Published in
Frehsee, D., Horn, W., and Bussman, K. (Eds.), 1996. VIOLENCE AGAINST CHILDREN.
Berlin and New York: Walter de Gruyter, pp. 91-105.
Murray A. Straus and Anita K. Mathur
Family Research Laboratory, University of New Hampshire
Durham, NH 03824 (603) 862-2594
Abstract
In 1968 there was almost complete consensus concerning the cultural norm which permitted and
expected parents to use corporal punishment. At that time 94% of the U.S. adult population
approved of spanking a child. We tested the hypothesis that approval of corporal punishment
declined since then. The findings, based on data from national surveys conducted in 1968, 1986,
1988, 1989, 1990, 1991, and 1994 show a decrease to 68% in 1994. The analysis included
controls for gender, race, region, age, income, and education. Consequently the decrease is more
likely to reflect a change in cultural norms than changes in demographic composition. The
appropriateness of corporal punishment was so deeply imbedded in 1968 that there was little
difference among races, regions, genders, and educational or income groups at that time. This
makes the decrease from 94% approval to 68% approval even more noteworthy. Moreover, the
downward trend was found within all socio-demographic groups. However, there was less
decrease within some sectors of American society, such as the South, those with low education,
and African-Americans. As a result, there is less normative consensus in 1994. Both the overall
decrease and the decline in normative consensus represent major changes in cultural norms
concerning child rearing. We suggest that these changes reflect the shift to a post- industrial
society and the change in labor force characteristics needed in a post-industrial society. The
resulting decrease in corporal punishment among middle class whites, and the lesser change
among lower class African- Americans, is likely to have important consequences for children and
for social stratification: children of the majority are likely to suffer fewer psychological problems,
whereas children of the minority who continue to be frequent recipients of corporal punishment,
are likely to suffer further restrictions on social mobility, and social class differences are likely to
increase. *********************
A major structural change in American society has been the shift toward a post-industrial
economic system. Employment in manufacturing has declined drastically, as has the demand for
unskilled manual workers. There is an increasing demand for professional, technical, and clerical
workers. These are positions which, at a minimum, demand literacy and computational skills. A
large proportion of these jobs also demand interpersonal and managerial skills. Kohn and others
(Kohn, 1969; Kohn and Schooler, 1983; Straus, 1971) have argued that corporal punishment is
inconsistent with the maximization of interpersonal and managerial skills. Consistent with that,
Straus and Mathur (1994) found that corporal punishment was linked to a lower probability of
college graduation. Moreover, of those who did graduate from college, Gimpel and Straus (1992)
found that corporal punishment decreased the probability of being in the top fifth of the U.S.
occupation/income distribution. Kohn (1969) suggests that parents who hope or expect their child
to attend college and be employed in non-manual work occupations tend to avoid using corporal
punishment, and thus provide anticipatory socialization for social roles in which what is needed is
information and negotiation skills rather than physical strength and compliance with the rigid
routines of the assembly line. To the extent that this is correct, the norms and practices of
American and other post-industrial societies should be moving away from use of corporal
punishment in child rearing. We therefore tested the hypothesis that approval of corporal
punishment has declined over the last 26 years, and that this decrease applies to all the
socio-demographic groups we examined. We will test these hypotheses using data from seven
national surveys which used identical or similar questions to determine approval of corporal
punishment.
CULTURAL NORMS SUPPORTING CORPORAL PUNISHMENT
Although cultural norms supporting corporal punishment of children may be changing, there is abundant evidence that these norms are deeply rooted and pervasive (Greven, 1991; Straus, 1994). One of the most fundamental ways in which cultural norms supporting corporal punishment are expressed is in the Criminal law on assault. In every state of the U.S. hitting a child for purposes of correction or control is exempt from the crime of assault, usually with the proviso that it is limited to "reasonable force." In practice, that includes the right to hit with belts and paddles, provided the child is not injured.
In the 1960s every state in the US passed legislation designed to protect children from physical abuse and to provide services for abused children. Ironically, in order to garner sufficient votes to pass the child abuse laws, it was typically necessary to include a provision declaring that parents continued to have the right to use corporal punishment. As a result, legislation intended to protect children from physical abuse contained provisions that further legitimated a practice that increases the risk of physical abuse (Straus and Yodanis in Straus, 1994).
