Influence of Physical Punishment on Adolescents' Self-esteem

Published Date: 23 Mar 2015 Last Modified: 15 Jan 2018

Introduction

Disciplinary incidents are central to moral development because disciplinary practices assist to inculcate "moral standards and values that provide the basis for self-controlled behaviour" within the child (Brody & Shaffer, 1982, p.32). Amongst the various disciplinary methods, physical punishment is widely practised across different cultures and countries. The present study focused on non-abusive physical punishment and adopted the definition by Straus (1994) that physical punishment “is the use of physical force with the intention of causing a child to experience pain, but not injury, for the purpose of correction or control of the child's behaviour” (p.4). This definition was used to delineate non-abusive physical punishment from harsher forms of abusive punishment. The term “corporal punishment” is synonymous and has been used interchangeably with physical punishment. We used the term “physical punishment” in this study because it specifically indicates that punishment is meted out in a physical and bodily manner.

A survey conducted in Jamaica revealed that physical punishment is frequently practiced in home and school (Smith & Mosby, 2003). Physical punishment is also common in south-west Ethiopia (Admassu, Belachew, & Haileamalak, 2006). This disciplinary method, however, is not peculiar to developing countries. Even in socially privileged countries, physical punishment is also used as a disciplinary method. Approximately 60% of Hong Kong Chinese parents admitted to using physical punishment as a form of discipline (Tang, 2006). In America, 94% of 3- and 4-year olds have been physically punished by their parents at least once during the past year (Straus & Stewart, 1999), and 85% of Americans believed that “a good hard spanking is sometimes necessary” (Bauman & Friedman, 1998). Beliefs in its positive disciplinary effects contributed to the widespread use of physical punishment (Straus, 1994) and there are evidence-based studies supporting the idea that physical punishment suppresses undesired behaviour (Gershoff, 2002; Larzelere, 2000; Paolucci & Violato, 2004). For example, studies in Larzelere's (2000) meta-analysis provided evidence that non-abusive spanking used by loving parents reduced subsequent noncompliance and fighting in 2- to 6-year olds. In relation to Larzelere's (2000) findings, Gershoff (2002) found a large mean effect size for immediate compliance following corporal punishment. However, as noted by Gershoff (2002), these beneficial outcomes are only temporarily because physical punishment neither teaches children the reasons for behaving correctly, nor does it communicate what effects their behaviours have on others. Hence, physical punishment may not facilitate moral internalisation of the intended disciplinary message (Gershoff, 2002). Moreover, the demerits may outweigh the merits of punishment because studies suggested that physical punishment carry with it unintended and adverse effects (Holden, 2002; Rohner, Kean, & Cournoyer, 1991; Straus, 1994). In response to the increasingly condemnatory international views about physical punishment, 25 states, to date, abolished all forms of physical punishment on children (Global Initiative to End All Corporal Punishment of Children, 2009).

Burgeoning research has related physical punishment to a variety of negative effects. These effects ranged from social-emotional and psychological problems, such as mental distress and withdrawal (Eamon, 2001), to behavioural problems, such as antisocial behaviour and increased aggression (Deater-Deckard, Dodge, Blates, & Pettit, 1996; Sim & Ong, 2005; Straus, Sugarman, & Giles-Sims, 1997; Tang, 2006). Straus et al. (1997) suggested that a “dose response” to physical punishment for children may exist, such that more frequent and longer usage of punishment will lead to increased probability of behaviour problems. These potentially adverse effects of physical punishment may also carry over into adulthood in the form of increased psychopathology and violent behaviour (Eron, 1996); substance abuse, depression, family violence, and suicide (Afifi, Brownridge, Cox, & Sareen, 2006; Straus, 1995; Straus & Kantor, 1994). Eron (1996) indicated that the more harshly 8- and 9-year olds were punished, the more aggressive and antisocial they were in late adolescence and young adulthood. Afifi and colleagues (2006) also found individuals who were physically punished, as compared to those who were not, had higher risk for major depression, alcohol abuse or dependence and externalising problems in adulthood, and these effects were not attenuated after controlling for sociodemographic variables and parental bonding. In addition, Straus (1995) found significant positive correlation between the level of punishment experienced as a child and level of depressive symptoms and thoughts of committing suicide in adulthood, after controlling for socioeconomic status (SES), martial violence, and witnessing violence as a child.

In the past decade, at least three meta-analyses were conducted to review research on the effects of physical punishment. Larzelere (2000) reviewed a total of 38 studies and found both beneficial (as discussed above) and negative effects of physical punishment. From 17 causally relevant studies, the author highlighted apparent detrimental effects of physical punishment. He first pointed out that physical punishment predicted increased subsequent negative externalising behaviour, supporting the “violence begets violence” viewpoint. One of the studies reviewed was the controlled longitudinal studies of the National Longitudinal Study of Youth (Larzelere & Smith, as cited in Larzelere, 2000), which provided not only consistent but also causally relevant evidence that physical punishment is associated to subsequent increase in antisocial behaviour. Secondly, Larzelere (2000) indicated that physical punishment predicted increased mental health problems. The links between physical punishment and both increased negative externalising behaviour and mental health problems were further supported by another meta-analysis, which evaluated 88 published works spanning a 62 year period. Gershoff (2002) found that physical punishment was indeed significantly associated to increased aggression, increased delinquency, increased antisocial behaviour, and decreased mental health, to name a few.

