Parenting styles and values: mechanisms of intergenerational continuity and discontinuity

PARENTING STYLES AND VALUES: MECHANISMS OF INTERGENERATIONAL CONTINUITY AND DISCONTINUITY A Thesis Submitted to the Faculty of the Graduate School of Western Carolina University in Partial Fulfillment of the Requirements for the Degree of Master of Arts in School Psychology. By Melissa Florence Littlewood Director: Dr. Bruce Henderson Professor of Psychology Psychology Department Committee Members: Dr. Lydia Aydlett, Psychology Dr. Marie Huff, Social Work March 2009 TABLE OF CONTENTS List of Tables………………………………………………………………….. List of Figures…………………………………………………………………. Abstract……………………………………………………………………….. Introduction…………………………………………………………………… Literature Review……………………………………………………………… Parenting Research…………………………………………………….. Baumrind’s parenting prototypes……………………………… Parenting trends in the 20th Century…………………………… Intergenerational Research…………………………………………….. Intergenerational transmission versus intergenerational continuity……………………………………………………… Child abuse and harsh punishment…………………………….. Mechanisms of Intergenerational Continuity and Discontinuity………. Infant temperament…………………………………………….. Parent-infant attachment……………………………………….. Social support and marital quality………………………………. Changing views of society……………………………………… Other variables………………………………………………….. Purpose of the Present Study…………………………………………… Method…………………………………………………………………………. Participants……………………………………………………………… Measures………………………………………………………………… Background questionnaire………………………………………. Warmth and control……………………………………………… Values……………………………………………………………. Design and Procedure……………………………………………………. Individualism……………………………………………………. Warmth and control……………………………………………… Background questionnaire………………………………………… Results……………………………………………………………………………. Intergenerational Continuity of Individualism…………………………… Intergenerational Continuity of Parental Care……………………………. Intergenerational Continuity of Parental Overprotection………………… Individualism and Parenting Practices……………………………………. Intergenerational Continuity of Family Practices………………………… Discussion………………………………………………………………………… Discontinuity of Individualism…………………………………………… Discontinuity of Parental Warmth………………………………………… Continuity of Parental Control……………………………………………. Individualism and Parenting Styles……………………………………….. Page 4 5 6 8 10 10 10 15 20 21 22 25 26 27 28 29 29 30 34 34 37 37 37 38 38 38 39 39 40 40 42 42 42 44 45 45 45 46 47 Family Practices…………………………………………………………… 48 Limitations………………………………………………………………… 48 Conclusions………………………………………………………………… 50 References…………………………………………………………………………. 51 Appendices………………………………………………………………………… 57 Appendix A………………………………………………………………… 58 Appendix B………………………………………………………………… 59 Appendix C………………………………………………………………… 60 Appendix D………………………………………………………………… 61 4 LIST OF TABLES Table 1. Frequencies of Highest Education Completed and Marital Status………. 2. Descriptive Statistics of Individualism and Perceived Parental Care for Each Generation………………………………………………………….. 3. Correlations for Individualism, Warmth, and Control…………………… Page 36 41 43 5 LIST OF FIGURES Figure 1. Baumrind’s parenting prototypes as part of the continuum of warmth and control……………………………………………………………… 2. Results for the three generations in Bengston’s (1975) study…………. 3. Mechanisms of continuity and discontinuity: variables which influence parent-child interactions and variables which interact with each other… 4. Hypothesized path of control and warmth for each generation………… Page 14 17 26 32 6 ABSTRACT PARENTING STYLES AND VALUES: MECHANISMS OF INTERGENERATIONAL CONTINUITY AND DISCONTINUITY Melissa Florence Littlewood, M. A. Western Carolina University (March 2009) Director: Dr. Bruce Henderson Do parenting styles continue from generation to generation? It is counter-intuitive to think that parenting styles do not continue from generation to generation, yet many researchers have found this to be true (Campbell & Gilmore, 2007; Covell, Grusec, & King, 1995; Staples & Warden Smith, 1954; Woods, Glavin, & Kettle, 1960). When we look at the major events of the 20th century, such as the Great Depression, World War II, and the rise of divorce rates, we can also see changes in social behavior and family structure, either as a direct or indirect result of these major events (e. g. Elder, 1974, 1994). Americans seem to be more individualistic today than they were 75 years ago, perhaps as one of the indirect results of these major events (Stearns, 2003). Many researchers have found strong correlations between parenting styles and cultural variables, such as collectivism and individualism (Baumrind, 1991). Research on intergenerational transmissions and continuity of parenting styles, behaviors, and values within families in the past 75 years provides strong evidence that parenting styles change over time, even from generation to generation within families. Therefore the present study asks the questions, have individualistic values increased with time over the past 75 7 years? And, have parenting style trends gone toward emphasizing higher warmth and lower control from generation to generation? Triads of grandmothers, mothers, and daughters were used, each generation representing a different cohort (the Children of the Great Depression, Baby Boomers, and women who grew up in the 1990s). Maternal warmth and control were measured by the Parental Bonding Instrument (PBI, Parker, Tupling, & Brown, 1979). Individualistic values, as measured by a rank-order scale (Bengston, 1975), rose significantly from generation to generation, as did parental warmth. There was no significant difference in the use of parental control from generation to generation within these families. The present study also found that although number of hours worked per week while raising their daughters did not increase significantly from grandmothers to mothers, there did exist a significant decrease (from the grandmother generation to the mother generation) in number of nights per week families ate dinner altogether while raising their daughters. These findings support much of the research which suggests that changing views of society may play a key role in the discontinuity of parenting practices from generation to generation. 