HM Medical Clinic

 

Childrearing discipline and violence in developing countries

Child Development, January/February 2012, Volume 83, Number 1, Pages 62–75 Childrearing Discipline and Violence in Developing Countries Jennifer E. Lansford Kirby Deater-Deckard Virginia Polytechnic Institute and State University The present study examined the prevalence and country-level correlates of 11 responses to children's behav-ior, including nonviolent discipline, psychological aggression, and physical violence, as well as endorsementof the use of physical punishment, in 24 countries using data from 30,470 families with 2- to 4-year-old chil-dren that participated in UNICEF's Multiple Indicator Cluster Survey. The prevalence of each response variedwidely across countries, as did the amount of variance accounted for by country in relation to each response.
Country-level indicators of life expectancy, educational attainment, and economic well-being were related toseveral responses to children's behavior. Country-level factors are widely related to parents' methods ofteaching children good behavior and responding to misbehavior.
Parents' responses to conflict with their children tural and economic contexts (Gershoff et al., 2010).
are an important part of the way that parents One could conceptualize parents' responses to chil- socialize children because their responses can cor- dren's behavior as falling into three broad catego- rect misbehavior and promote desired behaviors in ries: (a) nonviolent (e.g., offering explanations, the future. Parents hold a wide range of beliefs removing privileges), (b) psychologically aggressive regarding the acceptability and advisability of dif- (e.g., yelling, name calling, threatening), and (c) ferent forms of discipline and use a wide range of physically violent (e.g., slapping, beating with an actions to manage children's behavior (Mistry, object). These three categories are not necessarily Chaudhuri, & Diez, 2003). Although there are within-country differences in these beliefs and responses can occur simultaneously; it is also possi- behaviors, many beliefs and behaviors appear to be ble for an explanation to be given in an aggressive shaped by the cultural context in which parents live or threatening way.
(Bornstein & Lansford, 2009; Garcia-Coll & Magnu- Because most physical abuse takes place in the context of physical punishment (e.g., Durrant, Although physical punishment has received 2004), a growing body of literature identifies physi- more research attention than have other responses cal punishment as a risk factor in children's devel- to children's behavior, parents differ in their reac- opment (see Gershoff, 2002, for a review), and a tions to children not just in their use of physical sizable contingent of scholars and practitioners punishment but also in a variety of other ways. For categorizes any form of physical punishment as example, in some countries, calling the child derog- maltreatment (see Whipple & Richey, 1997), par- atory names is accepted and practiced as a means ents' actions toward their children have come of teaching the child right from wrong, whereas in under scrutiny in many countries. Because of con- other countries this kind of name calling would be cerns that children are vulnerable to abuse or strongly discouraged (Fung, 1999). Likewise, some neglect if parents' actions are left solely to their cultural groups rely more heavily on the manipula- own discretion, children's right to protection was tion of privileges to manage children's behavior recognized in the Convention on the Rights of the than do other cultural groups (Kim & Hong, 2007).
Child (CRC), which was adopted by the United Parents in several countries have been found to use Nations (UN) General Assembly in 1989. The CRC explanations regarding good and bad behavior as a sets out international standards that must be met to socialization tool designed to teach their children protect children from abuse and exploitation in a appropriate and expected behavior within their cul- Correspondence concerning this article should be addressed to  2012 The Authors Jennifer E. Lansford, Center for Child and Family Policy, Duke Child Development  2012 Society for Research in Child Development, Inc.
University, Box 90545, Durham, NC 27708. Electronic mail may All rights reserved. 0009-3920/2012/8301-0006 be sent to [email protected].
Childrearing Discipline and Violence 186 preindustrial societies, Ember and Ember (2005) Recently, the UN conducted a global study of vio- found that several societal-level factors were related lence against children, which concluded with a goal to the use of physical punishment in particular. For of putting ‘‘an end to adult justification of violence example, physical punishment was more prevalent against children, whether accepted as ‘tradition' or in societies with higher levels of social stratification disguised as ‘discipline''' (Pinheiro, 2006, p. 5). The and with undemocratic political decision making, UN defines physical and psychological violence which the authors suggested may support the the- toward children as a breach of their rights under ory that parents use physical punishment to social- the CRC. Therefore, those countries that have rati- ize children to live in a society with power fied the CRC (all but 2, the United States and Soma- inequalities. According to this theory, parents con- lia) are obligated to examine their policies, laws, sciously and unconsciously socialize their children and cultural norms to ensure that they uphold chil- to be more submissive and obedient in contexts in dren's right to protection. Indeed, more than 100 which parents believe these qualities are valued; countries have banned the physical punishment of physical punishment is sometimes used by parents children in schools, and 29 countries have legally as one method of promoting children's obedience.
banned the physical punishment of children in all Within the United States, low-socioeconomic-sta- settings (Durrant, 2008; http://www.endcorporal- tus (SES) families are more likely to value obedi- punishment.org). As a result, it is interesting to ence than are high-SES families (Peterson & Hann, examine and compare international trends in this 1999). It is possible that a similar pattern may be area at this particular point in history.
found at a country level if countries with fewer eco- The present study documents a range of disci- nomic and social resources value obedience more pline techniques and forms of violence caregivers and use physical punishment to try to instill obedi- reported using in 24 developing countries and ence in children. Likewise, around the world, a examines these responses in relation to indicators country's average life expectancy is highly corre- of the countries' life expectancy, educational lated with many indicators of resources available in achievements, and economic well-being. Under- the country (e.g., access to safe drinking water, san- standing within-country and between-country asso- itation, vaccines) that might relate to higher or ciations between sociodemographic factors and lower levels of stress experienced by the family. In parents' responses to children's behavior addresses developed countries, family stress has been found an important question: Are these associations uni- to predict more violence against children (e.g., versal or country specific? In the context of the cur- Annerba¨ck, Svedin, & Gustafsson, 2010; Crouch & rent global community's focus on protecting Behl, 2001); in comparisons of developing countries, children from all forms of violence, descriptive life expectancy might serve as a proxy for stressful information about both the prevalence and corre- living conditions experienced by families and might lates of different responses to children's behavior is also predict higher societal rates of violence against increasingly important as a first step toward imple- menting programs to prevent family violence Despite ethnographic evidence that parents' (United Nations Children's Fund [UNICEF], 2006).