All 20 studies involving mental health in Gershoff's (2000) meta-analysis found frequency of corporal punishment to be positively and significantly related to a decrease in children's mental health. Straus and Kantor (1994) reported that after controlling for low SES, those who experienced corporal punishment in adolescence were still at higher risk for depression, suicidal thoughts and alcohol abuse. Although Paolucci and Violato (2004), who conducted another meta-analyses and evaluated 70 studies between 1961 and 2000, did not find exposure to corporal punishment led to increased risk of developing cognitive problems (e.g., suicidal thoughts and attitudes toward violence), they found that people who experienced corporal punishment were at a small but increased risk for developing behavioural (e.g., aggression and antisocial behaviours) and affective problems (e.g., psychological impairment and low self-esteem).

Physical punishment appears to have a dampening effect on self-esteem in its victims. However, studies have been inconclusive and evidence-based literature in this area is much thinner, as compared to the large number of published articles on physical punishment and increased externalising behaviours, such as children's aggression, which is one of the most studied in the literature on parenting (Paolucci & Violato, 2004). Only 3 studies (Adams, 1995; Larzelere, Kein, Schumm, & Alibrano, 1989; Sears, 1970) cited in Larzelere's (2000) meta-analysis, investigated the association between physical punishment and self-esteem. Specifically, Larzelere et al. (1989) found the amount of spanking received negatively predicted self-esteem but the negative correlations between punishment, self-esteem, and perception of fairness of punishment were reduced to non-significance after controlling for parental positive communication. The other study included in the meta-analysis did not find a significant correlation between physical punishment and subsequent self-esteem (Sears, 1970). Joubert's (1991) study, which was not included in the meta-analysis, also found no evidence indicating spanking to have any effect on children's self-esteem scores, regardless whether spanking was administered by mothers or fathers, or both.

On the other hand, one of the three studies as cited in Larzelere's (2000) meta-analysis found lower self-esteem among 6- to 12-year olds, especially those who were hit with high frequency (twice a week), even after controlling for ethnicity, cognitive enrichment and poverty (Adams, 1995). Furthermore, recent studies, which were not included in the meta-analysis, also found similar results. Using data from 1,397 children, Eamon (2001) found 4- to 9-year-old children who received more frequent spanking exhibited more socio-emotional problems like low self-esteem. In another study, Amato and Fowler (2002) investigated the relationship between parental use of corporal punishment and children's self-esteem, using data collected from 3,400 households with a child within the age range of 5-18. Similarly, parent's use of corporal punishment was found to predict lower self-esteem.

Bauman and Friedman (1998) argued that physical punishment retards the development of self-esteem, and Paolucci and Violato (2004) used findings of corporal punishment being associated with psychosocial problems, such as depression, as supporting evidence that physical punishment is related to impaired self-esteem. Coercive disciplinary techniques are also linked to decreases in children's level of confidence and assertiveness, and increases in feelings of humiliation and helplessness (Gershoff, 2002). One explanation for these findings is the fear of punishment makes people attempt to escape. However, when escaping from punishment is not possible, feelings of learned helplessness and depression may develop (Paolucci & Violato, 2004).

Self-esteem, as defined by Rosenberg (1965), is a positive or negative attitude towards the self. Interestingly, self-esteem stability in childhood and adolescence does not differ between genders (Trzesniewski, Donnellan, & Robins, 2003). Research has shown that self-esteem has a pervasive impact on an individual's life in numerous areas. Emotionally, individuals with low self-esteem have a tendency to exhibit higher levels of anxiety, experience more frequent psychosomatic symptoms, feelings of depression, lack of personal acceptance and submissiveness (Battle, 1992). Low self-esteem is also an important predictor for disruptive and maladaptive behaviours (Aunola, Stattin, & Nurmi, 2000; Vandergriff & Rust, 1989). An individual's social functioning can also be affected by his level of self-esteem. Children with high self-esteem are more popular among peers and participate more actively in social groups, unlike those with low self-esteem, who experience more difficulties forming friendships (Battle, 1992; Growe, 1980). Self-esteem has also been supported by research evidence, to be positively related to academic self-efficacy. This significant relationship is found in Western countries (Jonson-Reid, Davis, Saunders, Williams, & Williams, 2005; Smith, Walker, Fields, Brookins, & Seary, 1999), as well as in Singapore (Ang, Neubronner, Oh, & Leong, 2006).

An important trend in the international research focuses on the effects of physical punishment on children (Ripoll- Núñez & Rohner, 2006). Considering that self-esteem has a wide range of influence on an individual's life and the current literature remains inconclusive on the effects physical punishment have on self-esteem, we chose to investigate the relationship between non-abusive physical punishment administered by adolescents' main disciplinarian and adolescents' level of self-esteem. Self-esteem plays a vital role in an individual's development, and if physical punishment has negative effects on adolescent's self-esteem, it is likely that his level of self-esteem will affect his psychosocial and educational development, and his overall well-being. For instance, his academic success and ability to socialise contribute to his current and future well-being.

The relationship between physical punishment and adolescents' outcome cannot be simply described as two distinct categories, such that physically punished adolescents will experience negative outcomes, and adolescents who have never been physically punished will not. Instead, this relationship may lie on a continuum and the frequency of physical punishment may play an important role in the punishment-outcome link, such that increase in frequency of punishment will lead to increased probability of negative outcomes. Since a "dose response" towards physical punishment was suggested by Straus et al. (1997), and a positive relationship between the frequency of physical punishment and negative outcomes was concurred by Larzelere (2000) and Gershoff (2002), and more specifically, Adams (1995) and Eamon (2001) found lower self-esteem, especially among those who experienced frequent use of physical punishment, we chose to investigate the punishment-self-esteem link by focusing on the frequency of punishment. With increased frequency of physical punishment, lower self-esteem can be expected.

