Measuring Parenting from an Epidemiologic Perspective Workshop 

October 4, 2004
Embassy Suites at the Chevy Chase Pavilion
Washington, DC

 

This meeting was held in conjunction with the National Children’s Study, which is led by a consortium of federal agency partners: the U.S. Department of Health and Human Services (including the National Institute of Child Health and Human Development [NICHD] and the National Institute of Environmental Health Sciences [NIEHS], two parts of the National Institutes of Health, and the Centers for Disease Control and Prevention [CDC]) and the U.S. Environmental Protection Agency (EPA).

 

Co-Chairs: Sarah S. Knox, Ph.D., NICHD, NIH, DHHS; Susan M. McHale. Ph.D., Pennsylvania State University; and Cathy Spatz Widon, Ph.D., University of Medicine and Dentistry of New Jersey

Purpose

The purpose of this workshop was to inform the protocol planning process of the National Children’s Study (Study) concerning the measurement of parenting in large cohort studies. Specifically, the goals were to identify the domains of parenting that relate to cognition, social/emotional development, and physical and mental health; to pinpoint the critical developmental time points for measurement; and to identify specific measures that are valid, reliable, and feasible under each domain.

Background

At the Gene-Environment Interactions in the Regulation of Behavior Study workshop held in June 2004, participants considered parenting to be of primary importance in early gene-environment interactions, citing well-controlled research in nonhuman primates in which this had been repeatedly demonstrated. During discussions of parenting in humans, participants at that workshop divided parental behaviors into four domains:

• Parental control of the macro-environment
• Parental stimulation of cognitive development
• Parents as protectors of the child’s emotional and physical states
• Parents as regulators of the child’s emotional and physical states.

They noted that parenting practices and family environments change over time and that a child’s needs and vulnerabilities also change, necessitating measurement of parenting at multiple points in time. The participants of the gene-environment workshop recommended a future workshop on parenting behavior that would deal with issues related to the measurement of parenting in longitudinal epidemiologic studies of child development.

A considerable amount of research has accumulated demonstrating the impact of parenting practices on children’s health, as well as behavioral and psychosocial outcomes. Although no single overarching theory of parenting exists (O’Connor, 2002), research in the effects of parenting on children’s outcomes has provided evidence of the importance of parenting in children’s long-term health and development. Infants who were observed to have a secure attachment to their mothers and who were followed-up later in adolescence, talked about their relationship with their parents in a more open, coherent, and secure manner than did adolescents who were not observed to have been securely attached as infants (Waters et al., 2000). In another study, insecurely attached infants later demonstrated more aggressive and contentious behavior with their peers than did securely attached infants (Cohn, 1990). The mechanism by which parenting of infants is theorized to impact later outcomes for children is through the quality of early parenting influencing the child’s feelings of trust in others, their sense of competence, and their mastery of the environment (Campbell, 1995), as well as their own feelings of self-worth (Bartholomew and Horowitz, 1991).

Aside from direct parent-child interactions that result in secure/insecure attachments, parents influence children’s lives through other means such as through cognitive stimulation and behavioral regulation. In addition, according to Parke (2004), parents act as advisors, coaches, and consultants to their children. In this role, they provide advice and guidance about the child’s relationships with peers, teachers, and others outside of the family. Parents provide the social opportunities in their children’s lives, encouraging and promoting relationships they view as positive while discouraging negative or destructive relationships. These two additional methods of parental influence allow parents to guide children even when they are not present. Over time, parental influence is significant with specific parental behaviors that are consistently experienced by the child, accumulating to produce larger outcomes in later years (Collins et al., 2000).

Charge to Participants

For the present workshop, experts in parenting research were brought together and charged with three primary responsibilities. First, they were asked to consider the domains of parenting as they relate to cognition, emotional development, and social development, as well as physical and mental health in children. Second, the participants were asked to suggest instruments to measure each parenting domain considering the appropriate age of assessment, the psychometric properties of the instruments, and the method of assessment, as well as the time and cost to administer the instrument. The final charge to the workshop participants was to consider the longitudinal nature of the assessment and analysis related to the parenting, specifically:

• The need for evaluation of developmental trajectories (normal and altered)
• Linking of exposures with early and latent effects
• Linking of altered effects at different developmental stages
• Analysis of longitudinal development, repeated measures, and linking exposures and outcomes.

Outcomes

To begin the discussion on domains of parenting, the participants decided to consider the relevant developmental outcomes influenced by parenting. Due to the complexity of the concepts and behaviors described, it was pointed out that there are no simple one-to-one relationships between parenting and specific outcomes or domains of functioning in children. Instead, the influence of parenting tends to be reflected on groups or clusters of related behaviors. These clusters include:

• Gender development
• Social behavior (that is, peer relations, aggression, social cognition, social competence, and empathy)
• Language
• Literacy and numeracy
• Memory
• Cognition
• Neurobehavioral development
• Psychological health
• Self-systems (that is, self-concept, self-esteem, self-efficacy, and self-regulation)
• Moral development
• Cultural competence
• Emotional development
• School readiness.

Understanding the link between parenting and developmental outcomes is complex because parenting interacts with genetic predisposition and gene expression, as well as with other environmental factors such as school, peers, and extended family. Nevertheless, a great deal of knowledge continues to accumulate to help sort out these interactions. For example, an intervention designed to increase effective parenting with 238 newly divorced mothers of sons in grades 1-3 found that the women participating in the program demonstrated less coercive behavior toward their children and that the children later showed better teacher-reported school adjustment than did children whose mothers were in the control group that did not receive the intervention (Forgatch and DeGarmo, 1999).