actions toward their children differ across cultural The anthropological literature has a history of contexts, the vast majority of studies in the quanti- comparing childrearing practices and value systems tative developmental science literature have been across cultural groups using qualitative, ethno- conducted using North American samples, and graphic approaches. For example, Beatrice and John studies of other cultural groups often have relied Whiting's Six Cultures Project incorporated obser- on families that have immigrated to North Amer- vations of children, interviews with mothers, and ica. However, there are a few notable exceptions.
ethnographic notes to understand parenting prac- For example, in a study of mothers and children in tices and children's behavior in Mexico, India, China, India, Italy, Kenya, the Philippines, and Kenya, the United States, Japan, and the Philip- Thailand, Lansford et al. (2005) found that although pines (Whiting & Whiting, 1975). One conclusion more frequent physical punishment was related to from this work was that parents differed across more child aggression and anxiety in all six coun- countries in the techniques they used to socialize tries, these links were weaker in countries in which their children, providing early evidence of the the use of physical punishment was more norma- importance of cultural context in understanding tive. In these same six countries, Gershoff et al.
parenting and child development. Using archival (2010) found that several forms of discipline were ethnographic data collected by anthropologists in related to higher levels of child aggression and Lansford and Deater-Deckard anxiety. Alyahri and Goodman (2008) found that GDP as parents with fewer financial resources in more than half of rural Yemeni caregivers and a the United Sates physically punish their children fourth of urban Yemeni caregivers reported using more than parents with more financial resources harsh forms of corporal punishment such as hitting within the United States? children with implements, tying them up, and bit- Mechanisms through which education and finan- ing them. Experiencing these forms of violence was cial resources within countries relate to parents' related to children's poor school performance and actions toward their children suggest that similar behavioral and emotional problems. In their review relations may be found between countries. For exam- of literature on parenting in Jamaica, Smith and ple, progressing through a formal education system Mosby (2003) reported that ‘‘flogging'' or beating may change parents' beliefs and attitudes in ways the child is the most common response to child that make them more likely to value their children's misbehavior. The endorsement and use of harsh independent thinking that is fostered by the parent's forms of physical punishment also have been verbal reasoning (Davis-Kean, 2005). In addition, reported in sub-Saharan Africa (Monyooe, 1996; country economic conditions and the nature of work Oburu & Palme´rus, 2003). One question is whether might create circumstances that move people to the prevalence of violence against children reported change from restrictive, controlling parental behav- in these studies is widespread in other countries.
ior aimed at inducing children's compliance to more Despite the exceptions described earlier, parents' reflective responses aimed at promoting children's responses to children's behaviors are virtually self-determination and autonomy. It could be that unknown for those countries that are underrepre- improving country-level economic conditions or sented in developmental science. The present study changing the nature of work from dull and hard fills a gap in the literature by examining discipline labor tasks to those requiring more independence and violence in 24 developing countries.
and decision making could move people both to seek Parents' actions are dependent, in part, on the more education and to change the ways in which age of the child. For example, in Western cultures they interact with their children (Conger & Donne- physical punishment has been found to peak dur- llan, 2007). Another possibility is that having more ing the toddler years and to decline thereafter financial resources is related to lower levels of paren- (Straus & Stewart, 1999). By contrast, parents' ver- tal stress and, therefore, less spillover of parental bal explanations increase as children grow older, stress into harsh parental behavior (Deater-Deckard, likely in response to children's cognitive develop- 2004; McLoyd, 1990).
ment, which enhances children's capacity for The present study documents the prevalence of understanding complex reasoning (Collins, Mad- 11 responses to children's behavior as well as sen, & Susman-Stillman, 2002). Given these age- endorsement of physical punishment in 24 develop- related differences in parents' responses to their ing countries. This work was guided by two ques- children and because of the particular salience of tions. First, how common is each response in each behavior management during early childhood (as country, and how do countries compare with opposed to aspects of parent–child relationships, respect to the prevalence of each response? Second, such as monitoring or more distal supervision that how is the prevalence of, and variability in, each take on increased importance as children age), the response related to country-level indicators of these present study focuses on parents' discipline and nations' life expectancy, educational achievements, violence toward children aged 2–4 years.
and economic well-being? We hypothesized that It is unclear to what extent within-country socio- socioeconomic differences in the use of physical economic variability in parents' use of physical punishment that have been reported within the punishment may extend to between-country vari- United States would also be found in comparisons ability among developing nations. For example, do between developing countries, with fewer caregiv- parents in countries with lower literacy rates physi- ers reporting the use of physical punishment in cally punish their children more than parents in countries with more sociodemographic resources.
countries with higher literacy rates, as less edu-cated parents in the United States physically punishtheir children more than more educated parents within the United States punish their children? Likewise, do parents in countries with lower grossdomestic product (GDP) physically punish their The Multiple Indicator Cluster Survey (MICS) is children more than parents in countries with higher a nationally representative and internationally Childrearing Discipline and Violence comparable household survey (UNICEF, 2006). It is implemented in a large number of developing Sample Size by Country and General Linear Model Parameter Esti- countries and provides a unique source of informa- mates for Country Deviation From the Overall Effect of Number of tion to examine protective and risk factors for child Responses Used in Last Month health, nutrition, education, development, and Number of responses well-being in different regions of the world.
UNICEF developed the MICS for countries to col- lect internationally comparable data to evaluate country-level progress on issues related to children and women in low- and middle-income countries.