Cultural norms supporting corporal punishment may also be observed through public opinion surveys which ask respondents whether they approve of spanking. A 1968 survey commission by the National Commission on the Causes and Prevention of Violence found that 94% of the population approved of spanking a child who misbehaves (Stark and McEvoy, 1970). The normative support for corporal punishment goes well beyond mere permission to hit children. It is a morally correct action, and parents who do not use corporal punishment are often thought of as bad parents whose children will "grow up wild." Carson (1986) found that non-spanking parents come under considerable pressure to spank from relatives, friends, and neighbors.
TRENDS IN APPROVAL OF CORPORAL PUNISHMENT
A number of surveys of attitudes toward spanking have been done since the late 1960's. All found very high rates of approval. However, the percentage who approve of spanking has varied considerably from study to study. Consequently, most of these studies cannot be compared to determine if there has been a change in support for corporal punishment because the questions used are so different. In addition several different and non-comparable types of populations have been surveyed. For example, a 1975 survey of a national sample of parents found that 77% believed that slapping or spanking a 12 year old who misbehaved is normal and necessary (Straus, Gelles, & Steinmetz, 1980). The 1986 General Social Survey found that 84% percent agreed that "It is sometimes necessary to discipline a child with a good hard spanking" (Lehman, 1989). A survey of pediatricians found that 77% supported use of corporal punishment (White, 1993).
Although the studies just cited cannot be compared to determine trends in cultural norms concerning corporal punishment, there is evidence that actual use of corporal punishment is decreasing (Daro and Gelles, 1992; Straus, 1994). Such a change is also consistent with the theory that parents tend to socialize children in ways that will help them fit into the social environment the parent perceives the child will experience as an adult, and specifically, as indicated above, a post-industrial society. Both the actual changes in use of corporal punishment and the presumed underlying changes in the social structure make it seem likely that there has also been a corresponding change in cultural norms.
DEFINITION OF CORPORAL PUNISHMENT
Most surveys, including the surveys analyzed for this paper, operationalized corporal punishment by asking whether the respondent approved of "spanking." If one takes this literally, then the present paper is about spanking, i.e., hitting a child on the buttocks. However, many parents use "spanking" to mean any form of corporal punishment that is not severe enough to be classified as physical abuse. Thus, one mother who had slapped her child's hand during the course of the interview, told the interviewer "Sometimes, spanking is the only way to stop him." We believe that most of the respondents in the surveys we analyzed used corporal punishment as a general term to refer not just to spanking, but also to other legal and socially permissible forms of corporal punishment. Consequently, to reflect the broader meaning that we think our respondents intended, we will often refer to the data analyzed in this paper as "approval of corporal punishment." A detailed conceptual analysis of corporal punishment is given in Straus (1994). For purposes of this paper, we defined corporal punishment as
An act by a parent or other caretaker which is
intended to cause physical pain, but not injury, for purposes of
correction or control.
GROUP DIFFERENCES IN APPROVAL OF CORPORAL PUNISHMENT
In addition to examining trends, another purpose of the study was to examine whether different socio-cultural groups in American society differ in support of corporal punishment, and if there are differences between groups in the trend over the 26 year period of this study. The evidence on this issue is contradictory. Flynn (1994,a,b) analyzed the 1988 General Social Survey and found that approval of corporal punishment varied among different regions of the United States. The West, Midwest, or the South had more favorable attitudes to corporal punishment than the Northeast. Flynn also found that blacks were more likely than whites to favor spanking, and that low education, being male, and being older are also associated with greater approval of corporal punishment. However, the 1968 survey by Stark and McEvoy (1970)found no differences on any of these variables. The fact that Flynn found differences between regional and other groups, and Stark and McEvoy did not, is puzzling since both were national surveys. Therefore, one of the objectives of this study was to reexamine these group differences. With seven different surveys available, it might be possible to uncover the reason or reasons for the discrepancy.
METHOD
The Samples
To carry out the study, we merged data from seven national surveys into a single data file. The merged data set and the statistical analyses were carried out using the program SPSS/PC The merged file has an N of 6794. Each of the surveys in the merged file is described below.
1968 National Violence Survey. The 1968 National Violence Survey was designed by Rodney Stark and James McEvoy III. The survey was conducted by the National Family Violence Commission. Louis Harris Associates carried out face-to-face interviews with a national probability sample of persons age 18 and over (N = 1160). Additional information on the survey may be found in Stark and McEvoy (1970), the Commission Reports (1969), and in Owens and Straus (1975). The data file used for this paper was obtained from the Interuniversity Consortium For Political And Social Research at the University of Michigan.