Parke (2004) suggests that sound parental advice is positively related to social competence, the selection of friends with low levels of antisocial behavior and higher levels of academic achievement. Monitoring, defined as a set of correlated parenting behaviors involving attention to and tracking of the child’s whereabouts, activities, and adaptations (Dishion and McMahon, 1998), appears to be an important factor. Poor monitoring is related to poorer academic skills, lower peer acceptance, and higher rates of delinquency and externalizing behavior (Sandstrom and Coie, 1999), as well as higher rates of drug and alcohol use (Dishion and McMahon, 1998). Warmth and appropriate limit setting by parents results in better internalization of control (self-regulation) and willingness to cooperate (Maccoby and Martin, 1983).

Adequate parental monitoring facilitates good physical health in children unrelated to any medical problems (O’Connor, 2002). The recent increase in childhood obesity has lead to a growing body of research on parenting’s effect on children’s body weight. Although this is a developing area, some findings have emerged. Strauss and Knight (1999) found that children from homes with low cognitive stimulation were at increased risk for obesity. Insecure attachment in the mother, possessiveness, and overprotection of children also appear to be related to overweight and obesity in children (Trombini et al., 2003).

Domains of Parenting

Parenting Theory. Theories of parenting were discussed in detail during the workshop. The participants recommended that any approach used in a longitudinal study should be based on well-documented parenting theory. The group discussed several specific theories but suggested that a heuristic model based on several theories rather than reliance on a specific model or theory would be most appropriate in a longitudinal study. Individual theories of parenting as well as a possible measurement framework are discussed below.

One of the earliest parenting theories was developed by Baumrind (1971), whose approach classified parents into three types: authoritative, authoritarian, and permissive, by crossing the dimensions of responsiveness/warmth and control/demandingness. Later, she added neglectful (or disengaged) as a fourth type of parent (Baumrind, 1989). Authoritative parents are high in both control and responsiveness, whereas authoritarian parents are high in control but low in responsiveness. Permissive parents are low in control but high in responsiveness, and neglectful parents are low on both dimensions. A considerable amount of research has been conducted using this model. As discussed in Mandara (2003), children from authoritative homes generally fair better than other children. They are more competent than other children, use fewer drugs as adolescents, and generally have a happier life than other children. Children from authoritarian homes tend to be more aggressive, less independent, more depressed, and have lower academic achievement. Permissive parents have children who are less achievement-oriented, have lower levels of self-control and self-esteem, and higher levels of aggression toward their parents. The children of neglectful parents are at higher risk for a variety of emotional and achievement problems.

Another theory of parenting proposed by Parke (2004; Parke and O’Neill, 1999) suggests that parents influence their children through multiple pathways. First, parents influence their children directly in their role as teachers, educators, or consultants and through parent-child interactions. Parents indirectly influence their children as managers of the environment, including directing their children’s social lives and regulating opportunities for social and cognitive experiences. Parke’s theory, although useful, was considered too simplistic by some workshop participants. Darling and Steinberg (1993) also posit a theory of direct and indirect parental influences on child adjustment. Adjustment in the child is predicted directly by both parenting style and parenting practices, but parenting style also indirectly influences adjustment through its influence on parenting practices. Parenting style, that is, the emotional climate of the home, includes sensitivity; warmth; intrusiveness; and parental beliefs, attitudes, and responsiveness. Parenting practices include sustenance, support, surveillance, stimulation, and structure, as described by Bradley.

Attachment theory posits that an infant’s experience of responsive and sensitive care early in life leads to a greater sense of self-efficacy and trust in significant others (Bowlby, 1968). This secure attachment leads to children who exhibit more exploratory behavior using their caregiver as a secure base and who are easier to comfort when distressed (Ainsworth et al., 1978). Conversely, infants experiencing inconsistent, unresponsive, or rejecting care later see themselves as unworthy and others as untrustworthy and unpredictable. The behavior of insecure children tends to be angry, withdrawn, explosive, and disorganized. Behavior with peers in later childhood tends to be more aggressive among insecurely attached children (Cohn, 1990).

Parenting Framework. Based on the theories described above, as well as their understanding of parenting research, the workshop participants developed a heuristic framework to help guide the measurement of parenting. The proposed framework includes four dimensions:

• Warmth. One of the two dimensions originally proposed by Baumrind and a key component in attachment theory, warmth includes responsiveness and sensitivity to the child’s needs.
• Control. The second of two dimensions proposed by Baumrind, control includes monitoring, limit setting, and discipline. Features of control are highly dependent on the child’s age.
• Stimulation. Includes the child’s exposure to experiences (cognitive, intellectual, and physical) by the parent and may be a key feature in school readiness as well as social development.
• Meta-parenting. Defined as parenting that occurs when a child is not present and includes such things a parental planning of the child’s activities, discussion between the parents about the child’s well-being, and choice of schools, as well as nonconstructive parental ruminating about the child (G. Holden).

Under the rubric of parenting, specific domains of activity have been identified. Table 1 shows the domains mentioned by the workshop participants within the parenting measurement framework. Several domains were discussed in depth during the workshop, particularly when they presented measurement issues. For example, because parenting behavior related to control varies with the age of the child, the dimension of control must include limit-setting, behavioral control, monitoring, and discipline.