The weight of the existing research seems to favour the viewpoint that non-abusive physical punishment carries with it negative baggage. However, unlike physical abuse, the conclusion that non-abusive physical punishment indeed has detrimental consequences on adolescents' well-being cannot be substantiated. Researchers at the opposite end of the debate cite conflicting evidence and physical punishment remains the most controversial topic in the domain of parental discipline (Holden, 2002; Larzelere, 1996).

The main debate remains on whether non-abusive physical punishment is completely harmful or it has negative effects only when used within certain conditions. As summarised by Ripoll- Núñez and Rohner (2006), the “conditional defenders” of corporal punishment argued that the effects of punishment may be positive, negative, or both depending on the conditions in which it was administered. As proposed in Gershoff's process-context model (2002), the link between physical punishment and its impact on the child is not direct and isolated. Instead, contextual factors of varying levels of influence may moderate the processes linking punishment and child constructs (Gershoff, 2002). This is supported by the fact that majority of the 94% of 3- and 4-year-old Americans who experienced physical punishment did not experience negative outcomes, such as developing into clinically aggressive adults or criminals. Critics of past research argued that many studies which linked physical punishment to negative effects have methodological flaws because they did not take into account the influence of moderating variables, which when included, tended to attenuate the relationship between punishment and negative outcomes (Rohner, Bourque, & Elordi, 1996). Since not all individuals who experienced non-abusive physical punishment experienced negative outcomes, the present study further examined two potential moderators of the punishment-self-esteem link: namely, adolescents' perceptions on the fairness of physical punishment and caregiver acceptance-rejection.

Typically, research in this area has relied on parental reports of physical punishment. However, parents may underreport the use of physical punishment due to social desirability. Parents may feel threaten to disclose the frequency with which they physically punish their children because it is not advocated in contemporary society, hence providing inaccurate data (Shum-Cheung, Hawkins, & Lim, 2006). Moreover, if parent is the source of data on both the punishment and children's behaviours, they may attempt to justify their punishment through the parental report of child behaviour (Bauman & Friedman, 1998). Following, we collected retrospective account of physical punishment from the recipients of the disciplinary practice, and further explored the possible moderating effect their cognitive perceptions on the punishment, may exert on the punishment-self-esteem link.

The impact of punishment on adolescents is not unidirectional because adolescents are not simply passive recipients of the punishment. Instead how adolescents perceive the punishment may affect the impact it has on their outcomes. As noted by Holden (2002), noticeably absent from research is studies of children's perceptions and reactions to punishment. It has been suggested that effects of physical punishment may be moderated by the meaning children ascribes to the punishment (Benject & Kazdin, 2003). Ignorance of this may lead to an inaccurate picture on the effects of punishment because the key to understanding how physical punishment affects its victims lies in understanding how they react to the punishment physiologically, affectively, and cognitively (Gershoff, 2002). Holden (2002) further posited that this reaction involves at least two processes, namely, immediate physiological and sensory reaction, followed by the secondary cognitive appraisal stage. In line with Ripoll- Núñez and Rohner's (2006) suggestions on variables that are important in the research of physical punishment and its effects on children, we explored the potential moderating effect of adolescents' perceptions of fairness of physical punishment, which has been considered to ameliorate the negative outcomes of punishment (Rohner et al., 1991; Rohner et al., 1996). Grusec and Goodnow (1994) suggested that children, who perceive punishment as fair, will be more willing to accept the intended disciplinary message, which then facilitates internalisation. Since adolescents are the recipients of parental disciplinary practices, the knowledge of their perceptions on the fairness of punishment will open the window to their internal mental processes, which is how they interpret and internalise the punishment. This provides a more complete understanding of the relationship between punishment and self-esteem. Concerns regarding whether adolescents are mature enough to make sensible judgments about the fairness of discipline can be allayed because Konstantareas and Desbois (2001) found 4-year-old preschoolers capable of making judgments about the fairness of discipline by mothers, and in a study conducted in Singapore, parents' and 10- to 12-year-old children's responses on fairness of discipline were similar (Shum-Cheung et al., 2006). Therefore, if adolescents perceive physical punishment as fair, the effects of punishment on their self-esteem may not be deleterious. Following, the negative association between physical punishment and self-esteem can be expected to be stronger at lower levels, as compared with higher levels of perceived fairness.

Little is also known about the conditions under which punishment occurs (Bauman & Friedman, 1998) and if information regarding the context in which the punishment is meted out is not captured, only a snapshot of the impact of punishment on adolescents will be known. Opponents of physical punishment have acknowledged that physical punishment by itself is unlikely to produce negative child outcomes. However, when combined with other risk factors in the family, negative effects of physical punishment may surface (Bauman & Friedman, 1998). Therefore, certain factors in the adolescent's family may influence the cognitive appraisal process of the punishment and, consequently, buffer the negative effects.

Corporal punishment is considered to be beneficial when administered by emotionally supportive parents who share positive interactions with their children (Paolucci & Violato, 2004). As discussed above, Larzelere et al. (1989) reported that positive parental communication moderated the punishment-self-esteem link. Therefore, information regarding other aspects of parenting, such as the warmth dimension, will provide a much fuller understanding towards the relationship between physical punishment and self-esteem.