1986-1991 General Social Surveys. The General Social Survey (GSS) is a yearly omnibus survey of a national probability sample of persons 18 and over (Davis & Smith, 1992). Each year had the following number of cases: 1986 N=1460, 1988 N=978, 1989 N=994, 1990 N=917 and 1991 N=283. The N for 1991 is low because GSS that year used a "split ballot" technique and asked the corporal punishment question only for this part of the sample. The 1987 GSS survey was not included in the study because it omitted the corporal punishment question.
1994 Gallup Survey. This survey was designed by David Moore and Murray A. Straus. The survey was conducted in January 1994 by the Gallup Organization using telephone interviews. The subjects were a national probability sample of persons age 18 and over (N = 1002). Additional information on this survey may be obtained from the Gallup Organization.
Comparability of Surveys. The surveys differ on several characteristics: the size of the sample (from 283 to 1,460 per year) the method of interviewing (face-to-face versus telephone), the organization conducting the survey, and the context in which the approval of violence questions were asked (a survey on violence, an omnibus survey, and a political opinion survey). The confounding of the year of the survey with these differences raises questions about the appropriateness of analyzing them as a single data set. In addition, as explained below, the question on approval of corporal punishment was slightly different in 1968. However, as will be shown, the findings follow a theoretically meaningful pattern over time, which suggests that the findings are probably not the result of confounding with one or more of the differences just discussed.
Measures
Approval of Corporal Punishment. The following question was asked in the five General Social Surveys and the 1994 Gallup Survey: "Do you strongly agree, agree, disagree, or strongly disagree that it is sometimes necessary to discipline a child with a good hard spanking?". A somewhat different question was used in the 1968 National Violence Survey: "Are there any situations that you can imagine in which you would approve of a parent spanking his or her child assuming the child is healthy and over a year old?" Respondents in this survey could answer only Yes(codes as 1) or No (coded as zero). In order to make the response categories of the seven surveys congruent, the General Social Survey and Gallup Survey answers were dichotomized by combining Strongly Agree and Agree into Yes (coded as 1), and combining Disagree and Strongly Disagree into No (coded as zero).
Independent Variables. The year in which the survey was conducted was the independent variable for testing the hypothesis concerning changes in approval of corporal punishment by parents. The variables used to investigate group differences in approval of corporal punishment were the race of the respondent, the region the respondent lived in, and the gender of the respondent. Control variables for age, income and education were also included in the form of Z-scores. The data for these three variables were transformed to Z scores before creating the merged data set. This was necessary because the various studies used different class intervals to code the data. Z scoring made them comparable in the sense that all express the score for each respondent as the number of standard deviations above or below the mean of all respondents in that survey.
Statistical Analysis
ANCOVA. The hypotheses were tested using a 7 X 2 X 3 X 4 analysis of covariance with three covariates. The seven category variable is year (1968, 1986, 1988, 1989, 1990, 1991, 1994), the two category variable is gender of the respondent, the three category variable is race (Euro-American, African-American, Other), and the four category variable is region (North East, North Central, South, West). The three covariates are age, income, and education.
The six socio-demographic factors in the analysis enable us to test the hypothesis that these groups do not differ significantly in approval of corporal punishment. In addition, their inclusion addresses the question of whether trends in approval of corporal punishment reflect a change in cultural norms per se, or a change in the demographic structure of American society. For example, the average age of the population has been increasing. If people aged 18 to 25 are more favorable to corporal punishment, the overall level of approval could decrease because there are fewer people in that age group, even though there might be no change in the extent to which persons 18-25 approve of corporal punishment. Thus, if the hypothesized decrease is found after controlling for these six demographic variables, it is plausible to infer that the decrease reflects a change in cultural norms rather than a change in demographic composition.
FINDINGS
Trends
Figure 1 shows an overall decrease in approval of corporal punishment from 1968 to 1994. The dashed line plots the rate of approval after adjusting for overlap with the other independent variables and the covariates. The dotted line is the unadjusted rate. Both lines reveal a decrease from 1968 to 1994, despite some minor deviations from this overall downward trend in 1990 and 1994. Table 1 shows that the relationship between year and approval of corporal punishment is significant at the .001 level.
