI

Revised model
Cross-validation
Children 45 years old
Children 67 years old
Children 89 years old
All females
All males
All Black children
All Hispanic children
All White children

698
699
463
406
528
667
730
246
317
834

262.16*
280.95*
199.07*
181.16*
278.98*
359.86*
182.86*
188.82*
231.71*
266.34*

3.40
3.65
2.59
2.35
3.62
4.67
2.37
2.45
3.00
3.46

.036
.044
.043
.041
.043
.042
.036
.069
.052
.037

.932
.923
.921
.917
.904
.903
.949
.861
.863
.942

.944
.929
.945
.941
.913
.907
.954
.901
.902
.945

Note: RMR root mean square residual; AGFI adjusted goodness of fit index; CFI comparative fit index.
* p .001.

than the 8- to 9-year-old sample; the model fit males better than females, and White children
better than Black and Hispanic children.
In every group sample, factor loadings for indicators of latent constructs were very
similar, and physical punishment was positively related to childrens socioemotional problems. Consistent with model results on the 21split sample, significant total effects of poverty,
maternal depression, and marital conflict on the use of physical punishment and on childrens
socioemotional problems were found in all group samples, with the exception of the Black
sample. In the Black sample, the total effects of only maternal depression and physical
punishment on socioemotional problems were statistically significant, results which may be
because of the small sample size (n 246). Significant total effects of maternal birth age and
parents education were less consistent across group samples.
Coefficients for the total effects of poverty, maternal depression and marital conflict on
physical punishment, and the effect of physical punishment on childrens socioemotional
problems indicated a few notable group differences. The effect of depression on spanking
was larger for older children (B .20 and B .29 for children 6 7 and 8 9 years old,
respectively) than for 4- to 5-year-old children (B .13), as was the effect of spanking on
socioemotional problems (B .16, B .30, and B .25 for children 4 5, 6 7, and 8 9
years old, respectively). The total effects of poverty, maternal depression and marital conflict
on the use of physical punishment, and the effect of physical punishment on childrens
socioemotional problems for males and females, and White and Hispanic children were
almost identical. The main exception was the larger effect of poverty on mothers frequent
use of spanking on girls (B .11) than on boys (B .05).

Discussion
The purpose of this study was to test a structural model of the antecedents and socioemotional consequences of physical punishment on a national sample of 4- to 9-year-old

M.K. Eamon / Child Abuse & Neglect 6 (2001) 787 802

797

children in two-parent families. The revised model fit the data well, and supported most of
the hypothesized relations between multiple antecedents and the socioemotional consequences of physical punishment, as well as of maternal depression and marital conflict. The
model cross-validated, and can be generalized to 4- to 9-year-old, male and female, and
Hispanic and White children. Findings for the Black sample were less consistent.
The studys findings contribute to understanding the family processes, and the relative
importance of parental and household characteristics that place children at risk for physical
punishment and externalizing and internalizing behavior problems. Consistent with previous
research and stress theory, poverty was positively related to maternal depression, which had
both direct and indirect influences on mothers frequent use of physical punishment and on
childrens socioemotional problems. Although physical abuse was not measured in this
study, these findings are consistent with theory and past research that suggest stress mediates
the effect of poverty on child abuse (Pelton, 1994; Whipple & WebsterStratton, 1991;
Wolfner & Gelles, 1993). This study also demonstrates the indirect effect of poverty,
suggesting that the use of traditional multivariate regression techniques, in which variables
related to poverty (such as financial stress or depression) are placed in models simultaneously, may mask the effect of income on child abuse (e.g., Cadzow, Armstrong, & Fraser,
1999).
The relatively large standardized total effects of maternal depression on physical punishment (B .244) and childrens socioemotional problems (B .429) indicate the important
influence of depression on mothers frequent use of spanking and on childrens socioemotional problems. The irritability and hostility commonly accompanying depressive symptoms
likely account for the relation between mothers depression and their frequent use of physical
punishment. The direct link between maternal depression and childrens socioemotional
problems might reflect genetic influences, or the childrens imitating or internalizing their
reactions to their mothers symptoms. Maternal depression also was related to frequent use
of physical discipline by influencing marital conflict. Spousal conflict might spill over into
mothers disciplinary practices; mothers divert their anger and the conflict away from the
spousal relationship and toward the child.
Consistent with past research, both marital conflict and physical punishment were linked
to childrens socioemotional problems. By their own conflictual and physically punishing
behaviors, parents might legitimize or provide role models for childrens externalizing
behaviors such as aggression or bullying (Straus, 1991). Children also can internalize their
reactions to physical punishment and parental conflict, and become sad, fearful, or withdrawn. Relations between maternal depression and marital conflict and maternal use of
physical punishment are also consistent with past studies that find relations between maternal
depression and marital conflict and harsh or abusive parenting practices (Conger et al., 1993;
Engfer & Schneewind, 1982; Whipple & WebsterStratton, 1991).
Fathers education played a minor role compared with other factors that resulted in
mothers frequent use of physical punishment and childrens socioemotional problems. More
educated mothers were less likely to use physical punishment (B .190), suggesting that
knowledge of alternative child disciplinary practices influences mothers use of physical
punishment. The positive relation between maternal birth age and physical discipline is not
theoretically consistent, but is consistent with another NLSY study (McLeod & Shannon,