Table 1. Domains of parenting within the measurement framework

Warmth Control Stimulation Meta-parenting Involvement Affection Nurturance Sustenance (feeding, grooming, cleaning) Sensitivity Responsiveness Maltreatment Intrusiveness Consistency Discipline Limit setting Monitoring Fun Cognitive stimulation Physical stimulation Perceptual stimulation School preparation Structuring of environment Knowledge of parenting and childcare Developmental knowledge Knowledge of and use of service networks

Parenting domains outside of the framework. Despite the usefulness of the framework, there are elements of parenting that do not fit neatly into these categories. The workshop experts identified the following domains as important:

• Parental expectations and goals. These constructs were discussed briefly, and although important, the participants felt that there were a number of complications in measuring them. For example, there tends to be racial, ethnic, and gender (of the child) differences in expectations; and they change for each stage of the child’s development.
• Perceived parental impact and self-efficacy. Parental self-efficacy should be combined with parental knowledge or even control because self-efficacy without appropriate knowledge of child development could be a misleading construct.
• Religious, cultural, and gender values. The workshop participants considered that values are important constructs because of their influence on all aspects of parental behavior. They noted, however, that values do not change substantially over time and probably would not need to be measured at multiple points in time.

Temperament. The workshop participants stressed that good parenting varies from child to child—one size does not fit all. Good or rather appropriate parenting is dependent on the child and his or her temperament. Temperament is defined as constitutionally based individual differences in emotional, motor, and attentional reactivity and self-regulation (Rothbart and Bates, 1998). Temperament emerges early and is relatively stable over time but can be modified by environmental factors. Differences in temperament interact with parenting to influence outcomes in children. Associations between parental behavior and temperamental characteristics reflect bidirectional interactive processes as well as genetic linkages (Collins et al., 2000). Difficult or irritable infants can evoke hostility, criticism, avoidance, and more coercive discipline from their parents. However, the effects of temperament in children can also be modified by effective parenting. It has, for instance, been shown that subtle parental socialization approaches are more effective with temperamentally fearful children, whereas close and restrictive control is more effective with bolder, fearless children (Kochanska, 1997).

Parenting not only interacts with temperament, it may influence gene expression. Animal research, which can carefully control for extraneous environmental influences as well as genotype, has demonstrated in multiple rat models that variations in maternal care in infancy can regulate the development of behavioral and neuroendocrine stress reactivity in ways that permanently alter adult reactivity (Caldji et al., 2000). In nonhuman primates such as rhesus monkeys, it has been shown that about 20 percent exhibit unusual behavioral disruption and enhanced physiological arousal when confronted with stimuli that other monkeys find interesting and want to explore. These high-reactive monkeys also tend to be shy and withdrawn in initial encounters with peers (Suomi, 1997). Offspring of high-reactive monkeys exhibit the same traits, indicating that this is an inherited temperament trait. However, if the shy monkeys are cross-fostered to other mothers in infancy, something interesting happens. Although those high-reactive monkeys cross-fostered to control mothers continue to exhibit the same behaviors as those who stay with their own mothers, those who are cross-fostered to nurturant mothers actually become behaviorally precocious (Suomi, 1997), even when compared with normally reactive infants cross-fostered to either type of mother. These are just a few of the examples that illustrate the ability of parenting style to influence gene expression.

Dimensional and Typological Approaches to Parenting. Dimensional approaches to parenting measure specific constructs or variables (for example, warmth or discipline) and their ability to predict outcomes (Mandara, 2003). However, they are limited because their focus is specific while ignoring the complexity in family systems. Despite this, supporters of dimensional approaches find this method preferable when designing and testing interventions (O’Connor, 2002). Typological approaches use categories to organize complex systems according to their similarities and dissimilarities. Individual constructs are first measured using the dimensional approach, but the dimensions are then combined to yield typologies. Typologies are considered to be preferable for describing family systems because they allow for the complexity inherent in these systems. For this reason, typologies were favored over dimensions by workshop participants.

Mandara (2003) classifies typologies into three groups: univariate, bivariate, and multivariate. Univariate typologies divide a sample into two groups based on a theoretical cutoff point on a dimension. As the simplest typology, these are the least useful. Bivariate typologies cross two dimensions to form quadrants. Baumrind’s crossing of the warmth and control dimensions to yield four parenting styles is an early example of a bivariate typology. Similarly, Bartholomew and Horowitz’s (1991) crossing of models of self and others (as positive or negative) yields four attachments styles: secure, preoccupied, fearful, and dismissing. Multivariate typologies attempt to classify objects using more than two dimensions by building a data matrix, examining the clusters of scores on all dimensions and then classifying clustered cases into types.

Parenting Measures Feasible in a Large Cohort

Using the measurement framework described above, the workshop participants recommended a set of instruments that would adequately assess each dimension in the framework. It was stressed that instruments used in longitudinal epidemiologic research should be reliable and well validated. It is especially important that they be appropriate for a broad range of socioeconomic status groups and that they be validated across cultures. Table 2 lists the instruments under the four domains listed above. Each of the instruments listed below is reviewed more fully in Attachment 1.

After reviewing the list shown in Table 2, the workshop participants recognized that there was some overlap in the measures and that some measures might not perform well in longitudinal cohort studies. For example:

• The HOME was considered by several participants to be too geared toward middle-class families. It was stated that low-income families would perform poorly on the HOME because of their lack of resources, rather than poor parenting. Some workshop participants were concerned about its cultural appropriateness. If a subscale of the HOME is used (for example, the Zaslow Modified HOME-SF), it should be reviewed for cultural appropriateness and perhaps supplemented with culturally sensitive items.
• The Emotional Availability Score is difficult to score and requires intensive interviewer training. There was concern about its appropriateness for low-income families.
• The Positive Interaction Subscale from the Canadian National Longitudinal Survey of Children and Youth is susceptible to ceiling effects.