As noted previously, the corporal punishment question in the 1968 survey was somewhat different from the question used from 1986 on. The extremely high rate of approval of corporal punishment in 1968 may be partly due to the wording of the question that year. But since the pattern continues down at about the same rate per year, from 1986 on, the rate shown in Figure 1 for 1968 may be correct.
(Insert Figure 1 and Table 1 about here)
In view of the deeply embedded nature of the belief that corporal punishment is sometimes necessary, and the clear legal and informal norms approving use of corporal punishment, the decrease of 26 percentage points from the high of 94% in 1968 to 68% in 1994 is a remarkable cultural change. Since controlling for changes in the demographic composition of the population made no difference in the downward trend, it seems likely that the decrease represents a change in culture rather than a change in demographics.
Group Differences
Since adjusting the rate of approval of corporal punishment made no important difference, the graphs showing differences in trends by gender, race and region show the original data, i.e., they are not adjusted for the independent variables and the covariates. However, the tests of significance in Table 1 take these factors into account.
Gender. The row for Gender in the Main Effects part of Table 1 shows that the difference between men and women in approval of corporal punishment is statistically significant. Examination of the rates themselves shows that although the difference is significant, it is not very large: 83% of men approved of corporal punishment compared to 76% of women.
(Figure 2 about here)
Figure 2 shows that approval of corporal punishment decreased for both men and women. However, the F test in Table 1 for the interaction of year and gender is significant. Examination of Figure 2 reveals the nature of the interaction effect: In 1968 slightly more women than men approved of corporal punishment (95% versus 92%). However, by 1994 this had reversed and fewer women than men approved of corporal punishment (61% of women versus 76% of men). Thus the effect of year on approval of corporal punishment to some degree depends upon whether the respondent is a man or woman. Men have held on to the old values concerning corporal punishment somewhat more than women.
Race. Respondents were divided into three ethnic categories: Euro-American, African-American, and Other. The row labeled ETHNIC in Table 1 shows that the difference between ethnic groups in approval of corporal punishment is significant at the .001 level. The highest rate of approval is for African Americans (91%), followed by Euro- Americans (90%), and the lowest approval percentage is for the Other category (78%).
(Insert Figure 3 about here) Trends for each ethnic group are shown in Figure 3. It shows large decreases in approval of corporal punishment for both Euro-Americans and Others over the 26 year period. African-Americans clearly do not follow the same trend. African-American respondents decreased in their approval of corporal punishment only 14 percentage points over the same period, whereas Euro-Americans and Others decreased about twice that amount (32 percentage points for Euro-Americans and 26 percentage points for the Other category). As a result of the nearly identical pattern for the Euro-American and Other groups, the F test for the interaction of Year and Ethnic Group, shown in Table 1 is not significant. However, since Figure 3 clearly shows that the decrease in approval of corporal punishment was much less among African-Americans than the rest of the population, it seemed appropriate to retest the hypothesis after combining Euro-Americans and Others into a single category. The results from that ANCOVA showed a significant interaction of Year with Ethnic Group (F = 2.99, p <.03).
Region The row for Region in the Main Effects part of Table 1 shows that there is a significant difference between regions in approval of corporal punishment. The region of the United States with the largest percentage approving corporal punishment is the South. The regions with the lowest rate of approval of corporal punishment are the North East and the West. The approval rates for the four regions are: 75% in the North East, 77% in the West, 81% in the North Central, and 85% in the South.
(Insert Figure 4 about here)
The trend lines in Figure 4 show that respondents from all four regions decreased in approval of corporal punishment over the 26 years of this study. However, respondents from the South decreased less than those in other regions. The North East started out in 1968 as having basically the same percent approval of corporal punishment as the South. By 1994, however, the North East was approximately 17 percentage points lower than the South in approval of corporal punishment. The North East decreased 35 percentage points from 1968 to 1994, whereas the South decreased only 21 percentage points during the same time interval. The regions thus grew more different from each other in the proportion approving corporal punishment mainly because the South did not decrease as much as other regions. These differences between regions in trends are reflected in the significant F test for the interaction of year and region in Table 1.
Covariates
Age. Age was included in the ANCOVA as a covariate rather than an independent variable because including age as an additional independent variable would have resulted in many empty cells. The row for AGE in Table 1 shows that the older the respondent, the more approval of corporal punishment.