As construed in the parental acceptance-rejection theory (PARTheory), parental acceptance and rejection form the warmth dimension of parenting (Rohner, 1991). Perceived parental acceptance-rejection may be one of the most important parenting dimensions to consider because no cultural or ethnic group was found where perceived parental acceptance-rejection failed to correlate with the predicted personality dispositions (Rohner & Britner, 2002). PARTheory predicted rejected children, as compared to children who perceived themselves as being accepted, are more likely to have an impaired sense of self-esteem, amidst other negative effects (Rohner, 1991; Rohner & Britner, 2002). Rohner (1991) used Mead's (1934) "significant other" concept to explain how parental rejection may affect self-esteem. PARTheory assumed that everyone tends to view ourselves as we imagine “significant others” view us. Therefore, if parents who are children's most significant other reject them, they are more likely to define themselves as unworthy, and consequently develop an overall sense of negative self-evaluation, including feelings of negative self-esteem and self-adequacy (Rohner, 1991). Although the term “parent” is used in PARTheory, Rohner (1991) explained it refers to the major caregiver of the child, not necessarily the parents. Therefore, we used the term “caregiver” instead of “parent” in this study.

Variations in perceived caregiver acceptance-rejection among adolescents may magnify or minimise the effects of physical punishment and this has been supported by cross-cultural evidence. Rohner et al. (1991), for example, found severe physical punishment to be related to psychological maladjustment among Kittitian youths and the effects became more substantial when it was paired with caregiver rejection. Similarly, results from another study conducted in Georgia showed that the association between perceived harshness of punishment and psychological maladjustment disappeared once perceptions of caregiver acceptance-rejection were accounted for (Rohner et al., 1996). In the context of Singapore, perceived parental acceptance-rejection was also found to play an important moderating role. Sim and Ong (2005) found perceived father's rejection moderated the link between slapping and daughter's level of aggression, and perceived mother's rejection moderated the canning-aggression link among Singapore Chinese preschoolers of both genders. All these studies uniformly showed that children's perception of caregiver acceptance-rejection has a significant impact on the association between physical punishment and its outcomes. Thus, at higher compared to lower levels of perceived caregiver rejection, a stronger negative association between physical punishment and self-esteem can be expected.

We collected data on adolescents' perceptions of caregiver acceptance and rejection, and frequency of physical punishment by their main disciplinarian, rather than their main caregiver. This is because our study used a Singapore Chinese sample, and it is common within this group that the main disciplinarian may not be the main caregiver. In Chinese societies, traditional roles of disciplinarian and caregiver are respectively played by fathers and mothers, and this role differentiation still applies in Singapore (Quah, 1999). In cases where the disciplinarian and caregiver are different persons, the adolescent may experience more punishment from the disciplinarian as compared to the caregiver, and the impact of punishment from the main disciplinarian will not be reflected if punishment administered by the caregiver was measured.

Since the main disciplinarian is the adult who administers punishment, effects of punishment may be moderated by the adolescent's perceived acceptance from his main caregiver, who plays the key caring role and spends the most time with him. Collecting data on adolescents' perceptions of caregiver acceptance-rejection allowed us to examine the punishment-self-esteem link through the relationship between caregiver and adolescent.

As pointed out by Larzelere (2000), one of the needs in the research on physical punishment is for studies to take a developmental perspective because reviews by Larzelere (2000) and Gershoff (2002) found outcomes of punishment varied by the child's age. For example, Gershoff (2002) found that with increased age, the association between corporal punishment and aggressive and antisocial behaviours became stronger. Following, we used a retrospective design to investigate the association between physical punishment and self-esteem, and the impact the two proposed moderators may have on this link, at two age frames, namely when the individual was 11- to 12-years old (early adolescence) and 15- to 16-years old (middle adolescence). Although physical punishment is at its zenith when children are aged 3-5 (Straus & Stewart, 1999), and its frequency decreases as children grow older, physical punishment is still prevalent during adolescence (Straus et al., 1997). Straus (1994) found more than 60% of parents in America reported hitting 10- to 12-year olds, and even at ages 15-17, one out of four adolescents is still physically punished.

The two age frames were chosen partly because this study was retrospective in nature, and memories of punishment incidents during early childhood may be weak due to the long time passage that passed. Additionally, an average Singapore student aged 11- to 12-years old and 15- to 16-years old, is in preparation for the national examinations, namely, the Primary School Leaving Examination and GCE ‘O' Levels, respectively. Being the periods of their major examinations, memories during these periods may be much clearer and distinct, and this will provide the study with more accurate data.

Children below age 8 have not developed the concept of global self-esteem, thus another reason for focusing at these two age frames is that at ages 11-12 and 15-16, adolescents will have developed the ability to “view themselves in terms of stable dispositions, which permits them to combine their separate self-evaluations into an overall sense of self-esteem” (Berk, 2006, p.449). Moreover, unlike in early childhood, individual differences in self-esteem from early to middle adolescence become increasingly stable (Trzesniweski et al., 2003), which allowed us to explore the punishment-self-esteem link more precisely.

According to Erikson's stages of psychosocial development (as cited in Berk, 2006), he organised life into eight stages that extend from birth to death, of which two stages were related to the present study. During the latency stage, where 11- and 12-year olds will be categorized, they enter school and are required to develop a sense of competence through the social interactions in school. With a wider range of socialisation opportunities, their relationships with parents may no longer be the most significant but it remains influential because little or no encouragement from parents, teachers, or peers may lead them to doubt their ability to succeed (Berk, 2006). However, 15- and 16-year olds are in the adolescence stage, where the need to develop an independent identity that is separated from the family, becomes the key developmental task, and relationships with peer groups become the most significant relationship. Hence the attenuation of familial influence for adolescents aged 15-16 may decrease as compared to when they were 11- to 12-years old.