Table 2. Parenting measures mapped onto the suggested measurement framework*

Warmth Control Stimulation Meta-parenting Home (Zaslow) Supportive Presence Index Positive Parenting Scale Canadian Scale Maternal Attachment Q-sort ECLSB Subscale CRPBI Subscale Structured Observations HOME (Zaslow) Emotional Availability Score Perceptions of Parenting Scale Canadian Scale CRPBI Subscale ECLSB Subscale Conflict Tactics Severe Item Subscale SRCAP HOME (Zaslow) ECLSB Subscale Structured Observations HOME (Zaslow) Meta-parenting Scale-SF

*Italicized scales were considered preferable by participants.

As a result of these discussions and consideration of the amount of time needed to administer all of the measures, the workshop participants pared down the list of suitable measures for warmth and control. Preferred scales are italicized in Table 2. To measure warmth, the Maternal Q Sort, a subscale from the Early Childhood Longitudinal Study (ECLSB), and a subscale from the Child Report of Parental Behavior Inventory (CRPBI) were considered sufficient, although scoring changes to the Maternal Q Sort are suggested (see Attachment 1). To measure control, subscales from the ECLSB, the CRPBI, and the lax discipline scale from the Canadian National Longitudinal Survey were recommended.

Two areas prompted particular discussion among the group. The validity of parental knowledge as a construct was questioned because it is highly correlated with parental education. The Concepts of Development Scale (Sameroff, 1994) and the Knowledge of Infant Development Index (KIDI; MacPhee, 1984) both measure parents’ knowledge of child development, but both are problematic. Scores from the Concepts of Development Scale can almost entirely be explained by the parent’s level of education, and the KIDI demonstrates poor psychometric properties and poorly defined subscales. In addition, the predictive validity of parenting knowledge is limited (Yarrow et al., 1968), and the experts noted that if parental knowledge is measured it should be combined with parental self-efficacy. Child maltreatment was another area promoting a great deal of discussion. The group suggested that attention be paid to sensitive measures of maltreatment and that children’s responses to maltreatment questions be sought as soon as feasible. A separate measure of child maltreatment is needed, and two meeting participants agreed to discuss this further and make a recommendation. They later agreed to the six severe violence items on the Conflict Tactics Scale.

Other Considerations

During the workshop, the participants discussed several additional issues of importance related to the measurement of parenting.

Measurement Issues. There were a number of suggestions regarding the measurement of parenting in epidemiologic studies. There was some discussion about how few domains in parenting are independently predictive because they complement each other. The instruments should be reviewed to avoid redundancy, but several instruments were proposed under each category as being complementary to each other. The participants considered minimally overlapping measures as the goal.

Related to this is the need to use multiple measures if possible. Even though there is often only modest agreement among different methods (O’Connor, 2002), there is sometimes a discrepancy between what parents say and what they do. An example of multiple measures would be including some combination of parental reports, teachers’ reports, and children’s reports of behavior. Interviewing children as soon as they are able to answer the questions is important, particularly in regard to maltreatment.

Structured and unstructured observational methods were discussed and encouraged where possible. Unstructured observations would have the interviewer observe natural interaction between parents and children and then complete a simple observation form during or immediately following the interview. One of the issues with this type of observation would be that it has to be long enough to accurately capture the parent-child relationship. The Maternal Q Sort attachment measure relies on this type of observation with the interviewer completing a "like mother-not like mother" observation form. Unstructured observations are useful with infants.

Structured observations would include observation of standardized events such as structured play, unstructured play, a teaching task, and a clean-up or compliance task. Structured observations would be best used with toddlers. The tasks usually last about 3-5 minutes, and there was discussion about whether the length was too short (particularly for free play) and would not provide enough opportunity to observe the behavior. However, those arguing for the short length believe that this challenges the relationship and will result in the emergence of more natural behaviors. A major problem with videotaping is the high cost of coding the tapes. However, participants raised the possibility of banking the tapes and coding them at a later date through a variety of funding mechanisms. They felt that the main issue was not to miss the opportunity of capturing the mother-child interaction at an early time in development. Although the participants generally liked observational tasks, there were several concerns about their use. For example, the time of day is an important consideration in the parent-child interaction, particularly if the child is tired or hungry. Structured observations are probably better done in the home rather than in a clinic because clinics are strange situations for the child. In addition, there needs to be consideration of when during the study protocol the observations occur. They might serve as a good transition or break from the standard interview protocol.

Measures of Parenting in Low-Income Families. There was strong concern in the group regarding the use of standard parenting measures in low-income families. Poor families may be defined as deficient on parenting dimensions because of a lack of resources. Examples used were neglect items on scales such as "having a quiet place to study" that are not appropriate for poor families. The HOME, in particular, was identified as a measure geared toward middle-income families.

Deficit Paradigm. There was discussion about the usefulness of standard measures of parenting in measuring good parenting. The usual paradigm in measurement looks at deficits in parenting or negative outcomes in children (for example, aggression, school unreadiness), and it is unclear how well these measures will perform in measuring good parenting. Several standard measures seem to have a ceiling effect because they cannot adequately measure supportive homes. Related to this point is the measurement of optimal parenting—the usual paradigm does not include the upper end of parenting behavior. The Positive Parenting Scale was cited as an example of this where 60 percent of a sample where this scale was used had a score of 10 on a 1-to-10 scale.