Table 1, however, cannot provide information on whether the trend from 1968 to 1994 was the same for all age groups. To investigate trends by age of respondent, and also trends by income and by education (the other two covariates), we computed an analysis of variance with these three variables categorized so that they could be used as blocks in the design. The other independent variables were year of the study, gender of the respondent, and region. We classified age of the respondents into quintiles. The results for age were again significant (F=4.067, p <.003), but inspection of the approval rates for the five age groups showed that the differences are not large: the maximum difference between any two age groups is only 6%. As for trends, inspection of the trends within each of the five age groups showed that all five followed the general downward trend over the 26 year period.
Income. There was no significant relationship found between income and approval of corporal punishment, and all five income groups followed a similar downward trend in approval of corporal punishment from 1968 to 1994.
Education. For purposes of examining trends within educational groups, the respondents were classified into four groups: more than one SD (standard deviation) below the mean, up to one SD below the mean, up to one SD above the mean, and more than one SD above the mean. The results of the ANCOVA show that with increasing amounts of education, there is a consistent decrease in approval of corporal punishment. This relationship between education and approval of corporal punishment is significant at the .001 level (F=12.63). Overall, the percent approving corporal punishment for the four educational groups are: 75% (highest education), 80%, 82%, and 84% (lowest education).
All educational groups decreased in approval of corporal punishment over the 26 year period, but the highest education group decreased at a faster rate. In 1968, the highest educational group had about the highest rate of approval of corporal punishment, and the lowest educational group was the least favorable to corporal punishment. The rates in 1968 ranged from 95% (highest education), 97%, 93%, 88% (lowest education). However, by 1994, the highest education group had a somewhat lower rate of approval than the other education groups: The rates in 1994 ranged from 64% (highest education group), 69%, 72%, 68% (lowest education group). Although the larger decrease for the highest education group is important, all education groups decreased, and the differences between education groups in 1994 are not large.
DISCUSSION
A 1968 national survey found that 94% of the U.S. population approved of parents spanking children. Except for the obligation to provide for the basic needs of a child, it is hard to think of any other aspect of child rearing on which there was or is such consensus. However, as argued in the introduction, corporal punishment interferes with children acquiring the skills and values needed for advancement in the labor market of a post-industrial society. This assumption, together with studies which found a decrease in use of corporal punishment by parents in recent years (Daro and Gelles, 1992; Straus, 1994) led to the hypothesis that cultural norms supporting use of corporal punishment have eroded. The findings from the analysis of national surveys from 1968 to 1994 reported in this paper are consistent with that hypothesis. The percent of the U.S. population who approve of corporal punishment declined over the 26 year period from 94% in 1968 to 68% in 1994.
Decline In Consensus
The findings also show that within this overall downward trend there are differences in the rate of decrease which have altered the former normative consensus on corporal punishment. The decline in approval of corporal punishment has been greater in certain parts of the population and smaller in others. These differences have produced a widening gulf in attitudes toward the appropriateness of spanking children. For example, although the South decreased in approval since 1968, the decrease was smaller than in other regions of the United States, so that by 1994 the South was more favorable to corporal punishment than other regions. Similarly, approval of corporal punishment by African-Americans declined less than in other ethnic groups. As a result, norms supporting corporal punishment are now stronger among African-Americans than other ethnic groups. Also, approval of corporal punishment by men has decreased less than approval by women. Therefore, men have emerged as being more approving than women. Similarly, approval of corporal punishment has declined less for people with lower education and for older age groups. As a result, less educated and older people are now more approving of corporal punishment than better educated or younger persons. These differences in the rate of change have resulted in a greater differentiation in norms between different sectors of American society.
The Backlash
A decrease from 94% approval of corporal punishment to 68% in 26 years is a major social change, as is the decrease in consensus on this aspect of child rearing. However, caution is needed concerning extrapolating this trend. Figure 1 shows a slight upward trend in the last three years. We have implicitly treated the slight increase as a random variation which does not indicate that the downward trend since 1968 has halted or reversed. But it is possible that the trend has halted or reversed. Given the rising youth crime rate, there may be a tendency to reconsider corporal punishment. Moreover, a change in a normative principle as deeply embedded as the idea that corporal punishment is necessary to properly socialize the next generation is likely to be threatening to many. Moreover, those who continue to favor corporal punishment are likely to perceive the change as threat to the family and morality. Some have organized to reverse the trend, as happened for the legislation passed by the Florida legislature in the spring of 1994. This bill would have prohibited child protective services from ruling that a child had been abused if the only evidence was that the parents used corporal punishment and the child exhibited welts or bruises. The bill was passed overwhelmingly but was vetoed by the governor. However, the reason for the veto was not a disagreement with the purpose of the bill; it was because the governor believed it would create administrative problems (Gainesville Sun, July 24, 1994, p. 1; Miami Herald, June 7, 1994, p. 1A). It is also possible that the Florida legislation was part of the backlash against child protective services (the state agencies charged with enforcing\ legislation to protect children from physical and sexual abuse) rather than (or in addition to) a defense of corporal punishment (Finkelhor, 1994).