798

M.K. Eamon / Child Abuse & Neglect 6 (2001) 787 802

1993). These findings might reflect the relatively young age of NLSY mothers. In this study,
maternal birth age ranged from 18 to 33 years.
Regardless of the childs age, gender, or race/ethnicity, physical punishment was related
to childrens socioemotional problems. The effect of depression on physical punishment and
the effect of physical punishment on socioemotional problems were larger for older than
younger children. Depressed mothers might be more likely to strike out at older children.
Because parental use of spanking declines with the childs age, older children might view
physical punishment as a more aversive experience, thus having more deleterious effects.
The total effects of poverty, maternal depression, and marital conflict on the use of physical
punishment, with the exception of the Black sample, were also consistent in all groups. These
findings suggest that household and parental factors that increase the likelihood that mothers
will use physical punishment can be generalized to most children in this age range. Why so
few relations were statistically significant in the Black sample is unclear, but may be because
of the small sample size. The larger effect of poverty on mothers use of physical punishment
for girls than boys suggests that same gender parent-child relations might be more adversely
affected by stressors, such as those associated with poverty.
This study has a number of limitations. Among them are the absence in the NLSY data
of measures of fathers depressive symptoms and disciplinary practices. Relying solely on
mothers reports of depressive symptoms, marital conflict, physical punishment, and childrens socioemotional problems may have introduced unmeasured bias. Mothers might not
accurately report their marital problems, spanking behaviors, and childrens behavior problems. Causal ordering among the variables was not established. For example, maternal
depression might interfere with the mothers ability to contribute to the familys economic
well being, not vice versa. Bidirectional influences between variables in the model, such as
between marital conflict and childrens socioemotional problems, likely exist but were not
estimated.
Despite these limitations, these results have practice and social policy implications.
Consistent with findings of previous national studies (e.g., Straus, 1991), physical punishment was a common practice in the present sample. Almost 35% of mothers in this sample
reported spanking their children during the previous week. Considering the lack of evidence
that physical punishment deters unwanted behavior better than other methods, educating
parents on alternative ways of discipline and the possible risks of using physical punishment
seem to be appropriate prevention and intervention strategies. This research suggests that
education programs, even if they were effective, would only partially reduce the use of
physical punishment and childrens socioemotional problems. Maternal depression and
marital conflict are important factors in influencing mothers use of physical punishment and
childrens socioemotional problems. Professionals who treat children with socioemotional
problems might find that assessing maternal depression and marital conflict provides valuable information for choosing appropriate interventions. The role that poverty plays in this
process cannot be neglected. The findings suggest that social programs that increase families
financial resources, but do not add stigma or additional stress, might indirectly decrease
physical punishment and childrens socioemotional problems. Whether these findings can be
generalized to child abuse is a future research question.