Contribution of Fathers. How to best include fathers in large cohort studies was discussed. Past studies of parenting have usually included only the mothers (or primary caregivers), and some participants recommended that the contribution of fathers be considered and measured. Fathers typically are less important in the earlier stages of development, but their importance increases as the child gets older. Fathers are frequently the entrée into the larger world outside of the family. One example cited of the unique role of fathers was in African-American families where with boys "moms protect and dads push," suggesting that measures of control and warmth for fathers would be quite different than for mothers.

One area that has received particular attention is a father’s role in the development of eating disorders in daughters. Low perceived mutuality (defined as active participation in the relationship and responsiveness to each others’ thoughts, feelings, and needs) with fathers is associated with bulimic and depressive symptoms in girls, whereas low perceived mutuality with mothers is associated only with depressive symptoms (Tantillo and Sanftner, 2003). Using discriminant analysis of parents’ combined character and temperament scores on the Temperament and Character Inventory (TCI), Fassino et al. (2003) found that fathers of bulimic girls were more different from fathers of nonbulimic girls (canonical correlation = .66; correct classification rate of 78 percent) and that mothers of bulimic girls were different from mothers of nonbulimic girls (canonical correlation = .39; correct classification rate of 60 percent). The TCI is a personality inventory that measures four temperament dimensions (novelty seeking, harm avoidance, reward dependence, and persistence) and three character dimensions (self-directedness, cooperativeness, and self-transcendence).

Several issues arose when considering the role of fathers in longitudinal epidemiologic research. One participant described an experience where mothers in a study had to give permission for fathers not resident in the household to take part in the study; many of the mothers choose not to have the nonresident fathers included. Other issues discussed included the role of biological fathers and/or functional fathers. Biological fathers contribute genetic information that is important to know, whereas functional fathers—who are not always biological parents—are actively parenting the child. Finally, there is the consideration of changing family situations across the life of longitudinal studies and how long to include fathers after they may have left the family or the family home. After some discussion, the participants recommended focusing on the primary caregiver (typically mothers) for most parenting measures but including a shorter examination (for example, 20-minute self-report) of the secondary parent later on in the child’s life.

References

Ainsworth, M.D., Bell, S.M., and Stayton, D.J. (1971). Individual differences in strange-situation behavior of one-year-olds. In: H.R. Schaffer (Ed.), The Origins of Human Social Relations (pp. 17-57). London and New York: Academic Press.

Ainsworth, M.D., Blehar, M.C., Waters, E., and Wall, S. (1978). Patterns of Attachment: A Psychological Study of the Strange Situation. Hillsdale, NJ: Erlbaum.

Bartholomew, K., and Horowitz, L.M. (1991). Attachment styles among young adults: A test of a four-category model. Journal of Personality and Social Psychology, 61, 236-244.

Baumrind, D. (1971). Current patterns of parental authority. Developmental Psychology Monographs, 4(1, pt. 2).

Baumrind, D. (1989). The permanence of change and the impermanence of stability. Human Development, 32, 187-195.

Bowlby, J. (1968). Attachment and Loss. Vol. 1. Attachment. New York: Basic Books.

Caldji, C., Diorio, J., and Meaney, M.J. (2000) Variation in maternal care in infancy regulate the development of stress reactivity. Biological Psychiatry, 48, 1164-1174.

Campbell, S.B. (1995). Behavior problems in preschool children: A review of recent research. Journal of Child Psychology, 36(1), 113-149.

Cohn, D.A. (1990). Child-mother attachment of six-year-olds and social competence at school. Child Development, 61, 152-162.

Collins, W.A., Maccoby, E.E., Steinberg, L., Hetherington, E.M., and Bornstein, M.H. (2000). Contemporary research on parenting: The case for nature and nurture. American Psychologist, 55(2), 218-232.

Darling, N., and Steinberg, L. (1993). Parenting style as context: An integrative model. Psychological Bulletin, 113, 487-496.

Dishion, T.J., and McMahon, R.J. (1998). Parental monitoring and the prevention of child and adolescent problem behavior: A conceptual and empirical foundation. Clinical Child and Family Psychology Review, 1(1), 61-75.

Fassino, S., Amianto, G., Abbate, D. P., Leombruni, L., Garzaro, M.L., and Rovera, G.G. (2003). Bulimic family dynamics: Role of parents’ personality—a controlled study with the Temperament and Character Inventory. Comprehensive Psychiatry, 44(1), 70-77.

Forgatch, M.S., and DeGarmo, D.S. (1999). Parenting through change: An effective prevention program for single mothers. Journal of Consulting and Clinical Psychology, 67, 711-724.

Holden, G.W., and Hawk, C.K. (2003, April). Meta-parenting as a key cognitive ingredient of successful parenting. In J. Valente (Chair), Cognitive perspectives on effective parenting. Symposium presented at the biennial meeting of the Society for Research in Child Development, Tampa, FL.

Holden, G.W., and Hawk, C.K. (2003). Meta-parenting in the journey of child rearing: A mechanism for change. In: L. Kuczynski (Ed.), Handbook of dynamics in parent-child relations (pp. 189-210). Thousand Oaks, CA: Sage.

Kochanska, G. (1997). Multiple pathways to conscience for children with different temperaments: From toddler to age 5. Developmental Psychology, 33, 228-240.

Maccoby, E.E., and Martin, J.A. (1983). Socialization in the context of the family: Parent-child interaction. In: E.M. Hetherington (Ed.), Mussen Manual of Child Psychology (Vol. 4, pp. 1-102). New York: Wiley.