Moreover, 15- and 16-year olds fall in Piaget's formal operational stage, which represents the apex of cognitive development (Siegler & Richards, 1982). Unlike the subsequent stage, 11- and 12-year olds are in the concrete operational stage and can only “operate on reality”. But formal operational adolescents developed the ability for abstract thinking and can engage in hypothetico-deductive reasoning and propositional thought, to conjure more general logical rules through internal reflection (Berk, 2006). Additionally, they can apply their abstract reason abilities to all areas of life (Siegler & Richards, 1982). Following, it may be the case that adolescents' perceptions of caregiver acceptance-rejection play a greater role, than perceived fairness of punishment, in moderating the link between punishment and self-esteem, when they are aged 11 to 12. Because their social circle though expanded, still centres around their parents and how accepted or rejected they perceived their caregiver to be may still play a significant role unlike during middle adolescence. At ages 15-16, adolescents' perceived fairness of punishment may matter more than perceived caregiver acceptance because their relationship with their caregiver is not the most critical factor in their psychosocial development. Additionally, their growing need for independence from their parents as well as their capacities to think through their own best interests with their greater cognitive awareness, may influence them to place more emphasis on their personal thoughts, and on their friends' views but less on what their caregiver thinks of them.

Within the realm of punishment research, it is also important to acknowledge the existing attitudes towards physical punishment within the particular culture. As pointed out by proponents of physical punishment, aside from the family, the cultural context also buffers potential negative consequences of physical punishment (Bauman & Friedman, 1998). Acceptance of physical punishment varies across cultures and it may contribute to variations in child outcomes across different groups because cultural values and beliefs affect whether punishment is used more instrumentally or emotionally, and how children emotionally respond to it (Gershoff, 2002; Larzelere, 2000). Larezelere's (2000) highlighted five studies which presented evidence of significantly differential effects of spanking by ethnicity. Deater-Deckard et al. (1996), for example, found maternal use of physical punishment predicted externalising behaviours only for European American, but not African American children. The authors suggested that this may be due to the stronger acceptance and preference for physical punishment among African American, in contrast to European American parents, hence affecting the manner in which punishment is used and children's perceptions of its appropriateness. Similarly, Gunnoe and Mariner (1997) found spanking to be negatively related to African American girls' later aggressive behaviours, but positively related to European American boys' later aggressive behaviours.

Majority of the studies, which investigated the link between physical punishment and self-esteem, were conducted in Western countries, such as America. However, attitudes towards childrearing in Western countries are different from those of the Asian cultures in Singapore (Tong, Elliot, & Tan, 1996). Unlike Western cultures, which display a lower tolerance of physical punishment, this form of discipline is popular within the Asian culture. “Spare the rod and spoil the child” is an old saying which reflects the prevalent parental attitude, especially among Singapore Chinese parents, who continue using caning to discipline children and view physical punishment as an effective disciplinary method (Elliot, Thomas, Chan, & Chow, 2000). Being a multi-ethnic society, ethnic differences exist in childrearing techniques, which may lead to differences in usage of physical punishment across ethnic groups. A study conducted by Quah (1999) on the Singapore family found Chinese parents tended to use physical punishment more than other parents, while Malay and Indian parents were most likely to use reasoning, and authority was most frequently used by parents in the group Other. Considering that ethnicity may affect the outcome of physical punishment, this research recruited only Singapore Chinese participants.

The higher percentage of physical punishment usage among Chinese parents may be due to its use to ensure obedience, which is emphasized in the concept of “training”, also known as “guan” in the Chinese culture (Chao, 1994). The importance of children respecting and obeying parents and other adults was demonstrated by immigrant Chinese mothers who placed it as one of their top three childrearing goals (Chao, 1995). This culturally valued goal of child obedience is also similarly viewed as important in Singapore (Shum-Cheung et al., 2006). Moreover, Chinese parents typically adopt stricter parenting practices and use physical punishment because they “believe that infliction of physical pain is necessary to train children's strength in enduring physical hardship in the future” (Tang, 2006, p. 894). With the global awakening to the potential negative outcomes of non-abusive physical punishment, as evident by the increasing number of countries taking the legal stand to ban its use, and in view of research evidence which suggested cultural diversity affects the punishment-outcome link, it is therefore important to study the relationship between physical punishment and self-esteem with a Singapore Chinese sample in order to obtain cultural-specific information to capture a clearer picture of the punishment-self-esteem link within this culture.

Furthermore, cultural values towards how children should be treated and which disciplinary practices are classified as physical abuse differ among nations (Tong et al., 1996). For many studies conducted in Western nations, the researchers considered spanking with an open hand as non-abusive and hitting the child with an object as abusive. However, this view is not commonly adopted in Singapore, where physical punishment is not only accepted, but also permitted by the law. To illustrate, judicial caning is a legal form of punishment for convicted male criminals. Caning is also not outlawed in other contexts, such as in the military, schools and homes, although laws are implemented against child abuse. Wong (1982) reported that “the Chinese use the cane as an instrument to chastise or punish a child, quite legitimately when he needs such punishment. It is sparingly used but so traditional that provision shops sell these canes looped at one end as a handle for the adult to hold” (p.147). Therefore, to incorporate the cultural views of Singapore Chinese, this study did not solely investigate the relationship between spanking and self-esteem, because other forms of physical punishment, such as caning, are not considered as abusive punishment and are used in Singapore. To ensure that only non-abusive physical punishment was investigated, we used the Physical Punishment Questionnaire (Rohner, 1991), which included 13 forms of physical punishment and did not include burning and typing, which were generally considered as abusive by Singaporeans. A study conducted by Singapore Children's Society surveyed members of the public and professionals, and found more than 90% of the respondents considered "burning child with cigarettes, hot water, or other hot things" and "typing child up" to be physical abuse (Elliot et al., 2000).