Three rounds of MICS have been implemented; we use data from the MICS3, which was conducted in 2005–2007 (questionnaires are available at http:// Each questionnaire is composed of core, additional, and optional modules, which are sets of standard- Bosnia and Herzegovina ized questions grouped by topics. Each country was responsible for designing and selecting a sam- ple. The survey sample was usually a probability sample in all stages of selection, national in cover- Syrian Arab Republic age, and designed in as simple a way as possible so that its field implementation could be easily and faithfully carried out with minimum opportunity for deviation from an overall standard design. Mul- tiple steps were taken to ensure data reliability.
MICS3 respondents were normally the mother or primary caregiver of the child.
Of the 28 countries that contributed data used in Central African Republic this Special Section, all but 4 (Bangladesh, Somalia, Thailand, and Uzbekistan) included the discipline module, which was administered in relation to chil- dren between the ages of 2 and 14 years. If there was more than one child between the ages of 2 and14 years in the household, the interviewer used a Note. All contrast estimates are significant at p < .05, except thosedesignated with a  . HDI = Human Development Index.
standardized protocol to select a target child ran-domly from the eligible children in the household.
Because parental actions toward children of differ-ent ages vary considerably, we included only those tion. Table 1 presents the number of families with households with a child between the ages of 2 and children between the ages of 2 and 4 years who 4 years whose female caregiver responded to the were selected to participate in the discipline mod- MICS3 discipline module (N = 30,470). Across ule in each country. Although the sample sizes and countries, the average age of the randomly selected proportionate representation of the country's popu- lation vary across countries, each sample was (M = 2.97, SD = 0.80), and the percentage of chil- selected to be nationally representative of the coun- dren who were female ranged from 46% to 53% try from which it was drawn.
(49% across all countries). Primary female caregiv-ers ranged in age from 13 to 95 years (M = 29.54, SD = 7.79). Twenty-five percent of caregivers hadno formal education, 25% had a primary school Discipline and violence items. Mothers or primary education, 40% had a secondary school education, female caregivers were told, ‘‘All adults use certain and 10% had more than a secondary school educa- ways to teach children the right behavior or to Lansford and Deater-Deckard address a behavior problem. I will read various of children who were spanked with a hand, hit on methods that are used and I want you to tell me if the extremities, shaken, or hit with an object. The you or anyone else in your household has used this severe physical violence scale reflected the percent- method with (name of child) in the past month.'' age of children who were hit on the head or who Caregivers were then asked whether they or any- were beaten with an implement. We recognize that one in their household had used each of 11 nonvio- views about the severity of specific forms of vio- lence likely vary across countries and therefore cau- violent responses with the target child in the last tion that the results should be interpreted with this month: (1) explained why something (the behavior) caveat in mind.
was wrong; (2) gave the child something else to do; (3) took away privileges, forbade something, or did Development Index (HDI) was developed by the not allow the child to leave the house; (4) shouted, UN as a measure of the social and economic status yelled at, or screamed at the child; (5) called the of a country (United Nations Development Pro- child dumb, lazy, or another name like that; (6) gramme, n.d.). It serves as a proxy for standard of spanked, hit, or slapped the child on the bottom living and is associated with the general level of with a bare hand; (7) hit or slapped the child on the purchasing power present within a country. The hand, arm, or leg; (8) shook the child; (9) hit the HDI ranges from 0 to 1 and has three major indices: child on the bottom or elsewhere on the body with life expectancy, education (composed of the adult something like a belt, hairbrush, stick, or other hard literacy rate and combined gross enrollment in pri- object; (10) hit or slapped the child on the face, mary, secondary, and tertiary school), and GDP.
head, or ears; (11) beat the child up with an imple- Countries with an HDI of .80 or greater are consid- ment (hit over and over as hard as one could). The ered high, .50 to .79 medium, and .00 to .49 low. The items were developed using an approach that countries in our study draw from high, medium, included convening an international panel of 25 and low ranges of HDI. (Additional information experts to identify candidate items from existing about the HDI is available in Bornstein et al., 2012.) validated measures of caregiving; field testing can-didate items via cognitive interviews and quantita-tive surveys in the Americas, South Asia, and Africa; and convening a second international panel of 27 experts to evaluate items' performance withinand across diverse cultures and settings (Kariger Child age in years (range = 2–4) and child gen- et al., 2010). These 11 items were adapted from the der (male = 1, female = 2) were included as covari- Parent–Child Conflict Tactics Scale (Straus, Hamby, ates in all analyses. Child age varied across Finkelhor, Moore, & Runyan, 1998) and the World- countries, v2(46) = 618.85, p < .001, and was mod- SAFE survey questionnaire (Sadowski, Hunter, estly but significantly correlated with all 11 Bangdiwala, & Munoz, 2004). A 12th item asked responses to the child in the last month, rs ranged whether the caregiver believes that to bring up, from .03 to .10, ps < .001; age was not significantly raise, or educate the target child properly it is nec- correlated with caregivers' belief in the necessity of essary to punish him or her physically; don't know using physical punishment. Child gender also var- and no opinion responses for this item were treated ied across countries, v2(23, N = 30,470) = 40.93, as missing. In Mongolia, only items 1, 2, 4, 5, 11, p < .05; correlations between gender and the disci- and 12 were asked. Each item was recoded as pline and violence items ranged from .01 to .04, ns 0 = no, 1 = yes.
to ps = .001.
Discipline and violence scales. We constructed four scales following the construction recommended by UNICEF (2006) for these items. The nonviolencescale reflected the percentage of children whose First, the total number of responses used, the caregivers explained why something was wrong, four scales, the 11 response items, and the physical gave the child something else to do, or took away punishment attitude item were explored, with privileges but did not engage in any of the other country as a predictor and child age and gender as covariates. For logistic regression models, we report reflected the percentage of children whose caregiv- Cox and Snell's and Nagelkerke's pseudo-R2 values ers yelled at the child or called the child a name.