MacPhee, D. (1984). The pediatrician as a source of knowledge about child development. Journal of Pediatric Psychology, 9, 87-100.

Mandara, J. (2003). The typological approach in child and family psychology: A review of theory, methods and research. Clinical Child and Family Psychology Review, 6(2), 129-146.

O’Connor, T.G. (2002). Annotation: The ’effects’ of parenting reconsidered: Findings, challenges and applications. Journal of Child Psychology and Psychiatry, 43(5), 555-572.

Parke, R.D., and O’Neill, R. (1999). Social relationships across contexts: Family-peer linkages. In: W.A. Collins and B. Laursen (Eds.) Minnesota Symposium on Child Psychology (Vol. 30, 211-239). Hillsdale, NJ: Erlbaum.

Parke, R.D. (2004). Development in the family. Annual Review of Psychology, 55, 365-399.

Rothbart, M.K., and Bates, J.E. (1998). Temperament. In: W. Damon and N. Eisenberg (Eds.) Handbook of Child Psychology, Vol. 3. Social, Emotional and Personality Development (Vol. 3, pp 105-176). New York: Wiley.

Sameroff, A.J. (1994). Developmental systems and family functioning. In: R.D. Parke and S.G. Kellam (Eds.) Exploring Family Relationships with Other Social Contexts (pp. 199-214). Hillsdale, NJ: Erlbaum.

Sandstrom, M.J., and Coie, H.D. (1999). A developmental perspective on peer rejection: Mechanisms of stability and change. Child Development, 70, 955-966.

Strauss, R.S., and Knight, J. (1999). Influence of the home environment on the development of obesity in children. Pediatrics, 103(6), 1-8.

Suomi, S.J. (1977). Early determinants of behavior: evidence from primate studies. British Medical Bulletin, 53(1), 170-184.

Tantillo, M., and Sanftner, J. (2003). The relationship between perceived mutuality and bulimic symptoms, depression and therapeutic change in groups. Eating Behavior, Jan: 3(4), 349-364.

Trombini, E., Baldaro, B., Bertaccini, R., Mattei, C., Montebarocci, O., and Rossi, N. (2003). Maternal attitudes and attachment styles in mothers of obese children. Perceptual and Motor Skills, 97(2), 613-620.

Waters, E., Merrick, S., Treboux, D., Crowell, J., and Albersheim, L. (2000). Attachment security in infancy and early adulthood: A twenty-year longitudinal study. Child Development, 71, 684-702.

Yarrow, M., Campbell, J., and Burton, R. (1968). Child Rearing: An Inquiry into Research and Methods. San Francisco: Jossey-Bass.

Participants

Christine A. Bachrach, Ph.D., NICHD, NIH, DHHS
John E. Bates, Ph.D., Indiana University
Harriet Becher, University of Maryland
Michel Boivin, Ph.D., Université Laval, Quebec, Canada
Marc H. Bornstein, Ph.D., NICHD, NIH, DHHS
Natasha Cabrera, Ph.D., University of Maryland
Alan R. Fleischman, M.D., NICHD, NIH, DHHS
Neal Halfon, M.D., M.P.H., University of California, Los Angeles
George W. Holden, Ph.D., University of Texas
Allison Holmes, University of Maryland
Sarah S. Knox, Ph.D., NICHD, NIH, DHHS
Dafna Kohen, Ph.D., University of Ottawa, Ottawa, Canada
Jelani Mandara, Ph.D., Northwestern University
Ellen L. Marks, RTI International
Susan M. McHale, Ph.D., Pennsylvania State University
Judith Smetana, Ph.D., University of Rochester
Douglas M. Teti, Ph.D., Pennsylvania State University
Cathy Spatz Widom, Ph.D., University of Medicine and Dentistry of New Jersey