In our study, we attempted to address some of the methodological shortcomings raised by Gershoff (2002), Holden (2002), Larzelere (2000), and Ripoll- Núñez & Rohner, (2006). First, we adopted Straus's (1994) definition and explicitly evaluated only non-abusive physical punishment. Second, instead of relying on parental reports, we asked recipients of the punishment to be the source of information because evidence supported children's perceptions to be strongly associated with specific outcomes, such as psychological adjustment and self-esteem (Gershoff, 2002; Ripoll- Núñez & Rohner, 2006). Finally, as critics argued that the context within which the child is punished is complex and different environments may affect the effects physical punishment have on children, and noticeably absent in studies are children's perceptions towards punishment, we examined the caregiver-child relationship through adolescents' perceptions of caregiver's acceptance or rejection, and adolescents' perceptions of the fairness of punishment.

Overall, our primary purpose in the current study was to investigate the relationship between non-abusive physical punishment and adolescent's self-esteem, at two age frames, namely at ages 11-12 (T1) and 15-16 (T2). Specifically, we aim to examine whether physical punishment meted out by adolescents' main disciplinarian has negative impact on their self-esteem, or is punishment perceived by adolescents as unfair and/or meted out in a perceived caregiver-rejection relationship that has this effect. We further sought to explore two related questions: (1) Do interrelationships among physical punishment, perceived caregiver acceptance-rejection, perceived fairness of punishment, and self-esteem vary between T1 and T2? (2) Do adolescents' perceptions at T1, concerning caregiver acceptance-rejection and fairness of punishment moderate the relationship between T1 punishment and T2 self-esteem? We hypothesized that higher frequency of physical punishment will be related to lower self-esteem at both age frames (Hypothesis 1). For moderation, perceived caregiver acceptance-rejection and perceived fairness of punishment were hypothesized to moderate the punishment-self-esteem link, at both age frames. However, due to differing cognitive and developmental trajectory at T1 and T2, we further posited that the moderators may play differing level of impact on this link. More specifically, we predicted that perceived caregiver acceptance-rejection will be a stronger moderator, than perceived fairness, at T1 (Hypothesis 2), such that the punishment-self-esteem link, of 11- and 12-year olds, would be stronger as rejection increased. At T2, we predicted that perceived fairness of punishment, as compared to perceived caregiver acceptance-rejection, will be a stronger moderator (Hypothesis 3), such that for 15- and 16-year olds, this link will be stronger as perceived unfairness increased.

Method

Participants

The sample for the current study included 200 Singaporean Chinese undergraduates (46 males and 154 females) enrolled in foundational psychology courses at the National University of Singapore (NUS). Their age ranges from 18 to 35, with the majority (99%) between the ages 18 to 25. Of these 200 participants, 12 reported a family monthly income of less than $1,000 (6%), 13 reported family monthly income ranging from $1,000 to $1,999 (6.5%), 29 reported the range as $2,000 to $2,999 (14.5%) and 26 reported it as within the range of $3,000 to $3,999 (13%). Another 39 (19.5%) and 81 (40.5%) participants, respectively, reported their family monthly income as within the range of $4,000 to $4,999 and $5,000 or more. The median monthly household income group of the current sample (Median = $4,000-$4,999) matched with that of the Singapore population (Median = $4,850) (Singapore Department of Statistics, 2010). No participants who signed up for the study refused to participate or withdrew from the study.

Procedure

Permission to collect data in NUS was received from the NUS's Institutional Review Board. Following this, we recruited participants to respond to an anonymous self-report online survey questionnaire. Although the questionnaire was administered via online, participants were required to respond to it at a stipulated time and location where the researcher was present to ensure that distracters were minimised and the surroundings of the location where the study was conducted were similar for all participants.

All participants were volunteers who received course credits in their foundational psychology courses for participation. The participant information sheet was provided on the first page of the survey and at end of this page, a question was posed to ask participants to give their consent to participate. All participants gave their consent to participate. During a 45-minutes period, participants responded to five self-report questionnaires. That is, the demographic form, Physical Punishment Questionnaire (PPQ), perceived fairness scale, Parental Acceptance-Rejection Questionnaire (PARQ), and the Rosenberg Self-Esteem Scale (RSES). To collect data for the two different age frames, participants responded to all the scales, except the demographic form, twice. In the first section of the survey, participants were asked to recall and answer the PPQ, perceived fairness scale, PARQ and RSES, as they were aged 11- to 12-years old (T1). The second section asked participants to reflect and indicate their experiences for the same four scales, but when they were aged 15- to 16-years old (T2). To minimise response bias, participants were randomly assigned to answer different versions of the questionnaires, which only differed by the varying order of presentation of the different scales. The order of the scales for the two sections of each version was the same.

Before participants answered the scales in the first section of the survey, they were asked to “Recall and describe briefly (1-2 lines) a significant event between your main caregiver and you when you were 11-12 years old. Tell us why it was significant and your feelings at that time”, before answering the scales pertaining to T1. Similarly, before the start of the second section of the survey (T2), participants were asked to describe a significant event that occurred between them and their main caregiver when they were 15- to 16-years old. This measure, to recall a significant event and to describe the emotions attached to it, was included to enhance recall of their adolescence years, specifically the interested age frame, and to increase the accuracy of recall. All participants answered this question and reported feelings related to the reported event, indicating that participants were conscientiously responding to the questionnaires.

After the procedure was explained to the participants, they proceeded with the survey and were asked to log off their computers upon completion to ensure the confidentiality of their responses. Each of the five measures is described below. English language version of all instruments was used because English is the medium of instruction in Singapore schools.

Measures

Demographics. The demographic form elicited information about the participants' sex, current age, and family monthly income.