Policy Implications
The backlash and the widening gulf between socioeconomic status groups in respect to corporal punishment are sources of concern. Ignoring these problems could interfere with the transition to new norms which prohibit rather than require corporal punishment. It could exacerbate class and race conflict if efforts to end corporal punishment are seen as an attempt to control and limit the lives of the under-privileged in the guise of "child protection." Such a perception would be ironic because of the evidence suggesting that violent socialization in the form of frequent use of corporal punishment reduces the probability of upward social mobility (Gimpel and Straus, 1992; Straus and Mathur, 1994). To help avoid either extreme (failing to take steps to end a practice which harms children and the society, or the perception of cultural imperialism), educational programs to reduce corporal punishment among disadvantaged groups need to be designed and implemented by members of those groups.
Despite the recent backlash approval of corporal punishment by parents has decreased drastically since 1968. A related change is occurring in corporal punishment in schools. In the forty year period from 1940 to 1980, only four states made the practice illegal. Since 1980, such legislation has been passed at an accelerating rate, so that by 1993, just over half the states of the U.S. prohibited corporal punishment in the schools (Hyman, 1990; ??, 19??). The decrease in the rate of approval and actual use of corporal punishment by parents, together with the dramatic decrease in the number of states which permit corporal punishment in the schools is encouraging given the evidence that corporal punishment is no more effective than other methods of discipline for young children, that it is less effective than other methods with older children, and that it increases the probability of many harmful side effects (Straus, 1994). Among the harmful side effects are a greater risk of the child being violent and delinquent, engaging in crime as an adult (Straus, 1991, 1994), being depressed (Straus, 1994), and decreased probability of upward social mobility (Gimpel and Straus, 1992; Straus and Mathur, 1994). Moreover, the shift away from corporal punishment among the majority, and the lesser change among low socioeconomic status groups, is likely to further stratify American society: The children of the majority are likely to suffer fewer psychological problems and that will facilitate their occupational success, whereas children of the minority, who continue to be frequent recipients of corporal punishment, are likely to suffer still further impediments to upward mobility, and social class differences are therefore likely to increase.
Legislative Approaches
Legislation to prohibit spanking or other form of corporal punishment was enacted in Sweden in
1979. It is important to note that this legislation does not include a criminal penalty. It is purely
for purposes of educating and assisting parents. Since then the other Scandinavian countries and
Austria have enacted legislation prohibiting corporal punishment by parents. In the United
Nations, ..... (??, 19??). However, in the United States, even among those who are opposed to
corporal punishment, all but a few doubt the possibility of such legislation for many years.
Moreover, there is a fear that attempts to enact a no-spanking law at this point would stiffen
opposition to the change. But social change is notoriously difficult to predict. Although the
decrease in normative support for corporal punishment has not yet progressed far enough to have
the issue raised in any of the 50 state legislatures of the U.S., the time may be closer than either
the critics or the advocates of legislation believe. There is evidence of a broad shift away from
norms tolerating interpersonal violence (Grasmick, Blackwell, Bursik, and Mitchel, 1993; Straus
and Kaufman Kantor, 1994; Straus, Kaufman Kantor, and Moore, 1994). If these trends continue,
the United States may not be far from the day when state legislatures in the United States will be
considering laws that are the opposite of the Florida legislation mentioned above, i.e., legislation
to prohibit corporal punishment by parents. In view of the many harmful side effects of corporal
punishment, such legislation would contribute to the well being of children, and also contribute to
reducing social inequality by avoiding the impaired opportunity for social mobility which occurs
when lower socioeconomic status groups socialize their children using corporal punishment.
REFERENCES
Anderson, Kathryn A. and David E. Anderson. 1976. "Psychologists and Spanking." Journal of Clinical Child Psychology 5(2): 46-49.
