Professional Documents
Culture Documents
Straus1979 (Teori)
Straus1979 (Teori)
Nut-ray A. Straus
University of New Hampshire, Durham, N.H. 03824, USA
ABSTRACT
Data on 1,146 families with a child age 3 through 17 at home are presented.
Child abuse information was obtained for a randomly selected child in each
family. Child abuse was defined as an attack by a parent involving punching,
kicking, biting, hitting with an object, “beating up,@ or using a knife or
gun. Over 14 out of every 100 American children 3-17 are subjected to
abusive violence each year. These figures are at least 26 times greater than
those of the National Center For Child Abuse and Neglect. Even so, for
reasons outlined in the paper, they are underestimates. Families in which
child abuse occurred are compared with other families. The results suggest
that child abuse is brought about by the very nature of the society and its
family system. This has profound implications for the prevention and
treatment of child abuse. Although psychotherapy may be appropriate in some
cases, a more fundamental approach lies in such things as a more equal
sharing of the burdens of child care, replacement of physical punishment with
non-violent methods of child care and training, eliminating the stresses and
insecurity which now characterize our economic system, and strengthening the
ties of individual families to the extended family and the community.
What can cause a parent to punch, kick, bite, burn, or stab a child? The
causes are complex in at least two ways.
*This paper is part of the Family Violence Research Program at the University
of New Hampshire. The Program is supported by NIMH grants T32 HH15161 and
1827557, and by the University of New Hampshire. A list of Program
publications is available on request.
I am grateful to Sieglinde Fizz for the care and skill with which she
arranged the computer typesetting of this article, to Shari Hagar for the
same qualities in her work on the innumberable computer runs for the
statistical analysis, and to Arnold Linsky for insightful comments and
suggestions on an earlier draft.
213
214 M. A. Straus
Both of these complicating factors must be kept in mind when evaluating what
has been written about the causes of child abuse because neither is usually
considered. Instead, there is a tendency to focus on one or two factors as
though they offered the key to explaining this baffling phenomenon. Even
when several factors are considered, the effects of the "interactive
combinationsi* of these factors is not. These criticism also apply to this
paper. I am stating them at the beginning to alert readers to the
limitations of what is to be presented.
Interviews were conducted with the father in a random half of the families,
and with the mother in the other half of the families. In each family, the
data on physical violence was obtained for only one child, and only for the
parent who was interviewed. When there was more than one child, the
"referent child" for the study was selected by a random number table.
The interview covered a great many aspects of family pattern8 and life
circumstances. It included the Conflict Tactics Scales (Straus, 1979). The
items in those scales which pertain to severe violence by the parent toward
the referent child were combined to form a Child Abuse Index. These are all
the items which refer to violence more severe than pushing, shoving,
slapping, and throwing things. Specifically, the list consists of whether,
Family Patterns and Child Abuse 215
during the 12 months up to the interview, the Parent had ever kicked, bit,
punched, hit mith an object, beaten up the child, or used a knife or gun (in
the sense of having actually tried to stab or shoot the child). A parent who
did any of these things was counted as having abused the child.
The rates of child abuse revealed by this method are truly astounding. Each
year, over 14 out of every TOO American parents of a child 3 through 17 is
violent enough to a child to be included in our Child Abuse Index. Thfs
means that of the 46 million children of this age group in the United States,
approximately six and a half million are abused each year.
Tt, might be objected that this index uses too lax a definition of child abuse
because one of the items is “hitting with an object.” For some parents that
could be the traditional strap or paddle rather than an out of control
assault. So we reoomputed the index leaving out the data on hitting with
objeots. The rates drop sharply to “only” three or four out of every hundred
Parents) and to an estimate of 1.7 million children per year.
The data just presented might overstate the amount of child abuse because a
family is included if even one isolated incident of abusive violence occurred
during the year. On the other hand, these rates may understate the extent to
which children are severely assaulted by their parents because the figures do
not take into account how often such assaults occurred. The answer to this
question is that ff one assault occurred, several were likely. In fact, a
single inrfdent occurred in only six percent of the child abuse cases. Mean
number of assaults per year was 10.5 and the median 4.5.