Physical Punishment.The frequency of physical punishment experienced by participants was assessed by the adult version of the Physical Punishment Questionnaire (PPQ: Rohner, 1991). On the first part of the questionnaire, participants indicated who their main disciplinarian was at the respective age frame (e.g., "Please tell us who your main disciplinarian was when you were aged 11-12. Your main disciplinarian is the adult who punishes you when you disobey or don't behave properly"). Although the PPQ was designed to assess nine conceptual issues that may affect the outcome of physical punishment, namely, frequency, severity, consistency, predictability, average incidence, fairness, deservedness, timing, and parents' use of explanation, for reasons explained above, the present study focused on frequency of punishment. The second part of the PPQ asked participants to recall and indicate the frequency of physical punishment received from their main disciplinarian. Instead of using the single-item scale that measures frequency, we adapted the 13 forms of punishment included in the PPQ that may or may not be experienced by the individual. These include spanking, slapping, shoving, yanking, kicking, beating severely with an object, hitting firmly but not severely with an object, pulling hair, twisting ear, kneeling on hard objects, standing for a long time, pinching, and shaking.

Participants responded to the 13 items on a 6-point Likert scale (1= never, 2 = almost never, 3 = sometimes, 4 = frequently, 5 = almost always, 6 = always), with higher scores indicating higher frequency of physical punishment. An example of a scale item is: “Recall the time when you were aged 11-12, and you did not behave right, did your main disciplinarian slapped you?” Cronbach's alphas for data collected for T1 was .84 and .93 for T2.

Perceived Fairness of Physical Punishment.Although the PPQ includes a measurement for perceived fairness of punishment, we did not use it as the perceived fairness scale because it is measured by a discrete single-item. Instead of asking participants to make a global judgment about their perception of the fairness of punishment, we adapted the 13 items from the PPQ and asked participants to rate the degree of fairness or unfairness of each specific form of punishment. An example of a scale item is phased as follows: "If you were naughty or didn't behave properly when you were aged 11-12, would it be fair for your main disciplinarian to slap you?" Participants rated using a 4-point Likert scale (1 = very unfair, 2 = unfair, 3 = fair, 4 = very fair), with higher scores indicating individuals perceiving physical punishment as more fair. Cronbach's alphas were .83 for T1 and .95 for T2.

Perceived Caregiver Acceptance-Rejection.After indicating who their main caregiver was at the specific age frame (e.g., "Please tell us who your main caregiver was when you were aged 11-12. Your main caregiver is the adult who took care of you the most when you were aged 11-12), participant's perceptions of caregiver acceptance-rejection were measured by the child version of the Parental Acceptance-Rejection Questionaire (PARQ: Rohner, 1991). Participants responded to 60 statements, which were distributed over four scales. These include warmth-affection, aggression-hostility, neglect-indifference, and undifferentiated rejection. Participants reflected on the way their main caregiver treated them at the respective age frames and responded by making judgments in terms of how well each statement described the way their caregiver's treated them. Examples of scale items for T1 include the following: “When I was aged 11-12, my caregiver made me feel wanted and needed” (perceived warmth/affection); “When I was aged 11-12, my caregiver went out of his/her way to hurt my feelings (perceived aggression/hostility); “When I was aged 11-12, my caregiver ignored me as long as I did not do anything to disturb her” (perceived neglect/indifference); “When I was aged 11-12, my caregiver did not really love me” (perceived undifferentiated rejection) (Rohner, 1991). Participants responded to similar items on a 4-point Likert-like scale ranging from 1 (almost never true) to 4 (almost always true).

Items of PARQ were keyed in the direction of perceived rejection such that the higher the score, the more rejection participants perceived. Reverse scoring was performed for some of the scale items, which were written in the direction of perceived acceptance to minimise potential acquiescence response bias (Rohner, 1991). The PARQ has established reliability and validity among American ethnic groups and internationally (Khaleque & Rohner, 2002). Cronbach's alphas for this sample were both .96 at T1 and T2

Self-Esteem.Participants completed the 10-item Rosenberg Self-Esteem Scale (RSES: Rosenberg, 1965), which was designed to assess the individual's self-reported global self-esteem. Equal number of positively worded (e.g. “I take a positive attitude toward myself”) and negatively worded (e.g. “I certainly feel useless at times”) items were presented to minimise the effect of respondent bias. Participants rated each statement on how well it described them on 4-point Likert scale ranging from 1 (strongly disagree) to 4 (strongly agree). Reverse scoring was done for the negatively worded items such that higher scores indicated higher self-esteem. Evidence regarding the reliability and validity of the RSES across 53 countries (Schmitt & Allik, 2005) demonstrated that this measure is psychometrically sound across cultures. In the present sample, the Cronbach's alpha was .92 for T1 and .91 for T2.

Results

Caregiver and Disciplinarian

As presented in Table 1, mothers were the main caregivers, with 68% and 82.5% of the participants reported being cared for by mothers at T1 and T2, respectively. At T1 and T2, approximately 30% of adolescents' main disciplinarian was their fathers and approximately 60% of the adolescents were disciplined by their mothers.

Caregiver and disciplinarian were reported to be the same person by 58.6% and 66.8% of the participants at T1 and T2, respectively. Mother was most often reported as the adult who played both roles at T1 and T2. At T1, 17.7% of the participants reported father as disciplinarian and mother as caregiver, and 22.8% reported this combination at T2. At T1 and T2, less than 1% of the participants reported father as caregiver and mother as disciplinarian.

Types of Physical Punishment

Figure 1 shows the frequency with which adolescents experienced different forms of physical punishments at T1 and T2. Among the different types of punishment, the most frequently experienced by adolescents aged 11-12 was hitting firmly but not severely with an object, followed by spanking, and beating severely with an object. The most frequently experienced form of physical punishment on adolescents aged 15-16 was slapping, followed by spanking, and hitting firmly but not severely with an object. Kicking and kneeling on hard objects were the two least frequently experienced types of physical punishment at both T1 and T2.