8 INTRODUCTION “ You’re just like your mother! ” This phrase many young women dread hearing may not be as true as it was once thought to be. As many would swear they ended up raising their children just as they themselves had been raised, many researchers have found that parenting does not actually tend to continue from generation to generation (Campbell & Gilmore, 2007; Covell, Grusec, & King, 1995; Staples & Warden Smith, 1954; Woods, Glavin, & Kettle, 1960). Parenting is often viewed as a mechanism of socialization which plays a role in human development (Hill, Mullis, Readdick, & Waters, 2000). When we look at the major events of the 20th century, such as the Great Depression, World War II, and the rise of divorce rates, we can also see changes in social behavior and family structure, either as a direct or indirect result of these major events (e. g. Elder, 1974, 1994). Many researchers have found strong correlations between parenting styles and cultural variables, such as collectivism and individualism (Baumrind, 1991). It seems possible, therefore, that as American culture and values change, parenting styles will follow similar trends. Research on intergenerational transmissions and continuity of parenting styles, behaviors, and values within families over the past 75 years provides strong evidence that parenting styles do change over time. The question arises, then: do the trends in parenting styles in the past 75 years mirror the social changes in America? This paper will discuss important parenting research of the past 100 years, including the significant contributions of Diana Baumrind (1991). As Baumrind’s four parenting prototypes are the parenting styles used in the present study, they will be 9 discussed in depth. This paper will then follow the parenting trends of the 20th century and how parenting has moved from the strict authoritarian to a more authoritative parenting style, to the more permissive parenting style which characterizes the trends of the past 20 years. Intergenerational research will then be discussed with relevance both to understanding parenting trends and the mechanisms of continuity and discontinuity from one generation to the next. The literature review will conclude with the purpose for the present study. 10 LITERATURE REVIEW Parenting Research Baumrind’s parenting prototypes. Baumrind’s parenting prototypes provide a good basis for discussing parenting, as they are multi-dimensional and have been used in much parenting research in the past 35 years (e. g., Brenner & Fox, 1999; Dominguez & Carton, 1997; Dornbusch, Ritter, Leiderman, Roberts, & Fraleigh, 1987). In addition, Baumrind’s research on parenting styles and their relationships to child outcomes such as self-esteem, well-being and school performance have greatly influenced the studies on parenting in the past 35 years (e. g., Brenner & Fox, 1999; Dominguez & Carton, 1997; Dornbusch et al., 1987; Milevsky, Schlechter, Klem, & Kehl, 2008). Therefore, an understanding of her research is necessary. A century of developing parenting research led to Baumrind’s conception of the four major parenting styles. At the turn of the 20th century, with the decline of infant mortality rates and the rise of child labor laws, American culture became more child-focused (Hulbert, 2003) and “ child experts” began to emerge. As researchers studied children more and more in the 1920s, they began to view children as more vulnerable, and therefore the value of parental warmth grew. Freud’s concept of the importance of the mother in early childhood development in the 1940s and Bowlby’s attachment models in the 1950s exemplified the integral roles parents had in the development of their children. In addition, “ child experts” were increasingly interested in parental control: Skinner and Spock told parents to exert much control over the development of their children; but then in the 1960s and 1970s, experts began suggesting that parents should lessen their control, encouraging their children to 11 make the right decisions for themselves (Hulbert, 2003). Furthermore, Maccoby’s research, which found that children in social situations tended to take on the roles of their same-sex parent (Cairns, 1998), was a great example of how children learn from their parents. From the increased interest and use of warmth and control in parenting research, Baumrind identified four main parenting styles: authoritative parenting, authoritarian parenting, permissive parenting, and neglecting/rejecting parenting (Baumrind, 1991). These four styles are defined by different levels of parental acceptance (nurturance or warmth), communication, reasoning, and control (strictness). Before parenting dimensions and styles are discussed, individualism and collectivism will be reviewed briefly, as they have been found to be strongly related to parenting and parenting styles (Baumrind, 1991; Jose, Huntsinger, Huntsinger, & Liaw, 2000). Markus and Kitayama (1991) described collectivism as a social pattern of closely linked individuals who define themselves as interdependent members of a collective unit and described individualism as individual autonomy and independence of the self. Markus and Kitayama stressed the importance of understanding these two different ways of viewing the self, as they may actually influence the way people perceive the world around them. Individualism is more typical in Western cultures, such as in the United States and England, and collectivism is more typical in Asian cultures, such as in China and Japan. Studies have found that in the past 50 years, people in the United States have become increasingly individualistic (e. g., Bengston, 1975; Elder, 1994; Greene, 2008) and as this shift towards an emphasis of independence and personal achievement has risen, so has the trend for parenting practices which emphasize less control and higher 12 warmth (e. g., Elder, 1974; Staples & Warden Smith, 1954; Stearns, 2003; Woods et al., 1960). Parental strictness and control have been used and studied as constructs in research for the past 60 years (Amato & Booth, 1997) and, as dimensions of parenting styles, seem to have the strongest correlations with child outcomes, including development and well-being. Parental strictness is often defined as consistent high expectations for behavior (Rankin, 2005). Rankin described strictness and permissiveness as two ends of a continuum of control whereas permissiveness can be considered to be very low control and strictness is very high control. Parental warmth is feelings and displays of affection for one’s children. Parental rejection is on the opposite end of the continuum of warmth. With a clear understanding of the warmth and control dimensions of parenting, we move now into descriptions of Baumrind’s parenting prototypes. Authoritative parenting is associated with moderate control, but high acceptance and warmth (Baumrind, 1991). Authoritative parents have clear, well-reasoned rules for their children, but they also allow their children age-appropriate independence, especially as they get older. Authoritative parents encourage their children to be independent, reasonable, and creative. They also encourage open communication between themselves and their children. It is well established in the research that authoritative parenting style is positively correlated with individualism (Dornbusch et al., 1987; Herz & Gullone, 1999; Kim & Rohner, 2002). In the United States, parents who use the authoritative parenting style also tend to have children with higher self-esteem, higher well-being, and higher grades in school than children whose parents use other parenting styles (Baumrind, 1991; Herz & Gullone, 1999; Kim & Rohner, 2002; Milevsky et al., 2008). 