Journal of Abnormal Child Psychology

, Volume 42, Issue 5, pp 717–730

Parental Physical Punishment and Adolescent Adjustment: Bidirectionality and the Moderation Effects of Child Ethnicity and Parental Warmth

Authors

    • School of Education, Department of Psychology, Learning Research & Development CenterUniversity of Pittsburgh
  • Sarah Kenny
    • Institute for Social ResearchUniversity of Michigan
Article

DOI: 10.1007/s10802-013-9827-8

Cite this article as:
Wang, M. & Kenny, S. J Abnorm Child Psychol (2014) 42: 717. doi:10.1007/s10802-013-9827-8

Abstract

This study used cross-lagged modeling to examine reciprocal relations between maternal and paternal physical punishment and adolescent misconduct and depressive symptoms, while accounting for stability in both physical punishment and adjustment problems over time. Data were drawn from a sample of 862 two-parent families and their adolescent children (52 % males; 54 % European American; 44 % African American; 2 % other ethnic backgrounds). Mothers’ and fathers’ physical punishment of their adolescents’ ages 12 and 14 predicted increased misconduct and depressive symptoms among these adolescents at ages 14 and 16. Adolescent misconduct, but not depressive symptoms, at ages 12 and 14 predicted increased physical punishment by their parents at ages 14 and 16. Neither parental warmth nor child ethnicity moderated the longitudinal relationship between parental physical punishment and adolescent adjustment. Patterns of findings were similar across mothers and fathers.

Keywords

AdolescencePhysical punishmentMisconductDepressive symptomsParental warmthTransactional process

The topic of physical punishment remains controversial among researchers and practitioners. Despite this controversy, great numbers of American parents continue to use physical punishment in order to obtain compliance from their children, even during adolescence (Bender et al. 2007; Collins et al. 2002). Straus and Stewart (1999) reported that approximately 50 % of American parents physically punished their children at age 12, 34 % at age 14, and 15 % at age 17. In 2010, a nationally representative survey reported that 77 % of American men and 65 % of American women agreed that sometimes an older child needs to be spanked (Child Trends 2012).

Physical punishment could be particularly detrimental during adolescence, when young people are developing their sense of identity (Gershoff 2002; Gunnoe and Mariner 1997). For the adolescent who experiences physical punishment, the family home may cease to represent a secure base, and he or she may develop a maladaptive view of relationships (Bender et al. 2007). The fragile and sometimes hostile parent–child relationship may make salient developmental tasks such as autonomy and relatedness difficult to accomplish (Allen et al. 1994). Indeed, a wide array of research has demonstrated that parents’ use of physical punishment is associated with high rates of externalizing and internalizing behaviors in adolescents (Sheehan and Watson 2008; Simons et al. 2000). However, much of this research fails to utilize a multi-wave longitudinal research design, which would account for the transactional nature of the associations between parent physical punishment and adolescent behavior. A transactional approach is particularly relevant for research investigating how parenting practices and adolescent problem behaviors influence each other over time. The bidirectional relationship between parents and children is likely to change as the child ages, given the developmental shift from unilateral authority to more egalitarian relationships (Hartup 1989; Hipwell et al. 2008). Yet, few studies have tested whether such reciprocal associations continue and can be traced from early to middle adolescence. Understanding the nature of the bidirectional relationship between parental disciplinary strategies and adolescent adjustment behavior is critical for the development of early interventions to prevent both the onset and acceleration of behavioral and emotional problems.

Furthermore, research investigating physical punishment has focused primarily on mothers. It is true that research tends to suggest that mothers engage in higher numbers of disciplinary acts than fathers (Straus and Stewart 1999). This difference is generally attributed to the fact that mothers typically spend more time with their children than do fathers. However, focusing solely on mothers’ physical punishment neglects the possibility that the father may also use physical punishment or may even be the primary disciplinarian. It also fails to evaluate whether maternal and paternal physical punishment may have differential effects on the behavioral outcomes of children. For example, Bender and colleagues (2007) found that both maternal and paternal physical discipline were related to greater adolescent behavioral problems at age 16, but that maternal physical discipline was associated with adolescent depressive symptoms over and above paternal physical discipline. As such, preliminary findings point to the importance of examining the roles of both mothers and fathers in the relationship between physical punishment and adolescent adjustment.

Finally, many studies ignore the fact that harsh discipline may occur within the context of varying degrees of positive parenting behaviors. Some researchers have suggested that parental warmth may moderate the link between harsh discipline and adolescent adjustment (e.g., McLoyd and Smith 2002; Simons et al. 2000). However, results showing the moderation effect of parental warmth on physical punishment have been mixed and no studies have investigated the moderation effect of parental warmth in the reciprocal relation between physical punishment and adolescent adjustment.

The present study used measures of paternal and maternal physical punishment and adolescent adjustment from independent sources at multiple time points to examine reciprocal associations between physical punishment and adolescents’ conduct problems and depressive symptoms in a longitudinal study following adolescents from ages 12 to 16. In addition, we investigated whether the longitudinal association between physical punishment and adolescent adjustment was moderated by child ethnicity or parental warmth.

Parental Physical Punishment and Adolescent Problem Behavior

A large body of research on physical punishment has provided empirical evidence for its negative influence upon adolescent adjustment. A comprehensive meta-analysis by Gershoff (2002) found that physical punishment is related to a variety of behavioral and psychological problems. Indeed, many recent studies suggest that physical punishment is associated with increased externalizing behaviors such as delinquency and aggression in adolescents (Dodge et al. 1990; Lansford et al. 2009; Lansford et al. 2012; Palmer and Hollin 2001). Frequent exposure to physical punishment in adolescence is also related to an increase in internalizing behaviors, including depressive symptoms and anxiety (Bender et al. 2007; Holmes and Robins 1987; Turner and Muller 2004). Although a large body of research has examined parental physical punishment as a predictor of negative adolescent behaviors, adolescents who exhibit problem behaviors may also elicit physical punishment from their parents. Parents may use more harsh forms of discipline in an effort to reduce increasingly difficult adolescent behavior. A number of studies suggest that adolescent externalizing behavior influences the development of diminishing levels of positive parenting over time (Larzelere 2000; Lytton 1990; Patterson et al. 1990; Shaw et al. 2004). For example, Pettit et al. (2001) found that high levels of child conduct problems predicted low levels of parental monitoring in early adolescence.