M.K. Eamon / Child Abuse & Neglect 6 (2001) 787 802

799

Acknowledgments
The author thanks Rachel Zuehl for her research assistance, and Sandra Kopels for her
helpful comments on an earlier version of this paper.

References
Aneshensel, C. S., Rutter, C. M., & Lachenbruch, P. A. (1991). Social structure, stress, and mental health:
competing conceptual and analytic models. American Sociological Review, 56, 166 178.
Belle, D. (1990). Poverty and womens mental health. American Psychologist, 45, 385389.
Blazer, D. G., Kessler, R. C., McGonagle, K. A., & Swartz, M. S. (1994). The prevalence and distribution of
major depression in a national community sample: the national comorbidity survey. American Journal of
Psychiatry, 151, 979 986.
Bruce, M. L., Takeuchi, D. T., & Leaf, P. J. (1991). Poverty and psychiatric status. Archives of General
Psychiatry, 48, 470 474.
Buehler, C., Anthony, C., Krishnakumar, A., Stone, G., Gerard, J., & Pemberton, S. (1997). Interparental conflict
and youth problem behaviors: a meta-analysis. Journal of Child and Family Studies, 6, 223247.
Cadzow, S. P., Armstrong, K. L., & Fraser, J. A. (1999). Stressed parents with infants: reassessing physical abuse
risk factors. Child Abuse & Neglect, 23, 845 853.
Caldwell, B. M., & Bradley, R. H. (1979). Home observation for measurement of the environment. Center for
Child Development and Education. Little Rock, AK: University of Arkansas.
Center for Human Resource Research. (1994). NLS users guide 1994. Columbus, OH: Author.
Center for Human Resource Research. (1997). 1994 NLSY79 child and young adult data: users guide. Columbus,
OH: Author.
Conger, R. D., Conger, K. J., Elder, G. H., Lorenz, F. O., Simons, R. L., & Whitbeck, L. B. (1993). Family
economic stress and adjustment for early adolescent girls. Developmental Psychology, 29, 206 219.
Cummings, E. M., & Davies, P. T. (1994). Maternal depression and child development. Journal of Child
Psychology and Psychiatry, 35, 73112.
Davies, P. T., & Cummings, E. M. (1994). Marital conflict and child adjustment: an emotional security
hypothesis. Psychological Bulletin, 116, 387 411.
Davies, P. T., Dumenci, L., & Windle, M. (1999). The interplay between maternal depressive symptoms and
marital distress in the prediction of adolescent adjustment. Journal of Marriage and the Family, 61, 238 255.
Davies, P. T., & Windle, M. (1997). Gender-specific pathways between maternal depressive symptoms, family
discord, and adolescent adjustment. Developmental Psychology, 33, 657 668.
DeaterDeckard, K., Dodge, K. A., Bates, J. E., & Pettit, G. S. (1996). Physical discipline among African
American and European American mothers: links to childrens externalizing behaviors. Developmental
Psychology, 32, 10651072.
Dodge, K. A. (1990). Developmental psychopathology in children of depressed mothers. Developmental Psychology, 16, 3 6.
Downey, G., & Coyne, J. C. (1990). Children of depressed parents: an integrative review. Psychological Bulletin,
108, 50 76.
Emery, R. E. (1982). Interparental conflict and the child of discord and divorce. Psychological Bulletin, 92,
310 330.
Emery, R. E., Fincham, F. D., & Cummings, E. M. (1992). Parenting in context: systemic thinking about parental
conflict and its influence on children. Journal of Consulting and Clinical Psychology, 60, 909 912.
Engfer, A., & Schneewind, K. A. (1982). Causes and consequences of harsh parental punishment: an empirical
investigation in a representative sample of 570 German families. Child Abuse & Neglect, 6, 129 139.
Erel, O., & Burman, B. (1995). Inter-relatedness of marital relations and parent-child relations: a meta-analytic
review. Psychological Bulletin, 118, 108 132.