The physical violence scale reflected the percentage accounted for by country. Next, for each country, Childrearing Discipline and Violence we computed the average of caregivers' responses behavior used by caregivers in the last month. The to each of the scales and items, creating the overall general linear model main effect of country was percentages of caregivers who indicated that they significant, F(23, 29,919) = 251.57, p < .001, partial or someone in their household had responded to g2 = .16, controlling for child age and gender. As the child in each way in each country and who shown in Table 1, countries ranged from a low of believed it was necessary to physically punish their 2.01 responses in Kazakhstan to a high of 5.82 children. This procedure reduced the number of responses in Yemen. All of the countries except ‘‘observations'' to 24 countries instead of 30,470 Belize, Vietnam, Tajikistan, and Ghana differed sig- children. Country averages were then correlated nificantly from the grand mean. All except one of with country HDI and its three indices (life expec- the high-HDI countries were below the average tancy, education, and GDP), controlling for child effect, whereas all of the low-HDI countries were age and child gender. For any significant correla- above the average effect, meaning that caregivers in tion with a component of the HDI, we then con- low-HDI countries reported using more responses trolled the other two indices of the HDI to remove to child behavior (accounted for by using more the shared variance and isolate the effect to the forms of violence) than did caregivers in the high- independent contribution of the component in HDI countries.
Nonviolence. Overall, 18% of caregivers reported that no one in their household had used anyaggression or violence toward their child in the last Deviation From the Grand Mean for Discipline and month, but there was wide variability across coun- tries. For example, no caregivers in Mongolia Figure 1 shows the percentage of the entire sam- reported that the members of their household had ple of caregivers reporting that they or someone in only responded nonviolently in the last month, their household had engaged in each action toward whereas 49% of caregivers in Albania did so. All of their child in the last month. Eighty percent of care- the countries except Serbia differed significantly givers reported that they or someone else in their from the average effect of country. Country household had explained to the child why some- explained between 11.3% (Cox & Snell R2) and thing was wrong. Six percent of caregivers reported 18.3% (Nagelkerke R2) of the variance in the nonvi- that anyone in their household had beaten the child olence scale, over and above the variance associated with an implement.
with child age and child gender. Caregivers in all Total number of responses to child behavior. We first of the high-HDI countries except Serbia and Belarus examined the total number of responses to child were more likely than the average effect to report Figure 1. Grand means of individual items showing the percentage of the entire sample of caregivers reporting that they or someone intheir household had used each response with their child in the last month.
Lansford and Deater-Deckard responding only with nonviolence; caregivers in all effect; caregivers in only one of the six low-HDI of the low-HDI countries except Guinea-Bissau countries were less likely to report psychological were less likely than the average effect to report aggression than the average effect (see Table 2, sec- that their child had experienced only nonviolence ond ‘‘Odds ratio'' column).
(see Table 2, first ‘‘Odds ratio'' column).
Physical violence. On average, 63% of caregivers Psychological aggression. Across all countries, 66% reported that they or someone in their household of caregivers reported that their children had expe- had used physical violence with their child in the rienced psychological aggression in the last month.
last month, with a range from 28% in Bosnia and The range was large, from 7% of caregivers in Alba- Herzegovina to 84% in Jamaica. All countries nia to 89% of caregivers in Yemen. All countries except Serbia, Belize, and Guinea-Bissau differed except Serbia differed significantly from the aver- significantly from the average effect of country.
age effect of country. Country accounted for Country explained between 9.9% (Cox & Snell R2) between 13.6% (Cox & Snell R2) and 18.8% and 13.5% (Nagelkerke R2) of the variance in physi- (Nagelkerke R2) of the variance in psychological cal violence over and above the variance associated aggression over and above the variance associated with child age and child gender. Caregivers in five with child age and child gender. Caregivers in five of the seven high-HDI countries were less likely to of the seven high-HDI countries were less likely to report that their child had experienced physical report psychological aggression than the average violence than the average effect; caregivers in the Table 2Nonviolence and Psychological Aggression: Logistic Regression Parameter Estimates for Country Deviation From the Overall Effect Nonviolence scale Psychological aggression scale Bosnia and Herzegovina Syrian Arab Republic Central African Republic Note. All odds ratios are significant at p < .05, except those designated with a  . HDI = Human Development Index.
aResults for Mongolia did not converge in the model because there is no variance in that country.
Childrearing Discipline and Violence low-HDI countries were more likely to report that child gender. Caregivers in all high-HDI countries their child had experienced physical violence than except Macedonia were less likely than the average the average effect. Caregivers in the medium-HDI effect to use severe physical violence, and all low- countries varied in whether they were more or less HDI countries were more likely than the average likely to report that their child had experienced effect to use severe physical violence (see Table 3, physical violence than the average effect (see second ‘‘Odds ratio'' column).
Table 3, first ‘‘Odds ratio'' column).
Severe physical violence. Overall, 16% of caregiv- Deviation From the Grand Mean for Discipline and ers reported that they or someone in their house- hold had used severe physical violence with theirchild in the last month. Use of severe physical vio- Across countries, 29% of caregivers believed that lence in the last month ranged from a low of 1% in physical punishment is necessary to rear a child Belarus, Kazakhstan, and Ukraine to a high of 40% properly. There was wide variability across coun- in Mongolia and Yemen. All countries except Serbia tries in the percentages of caregivers reporting the differed significantly from the average effect of belief that physical punishment is necessary, rang- country. Country accounted for between 10.6% ing from 4% in Albania to 93% in the Syrian Arab (Cox & Snell R2) and 17.9% (Nagelkerke R2) of the Republic. All countries except Belize and Guinea- variance in using severe physical violence over and Bissau differed significantly from the average effect above the variance associated with child age and of country. Country accounted for between 26.6% Table 3Physical Violence and Severe Physical Violence: Logistic Regression Parameter Estimates for Country Deviation From the Overall Effect Physical violence scale Severe physical violence scale Bosnia and Herzegovina Syrian Arab Republic Central African Republic Note. All odds ratios are significant at p < .05, except those designated with a  . HDI = Human Development Index.