Calhoun, Craig. 1992. "Social Change." Pp 1807-1812 in Encyclopedia of Sociology, vol 4, edited by Edgar F. Borgatta and Marie L. Borgatta. NY:Macmillan.
Carson, Barbara A. 1986. Parents Who Don't Spank: Deviation in the Legitimation of Physical Force. Durham, NH: Ph.D dissertation. University of New Hampshire.
Daro, Deborah and Richard J. Gelles. 1992. "Public Attitudes and behaviors with Respect to Child Abuse Prevention." Journal of Interpersonal Violence. 7:517-531.
Davis, James A. and Tom W. Smith. 1992. The General Social Surveys, 1972-1991: Cumulative Codebook. Chicago: National Opinion Research Center.
Finkelhor, David. 1984. "The 'Backlash' and the future of child protection advocacy: Insights from the study of social Issues." Pp 1-16 in Myers, John E.B. The Backlash:Child Protection Under fire. Thousand Oaks, CA: Sage, 1994
Flynn, Clifton P. 1994. "Regional Differences in Attitudes Toward Corporal Punishment." Journal of Marriage and th Family. 56(2):314-324.
Gimpel, Holly, and Murray A. Straus. 1992. "Corporal punishment and economic achievement: A theoretical model and some preliminary data." Paper presented at the 1992 meeting of the American Sociological Association. An abbreviated version is in Straus, 1994, Chapter 9.
Grasmick, Harold G., Brenda Sims Blackwell., Robert J. Bursik, Jr., and Suzanne Mitchell. 1993. "Changes in Perceived Threats of Shame, Embarrassment, and Legal Sanctions for Interpersonal Violence, 1982-1992." Violence and Victims. 8(4):313-325.
Greven, Philip. 1991. Spare the Child: The Religious Roots of Physical Punishment and the Psychological Impact of Physical Abuse. New York: Alfred A. Knopf.
Harris, Marvin. 1979. Cultural Materialism. NY:Vintage.
Hyman, Irwin A. 1990. Reading, Writing, and the Hickory Stick: The Appalling Story of Physical and Psychological Abuse in American Schools. Lexington, MA: D.C. Heath and Company.
Kohn, Melvin L. 1969. Class and conformity: A study in values. Homewood, Il: Dorsey Press.
Kohn, Melvin L. and C. Schooler. 1983. Work and Personality: An Inquiry into the Impact of Social Stratification. Norwood, NJ: Ablex Publishing Corporation.
Lehman, Betsey A. 1989. "Making a case against spanking." The Washington Post. Thursday, March 23, p. D5.
Owens, David and Murray A. Straus. 1975. "The Social Structure of Violence in Childhood and Approval of Violence as an Adult. Aggressive Behavior 1:193-211. Also in Corporal Punishment in American Education: Readings in History, Practice, and Alternatives, edited by I. H. Hyman and J. H. Wise. Philadelphia: Temple University Press, 1979.
Stark, Rodney, and James McEvoy, III. 1970. Middle Class Violence. Psychology Today, 4, 52-65.
Straus, Murray A. 1971. "Some social antecedents of physical punishment: A linkage theory interpretation." Journal of Marriage and the Family. 70(2):658-663. Also reprinted in S. K. Steinmetz & M. A. Straus (Eds.), Violence in the family (pp. 159-166). New York: Harper and Row.
Straus, Murray A. 1994. Beating the Devil Out of Them: Corporal Punishment In American Families. Lexington, Mass: Lexington/Macmillian Books.
Straus, Murray A. and Anita K. Mathur. 1994. "Corporal Punishment by Parents and Later Occupational and Economic Achievement of Children." Durham, NH: University of New Hampshire, Family Research Laboratory.
Straus, Murray A. and Denise Donnelly. 1993. "Corporal Punishment of Teen Age Children in the United States." Youth and Society24:419-42.
Straus, Murray A., Richard J. Gelles, and Suzanne K. Steinmetz. 1980. Behind Closed Doors: Violence in the American Family. NY: Doubleday /Anchor.