(1) THE NCCAN figures are based on incidents which come to official
attention. This leaves out the vast number of cases in which physical abuse
is suspected and not reported, as well as the equally vast number of cases in
which a child is injured but there is no suspicion of abuse.
(21 Probably the most important reason why our rates are so much higher is
that our data is based on violent acts carried out, rather than on injuries
produced. Fortunately, children are resilient. Many is the child who has
been thrown against a wall and who simply bounces off with at most a bruise.
Only the relatively rare instances fn which a concussion occurs stand even a
chance of being suspected of parental abuse.
fn describing the sample, I said that the omission of children in the high
abuse risk age of birth to age three is one of several factors which make
even our very high rates of child abuse an underestimate. There are several
other factors whioh push in the same direction. The second such factor is
that these are self-reports by parents to a stranger doing a survey. Not
every parent who had punched or kicked a child is going to admit that In such
an interview. Third) the Conflict Tactics Scales include only a limited list
Of all the possible abusive acts. For example, we omitted burning a child,
wiping out the child’s mouth with soap or more noxious substances, and sexual
abuse. Fourth, we interviewed either the father or the mother and have data
only on that person’s abuse of the child, But most children have two parents
and therefore twice the risk --or at least a higher risk --of being abused than
our figures show. A fifth factor making these underestimates is that our
data are based on children living with two parents. The two parents need not
be the child’s natural parents. However, the omission of children living in
216 M. A. Straus
I will start this examination of the social caumea of child abuse with tuo of
the most elementary but also two of the most important characteristics which
are associated with child abuser sex and socioeconomic status.
It is widely known that women are less violent than men. The3 assault and
murder rates of women are a fraction of the rates for men. But in the family
it is different. Using our index of child abuse, the rate for men is 10.1
per hundred childred whereas the rate for women is 75% greater: 17.7 per
hundred children.
I have started with the sfmple fact of the sex of the parent to emphasize the
importance of social factors as compared to psychological factors in
understanding child abuse. These, after all, are the same women who, outside
the family, are much less violent than men. So the reasons underlying their
much greater violence toward their own children is unlikely to be anything in
tile personality or other mental. charaoteristics of women as compared to men.
Rather, the reasans start with the simple fact that husbands and wives do not
have equal responsibility for care of children. The way our type of society
is organized r child care is the responsibility of women, So women are simply
exposed more to both the joys and the trials and tribulations of oaring for
children--they experience more time "at risk,* Moreover, since our sample of
chtldren did not inolude any infants, it is even more clear that this is a
matter of culturally determined rules and arrangements. Fathers are just as
capable-- biologically, even if not by training-- of caring for children of
this age.
*Time at risk,* however is not the whole story, The factors underlying the
much greater frequency of child abuse by mothers go well beyond that. At
least tuo other factors need to be considered. First, it is that mothers who
tend to be blamed if the child misbehaves or does not achieve what is
expected of children at a given age, Since all children misbehave, and since
standards of achievement are ambfguous, almost afl mothers tend to feel
anxiety, frustration, and guilt about their ahildren and their adequacy as
mothers. This is not because women are any more anxiety prooe than men, but
because society sets up a situation in which a high level of anxiety and
frustration is just about inevitable.
A third factor which might account for the higher chil.d abuse rate of women
is that the unequal division of labor and the responsibility for the child's
conduct is not a voluntary choice. These are roles assigned to women by long
historical tradition, and on which most husbands insist. It is no problem
for those women who wish to focus their lives on the role of mother and
housewife. Rut in a society where other opportunities beckon, millions of
women feel blocked and frustrated by the fact that they--not their
husbands--will have the overwhelmlng responsibility for the children.
Most of us are so imbued with the cuI.turai ideology of women as mothers and
housewives that it may be diffioul.t. to see the argument just presented.
There is neither the space nor the evidence to prove tbe point. But there is
a way of getting at this issue indirectly. Ye can compare mothers who are
full time housewives with mothers uho are also employed outside the home.
From one point of view, the highest rate of child abuse should be among women
who have paid employment in addition to their work as mothers and
housekeepers. This is because the research shows that such women continue to
have the major burden of housekeeping and child care. They therefore carry a
Family Patterns and Child Abuse 217
double burden. The opposite point of view is also plausible. Women who have
jobs outside the home may have a Jower rate of child abuse because the hours
the child spends at a nursery or with a baby-sitter reduces their time at
risk, because they can escape the stress and frustration of being employed
full time in a role which is not of their own chOOBing, and because being
employed for vages gives b?0Slen mm power in the family and control over
their own lives.