800

M.K. Eamon / Child Abuse & Neglect 6 (2001) 787 802

Fauber, R., Forehand, R., Thomas, A. M., & Wierson, M. (1990). A mediational model of the impact of marital
conflict on adolescent adjustment in intact and divorced families: the role of disrupted parenting. Child
Development, 11121123.
Fincham, F. D. (1998). Child development and marital relations. Child Development, 69, 543574.
Gelfand, D. M., & Teti, D. M. (1990). The effects of maternal depression on children. Clinical Psychology
Review, 10, 329 353.
Harnish, J. D., Dodge, K. A., & Valente, E. (1995). Mother-child interaction quality as a partial mediator of the
roles of maternal depressive symptomatology and socioeconomic status in the development of child behavior
problems. Child Development, 66, 739 753.
Harold, G. T., Fincham, F. D., Osborne, L. N., & Conger, R. D. (1997). Mom and dad are at it again: adolescent
perceptions of marital conflict and adolescent psychological distress. Developmental Psychology, 33, 333
350.
Harrell, W. A. (1990). Husbands masculinity, wifes power, and marital conflict. Social Behavior and Personality, 18, 207216.
Harrist, A. W., & Ainslie, R. C. (1998). Marital discord and child behavior problems: parent-child relationship
quality and child interpersonal awareness as mediators. Journal of Family Issues, 19, 140 163.
Joreskog, K. G., & Sorbom, D. (1996a). LISREL 8: users reference guide. Chicago, IL: Scientific Software
International.
Joreskog, K. G., & Sorbom, D. (1996b). Prelis 2: users reference guide. Chicago, IL: Scientific Software
International.
Katz, L. F., & Gottman, J. M. (1993). Patterns of marital conflict predict childrens internalizing and externalizing
behaviors. Developmental Psychology, 29, 940 950.
Kessler, R. C. (1982). A desegregation of the relationship between socioeconomic status and psychological
distress. American Sociological Review, 47, 752764.
Kessler, R. C., McGonagle, K. A., Zhao, S., Nelson, C. B., Hughes, M., Eshleman, S., Wittchen, H., & Kendler,
K. S. (1994). Lifetime and 12month prevalence of DSMIIIR psychiatric disorders in the United States.
Archives of General Psychiatry, 51, 8 19.
Larzelere, R. E. (1986). Moderate spanking: model or deterrent of childrens aggression in the family? Journal
of Family Violence, 1, 2736.
Luthar, S. S. (1999). Poverty and childrens adjustment. Thousand Oaks, CA: Sage.
McCabe, K. M., Clark, R., & Barnett, D. (1999). Family protective factors among urban African American youth.
Journal of Clinical Child Psychology, 28, 137150.
McLeod, J. D., & Kessler, R. C. (1990). Socioeconomic status differences in vulnerability to undesirable life
events. Journal of Health and Social Behavior, 31, 162172.
McLeod, J. D., & Shanahan, M. J. (1993). Poverty, parenting, and childrens mental health. American Sociological Review, 58, 351366.
Miller, N. B., Cowan, P. A., Cowan, C. P., Hetherington, E. M., & Clingempeel, W. G. (1993). Externalizing in
preschoolers and early adolescents: a cross-study replication of a family model. Developmental Psychology,
29, 318.
Muntaner, C., Eaton, W. W., Diala, C., Kessler, R. C., & Sorlie, P. D. (1998). Social class, assets, organizational
control and the prevalence of common groups of psychiatric disorders. Social Science and Medicine, 47,
20432053.
Nix, R. L., Pinderhughes, E. E., Dodge, K. A., Bates, J. E., Pettit, G. S., & McFadyenKetchum, S. A. (1999).
The relation between mothers hostile attribution tendencies and childrens externalizing behavior problems:
the mediating role of mothers harsh discipline practices. Child Development, 70, 896 909.
Ogbu, J. U. (1981). Origins of human competence. A cultural-ecological perspective. Child Development, 52,
413 429.
Pelton, L. H. (1994). The role of material factors in child abuse and neglect. In G. B. Melton, & F. D. Barry (Eds.),
Protecting children from abuse and neglect (pp. 131181). New York: Guilford Press.
Radloff, L. S., & Locke, B. A. (1986). The community mental health assessment survey and the CES-D scale.