Lansford and Deater-Deckard (Cox & Snell R2) and 37.9% (Nagelkerke R2) of the dren's behavior in the last month ranged from a variance in caregivers' belief in the need to punish low of 6% of caregivers for beating the child with the child physically, over and above the variance an implement to a high of 80% of caregivers for associated with child age and child gender. Care- explaining to the child why something was wrong.
givers in all of the countries with high HDI were Table 5 also depicts the range of the percentages of less likely to report that it was necessary to punish caregivers who reported that they or someone in their child physically than the average effect, but their household had used each response to chil- caregivers in all of the countries with low HDI were dren's behavior in the last month by country. For more likely to report that it was necessary to pun- example, explaining why something was wrong ish their child physically than the average effect was used by as few as 65% of the caregivers (in (see Table 4, ‘‘Odds ratio'' column).
Kazakhstan) and as many as 92% of the caregivers Results of deviation contrasts for individual (in Ukraine). The percentage of variance accounted items reflecting caregivers' responses to children for by country ranged from a low of 3.9% (Cox & are summarized below. As shown in Table 5 and Snell R2) and 6.2% (Nagelkerke R2) in explaining Figure 1, across all countries, the percentages of why something was wrong to a high of 17.9% (Cox caregivers who reported that they or someone in & Snell R2) and 25.2% (Nagelkerke R2) in giving the their household had used each response to chil- child something else to do, over and above the vari-ance associated with child age and child gender(see Table 5).
Table 4Need to Punish Physically: Logistic Regression Parameter Estimatesfor Country Deviation From the Overall Effect Associations With the HDI Need to punish physically To evaluate relations of the discipline and vio- lence items with the HDI and its indices, we aggre- gated the data across countries, resulting in an N of 24 countries. Because the participants were aver- aged across countries, the power for the following Logistic Regression Parameter Estimates for Country Deviation From the Overall Effect of Responses to Child's Behavior in the Last Month Bosnia and Herzegovina Response to child's something was wrong Syrian Arab Republic Gave child something Took away privileges Spanked with a hand Hit child's extremities Hit child with an Hit child on the head Beat child with an Central African Republic Note. The first number in each R2 range is the Cox and Snell R2; the second number in each R2 range is the Nagelkerke R2.
aThe percentage of the entire sample of caregivers reporting that Note. All odds ratios are significant at p < .05, except those they or someone in their household had used each response with designated with a  . HDI = Human Development Index.
their child in the last month.
Childrearing Discipline and Violence tests is low, and they should be interpreted accord- that they and others in their household had ingly. Table 6 displays partial correlations of the responded to their child in the last month using HDI and its indices with the discipline and violence psychological aggression, physical violence, or scales and individual items, controlling for average severe physical violence. Reporting nonviolence child age and the percentage of boys in each coun- was correlated with a higher education index, a try's sample. If partial correlations with an HDI higher literacy rate, and a higher GDP index. Psy- component were significant, we then also con- chological aggression was correlated with a lower trolled for the other two major indices of the HDI education index and lower literacy rate. Physical (presented after the slash in Table 6). The HDI is violence was correlated with a lower education multidimensional and, although life expectancy, index, a lower literacy rate, and lower schooling.
education, and GDP are related to one another, it is Severe physical violence was correlated with a also possible that they relate in different ways to lower GDP index. With the exception of the correla- child development and parenting. Controlling the tion between physical violence and the education other two indices of the HDI allowed us to remove index, these correlations were no longer significant the shared variance among indices and give a more after controlling for the other indices of the HDI.
precise estimate of the effects of the component in The HDI and its indices were not significantly question. Because literacy and schooling could have correlated with 5 of the 11 individual response different effects, we also included these two compo- items, namely, whether the caregiver or someone nents of the education index.
else in the household in the last month had: (a) As shown in Table 6, the HDI was significantly given the child something else to do, (b) taken correlated with all four discipline and violence away privileges, (c) yelled, (d) spanked on the bot- scales. Higher HDI was correlated with a higher tom with a hand, or (e) beaten the child with an percentage of caregivers reporting that they and implement. The HDI was significantly correlated others in their household had responded to their with the remaining six items: (a) explaining why child in the last month only with nonviolence (i.e., something was wrong, (b) name calling, (c) hitting no psychological aggression or physical violence) the child's extremities, (d) shaking, (e) hitting with and with a lower percentage of caregivers reporting an object, and (f) belief in the necessity of physical Table 6Partial Correlations of Human Development Index (HDI) With Discipline and Violence, Controlling for Average Child Age and Percent Male Chil-dren Psychological aggression Physical violence Severe physical violence Explained why something was wrong Gave child something else to do Took away privileges ).82*** ⁄ ).48* Spanked with a hand Hit child's extremities Hit child with an object ).75*** ⁄ ).69** ).78*** ⁄ ).74*** Hit child on the head Beat child with an implement Need to physically punish Note. N = 24 countries. Partial correlations after the slash are controlling for child age and gender and the other two indices thatcompose the HDI. GDP = gross domestic product.
*p < .05. **p < .01. ***p < .001.
Lansford and Deater-Deckard punishment. When the separate indices of the HDI The most frequently reported discipline technique were examined, the life expectancy index was signif- in all countries was explaining to the child why icantly correlated with three items: name calling, something was wrong. Overall, this finding suggests shaking, and hitting with an object. The education that caregivers in developing countries appear to index was significantly correlated with six items: recognize the value in socializing children with name calling, hitting the child's extremities, shaking, explanations for why their misbehaviors are wrong, hitting with an object, hitting the child's head, and with country accounting for only between 4% and believing that physical punishment is necessary.