Evidence of both unidirectional parent and child effects provides a foundation for the examination of a transactional model, which highlights bidirectionality in the parent–child relationship. Sameroff’s (1975) transactional model asserts that there is a continuous dynamic interplay between children’s behavior and parents’ reactions to this behavior, such that the child’s behavior elicits a parental behavior, which in turn elicits future child behavioral responses. This echoes Patterson’s social coercion theory (Patterson 2002; Patterson et al. 1992), which points to an escalating cycle of misbehavior on the part of the child and negative reactions on the part of the parents. Higher levels of antisocial behavior may lead parents to resort to physical punishment, in turn increasing the child’s problematic behavior. It is clear that a focus on both parent and child effects is necessary in order to understand the development of interpersonal dynamics in a family. Given that adolescence is a period in which there is generally an emergence and escalation of problem behaviors (Hipwell et al. 2008), focusing on both parent and child effects will allow us to further comprehend the onset and course of problematic outcomes resulting from parental physical punishment.

Despite the theoretical framework indicating a reciprocal association between parenting and adolescent behavior, current empirical evidence from literature concerning parental physical punishment and the bi-directionality between parent and child behaviors is inconsistent and limited. Regarding externalizing behavior, Pardini et al. (2008) found that conduct problems among males aged 6-16 years influenced changes in parenting behaviors (including physical punishment) as strongly as the parenting behaviors influenced changes in conduct problems. Yet, this finding is limited as the authors did not model the bidirectional relations simultaneously but instead examined the parent and child effects separately. In contrast, Sheehan and Watson (2008) found that mothers’ use of aggressive discipline predicted an increase in child aggression at all ages (7-19 years), whereas child aggression predicted an increase in maternal use of aggressive discipline at younger ages (8-15), but not at older ages (12-19). However, some of the items used to measure severe aggressive discipline could be characterized as physical abuse and may have confounded findings, as it is likely that the effects of physical discipline differ by the degree of severity (Larzelere 2000). Lansford et al. (2011) recently conducted the first long-term longitudinal study of the reciprocal relations between parents’ physical discipline and youth’s externalizing behavior. In a sample of high risk adolescent boys aged 10-15 years, they found that physical discipline in a given year predicted antisocial behavior in the next year, but antisocial behavior did not predict parents’ use of physical discipline. However, their use of a sample of high-risk adolescent boys from low-income families limits the generalizability of its conclusions.

Less research has examined the reciprocal relationships between parental physical punishment and adolescent internalizing behavior. Indeed, we found only one study that investigated the reciprocal relationship between parental harsh punishment and youth’s depressive symptoms (Hipwell et al. 2008). This study found that parental harsh punishment predicted increases in depressed mood in adolescent girls, whereas depressed mood did not predict increases in harsh punishment over time. It has been shown that depressive symptoms in young people may elicit stress in interpersonal relationships (Hammen 2006). It is possible that as an internalizing problem, adolescent depressive symptoms provoke a different response from parents than do conduct problems. Specifically, externalizing problems may lead to a harsh disciplinary response from parents, whereas internalizing problems may instead lead parents to withdraw and show less affection (Pettit and Ariswalla 2008). However, Hipwell and colleagues’ (2008) findings are limited to younger girls, and child effects may become more powerful during adolescence as youth become more actively involved in initiating and controlling family interactions (Stice and Barrera 1995). Another notable limitation is that they combined items from a psychological aggression scale with spanking to create a composite of harsh punishment, making comparison and integration of the literature on physical discipline difficult.

Ethnic Differences

Research suggests that African American parents administer physical punishment to a greater extent than European American parents (Child Trends 2012; Lansford et al. 2004; McLoyd and Smith 2002). This may be because African American families are more likely to endorse an authoritarian parenting style, which emphasizes the importance of parental control and child compliance (Baumrind 1971; Giles-Sims et al. 1995; Wang & Huguley 2012). Furthermore, there is evidence to suggest that physical punishment is less psychologically harmful in communities where it is more prevalent and considered more normative (Simons et al. 2002). It is possible that the cultural normative context in which the discipline is administered changes the meaning of the discipline for the child and in turn moderates its adverse effects (Deater-Deckard and Dodge 1997). For example, parental physical punishment has been shown to have differential effects on behavioral outcomes in European American and African American children. Some studies have found that African American children experience fewer negative outcomes as a result of physical discipline than European American children (Deater-Deckard et al. 1996). For example, Gunnoe and Mariner (1997) found that whereas corporal punishment was related to more instances of aggressive behavior among European American children aged 4-11 years, it was actually related to fewer instances among African American children. Lansford et al. (2004) reported similar findings in a longitudinal study of children aged 5-12 years old where the experience of physical discipline at each time point was related to higher levels of subsequent externalizing behaviors among European American adolescents but lower levels of externalizing behaviors among African American adolescents.

In contrast, other researchers report contradictory findings. For example, McLoyd and Smith (2002) found no differences in 4-12 year old child outcomes over time by ethnicity, despite a greater frequency of spanking in African American families compared to European American and Latino families. Similar findings were reported by Grogan-Kaylor (2004) and Gershoff et al. (2012). Gershoff et al. (2012) found ethnic differences in the frequency of spanking among the mothers of 5-8 year olds, but not in associations between spanking and externalizing behavior. However, it is important to note that most of the research examining ethnic differences has focused on children, not adolescents. Adolescents may be particularly sensitive to the effects of physical discipline (Turner and Finkelhor 1996) and any potential moderating effects of ethnicity may be less likely to operate during this developmental period. For example, it is possible that during adolescence, African American youth are unable to interpret physical discipline within the protective framework of cultural normativity, viewing it instead as indicative of parental hostility and rejection. Given the conflicting evidence from prior studies, we examine whether the reciprocal association between physical discipline and adolescent problem behaviors varies by ethnicity.