13 Authoritarian parenting practices are high in strictness and low in warmth and acceptance (Baumrind, 1991). These parents demand their children to follow their rules without question. Authoritarian parents encourage obedience, respect for authority, and discourage open communication between their children. Many researchers have found strong correlations between authoritarian parenting style and collectivism (e. g., Herz & Gullone, 1999; Kim & Rohner, 2002). Permissive parents are highly warm and accepting, but are very low in control (Baumrind, 1991). Permissive parents are very supportive and responsive to their children, but do not put many limits to their children’s behavior and punishment is less severe and less frequent. Both self-esteem and school performance tend to be lower in children of permissive parents than in children of authoritative or authoritarian parents. Rejecting/Neglecting parents are extremely low in both permissiveness and acceptance (Baumrind, 1991). Rejecting/Neglecting parents do not pay much attention to the needs of their children and are often rejecting towards them. Not surprisingly, children whose parents are rejecting/neglecting tend to have the lowest self-esteem and school performance of children of all four parenting types (Milevsky et al., 2008). A few recent studies have criticized Baumrind’s parenting prototypes for not being able to classify all parenting styles (Brenner & Fox, 1999; Dominguez & Carton, 1997; Kim & Rohner, 2002). Many studies using Baumrind’s parenting prototypes have found that the majority of their samples do not fit into any of the four styles (e. g., Dornbusch et al., 1987; Kim & Rohner, 2002). For instance, there exist parents who are both highly accepting and highly punitive with their children; however these parents cannot be classified within any of Baumrind’s four parenting prototypes. Because of this, 14 many recent researchers, when investigating correlates of parenting, use levels of parenting dimensions and behaviors, rather than specific parenting styles, in order to include all participants (e. g., Brenner & Fox, 1999; Herz & Gullone, 1999). However, the wide use of Baumrind’s parenting styles in research over the past 35 years suggests these styles are still valuable. It is also possible to measure the dimensions within the styles in order to define styles; for instance, higher scores on a measure of warmth and lower scores on a measure of control may better characterize the permissive parenting style (as well as include more parents) than a simple unidimensional measure of permissive parenting style (see Figure 1). The rejecting/neglecting parenting style is not included here because the nature of the present study makes it unlikely we would encounter this parenting style. Figure 1. Baumrind’s parenting prototypes as part of the continuum of warmth and control. The solid line indicates rising levels of warmth; broken line indicates declining levels of control. 15 Parenting trends in the 20th century. With a better understanding of the research on parenting and the structure of Baumrind’s parenting prototypes, we move now into a review of the parenting trends in the past 100 years. The following research suggests many changes in families and parenting over the past century which seem to follow the historical and social changes which occurred with them. As society has become less collectivistic and more individualistic, parenting styles have risen in warmth and declined in control. Economic and social change in the early 20th century took the role of children within the family from economic asset to economic burden (Stearns, 2003). Elder’s (1974) Children of the Great Depression study explicitly showed how economic and social change could transform family structures and parenting styles. Elder’s was a longitudinal study on a group of people from childhood through their 40s who grew up during the Great Depression. What Elder found was that two large historical events (the Great Depression and World War II) had large impacts on these children as they grew up. When Elder looked at the (now grown up) children as a cohort, he found many discontinuities from their parents’ generation. For instance, the children of the Great Depression got married at a much younger age than their parents and had significantly more children. Other longitudinal studies around this time period found similar results (Amato & Booth, 1997). Amato and Booth (1997) and Elder (1974) both provided two explanations for the younger age of marriage: first, parents of the children of the Great Depression waited longer to get married because they could not afford to get married at a younger age and second, by the time these children were just old enough to get married (in the early 1950s), the economy was booming and there was a surplus of jobs. These 16 longitudinal studies provide ample evidence for the existence of a possible relationship between historical events and social behavior. Elder’s study also set up the beginning of the major changes in parenting styles. Elder (1974) found that this cohort as parents used more reasoning and autonomygranting, (which fits well into the definition of authoritative parenting) than their own parents. The children of the Great Depression allowed their children to have a voice as part of the family and they emphasized the importance of achievement. This study showed how historical events can change family dynamics, indicating that intergenerational transmissions of parenting styles can be quite complex. It also showed the beginning of a shift in parenting styles from authoritarian to authoritative. Hareven (1978) discussed the differences in the family cycle during the early 20th century compared to the trends in the latter part of the 20th century. In the early 20th century, people were parenting for most of their adulthood, due to later age of marriage, the larger number of children within families, and shorter life spans. Parents usually did not have an “ empty nest” when their children were launched. Because of this, families were more of a collective unit during this time than they were in the late 70s. Many other researchers have also found differences in families and society that came about after World War II. Mills (1987) argued that the different values and attitudes of Baby Boomers (those born from 1946-1966) reflect the changes of society after World War II, whereas the attitudes of the parents of the Baby Boomers are more reflective of the Great Depression era. Baby Boomers are more individualistic than their parents and Mills argued that without major historical events, generations would have many fewer differences, regardless of age. 17 Several studies in the 1970s found similar trends in the shift of values. Bengston (1975) did a study in which he looked at individualism and collectivism in three generations of families. He found a significant increase in individualism from the grandparents (the first generation) to the grandchildren (the third generation). The first generation’s means on the individualism-collectivism index were very close to the collectivism pole, whereas the third generation’s means were very close to the individualism pole (see Figure 2). Second generation means were between the first and the third on the individualism and collectivism measure. Bengston discussed the implications of these results as between-generation effects rather than within-family effects; in other words, although the participants of this study were the triads of family members, he found it was the cohorts who changed over time. The progression of time, whether related to age or to historical change, seemed to be related to the effects in this study. The results of Bengston’s study mirror the trends in other intergenerational studies which measured parenting styles. Figure 2. Results for the three generations in Bengston’s (1975) study. G1 are grandparents; G2 are parents; G3 are grandchildren. A new emphasis from parenting experts in parenting magazines