6% of the variance in whether caregivers reported The GDP index was significantly correlated with six that someone had explained to their child why items: explaining why something was wrong, name something was wrong. In all except one country (Ka- calling, hitting the child's extremities, shaking, hit- zakhstan), more than 70% of caregivers reported that ting with an object, and believing that physical pun- someone in their household had explained to their ishment is necessary. However, after controlling for child why something was wrong at least once in the the other HDI indices, the only HDI index that last month. Reasoning and explanations have been remained significantly correlated with the use of shown to contribute to children's empathy, prosocial any responses was the education index in relation to behavior, and moral internalization of parents' mes- three items: name calling, hitting the child's extremi- sages (e.g., Krevans & Gibbs, 1996). Thus, it is reas- ties, and hitting with an object; caregivers in coun- suring that offering explanations is the most tries with higher education indices were less likely frequently reported response to children's behavior to report name calling, hitting the child's extremi- across countries. Also, this near universality of use ties, and hitting with an object in the last month.
of reasoning and explanations probably reflects par-ents' acknowledgment that as a discipline method,reasoning is effective most of the time. However,20% of caregivers across the 24 countries, on aver- age, could not recall a time during the last month This study examined the prevalence and country- when they or anyone in their household had offered level correlates of 11 forms of discipline and the child an explanation regarding why something violence, as well as the belief in the necessity of was wrong. Other nonviolent discipline techniques, physical punishment, in 24 countries using data such as giving the child something else to do or from 30,470 families. The prevalence of each removing privileges, were less common. This overall response and beliefs regarding the necessity of pattern suggests the need for countries to promote physical punishment varied widely across coun- the use of positive, nonviolent verbal communica- tries. Country-level indicators of life expectancy, tion and behavioral management techniques.
educational attainment, and economic well-being Country also accounted for relatively little vari- were unrelated to giving the child something else ance in taking away privileges (between 7% and to do, taking away privileges, shouting, spanking 9%) and spanking on the bottom with a hand with a hand, and beating the child with an imple- (between 6% and 9%). The small percentage of vari- ment. By contrast, the three country-level indicators ance accounted for by country for these two (particularly the education index and literacy) were responses to children's behavior suggests that these related to explaining why something was wrong, practices do not vary by country-level factors. In all name calling, hitting the child on the extremities, except Macedonia, more than 20% of caregivers shaking, hitting the child with an object, and hitting reported that privileges had been taken away from the child on the head. The finding that the educa- their child in the last month. In all countries, at tion components of the HDI were the only compo- least 20% of caregivers reported that their child had nents that remained significantly correlated with been spanked on the bottom with a hand in the last discipline and violence after controlling for the month. Considerably more variance could be other indices of the HDI suggests the primacy of accounted for by country of residence in the case of education in shaping this domain of parenting in the other responses to child behavior. The variabil- developing countries. Overall, the findings suggest ity associated with country was especially large for that country-level factors are related to discipline harsher physical violence (e.g., from 1% to 44% of and violence; in particular, harsh forms of physical caregivers reported that the target child had been violence (with the exception of beating the child hit with an object in the last month).
with an implement, which was rare in most coun- This particularly wide variability across coun- tries) were related to country-level factors.
tries in the use of violence suggests that country- Childrearing Discipline and Violence level factors (e.g., cultural norms, policies) warrant remorse, and agitation when dealing with their additional attention to understand more clearly why the use of violence is widespread in some beliefs and behavior may also imply that the cul- countries but not in others. For example, the preva- ture is experiencing a fairly rapid shift in prevailing lence of severe violence ranged from 1% in Ukraine beliefs about appropriate ways to respond to chil- to 40% in Mongolia. What differs between these dren's behavior. That is, in cultures where norms countries that could account for such a vast differ- are shifting, parents may still use responses that ence? Ukraine adopted a full prohibition of were used in the past but may report greater physical punishment in 2003 (http://www.endcor- ambivalence about these responses, on average. For poralpunishment.org); both the legal prohibition as example, Straus (1996) found that 54% of mothers well as attitudes related to adults' responses to chil- in an American sample reported that in over half of dren that enabled the prohibition to be passed the times in which they had used physical punish- appear to be reflected in rates of violence against ment, it was the wrong response to have used. The children in Ukraine. By contrast, violence against gap between beliefs about the necessity of physical children is much more common in some countries, punishment and actual behavior was generally as evidenced by the fact that 40% of Mongolian smaller in countries where caregivers were more caregivers had seen someone in the home beat a likely to report that they believed physical punish- child in the last month, and 44% of Gambian care- ment was necessary. For example, in Ghana, 46% of givers witnessed a child being hit with an object in caregivers believed physical punishment was nec- the last month. However, just because an act is essary, and 50% reported that their child had been common or is not seen by parents as having delete- spanked in the last month. In these countries, it is rious effects on children does not mean that it possible that physical punishment is used more should be accepted. Indeed, one of the major tenets instrumentally rather than as an angry emotional in the international community, as recognized in response. It would be useful for future research to the CRC, is that children in all countries have the explore how differences between caregiver beliefs right to protection from violence.
and behaviors might contribute to how parents' Within the United States, differences among responses affect children.
ethnic groups have been found in beliefs about Just as within-country analyses based on Ameri- the appropriateness and effectiveness of physical can samples have shown that more educated par- punishment (Maker, Shah, & Agha, 2005; Mosby, ents are less likely to use physical punishment than Rawls, Meehan, Mays, & Pettinari, 1999). In the are less educated parents (Straus & Stewart, 1999), present study, country of residence accounted for the present between-country analyses support the between 27% and 38% of the variance in whether hypothesis that caregivers in countries with higher caregivers believed that it was necessary to physi- literacy rates and higher rates of education are less cally punish children. This finding suggests that a likely to report that their child has experienced psy- great deal of variability in caregivers' beliefs about chological aggression or physical violence. Parents' physical punishment can be accounted for by coun- knowledge and values may change as they advance try-level factors and that future research should in formal education (Davis-Kean, 2005), which may delve into what such country-level factors may be.