Parental Warmth as a Moderator

Some researchers argue that the effects of physical punishment on child behavior may be buffered by positive parenting behaviors such as parental warmth (Lansford et al. 2010; McLoyd and Smith 2002). Positive and negative parenting behaviors coalesce to construct the family and relational context in which adolescents interpret parenting behavior. The child may be less likely to see physical punishment as unjust or harsh and indicative of rejection if it is administered in a general context of parental warmth and acceptance. In support of this hypothesis, some studies have found that parental physical punishment is not associated with increased adolescent behavior problems when it is administered with higher levels of parental warmth (Rohner et al. 1996; Simons et al. 2000). Simons et al. (2000) found that the relationship between physical punishment and 7th graders’ aggression was strongest when maternal warmth, induction, and monitoring were low. Moreover, parental warmth was particularly protective against harsh discipline for internalizing compared to externalizing problems. Specifically, the positive association between corporal punishment and problematic emotional adjustment was strongest for children aged 10-14 from families low in warmth and support (Aucoin et al. 2006).

However, other research indicates that parental warmth does not alleviate the stress of frequent corporal punishment on youth and may even exacerbate the negative impact of physical punishment (Lee et al. 2013). For example, Turner and Finkelhor (1996) found that the strength of the positive association between frequent physical punishment and emotional distress was stronger among 10-16 years old who reported high parental support. The authors suggest that this may be due to the fact that parents who are supportive but administer frequent corporal punishment are inconsistent in their parenting, causing a sense of unpredictability and increasing distress in their children. In addition, using a younger sample of 6-9 years old, Straus et al. (1997) found that corporal punishment was associated with an increase in antisocial behavior, regardless of the extent to which parents provided emotional support. Unfortunately, studies concerning the moderation effect of parental warmth in the relationship between parental physical punishment and adolescent outcomes tend to be cross-sectional, precluding examination of transactional processes between parent and child. Most studies also use a composite of harsh punishment that includes other forms of discipline (e.g., harsh verbal discipline) or potentially abusive punishment. The current study advances our understanding of the potential moderating role of parental warmth by using a clearly defined construct of parental physical punishment and by framing parental and child behaviors as transactional processes.

The Current Study

The main goal of the present study is to examine whether parental physical punishment and adolescents’ misconduct and depressive symptoms are reciprocally related over the course of adolescent development from ages 12 to 16. We explicitly discriminate between physical punishment and other forms of punitive parental behavior, such as physical abuse, harsh verbal discipline, and psychological aggression. We also exclude severe types of physical punishment (e.g., hitting with a fist or object and beating up). In our study, physical punishment is defined specifically as a parent slapping, pushing, grabbing, or shoving his or her child in response to a misdemeanor. The exclusion of overly severe types of physical punishment enables us to examine whether these less severe types of physical punishment still lead to adolescent maladjustment. We hypothesize that there is a bidirectional process that develops over time in which parental physical punishment predicts an increase in misconduct and adolescent misconduct elicits increases in parental physical punishment. We also investigate whether parent effects are as consistently prevalent as child effects and whether the strength of both of these effects changes over the course of adolescent development. In addition, we examine the potential moderating effects of child ethnicity and parental warmth on the longitudinal, reciprocal associations between parental physical punishment and adolescents’ misconduct and depressive symptoms.

In this study, we expand the physical punishment literature in several ways. First, we employ a large, community-based, socioeconomically and ethnically diverse sample of 862 adolescents aged 12-16 years, and include their parents. Second, we use cross-lagged modeling techniques to examine the transactional nature of the associations between parental physical punishment and both adolescent misconduct and depressive symptoms over time. This model accounts for both initial levels and the continuity of problem behavior and physical punishment and, as such, monitors changes in the longitudinal associations. Third, our use of both maternal and paternal ratings of physical punishment and adolescents’ ratings of misconduct and depressive symptoms allows us to compare potentially differential adolescent outcomes as a function of parent gender while preventing shared method variance. Fourth, given the conflicting evidence from prior studies, we examine whether child ethnicity and parental warmth moderate the longitudinal, reciprocal association between physical punishment and adolescent problem behaviors. Finally, the model controls for several important confounds at the first wave in the association between parental physical punishment and adolescent outcomes, including demographic- (child gender, ethnicity, family socioeconomic status) and parent-risk variables (parenting stress and parental depression) that have been found to predict both parental use of harsh discipline and adolescent problem behaviors (Hipwell et al. 2008; Lovejoy et al. 2000).

Method

Participants

The families in the current study were participants in a longitudinal study of youth development. Two-parent families with children attending middle school were recruited from the Eastern United States. Ninety percent of these families included biological parents of the child who were married, 7 % included the biological mother and a stepfather who were married, 2 % included the biological father and a stepmother who were married, and 1 % included biological parents of the child who were cohabiting. We examined three waves of data: Wave 1, collected when the adolescents were at age 12; Wave 2, collected when the adolescents were at age 14; and Wave 3, collected when the adolescents were at age 16. During the first assessment at age 12, data were collected from 862 families (52 % male; 54 % European American, 44 % African American, 2 % other ethnic backgrounds). The sample is broadly representative of varying socioeconomic levels, with a mean pre-tax family annual income of $43,342. Eighty-six percent of families reported employment of at least one parent. Sixty-five percent of parents were high school graduates, 47 % of whom were college graduates.

Eighty-two percent of adolescents had data for all three waves, and an additional 8 % had data for two waves. To explore possible differences in the sample due to attrition, we conducted a series of chi-square tests and t-tests to compare those adolescents with and without complete waves of data on all study variables at Wave 1. Results revealed that those who dropped out of the study did not differ significantly in the outcome variables, predictors, and demographic characteristics from those who started the study. The data were missing completely at random, as evidenced by non-significant results derived from the generalized least squares combined test of homogeneity of means and covariance matrices representing complete and incomplete data, χ2 (862) = 247.52, ns. Missing data were handled through full information maximum likelihood estimation (FIML).