Attachment 1: Detailed Description of Parenting Measures

Measure 1:		Mother-Child Relations for Middle Childhood (Zaslow Modified HOME-SF)
Age Appropriateness:		Consensus at the Study October parenting workshop was that the SF was intended for middle childhood but could be extended younger.
Content:		The Mother-Child Relations for Middle Childhood is based on the HOME-SF. Seven items from the HOME-SF remain in the Mother-Child Relations measure. However, the scoring of these seven items has been modified. Maternal-Child Relations consists of nine measures, four of which reflect the dyadic characteristics of the mother child relationship (expression of affection, reasoning discipline, eliciting, and joint activities) and five that reflect the managerial role of the mother in guiding and monitoring the child’s environment and activities (provision of stimulating materials, mother-school interface, expectations of responsibilities, monitoring, and socialization regarding racial/ethnic issues).
Length:		52 items
Psychometrics:		The nine scales had Cronbach’s alphas ranging from 0.56 to 0.75. Across ethnic groups (African American, Mexican American, European American) there were relatively consistent internal consistencies for each of the scales. The Eliciting scale showed a marginal alpha particularly among Mexican Americans. Mother-School Interface showed a lower alpha among African Americans than among European Americans. Results give preliminary evidence of predictive validity. In general, the scales show similar patterns in their prediction of child well-being across racial/ethnic groups. There exists some differences in mean scale scores across groups.
Method:		Maternal report and observation.
Suggestions for Changes in Measures:
Dimension(s):		Warmth, Control, Stimulation, Meta-Parenting
Comments:		Further investigation of internal validity is important especially within minority populations. More information regarding the validity of the Provision of Stimulating Materials among Mexican Americans is needed (authors’ comment). Workshop participants expressed concern about the scale’s cultural appropriateness and that it may not perform well in low-income situations.
Measure 2:		Supportive Presence Index (Sroufe)
Age Appropriateness:		Young children aged 24 months and older
Content:		A scale to assess the degree to which the parent provides emotional, physical, and affective support to child.
Length:		Varies depending on use
Psychometrics:
Limitations:		Requires interviewer training specific to the content of the instrument.
Advantages:
Method:		Observation of parent and child, usually (but not always) during structured tasks. Parental behaviors are rated on 1 (not supportive) to 7 (very supportive) scale.
Suggestions for Changes in Measures:
Dimension(s):		Warmth
Comments:		Instrument is context specific. Coding is challenging.
Measure 3:		Positive Parenting Scale
Age Appropriateness:		Middle childhood with subset acceptable for very young children
Content:		Measures warmth, affection and appropriate rewards.
Length:		Six items (three items for use with very young children)
Psychometrics:		Good internal consistency for both child (alpha = 0.85) and parent reports (alpha = 0.84).
Method:		Parent-child self-report
Suggestions for Changes in Measures:
Dimension(s):		Warmth
Comments:		Hard to get a good range of responses and may have ceiling effective for supportive parenting. On a 1-to-10 response scale, 60 percent of respondents had score of 10.
Measure 4:		Canadian Parenting Scale (D. Kohen)
Age Appropriateness:		0-11 years
Content:		The objective of the parenting scale is to measure certain parental behaviors. There are two scales. One scale is designed to measure positive interaction, hostile/ineffective parenting and consistent parenting. A second scale is designed to measure aversive and non-aversive parental management techniques.
Length:		30 items
Psychometrics:		Subscale: Cronbach alphas range from 0.56 to 0.75. The scale has been significantly associated with measures of parents’ psychological and social health, concurrent measures of parents’ behavior with their children, and child behavior variables, especially those related to oppositionalism and attention deficit.
Method:		Questionnaire for parents
Suggestions for Changes in Measures:
Dimension(s):		Warmth, Control
Comments:		Problem with response bias. The positive interaction subscale is susceptible to ceiling effects.
Measure 5:		Maternal Q-Sort (Modified)
Age Appropriateness:		Infant through toddler
Content:		Items describe a wide range of maternal behaviors including interactive style, sensitivity to the infant’s state, feeding interactions, and extent to which the home reflects the infant’s needs.
Length:		Full scale is 90 items. Observation is 2 hours or more. Could shorten scale by selecting items with two top- and two bottom-scored items for security on each subscale.
Psychometrics:		Reliable and correlates strongly with security of infant-parent attachment.
Limitations:		Observers must be familiar with attachment relationship constructs as described by Ainsworth et al. (1971) in the home and Ainsworth et al. (1978) in the strange situation.
Method:		Q sorting system in which an observer watches the parent with the infant in a naturalistic setting and then sorts a set of descriptors about the parent according to how like and how unlike the descriptor was of the parent’s behavior during the observation. Categories range from very much like the parent to very much unlike the parent.
Suggestions for Changes in Measures:		Interviewer could do incidental observations while in the home. Put items on single sheet of paper with ratings that can simply be checked during or immediately following the interview, rather than use card sorting method. Use short version as described above.
Dimension(s):		Warmth
Comments:		Need to train interviewers and test inter-rater reliability before attempting. Need a large enough window of behavior to observe interactions.
Measure 6:		Early Childhood Longitudinal Study (ECLS-B) Subscale
Age Appropriateness:		Assessments take place at 9 months, 2 years, 4 years (preschool), kindergarten, first grade
Content:		Comprehensive parenting measures with addition of HOME. The control items of interest in the ECLS-B assess areas such as parent authoritarianism (9 months), parenting style (24 months), enforcement of rules (24 months), frequency that parent spanked child in past week (24 months), frequency that parent used time out in past week (24 months), and the use of different modes of discipline (24 months).
Length:		14 items
Psychometrics:
Method:		The ECLS-B assesses children and interviews primary caregivers (usually the mother) in the children’s homes. Fathers complete a self-administered questionnaire. The in-home visit includes a videotaping of parent-child interaction.
Suggestions for Changes in Measures:
Dimension(s):		Warmth, Control, Stimulation, Meta-Parenting, Parental values, expectations and goals (achievement oriented items)
Comments:		Appears to be well-designed and was recommended during workshop
Measure 7:		Child Report of Parental Behavior Inventory (CRPBI)
Age Appropriateness:		Middle childhood through adolescence