help account for the relations we found between Fewer caregivers believed that it was necessary prevalence rates of violence against children and to physically punish children than actually did so.
country-wide literacy and other educational oppor- For example, only 5% of caregivers in Montenegro believed that using physical punishment was neces- Several limitations should be acknowledged.
sary, but 47% reported that they or someone in First, because the data were reported by primary their household had spanked their child in the last female caregivers, they may reflect underreporting month. When beliefs and practices related to physi- if someone in the household besides the respondent cal punishment are discordant, this might suggest had interacted with the child in a particular way that physical punishment is being used in an emo- without the respondent's knowledge or if the tional rather than instrumental manner (e.g., as an respondent did not report psychological aggression angry response executed in the heat of the or physical violence because of concerns about moment). Indeed, Graziano and Hamblen (1996) social desirability. The concern about social desir- found that 85% of the middle-class, primarily Euro- ability is somewhat mitigated by the fact that pean American parents in their sample reported respondents reported on the actions of everyone in experiencing moderate to high levels of anger, the household, not just their own behaviors, but Lansford and Deater-Deckard they may still have wanted to present the house- The CRC has focused governments around the hold in a favorable light. Second, each item refer- world on reducing violence against children and enced behavior toward the child during the last month. Therefore, it is important to keep in mind responses to children. All 24 countries that provided that the percent of children in each country who data for this study have ratified the CRC, indicating had ever experienced each caregiver response their commitment to child protection. The findings would be higher than the figures presented here.
reported here suggest wide variation across coun- Third, the responses were dichotomous, providing tries with respect to the use of violence against chil- information about whether a given response had dren; countries with low levels of educational been used at all in the last month but not how attainment are at particularly high risk for violence frequently it was used or under what circum- against children. Given both the widespread use of stances. Fourth, we do not have available informa- violence and the widespread belief in the necessity tion about change over time in responses to of using physical punishment in some countries, children. Although the analyses present a snapshot efforts to eliminate violence against children will of responses used during the course of a single need to alter the belief that physical punishment is month, we recognize that caregivers' responses necessary to rear a child as well as provide caregiv- change over time, both as a function of the child's ers with nonviolent alternatives to replace violence.
development and as a function of historical andcontextual changes in societal standards related toresponding to children's behavior. Fifth, our ability to interpret and contextualize the findings wouldbe enhanced by a rich ethnographic understanding Alyahri, A., & Goodman, R. (2008). Harsh corporal pun- of the cultural beliefs, policies, and laws that con- ishment of Yemeni children: Occurrence, type and tribute to parenting in each country. Sixth, the associations. Child Abuse and Neglect, 32, 766–773.
number of caregivers sampled from different coun- Annerba¨ck, E.-M., Svedin, C.-G., & Gustafsson, P. A.
tries varied widely, ranging from a low of 202 to a (2010). Characteristic features of severe child physicalabuse—A multi-informant approach. Journal of Family high of 3,316. Finally, family patterns can and do Violence, 25, 165–172.
vary considerably both within and across samples.
Bornstein, M. H., Britto, P. R., Nonoyama-Tarumi, Y., Diversity in family socialization environments was Ota, Y., Petrovic, O., & Putnick, D. L. (2012). Child apparent in the earliest societies and continues in development in developing countries: Introduction and contemporary societies. Although we have provided methods. Child Development, 83, 16–31.
standard deviations for our measures, this study Bornstein, M. H., & Lansford, J. E. (2009). Parenting. In focuses on variation across, not within, countries.
M. H. Bornstein (Ed.), The handbook of cross-cultural This study also has important strengths. The developmental science (pp. 259–277). New York: Taylor & availability of data from 24 countries that are under- represented in developmental science is a major Collins, W. A., Madsen, S. D., & Susman-Stillman, A. (2002).
strength. Indeed, 96% of research participants in Parenting during middle childhood. In M. H. Bornstein(Ed.), Handbook of parenting; Vol. 3. Being and becoming a psychological studies are from countries with only parent (2nd ed., pp. 73–101). Mahwah, NJ: Erlbaum.
12% of the world's population (Henrich, Heine, & Conger, R. D., & Donnellan, M. B. (2007). An interaction- Norenzayan, 2010). Our study makes a contribution ist perspective on the socioeconomic context of to the growing literature that attempts to increase human development. Annual Review of Psychology, 58, international perspectives and bring cultural and contextual awareness to the study of parenting and Crouch, J. L., & Behl, L. E. (2001). Relationships among child development (Bornstein & Lansford, 2009).
parental beliefs in corporal punishment, reported Another strength is the large and nationally repre- stress, and physical child abuse potential. Child Abuse sentative samples available in each country. No indi- & Neglect, 25, 413–419.
vidual research team would be able to execute the Davis-Kean, P. E. (2005). The influence of parent educa- probability sampling used in all of the participating tion and family income on child achievement: The indi-rect role of parental expectations and the home countries, but the cooperation of UNICEF and gov- environment. Journal of Family Psychology, 19, 294–304.
ernment agencies within each country made possi- Deater-Deckard, K. (2004). Parenting stress. New Haven, ble this sampling success. Yet another strength is the CT: Yale University Press.
conceptual focus on violence at a historical period Durrant, J. E. (2004). Physical punishment and physical that is marked by increasing efforts by the global abuse. Children: A Newsletter From the Office of the Chil- community to eliminate violence against children.
dren's Commissioner, 50, 4–6.
Childrearing Discipline and Violence Durrant, J. E. (2008). Physical punishment, culture, and families: Historical and theoretical foundations (pp. 233– rights: Current issues for professionals. Journal of Devel- 257). Thousand Oaks, CA: Sage.
opmental & Behavioral Pediatrics, 29, 55–66.