M.K. Eamon / Child Abuse & Neglect 6 (2001) 787 802

801

In M. M. Weissman, J. K., Myers, & C. E. Ross (Eds.), Community surveys of psychiatric disorders (pp.
177189). New Brunswick, NJ: Rutgers University Press.
Reid, W. J., & Crisafulli, A. (1990). Marital discord and child behavior problems: a meta-Analysis. Journal of
Abnormal Child Psychology, 18, 105117.
Rogers, S. J. (1996). Marital quality, mothers parenting and childrens outcomes: a comparison of mother/father
and mother/stepfather families. Sociological Focus, 29, 325340.
Ross, C. E., & Huber, J. (1985). Hardship and depression. Journal of Health and Social Behavior, 26, 312327.
Ross, C. E., & Mirowsky, J. (1989). Explaining the social patterns of depression: control and problem solving or
support and talking? Journal of Health and Social Behavior, 30, 206 219.
Rothbaum, F., & Weisz, J. R. (1994). Parental caregiving and child externalizing behavior in nonclinical samples:
a meta-analysis. Psychological Bulletin, 116, 5574.
Rowe, D. C., Vazsonyi, A. T., & Flannery, D. J. (1994). No more than skin deep: ethnic and racial similarity in
developmental process. Psychological Review, 19, 396 413.
Snyder, J. R. (1998). Marital conflict and child adjustment: what about gender? Developmental Review, 18,
390 420.
Stets, J. E. (1997). Status and identity in marital interaction. Social Psychology Quarterly, 60, 185217.
Strassberg, Z., Dodge, K. A., Pettit, G. S., & Bates, J. E. (1994). Spanking in the home and childrens subsequent
aggression toward kindergarten peers. Development and Psychopathology, 6, 445 461.
Straus, M. A. (1991). Beating the devil out of them. New York: Lexington Books.
Straus, M. A., & Kantor, G. K. (1994). Corporal punishment of adolescents by parents: a risk factor in the
epidemiology of depression, suicide, alcohol abuse, child abuse, and wife beating. Adolescence, 29, 543561.
Straus, M. A., Sugarman, D. B., & GilesSims, J. (1997). Spanking by parents and subsequent antisocial behavior
of children. Archives of Pediatrics & Adolescent Medicine, 151, 761767.
Turner, J., & Finkelhor, D. (1996). Corporal punishment as a stressor among youth. Journal of Marriage and the
Family, 58, 155166.
Turner, R. J., Wheaton, B., & Lloyd, D. A. (1995). The epidemiology of social stress. American Sociological
Review, 60, 104 125.
Ullman, J. B. (1996). Structural equation modeling. In B. G. Tabachnick, & L. S. Fidell (Eds.), Using multivariate
statistics (pp. 709 812). New York: Harper Collins.
WebsterStratton, C., & Hammond, M. (1999). Marital conflict management skills, parenting style, and earlyonset conduct problems: processes and pathways. Journal of Child Psychology and Psychiatry, 40, 917927.
Whipple, E. E., & Richey, C. A. (1997). Crossing the line from physical discipline to child abuse: how much is
too much? Child Abuse & Neglect, 21, 431 444.
Whipple, E. E., & WebsterStratton, C. (1991). The role of parental stress in physically abusive families. Child
Abuse & Neglect, 15, 279 291.
Wolfner, G. D., & Gelles, R. J. (1993). A profile of violence toward children: a national study. Child Abuse &
Neglect, 17, 197212.