and books in the 1950s and 60s was on less strictness (Hulbert, 2003; Stearns, 2003). Experts at this time were beginning to advise parents to discourage strict obedience in their children. Instead, 18 they said parents should let their children do what is right and make the right decisions on their own, giving children more independence. Two studies during this time reflected this trend. Staples and Warden Smith (1954) and Woods et al., (1960) did intergenerational studies looking at grandmothers’ (both paternal and maternal) and mothers’ parenting attitudes toward child rearing practices. Both studies found that mothers were significantly more permissive than grandmothers, both as a cohort and within the family. In Staples and Warden Smith’s study, child rearing practices were organized into eight subscales: general home standards, verbal standards, expression of hostility, weaning, feeding, and thumb sucking, toilet training, sexual behavior, boy-girl differences, and crying. They found that in attitudes towards these practices, grandmothers had significantly stricter, authoritarian parenting attitudes than the mothers, who tended to have more permissive attitudes toward child-rearing practices. Woods et al. (1960) looked at seven areas of child rearing practices in this study: sucking and feeding, toilet training, dependency, aggression, sex and modesty, bedtime restrictions, and manners. They found that within these seven areas, the daughters tended to have more permissive attitudes than their mothers. The researchers concluded that these results tend to coincide with shifting parenting attitudes since the 1920s and 1930s, meaning that the trend of parenting attitudes was becoming more permissive. In the late 20th century, parents were much more permissive and there was a large trend away from physical and harsh punishment (Stearns, 2003). Rather than strict rules and emphasis on obedience, more parents were using talking, reasoning, and rationalization to shape behavior. Martin, Halverson, Wampler, and Hollett-Wright (1991) saw this trend when they looked at mothers’ and grandmothers’ reports of their 19 parenting goals and styles (the grandmothers had to reflect on when they themselves were mothers). They found that grandmothers viewed children as less self-sufficient than mothers did and grandmothers’ parenting styles were more controlling and less nurturing than mothers, which well characterized the authoritarian parenting style. Covell et al. (1995) looked at mothers’ and grandmothers’ endorsements of different disciplinary techniques and they found that mothers used explanation significantly more than grandmothers, which is a common characteristic of the authoritative parenting style. During the 1960s and 1970s, more women began to work full-time outside the home and fulfill other roles besides housewife and mother (Amato & Booth, 1997). In the 1980s and 1990s, another significant social change occurred in America which changed families: the rise in divorce and single-parenthood (Hulbert, 2003). In 1997, almost onethird of children lived in single-parent families in the United States. Along with singleparenthood came smaller income, which led to lower economic stability, and therefore lower parental availability (both physically and emotionally). More children were put in day-care and after-school care in the past 20 years. Stearns (2003) argued that these parents, who spend more time working and therefore less time with their children, feel obligated to make their time spent with their children as enjoyable as possible. It seems likely, then, that as these parents are trying to make their time as enjoyable as possible with their children, they are likely putting less effort into being strict, thus the rise of the permissive parenting style in the past 20 years. It seems clear there has been a shift in the past 100 years away from the traditional collectivistic authoritarian parenting styles. Biblarz, Bengston, and Bucur (1996) identified four important social changes over the 20th century: increasing 20 availability of nonmanual jobs, a shift in childrearing values from obedience to autonomy, the growth of alternate family structures, and changing gender roles. These social changes, coupled with the major historical events in the 20th century, seem to reflect society’s increase in individualistic values and this is evident in the shift in parenting style trends over this time. Intergenerational Research Intergenerational research is a useful way to observe historical trends over time. When continuity or discontinuity exists between generations in families, it suggests there are forces outside the family which play a role in shaping values and behavior. Transmission and continuity are often the terms used in research on parenting behavior and values and their existence (or lack thereof) from one generation to the next. However, understanding the theoretical differences between the definitions of transmission and continuity in intergenerational research is important. Intergenerational transmission is the process of one generation, either intentionally or unintentionally, influencing the values and behaviors of the next generation (Van Ijzendoorn, 1992). Van Ijzendoorn provided three ways in which intergenerational transmission of parenting behavior occurs. In the first way, the child learns by observing her mother with other children. In the second way, the child learns by her own experiences with her mother. In these two ways, behavior is transmitted because the mother is the model from which the child learns and the child imitates the behavior when she has her own children. In the third way, the child learns the parenting behavior by her mother coaching her while she is interacting with a child. In this third way, the mother is purposefully shaping the child’s behavior and the child continues this reinforced behavior with her own children. Continuity of parenting 21 includes the transmission of parenting from one generation to the next through socialization and genetics, but its definition also includes other mechanisms for existence from one generation to the next, including shared environments (Van Ijzendoorn, 1992). Intergenerational transmission versus intergenerational continuity. There are unique family factors that persist from generation to generation, but there are also societal factors that influence whether or not something continues to the next generation (Hareven, 1978). This can be seen in similarities between cohorts. Hareven gave the example of the age at which women have children. If in a particular family from generation to generation the women have their first child at 18 years old, then in a later generation a woman has her first child at 25 years old, Hareven argued it is likely that others in her cohort have also started having children at around 25. It is also likely that there is something in history accounting for this discontinuity between generations, such as the increased number of women who go to college. There are within-family effects and between-generation effects in intergenerational research (Bengston, 1975). Within-family effects are those which are passed from one generation to the next within a particular family. These are either genetic or have been learned behaviorally from parents. Studies which