Content:		Three subscales: Acceptance (24 items), Control through guilt (16 items), Lax Discipline (16 items). Assesses children’s perceptions with respect to parental acceptance, permitted psychological autonomy, and level of parental control. Nearly identical forms are available for indicating behaviors of mothers and fathers.
Length:		192 items with 18 factors, 3 dimensions. There are other versions with fewer items. 3-point response option.
Psychometrics:		Internal consistency of the 18 individual scales is adequate (mean alpha = 0.71). Low agreement among family members on each scale behavior is concern,
Method:		Both children and parents complete
Suggestions for Changes in Measures:		Recommend using short form and a 5-point response scale. Possible to use some items with very young children with adaptation of response scale
Dimension(s):		Warmth, Control (subscale)
Comments:
Measure 8:		Emotional Availability Score (EAS)
Age Appropriateness:		There are different versions of the EAS for infancy/early childhood and for middle childhood
Content:		Measures parental emotional availability on four dimensions: sensitivity (responsiveness to child in terms of timing, appropriateness, flexibility, quality of affect, ability to resolve conflict), structuring (quality of scaffolding of child behavior, particularly during play, and limit setting), non-intrusiveness (parent’s availability to the infant without being intrusive or overprotective), and nonhostility. Also assesses child behavior along dimensions of responsiveness to and involvement with the parent
Length:
Psychometrics:		Little availability on the use and validity of these scales for parents of preschoolers.
Limitations:		Dimensions of sensitivity and structuring /intrusiveness are more optimal among mothers of securely attached 12-month-old infants, relative to mothers of insecure infants. EAS explicitly takes into consideration the behavior of the child. Parental emotional availability is thus construed as a product of mutual parent-child behavioral and affective regulation. This point is important because of the prior assumptions that maternal sensitivity is wholly characteristic of the mother. Also must take into account child’s developmental level
Method:		Observation
Suggestions for Changes in Measures:
Dimension(s):		Control
Comments:		Difficult to learn scoring and would require interviewer training. Some concern about appropriateness for low-income families
Measure 9		Perceptions of Parenting Scale (POPA)
Age Appropriateness:		Older children because of emphasis on discipline
Content:		Attitudinal measures that focus on concrete assessment of reactions to parental disciplinary acts. Assesses the perceptions of the harshness, appropriateness and effectiveness of parental behaviors. Parenting practices evaluated include verbally abusive, physically punitive, and positive care-giving acts.
Length:		61 items
Psychometrics:		Some validity for verbal physical abuse. Has demonstrated adequate test-retest reliability and internal consistency in past research with college students and mothers of twins.
Method:		Paper-and-pencil survey of parent
Suggestions for Changes in Measures:
Dimension(s):		Control
Comments:
Measure 10:		Conflict Tactics Scale (Severe items)
Age Appropriateness:		10-18 years
Content:		Measures aggression and violence. The full scale assesses physical abuse in childhood. Subscales: Non-Violent Discipline, Psychological Aggression, Physical Assault, Neglect. Likert scale. There is also a revised conflict tactics scale which incorporates more items to increase the reliability and validity of the measure.
Length:		22 items, 10-15 minutes
Psychometrics:		Info on criterion and construct validity is adequate but limited. Range of internal consistency: 0.22-0.70. Criterion validity acceptable.
Method:		Administer group and individual. Parent is assessed. There are also child report and sibling versions.
Suggestions for Changes in Measures:		Use only severe violence questions (VSV, 6 items).
Dimension(s):		Control
Comments:		It was suggested that an adapted measure of Physical Abuse from the Conflict Tactics Scale be used. The five items are: (1) Hit him/her on the bottom with something like a belt, hairbrush, a stick or some other hard object; (2) Slapped him/her on the hand, arm, leg; (3) Shook him/her; (4) Slapped him/her on the face or head or ears; (5) Hit him/her on some other part of the body besides the bottom with something like a belt, hairbrush, a stick, or some other hard object. Parents answer how many times they have done these things in the past 6 months.
Measure 11:		Self-Report Childhood Abuse Physical (SRCAP)
Age Appropriateness:		Developed for adults but can be used with older children.
Content:		Retrospectively assesses childhood physical abuse.
Length:		Six items
Psychometrics:		Good discriminant validity. Construct validity has not been thoroughly demonstrated and requires more investigation. Internal consistency = .75.
Limitations:		Issues with social desirability. Consider validating with at Child Protective Service reports. Ask child as soon as possible.
Advantages:
Method:		Structured interview. Dichotomous "yes/no" response.
Suggestions for Changes in Measures:
Dimension(s):		Control
Comments:		Good metric, short. Could be used as part of multiple assessment of abuse at different points in time—people more likely to admit at second assessment.
Measure 12:		Meta-Parenting Profile Questionnaire (MMPQ)
Age Appropriateness:		6 months-12 years
Content:		Designed to measure meta-parenting thoughts. Meta-parenting is defined as thoughts occurring outside ongoing interactions with children (Holden and Hawk, 2003a, 2003b). These thoughts are grouped into four areas (and subscales on the MPPQ): Assessing, Anticipating, Problem solving, and Reflecting. Respondents to the questionnaire are instructed to answer questions with a particular child in mind. Most of items address "how often" or "to what extent" certain thoughts occur.
Length:		21 items
Psychometrics:		Cronbach alphas for subscales include: Assessing = 0.71, Anticipating = 0.67; Problem solving = 0.63, and Reflecting = 0.67. Test-retest reliability across a 30-day span indicated moderate levels of reliability: Assessing ( r[52] = .72, p < .01), Anticipating ( r[52] = .61, p < .01), Problem solving ( r[52] = .59, p < .01), and Reflecting ( r[52] = .71, p < .01). The questionnaires is currently undergoing further refinement and testing. No cognitive testing to date.
Method:		Each parent completes. Self-report instrument. Responses occur on 5-point Likert scales, involving reports of frequency or extent of the thinking.
Suggestions for Changes in Measures:		Could develop a short form for use in the Study
Dimension(s):		Meta-Parenting
Comments:		Relatively new scale. Unclear about its appropriateness across populations.
Measure 13:		Parental Impact Scale
Age Appropriateness:
Content:		The parent’s evaluation of the effect of his/her behavior on the developing child.
Length:		Five items
Psychometrics:		Good internal consistency, alpha > 0.75
Method:		Questionnaire for parents
Suggestions for Changes in Measures:
Dimension(s):		Perceived parental impact and self efficacy
Comments:		May have ceiling effect