Monyooe, L. A. (1996). Teachers' views towards corporal Ember, C. R., & Ember, M. (2005). Explaining corporal punishment in Lesotho schools. Psychological Reports, punishment of children: A cross-cultural study. Ameri- 79, 121–122.
can Anthropologist, 107, 609–619.
Mosby, L., Rawls, A. W., Meehan, A. J., Mays, E., & Fung, H. (1999). Becoming a moral child: The socializa- Pettinari, C. J. (1999). Troubles in interracial talk about tion of shame among young Chinese children. Ethos, discipline: An examination of African American child 27, 180–209.
rearing narratives. Journal of Comparative Family Studies, Garcia-Coll, C., & Magnuson, K. (1999). Cultural influ- 30, 489–521.
ences on child development: Are we ready for a para- Oburu, P. O., & Palme´rus, K. (2003). Parenting stress and digm shift? In A. S. Masten (Ed.), Cultural influences in self reported discipline strategies of Kenyan grand- child development (pp. 1–24). Mahwah, NJ: Erlbaum.
mothers taking care of their orphaned grandchildren.
Gershoff, E. T. (2002). Corporal punishment by parents International Journal of Behavioral Development, 27, 505– and associated child behaviors and experiences: A meta-analytic and theoretical review. Psychological Bul- Peterson, G. W., & Hann, D. (1999). Socializing children letin, 128, 539–579.
and parents in families. In M. B. Sussman, S. K. Stein- Gershoff, E. T., Grogan-Kaylor, A., Lansford, J. E., Chang, metz, & G. W. Peterson (Eds.), Handbook of marriage and L., Zelli, A., Deater-Deckard, K., et al. (2010). Parent the family (2nd ed., pp. 327–370). New York: Plenum.
discipline practices in an international sample: Associa- Pinheiro, P. S. (2006). Report of the independent expert for tions with child behaviors and moderation by per- the United Nations study on violence against children.
ceived normativeness. Child Development, 81, 487–502.
United Nations General Assembly. Retrieved January Graziano, A. M., & Hamblen, J. L. (1996). Subabusive vio- 10, 2010, from http://www.unicef.org/violencestudy/ lence in child rearing in middle-class American fami- lies. Pediatrics, 98, 845–848.
Sadowski, L. S., Hunter, W. M., Bangdiwala, S. E., & Henrich, J., Heine, S. J., & Norenzayan, A. (2010). The Munoz, S. R. (2004). The World Studies of Abuse in the weirdest people in the world? Behavioral and Brain Sci- Family Environment: A model of a multi-national ences, 33, 1–75.
study of family violence. Injury Control and Safety Pro- Kariger, P., Frongillo, E. A., Engle, P., Rebello Britto, P.
motion, 11, 81–90.
M., Sywulka, S. M., & Menon, P. (2010). Indicators of Smith, D. E., & Mosby, G. (2003). Jamaican child-rearing family care for development for use with multi-country sur- practices: The role of corporal punishment. Adolescence, veys. Manuscript submitted for publication.
38, 370–381.
Kim, E., & Hong, S. (2007). First generation Korean Straus, M. A. (1996). Spanking and the making of a vio- American parents' perceptions of discipline. Journal of lent society. Pediatrics, 98, 837–842.
Professional Nursing, 23, 60–68.
Straus, M. A., Hamby, S. L., Finkelhor, D., Moore, D. W., Krevans, J., & Gibbs, J. C. (1996). Parents' use of inductive & Runyan, D. (1998). Identification of child maltreat- discipline: Relations to children's empathy and proso- ment with the Parent-Child Conflict Tactics Scales: cial behavior. Child Development, 67, 3263–3277.
Development and psychometric data for a national sam- Lansford, J. E., Chang, L., Dodge, K. A., Malone, P. S., ple of American parents. Child Abuse and Neglect, 22, Oburu, P., Palme´rus, K., et al. (2005). Cultural norma- tiveness as a moderator of the link between physical Straus, M. A., & Stewart, J. H. (1999). Physical punish- discipline and children's adjustment: A comparison of ment by American parents: National data on pre- China, India, Italy, Kenya, Philippines, and Thailand.
valence, chronicity, severity, and duration, in relation Child Development, 76, 1234–1246.
to child and family characteristics. Clinical Child and Maker, A. H., Shah, P. V., & Agha, Z. (2005). Child physi- Family Review, 2, 55–70.
cal abuse: Prevalence, characteristics, predictors, and United Nations Children's Fund. (2006). Multiple Indicator beliefs about parent-child violence in South Asian, Mid- Cluster Survey manual 2005: Monitoring the situation of dle Eastern, East Asian, and Latina women in the Uni- children and women. New York: Author.
ted States. Journal of Interpersonal Violence, 20, 1406–1428.
United Nations Development Programme. (n.d.). The human McLoyd, V. C. (1990). The impact of economic hardship development index (HDI). Retrieved July 10, 2009, from on Black families and children: Psychological distress, parenting, and socioemotional development. Child Whipple, E. E., & Richey, C. A. (1997). Crossing the line Development, 61, 311–346.
from physical discipline to child abuse: How much is Mistry, J., Chaudhuri, J. H., & Diez, V. (2003). Ethnotheo- too much? Child Abuse and Neglect, 21, 431–444.
ries of parenting: At the interface between culture and Whiting, B. B., & Whiting, J. W. (1975). Children of six cul- child development. In R. M. Lerner, D. Wertlieb, & tures: A psycho-cultural analysis. Cambridge, MA: Har- F. Jacobs (Eds.), Handbook of applied developmental sci- vard University Press.
ence: Vol. I. Applying developmental science for youth and

Source: https://childandfamily.staging.properdesign.co.uk/wp-content/uploads/2014/12/Lansford-research2.pdf