The results of this study clearly support the secand line of reasoning. The
child abuse rate for women who are full time housewives is half again higher
than the rate for women who, in addition to their responsibilities as
housekeeper and mother, also have a job outside the home f'15.T versus lO.3).
One can also see that the ohild abuse rate for women with paid employment is
almost identical to the rate far men (tO.ff.*l
Socioecow StatuR
One should not be misled by the above it-it.0 thinking of the workplace as
heaven. HiLliOIlS af men and nomen work at boring, demeaning, and underpaid
work. But far women, even these jobs may be a lesser evil than the confines
of the type of family typical aF industrial societies. Even inadequate wages
oan help alleviate the economic stress faced by working class families. For
men I however, we uilf, See that the situetion is quite different,
Our data do not depend on official reports. Yet they reveal. a child abuse
rate in families where the husband is a manual worker which is 41% higher
than for families in which the husband is a white-collar uorker (16.4 versus
11,6). blthaugh this is a substantial difference, it is a much smaller
difference than is typical when officially reported child abuse sates are
used to compare social classes, ThUS, part of the social class difference in
officially reported ohild abuse does seem to be the resuit of biases in the
system of reporting, At the ssme time, the data also suggest that there is a
considerably higher rate of actual child abuse in families in which the
husband is a manual worker.
There are
many factors which could produce the higher rate of chifd abuse
found in
manual worker families v One such factor is the stress of
maintaining a family on a low or sometimes entirely inadequate income. That
this is one of the underlying factors is indicated by the child abuse rate
for families with incomes under $6,000 a year. It is 62% greater than for
other families f21.1 versus 13.4). The same pattern of higher rates low
income families applies to abuse by the f’ather (15.6 versus 9.6) and abuse by
mothers (24.4 versus 16.7).
versus 12.6).
Finally, there are a number of factors on which we do not have data but are
likely to enter the picture. Uorking class parents are known to be more
authoritarian with their children, to have greater faith in physical
punishment as a means of child rearing, and a lesser understanding of child
psychology. In addition, low income areas of American cities have much
higher rates of violence outside the family. Each of these in their own way
are conditions making for parental violence.
Some analysts of child abuse, for example David Gil (1975) write as though
the direct and indirect effects of an unjust and unstable economic system,
and its associated oppression of women and minorities, fully explain the
paradox of child abuse. Unfortunately, the etiology of child abuse is far
more complicated. One can see this from a comparison of child abuse in Black
and in white families. Blacks are one of the most economically and socially
oppressed groups in American society. Yet both this study and some studies
of officially reported child abuse (Billingsley, 1969; Young, 1963) show
that Blacks do n_8fihave a significantly higher rate of child abuse than
whites. Blacks in this sample have a rate of 15.7, which is only 11% greater
than the white rate of 74.1. One reason why Blacks have a rate of child
abuse that is much lower than expected on the basis of their low income, high
unemployment, and rejection by the rest of the society, seems to be the aid
and support, especially in the care of children, provided by Black extended
families (Cazenave and Straus, 1978).
Cazenave and Straus's findings about Black extended families points to the
structure of the family as another important set of factors leading to or
insulating parents from child abuse: the patterns of interaction within the
family.
Years Married
Uife Conflict
Even more important are the tactics used when a couple has a conflict. Some
family therapists argue that the best tactic is to let go and not repress
one's anger. "Don't be afraid to be a real shrew, a real bitch. Tell them
where your really at. Let it be total, vicious, exaggerated, hyperbole..."
as one advocate of this approach put it (Howard, 1970:54). Venting one's
anger in this way is claimed to provide a release from the tension of a
dispute, and therefore helps avoid physical aggression. On the contrary, the
research evidence shows that the more husbands and wives are verbally
aggressive to each other the higher the rate of violence (Straus, 1974). The
main reason for this is that verbal aggression, no matter how emotionally
satisfying it may be, does not come to grips with the substance of the
dispute. Rather, it creates additional animosity which makes it even more
difficult to deal with the original source of the conflict.