Adolescents and their parents in the 12 participating schools were sent a letter through schools endorsing the research project. Those families interested in participating in the study were asked to sign and return a consent form. Adolescents completed questionnaires at ages 12, 14, and 16. This data collection process took place in adolescents’ schools. The questionnaire for adolescents took approximately 50 min to complete. Parents also completed surveys taking approximately 30 min to complete. At each wave, adolescents were offered $20, and parents were offered $40 for their participation. A review conducted by the IRB approved the study to be consistent with the protection of the rights and welfare of human subjects.

Measures

Physical Punishment

Parental physical punishment was assessed via survey when the adolescents were 12, 14, and 16 years of age. Mothers and fathers reported their own discipline strategies. They were asked: “In the past six months, if your child did something that he/she is not allowed to do or something that you did not like, how often did you: a) slap, b) push, c) grab, or d) shove him/her?” The four items were scaled using a 5-point answer format (0 = never, 1 = 1 ~ 2 times, 2 = 3 ~ 5 times, 3 = 6 ~ 10 times, and 4 = more than 10 times). All the items were averaged to create the construct of physical punishment (mother: αs = 0.72, 0.71, and 0.70; father: αs = 0.69, 0.71, and 0.70).

Depressive Symptoms

Adolescents’ depressive symptoms were assessed at ages 12, 14, and 16 based on the 20 items from the Children’s Depression Inventory (CDI; Kovacs 1992). Adolescents were presented with multiple sets of 3-response choices (increasing in symptom frequency) and asked to select the one that best described their feelings during the previous two weeks. This scale began with the phrase “In the past two weeks, how often have you had these feelings?” Example items were “I am sad,” “I feel like crying,” “I feel like I hate myself,” “I feel like nothing will ever work out for me,” and “I am worthless.” All items were rated on a 3-point scale ranging from 1 (once in a while) to 3 (all the time). All the items were averaged to measure adolescents’ depressive symptoms. The CDI has been used extensively with adolescents, and reliability and validity with populations in middle school and high school have been established (Kovacs 1992). In the present study, this measure demonstrated good internal consistency at each of the three time points (αs = 0.85, 0.84, and 0.86, respectively).

Misconduct

Adolescents’ conduct problems were assessed at ages 12, 14, and 16 through adolescent self-reports on five items based on the work of Elliott et al. (1989). Example items were “In the past year, how often have you: a) been disobedient in school; b) lied to your parents; c) stolen from a store; d) been involved in a physical fight; e) damaged public or private property for fun?” The response format ranged from 1 (never) to 5 (10 or more times). The scores of the six items were averaged to form the construct of misconduct. The measure has good estimated internal consistency reliability and has been used in previous research to measure youth problem behavior (Gutman et al. 2003). In this study, the misconduct measure yielded good internal consistency at each of the three time points (αs = 0.78, 0.80, and 0.79, respectively).

Parental Warmth

Five items from the Family Management Study (Furstenberg et al. 1999) were used to assess mothers’ and fathers’ warmth at adolescent age 12. Items involved the degrees of love, emotional support, affection, and care between parents and adolescents in the past year. This scale began with the phrase “How true are the following statements for you and your child?” with respondents answering on a 5-point scale ranging from 1 (never) to 5 (always true). Example items were “I let my child know I really care about him/her” and “I say something funny to get my child’s mind off his/her troubles.” Responses to these five items were averaged to create the construct of the mother’s and father’s warmth. The constructs of the mother’s and father’s warmth yielded good internal consistency (mother: α = 0.85; father: α = 0.83).

Parental Psychological Risk Factors

We included two psychological risk factors assessed at Wave 1 when adolescents were age 12 as control variables. Eight items from The Parenting Stress Scale (Abidin 1995) were used to measure mothers’ and fathers’ parenting stress. An example item was, “Being a mother/father is harder than I thought it would be.” Responses were averaged to create the construct of parenting stress (mother: α = 0.78; father: α = 0.77). Depression for both mothers and fathers was measured when adolescents were age 12 using an abbreviated form of the Center for Epidemiological Studies-Depression (CES-D; Radloff 1977). The CES-D Short Form contains 12 items with each item coded on a 4-point scale between 0 and 3 (mother: α = 0.80; father: α = 0.79).

Socioeconomic and Demographic Variables

We used socioeconomic and demographic characteristics of participating adolescents and their families at Wave 1 as statistical controls. These covariates included adolescent gender, ethnicity, and family socioeconomic status (SES). The SES index included parental reports of highest level of educational attainment (in years) that both family members achieved and combined family income. We standardized household income and education and combined these scores to form an overall measure of SES.

Results

Means, standard deviations, and bivariate correlations among the key variables are shown in Table 1. Approximately 38 % of mothers and 27 % of fathers reported that they physically punished their children at age 12, 28 % of mothers and 22 % of fathers at age 14, and 12 % of mothers and 10 % of fathers at age 16. Both mothers’ and fathers’ frequency of using physical punishment decreased over time. At each wave, mothers reported greater use of physical punishment than fathers [t(860) = 11.54, p < 0.05 at age 12; t(860) = 7.46, p < 0.05 at age 14; and t(860) = 4.87, p < 0.05 at age 16]. There were low positive correlations between mothers’ and fathers’ use of physical punishment (ranging from 0.16 to 0.21; see Table 1). We found positive associations between mothers’ and fathers’ physical punishment and adolescent misconduct at all ages. Similarly, associations between mothers’ and fathers’ physical punishment and adolescent depressive symptoms were positive at all ages.
Table 1

Means, standard deviations, and correlations among parental physical punishment and adolescent misconduct and depressive symptom variables

 

1

2

3

4

5

6

7

8

9

10

11

12

13

14

1. Maternal PP: age 12

1.00

             

2. Maternal PP: age 14

0.37***

1.00

            

3. Maternal PP: age 16

0.24***

0.23***

1.00

           

4. Paternal PP: age 12

0.17***

0.12***

0.14***

1.00

          

5. Paternal PP: age 14

0.13***

0.21***

0.14***

0.32***

1.00

         

6. Paternal PP: age 16

0.03

0.02

0.16***

0.14***

0.17***

1.00

        

7. Misconduct: age 12

0.25***

0.15**

0.03

0.18***

0.16**

0.06

1.00

       