look at the intergenerational transmission of child abuse, for instance, show within-family effects; parents who abuse their children are extremely likely to have been abused by their own parents. Therefore, the abusive behavior continued within the family to the next generation. Between-generation effects are the changes from one generation cohort to the next generation cohort. These often mirror historical, societal, and cultural changes. Elder’s (1974) longitudinal study on the children who grew up during the Great 22 Depression showed many between-generation effects, such as the increased number of children that this cohort had as parents compared to the number of children their own parents had. Intergenerational transmission only pertains to within-family effects. Using the term generational continuity includes intergenerational transmission; however, using the term intergenerational transmission by itself is limiting. The abbreviations G1, G2, and G3 are used in intergenerational research and in the present study. These abbreviations represent three generations, either as three separate cohorts, or as three generations within a particular family. G1 is the first generation, or grandparent generation. G2 is the second, parent generation. G3, the third generation, is the grandchild of G1. G3 may or may not be a parent him- or herself. Child abuse and harsh punishment. It is important to look at child abuse studies when trying to understand the nature of the continuity and discontinuity of parenting practices. When participants in these studies report that they were either recipients of harsh punishment or that they use harsh punishment on their own children, it is unlikely they are being dishonest, even though the subjectivity of the definition of harsh punishment may be present. It seems even more likely, in fact, that subjects withhold information about being abused or about being abusers. Therefore when continuity exists in abuse and harsh punishment between generations, there is less ambiguity about the nature of the construct than in other studies which look at continuities such as parenting style. Lunkenheimer, Kittler, Olson, and Kleinberg (2006) found that maternal physical punishment was highly likely to be transmitted from generation to generation and Egeland, Jacobvitz, and Sroufe (1988) concluded that parents who physically abuse their 23 children are highly likely to have been abused themselves; however children who have been abused are not statistically likely to abuse their own children when they become parents. It has been well established that the design of the study makes a big difference in whether or not harsh punishment can be said to continue or discontinue from generation to generation (Egeland et al., 1988; Rutter, 1989). This means studies which are retrospective find that parents who use harsh punishment with their children are very likely to have had parents who used harsh punishment on them; however studies which are prospective find that parents who use harsh punishment on their children do not tend to have children who use the same practices on their own children. Clearly the methods of intergenerational abuse research must be considered when drawing conclusions on the data. What these studies do show, however, is that transmission of parenting behaviors is complex and there may be one or many intervening variables that play a role in both continuity and discontinuity. Many studies have found intervening variables in both the continuity and discontinuity of abuse and harsh punishment. Putallaz, Costanzo, Grimes, and Sherman (1998) found that many researchers have concluded that the underpinnings of continuity of child abuse is an interaction of social learning theory and lack of outside support; meaning that for the abuse to continue, other important factors must exist. Other researchers have found that abusive families tend to socially isolate themselves and parents tend to discourage openness to experiences (Grusec, Hastings, & Mammone, 1994). De-emphasis on societal values may play a role in the continuity of abuse from generation to generation. Finally, researchers have also discussed the possibility of third variables in continuity such as mental illness (Rutter, 1989). Mental illness and problems 24 with psychosocial functioning can be highly heritable. In addition, poor parenting is often likely to be a manifestation of these types of illnesses. Therefore as mental illness tends to be genetically passed from one generation to another, so are poor parenting practices. Research on child abuse and harsh punishment has also revealed many protective factors which may account for the discontinuity to the next generation. Egeland et al., (1988) discussed the importance and positive influence that supportive non-abusing adults have on children who have been abused, including that this may be a protective factor that contributes to the discontinuity of abuse to another generation. They also found that mothers who had been abused themselves as children and who did not abuse their own children had likely undergone therapy and/or had forgiven their own parents. Other researchers have found that parents who were abused themselves but do not abuse their own children were more expressive about their pasts than parents who continued the abuse (Putallaz et al., 1998). The discontinuing parents were more able to clearly express their own abuse and they were expressive about intentionally discontinuing the abuse for another generation. The continuity of abuse, therefore, is not simple; many outside variables may play direct and indirect roles in whether or not abuse continues. What these studies on the continuity and discontinuity of child abuse and harsh punishment tell us is that passing parenting behaviors from generation to generation is the result of complex interactions. Many other variables play complex, interacting roles in whether or not a child who has been abused will be an abuser some day to his or her own children. These studies also tell us that the interactions between parent and child play a large role in shaping children’s future behavior, even if they do not model the same behavior. Parental socialization is an important component in shaping the future behavior 25 of children, but it is not the only variable in the equation. Outside forces and variables may be more powerful than parenting, which perhaps is the key in discontinuity of parenting practices. Mechanisms of Intergenerational Continuity and Discontinuity This section reviews the mechanisms of intergenerational continuity and discontinuity of parenting styles, behaviors, and values. There are many variables (temperament, attachment, social support, etc.) which researchers have found to influence parent-child interactions (including styles, behaviors, and values) and each other (Grusec et al., 1994; Perren, Von Wyl, Burgin, Simoni, & Von Klitzing, 2005; Putnam, Sanson, & Rothbart, 2002). However, as the continuity and discontinuity of child abuse is mediated and moderated by outside variables, it seems the same mechanisms exist in other parentchild interactions. Figure 3 represents the variables discussed in this review, their relationship to parent-child interactions, their relationship to each other, and whether or not they tend to continue from