Exactly the same results were found in this study for the relation between
verbal aggression and child abuse. Parents who were verbally aggressive to
the referent child.3 have a child abuse rate which is six times that of other
parents (21.0 versus 3.6).
Number f Children
We expected the rate of child abuse to increase with each additional child.
It turns out that there is a 42% greater incidence of child abuse among
couples with two or more children at home as compared to those with only one
child at home (15.1 versus 10.6). However, the rate for those with three or
more children is not greater than the rate for those with two children. Just
why a second child seems to be a critical threshold after which further
children make little difference is an issue which will be studied in the next
phase of the research.
One of the clearest finding8 to emerge from the Family Violence Research
Program, as well as from the findings of the study reported in this paper, is
that violence in one sphere of life is related to violence in other sphere8
of life (Straus, 1977). Early in life, most of us receive a kind of basic
training in violence in the form of physical punishment. Mommy slaps an
infant’s hand to teach the child not to put dirty things in his or her mouth.
But this also teaches the child that love and violence go together.
Moreover, it does more than establish the empirical fact that those uho love
you are those who hit you. Ironically, it also teaches that the use of
viol.ence within the family is morally right.
The ideas expressed in the previous two paragraphs are based on plausible
reasoning, and on the results of our pilot studies. To what extent are they
supported by the data for this nationally representative sample of parents?
To find out we asked the parents how much their own parents had used physical
punishment when they were 13 or older. Those who said their mother had used
physical punishment twice or more a year had a child abuse rate of 18.5,
which is 57% greater than the rate for parents who experienced less physical
punishment (11.8). Physical punishment by the father% of the parents in this
sample made less difference: the child abuse rate for those whose father8
punished them two or more times was 16.7, compared to 13.2 for other parents.
The data just presented suggest that children learn to be violent to others
by being the victim8 of violence by their parents. Ironically, the learning
effect is probably enhanced because, by and large, parental violence is done
out of concern for the child and for other morally desirable ends. Parents
also teach violence to their children in number of other ways, for example,
by teaching boys to "stand up and fight like a man" (Stark and McEvoy, 1970),
and by example through violence towards each other.
More than one out Of ten of the parents in this sample (11.5%) could remember
at least one instance when they saw their own parents hitting each other.
Family Patterns and Child Abuse 221
The effects of observing one's father hit one's mother are greater than
effects of observing one’s mother hit one’s father. Being the child of a
father who hi& && yif_e is associated with a 39% greater rate of child abuse
compared to DS!l whose fathers did not hit their wives (13.3 versus 9.7).
However, being the daughter of a father who hit his wife is associated with
only a slightly greater rate of child abuse (19.7 versus 17.4). On the other
hand, being the daughter of a mother &.Q u m husband is associated with a
substantially higher incidence of child abuse as compared to the abuse rates
of other women (24.4 versus 17.2). The greater effect of father's violence
to his wife on sons, and of mother's violence to her husband on daughters,
suggests that violence by a parent of the same sex as the child provides the
strongest role-model.*4
Now let us turn to another way in which the principle that child abuse tends
to be associated with violence in other spheres of family life. In this
case, the question is whether couples who hit each other, also tend to abuse
their children. The answer is an emphatic nye8.n
In families where the husband had hit his wife during the year of the survey,
even if the violence was restricted to slaps, pushes, and throwing things,
the incidence of child abuse was 129% greater than in other families (28.0
versus 12.2). If the violence was more severe, the difference was even
greater: Almost a third of the couples (31.9%) in which there was an
incident of wife-beating also abused a child that year. If it is the wife
who hits the husband, this is associated with a 120% greater incidence of
child abuse (27.7 versus 12.6). But surprisingly, if the wife was very
violent, this did not produce the same further increase in child abuse as
occurred when the husband went beyond "ordinary" pushing, slapping, and
throwing things.