8. Misconduct: age 14

0.18***

0.23***

0.13**

0.17**

0.24***

0.19***

0.40***

1.00

      

9. Misconduct: age 16

0.05

0.14**

0.17***

0.09*

0.19***

0.25***

0.25***

0.31***

1.00

     

10. Depression: age 12

0.23***

0.12**

0.09**

0.20***

0.09*

0.05

0.18***

0.08*

0.05

1.00

    

11. Depression: age 14

0.18***

0.21***

0.08*

0.16***

0.19***

0.17**

0.18***

0.23***

0.11**

0.32***

1.00

   

12. Depression: age 16

0.03

0.16***

0.18***

0.06

0.15***

0.21***

0.10**

0.08*

0.22***

0.18***

0.34***

1.00

  

13. Maternal warmth: age 12

-0.15**

-0.07*

-0.07*

-0.11**

-0.02

-0.04

-0.15***

-0.09**

-0.03

-0.15***

-0.07*

-0.02

1.00

 

14. Paternal warmth: age 12

-0.10**

-0.04

-0.06*

-0.14**

-0.09*

-0.05

-0.12***

-0.08**

-0.02

-0.09***

-0.08**

-0.08*

0.15***

1.00

Mean

1.63

1.37

1.15

1.16

1.13

1.03

1.62

1.73

1.42

1.30

1.54

2.21

4.00

3.67

SD

1.04

0.76

0.57

0.62

0.54

0.47

0.62

0.71

0.54

0.45

0.47

0.60

0.78

1.25

PP physical punishment

*p < 0.05, **p < 0.01, ***p <0.00

Boys received more physical punishment than girls at age 12 from fathers [girl = 1.10, boy = 1.25, t(860) = 5.24, p < 0.001]. In addition, African American adolescents received more physical punishment than European American adolescents from mothers at ages 12, 14, and 16 [African American = 1.82, European American = 1.31, t(860) = 9.95, p < 0.001 at age 12; African American = 1.52, European American = 1.15, t(860) = 7.56, p < 0.05 at age 14; African American = 1.24, European American = 1.09, t(860) = 4.43, p < 0.001 at age 16] and fathers at ages 12 [African American = 1.22, European American = 1.13, t(860) = 2.57, p < 0.01 at age 12].

We conducted cross-lagged and multi-group analyses in Mplus (Muthén and Muthén 2010). Models were estimated with a procedure (CLUSTER) designed to address violations of independence assumptions related to the multilevel nature of the data (students nested in schools), thereby achieving robust standard errors. We used cross-lagged modeling techniques to test whether there were reciprocal relations between mothers’ and fathers’ use of physical punishment and adolescent adjustment from ages 12 to 16, controlling for stability in parental physical punishment and adolescent adjustment from ages 12 to 16, as well as demographic variables, parenting stress, and parental depression at age 12.

Physical Punishment and Misconduct

The model with adolescent misconduct in relation to parental physical punishment fit the data well, χ2(12) = 10.53, ns, CFI = 0.99, RMSEA = 0.02, SRMR = 0.02 (see Fig. 1). We found stable and significant relationships for mothers’ and fathers’ physical punishment and adolescent misconduct over both concurrent and distant waves. The cross-lagged model showed that physical punishment in one year predicted adolescent misconduct in the next year, such that higher levels of mothers’ and fathers’ physical punishment at ages 12 and 14 predicted greater misconduct at ages 14 and 16, respectively. The reciprocal relation also held: higher levels of misconduct at ages 12 and 14 predicted mothers’ and fathers’ greater use of physical punishment at ages 14 and 16, respectively. Each of the cross-lagged paths was significant and they were nearly equal to each other in magnitude. To determine where these cross-lagged paths between parent and child effects were not significantly different, we compared the model with a second model in which the four cross-lagged paths were constrained to be equal. We found no significant difference, χ2(4) = 6.82, ns, suggesting that the parent (i.e., mother and father) and child effects could be treated as equal. In addition, family SES was negatively associated with maternal and paternal physical punishment at ages 12, 14, and 16. Parenting stress and parent depression were positively associated with maternal and paternal physical punishment at ages 12, 14, and 16.
https://static-content.springer.com/image/art%3A10.1007%2Fs10802-013-9827-8/MediaObjects/10802_2013_9827_Fig1_HTML.gif
Fig. 1

Cross-lagged associations between parental physical punishment and adolescent misconduct across three waves. Standardized path coefficients are presented. The model controlled for adolescent gender, ethnicity, parenting stress, parental depression, parent education, and family income; *p < 0.05, **p < 0.01, ***p < 0.001

To test whether cross-lagged associations varied across ages, we imposed equality constraints on path coefficients across the two lags (i.e., age 12 and 14, and age 14 and 16). The model specifying temporally invariant cross-lagged paths provided a good fit to the data, χ2 (16) = 18.79, ns, CFI = 0.98, RMSEA = 0.02, SRMR = 0.02, and the model fit did not differ significantly from the model without the temporal invariant constraints, ∆χ2 (4) = 7.18, ns, indicating that the reciprocal associations between physical punishment and misconduct did not differ significantly across ages 12 to 16.

To examine whether the reciprocal relations between physical punishment and misconduct were moderated by adolescent ethnicity and parental warmth, we conducted two sets of multi-group analyses separately, which involved a series of comparative models whereby the cross-lagged model was fit separately for each group and fit with different numbers of equality constraints for the cross-lagged paths of the model. Specifically, the reciprocal relations were constrained to be equal for European American and African American adolescents and for high, moderate, and low parental warmth and then compared to unconstrained path models.

To examine the possible moderation effect of ethnicity, we compared the fit of models in which cross-lagged paths were free to differ between African American and European American adolescents with the fit of models in which the cross-lagged paths were held to be equal across African American and European American adolescents. The fully constrained model, in which all of the cross-lagged paths were set to be equal across the ethnicity groups, provided a good fit to the data, χ2 (24) = 31.79, ns, CFI = 0.97, RMSEA = 0.03, SRMR = 0.03. Freeing any or all of the cross-lagged paths between physical punishment and misconduct did not improve model fit over the model in which all parameters were held equal across ethnicity groups, ∆χ2 (8) = 12.17, ns. The finding suggests that the reciprocal relations did not differ for African American and European American adolescents.