generation to generation. Variables other studies have found that may account for the discontinuity of parenting practices from generation to generation also include age of parent at child’s birth, existence of antisocial behavior in adolescence, discipline practices of the other parent (Capaldi, Pears, Kerr, & Owen, 2008), and personality (Olsen, Martin, & Halverson, 1999). As there are many variables which influence parenting styles, behaviors, and values, it seems that only current societal views are inconsistent from one generation to the next. 26 Figure 3. Mechanisms of continuity and discontinuity: variables which influence parentchild interactions and variables which interact with each other. Goodnow (1994) termed continuity and discontinuity as convergence and divergence. Goodnow proposed that convergence and divergence of values between generations has to do with the explicitness or implicitness of parental communication, how the child perceives the message, and whether or not the child rejects the message in relation to his or her own world view. The explicitness or implicitness of the parent’s interactions or messages may therefore be an additional variable in the mechanisms of continuity and discontinuity. Infant temperament. Temperament studies have many implications for intergenerational research. Temperament is both passed genetically from generation to generation and is something that elicits environmental changes (Putnam et al., 2002). 27 Many studies found that infants with an easy-going sociable temperament are treated differently by their parents and others than infants who are more fussy and irritable; biological temperament, therefore, may play an important role in creating one’s social environment. Researchers have even postulated that parents raise each of their children differently due in large part to the different temperaments of their children. Temperament studies have also found that parent temperament is related to parenting styles (Putnam et al., 2002). For instance, mothers who have a high negative affect tend to have low responsiveness, low sensitivity, and high power-assertion in their parenting styles. Therefore, as temperament is seen as something passed from generation to generation, parenting styles–at least dimensions of parenting styles which are related to temperament–should continue somewhat as well from generation to generation. Parent-infant attachment. Attachment between parents and their infants is another important mechanism of intergenerational transmission and continuity (Grusec et al., 1994). Bowlby’s working model of attachment (Goldberg, 1999) postulated that parentinfant attachment relationships were the basis of the infant’s social relationships for the rest of his or her life. Relationships with other family members, friends, spouses, and their own children stem from their attachment to their parents as an infant. In other words, if a mother forms a secure attachment with her infant, that child is more likely to form healthy relationships with friends and his or her spouse one day. However, when mothers have insecure or disorganized attachments with their infants, the infants develop similar working models of relationships as they get older as well. Putallaz et al. (1998) reviewed literature which supported that mothers who reported secure attachments in their childhood were more responsive to their own children’s needs and formed secure 28 attachments with them. Like temperament, attachment styles are also often present from generation to generation and have an influence on parenting styles. In addition, some researchers have found strong evidence that temperament of infants may play a role in mothers’ attachment styles, especially when mothers have outside social and emotional stressors (Putnam et al., 2002). Parental social support and marital quality are discussed more in depth in the next section. Social support and marital quality. Social support and marital quality are two more variables which seem to play complex roles in the development of parenting styles and behaviors. Putnam et al. (2002) discussed several studies which found that parents who had low social support (coupled with temperamentally irritable babies) were less responsive to their babies’ needs, even as the babies grew out of infancy and became less irritable. Other studies have found marital satisfaction to be a moderating factor in the discontinuity of physical punishment, maternal anger, and punitive control from one generation to another (Lunkenheimer, et al., 2006). Interestingly, Perren et al. (2005) found that marital satisfaction is transmitted maternally from generation to generation, but only when the daughters have had children of their own, suggesting that the addition of children into the family system changes the family dynamics. Researchers who have applied the life course perspective to the interpretation of the findings of these studies postulate that social support and marital quality may have long term consequences, even for later generations (Amato & Cheadle, 2005). Socioeconomic status has also been linked to lower social support and marital quality (Amato & Booth, 1997; Capaldi et al., 2008). Researchers have also postulated this correlation goes in both directions: perhaps poverty may adversely affect marital 29 quality and poor marital quality (especially if it leads to divorce, which is likely) may increase the likelihood of economic hardships. Elder (1994) argued that the link between economic instability and poor parenting was marital discord and individual distress. All of these variables seem to go together–each having some reciprocal effect on the others (see Figure 3). Changing views of society. Research has found many links between the changing views of society and the discontinuity of parenting styles from generation to generation; the changing views of society stem from historical and economic change (Elder, 1994). Furthermore, education level has been found to be positively correlated with more democratic, authoritative parenting styles (Campbell & Gilmore, 2007) and as education levels in men and especially in women have risen in the past 30 years, so has the increased use of authoritative parenting practices. Grusec et al. (1994) argued that societal cultural values had great influences on parenting values and practices. This included both historical and socioeconomic trends in parenting. For example, middle class Caucasian American parents emphasize more autonomy than the lower and working classes (Elder, 1994). However, parents in all socioeconomic classes emphasize more autonomy than previous generations. In addition to historical change, peers, other authority figures, and media also seem to play important roles in shaping behavior and values (Elder, 1994; Putallaz et al., 1998; Stearns, 2003). Other variables. Putallaz et al. (1998) and Capaldi et al. (2008) discussed age at becoming a parent as a variable which may interfere with one’s parenting attitudes. Putallaz et al. (1998) reviewed several studies which found that as people get older, their parenting attitudes and values change. Putallaz et al. also reviewed studies which 30 examined changes in attitudes as one becomes a parent; however the research is limited in that area. Furthermore, Capaldi et al. (2008) argued that younger