White, K. (1993). Where pediatricians stand on spanking. Pediatric Management, September,
11-15
Source Regr. Sum of Mean Signif Variation Coef. Squares DF Square F of F -------------------------------------------------------------------------------- Main Effects 574435.808 12 47869.651 32.447 .001 YEAR 361073.162 6 60178.860 40.790 .001 GENDER 53834.017 1 53834.017 36.489 .001 ETHNIC 55619.047 2 27809.524 18.850 .001 REGION 66280.103 3 22093.368 14.975 .001 2-way Interactions* 175211.069 47 3727.895 2.527 .001 YEAR GENDER 43286.514 6 7214.419 4.890 .001 YEAR ETHNIC 19565.840 12 1630.487 1.105 .350 YEAR REGION 67693.330 18 3760.741 2.549 .001 Covariates 71120.117 3 23706.706 16.069 .001 AGE .783 7426.471 1 7426.471 5.034 .025 INCOME .122 148.161 1 148.161 .100 .751 EDUC -.3058 48973.442 1 48973.442 33.195 .001 Explained 820766.994 62 13238.177 8.973 .001 Residual 9346222.690 6335 1475.331 Total 10166989.684 6397 1589.337 _______________________________________________________________________________ YEAR = YEAR OF THE STUDY GENDER = GENDER OF RESPONDENT (0=MALE 1=FEMALE) ETHNIC = ETHNIC GROUP OF RESPONDENT (1=EURO-AMERICAN 2=AFRICAN-AMERICAN 3=OTHER) REGION = REGION OF RESPONDENT:FOUR CENSUS REGIONS AGE = AGE OF RESPONDENT, Z SCORE, 5 CATEGORIES INCOME = INCOME OF RESPONDENT, Z SCORE, 5 CATEGORIES EDUC = EDUCATION OF RESPONDENT, Z SCORE * Higher order interactions were not tested because we had no hypotheses about higher order interactions. In addition, even with an N of this size, there would be empty cells.
A. APPROVAL BY YEAR AND GENDER OF RESPONDENT GENDER OF RESPONDENT YEAR MEN WOMEN 1968 92.24 95.31 (541) (512) 1986 85.47 82.01 (592) (795) 1988 83.14 76.60 (421) (517) 1989 81.77 74.40 (406) (539) 1990 82.17 76.43 (387) (488) 1991 70.64 63.75 (109) (160) 1994 75.73 60.49 (478) (453) B. APPROVAL BY YEAR AND ETHNIC GROUP OF RESPONDENT ETHNIC GROUP OF RESPONDENT YEAR Euro- African- Other American American 1968 93.05 96.69 100.00 (863) (181) ( 9) 1986 82.53 89.35 87.88 (1185) (169) ( 33) 1988 77.79 91.53 78.13 (788) (118) ( 32) 1989 76.85 84.54 75.00 (812) ( 97) ( 36) 1990 77.13 91.43 80.65 (739) (105) ( 31) 1991 62.86 83.67 60.00 (210) ( 49) ( 10) 1994 66.91 83.12 68.29 (813) ( 77) ( 41) C. APPROVAL BY YEAR AND REGION OF RESPONDENT REGION YEAR NORTH NORTH SOUTH WEST EAST CENTRAL 1968 94.29 91.58 94.02 96.00 (280) (297) (301) (175) 1986 79.36 86.09 86.94 77.91 (281) (381) (467) (258) 1988 68.54 79.68 86.54 77.47 (178) (251) (327) (182) 1989 73.60 71.14 85.63 75.40 (178) (246) (334) (187) 1990 71.98 82.91 81.57 76.51 (182) (234) (293) (166) 1991 67.89 66.07 68.42 61.70 (109) ( 56) ( 57) ( 47) 1994 58.95 71.95 76.14 63.78 (229) (221) (285) (196)
Table 3. ADJUSTED RATES OF APPROVAL
YEAR % APPROVE 1968 93.26 (1053) 1986 83.38 (1387) 1988 79.29 (938) 1989 77.67 (945) 1990 79.08 (875) 1991 68.06 (269) 1994 68.59 (931)
Adjusted for Independents Unadjusted + Covariates Variable + Category N Dev'n Eta Dev'n Beta
Year 1968 VN 1053 13.55 13.08 1986 GSS 1387 3.31 3.20 1988 GSS 938 -.65 -.89 1989 GSS 945 -2.62 -2.51 1990 GSS 875 -1.21 -1.10 1991 GSS 269 -13.64 -12.12 1994 GA2 931 -11.87 -11.59 .20 .19 Grand Mean = 80.181 Multiple R Squared .063 Multiple R .252