To test for the moderation effect of parental warmth, mothers and fathers were divided into three groups. The high warmth group included mothers and fathers who had levels of parental warmth 0.5 standard deviation above the mean. The group means were 4.40 for mothers’ warmth (n = 420) and 4.30 for fathers’ warmth (n = 385). The low warmth group included mothers and fathers who had levels of parental warmth 0.5 standard deviation below the mean. The group means were 3.61 for mothers’ warmth (n = 210) and 3.06 for fathers’ warmth (n = 232). The moderate warmth group included mothers and fathers who had levels of mothers’ and fathers’ warmth within 0.5 standard deviation of the mean. The group means were 4.01 for mothers’ warmth (n = 232) and 3.68 for fathers’ warmth (n = 245).

To determine whether maternal and paternal warmth were significant moderators, we compared the fit of models in which cross-lagged paths were free to differ across parental warmth with the fit of models in which the cross-lagged paths were constrained to be equal across parental warmth. Two separate sets of constraints were imposed. First, we tested the moderation effect of maternal warmth between parental physical punishment and adolescent misconduct, finding the fully constrained model was not different from the fully free model, ∆χ2 (8) = 11.52, ns. This indicates that constraining any or all of the paths to be the same across groups did not reduce the model fit. The fully constrained model, in which all of the cross-lagged paths were set to be equal across the three maternal warmth groups, was accepted as the best fitting model, χ2 (36) = 38.37, CFI = 0.97, RMSEA = 0.04, SRMR = 0.04. We then tested the moderation effect of paternal warmth between parental physical punishment and misconduct. There was no significant difference between the fully constrained model and the fully free model, ∆χ2(8) = 10.75, ns. The fully constrained model was thus accepted as the best fitting model, χ2 (36) = 39.23, CFI = 0.97, RMSEA = 0.04, SRMR = 0.04. In summary, the multi-group findings suggest that physical punishment predicted an increase in adolescent misconduct over time, and adolescent misconduct elicited more parental physical punishment over time equally across the three levels of paternal and maternal warmth.

Physical Punishment and Depressive Symptom

We ran the same cross-lagged analysis for the model with adolescent depressive symptoms in relations to parental physical punishment, which resulted in a similarly good fit for the data, χ2(12) = 16.38, ns, CFI = 0.97, RMSEA = 0.03, SRMR = 0.03 (see Fig. 2). Indeed, several similarities to the results for misconduct were found. For example, as was the case with misconduct, stable and significant relationships between mothers’ and fathers’ physical punishment and adolescent depressive symptoms over both concurrent and distant waves were found. The cross-lagged model showed that physical punishment in one year predicted adolescent depressive symptoms in the next year, such that high levels of parental physical punishment at ages 12 and 14 predicted greater depressive symptoms at ages 14 and 16, respectively. However, higher levels of depressive symptoms at ages 12 and 14 did not significantly predict greater use of mothers’ and fathers’ physical punishment at ages 14 and 16, respectively. Therefore, we found more evidence of parent effects than child effects between parental physical punishment and adolescent depressive symptoms.
https://static-content.springer.com/image/art%3A10.1007%2Fs10802-013-9827-8/MediaObjects/10802_2013_9827_Fig2_HTML.gif
Fig. 2

Cross-lagged associations between parental physical punishment and adolescent depressive symptoms across three waves. Standardized path coefficients are presented. The model controlled for adolescent gender, ethnicity, parenting stress, parental depression, parent education, and family income; *p < 0.05, **p < 0.01, ***p < 0.001

To test whether cross-lagged associations varied across ages, we again imposed equality constraints on path coefficients across the two lags (i.e., age 12 and 14, and age 14 and 16). The model specifying temporally invariant cross-lagged paths provided a good fit to the data, χ2 (16) = 24.02, ns, CFI = 0.97, RMSEA = 0.03, SRMR = 0.03, and the model fit did not differ significantly from the model without the temporal invariant constraints, ∆χ2 (4) = 7.47, ns, indicating that, similar to the relationship between physical punishment and misconduct, the longitudinal associations between physical punishment and depressive symptoms did not differ significantly across ages 12 to 16.

To test the moderation effect of adolescent ethnicity on the association between physical punishment and depressive symptoms, we compared the fit of models in which cross-lagged paths were free to differ across ethnicity to the fit of models in which the cross-lagged paths were constrained to be equal across adolescent ethnicity. We did not find any of these model comparisons to be significant, indicating that freeing any or all of the cross-lagged paths between physical punishment and depressive symptoms did not improve model fit over the model in which all parameters were held equal across ethnic groups, ∆χ2 (8) = 11.28, ns. The fully constrained model, in which all of the cross-lagged paths were set to be equal across the ethnic groups, was accepted as the best fitting model, χ2 (24) = 34.45, ns, CFI = 0.97, RMSEA = 0.03, SRMR = 0.02. The finding suggests that the association between physical punishment and depressive symptoms once again did not differ for African American and European American adolescents.

Finally, we tested the moderation effect of maternal warmth between parental physical punishment and adolescent depressive symptoms, finding the fully constrained model was not different from the fully free model, ∆χ2 (8) = 12.04, ns. The fully constrained model, in which all of the cross-lagged paths were set to be equal across the three maternal warmth groups, was accepted as the best fitting model, χ2 (36) = 40.53, ns, CFI = 0.97, RMSEA = 0.03, SRMR = 0.03. We then tested the moderation effect of paternal warmth between parental physical punishment and depressive symptoms. There was no significant difference between the fully constrained model and the fully free model, ∆χ2(8) = 10.12, ns. The fully constrained model was thus accepted as the best fitting model, χ2 (36) = 42.75, CFI = 0.97, RMSEA = 0.03, SRMR = 0.03. Thus the multi-group findings suggest that physical punishment predicted an increase in adolescent depressive symptoms over time equally across all three levels of maternal and paternal warmth.