parents tend to be less financially stable and are less mature than older parents. Purpose of Present Study As the review of the literature indicated, there are many parenting styles, behaviors, and values which continue and discontinue from generation to generation and there are many different mechanisms of their continuity and discontinuity. However, as many things continue from generation to generation, historical events and changing societal views seem to be the biggest players in discontinuity (Campbell & Gilmore, 2007; Goodnow, 1994). As the social values of individualism and collectivism are strongly correlated to parenting styles (Baumrind, 1991), and parenting styles seemed to have shifted from authoritarian to authoritative, to permissive (e. g., Campbell & Gilmore, 2007; Elder, 1974; Staples & Warden Smith, 1954; Stearns, 2003; Woods et al., 1960) over the past 100 years, it seems likely that individualism and collectivism followed the same path; that is, as parenting styles reflected more individualism over time, have people become less collectivistic and more individualistic? And furthermore, what are the consequences of this big change in self-construal? The present study seeks to answer this question. Based on the research on parenting styles and individualism and collectivism, authoritarian parenting style has been associated with higher levels of collectivism and authoritative parenting style has been associated with higher levels of individualism (Baumrind, 1991). Our review of the parenting trends and values in the past century has shown a trend from authoritarian to authoritative to permissive parenting (e. g., Campbell 31 & Gilmore, 2007; Elder, 1974; Staples & Warden Smith, 1954; Stearns, 2003; Woods et al., 1960), and another trend from collectivistic to increasing individualistic values (e. g., Bengston, 1975; Elder, 1994). Based on this research: Hypothesis 1 (H1) is that participant values should reflect increasing individualism from G1-G3 within families and values should be significantly different from generation to generation within families. Hypothesis 2 (H2) is that the warmth dimension of parenting styles should increase from G1-G3 (see Figure 4). Hypothesis 3 (H3) is that the control dimension of parenting styles should decrease from G1-G3; these should also be significantly different from generation to generation within families (see Figure 4). Hypothesis 4 (H4) is that individualism in grandmothers (G1) should be negatively correlated with parental control as reported by their daughters (G2) and individualism in mothers (G2) should be negatively correlated with parental control as reported by their daughters (G3). Individualism in G1 and G2 should likewise be positively correlated with parental warmth as reported by their daughters. 32 Figure 4. Hypothesized path of control and warmth for each generation. Hypothesis 5 (H5) is that we expect the amount of time mothers have worked per week while raising their daughters will increase significantly from G1-G2, based on findings from Stearns (2003) and Hulbert (2003). Amount of time worked per week represents the availability of the mother (G1 and G2) to her daughter (G2 and G3, respectively) while the daughter was being raised. Furthermore, we expect (perhaps as a consequence of mothers working more) number of times per week families ate dinner together will decrease from generation to generation. These trends we expect to find will support increasing individualism, and a greater trend towards permissive parenting styles. Why use mothers and daughters? The present study uses triads of females rather than males due to the unique ways families influence females versus males (e. g., Amato & Booth, 1997; Amato & Cheadle, 2005; Campbell & Gilmore, 2007; Caspi & Elder, 33 1988). Families have a stronger social influence on daughters than they do sons. Parents tend to encourage their sons to go out and be independent, whereas they encourage their daughters to stay closer to the family and participate in more family events. Furthermore, the influences of mothers and fathers tend to be different with their daughters and sons; for instance, daughters tend to be closer to their mothers than sons are. Therefore, the present study controls for gender without using statistics to avoid possible complex interactions. Controlling for gender also makes differences between generations more powerful, and the differences found are more likely to be the result of historical events and societal views. The three generations were chosen to represent a cohort who were Children of the Great Depression (G1), a cohort of Baby Boomers (G2), and a cohort who grew up during the 1990s (G3). 34 METHOD Participants Ninety-two women participated in the present study. Of these participants, 31 were grandmothers (G1), 30 were mothers (G2), and 31 were daughters (G3). Although 35 families participated, only 27 families were complete triads consisting of a daughter, her mother, and her mother’s mother. In cases where only two generations within one family participated (such as mothers and daughters), their data was included in analyses when appropriate. Participants were recruited to create a sample of convenience. About half of the daughters were recruited from the psychology department at Western Carolina University in North Carolina. Other daughters, mothers, and grandmothers were recruited through mass e-mail, online social networking posts, and through the efforts of friends and family. When a daughter, mother, or grandmother indicated interest in participating, the researcher requested contact information for her and the other two generations and subsequently mailed out surveys (see Appendices B, C, & D), consent forms (see Appendix A), and stamped, self-addressed envelopes to each of all three generations of women within the family. Participants who did not return surveys after 30 days were sent a reminder via e-mail. All participants were treated ethically and signed a consent form (see Appendix A). Returned self-addressed envelopes indicated participants were from 13 different states and from the coastal to mountain regions of North Carolina. Twenty-one grandmothers were in the 61-75 years old age bracket and 10 grandmothers reported being in the 76-90 years old age bracket. One mother reported being in the 26-40 years old age bracket and 29 mothers reported being in the 41-60 years old age bracket. Finally, 35 27 daughters reported being in the 18-25 years old age bracket, three were in the 26-40 years old age bracket, and one daughter was 17 years old (consent was obtained from her mother). Frequencies of marital status and highest education completed can be found in Table 1. 