Resume
Objectif: Le but principal de cette etude fut de verifier la validite dun mode`le structurel portant sur
les antecedents et les consequences sociales et emotives qui sensuivent lorsque des me`res ont recours
a` des methodes de discipline corporelle.
Methode: Un echantillon de 1.397 enfants ages entre 4 et 9 ans, obtenu par le biais dune enquete
longitudinale nationale de jeunes, a servi dans lapplication du mode`le, lequel avait ete elabore par les
auteurs dans des recherches anterieures. Le mode`le hypothetique a ete revise; on a verifie ce nouveau
mode`le en comparant une moitie de lechantillon avec lautre selon lage, le sexe et la race/ethnie.
Resultats: Le mode`le revise se preta tre`s bien aux donnees et la verification a produit des resultats
positifs. La pauvrete, lage maternel a` la naissance, la scolarite des parents, la depression maternelle

802

M.K. Eamon / Child Abuse & Neglect 6 (2001) 787 802

et les conflits de couple etaient associes directement et/ou indirectement a` lusage de punitions
corporelles frequentes par la me`re. En revanche, on a note un lien entre ces punitions et les proble`mes
sociaux et emotionnels des enfants, la depression maternelle et les conflits du couple. On a remarque
peu de differences entre les divers groupes.
Conclusions: Letude indique un lien indirect entre les effets de la pauvrete et les punitions
corporelles lorsquil y a depression maternelle et conflits au niveau du couple. Les me`res deprimees
frappent leurs enfants plus souvent et connaissent des niveaux de conflit marital plus eleve, ce qui en
revanche les ame`nent a` utiliser des punitions corporelles. Les jeunes me`res mieux instruites frappent
leurs enfants moins souvent. Les enfants qui sont frappes frequemment demontrent des troubles
socio-affectifs.

Resumen
Objetivo: El principal objetivo de este estudio fue probar un modelo estructural de los antecedentes
y consecuencias socioemocionales del uso del castigo fsico en madres con familias biparentales.
Metodo: Se utilizaron los datos de madre-hijo de una Encuesta Nacional Longitudinal sobre Juventud
con una muestra de 1,397 ninos de 4 a 9 anos de edad para probar un modelo estructural derivado por
el autor de investigaciones anteriores. Se reviso el modelo de la hipotesis, el modelo revisado fue
validado con una muestra dividida por la mitad y estimada por separado en grupos de edad, genero
y raza/etnicidad.
Resultados: El modelo revisado se relaciono bien con los datos y fue comprobado por la validacion
cruzada. La pobreza, la edad de la madre al nacimiento, la educacion parental, la depresion materna,
y los conflictos maritales estaban directamente y/o indirectamente relacionados con el uso frecuente
del castigo fsico en la madre. El uso frecuente del castigo fsico estaba directamente relacionado con
los problemas socioemocionales de los ninos, como son la depresion maternal, y los conflictos
maritales. Se encontraron pocas diferencias en los subgrupos.
Conclusiones: Los principales hallazgos indican que el efecto de la pobreza en el uso del castigo
fsico en la madre es indirecto y esta mediado por la depresion maternal y los conflictos maritales. Las
madres deprimidas golpean a sus hijos mas frecuentemente y viven niveles de conflicto marital mas
altos, los que, a su vez, estan directamente relacionados con su uso del castigo fsico. Las madres mas
jovenes y mas educadas golpean a sus hijos menos frecuentemente. Los ninos que son golpeados mas
frecuentemente exhiben mas problemas socioemocionales.