Additional Analyses for Gender as A Moderator

Based on the observed gender differences in receipt of physical punishment, we tested whether gender moderated the reciprocal association between physical punishment and adolescent adjustment as a post-hoc analysis. Specifically, we examined adolescent gender differences by comparing the fit of models in which paths were free to differ across adolescent gender with the fit of models in which the structural paths were constrained to be equal across gender. The fit of the constrained models was not significantly worse than the unconstrained models, χ2(8) = 5.45, ns, for misconduct and χ2(8) = 7.17, ns, for depressive symptoms. The findings suggest that the relations between physical punishment and adolescent adjustment did not differ for boys and girls.

Discussion

The present study used cross-lagged modeling techniques to examine reciprocal relations between maternal and paternal physical punishment and adolescents’ misconduct and depressive symptoms between ages 12 and 16, accounting for stability in both parental physical punishment and adolescent problem behaviors over time. We found that mothers’ and fathers’ physical punishment in a given year predicted an increase in adolescents’ misconduct and depressive symptoms in the next year. Adolescent misconduct, but not depressive symptoms, in a given year also predicted increases in mothers’ and fathers’ physical punishment in the next year. The pattern of findings in the relationships between physical punishment and adolescent misconduct and depressive symptoms were similar across mothers and fathers. Parental warmth and adolescent ethnicity did not moderate the longitudinal, reciprocal relations between parental physical punishment and adolescents’ adjustment.

In our sample of 862 American two-parent families, physical punishment appears to be relatively common during adolescence (ranging from 9 % to 38 % between ages 12 and 16) though frequency of use by both mothers and fathers decreased over time. Consistent with previous research (Straus and Stewart 1999), mothers reported greater use of physical punishment than fathers. There was a low correlation between both mothers’ and fathers’ use of physical punishment. This suggests that these parents did not necessarily adopt one another’s discipline techniques, as has been suggested by other parenting researchers (Fletcher et al. 1999). Results further showed that mothers used similar levels of physical punishment with girls and boys, but that fathers used greater levels of physical punishment with boys than with girls. It is possible that fathers feel greater responsibility for the socialization of their male children (Huston 1983). Fathers may also possess a more stereotypical view of boys than do mothers, perhaps believing that greater levels of physical punishment are required in order to address boys’ misbehavior (Chang et al. 2003; McKee et al. 2007).

Our findings on ethnic differences in the levels of use of physical punishment are consistent with previous studies (Lansford et al. 2004; McLoyd and Smith 2002) and indicate that African American adolescents receive more physical punishment than European American adolescents. Research suggests that parental use of physical punishment is related to family stress and to worries that one’s children are growing up in risky environments (Pinderhughes et al. 2000). African American parents are likely to be more affected by such factors than European American parents, potentially contributing to the higher levels of physical discipline within that community (Lansford et al. 2004). Moreover, African American families often value interdependence, communal effort for goal realization, and perseverance in the face of adversity (Hill 2001). These family values can lead to ‘no-nonsense’ parenting practices that are sometimes characterized as harsh and authoritarian (Garcia Coll et al. 1994).

Parental Physical Punishment and Adolescent Adjustment

In this socioeconomically and ethnically diverse sample, we found that parental physical punishment is associated with increased problematic outcomes in youth. It appears that the use of physical punishment does not function as parents generally intend, which is to teach children standards of acceptable behavior and the repercussions of not adhering to these standards. Rather, our findings suggest that physical punishment increases both the misconduct and depressive symptoms of adolescents over time and beyond initial levels. During adolescence, young people are charged with the task of developing a sense of identity and autonomy (Bender et al. 2007; Gunnoe and Mariner 1997). Physical punishment may be interpreted as being indicative of hostility and rejection. Adolescence is also a time when young people are acquiring the skills needed to reason with their caregivers (Collins et al. 2002) which should lead to a more egalitarian relationship. Physical punishment may be experienced by the adolescent as unjust and degrading, leading them to feel angry and rejected—feelings that may lead to problem behaviors (Evans et al. 2012).

We found that there was a reciprocal relationship between parental physical punishment and adolescents’ misconduct, and this transactional process was consistent over time. Both parent and child effects had similar effect sizes. This finding provides support for Sameroff’s (1975) transaction model and Patterson’s social coercion model (Patterson et al. 1992) which point to the importance of considering child effects when examining parental use of discipline. It is likely that an escalating cycle of negative behavior on behalf of both the child and parents occurs. For example, the parent is likely to react to their child’s misbehavior with increasing physicality, which may in turn promote even more misbehavior.

It is noteworthy that we examined adolescents aged 12 – 16 years old in this study. Most adolescents had just experienced school transition (from elementary to middle school and middle to high school), alongside the transition into adolescence itself. These transition points are particularly stressful for adolescents as they experience discontinuity simultaneously in their academic lives and their social relationships (Gross et al. 2008; Wang and Eccles 2012). These changes may strain the parent–child relationship and create more parent–child conflict (Steinberg and Steinberg 1994; Wang and Kenny 2013). How parents and adolescents manage these changes and maintain the quality of the parent–child relationship has an impact on adolescents’ behavioral and emotional development (Steinberg and Steinberg 1994; Wang et al. 2011; Wang and Kenny 2013). If the parent–child dyad is unable to adapt to the potentially difficult transitions and challenges that are part and parcel of adolescence, it is likely that a strain will be put upon the parent–child relationship. Parents may resort to physical punishment in the face of their adolescent child’s misconduct, for example. Not only is physical punishment likely to be particularly detrimental during this developmental period when the adolescent has a growing capacity to reason (Collins et al. 2002) and a need for greater levels of autonomy, but it is also likely to cause higher levels of misconduct.

On the contrary, we did not find a reciprocal relationship between parental physical punishment and adolescent depressive symptoms. As hypothesized, in this relationship we found evidence of parent effects only, and these were consistent over time for both mothers and fathers. It appears that depressive symptoms in adolescents do not elicit a parental inclination to resort to physical punishment, unlike externalizing behaviors. This is likely explained by the tendency of internalizing behaviors to lead to social withdrawal, perhaps provoking concern or frustration, rather than anger that might provoke physical discipline.