As illustrated in Figure 2, adolescents at T1 perceived punishment through the means of standing for a long time, being hit firmly but not severely with an object, and spanking as more fair, in comparison to the other forms of physical punishment. In general, physical punishment was perceived as unfair at T2 but being kicked and beaten severely with an object were perceived as the two most unfair forms of punishment to receive at ages 15-16.
Table 1

Main disciplinarian and main caregiver of adolescents at T1 and T2

Figure 1

Frequency with which different forms of physical punishments were experienced by adolescents at T1 and T2

Figure 2

Adolescents' perceived fairness of the different forms of physical punishments at T1 and T2

Preliminary Analyses

Mean frequency of physical punishment, mean perceived fairness of punishment, mean perceived caregiver acceptance-rejection, and mean self-esteem scores were calculated and derived as our variables for this study. Possible differences between participants on the variables of interest, due to family monthly income and sex, were first examined by respectively conducting ANOVA and t tests. Results showed that frequency of physical punishment, perceived fairness of punishment, perceived caregiver acceptance-rejection and self-esteem, reported at T1 and T2, did not vary significantly by family monthly income or sex.

In view that retrospective data is susceptible to biases, we checked if our memory-based data showed patterns that are consistent with previous studies that used prospective data, by subjecting T1 and T2 means of the variables to two tests. We first examined the differences between T1 and T2 means for each variable using the dependent-samples t test. Table 2 presents the results for this test and the descriptive statistics of each variable. On the average, participants reported almost never been physically punished at T1, and never been physically punished at T2. As shown in Table 2, physical punishment occurred more frequently at T1 than at T2, t(199) = 8.78, p < .001. This significant decrease in mean level of punishment from T1 to T2 is similar to previous studies, which found fewer children being hit as they grow older (Straus, 1994; Straus et al., 1997). Table 2 also shows that participants, on the average, perceived physical punishment as unfair at T1, and very unfair at T2. As would be expected, participants perceived physical punishment at T2 to be more unfair than at T1 (t(199) = 7.83, p < .001), and perceived more caregiver rejection at T1 than at T2 (t(199) = 1.98, p < .05). With respect to self-esteem, no significant differences were found between T1 and T2 means, t(199) = -.35, p > .05.

Table 2

Descriptive statistics and comparisons of means of study variables between T1 and T2

* p < .05. *** p < .001.

Pearson product-moment correlation coefficients were then computed to assess the relationships between T1 and T2 means of the study variables.

Since the present study used a within-participants design, participants' reports for T1 and T2 variables should be correlated because of similar environments and factors which they were exposed to. As expected, four significant correlations were found. Table 3 shows that T1 and T2 means were moderately correlated for physical punishment (r = .50, p < .001), perceived fairness (r = .48, p < .001), self-esteem (r = .44, p < .001), and strongly correlated for perceived caregiver acceptance-rejection (r = .70, p < .001).

Proposed Predictor and ModeratorVariables

Table 3 also provides the intercorrelations among the study variables. Perceived fairness was positively correlated with punishment at T1 (r = .39, p < .001) and T2 (r = .21, p < .001). That is, physical punishment was perceived as more fair as adolescents were punished more frequently. Perceived caregiver acceptance-rejection, which was the other proposed moderator, was also positively correlated with physical punishment at T1 (r = .15, p = .02) and T2 (r= .25, p < .001). Children who perceived more caregiver rejection tended to experience more frequent physical punishment.

Self-Esteem and Physical Punishment

Although no significant correlation was found between self-esteem and punishment at T2 (r = -.01, p = .43), the negative correlation between self-esteem and punishment at T1 was marginally significant (r = -.11, p = .06). That is, as frequency of physical punishment increases, level of self-esteem decreases, only for adolescents aged 11-12 but not for adolescents aged 15-16. Hypothesis 1 that higher levels of physical punishment will be related to lower self-esteem was thus supported only at T1 but not supported at T2.
Table 3

Intercorrelations among the study variables

+p < .10. *p < .05. ***p < .001.

Moderators of the Punishment-Self-Esteem Link

To take into account the developmental trajectory, we conducted separate moderator analyses for T1 and T2. We used the hierarchical multiple regression approach described by Frazier, Tix, and Barron (2004) to test the moderator effects in the present study. The moderator hypothesis is supported if the interaction between the predictor and moderator variables is significant. To minimize the problems associated with multicollinearity among the variables, we first standardized our proposed predictor and moderator variables by creating z scores. Following, we multiplied the predictor and moderator variables using the standardized variables to create four interaction terms (i.e., T1 Punishment x T1 Perceived Acceptance-Rejection,T1 Punishment x T1 Perceived Fairness, T2 Punishment x T2 Perceived Acceptance-Rejection, and T2 Punishment x T2 Perceived Fairness). After creating the interaction terms, we structured the multiple regression equation. To prevent confounding the moderator effect with the individual effects of the predictor and moderator variables (Frazier et al., 2004), we entered the interaction term into the regression equation after entering the standardized predictor and moderator variables from which the interaction term was created.

We first explored the question on whether the influence of physical punishment on adolescents' self-esteem was moderated by their perceptions regarding the fairness of punishment. As presented in Table 4, no significant interactions involving perceived fairness at T1 (t(196)= -.36, p = .72) and T2 (t(196) = .42, p = .68) was found. There were also no significant relationships between perceived fairness and self-esteem at T1 and T2. Thus it appears that adolescents' perceptions concerning fairness of punishment did not moderate the effects of physical punishment on self-esteem and were not directly related to adolescents' self-esteem.