36 Table 1 Frequencies of Highest Education Completed and Marital Status G1 G2 G3 Highest Education Less than high school High school diploma or equivalent High school, plus business or trade school 1-4 years of college, but did not graduate Graduated from college with BA, BS, or equivalent Postgraduate professional degree such as MA, MS, or PhD Total 31 30 30 2 6 0 5 8 5 6 5 1 5 4 5 9 2 10 1 7 10 Marital Status Married and Living with Spouse Not Married, but Living with Someone Divorced Widowed Never Married Total 21 0 2 8 0 31 26 1 3 0 0 30 3 3 0 0 25 31 Note. G1 = Grandmothers, G2 = Mothers, G3 = Daughters 37 Measures Background questionnaire. All participants filled out a background questionnaire (see Appendices B and C). This questionnaire included demographic information about participants, as well as information about when they were growing up and when the mothers and grandmothers were raising their daughters. Several participants contacted the researchers about their concerns that grandmothers (G1) had some difficulties with understanding how to fill out parts of the questionnaires. In these cases, the researchers encouraged granddaughters (G3) to assist their grandmothers (G1) by asking questions orally while G3 wrote in responses in the questionnaires. Because questions on G1’s questionnaires did not pertain directly to G3, the researchers felt this would not confound the results. Warmth and control. All participants completed the Parental Bonding Instrument (PBI, Parker, Tupling, & Brown, 1979, see Appendix D) in regards to their perceptions of their mothers’ parenting until they were 16 years old. The PBI has 25 items with two scales: perceived maternal care (12 items) and perceived overprotectiveness (13 items). Items are scored on a four-point Likert scale. High scores on the care scale indicate high perceived maternal affection, emotional warmth, empathy, and closeness (“ My mother spoke to me with a warm and friendly voice”). Low scores on the care scale indicate perceived maternal emotional coldness, indifference, and neglect (“ My mother did not talk with me very much”). High scores on the overprotectiveness scale indicate perceived maternal control, overprotection, intrusion, excessive contact, infantilization, and prevention of independent behavior (“ My mother tried to control everything I did”). Low scores on the overprotectiveness scale indicate perceived maternal allowance of 38 independence and autonomy (“ My mother let me do those things I liked doing”). Both scales have adequate reliability and validity and have been widely used in research for the past 25 years (e. g., Leerkes & Crockenberg, 2006; Lindhout, Markus, Hoogendijk, Borst, Maingay, Spinfoven, van Dyck, & Boer, 2006; Manian, Papadakis, Strauman, & Essex, 2006; Parker et al., 1979). Of the 27 families in which all three generations participated, two grandmothers (G1) and one mother (G2) did not fill out the PBI appropriately (they marked several lines with “ X’s” instead of writing down the corresponding numbers), one grandmother (G1) left an item blank on the overprotection scale, and one mother (G2) left an item blank on the care scale. As a result, analyses were only computed for 23 families on each scale. Values. Participants rank-ordered eight items which reflect differing levels of individualism (four items representing highly individualistic values, and four items representing low individualistic values, see Appendix D). These items were used in an intergenerational study by Bengston (1975). Each item was multiplied by ranks, and then scores were added up to get a global individualism score. Scores are on a continuum of individualism so that possible score ranges are from -80 (indicating very low individualism) to 80 (very high in individualism). Design and Procedure Individualism. To look at the trends of individualism in the three generations, linear trend analyses were performed. If a trend existed, one-way within-subjects ANOVAs were computed as well as follow-up tests for differences in individualism between generations, p < . 05. One-tailed Pearson correlations were run between grandmothers’ individualism and grandmothers’ warmth and control as reported by 39 mothers as well as between mothers’ individualism and mothers’ warmth and control as reported by daughters. Warmth and control. To look at the trends of perceived warmth and control in our three generations, linear trend analyses were performed. If a trend existed, one-way within-subjects ANOVAs were computed as well as follow-up tests for differences and direction in perceived warmth and control between generations, p < . 05. Background questionnaire. To provide support for why discontinuities in parenting practices and individualism from generation to generation might occur, data about other family practices indicative of a changing society, (specifically hours per week mothers spent at work and number of days per week families ate dinner together while daughters were growing up) were taken into account. T-tests were computed between grandmothers (G1) and mothers (G2) for amount of time mothers worked per week and amount of days families ate dinner together. 40 RESULTS Intergenerational Continuity of Individualism A linear trend analysis was performed on individualism scores within families for grandmothers, mothers, and daughters. A significant linear trend existed within the three generations, F(1, 26) = 14. 04, p < . 005. A one-way within-subjects analysis of variance (ANOVA) was then conducted to compare scores for each generation within families for individualism. The means and standard deviations for the individualism scale are presented in Table 2. There was a significant effect for generation, F(2, 25) = 7. 38, p < . 005, which supported H1. Post hoc comparisons indicated a significant (p < . 05) increase in individualism from grandmothers (G1) to daughters (G3) and from mothers (G3) to daughters (G3), but no significant difference between grandmothers and mothers (p > . 05) for individualism. 41 Table 2 Descriptive Statistics of Individualism and Perceived Parental Care for Each Generation G1 G2 G3 Individualism n Mean Standard Deviation 27 -36. 67 33. 77 27 -18. 52 34. 08 27 -. 37 39. 47 Parental Care (Warmth) n Mean Standard Deviation 23 24. 65 9. 76 23 26. 61 8. 60 23 31. 48 3. 58 Parental Overprotection (Control) n Mean Standard Deviation 23 13. 13 5. 70 23 11. 83 4. 37 23 12. 65 3. 77 Note. Maximum Parental Care score = 36. Maximum Parental Overprotection score = 39. G1 = Grandmothers; G2 = Mothers; G3 = Daughters. 42 Intergenerational Continuity of Parental Care A linear trend analysis was performed on perceived parental care (otherwise described as parental warmth) scores within families for grandmothers, mothers, and daughters as they reported about their own mothers. A significant linear trend existed within the three generations, F(1, 22) = 10. 95, p < . 005. A one-way within-subjects analysis of variance (ANOVA) was then conducted to compare scores for each generation within families for care. The means and standard deviations are presented in Table 2. There was a significant effect for generation, F(2, 21) = 6. 84, p < . 01, supporting H2. Post hoc comparisons indicated a significant (p < . 01) increase in parental care from grandmothers (G1) to daughters (G3) and from mothers (G3) to daughters (G3), but no significant difference between grandmothers and mothers (p > . 05) for perceived parental care. Intergenerational Continuity of Parental Overprotection A linear trend analysis was performed on perceived parental overprotection (otherwise described as parental control) scores within families for grandmothers, mothers, and daughters as they reported about their own mothers. No significant linear trend existed within the three generations, F(1, 22) = . 107, p > . 05, which did not support H3, and therefore an ANOVA was not performed on the means. The means and standard deviations are presented in