Moderation Effect

Ethnicity did not moderate the relationship between physical punishment and adolescent outcomes nor did it buffer against the detrimental effects of physical punishment. Specifically, we found a bidirectional relationship (both parent and child effects) between physical punishment and misconduct and a unidirectional relationship (parent effects only) between physical punishment and depressive symptoms, operating similarly in both African American and European American families. This finding implies that African American adolescents are equally likely to experience harmful outcomes as a result of physical punishment as European American adolescents. Previous research has tended to focus on the association between physical punishment and adjustment outcomes in young children. It is possible that cultural normativity may buffer children against the effects of physical punishment (Simons et al. 2002), but that adolescents of all ethnicities may see it as indicative of rejection and hostility. Given that greater levels of physical punishment are reported in African American families (Lansford et al. 2004; McLoyd and Smith 2002), it may be important to design culturally sensitive and informative interventions that dispel the notion that physical punishment is a deterrent of antisocial behavior and other problematic outcomes in youth.

Both positive and negative parenting characteristics can operate simultaneously, and some prior research concludes that a positive parenting style may consist of both strict discipline and warmth (Darling and Steinberg 1993; Stice and Barrera 1995). However, the results of our study found that parental warmth did not moderate the negative effects of physical punishment on adolescent misconduct and depressive symptoms. Regardless of whether mothers or fathers demonstrated low, medium, or high levels of warmth, physical punishment was ineffective at reducing misconduct, and in fact led to increased misconduct and depressive symptoms in our sample of 12-16 year olds. While parental warmth creates trust and reciprocity between parent and child (Amato 1990), physical punishment may compromise those bonds and thus contribute to coercive processes that reinforce the adolescent’s problem behaviors. Moreover, adolescents exposed to physical discipline may suffer from low self-esteem (Donovan and Brassard 2011) and it may be particularly difficult to moderate the effects that parental corporal punishment has on the adolescent’s developing sense of self.

This finding conflicts with some previous research which has reported that a high level of parental warmth is protective and buffers against the negative impacts of physical discipline (Deater-Deckard et al. 2006; McLoyd and Smith 2002). However, many of these previous studies employed cross-sectional designs and measured a composite of physical punishment which may have confounded physical punishment with more severe or even potentially abusive parenting (Aucoin et al. 2006). Moreover, their samples were limited in generalizability, with little socioeconomic or ethnic diversity (e.g., McKee et al. 2007). Our use of a longitudinal design and an ethnically and socioeconomically diverse community sample extends our understanding of the effect of parental warmth in the interaction between physical punishment and adolescent outcomes. Finally, these inconsistent findings may result from the lack of a consistent definition or measure for parental warmth.

Limitations and Conclusions

Some limitations of our study should be noted. First, our use of self-reported physical punishment from mothers and fathers may introduce reporter bias. It is possible that parents underreported their use of physical punishment as a function of social desirability bias. However, since our findings demonstrate that physical punishment predicts increased problem behaviors, this pattern would only strengthen if parents reported more use of physical punishment. Future research should examine both parental and child reports of physical punishment and compare their differences in relation to child outcomes. Second, our analyses began at age 12, and as a result, we do not capture the beginning of the developmental trajectory of problem behaviors nor the origins of the parent and child influence. Future research will benefit from examining the transactional interaction between parental physical punishment and child outcomes at all ages of child and youth development. Third, we measured the frequency of incidences of specific acts of physical punishment, such as slapping, grabbing, and pushing, to the exclusion of more severe acts of physical punishment. Although this means that we have avoided confounding severity of physical punishment, future research will benefit from separating and comparing the deleterious effects of physical punishment as a function of both severity and frequency (Kazdin and Benjet 2003). Finally, the adolescents in this study were from two-parent families. The findings may not generalize to other family structures, such as divorced or single parents. Additionally, same-sex and mixed family arrangements (e.g., mother and grandmother share co-parenting responsibilities) were also not explored in the present study. Given the growing complexity of family formation in American society, researchers should expand their focus to examine how differently composed parenting dyads are associated with adolescent outcomes.

Our findings add fuel to the ongoing controversy that surrounds the use of physical punishment by parents. A great deal of research indicates that physical punishment is associated with negative outcomes in youth (e.g., Bender et al. 2007; Gershoff 2002). As a result, 33 countries have banned physical punishment as a method of child discipline (Global Initiative to End Corporal Punishment of Children 2012). Even so, parental physical punishment appears to be relatively common amongst the American public, and its use remains high even in adolescence (Child Trends 2012). Using a large, diverse sample of adolescents and their mothers and fathers, the findings of the current study lend support to the growing evidence against the use of physical punishment as a disciplinary strategy and suggest that even the relatively less severe types of physical punishment should be eliminated. In addition, we expand the literature by showing that parental warmth cannot moderate the negative effects of physical punishment on adolescent misconduct and depressive symptoms. One important implication from this finding is that it is not sufficient for interventions to focus on increasing positive parenting in the home; rather, they need to target physical punishment specifically, and to focus on its elimination as a disciplinary tactic. Furthermore, the promotion of the appropriate use of non-physical punishment alongside communication and warmth will help parents to manage difficult adolescent behavior. Physical punishment should not be resorted to at any age. However, given that adolescents may be particularly sensitive to the effects of physical punishment, it is highly important that parents are made aware that this form of punishment is likely to elicit increasing levels of problem behavior in their developing adolescent child—exacerbating the behavior that a parent initially sought to eliminate.

Adolescence can be a challenging time for parents, but it is important for parents to bear in mind that it is also often a challenging time for adolescents themselves. School, peer, and biological transitions have the potential to cause stress and anxiety for the growing adolescent (Wang, Brinkworth, & Eccles 2013). Parents who can empathize, rationalize, and reason with their developing adolescent children are likely to set the tone for a healthier and happier parent–child relationship. Alongside this, other forms of disciplinary practices, such as grounding or allocating chores may also help parents to set boundaries and discourage problematic behaviors. Parents who use physical force, on the other hand, disregard adolescents’ need for autonomy, their growing capacity to discuss and understand their parents’ and their own fears and worries, and their need to grow up in a safe environment.

Copyright information

© Springer Science+Business Media New York 2014