Earlier in this paper I referred to the important role played by the extended
family among American Blacks. The same process seems to be a work for the
entire sample, but the differences in child abuse rates between couples with
weak ties to the community and other couples is not as large as for Black
families, and fewer whites than Blacks have such ties. Taking the sample as
a whole, parents who lived in the neighborhood for a relatively short period
(0 to 3 years) have a 54% higher rate of child abuse (18.5 versus 12.03).
as a whole. Both these data and the results of other studies (see the review
in Maden and Wrench, 1977; Smith, 1975) all point to a strong association
between child abuse and sccial isolation.
ecHlrDAsUsE CHECK m
Up to this point, each of the factors associated with child abuse has been
considered separately from the others. This is clearly inadequate. These
factors do not exist in isolation. Some overlap with the others, and the
existence of certain combinations may be particularly important. As a first
approach to at least partly overcoming these limitations, a child abuse
checklist score was computed.
The procedure started with a wdiscriminant analysis" using the SPSS program
DISCRIMINANT (Nie f& j&L, 1975). This identified 16 of the variables
described in this paper as the most useful in distinguishing between abusing
parents and other parents, and which do not significantly overlap with each
oTher. These 16 variables are listed in Table 1.
Each couple was then given a score by assigning a point for any of the 16
variables on which their characteristics matched that of the abusing parents.
A couple could therefore have a score ranging from 0 (none of the factors is
present) to 76 fall of the factors present).
Family Patterns and Child Abuse 223
40.0 _
Child
l .
Abuse l
30.0 1 -
Rate
Per
l
100 20.0 _
Children
l
Age
10.0 _ -
3-17 l
...l ...........................................
o-2 3 4 5 6 7 8 9 lO+
Score on Child Abuse Checklist
The fact that the child abuse rate for parents with scores of ten or more is
“only” 33 per hundred children also indicates that we still have a long way
to go in pinpointing the causes of child abuse. As suggested in the
introduction, there is an obvious need to include data on the psychological
characteristics of the parents u the characteristics of the child. If such
characteristics were included, and if we were to use more adequate methods of
taking account of the unique combinations of factors, it might be possible to
account for even more of the cases of child abuse. But even limiting the
study to purely social factors, and using the simple methodology which there
has so far been time to apply, it is clear that we have been able to isolate
many of the factors associated with child abuse.
224 M. A. Straus
The main part of the paper examined social factors which might account for
the extremely high incidence of chi1.l abuse. The findings suggest that the
causes of child abuse can be found in (but are not limited to) the follouing
factors: (1) The structure of the contemporary American family, for example,
the practice of placing almost the whole burden on child care on mothers.
This is a main reason why women have a much higher rate of child abuse than
men, despite lower rates of violence outside the family. (2) The economic
and psychological stress created by poverty and an unstable economic system.
Illustrative of this is the finding of a higher incidence of child abuse
among manual workers and among the unemployed. (3) Isolation from the help
and social control which occurs when a family is embedded in a network of kin
and community. This is illustrated by the finding that short term residents
of a neighborhood have a higher incidence of child abuse when compared to
longer established residents. (4) Unintended but powerful training in the
use of violence as a means of teaching and resolving conflicts. Parents who
had been physically punished abuse their children much more often, as do
parents who engage in physical fights with each other, Parents who saw their
parents hit each other have a much higher rate of child abuse than other
parents.
Overall, the results of this study suggest that child abuse is brought about
by the very nature of the society and its family system. This has profound
implications for the prevention and treatment of child abuse. Although
psychotherapy may be appropriate in some cases, a more fundamental approach
lies in such things as a more equal sharing of the burdens of child care,
replacement of physical punishment with non-violent methods of child care and
training, eliminating the stresses and insecurity which now characterize our
economic system, and strengthening the ties of individual families to the
extended family and the community.
Nie, Norman, H., C. Hadlai Hull, Jean G. Jenkins, Karin Steinbrenner, and
Dale H. Brent, (1975) StatisticalPaekaaeXQKLBf:SocialSciences.Second
Edition, McGraw-Hill, New York.
Stark, Rodney and James HcEvoy III, Middle class violence, EQvcholQgy &&y
4, 52-65, (1970).
Straus, Murray A., Measuring intrafamily conflict and violence: the Conflict
Tactics (CT) scales, dournd nf Marriane anp m FalPllv 41 (1979).
Young, Leontfne R., The behavior syndromes of parents who neglect and abuse
their children, Doctoral Dissertation, Columbia University School of
Social Work (1963).